writing

profileTAZA
epdf.tips_managerial-economics-applications-strategy-and-tac1.pdf

Managerial Economics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Australia • Brazil • Japan • Korea • Mexico • Singapore • Spain • United Kingdom • United StatesAustralia • Brazil • Japan • Korea • Mexico • Singapore • Spain • United Kingdom • United States

Managerial Economics Applications, Strategy, and Tactics

TWELFTH EDITION

JAMES R . MCGU IGAN JRM Investments

R . CHARL ES MOYER University of Louisville

F R EDER I CK H . d e B . HARR I S Schools of Business

Wake Forest University

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

52609_00_fm_pi-pxxvi.indd ii52609_00_fm_pi-pxxvi.indd ii 2/1/10 11:37:43 PM2/1/10 11:37:43 PM

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

This is an electronic version of the print textbook. Due to electronic rights

restrictions, some third party may be suppressed. Edition review has deemed that any suppres ed content does not materially

affect the over all learning experience. The publisher reserves the right to remove the contents from this title at any time if subsequent rights restrictions require it. For valuable information on pricing, previous editions, changes to current editions, and alternate format, please visit www.cengage.com/highered to search by ISBN#, author, title, or keyword for materials in your areas of interest.

s

Managerial Economics: Applications, Strategy, and Tactics, 12th Edition

James R. McGuigan, R. Charles Moyer, Frederick H. deB. Harris

Vice President of Editorial, Business: Jack W. Calhoun

Publisher: Joe Sabatino

Sr. Acquisitions Editor: Steven Scoble

Sr. Developmental Editor: Jennifer Thomas

Marketing Manager: Betty Jung

Assoc. Content Project Manager: Jana Lewis

Manager of Technology, Editorial: Matthew McKinney

Media Editor: Deepak Kumar

Sr. Frontlist Buyer, Manufacturing: Sandra Milewski

Production Service: KnowledgeWorks Global Limited

Compositor: KnowledgeWorks Global Limited

Sr. Art Director: Michelle Kunkler

Production Technology Analyst: Starratt Scheetz

Internal Designer: Juli Cook/Plan-It- Publishing

Cover Designer: Rokusek Design

Cover Image: © Media Bakery

Text Permissions Manager: Mardell Glinski-Schultz

Photography Permissions Manager: Deanna Ettinger

Photo Researcher: Scott Rosen, Bill Smith Group

© 2011, 2008 South-Western, Cengage Learning

ALL RIGHTS RESERVED. No part of this work covered by the copyright herein may be reproduced, transmitted, stored, or used in any form or by any means graphic, electronic, or mechanical, including but not limited to photocopying, recording, scanning, digitizing, taping, web distribution, information networks, or information storage and retrieval systems, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without the prior written permission of the publisher.

For product information and technology assistance, contact us at Cengage Learning Customer & Sales Support, 1-800-354-9706

For permission to use material from this text or product, submit all requests online at www.cengage.com/permissions

Further permissions questions can be emailed to [email protected]

ExamView® is a registered trademark of eInstruction Corp. Windows is a registered trademark of the Microsoft Corporation used herein under license. Macintosh and Power Macintosh are registered trademarks of Apple Computer, Inc. used herein under license.

© 2008 Cengage Learning. All Rights Reserved.

Library of Congress Control Number: 2010929867 Package ISBN-13: 9781439079232 Package ISBN-10: 1-4390-7923-4 Student Edition ISBN 13: 978-1-4390-7939-3 Student Edition ISBN 10: 1-4390-7939-0

South-Western Cengage Learning Cengage Learning 5191 Natorp Boulevard Mason, OH 45040 USA

Cengage Learning products are represented in Canada by Nelson Education, Ltd.

For your course and learning solutions, visit www.cengage.com Purchase any of our products at your local college store or at our preferred online store www.CengageBrain.com

Printed in the United States of America 1 2 3 4 5 6 7 14 13 12 11 10

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

To my family J.R.M.

To Sally, Laura, and Craig R.C.M.

To my family, Roger Sherman, and Ken Elzinga F.H.B.H.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Brief TABLE OF CONTENTS

Preface, xvii About the Authors, xxi

PART I

INTRODUCTION 1

1 Introduction and Goals of the Firm 2

2 Fundamental Economic Concepts 26

PART II

DEMAND AND FORECASTING 61

3 Demand Analysis 62

4 Estimating Demand 95

4A Problems in Applying the Linear Regression Model 126

5 Business and Economic Forecasting 137

6 Managing in the Global Economy 175

6A Foreign Exchange Risk Management 227

PART III

PRODUCTION AND COST 229

7 Production Economics 230

7A Maximization of Production Output Subject to a Cost Constraint 265

7B Production Economics of Renewable and Exhaustible Natural Resources 267

8 Cost Analysis 275

8A Long-Run Costs with a Cobb-Douglas Production Function 301

9 Applications of Cost Theory 305

PART IV

PRICING AND OUTPUT DECISIONS: STRATEGY AND TACTICS 333

10 Prices, Output, and Strategy: Pure and Monopolistic Competition 334

11 Price and Output Determination: Monopoly and Dominant Firms 382

12 Price and Output Determination: Oligopoly 409

13 Best-Practice Tactics: Game Theory 444

13A Entry Deterrence and Accommodation Games 488

14 Pricing Techniques and Analysis 499

PART V

ORGANIZATIONAL ARCHITECTURE AND REGULATION 545

15 Contracting, Governance, and Organizational Form 546

15A Auction Design and Information Economics 580

16 Government Regulation 610

17 Long-Term Investment Analysis 644

APPENDICES

A The Time Value of Money A-1

B Tables B-1

C Differential Calculus Techniques in Management C-1

D Check Answers to Selected End-of-Chapter Exercises D-1

Glossary G-1

Index I-1

Notes

WEB APPENDICES

A Consumer Choice Using Indifference Curve Analysis

B International Parity Conditions

C Linear-Programming Applications

D Capacity Planning and Pricing Against a Low-Cost Competitor: A Case Study of Piedmont Airlines and People Express

E Pricing of Joint Products and Transfer Pricing

F Decisions Under Risk and Uncertainty

v i i

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Contents

Preface, xvii About the Authors, xxi

PART I

INTRODUCTION 1

1 Introduction and Goals of the Firm 2

Chapter Preview 2 Managerial Challenge: How to Achieve

Sustainability: Southern Company 2 What is Managerial Economics? 4 The Decision-Making Model 5

The Responsibilities of Management 5 The Role of Profits 6

Risk-Bearing Theory of Profit 7 Temporary Disequilibrium Theory of Profit 7 Monopoly Theory of Profit 7 Innovation Theory of Profit 7 Managerial Efficiency Theory of Profit 7

Objective of the Firm 8 The Shareholder Wealth-Maximization

Model of the Firm 8 Separation of Ownership and Control: The

Principal-Agent Problem 9 Divergent Objectives and Agency Conflict 10 Agency Problems 11

What Went Right/What Went Wrong: Saturn Corporation 13 Implications of Shareholder Wealth

Maximization 13

What Went Right/What Went Wrong: Eli Lilly Depressed by Loss of Prozac Patent 14

Caveats to Maximizing Shareholder Value 16 Residual Claimants 17 Goals in the Public Sector and Not-for-Profit

Enterprises 18 Not-for-Profit Objectives 18 The Efficiency Objective in Not-for-Profit

Organizations 19 Summary 19 Exercises 20

Case Exercise: Designing a Managerial Incentives Contract 21

Case Exercise: Shareholder Value of Wind Power at Hydro Co.: RE < C 23

2 Fundamental Economic Concepts 26

Chapter Preview 26 Managerial Challenge: Why Charge

$25 per Bag on Airline Flights? 26 Demand and Supply: A Review 27

The Diamond-Water Paradox and the Marginal Revolution 30

Marginal Utility and Incremental Cost Simultaneously Determine Equilibrium Market Price 30

Individual and Market Demand Curves 31 The Demand Function 32 Import-Export Traded Goods 34 Individual and Market Supply Curves 35 Equilibrium Market Price of Gasoline 36

Marginal Analysis 41 Total, Marginal, and Average Relationships 41

The Net Present Value Concept 45 Determining the Net Present Value of an

Investment 46 Sources of Positive Net Present Value

Projects 48 Risk and the NPV Rule 48

Meaning and Measurement of Risk 49 Probability Distributions 49 Expected Values 50 Standard Deviation: An Absolute Measure

of Risk 51 Normal Probability Distribution 51 Coefficient of Variation: A Relative Measure

of Risk 53

What Went Right/What Went Wrong: Long-Term Capital Management (LTCM) 53 Risk and Required Return 54 Summary 56 Exercises 56 Case Exercise: Revenue Management at

American Airlines 58

v i i i

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PART II

DEMAND AND FORECASTING 61

3 Demand Analysis 62

Chapter Preview 62 Managerial Challenge: Health Care

Reform and Cigarette Taxes 62 Demand Relationships 64

The Demand Schedule Defined 64 Constrained Utility Maximization and

Consumer Behavior 65

What Went Right/What Went Wrong: Chevy Volt 69 The Price Elasticity of Demand 69

Price Elasticity Defined 70 Arc Price Elasticity 72 Point Price Elasticity 73 Interpreting the Price Elasticity:

The Relationship between the Price Elasticity and Revenues 73

The Importance of Elasticity-Revenue Relationships 78

Factors Affecting the Price Elasticity of Demand 80

International Perspectives: Free Trade and the Price Elasticity of Demand: Nestlé Yogurt 82

The Income Elasticity of Demand 83 Income Elasticity Defined 83 Arc Income Elasticity 84 Point Income Elasticity 85

Cross Elasticity of Demand 87 Cross Price Elasticity Defined 87 Interpreting the Cross Price Elasticity 87 Antitrust and Cross Price Elasticities 87 An Empirical Illustration of Price, Income,

and Cross Elasticities 89 The Combined Effect of Demand

Elasticities 89 Summary 90 Exercises 91 Case Exercise: Polo Golf Shirt Pricing 93

4 Estimating Demand 95

Chapter Preview 95 Managerial Challenge: Global Warming

and the Demand for Public Transportation 95

Estimating Demand Using Marketing Research Techniques 98

Consumer Surveys 98 Consumer Focus Groups 98 Market Experiments in Test Stores 99

Statistical Estimation of the Demand Function 99 Specification of the Model 99

A Simple Linear Regression Model 101 Assumptions Underlying the Simple Linear

Regression Model 102 Estimating the Population Regression

Coefficients 103 Using the Regression Equation to Make

Predictions 106 Inferences about the Population Regression

Coefficients 108 Correlation Coefficient 111 The Analysis of Variance 112

Multiple Linear Regression Model 114 Use of Computer Programs 115 Estimating the Population Regression

Coefficients 115 Using the Regression Model to Make

Forecasts 115 Inferences about the Population Regression

Coefficients 115 The Analysis of Variance 118

Summary 118 Exercises 119 Case Exercise: Soft Drink Demand

Estimation 124

4A Problems in Applying the Linear Regression Model 126

Introduction 126 Nonlinear Regression Models 132 Summary 135 Exercises 135

5 Business and Economic Forecasting 137

Chapter Preview 137 Managerial Challenge: Excess Fiber

Optic Capacity at Global Crossing Inc. 137

The Significance of Forecasting 139 Selecting a Forecasting Technique 139

Hierarchy of Forecasts 139 Criteria Used to Select a Forecasting

Technique 140 Evaluating the Accuracy of Forecasting

Models 140

What Went Right/What Went Wrong: Crocs Shoes 140

Contents ix

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

x Contents

Alternative Forecasting Techniques 141 Deterministic Trend Analysis 141

Components of a Time Series 141 Some Elementary Time-Series Models 142 Secular Trends 143 Seasonal Variations 146

Smoothing Techniques 147 Moving Averages 148 First-Order Exponential Smoothing 151

Barometric Techniques 154 Leading, Lagging, and Coincident Indicators 154

Survey and Opinion-Polling Techniques 155 Forecasting Macroeconomic Activity 158 Sales Forecasting 159

Econometric Models 159 Advantages of Econometric Forecasting

Techniques 159 Single-Equation Models 160 Multi-Equation Models 160 Consensus Forecasts: Blue Chip Forecaster

Surveys 162 Stochastic Time-Series Analysis 163 Forecasting with Input-Output Tables 166 International Perspectives: Long-Term

Sales Forecasting by General Motors in Overseas Markets 167

Summary 167 Exercises 168 Case Exercise: Cruise Ship Arrivals in

Alaska 172 Case Exercise: Lumber Price Forecast 173

6 Managing in the Global Economy 175

Chapter Preview 175 Managerial Challenge: Financial

Crisis Crushes U.S. Household Consumption and Business Investment: Will Exports to China Provide the Way Out? 175

Introduction 178

What Went Right/What Went Wrong: Export Market Pricing at Toyota 179 Import-Export Sales and Exchange

Rates 179 Foreign Exchange Risk 180

International Perspectives: Collapse of Export and Domestic Sales at Cummins Engine 181

Outsourcing 183 China Trade Blossoms 185

China Today 186 The Market for U.S. Dollars as Foreign

Exchange 187

Import/Export Flows and Transaction Demand for a Currency 189

The Equilibrium Price of the U.S. Dollar 189 Speculative Demand, Government

Transfers, and Coordinated Intervention 189 Short-Term Exchange Rate Fluctuations 190

Determinants of Long-Run Trends in Exchange Rates 191 The Role of Real Growth Rates 191 The Role of Real Interest Rates 194 The Role of Expected Inflation 194

Purchasing Power Parity 195 PPP Offers a Better Yardstick of

Comparative Growth 196 Relative Purchasing Power Parity 197 Qualifications of PPP 198

What Went Right/What Went Wrong: GM, Toyota, and the Celica GT-S Coupe 199

The Appropriate Use of PPP: An Overview 200 Big Mac Index of Purchasing Power Parity 201 Trade-Weighted Exchange Rate Index 201

International Trade: A Managerial Perspective 204 Shares of World Trade and Regional

Trading Blocs 204 Comparative Advantage and Free Trade 207 Import Controls and Protective Tariffs 209 The Case for Strategic Trade Policy 211 Increasing Returns 213 Network Externalities 214

Free Trade Areas: The European Union and NAFTA 214 Optimal Currency Areas 216 Intraregional Trade 216 Mobility of Labor 216 Correlated Macroeconomic Shocks 217

Largest U.S. Trading Partners: The Role of NAFTA 217 A Comparison of the EU and NAFTA 219 Gray Markets, Knockoffs, and Parallel

Importing 220

What Went Right/What Went Wrong: Ford Motor Co. and Exide Batteries: Are Country Managers Here to Stay? 222 Perspectives on the U.S. Trade Deficit 222 Summary 224 Exercises 225 Case Exercise: Predicting the Long-Term

Trends in Value of the U.S. Dollar and Euro 226

Case Exercise: Elaborate the Debate on NAFTA 226

6A Foreign Exchange Risk Management 227

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Contents xi

PART III

PRODUCTION AND COST 229

7 Production Economics 230

Chapter Preview 230 Managerial Challenge: Green Power

Initiatives Examined: What Went Wrong in California’s Deregulation of Electricity? 230

The Production Function 232 Fixed and Variable Inputs 234

Production Functions with One Variable Input 235 Marginal and Average Product Functions 235 The Law of Diminishing Marginal Returns 236

What Went Right/What Went Wrong: Factory Bottlenecks at a Boeing Assembly Plant 237

Increasing Returns with Network Effects 237 Producing Information Services under

Increasing Returns 239 The Relationship between Total, Marginal,

and Average Product 239 Determining the Optimal Use of the

Variable Input 242 Marginal Revenue Product 242 Marginal Factor Cost 242 Optimal Input Level 243

Production Functions with Multiple Variable Inputs 243 Production Isoquants 243 The Marginal Rate of Technical Substitution 245

Determining the Optimal Combination of Inputs 248 Isocost Lines 248 Minimizing Cost Subject to an Output

Constraint 249 A Fixed Proportions Optimal Production

Process 250 Production Processes and Process Rays 251

Measuring the Efficiency of a Production Process 252

Returns to Scale 253 Measuring Returns to Scale 254 Increasing and Decreasing Returns to Scale 255 The Cobb-Douglas Production Function 255 Empirical Studies of the Cobb-Douglas

Production Function in Manufacturing 256 A Cross-Sectional Analysis of U.S.

Manufacturing Industries 256 Summary 259 Exercises 260

Case Exercise: The Production Function for Wilson Company 263

7A Maximization of Production Output Subject to a Cost Constraint 265

Exercise 266

7B Production Economics of Renewable and Exhaustible Natural Resources 267

Renewable Resources 267 Exhaustible Natural Resources 270 Exercises 274

8 Cost Analysis 275

Chapter Preview 275 Managerial Challenge: US Airways Cost

Structure 275 The Meaning and Measurement of Cost 276

Accounting versus Economic Costs 276 Three Contrasts between Accounting and

Economic Costs 277 Short-Run Cost Functions 281

Average and Marginal Cost Functions 281 Long-Run Cost Functions 286

Optimal Capacity Utilization: Three Concepts 286

Economies and Diseconomies of Scale 287 The Percentage of Learning 289 Diseconomies of Scale 291

International Perspectives: How Japanese Companies Deal with the Problems of Size 292 The Overall Effects of Scale Economies and

Diseconomies 293 Summary 295 Exercises 295 Case Exercise: Cost Analysis 298

8A Long-Run Costs with a Cobb-Douglas Production Function 301

Exercises 304

9 Applications of Cost Theory 305

Chapter Preview 305 Managerial Challenge: How Exactly

Have Computerization and Information Technology Lowered Costs at Chevron, Timken, and Merck? 305

Estimating Cost Functions 306 Issues in Cost Definition and Measurement 307 Controlling for Other Variables 307

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Form of the Empirical Cost-Output Relationship 308

What Went Right/What Went Wrong: Boeing: The Rising Marginal Cost of Wide-Bodies 309

Statistical Estimation of Short-Run Cost Functions 310

Statistical Estimation of Long-Run Cost Functions 310

Determining the Optimal Scale of an Operation 311

Economies of Scale versus Economies of Scope 314

Engineering Cost Techniques 314 The Survivor Technique 317 A Cautionary Tale 317

Break-Even Analysis 317 Graphical Method 318 Algebraic Method 319 Doing a Break-Even versus a Contribution

Analysis 323 Some Limitations of Break-Even and

Contribution Analysis 323 Operating Leverage 324 Business Risk 326 Break-Even Analysis and Risk Assessment 326

Summary 327 Exercises 328 Case Exercise: Cost Functions 330 Case Exercise: Charter Airline Operating

Decisions 331

PART IV

PRICING AND OUTPUT DECISIONS: STRATEGY AND TACTICS 333

10 Prices, Output, and Strategy: Pure and Monopolistic Competition 334

Chapter Preview 334 Managerial Challenge: Resurrecting

Apple 334 Introduction 335 Competitive Strategy 336

What Went Right/What Went Wrong: Xerox 337

Generic Types of Strategies 337 Product Differentiation Strategy 338 Cost-Based Strategy 339 Information Technology Strategy 339 The Relevant Market Concept 341

Porter’s Five Forces Strategic Framework 342

The Threat of Substitutes 342 The Threat of Entry 343 The Power of Buyers and Suppliers 346 The Intensity of Rivalrous Tactics 347 The Myth of Market Share 351

A Continuum of Market Structures 352 Pure Competition 352 Monopoly 353 Monopolistic Competition 354 Oligopoly 355

Price-Output Determination under Pure Competition 355 Short Run 355 Long Run 358

Price-Output Determination under Monopolistic Competition 361

What Went Right/What Went Wrong: The Dynamics of Competition at Amazon.com 362

Short Run 362 Long Run 362

Selling and Promotional Expenses 363 Determining the Optimal Level of Selling

and Promotional Outlays 363 Optimal Advertising Intensity 366 The Net Value of Advertising 367

Competitive Markets under Asymmetric Information 368 Incomplete versus Asymmetric Information 368 Search Goods versus Experience Goods 368 Adverse Selection and the Notorious Firm 369 Insuring and Lending under Asymmetric

Information: Another Lemons Market 371 Solutions to the Adverse Selection

Problem 372 Mutual Reliance: Hostage Mechanisms

Support Asymmetric Information Exchange 372

Brand-Name Reputations as Hostages 373 Price Premiums with Non-Redeployable

Assets 374 Summary 377 Exercises 378 Case Exercise: Blockbuster, Netflix, and

Redbox Compete for Movie Rentals 380 Case Exercise: Saving Sony Music 381

11 Price and Output Determination: Monopoly and Dominant Firms 382

Chapter Preview 382 Managerial Challenge: Dominant

Microprocessor Company Intel Adapts to Next Trend 382

xii Contents

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Monopoly Defined 383 Sources of Market Power for a

Monopolist 383 Increasing Returns from Network Effects 384

What Went Right/What Went Wrong: Pilot Error at Palm 387 Price and Output Determination for a

Monopolist 388 Spreadsheet Approach 388 Graphical Approach 389 Algebraic Approach 390 The Importance of the Price Elasticity of

Demand 391 The Optimal Markup, Contribution

Margin, and Contribution Margin Percentage 393 Components of the Gross Profit Margin 394 Monopolists and Capacity Investments 396 Limit Pricing 396

Regulated Monopolies 397 Electric Power Companies 397 Natural Gas Companies 399

What Went Right/What Went Wrong: The Public Service Company of New Mexico 400

Communications Companies 400 The Economic Rationale for Regulation 400

Natural Monopoly Argument 401 Summary 402 Exercises 403 Case Exercise: Differential Pricing of

Pharmaceuticals: The HIV/AIDS Crisis 406

12 Price and Output Determination: Oligopoly 409

Chapter Preview 409 Managerial Challenge: Are Nokia’s

Margins on Cell Phones Collapsing? 409 Oligopolistic Market Structures 411

Oligopoly in the United States: Relative Market Shares 411

Interdependencies in Oligopolistic Industries 415 The Cournot Model 415

Cartels and Other Forms of Collusion 417 Factors Affecting the Likelihood of

Successful Collusion 419 Cartel Profit Maximization and the

Allocation of Restricted Output 421 International Perspectives: The OPEC

Cartel 422 Cartel Analysis: Algebraic Approach 426

Price Leadership 429 Barometric Price Leadership 430 Dominant Firm Price Leadership 430

The Kinked Demand Curve Model 434 Avoiding Price Wars 434

What Went Right/What Went Wrong: Good-Better-Best Product Strategy at Kodak and Marriott 437 Summary 440 Exercises 440 Case Exercise: Cell Phones Displace

Mobile Phone Satellite Networks 442

13 Best-Practice Tactics: Game Theory 444

Chapter Preview 444 Managerial Challenge: Large-Scale Entry

Deterrence of Low-Cost Discounters: Southwest, People Express, Value Jet, Kiwi, and JetBlue 444

Oligopolistic Rivalry and Game Theory 445 A Conceptual Framework for Game Theory

Analysis 446 Components of a Game 447 Cooperative and Noncooperative Games 449 Other Types of Games 449

Analyzing Simultaneous Games 450 The Prisoner’s Dilemma 450 Dominant Strategy and Nash Equilibrium

Strategy Defined 452 The Escape from Prisoner’s Dilemma 455

Multiperiod Punishment and Reward Schemes in Repeated Play Games 455

Unraveling and the Chain Store Paradox 456 Mutual Forbearance and Cooperation in

Repeated Prisoner’s Dilemma Games 458 Bayesian Reputation Effects 459 Winning Strategies in Evolutionary

Computer Tournaments: Tit for Tat 459 Price-Matching Guarantees 461 Industry Standards as Coordination Devices 463

Analyzing Sequential Games 464 A Sequential Coordination Game 465 Subgame Perfect Equilibrium in Sequential

Games 466 Business Rivalry as a Self-Enforcing

Sequential Game 467 First-Mover and Fast-Second Advantages 469

Credible Threats and Commitments 471 Mechanisms for Establishing Credibility 472 Replacement Guarantees 473

Hostages Support the Credibility of Commitments 475

Contents xiii

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

xiv Contents

Credible Commitments of Durable Goods Monopolists 476

Planned Obsolescence 476 Post-Purchase Discounting Risk 477 Lease Prices Reflect Anticipated Risks 480

Summary 480 Exercises 481 Case Exercise: International Perspectives:

The Superjumbo Dilemma 485

13A Entry Deterrence and Accommodation Games 488

Excess Capacity as a Credible Threat 488 Pre-commitments Using Non-Redeployable

Assets 488 Customer Sorting Rules 491 Tactical Insights about Slippery Slopes 495 Summary 497 Exercises 497

14 Pricing Techniques and Analysis 499

Chapter Preview 499 Managerial Challenge: Pricing of Apple

Computers: Market Share versus Current Profitability 499

A Conceptual Framework for Proactive, Systematic-Analytical, Value-Based Pricing 500

Optimal Differential Price Levels 503 Graphical Approach 504 Algebraic Approach 505 Multiple-Product Pricing Decision 506 Differential Pricing and the Price Elasticity

of Demand 507 Differential Pricing in Target Market

Segments 512 Direct Segmentation with “Fences” 513 Optimal Two-Part Tariffs 515

What Went Right/What Went Wrong: Two-Part Pricing at Disney World 517

Couponing 517

What Went Right/What Went Wrong: Price-Sensitive Customers Redeem 517

Bundling 518 Price Discrimination 521

Pricing in Practice 523 Product Life Cycle Framework 523 Full-Cost Pricing versus Incremental

Contribution Analysis 525 Pricing on the Internet 527

The Practice of Revenue Management, Advanced Material 529

A Cross-Functional Systems Management Process 531

Sources of Sustainable Price Premiums 531 Revenue Management Decisions, Advanced

Material 533 Summary 540 Exercises 541

PART V

ORGANIZATIONAL ARCHITECTURE AND REGULATION 545

15 Contracting, Governance, and Organizational Form 546

Chapter Preview 546 Managerial Challenge: Controlling the

Vertical: Ultimate TV 546 Introduction 547 The Role of Contracting in Cooperative

Games 547 Vertical Requirements Contracts 549 The Function of Commercial Contracts 549 Incomplete Information, Incomplete

Contracting, and Post-Contractual Opportunism 553

Corporate Governance and the Problem of Moral Hazard 553 The Need for Governance Mechanisms 555

What Went Right/What Went Wrong: Moral Hazard and Holdup at Enron and WorldCom 556 The Principal-Agent Model 557

The Efficiency of Alternative Hiring Arrangements 557

Work Effort, Creative Ingenuity, and the Moral Hazard Problem in Managerial Contracting 558

Formalizing the Principal-Agent Problem 561 Screening and Sorting Managerial Talent

with Optimal Incentives Contracts 561

What Went Right/What Went Wrong: Why Have Restricted Stock Grants Replaced Executive Stock Options at Microsoft? 562 Choosing the Efficient Organizational

Form 564

What Went Right/What Went Wrong: Cable Allies Refuse to Adopt Microsoft’s WebTV as an Industry Standard 567

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

International Perspectives: Economies of Scale and International Joint Ventures in Chip Making 568 Prospect Theory Motivates Full-Line

Forcing 568 Vertical Integration 571

What Went Right/What Went Wrong: Dell Replaces Vertical Integration with Virtual Integration 573

The Dissolution of Assets in a Partnership 574 Summary 575 Exercises 576 Case Exercise: Borders Books and

Amazon.com Decide to Do Business Together 578

Case Exercise: Designing a Managerial Incentive Contract 578

Case Exercise: The Division of Investment Banking Fees in a Syndicate 578

15A Auction Design and Information Economics 580

Optimal Mechanism Design 580 First-Come, First-Served versus Last-Come,

First-Served 581 Auctions 583 Incentive-Compatible Revelation

Mechanisms 598 International Perspectives: Joint Venture

in Memory Chips: IBM, Siemens, and Toshiba 603

International Perspectives: Whirlpool’s Joint Venture in Appliances Improves upon Maytag’s Outright Purchase of Hoover 605

Summary 605 Exercises 607 Case Exercise: Spectrum Auction 608 Case Exercise: Debugging Computer

Software: Intel 608

16 Government Regulation 610

Chapter Preview 610 Managerial Challenge: Cap and Trade,

Deregulation, and the Coase Theorem 610 The Regulation of Market Structure and

Conduct 611 Market Performance 611 Market Conduct 612 Market Structure 612 Contestable Markets 614

Antitrust Regulation Statutes and Their Enforcement 614 The Sherman Act (1890) 614 The Clayton Act (1914) 615 The Federal Trade Commission Act

(1914) 615 The Robinson-Patman Act (1936) 615 The Hart-Scott-Rodino Antitrust

Improvement Act (1976) 616 Antitrust Prohibition of Selected Business

Decisions 617 Collusion: Price Fixing 617 Mergers That Substantially Lessen

Competition 617 Merger Guidelines (1992 and 1997) 619 Monopolization 619 Wholesale Price Discrimination 620 Refusals to Deal 622 Resale Price Maintenance Agreements 622

Command and Control Regulatory Constraints: An Economic Analysis 622 The Deregulation Movement 624

What Went Right/What Went Wrong: The Need for a Regulated Clearinghouse to Control Counterparty Risk at AIG 625 Regulation of Externalities 626

Coasian Bargaining for Reciprocal Externalities 626

Qualifications of the Coase Theorem 628 Impediments to Bargaining 629 Resolution of Externalities by Regulatory

Directive 630 Resolution of Externalities by Taxes and

Subsidies 630 Resolution of Externalities by Sale of

Pollution Rights: Cap and Trade 632 Governmental Protection of Business 633

Licensing and Permitting 633 Patents 633

The Optimal Deployment Decision: To License or Not 634

What Went Right/What Went Wrong: Delayed Release at Aventis 635

Pros and Cons of Patent Protection and Licensure of Trade Secrets 636

What Went Right/What Went Wrong: Technology Licenses Cost Palm Its Lead in PDAs 637

What Went Right/What Went Wrong: Motorola: What They Didn’t Know Hurt Them 638

Conclusion on Licensing 639

Contents xv

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Summary 640 Exercises 641 Case Exercise: Microsoft Tying

Arrangements 643 Case Exercise: Music Recording Industry

Blocked from Consolidating 643

17 Long-Term Investment Analysis 644

Chapter Preview 644 Managerial Challenge: Multigenerational

Effects of Ozone Depletion and Greenhouse Gases 644

The Nature of Capital Expenditure Decisions 647

A Basic Framework for Capital Budgeting 647 The Capital Budgeting Process 647

Generating Capital Investment Projects 648 Estimating Cash Flows 649 Evaluating and Choosing the Investment

Projects to Implement 650 Estimating the Firm’s Cost of Capital 653

Cost of Debt Capital 654 Cost of Internal Equity Capital 655 Cost of External Equity Capital 656 Weighted Cost of Capital 657

Cost-Benefit Analysis 658 Accept-Reject Decisions 658 Program-Level Analysis 659

Steps in Cost-Benefit Analysis 660 Objectives and Constraints in Cost-Benefit

Analysis 660 Analysis and Valuation of Benefits and

Costs 662 Direct Benefits 662 Direct Costs 663 Indirect Costs or Benefits and Intangibles 663

The Appropriate Rate of Discount 663 Cost-Effectiveness Analysis 664

Least-Cost Studies 665 Objective-Level Studies 665

Summary 666 Exercises 667 Case Exercise: Cost-Benefit Analysis 670 Case Exercise: Industrial Development Tax

Relief and Incentives 672

APPENDICES

A The Time Value of Money A-1 B Tables B-1 C Differential Calculus Techniques in

Management C-1 D Check Answers to Selected

End-of-Chapter Exercises D-1

Glossary G-1 Index I-1 Notes

WEB APPENDICES

A Consumer Choice Using Indifference Curve Analysis

B International Parity Conditions C Linear-Programming Applications D Capacity Planning and Pricing Against a

Low-Cost Competitor: A Case Study of Piedmont Airlines and People Express

E Pricing of Joint Products and Transfer Pricing F Decisions Under Risk and Uncertainty

xvi Contents

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Preface

ORGANIZATION OF THE TEXT The 12th edition has been thoroughly updated with more than 50 new applications. Although shortened to 672 pages, the book still covers all previous topics. Responding to user request, we have expanded the review of microeconomic fundamentals in Chap- ter 2, employing a wide-ranging discussion of the equilibrium price of crude oil and gas- oline. A new Appendix 7B on the Production Economics of Renewable and Exhaustible Natural Resources is complemented by a new feature on environmental effects and sus- tainability. A compact fluorescent lightbulb symbol highlights these discussions spread throughout the text. Another special feature is the extensive treatment in Chapter 6 of managing global businesses, import-export trade, exchange rates, currency unions and free trade areas, trade policy, and an extensive new section on China.

There is more comprehensive material on applied game theory in Chapter 13, 13A, 15, 15A, and Web Appendix D than in any other managerial economics textbook, and a unique treatment of yield (revenue) management appears in Chapter 14 on pricing. Part V includes the hot topics of corporate governance, information economics, auction design, and the choice of organization form. Chapter 16 on economic regulation includes a broad discussion of cap and trade policy, pollution taxes, and the optimal abatement of externalities. By far the most distinctive feature of the book, however, is its 300 boxed examples, Managerial Challenges, What Went Right/What Went Wrong explorations of corporate practice, and mini-case examples on every other page demonstrating what each analytical concept is used for in practice. This list of concept applications is highlighted on the inside front and back covers.

STUDENT PREPARATION The text is designed for use by upper-level undergraduates and first-year graduate stu- dents in business schools, departments of economics, and professional schools of man- agement, public policy, and information science as well as in executive training programs. Students are presumed to have a background in the basic principles of micro- economics, although Chapter 2 offers an extensive review of those topics. No prior work in statistics is assumed; development of all the quantitative concepts employed is self- contained. The book makes occasional use of elementary concepts of differential calculus. In all cases where calculus is employed, at least one alternative approach, such as graph- ical, algebraic, or tabular analysis, is also presented. Spreadsheet applications have be- come so prominent in the practice of managerial economics that we now address optimization in that context.

PEDAGOGICAL FEATURES OF THE 12TH EDITION The 12th edition of Managerial Economics makes extensive use of pedagogical aids to enhance individualized student learning. The key features of the book are:

xv i i

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

1. Managerial Challenges. Each chapter opens with a Managerial Challenge (MC) illuminating a real-life problem faced by managers that is closely related to the topics covered in the chapter. Instructors can use the new discussion questions fol- lowing each MC to “hook” student interest at the start of the class or in pre-class preparation assignments.

2. What Went Right/What Went Wrong. This feature allows students to relate busi- ness mistakes and triumphs to what they have just learned, and helps build that elusive goal of managerial insight.

3. Extensive Use of Boxed Examples. More than 300 real-world applications and ex- amples derived from actual corporate practice are highlighted throughout the text. These applications help the analytical tools and concepts to come alive and thereby enhance student learning. They are listed on the inside front and back covers to highlight the prominence of this feature of the book.

4. Environmental Effects Symbol. A CFL bulb symbol highlights numerous passages throughout the book that address environmental effects and sustainability.

5. Exercises. Each chapter contains a large problem analysis set. Check answers to se- lected problems color-coded in blue type are provided in Appendix C at the end of the book. Problems that can be solved using Excel are highlighted with an Excel icon. The book’s Web site (www.cengage.com/economics/mcguigan) has answers to all the other textbook problems.

6. Case Exercises. Most chapters include mini-cases that extend the concepts and tools developed into a deep fact situation context of a real-world company.

7. Chapter Glossaries. In the margins of the text, new terms are defined as they are introduced. The placement of the glossary terms next to the location where the term is first used reinforces the importance of these new concepts and aids in later studying.

8. International Perspectives. Throughout the book, special International Perspec- tives sections are provided that illustrate the application of managerial economics concepts to an increasingly global economy. A globe symbol highlights this internationally-relevant material.

9. Point-by-Point Summaries. Each chapter ends with a detailed, point-by-point summary of important concepts from the chapter.

10. Diversity of Presentation Approaches. Important analytical concepts are presented in several different ways, including tabular analysis, graphical analysis, and alge- braic analysis to individualize the learning process.

ANCILLARY MATERIALS A complete set of ancillary materials is available to adopters to supplement the text, in- cluding the following:

Instructor’s Manual and Test Bank Prepared by Richard D. Marcus, University of Wisconsin–Milwaukee, the instructor’s manual and test bank that accompany the book contain suggested answers to the end- of-chapter exercises and cases. The authors have taken great care to provide an error- free manual for instructors to use. The manual is available to instructors on the book’s Web site as well as on the Instructor’s Resource CD-ROM (IRCD). The test bank, con- taining a large collection of true-false, multiple-choice, and numerical problems, is avail- able to adopters and is also available on the Web site in Word format, as well as on the IRCD.

xviii Preface

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

ExamView Simplifying the preparation of quizzes and exams, this easy-to-use test creation software includes all of the questions in the printed test bank and is compatible with Microsoft Windows. Instructors select questions by previewing them on the screen, choosing them randomly, or picking them by number. They can easily add or edit questions, in- structions, and answers. Quizzes can also be created and administered online, whether over the Internet, a local area network (LAN), or a wide area network (WAN).

Textbook Support Web Site When you adopt Managerial Economics: Applications, Strategy, and Tactics, 12e, you and your students will have access to a rich array of teaching and learning resources that you won’t find anywhere else. Located at www.cengage.com/economics/mcguigan, this out- standing site features additional Web Appendices including appendices on indifference curve analysis of consumer choice, international parity conditions, linear programming applications, a capacity planning entry deterrence case study, joint product pricing and transfer prices, and decision making under uncertainty. It also provides links to addi- tional instructor and student resources including a “Talk-to-the-Author” link.

PowerPoint Presentation Available on the product companion Web site, this comprehensive package provides an excellent lecture aid for instructors. Prepared by Richard D. Marcus at the University of Wisconsin–Milwaukee, these slides cover many of the most important topics from the text, and they can be customized by instructors to meet specific course needs.

CourseMate Interested in a simple way to complement your text and course content with study and practice materials? Cengage Learning’s Economics CourseMate brings course concepts to life with interactive learning, study, and exam preparation tools that support the printed textbook. Watch student comprehension soar as your class works with the printed text- book and the textbook-specific Web site. Economics CourseMate goes beyond the book to deliver what you need! You and your students will have access to ABC/BBC videos, Cengage’s EconApps (such as EconNews and EconDebate), unique study guide content specific to the text, and much more.

ACKNOWLEDGMENTS A number of reviewers, users, and colleagues have been particularly helpful in providing us with many worthwhile comments and suggestions at various stages in the develop- ment of this and earlier editions of the book. Included among these individuals are:

William Beranek, J. Walter Elliott, William J. Kretlow, William Gunther, J. William Hanlon, Robert Knapp, Robert S. Main, Edward Sussna, Bruce T. Allen, Allen Moran, Edward Oppermann, Dwight Porter, Robert L. Conn, Allen Parkman, Daniel Slate, Richard L. Pfister, J. P. Magaddino, Richard A. Stanford, Donald Bumpass, Barry P. Keating, John Wittman, Sisay Asefa, James R. Ashley, David Bunting, Amy H. Dalton, Richard D. Evans, Gordon V. Karels, Richard S. Bower, Massoud M. Saghafi, John C. Callahan, Frank Falero, Ramon Rabinovitch, D. Steinnes, Jay Damon Hobson, Clifford Fry, John Crockett, Marvin Frankel, James T. Peach, Paul Kozlowski, Dennis Fixler, Steven Crane, Scott L. Smith, Edward Miller, Fred Kolb, Bill Carson, Jack W. Thornton, Changhee Chae, Robert B. Dallin, Christopher J. Zappe, Anthony V. Popp, Phillip M. Sisneros, George Brower, Carlos Sevilla, Dean Baim, Charles Callahan, Phillip Robins,

Preface xix

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Bruce Jaffee, Alwyn du Plessis, Darly Winn, Gary Shoesmith, Richard J. Ward, William H. Hoyt, Irvin Grossack, William Simeone, Satyajit Ghosh, David Levy, Simon Hakim, Patricia Sanderson, David P. Ely, Albert A. O’Kunade, Doug Sharp, Arne Dag Sti, Walker Davidson, David Buschena, George M. Radakovic, Harpal S. Grewal, Stephen J. Silver, Michael J. O’Hara, Luke M. Froeb, Dean Waters, Jake Vogelsang, Lynda Y. de la Viña, Audie R. Brewton, Paul M. Hayashi, Lawrence B. Pulley, Tim Mages, Robert Brooker, Carl Emomoto, Charles Leathers, Marshall Medoff, Gary Brester, Stephan Gohmann, L. Joe Moffitt, Christopher Erickson, Antoine El Khoury, Steven Rock, Rajeev K. Goel, Lee S. Redding, Paul J. Hoyt, Bijan Vasigh, Cheryl A. Casper, Semoon Chang, Kwang Soo Cheong, Barbara M. Fischer, John A. Karikari, Francis D. Mummery, Lucjan T. Orlowski, Dennis Proffitt, and Steven S. Shwiff.

People who were especially helpful in the preparation of the 12th edition include Robert F. Brooker, Kristen E. Collett-Schmitt, Simon Medcalfe, Dr. Paul Stock, Shahab Dabirian, James Leady, Stephen Onyeiwu, and Karl W. Einoff. A special thanks to B. Ramy Elitzur of Tel Aviv University for suggesting the exercise on designing a mana- gerial incentive contract.

We are also indebted to Richard D. Marcus, Bob Hebert, Sarah E. Harris, Wake Forest University, and the University of Louisville for the support they provided and owe thanks to our faculty colleagues for the encouragement and assistance provided on a continuing basis during the preparation of the manuscript. We wish to express our appreciation to the members of the South-Western/Cengage Learning staff—particularly, Betty Jung, Jana Lewis, Jennifer Thomas, Deepak Kumar, Steve Scoble, and Joe Sabatino—for their help in the preparation and promotion of this book. We are grateful to the Literary Executor of the late Sir Ronald A. Fisher, F.R.S.; to Dr. Frank Yates, F.R.S.; and to Longman Group, Ltd., London, for permission to reprint Table III from their book Statistical Tables for Bio- logical, Agricultural, and Medical Research (6th ed., 1974).

James R. McGuigan R. Charles Moyer

Frederick H. deB. Harris

xx Preface

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

About the Authors

James R. McGuigan James R. McGuigan owns and operates his own numismatic investment firm. Prior to this business, he was Associate Professor of Finance and Business Economics in the School of Business Administration at Wayne State University. He also taught at the University of Pittsburgh and Point Park College. McGuigan received his undergraduate degree from Carnegie Mellon University. He earned an M.B.A. at the Graduate School of Business at the University of Chicago and his Ph.D. from the University of Pittsburgh. In addition to his interests in economics, he has coauthored books on financial management. His re- search articles on options have been published in the Journal of Financial and Quantitative Analysis.

R. Charles Moyer R. Charles Moyer earned his B.A. in Economics from Howard University and his M.B.A. and Ph.D. in Finance and Managerial Economics from the University of Pittsburgh. Pro- fessor Moyer is Dean of the College of Business at the University of Louisville. He is Dean Emeritus and former holder of the GMAC Insurance Chair in Finance at the Babcock Graduate School of Management, Wake Forest University. Previously, he was Professor of Finance and Chairman of the Department of Finance at Texas Tech University. Profes- sor Moyer also has taught at the University of Houston, Lehigh University, and the Uni- versity of New Mexico, and spent a year at the Federal Reserve Bank of Cleveland. Professor Moyer has taught extensively abroad in Germany, France, and Russia. In addi- tion to this text, Moyer has coauthored two other financial management texts. He has been published in many leading journals including Financial Management, Journal of Financial and Quantitative Analysis, Journal of Finance, Financial Review, Journal of Financial Re- search, International Journal of Forecasting, Strategic Management Journal and Journal of Economics and Business. Professor Moyer is a member of the Board of Directors of King Pharmaceuticals, Inc., Capital South Partners, and the Kentucky Seed Capital Fund.

Frederick H. deB. Harris Frederick H. deB. Harris is the John B. McKinnon Professor at the Schools of Business, Wake Forest University. His specialties are pricing tactics and capacity planning. Professor Harris has taught integrative managerial economics core courses and B.A., B.S., M.S., M.B.A., and Ph.D. electives in business schools and economics departments in the United States, Europe, and Australia. He has won two school-wide Professor of the Year teaching awards and two Researcher of the Year awards. Other recognitions include Outstanding Faculty by Inc. magazine (1998), Most Popular Courses by Business Week Online 2000– 2001, and Outstanding Faculty by BusinessWeek’s Guide to the Best Business Schools, 5th to 9th eds., 1997–2004.

Professor Harris has published widely in economics, marketing, operations, and finance journals including the Review of Economics and Statistics, Journal of Financial and Quanti- tative Analysis, Journal of Operations Management, Journal of Industrial Economics, and Journal of Financial Markets. From 1988–1993, Professor Harris served on the Board of Associate Editors of the Journal of Industrial Economics. His current research focuses on

xx i

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the application of capacity-constrained pricing models to specialist and electronic trading systems for stocks. His path-breaking work on price discovery has been frequently cited in leading academic journals, and several articles with practitioners have been published in the Journal of Trading. In addition, he often benchmarks the pricing, order processing, and capacity planning functions of large companies against state-of-the-art techniques in revenue management and writes about his findings in journals like Marketing Management and INFORMS’s Journal of Revenue and Pricing Management.

xxii About the Authors

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Managerial Economics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PART1 INTRODUCTION

ECONOMIC ANALYSIS AND DECISIONS

1. Demand Analysis 2. Production and Cost Analysis 3. Product, Pricing, and Output

Decisions 4. Capital Expenditure Analysis

ECONOMIC, POLITICAL, AND SOCIAL ENVIRONMENT

1. Business Conditions (Trends, Cycles, and Seasonal Effects)

2. Factor Market Conditions (Capital, Labor, and Raw Materials)

3. Competitors’ Reactions and Tactical Response

4. Organizational Architecture and Regulatory Constraints

Cash Flows Risk

Firm Value (Shareholders’ Wealth)

1

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

1 CHAP T E R

Introduction and Goals of the Firm CHAPTER PREVIEW Managerial economics is the application of microeconomics to problems faced by decision makers in the private, public, and not-for-profit sectors. Managerial economics assists managers in efficiently allocating scarce resources, planning corporate strategy, and executing effective tactics. In this chapter, the responsibilities of management are explored. Economic profit is defined and the role of profits in allocating resources in a free enterprise system is examined. The primary goal of the firm, namely, shareholder wealth maximization, is developed along with a discussion of how managerial decisions influence shareholder wealth. The problems associated with the separation of ownership and control and principal-agent relationships in large corporations are explored.

MANAGERIAL CHALLENGE How to Achieve Sustainability: Southern Company1

In the second decade of the twenty-first century, com- panies all across the industrial landscape are seeking to achieve sustainability. Sustainability is a powerful meta- phor but an elusive goal. It means much more than aligning oneself with environmental sensitivity, though that commitment itself tests higher in opinion polling of the latent preferences of American and European custo- mers than any other response. Sustainability also im- plies renewability and longevity of business plans that are adaptable to changing circumstances without up- rooting the organizational strategy. But what exactly should management pursue as a set of objectives to achieve this goal?

Management response to pollution abatement illus- trates one type of sustainability challenge. At the insis- tence of the Prime Minister of Canada during the Reagan Administration, the U.S. Congress wrote a bi- partisan cap-and-trade bill to address smokestack emis- sions. Sulfur dioxide and nitrous oxide (SOX and NOX) emissions precipitate out as acid rain, mist, and ice, im-

posing damage downwind over hundreds of miles to painted and stone surfaces, trees, and asthmatics. The Clean Air Act (CAA) of 1990, amended in 1997 and 2003, granted tradable pollution allowance assets (TPAs) to known polluters. The CAA also authorized an auction market for these TPA assets. The EPA Web site (www.epa.gov) displays on a daily basis the equilibrium, market-clearing price (e.g., $250 per ton of soot) for the use of what had previously been an un- priced common property resource—namely, acid-free air and rainwater. Thereby, large point-source polluters like power plants and steel mills earned an actual cost per ton for the SOX and NOX–laden soot by-products of burning lots of high sulfur coal. These amounts were promptly placed in spreadsheets designed to find ways of minimizing operating costs.2 No less importantly, each polluter felt powerful incremental incentives to mitigate compliance cost by reducing pollution. And an entire industry devoted to developing pollution abatement technology sprang up.

Cont.

2

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The TPAs granted were set at approximately 80 per- cent of the known pollution taking place at each plant in 1990. For example, Duke Power’s Belews Creek power plant in northwestern North Carolina, generating 82,076 tons of sulfur dioxide acidic soot annually from burning 400 train carloads of coal per day, was granted 62,930 tons of allowances (see Figure 1.1 displaying the 329 × 365 = 120,085 tons of nitrous oxide). Although this approach “grandfathered” a substantial amount of

pollution, the gradualism of the 1990 cap-and-trade bill was pivotally important to its widespread success. In- dustries like steel and electric power were given five years of transition to comply with the regulated emis- sions requirements, and then in 1997, the initial allow- ances were cut in half. Duke Power initially bought 19,146 allowances for Belews Creek at prices ranging from $131 to $480 per ton and then in 2003 built two 30-story smokestack scrubbers that reduced the NOX emissions by 75 percent.

Another major electric utility, Southern Company, analyzed three compliance choices on a least-cost cash flow basis: (1) buying allowances, (2) installing smoke- stack scrubbers, or (3) adopting fuel switching technol- ogy to burn higher-priced low-sulfur coal or even cleaner natural gas. In a widely studied case, the South- ern Company’s Bowen plant in North Georgia necessi- tated a $657 million scrubber that after depreciation and offsetting excess allowance revenue was found to cost $476 million. Alternatively, continuing to burn high- sulfur coal from the Appalachian Mountain region and buying the requisite allowances was projected to cost

FIGURE 1.1 Nitrous Oxide from Coal-Fired Power Plants (Daily Emissions in Tons, pre Clean Air Act)

Asheville CP&L

Cliffside Duke

Duke Allen

Marshall Duke

Riverbend Duke

Belews Creek Duke

Buck Duke44

39 5924

164

329 tons NOx

14

13 55

194

17

13

Cape Fear

CP&L

Weatherspoon CP&L

Sutton CP&L

Lee CP&L

Mayo CP&L

Roxboro CP&L

Dan River Duke

55

27

Source: NC Division of Air Quality.

MANAGERIAL CHALLENGE Continued ©

AP Im ag es /S te ph en

M or to n

Chapter 1: Introduction and Goals of the Firm 3

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

WHAT IS MANAGERIAL ECONOMICS? Managerial economics extracts from microeconomic theory those concepts and tech- niques that enable managers to select strategic direction, to allocate efficiently the re- sources available to the organization, and to respond effectively to tactical issues. All such managerial decision making seeks to do the following:

1. identify the alternatives, 2. select the choice that accomplishes the objective(s) in the most efficient manner, 3. taking into account the constraints 4. and the likely actions and reactions of rival decision makers.

For example, consider the following stylized decision problem:

$266 million. And finally, switching to low-sulfur coal and adopting fuel switching technology was found to cost $176 million. All these analyses were performed on a present value basis with cost projections over 25 years.

Southern Company’s decision to switch to low-sulfur coal was hailed far and wide as environmentally sensi- tive. Today, such decisions are routinely described as a sustainability initiative. Many electric utilities support these sustainable outcomes of cap-and-trade policies and even seek 15 percent of their power from renewable energy (RE). In a Case Study at the end of the chapter, we analyze several wind power RE alternatives to burn- ing cheap high-sulfur large carbon footprint coal.

The choice of fuel-switching technology to abate smoke- stack emissions was a shareholder value-maximizing choice for Southern Company for two reasons. First, switching to low-sulfur coal minimized projected cash flow compliance costs but, in addition, the fuel-switching technology created a strategic flexibility (a “real option”) that created additional shareholder value for the Southern Company. In this chapter, we will see what maximizing capitalized value of equity (shareholder value) is and what it is not.

Discussion Questions

� What’s the basic externality problem with acid rain? What objectives should management serve in responding to the acid rain problem?

� How does the Clean Air Act’s cap-and-trade approach to air pollution affect the Southern Company’s analysis of the previously unpriced common property air and water resources damaged by smokestack emissions?

� How should management comply with the Clean Air Act, or should the Southern Com- pany just pay the EPA’s fines? Why? How would you decide?

� Which among Southern Company’s three alternatives for compliance offered the most strategic flexibility? Explain.

1Based on Frederick Harris, Alternative Energy Symposium, Wake Forest Schools of Business (September 2008); and “Acid Rain: The Southern Com- pany,” Harvard Business School Publishing, HBS: 9-792-060. 2EPA fines for noncompliance of $2,000 per ton have always far exceeded the auction market cost of allowances ($131–$473 in recent years).

Example Capacity Expansion at Honda, N.A., and Toyota Motors, N.A. Honda and Toyota are attempting to expand their already substantial assembly op- erations in North America. Both companies face increasing demand for their U.S.-manufactured vehicles, especially Toyota Camrys and Honda Accords. Camrys and Accords rate extremely highly in consumer reports of durability and reliability. The demand for used Accords is so strong that they depreciate only 45 percent in their first four years. Other competing vehicles may depreciate as much

(Continued)

MANAGERIAL CHALLENGE Continued

4 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

THE DECISION-MAKING MODEL The ability to make good decisions is the key to successful managerial performance. All decision making shares several common elements. First, the decision maker must establish the objectives. Next, the decision maker must identify the problem. For example, the CEO of electronics retailer Best Buy may note that the profit margin on sales has been decreas- ing. This could be caused by pricing errors, declining labor productivity, or the use of out- dated retailing concepts. Once the source or sources of the problem are identified, the manager can move to an examination of potential solutions. The choice between these al- ternatives depends on an analysis of the relative costs and benefits, as well as other organi- zational and societal constraints that may make one alternative preferable to another.

The final step in the decision-making process, after all alternatives have been evalu- ated, is to analyze the best available alternative under a variety of changes in the assump- tions before making a recommendation. This crucial final step is referred to as a sensitivity analysis. Knowing the limitations of the planned course of action as the deci- sion environment changes, the manager can then proceed to an implementation of the decision, monitoring carefully any unintended consequences or unanticipated changes in the market. This six-step decision-making process is illustrated in Figure 1.2.

The Responsibilities of Management In a free enterprise system, managers are responsible for a number of goals. Managers are responsible for proactively solving problems before they become crises and for selecting strat- egies to assure the more likely success of the current business model. Managers create organi- zational structure and culture based on the organization’s mission. Senior management especially is responsible for establishing a vision of new business directions and setting stretch goals to get there. In addition, managers monitor, motivate, and incentivize teamwork and coordinate the integration of marketing, operations, and finance functions. In pursuing all of these responsibilities, managers in a capitalist economy are ever conscious of their over- arching goal to maximize returns to the owners of the business—that is, economic profits.

as 65 percent in the same period. Toyota and Honda have identified two possible strategies (S1NEW and S2USED) to meet the growing demand for Camrys and Ac- cords. Strategy S1NEW involves an internal expansion of capacity at Toyota’s $700 million Princeton, Indiana, plant and Honda’s Marysville, Ohio, plant. Strategy S2USED involves the purchase and renovation of assembly plants now owned by General Motors. The new plants will likely receive substantial public subsidies through reduced property taxes. The older plants already possess an enormous infrastructure of local suppliers and regulatory relief.

The objective of Toyota’s managers is to maximize the value today (present value) of the expected future profit from the expansion. This problem can be sum- marized as follows:

Objective function: Maximize the present value (P.V.) of profit (S1NEW, S2USED)

Decision rule: Choose strategy S1NEW if P.V.(Profit S1NEW) > P.V.(Profit S2USED) Choose strategy S2USED if the reverse.

This simple illustration shows how resource-allocation decisions of managers attempt to maximize the value of their firms across forward-looking dynamic strat- egies for growth while respecting all ethical, legal, and regulatory constraints.

Chapter 1: Introduction and Goals of the Firm 5

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Economic profit is the difference between total sales revenue (price times units sold) and total economic cost. The economic cost of any activity may be thought of as the highest valued alternative opportunity that is forgone. To attract labor, capital, intellectual property, land, and materiel, the firm must offer to pay a price that is suffi- cient to convince the owners of these resources to forego other alternative activities and commit their resources to this use. Thus, economic costs should always be thought of as opportunity costs—that is, the costs of attracting a resource such as investment capital from its next best alternative use.

THE ROLE OF PROFITS In a free enterprise system, economic profits play an important role in guiding the deci- sions made by the thousands of competing independent resource owners. The existence of profits determines the type and quantity of goods and services that are produced and sold, as well as the resulting derived demand for resources. Several theories of profit indicate how this works.

FIGURE 1.2 The Decision-Making Process

Analyze alternatives

and select the best

Implement and monitor the

decision

Consider societal

constraints

Consider organizational and input constraints

Establish objectives

Identify the problem

Examine possible alternative solutions

Perform a sensitivity analysis

economic profit The difference between total revenue and total economic cost. Economic cost includes a “normal” rate of return on the capital contributions of the firm’s partners.

6 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Risk-Bearing Theory of Profit Economic profits arise in part to compensate the owners of the firm for the risk they assume when making their investments. Because a firm’s shareholders are not entitled to a fixed rate of return on their investment—that is, they are claimants to the firm’s residual cash flows after all other contractual payments have been made—they need to be compensated for this risk in the form of a higher rate of return.

The risk-bearing theory of profits is explained in the context of normal profits, where normal is defined in terms of the relative risk of alternative investments. Normal profits for a high-risk firm, such as Las Vegas hotels and casinos or a biotech pharmaceutical company or an oil field exploration well operator, should be higher than normal profits for firms of lesser risk, such as water utilities. For example, the industry average return on net worth for the hotel/gaming industry was 12.6 percent in 2005, compared with 9 percent for the water utility industry.

Temporary Disequilibrium Theory of Profit Although there exists a long-run equilibrium normal rate of profit (adjusted for risk) that all firms should tend to earn, at any point in time, firms might earn a rate of return above or below this long-run normal return level. This can occur because of temporary dislocations (shocks) in various sectors of the economy. Rates of return in the oil indus- try rose substantially when the price of crude oil doubled from $75 in mid-2007 to $146 in July 2008. However, those high returns declined sharply by late 2008, when oil market conditions led to excess supplies and the price of crude oil fell to $45.

Monopoly Theory of Profit In some industries, one firm is effectively able to dominate the market and persistently earn above-normal rates of return. This ability to dominate the market may arise from economies of scale (a situation in which one large firm, such as Boeing, can produce ad- ditional units of 747 aircraft at a lower cost than can smaller firms), control of essential natural resources (diamonds), control of critical patents (biotech pharmaceutical firms), or governmental restrictions that prohibit competition (cable franchise owners). The conditions under which a monopolist can earn above-normal profits are discussed in greater depth in Chapter 11.

Innovation Theory of Profit The innovation theory of profit suggests that above-normal profits are the reward for successful innovations. Firms that develop high-quality products (such as Porsche) or successfully identify unique market opportunities (such as Microsoft) are rewarded with the potential for above-normal profits. Indeed, the U.S. patent system is designed to en- sure that these above-normal return opportunities furnish strong incentives for contin- ued innovation.

Managerial Efficiency Theory of Profit Closely related to the innovation theory is the managerial efficiency theory of profit. Above-normal profits can arise because of the exceptional managerial skills of well- managed firms. No single theory of profit can explain the observed profit rates in each industry, nor are these theories necessarily mutually exclusive. Profit performance is in- variably the result of many factors, including differential risk, innovation, managerial skills, the existence of monopoly power, and chance occurrences.

Chapter 1: Introduction and Goals of the Firm 7

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

OBJECTIVE OF THE FIRM These theories of simple profit maximization as an objective of management are insight- ful, but they ignore the timing and risk of profit streams. Shareholder wealth maximiza- tion as an objective overcomes both these limitations.

The Shareholder Wealth-Maximization Model of the Firm To maximize the value of the firm, managers should maximize shareholder wealth. Shareholder wealth is measured by the market value of a firm’s common stock, which is equal to the present value of all expected future cash flows to equity owners dis- counted at the shareholders’ required rate of return plus a value for the firm’s embedded real options:

V0 · ðShares OutstandingÞ = π1ð1+keÞ1 +

π2

ð1+keÞ2 +

π3

ð1+keÞ3 + . . . +

π∞ ð1+keÞ∞

+ Real Option Value

V0 · ðShares OutstandingÞ = ∑ ∞

t=1

πt ð1+keÞt

+ Real Option Value [1.1]

where V0 is the current value of a share of stock (the stock price), πt represents the eco- nomic profits expected in each of the future periods (from period 1 to ∞), and ke equals the required rate of return.

A number of different factors (like interest rates and economy-wide business cycles) influence the firm’s stock price in ways that are beyond the manager’s control, but many factors (like innovation and cost control) are not. Real option value represents the cost savings or revenue expansions that arise from preserving flexibility in the business plans the managers adopt. For example, the Southern Company saved $90 million in comply- ing with the Clean Air Act by adopting fuel-switching technology that allowed burning of alternative high- and low-sulfur coals or fuel oil whenever the full cost of one input became cheaper than another.

Note that Equation 1.1 does take into account the timing of future profits. By discount- ing all future profits at the required rate of return, ke, Equation 1.1 shows that a dollar

Example Shareholder Wealth Maximization at Berkshire Hathaway Warren E. Buffett, chairman and CEO of Berkshire Hathaway, Inc., has described the long-term economic goal of Berkshire Hathaway as follows: “to maximize the average annual rate of gain in intrinsic business value on a per-share basis.”3 Berk- shire’s book value per share has increased from $19.46 in 1964, when Buffett ac- quired the firm, to $91,485 at the end of 2005, a compound annual rate of growth of 21.5 percent. The Standard and Poor’s 500 companies experienced 10.3 percent growth over this same time period.

Berkshire’s directors are all major stockholders. In addition, at least four of the di- rectors have over 50 percent of their family’s net worth invested in Berkshire. Man- agers and directors own over 47 percent of the firm’s stock. As a result, Buffet’s firm has always placed a high priority on the goal of maximizing shareholder wealth.

3Annual Report, Berkshire Hathaway, Inc. (2005).

shareholder wealth A measure of the value of a firm. Shareholder wealth is equal to the value of a firm’s common stock, which, in turn, is equal to the present value of all future cash returns expected to be generated by the firm for the benefit of its owners.

8 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

received in the future is worth less than a dollar received immediately. (The techniques of discounting to present value are explained in more detail in Chapter 2 and Appendix A at the end of the book.) Equation 1.1 also provides a way to evaluate different levels of risk since the higher the risk the higher the required rate of return ke used to discount the future cash flows, and the lower the present value. In short, shareholder value is deter- mined by the amount, timing, and risk of the firm’s expected future profits.

SEPARATION OF OWNERSHIP AND CONTROL: THE PRINCIPAL-AGENT PROBLEM Profit maximization and shareholder wealth maximization are very useful concepts when alternative choices can be easily identified and when the associated costs and revenues can be readily estimated. Examples include scheduling capacity for optimal production runs, determining an optimal inventory policy given sales patterns and available produc- tion facilities, introducing an established product in a new geographic market, and choosing whether to buy or lease a machine. In other cases, however, where the alterna- tives are harder to identify and the costs and benefits less clear, the goals of owners and managers are seldom aligned.

Example Resource-Allocation Decisions and Shareholder Wealth: Apple Computer4

In distributing its stylish iMac personal computers and high tech iPods, Apple has considered three distribution channels. On the one hand, copying Dell’s direct- to-the-consumer approach would entail buying components from Motorola, AMD, Intel, and so forth and then hiring third-party manufacturers to assemble what each customer ordered just-in-time to fulfill Internet or telephone sales. In- ventories and capital equipment costs would be very low indeed; almost all costs would be variable. Alternatively, Apple could enter into distribution agreements with an independent electronics retailer like Computer Tree. Finally, Apple could retail its own products in Apple Stores. This third approach entails enormous cap- ital investment and a higher proportion of fixed cost, especially if the retail chain sought high visibility locations and needed lots of space.

Recently Apple opened its 147th retail store on Fifth Avenue in New York City. The location left little doubt as to the allocation of company resources to this new distribution strategy. Apple occupies a sprawling subterranean space topped by a glass cube that Steve Jobs himself designed, across from Central Park, opposite the famed Plaza Hotel. Apple’s profits in this most heavily trafficked tourist and retail corridor will rely on several initiatives: (1) in-store theatres for workshop training on iMac programs to record music or edit home movies, (2) numerous technical experts available for troubleshooting with no waiting time, and (3) con- tinuing investment in one of the world’s most valuable brands. In 2005, Apple made $151 million in operating profits on $2.35 billion in sales at these Apple Stores, a 6.4 percent profit margin relative to approximately a 2 percent profit mar- gin company-wide.

4Based on Nick Wingfield, “How Apple’s Store Strategy Beat the Odds,” Wall Street Journal (May 17, 2006), p. B1.

Chapter 1: Introduction and Goals of the Firm 9

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Divergent Objectives and Agency Conflict As sole proprietorships and closely held businesses grow into limited liability corpora- tions, the owners (the principals) frequently delegate decision-making authority to pro- fessional managers (the agents). Because the manager-agents usually have much less to lose than the owner-principals, the agents often seek acceptable levels (rather than a maximum) of profit and shareholder wealth while pursuing their own self-interests. This is known as a principal-agent problem or “agency conflict.”

For example, as oil prices subsided with the collapse of the OPEC cartel in the 1990s, Exxon’s managers diversified the company into product lines like computer software development—an area where Exxon had little or no expertise or competitive advantage. The managers were hoping that diversification would smooth out their executive bonuses tied to quarterly earnings, and it did. However, the decision to diversify ended up caus- ing an extended decline in the value of Exxon’s stock.

Pursuing their own self-interests can also lead managers to focus on their own long-term job security. In some instances this can motivate them to limit the amount of risk taken by the firm because an unfavorable outcome resulting from the risk could lead to their dismissal. Kodak is a good example. In the early 2000s, Kodak’s executives didn’t want to risk developing immature digital photography products. When the demand for digital camera products subsequently soared, Kodak was left with too few markets for its traditional film products. Like Exxon, its stock value plummeted.

Finally, the cash flow to owners erodes when the firm’s resources are diverted from their most productive uses to perks for managers. In 1988, RJR Nabisco was a firm that had become bloated with corporate retreats in Florida, an extensive fleet of corporate airplanes and hangars, and an executive fixation on an awful-tasting new product (the “smokeless” cigarette Premier). This left RJR Nabisco with substantially less value in the marketplace than would have been possible with better resource allocation decisions. Recognizing the value enhancement potential, Kohlberg Kravis Roberts & Co. (KKR) initiated a hostile takeover bid and acquired RJR Nabisco for $25 billion in early 1989. The purchase price offered to common stockholders by KKR was $109 per share, much better than the $50 to $55 pre-takeover price. The new owners moved quickly to sell many of RJR’s poorly performing assets, slash op- erating expenses, and cancel the Premier project. Although the deal was heavily lev- eraged with a large amount of debt borrowed at high interest rates, a much-improved cash flow allowed KKR to pay down the debt within seven years, substantially ahead of schedule.

To forge a closer alliance between the interests of shareholders and managers, some companies structure a larger proportion of the manager’s compensation in the form of performance-based payments. For example, in 2002, Walt Disney’s Michael Eisner re- ceived over $20.2 million in long-term compensation (in addition to his $750,000 salary) as a reward for increasing Walt Disney’s market value 10-fold from $2 billion to $23 billion during his first 10 years as CEO.5 Other firms like Hershey Foods, CSX, Union Carbide, and Xerox require senior managers and directors to own a substantial amount of company stock as a condition of employment. The idea behind this is to align the pocketbook interests of managers directly with those of stockholders. In sum, how moti- vated a manager will be to act in the interests of the firm’s stockholders depends on the structure of his or her compensation package, the threat of dismissal, and the threat of takeover by a new group of owners.

5J. Steiner, Business, Society, and Government (New York: McGraw-Hill, 2003), pp. 660–662.

10 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Agency Problems Two common factors that give rise to all principal-agent problems are the inherent un- observability of managerial effort and the presence of random disturbances in team pro- duction. The job performance of piecework garment workers is easily monitored, but the work effort of salespeople and manufacturer’s trade representatives may not be observ- able at less-than-prohibitive cost. Directly observing managerial input is even more prob- lematic because managers contribute what one might call “creative ingenuity.” Creative ingenuity in anticipating problems before they arise is inherently unobservable. Owners know it when they see it, but often do not recognize when it is missing. As a result, in explaining fluctuations in company performance, the manager’s creative ingenuity is often inseparable from good and bad luck. Owners therefore find it difficult to know when to reward managers for upturns and when to blame them for poor performance.

To an attempt to mitigate these agency problems, firms incur several agency costs, which include the following:

1. Grants of restricted stock or deferred stock options to structure executive compensa- tion in such a way as to align the incentives for management with shareholder interests.

Separation of ownership (shareholders) and control (management) in large cor- porations permits managers to pursue goals, such as maximization of their own personal welfare, that are not always in the long-term interests of shareholders. As a result of pressure from large institutional shareholders, such as Fidelity Funds, from statutes such as Sarbanes-Oxley mandating stronger corporate governance, and from federal tax laws severely limiting the deductibility of executive pay, a growing num- ber of corporations are seeking to assure that a larger proportion of the manager’s pay occurs in the form of performance-based bonuses. They are doing so by (1) tying executive bonuses to the performance of comparably situated competitor companies, (2) by raising the performance hurdles that trigger executive bonuses, and (3) by eliminating severance packages that provide windfalls for executives whose poor per- formance leads to a takeover or their own dismissal.

In 2005, CEOs of the 350 largest U.S. corporations were paid $6 million in median total direct compensation. The 10 companies with the highest shareholder returns the previous five years paid $10.6 million in salary, bonus, and long-term

Example Agency Costs and Corporate Restructuring: O.M. Scott & Sons6

The existence of high agency costs sometimes prompts firms to financially restruc- ture themselves to achieve higher operating efficiencies. For example, the lawn pro- ducts firm O.M. Scott & Sons, previously a subsidiary of ITT, was purchased by the Scott managers in a highly leveraged buyout (LBO). Faced with heavy interest and principal payments from the debt-financed LBO transaction and having the poten- tial to profit directly from more efficient operation of the firm, the new owner- managers quickly put in place accounting controls and operating procedures designed to improve Scott’s performance. By monitoring inventory levels more closely and negotiating more aggressively with suppliers, the firm was able to reduce its average monthly working capital investment from an initial level of $75 million to $35 million. At the same time, incentive pay for the sales force caused revenue to increase from $160 million to a record $200 million.

6A more complete discussion of the Scott experience can be found in Brett Duval Fromson, “Life after Debt: How LBOs Do It,” Fortune (March 13, 1989), pp. 91–92.

agency costs Costs associated with resolving conflicts of interest among shareholders, managers, and lenders. Agency costs include the cost of monitoring and bonding performance, the cost of constructing contracts designed to minimize agency conflicts, and the loss in efficiency resulting from unresolved agent- principal conflicts.

Chapter 1: Introduction and Goals of the Firm 11

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

incentives. The 10 companies with the lowest shareholder returns paid $1.6 million. Figure 1.3 shows that across these 350 companies, CEO total compensation has mirrored corporate profitability, spiking when profits grow and collapsing when profits decline. In the global economic crisis of 2008–2009, CEO salaries declined in 63 percent of NYSE Euronext companies, and bonuses and raises were frozen, cut, or eliminated in 47 percent and 52 percent, respectively.7

Example Executive Performance Pay: General Electric8

As a representative example of a performance-based pay package, General Electric CEO Jeff Immelt had a 2006 salary of $3.2 million, a cash bonus of $5.9 million, and gains on long-term incentives that converted to stock options of $3.8 million. GE distributes stock options to 45,000 of its 300,000 employees, but decided that one-half of CEO Jeff Immelt’s 250,000 “performance share units” should only con- vert to stock options if GE cash flow grew at an average of 10 percent or more for five years, and the other one-half should convert only if GE shareholder return ex- ceeded the five-year cumulative total return on the S&P 500 index.

Basing these executive pay packages on demonstrated performance relative to in- dustry and sector benchmarks has become something of a cause célèbre in the United States. The reason is that by 2008 median CEO total compensation of $7.3 million had grown to 198 times the $37,000 salary of the average U.S. worker. In Europe, the comparable figure was $900,000, approximately 33 times the median worker sal- ary of $27,000.9 And similar multipliers to those in Europe apply in Asia. So, what U.S. CEOs get paid was the focus of much public policy discussion even before the pay scandals at AIG and Merrill Lynch/Bank of America in the fall of 2009.

8Based on http://people.forbes.com/rankings/jeffrey-r-immelt/36126 9Mercer Human Resources Consulting, “Executive Compensation” (2006).

FIGURE 1.3 CEO Pay Trends

+25%

+15%

+5%

–15%

–25%

0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2009

–5%

Corporate profits CEO compensation

2008

Source: Mercer Human Resource Consulting.

7“NYSE Euronext 2010 CEO Report,” NYSEMagazine.com (September 2009), p. 27.

12 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

2. Internal audits and accounting oversight boards to monitor management’s actions. In addition, many large creditors, especially banks, now monitor financial ratios and investment decisions of large debtor companies on a monthly or even biweekly basis. These initiatives strengthen the firm’s corporate governance.

3. Bonding expenditures and fraud liability insurance to protect the shareholders from managerial dishonesty.

4. Lost profits arising from complex internal approval processes designed to limit managerial discretion, but which prevent timely responses to opportunities.

IMPLICATIONS OF SHAREHOLDER WEALTH MAXIMIZATION Critics of those who want to align the interests of managers with equity owners often allege that maximizing shareholder wealth focuses on short-term payoffs—sometimes to the detriment of long-term profits. However, the evidence suggests just the opposite. Short-term cash flows reflect only a small fraction of the firm’s share price; the first 5 years of expected dividend payouts explain only 18 percent and the first 10 years only 35 percent of the share prices of NYSE stocks.11 The goal of shareholder wealth maximi- zation requires a long-term focus.

WHAT WENT RIGHT • WHAT WENT WRONG

Saturn Corporation10

When General Motors rolled out their “different kind of car company,” J.D. Powers rated product quality 8 per- cent ahead of Honda, and customers liked the no-haggle selling process. Saturn achieved the 200,000 unit sales en- joyed by the Honda Civic and the Toyota Corolla in two short years and caught the 285,000 volume of the Ford Escort in Saturn’s fourth year. Making interpersonal as- pects of customer service the number-one priority and possessing superior inventory and MIS systems, Saturn dealerships proved very profitable and quickly developed a reputation for some of the highest customer loyalty in the industry.

However, with pricing of the base Saturn model $1,200 below the $12,050 rival Japanese compact cars, the GM parent earned only a $400 gross profit margin per vehicle. In a typical year, this meant GM was recovering only about $100 million of its $3 billion capital investment, a paltry 3 percent return. Netting out GM’s 11 percent cost of capital, each Saturn was losing approximately $1,000. These figures compare to a $3,300 gross profit margin per vehicle in some of GM’s other divisions. Consequently, cash flow was not reinvested in the Saturn division, products were not updated, and the models stagnated. By 1997, sales

were slumping at −9 percent and in 1998 they fell an ad- ditional 20 percent. In 2009, GM announced it was perma- nently closing the Saturn division.

What problems appear responsible for Saturn’s mid-life crisis? GM failed to adopt a change-management view of what would be required to transfer the first-time Saturn owners to more profitable GM divisions. The corporate strategy was that price-conscious young Saturn buyers would eventually trade up to Buick and Oldsmobile. In- stead, middle-aged loyal Saturn owners sought to trade up within Saturn, and finding no sporty larger models available, they switched to larger Japanese imports like the Honda Accord and Toyota Camry. Saturn has now learned that companies whose products are exposed to competition from foreign producers must plan product in- troductions and marketing campaigns to account for this global competitive environment. Recent product introduc- tions have included a sport wagon, an efficient SUV, and a high-profile sports coupe.

10Based on M. Cohen, “Saturn’s Supply-Chain Innovation,” Sloan Manage- ment Review (Summer 2000), pp. 93–96; “Small Car Sales Are Back” and “Why Didn’t GM Do More for Saturn?” BusinessWeek, September 22, 1997, pp. 40–42, and March 16, 1998, p. 62.

11J.R. Woolridge, “Competitive Decline: Is a Myopic Stock Market to Blame?” Journal of Applied Corporate Finance (Spring 1988), pp. 26–36.

Chapter 1: Introduction and Goals of the Firm 13

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Admittedly, value-maximizing managers must manage change—sometimes radical changes in competition (free-wheeling electric power), in technology (Internet signal compression), in revenue collection (music), and in regulation (cigarettes)—but they must do so with an eye to the long-run sustainable profitability of the business. In short, value-maximizing managers must anticipate change and make contingency plans.

Shareholder wealth maximization also reflects dynamic changes in the information available to the public about a company’s expected future cash flows and foreseeable risks. An accounting scandal at Krispy Kreme caused the stock price to plummet from $41 to $20 per share in one month. Stock price also reflects not only the firm’s preexist- ing positive net present value investments, but also the firm’s strategic investment oppor- tunities (the “embedded real options”) a management team develops. Amgen, a biotechnology company, had shareholder value of $42 million in 1983 despite no sales, no cash flow, no capital assets, no patents, and poorly protected trade secrets. By 1993, Amgen had sales of over $1.4 billion and cash flow of $408 million annually. Amgen had developed and exercised enormously valuable strategic opportunities.

WHAT WENT RIGHT • WHAT WENT WRONG

Eli Lilly Depressed by Loss of Prozac Patent12

Pharmaceutical giants like GlaxoSmithKline, Merck, Pfizer, and Eli Lilly expend an average of $802 million to develop a new drug. It takes 12.3 years to research and test for efficacy and side effects, conduct clinical trials, and then produce and market a new drug. Only 4 in 100 candidate molecules or screening compounds lead to investigational new drugs (INDs). Only 5 in 200 of these INDs display sufficient efficacy in animal testing to warrant human trials. Clinical failure occurs in 6 of 10 human trials, and only half of the FDA-proposed drugs are ultimately ap- proved. In sum, the joint probability of successful drug discovery and development is just 0.04 × 0.025 × 0.4 × 0.5 = 0.0002, two hundredths of 1 percent. Those few pat- ented drugs that do make it to the pharmacy shelves, espe- cially the blockbusters with several billion dollars in sales, must contribute enough operating profit to recover the cost of all these R & D failures.

In 2000, one of the key extension patents for Eli Lilly’s blockbuster drug for the treatment of depression, Prozac,

was overturned by a regulator and a U.S. federal judge. Within one month, Eli Lilly lost 70 percent of Prozac’s sales to the generic equivalents. Although this company has several other blockbusters, Eli Lilly’s share price plum- meted 32 percent. CEO Sidney Taurel said he had made a mistake in not rolling out Prozac’s successor replacement drug when the patent extension for Prozac was first chal- lenged. Taurel then moved quickly to establish a new man- agement concept throughout the company. Now, each new Eli Lilly drug is assigned a team of scientists, marketers, and regulatory experts who oversee the entire life cycle of the product from research inception to patent expiration. The key function of these cross-functionally integrated teams is contingency analysis and scenario planning to deal with the unexpected.

12C. Kennedy, F. Harris, and M. Lord, “Integrating Public Policy and Public Affairs into Pharmaceutical Marketing: Differential Pricing and the AIDS Pandemic,” Journal of Public Policy and Marketing (Fall 2004), pp. 1–23; and “Eli Lilly: Bloom and Blight,” The Economist (October 26, 2002), p. 60.

Example Amgen’s Potential Profitability Is Realized Amgen, Inc. uses state-of-the-art biotechnology to develop human pharmaceutical and diagnostic products. After a period of early losses during their start-up phase, profits increased steadily from $19 million in 1989 to $355 million in 1993 to $670 million in 1996. On the strength of royalty income from the sale of its Epogen prod- uct, a stimulator of red blood cell production, profits jumped to $900 million per year by 1999. In 2009, Amgen was valued at $60 billion with revenues and cash flows hav- ing continued to grow throughout the previous 10 years at 19 percent annually.

14 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

In general, only about 85 percent of shareholder value can be explained by even 30 years of cash flows.13 The remainder reflects the capitalized value of strategic flexibil- ity to expand some profitable lines of business, to abandon others, and to retain but de- lay investment in still others until more information becomes available. These additional sources of equity value are referred to as “embedded real options.”

We need to address why NPV and option value are additive concepts. NPV was in- vented to value bonds where all the cash flows are known and guaranteed by contract. As a result, the NPV analysis adjusts for timing and for risk but ignores the value of flexibility present in some capital budgeting projects but not others. These so-called em- bedded options present the opportunity but not the obligation to take actions to maxi- mize the upside or minimize the downside of a capital investment. For example, investing in a fuel-switching technology in power plants allows Southern Company to burn fuel oil when that input is cheap and burn natural gas when it is cheaper. Similarly, building two smaller assembly plants, one in Japan and another in the United States, al- lows Honda Camry production to be shifted as currency fluctuations cause costs to fall in one plant location relative to the other. In general, a company can create flexibility in their capital budgeting by: (1) facilitating follow-on projects through growth options, (2) exiting early without penalty through abandonment options, or (3) staging investment over a learning period until better information is available through deferral options. The scenario planning that comes from such financial thinking compares the value of expanding, leaving, or waiting to the opportunity loss from shrinking, staying, or imme- diate investment. Flexibility of this sort expands upon the NPV from discounted cash flow alone.

Value-maximizing behavior on the part of managers is also distinguishable from satisficing behavior. Satisficers strive to “hit their targets” (for example, on sales growth, return on investment, or safety rating targets). Not value maximizers. Rather than trying to meet a standard like 97 percent, 99 percent, or 99.9 percent error-free takeoffs and landings at O’Hare field in Chicago, or deliver a 9, 11, or 12.1 percent return on share- holders’ equity, the value-maximizing manager will commit himself or herself to contin- uous incremental improvements. Any time the marginal benefits of an action exceed its marginal costs, the value-maximizing manager will just do it.

Example Real Option Value Attributable to Fuel-Switching Technology at Southern Company Ninety-six percent of all companies employ NPV analysis.14 Eighty-five percent employ sensitivity analysis to better understand their capital investments. Only 66.8 percent of companies pursue the scenario planning and contingency analysis that underlies real option valuation. A tiny 11.4 percent of companies formally cal- culate the value of their embedded real options. That suggests an opportunity for recently trained managers to introduce these new techniques of capital budgeting to improve stockholder value. Southern Company found its embedded real option from fuel switching technology was worth more than $45 million.

14Based on P. Ryan and G. Ryan, “Capital Budgeting Practices of the Fortune 1000: How Have Things Changed?” Journal of Business and Management (Fall 2002).

13Woolridge, op. cit.

Chapter 1: Introduction and Goals of the Firm 15

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Caveats to Maximizing Shareholder Value Managers should concentrate on maximizing shareholder value alone only if three con- ditions are met. These conditions require: (1) complete markets, (2) no significant asym- metric information, and (3) known recontracting costs. We now discuss how a violation of any of these conditions necessitates a much larger view of management’s role in firm decision making.

Complete Markets To directly influence a company’s cash flows, forward or futures markets as well as spot markets must be available for the firm’s inputs, output, and by- products. For example, forward and futures markets for crude oil and coffee bean inputs allow Texaco and Starbuck’s Coffeehouses to plan their costs with more accurate cash flow projections. For a small 3 to 5 percent fee known in advance, value-maximizing managers can lock in their input expense and avoid unexpected cost increases. This com- pleteness of the markets allows a reduction in the cost-covering prices of gasoline and cappuccino.

Example Tradable Pollution Permits at Duke Power15

By establishing a market system for tradable air pollution permits, the Clean Air Act set a price on the sulfur dioxide (SO2) by-product from burning high-sulfur coal. SO2 emissions from coal-fired power plants in the Midwest raised the acidity of rain and mist in eastern forests from Maine to Georgia to levels almost 100 times higher than the natural acidity of rainfall in the Grand Tetons in the far northwestern United States. Dead trees, peeling paint, increased asthma, and stone decomposition on buildings and monuments were the result.

To elicit substantial pollution abatement at the least cost, the Clean Air Act of 1990 authorized the Environmental Protection Agency to issue tradable pollution allowances (TPAs) to 467 known SO2 polluters for approximately 70 percent of the previous year’s emissions. The utility companies doing the polluting then began to trade the allowances. Companies that were able to abate their emissions at a low cost (perhaps because they had smokestack scrubbing equipment) sold their allow- ances to plants that couldn’t abate their emissions as cost effectively. In other words, the low-cost abaters were able to cut their emissions cheaply and then sell the permits they didn’t need to high-cost abaters. The result was that the nation’s air got 30 percent cleaner at the least possible cost.

As a result of the growing completeness of this market, electric utilities like Duke Power now know what expense line to incorporate in their cash flow projec- tions for the SO2 by-products of operating with high-sulfur coal. TPAs can sell for more than $100 per ton, and a single utility plant operation may require 15,000 tons of permits or more. The continuous tradeoff between installing 450- million-dollar pollution abatement equipment, utilizing higher-cost alternative fuels like low-sulfur coal and natural gas, or paying the current market price of these EPA-issued pollution permits can now be explicitly analyzed and the least- cost solutions found.

15Based on “Acid Rain: The Southern Company,” Harvard Business School Publishing, HBS: 9-792-060; “Cornering the Market,” Wall Street Journal (June 5, 1995), p. B1; and Economic Report of the President, February 2000 (Washing- ton, DC: U.S.G.P.O., 2000), pp. 240–264.

16 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

No Asymmetric Information Monitoring and coordination problems within the corporation and contracting problems between sellers and buyers often arise because of asymmetric information. Line managers and employees can misunderstand what senior executives intend and miscommunicate these intentions to customers. A Food Lion memo challenging employees to find a thousand different ways to save 1 percent of their own costs elicited undesirable shortcuts in food preparation and storage. Dianne Sawyer then secretly recorded seafood counter employees spraying old salmon with a light con- centration of ammonia to restore the red appearance of fresh fish. Clearly, this was not what the senior executives at Food Lion intended.

Building a good reputation with customers, workers, and the surrounding tax jurisdic- tion is one way companies deal with the problem of asymmetric information, and man- agers must attend to these reputational effects on shareholder value. We discuss the implications of asymmetric information in competitive markets in Chapter 10.

Known Recontracting Costs Finally, to focus exclusively on the discounted pres- ent value of future cash flows necessitates that managers obtain not only sales revenue and expense estimates but also forecasts of future recontracting costs for pivotal inputs. Owners of professional sports teams are acutely aware of how unknown recontracting costs with star players can affect the value of their franchises. The same thing can occur with a piv- otal corporate executive. A star CFO, COO, CMO, or CIO can often “hold up” the firm’s owners when the time comes for contract renewals. In another arena, Westinghouse en- tered into long-term supply contracts to provide fuel rods to nuclear power plants across the country. Thereafter, when the market price of uranium quadrupled, Westinghouse re- fused to deliver the promised fuel rods and recontracting costs skyrocketed. Value- maximizing managers must anticipate and mitigate these recontracting problems.

To the extent markets are incomplete, information is asymmetric, or recontracting costs are unknown, managers must attend to these matters in order to maximize share- holder wealth rather than simply focus myopically on maximizing the net present value of expected future cash flows.

Residual Claimants Why is it that the primary duty of management and the board of directors of a company is to the shareholders themselves? Shareholders have a residual claim on the firm’s net cash flows after all expected contractual returns have been paid. All the other stake- holders (employees, customers, bondholders, banks, suppliers, the surrounding tax juris- dictions, the community in which plants are located, etc.) have contractual expected returns. If expectations created by those contracts are not met, any of these stakeholders has access to the full force of the contract law in securing what they are due. Share- holders have contractual rights, too, but those rights simply entitle them to whatever is left over, that is, to the residual. As a consequence, when shareholder owners hire a CEO and a board, they create a fiduciary duty to allocate the company’s resources in such a way as to maximize the net present value of these residual claims. This is what consti- tutes the objective of shareholder wealth maximization.

Be very clear, however, that the value of any company’s stock is quite dependent on repu- tation effects. Underfunding a pension plan or polluting the environment results in massive losses of capitalized value because the financial markets anticipate (correctly) that such a com- pany will have reduced future cash flows to owners. Labor costs to attract new employees will rise; tax jurisdictions will reduce the tax preferences offered in new plant locations; customers may boycott; and the public relations, lobbying, and legal costs of such a company will surely rise. All this implies that wealth-maximizing managers must be very carefully attuned to stakeholder interests precisely because it is in their shareholders’ best interests to do so.

Chapter 1: Introduction and Goals of the Firm 17

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Goals in the Public Sector and Not-for-Profit Enterprises16 The value-maximization objective developed for private sector firms is not an appropri- ate objective in the public sector or in not-for-profit (NFP) organizations. These organi- zations pursue a different set of objectives because of the nature of the goods and services they supply and the manner in which they are funded.

There are three characteristics of NFP organizations that distinguish them from for- profit enterprises and influence their decision making. First, no one possesses a right to receive profit or surpluses in an NFP enterprise. The absence of a profit motive can have a serious impact on the incentive to be efficient. Second, NFP enterprises are exempt from taxes on corporate income. Finally, donations to NFPs are tax deductible, which gives NFP enterprises an advantage when competing for capital.

Not-for-profit organizations include performing arts groups, museums, libraries, hos- pitals, churches, volunteer organizations, cooperatives, credit unions, labor unions, pro- fessional societies, foundations, and fraternal organizations. Some of these organizations offer services to a group of clients, such as the patients of a hospital. Others provide ser- vices primarily to their members such as tennis clubs or credit unions. Finally, some NFP organizations produce products to benefit the general public. Local symphony and theater companies are examples.

Public sector (government) agencies tend to provide services that have significant public-good characteristics. In contrast to private goods, like a bite-sized candy bar, a public good can be consumed by more than one person. Moreover, excluding those who do not pay can only be done at a prohibitively high cost. Examples of public goods include national defense and flood control. If an antiballistic missile system or a flood control levy is constructed, no one can be excluded from its protection even if they refuse to contribute to the cost. Even if exclusion were feasible, the indivisibility of mis- sile defense or flood control consumption makes the incremental cost (and therefore the efficient price) of adding another participant quite low.

Some goods, such as recreational facilities and the performing arts, have both private- good and public-good characteristics. For example, concerts and parks may be shared (within limits) and are partially nonexcludable in the sense that they convey prestige and quality-of-life benefits to the entire community.17 The more costly the exclusion, the more likely the good or service will be provided by the public sector rather than the pri- vate sector. Portrait artists and personal fitness trainers offer pay-as-you-go private fee arrangements. Chamber music fans and tennis court users often organize in consumption-sharing and cost-sharing clubs. At the end of the spectrum, open-air sym- phony concerts and large parks usually necessitate some public financing.

Not-for-Profit Objectives Several organizational objectives have been suggested for the NFP enterprise. These in- clude the following:

1. Maximizing the quantity and quality of output subject to a break-even budget constraint.

2. Maximizing the outcomes preferred by the NFP’s contributors. 3. Maximizing the longevity of the NFP’s administrators.

16This section draws heavily on Burton A. Weisbrod, The Nonprofit Economy (Cambridge, MA: Harvard Uni- versity Press, 1988).

public goods Goods that may be consumed by more than one person at the same time with little or no extra cost, and for which it is expensive or impossible to exclude those who do not pay.

17William J. Baumol and W.G. Bowen, Performing Arts: The Economic Dilemma (Brookfield, VT: Ashgate Publishing Co., 1993).

18 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Efficiency Objective in Not-for-Profit Organizations Cost-benefit analysis has been developed to more efficiently allocate public and NFP resources among competing uses. Because government and NFP spending is normally constrained by a budget ceiling, the goals actually used in practice can be any one of the following:

1. Maximize the benefits for given costs. 2. Minimize the costs while achieving a fixed level of benefits. 3. Maximize the net benefits (benefits minus costs).

Cost-benefit analysis is only one factor in the final decision, however. It does not in- corporate many of the more subjective considerations or less easily quantifiable objec- tives, like how fair it might be. Such matters must be introduced at a later stage in the analysis, generally through the political process.

SUMMARY

� Managers are responsible for proactively solving problems in the current business model, for setting stretch goals, establishing the vision, and setting strategy for future business, for monitoring team- work, and integrating the operations, marketing, and finance functions.

� Economic profit is defined as the difference between total revenues and total economic costs. Economic costs include a normal rate of return on the capital contributed by the firm’s owners. Economic profits exist to compensate investors for the risk they assume, because of temporary disequilibrium conditions that may occur in a market, because of the existence of monopoly power, and as a reward to firms that are especially innovative or highly efficient.

� As an overall objective of the firm, the shareholder wealth-maximization model is flexible enough to account for differential levels of risk and timing differences in the receipt of benefits and the incur- ring of future costs. Shareholder wealth captures the net present value of future cash flows to owners from positive NPV projects plus the value of em- bedded real options.

� Managers may not always behave in a manner con- sistent with the shareholder wealth-maximization ob- jective. The agency costs associated with preventing or at least mitigating these deviations from the owner-principal’s objective are substantial.

� Changes in the firm’s performance, perhaps un- related to a manager’s effort, combined with the unobservable nature of their creative ingenuity pre- sents a difficult principal-agent problem to resolve. This combination makes it difficult for owner- principals to know when to blame manager- agents for weak performances versus giving them credit for strong performances.

� Shareholder wealth maximization implies forward- looking, long-run-oriented, dynamic strategies that anticipate change in a risky market environment. Managers can focus on maximizing the discounted present value of the firm’s cash flows if three con- ditions hold: complete markets, no asymmetric information, and known recontracting costs. Oth- erwise, they must attend to these complications as well.

� Governance mechanisms (including internal moni- toring by subcommittees appointed by boards of directors and large creditors, internal/external monitoring by large block shareholders, auditing and variance analysis) can be used to mitigate agency problems by limiting managerial discretion.

� Shareholder wealth maximization implies a firm should be forward-looking, dynamic, and have a long-term outlook; anticipate and manage change; acquire strategic investment opportunities; and maximize the present value of expected cash flows

cost-benefit analysis A resource-allocation model that can be used by public sector and not-for-profit organizations to evaluate programs or investments on the basis of the magnitude of the discounted costs and benefits.

Chapter 1: Introduction and Goals of the Firm 19

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

to owners within the boundaries of the statutory law, administrative law, and ethical standards of conduct.

� Shareholder wealth maximization will be difficult to achieve when firms suffer from problems related to incomplete markets, asymmetric information, and unknown recontracting costs. In the absence of these complications, managers should maximize the present value of the discounted future net cash flows to residual claimants—namely, equity own- ers. If any of the complicating factors is present, managers must first attend to those issues before attempting to maximize shareholder wealth.

� Not-for-profit enterprises exist to supply a good or service desired by their primary contributors.

� Public sector organizations often provide services having significant public-good characteristics. Pub- lic goods are goods that can be consumed by more than one person at a time with little additional cost, and for which excluding those who do not pay for the goods is exceptionally difficult or pro- hibitively expensive.

� Regardless of their specific objectives, both public and private institutions should seek to furnish their goods or services in the most efficient way, that is, at the least cost possible.

Exercises 1. One of the approaches for the Southern Company to comply with the Clean Air Act is to adopt fuel-switching technology. Do you think this strategic flexibility would have value to Southern Company’s shareholders? Why?

2. Explain several dimensions of the shareholder-principal conflict with manager- agents known as the principal-agent problem. To mitigate agency problems be- tween senior executives and shareholders, should the compensation committee of the board devote more to executive salary and bonus (cash compensation) or more to long-term incentives? Why? What role does each type of pay play in motivating managers?

3. Corporate profitability declined by 20 percent from 2008 to 2009. What perfor- mance percentage would you use to trigger executive bonuses for that year? Why? What issues would arise with hiring and retaining the best managers?

4. In the Southern Company Managerial Challenge, which alternative for complying with the Clean Air Act creates the greatest real option value? How exactly does that alternative save money? Why? Explain why installing a scrubber “burns” this option.

5. In 2006, firms in the drug industry earned an average return on net worth of 22 percent, compared with an average return of 14 percent earned by over 1,400 firms followed by Value Line. Which theory or theories of profit do you think best explain(s) the performance of the drug industry?

6. In the context of the shareholder wealth-maximization model of a firm, what is the expected impact of each of the following events on the value of the firm? Explain why. a. New foreign competitors enter the market. b. Strict pollution control requirements are enacted. c. A previously nonunion workforce votes to unionize. d. The rate of inflation increases substantially. e. A major technological breakthrough is achieved by the firm, reducing its

costs of production.

Answers to the exercises in blue can be found in Appendix D at the back

of the book.

20 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

7. In 2008–2009, the price of jet and diesel fuel used by air freight companies de- creased dramatically. As the CEO of FedEx, you have been presented with the fol- lowing proposals to deal with the situation: a. Reduce shipping rates to reflect the expense reduction. b. Increase the number of deliveries per day in some markets. c. Make long-term contracts to buy jet fuel and diesel at a fixed price for the

next two years and set shipping rates to a level that will cover these costs. Evaluate these alternatives in the context of the decision-making model presented in the text.

8. How would each of the following actions be expected to affect shareholder wealth? a. Southern Company adopts fuel-switching technology at its largest power

plants. b. Ford Motor Company pays $2.5 billion for Jaguar. c. General Motors offers large rebates to stimulate sales of its automobiles. d. Rising interest rates cause the required returns of shareholders to increase. e. Import restrictions are placed on the French competitors of Napa wineries. f. There is a sudden drop in the expected future rate of inflation. g. A new, labor-saving machine is purchased by Wonder Bread and results in

the layoff of 300 employees.

Case Exercises DESIGNING A MANAGERIAL INCENTIVES

CONTRACT Specific Electric Co. asks you to implement a pay-for-performance incentive contract for its new CEO. The CEO can either work really hard with a personal opportunity cost of $200,000 in reduced personal entrepreneurship and increased stress-related health care costs or she can reduce her effort, thereby avoiding the personal costs. The CEO faces three possible outcomes: the probability of her company experiencing good luck is 30 percent, medium luck is 40 percent, and bad luck is 30 percent. Al- though the management team can distinguish the three “states” of luck as the quarter unfolds, the Compensation Committee of the Board of Directors (and the share- holders) cannot do so. Once the board designs an incentive contract, the CEO decides to expend high or low work effort, and soon thereafter the good, medium, or bad luck occurs. One of the observable shareholder values listed below then results.

SHAREHOLDER VALUE

GOOD LUCK (30%)

MEDIUM LUCK (40%)

BAD LUCK (30%)

High CEO Effort $1,000,000,000 $800,000,000 $500,000,000

Low CEO Effort $ 800,000,000 $500,000,000 $300,000,000

Chapter 1: Introduction and Goals of the Firm 21

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Assume the company has 10 million shares outstanding offered at a $65 initial share price, implying a $650,000,000 initial shareholder value. Since the CEO’s effort and the company’s luck are unobservable to the owners and company directors, it is not possible when the company’s share price falls to $50 and the company’s value to $500,000,000 to distinguish whether the company experienced low CEO effort and medium luck or high CEO effort and bad luck. Similarly, it is not possible to distin- guish low CEO effort and good luck from high CEO effort and medium luck.

Answer the following questions from the perspective of a member of the Compen- sation Committee of the board of directors who is aligned with shareholders’ interests and is deciding on a performance-based pay plan (an “incentive contract”) for the CEO.

Questions 1. What is the maximum amount it would be worth to shareholders to elicit high

CEO effort all of the time rather than low CEO effort all of the time? 2. If you decide to pay 1 percent of this amount (in Question 1) as a cash bonus,

what performance level (what share price or shareholder value) in the table should trigger the bonus? Suppose you decide to elicit high CEO effort when, and if, medium luck occurs by paying a bonus should the company’s value rise to $800,000,000. What criticism can you see of this incentive contract plan?

3. Suppose you decide to elicit high CEO effort when, and if, good luck occurs by paying a bonus only for an increase in the company’s value to $1,000,000,000. What criticism can you see of this incentive contract plan?

4. Suppose you decide to elicit high CEO effort when, and if, bad luck occurs by paying the bonus when the company’s value falls to $500,000. What criticism can you see of this incentive contract plan?

5. In an effort to identify the share price that should trigger a bonus, the payment for the CEO, and maximize shareholder value, how much would you, the Com- pensation Committee, be willing to pay an auditor to examine the expense and revenue flows in real time and deliver perfect forecasting information about the “luck” the firm is experiencing? Compare shareholder value with this perfect in- formation relative to the best choice among the cash bonus plans in Questions 2, 3, and 4.

6. Design a stock option-based incentive plan to elicit high effort. Show that 1 mil- lion stock options at a $70 exercise price improves shareholder value relative to the best of the cash bonus plans in Questions 2, 3, or 4.

7. Design an incentive plan that seeks to elicit high effort by granting restricted stock. Show that one-half million shares granted at $70 improves shareholder value relative to all prior alternatives.

8. Financial audits are basically sampling procedures to verify with a predetermined accuracy the sources and uses of the company receipts and expenditures; the larger the sample, the higher the accuracy. What’s the maximum amount the Compensation Committee of the board will be willing to pay for a perfect forecast if it were possible for the auditors to distinguish good from medium luck? What about medium from bad luck?

22 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

SHAREHOLDER VALUE OF WIND POWER AT HYDRO CO.:18 RE < C Wind farms and massive solar collector arrays are spreading across the globe. Wind produces enough electricity today in the United States to completely power 2 million homes. Wind and solar energy together provide less than 1 percent of the electric power worldwide, but already much more in some locations—for example, 19 percent in Denmark and 15 percent in Germany. Hydro, a Norwegian aluminum company, has established wind turbine pilot projects where entire communities are electricity self-sufficient. At 80 meters of elevation, class 3 wind energy (steady 22 kph breeze) is available almost everywhere on the planet, implying wind power potential world- wide of 72 million megawatts. Harvesting just the best 5 percent of this wind energy (3.6 million megawatts) would make it possible to retire several thousand coal-fired power plants, 617 of which operate in the United States today.19 Britain’s 2008 Re- newable Energy Strategy calls for renewable energy to account for 47 percent of total electricity output by 2020, 19 percent from offshore and 13 percent from onshore wind power.

So-called “alternative energy” is: (1) renewable, (2) in abundant local supply, and (3) generates a low carbon footprint. Renewables are naturally replenishing sources including wind, solar, hydro, biofuel, biomass, geothermal, tidal, ocean current, and wave energy. Nuclear energy is not renewable because of the waste disposal issues. To date, by far the most successful renewables are hydroelectric power plants and ethanol-based biofuels, each accounting for about 2 percent of energy worldwide. New sources of renewable energy such as wind and solar power are often judged against fuel oil at $15, natural gas at $6, and coal at $4 per million BTUs (see Figure 1.4). One ton of plentiful high-sulfur-content coal generates approximately a mega- watt of electricity and a ton of carbon dioxide (CO2). In 2008, the European Union’s cap-and-trade legislation to reduce carbon emissions imposed a $23 per ton addi- tional CO2 emissions charge atop the $85 purchase price of coal. Finding renewable energy sources that have full costs lower than coal’s $23 + $85 = $108 for a megawatt hour (RE < C) is a reasonable objective of energy policy.20

Why pursue wind and solar power rather than other alternative energy sources? Nu- clear energy has a decades-long timeline for construction and permitting especially of nuclear waste disposal sites. Corn-based ethanol runs up the cost of animal feedstocks and raises food prices. In addition, corn contains only one-eighth the BTUs of sugar- cane, which is in abundant supply in the Caribbean and Brazil. Unfortunately, the U.S. Congress has placed a $0.54 per gallon tariff on sugarcane-based ethanol. Natural gas is 80 percent cleaner than coal and extraordinarily abundant in the United States, the world’s biggest energy user at 21 million barrels per day (mbd), 13 mbd being imported.

18Based on Frederick Harris, Alternative Energy Symposium, Wake Forest University (September 19, 2008). 19Older, smaller 500-megawatt coal-fired plants have adopted little pollution abatement technology. Nu- clear power plants are much larger, generating typically 2,000 megawatts of electricity. Duke Power’s Be- lews Creek plant at 2,200 megawatts is one of the largest coal-fired power plants in the United States (see Figure 1.1). Following the installation of a $450-million smokestack scrubber, it is also one of the cleanest. 20France has added another €17 ($24) per ton of CO2 emissions tax on households and businesses using coal-based and oil-based electricity. See “France Moves to Levy Carbon Tax on Fossil Fuels,” Wall Street Journal (September 11, 2009), p. A10.

Chapter 1: Introduction and Goals of the Firm 23

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The United States contains almost 30 percent of the known deposits worldwide of nat- ural gas (and coal) but only 3 percent of the proven reserves of crude oil.

A 0.6 megawatt wind turbine that costs $1.2 million today will generate $4.4 mil- lion in discounted net present value of electricity over a 15-year period, sufficient to power 440 Western European or American households with 100 percent capacity uti- lization and continuous 15 mph wind.21 Mechanical energy in the turbine is con- verted directly into electrical potential energy with a magnetic coil generator. When the wind does not blow, Hydro has demonstrated and patented a load-shifting tech- nology that consists of a hydrolysis electrolyzer splitting water into oxygen and hydro- gen, a hydrogen storage container, and a fuel cell to convert the hydrogen chemical energy back to electrical current (see Figure 1.5). With the three extra pieces of equip- ment, the capital investment rises from $1.2 million to $2.7 million. Even so, wind power can be quite profitable with full cost recovery periods as short as seven years under ideal operating conditions.

Of course, frequently the operating conditions with wind power are far less than ideal. Despite the presence of wind at elevation across the globe, few communities want 80+ meter wind turbines as tall as a football field in their backyard sight lines. Lower installations result in less wind and therefore less electricity. In addition, the conversion of one form of energy to another always burns energy. In Hydro’s load- shifting process of converting mechanical energy from the turbine to chemical energy in the electrolyzer and then to electrical energy in the hydrogen fuel cell, about 30 percent of the maximum energy coming directly to the electrical grid from the tur- bine’s generator when the wind is blowing hard and steady is lost. Experiments in many wind conditions at the Utsira site suggest that baseline output of Hydro’s pilot project in Norway has a maximum energy conversion factor (CF) of 70 percent with 60 percent more typical. Even lower 45 percent CFs are expected in typical operating conditions elsewhere. Seventy percent CF realizes $3.1 million of electricity.

FIGURE 1.4 RE < C? Renewable Energy Less Than Coal Cost?

1999 2001 2003 2005

Coal

Natural gas

USD price per million BTU

Fuel oil

2007 2009

5

10

15

20

Source: Thomson Datastream; U.S. Energy Information Administration.

21600,000 kilowatt hours × $0.11 average electricity rates × 24 hours × 365 days equals $578,160 per year for 15 years of expected working life of the turbine. Based on “Hydro: From Utsira to Future Energy Solu- tions,” Ivey School of Business, Case #906M44, 2006.

24 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Questions 1. Should Hydro as an aluminum producer invest in wind power in light of the

Utsira pilot project? Why or why not? 2. Should value-maximizing managers more generally invest in wind power? Why

or why not? 3. Larger-scale turbines increase the electricity more than proportionately to the in-

crease in costs. A 1 megawatt turbine costs $2.5 million, with the remaining equipment costs unchanged, for a total required investment of $4 million to power approximately 760 households. Electricity revenue over 15 years rises to $7.2 million in discounted present value. What conversion factor allows cost re- covery of this larger-scale turbine?

4. If the net present value of the Utsira project is negative, yet Hydro goes ahead and funds the investment anyway, what ethical obligations does Hydro have to its shareholders?

5. On what basis could shareholder value possibly rise if Hydro invests in wind power? Would more or less disclosure to financial analysts improve the chances of this outcome?

6. In 2009, 41 percent of all energy consumption in the United States comes from electric power generation. Coal provides the preponderant fuel (51 percent), with nuclear power (21 percent) and natural gas (17 percent) providing most of the rest. Renewable energy provides only 9 percent. Recently, T. Boone Pickens pro- posed converting the trucking fleet in the United States to liquefied natural gas (LNG) and using wind power to replace the missing LNG in electric power production. What issues do you see that must be resolved before the Pickens plan could be adopted?

FIGURE 1.5 Wind Turbine Cost Recovery: Wind-to-H2 Load-Shift Technology

Wind Turbine

(0.6 MWh)

Electrolyzer H2O → H2 + O

H2 Fuel Cell H2 + O → H2O

Hydro’s Patented

Control & Regulating

System

H2 Storage

Electric Power Grid

80%→ $3.5 mil 70% → $3.1 mil 60%→ $2.6 mil

$2.7 mil Investment

CF:

Chapter 1: Introduction and Goals of the Firm 25

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

2 CHAP T E R

Fundamental Economic Concepts CHAPTER PREVIEW A few fundamental microeconomic concepts provide cornerstones for all of the analysis in managerial economics. Four of the most important are demand and supply, marginal analysis, net present value, and the meaning and measurement of risk. We will first review how the determinants of demand and supply establish a market equilibrium price for gasoline, crude oil, and hybrid electric cars. Marginal analysis tools are central when a decision maker is seeking to optimize some objective, such as maximizing cost savings from changing a lightbulb (e.g., from normal incandescent to compact fluorescent [CFL]). The net present value concept makes directly comparable alternative cash flows occurring at different points in time. In so doing, it provides the linkage between the timing and risk of a firm’s projected profits and the shareholder wealth-maximization objective. Risk-return analysis is important to an understanding of the many trade-offs that managers must consider as they introduce new products, expand capacity, or outsource overseas in order to increase expected profits at the risk of greater variation in profits.

Two appendices elaborate these topics for those whowant to knowmore analytical details and seek exposure to additional application tools. Appendix C develops the relationship between marginal analysis and differential calculus. Web Appendix F shows how managers incorporate explicit probability information about the risk of various outcomes into individual choice models, decision trees, risk-adjusted discount rates, simulation analysis, and scenario planning.

MANAGERIAL CHALLENGE Why Charge $25 per Bag on Airline Flights?

In May 2008, American Airlines (AA) announced that it would immediately begin charging $25 per bag on all AA flights, not for extra luggage but for the first bag! Crude oil had doubled from $70 to $130 per barrel in the previ- ous 12 months, and jet fuel prices had accelerated even faster. AA’s new baggage policy applied to all ticketed passengers except first class and business class. On top of incremental airline charges for sandwiches and snacks

introduced the previous year, this new announcement stunned the travel public. Previously, only a few deep- discount U.S. carriers with very limited route structures such as People Express had charged separately for both food and baggage service. Since American Airlines and many other major carriers had belittled that policy as part of their overall marketing campaign against deep discounters, AA executives faced a dilemma.

26

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

DEMAND AND SUPPLY: A REVIEW Demand and supply simultaneously determine equilibrium market price (Peq). Peq equates the desired rate of purchase Qd/t with the planned rate of sale Qs/t. Both con- cepts address intentions—that is, purchase intentions and supply intentions. Demand is therefore a potential concept often distinguished from the transactional event of “units sold.” In that sense, demand is more like the potential sales concept of customer traffic than it is the accounting receivables concept of revenue from completing an actual sale. Analogously, supply is more like scenario planning for operations than it is like actual

Jet fuel surcharges had recovered the year-over-year average variable cost increase for jet fuel expenses, but incremental variable costs (the marginal cost) re- mained uncovered. A quick back-of-the-envelope calcu- lation outlines the problem. If total variable costs for a 500-mile flight on a 180-seat 737-800 rise from $22,000 in 2007 Q2 to $36,000 in 2008 Q2 because of $14,000 of additional fuel costs, then competitively priced carriers would seek to recover $14,000/180 = $78 per seat in jet fuel surcharges. The average variable cost rise of $78 would be added to the price for each fare class. For example, the $188 Super Saver airfare restricted to 14-day advance purchase and Saturday night stay overs would go up to $266. Class M airfares requiring 7-day advance purchase but no Saturday stay overs would rise from $289 to $367. Full coach economy airfares without purchase restrictions would rise from $419 to $497, and so on.

The problem was that by 2008 Q2, the marginal cost for jet fuel had risen to approximately $1 for each pound transported 500 miles. Carrying an additional 170-pound passenger in 2007 had resulted in $45 of additional fuel costs. By May 2008, the marginal fuel cost was $170 – $45 = $125 higher! So although the $78 fuel surcharge was offsetting the accounting expense increase when one averaged in cheaper earlier fuel pur- chases, additional current purchases were much more expensive. It was this much higher $170 marginal cost that managers realized they should focus upon in decid- ing upon incremental seat sales and deeply discounted prices.

And similarly, this marginal $1 per pound for 500 miles became the focus of attention in analyzing bag- gage cost. A first suitcase was traveling free under the prior baggage policy as long as it weighed less than 42 pounds. But that maximum allowed suitcase imposed $42 of marginal cost in May 2008. Therefore, in

mid-2008, American Airlines (and now other major car- riers) announced a $25 baggage fee for the first bag in order to cover the marginal cost of the representative suitcase on AA, which weighs 25.4 pounds.

Discussion Questions

� How should the airline respond when presented with an overweight bag (more than 42 pounds)?

� Explain whether or not each of the following should be considered a variable cost that in- creases with each additional airline seat sale: baggage costs, crew costs, commissions on ticket sales, airport parking costs, food costs, and additional fuel costs from passenger weight.

� If jet fuel prices reverse their upward trend and begin to decline, fuel surcharges based on av- erage variable cost will catch up with and sur- pass marginal costs. How should the airlines respond then?

MANAGERIAL CHALLENGE Continued

© AP

Im ag es /J ef f Ro be rs on

Chapter 2: Fundamental Economic Concepts 27

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

production, distribution, and delivery. In addition, supply and demand are explicitly rates per unit time period (e.g., autos per week at a Chevy dealership and the aggregate purchase intentions of the households in the surrounding target market). Hence, Peq is a market-clearing equilibrium concept, a price that equates the flow rates of intended pur- chase and planned sale.

When the order flow to buy at a given price (Qd/t) in Figure 2.1 just balances against the order flow to sell at that price (Qs/t), Peq has emerged, but what ultimately deter- mines this metric of “value” in a marketplace? Among the earliest answers can be found in the Aristotelian concept of intrinsic use value. Because diamonds secure marriage covenants and peace pacts between nations, they provide enormous use value and should therefore exhibit high market value. The problem with this theory of value taken alone arises when one considers cubic zirconium diamonds. No one other than a jewel mer- chant can distinguish the artificial cubic zirconium from the real thing, and therefore the intrinsic uses of both types are identical. Yet, cubic zirconium diamonds sell for many times less than natural stones of like grade and color. Why? One clue arose at the end of the Middle Ages, when Catholic monasteries produced beautiful hand- copied Bibles and sold them for huge sums (i.e., $22,000 in 2010 dollars) to other mon- asteries and the nobility. In 1455, Johannes Guttenberg offered a “mass produced” printed facsimile that could be put to exactly the same intrinsic use, and yet, the market value fell almost one-hundred-fold to $250 in 2010 dollars. Why?

Equilibrium market price results from the interaction of demanders and suppliers in- volved in an exchange. In addition to the use value demanders anticipate from a product, a supplier’s variable cost will also influence the market price observed. Ultimately, there- fore, what minimum asking price suppliers require to cover their variable costs is just as pivotal in determining value in exchange as what maximum offer price buyers are willing to pay. Guttenberg Bibles and cubic zirconium diamonds exchange in a marketplace at lower “value” not because they are intrinsically less useful than prior copies of the Bible

FIGURE 2.1 Demand and Supply Determine the Equilibrium Market Price

0

Equilibrium price ($/unit)

Peq

St

Dt

Planned rate of sale

Desired rate of purchase

Qdt = Q s t

Quantity (units/time)

28 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

or natural stones but simply because the bargain struck between buyers and sellers of these products will likely be negotiated down to a level that just covers their lower vari- able cost plus a small profit. Otherwise, preexisting competitors are likely to win the business by asking less.

Even when the cost of production is nearly identical and intrinsic use value is nearly identical, equilibrium market prices can still differ markedly. One additional determinant of value helps to explain why. Market value depends upon the relative scarcity of re- sources. Hardwoods are scarce in Japan but plentiful in Sweden. Even though the cost of timber cutting and sawmill planing is the same in both locations, hardwood trees have scarcity value as raw material in Japan that they do not have in Sweden where they are plentiful. To take another example, whale oil for use in lamps throughout the nineteenth and early twentieth centuries stayed at a nearly constant price until whale species began to be harvested at rates beyond their sustainable yield. As whale resources became scarcer, the whalers who expended no additional cost on better equipment or longer voyages came home with less oil from reduced catches. With less raw material on the market, the input price of whale oil rose quickly. Consequently, despite un- changed other costs of production, the scarcer input led to a higher final product price. Similar results occur in the commodity market for coffee beans or orange juice when climate changes or insect infestations in the tropics cause crop projections to decline and scarcity value to rise.

Example Discovery of Jojoba Bean Causes a Collapse of Whale Oil Lubricant Prices1

Until the last decade of the twentieth century, the best-known lubricant for high- friction machinery with repeated temperature extremes like fan blades in aircraft jet engines, contact surfaces in metal cutting tools, and gearboxes in auto transmis- sions was a naturally occurring substance—sperm whale oil. In the early 1970s, the United States placed sperm whales on the endangered species list and banned their harvest. With the increasing scarcity of whales, the world market price of whale oil lubricant approached $200 per quart. Research and development for synthetic oil substitutes tried again and again but failed to find a replacement. Finally, a Califor- nia scientist suggested the extract of the jojoba bean as a natural, environmentally friendly lubricant. The jojoba bean grows like a weed throughout the desert of the southwestern United States on wild trees that can be domesticated and cultivated to yield beans for up to 150 years.

After production ramped up from 150 tons in 1986 to 700 tons in 1995, solvent-extracted jojoba sold for $10 per quart. When tested in the laboratory, jojoba bean extract exhibits some lubrication properties that exceed those of whale oil (e.g., thermal stability over 400°F). Although 85 to 90 percent of jojoba bean output is used in the production of cosmetics, the confirmation of this plentiful substitute for high-friction lubricants caused a collapse in whale lubricant prices. Sperm whale lubricant has the same cost of production and the same use value as before the discovery of jojoba beans, but the scarcity value of the raw material in- put has declined tenfold. Consequently, a quart of sperm whale lubricant now sells for under $20 per quart.

1Based on “Jojoba Producers Form a Marketing Coop,” Chemical Marketing Reporter (January 8, 1995), p. 10.

Chapter 2: Fundamental Economic Concepts 29

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Diamond-Water Paradox and the Marginal Revolution So equilibrium price in a marketplace is related to (1) intrinsic use value, (2) production cost, and (3) input scarcity. In addition, however, most products and services have more than one use and more than one method of production. And often these differences re- late to how much or how often the product has already been consumed or produced. For example, the initial access to e-mail servers or the Internet for several hours per day is often essential to maintaining good communication with colleagues and business associ- ates. Additional access makes it possible to employ search engines such as Google for information related to a work assignment. Still more access affords an opportunity to meet friends in a chat room. Finally, some households might purchase even more hours of access on the chance that a desire to surf the Web would arise unexpectedly. Each of these uses has its own distinct value along a continuum starting with necessities and end- ing with frivolous non-essentials. Accordingly, what a customer will pay for another hour of Internet access depends on the incremental hour in question. The greater the utilization already, the lower the use value remaining.

This concept of amarginal use value that declines as the rate of consumption increases leads to a powerful insight about consumer behavior. The question was posed: “Why should something as essential to human life as water sell for low market prices while something as frivolous as cosmetic diamonds sell for high market prices?” The initial an- swer was that water is inexpensive to produce in most parts of the world while diamonds require difficult search and discovery, expensive mining, and extensive transportation and security expenses. In other words, diamonds cost more than water, so minimum asking prices of suppliers dictate the higher market value observed for diamonds. However, recall that supply is only one of what Alfred Marshall famously called “two blades of the scis- sors” representing demand and supply. You can stab with one blade but you can’t cut paper, and using supply alone, you can’t fully explain equilibrium market price.

The diamond-water paradox was therefore restated more narrowly: “Why should con- sumers bid low offer prices for something as essential as water while bidding high offer prices for something as frivolous as diamonds?” The resolution of this narrower paradox hinges on distinguishing marginal use value (marginal utility) from total use value (total utility). Clearly, in some circumstances and locales, the use value of water is enormous. At an oasis in the desert, water does prevent you from thirsting to death. And even in the typical city, the first couple of ounces of some liquid serve this same function, but that’s the first couple of ounces. The next couple of dozen gallons per day remain at high use value for drinking, flushing indoor plumbing, cooking, body washing, and so forth. Thereafter, water is used for clothes washing, landscape watering, car washing, and sundry lesser purposes. Indeed, if one asks the typical American household (which consumes 80–100 gallons per person per day) to identify its least valuable use of water each day, the answer may come back truly frivolous—perhaps something like the water that runs down the sink drain while brushing teeth. In other words, the marginal use value of water in most developed countries is the water that saves the consumer the in- convenience of turning the water taps (on and off) twice rather than just once. And it is this marginal use value at the relevant margin, not the total utility across all uses, that determines a typical water consumer’s meager willingness to pay.

Marginal Utility and Incremental Cost Simultaneously Determine Equilibrium Market Price Alfred Marshall had it right: demand and supply do simultaneously determine market equilibrium price. On the one hand, marginal utility determines the maximum offer

marginal use value The additional value of the consumption of one more unit; the greater the utilization already, the lower the use value remaining.

marginal utility The use value obtained from the last unit consumed.

30 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

price consumers are willing to pay for each additional unit of consumption on the de- mand side of the market. On the other hand, variable cost at the margin (an incremental cost concept sometimes referred to as “marginal cost”) determines the minimum asking price producers are willing to accept for each additional unit supplied. Water is both cheaper to produce and more frivolous than diamonds at the relevant margin, and hence water’s market equilibrium price is lower than that of diamonds. Figure 2.2 illustrates this concept of marginal use value for water varying from the absolutely essential first few ounces to the frivolous water left running while brushing one’s teeth.

At the same time, the marginal cost of producing water remains low throughout the 90- gallon range of a typical household’s consumption. In contrast, diamonds exhibit steeply rising marginal cost even at relatively small volume, and customers continue to employ cos- metic diamonds for highly valuable uses even out to the relevant margin (one to three car- ats) where typical households find their purchases occurring. Therefore, diamonds should trade for equilibrium market prices that exceed the equilibrium market price of water.

Individual and Market Demand Curves We have seen that the market-clearing equilibrium price (Peq) that sets the desired rate of purchase (Qd/t) equal to the planned rate of sale (Qs/t) is simultaneously both the maximum offer price demanders are willing to pay (the “offer”) and the minimum ask- ing price sellers are willing to accept (the “ask”). But what determines the desired rate of purchase Qd/t and planned rate of sales Qs/t? The demand schedule (sometimes called the “demand curve”) is the simplest form of the demand relationship. It is merely a list of prices and corresponding quantities of a commodity that would be demanded by some individual or group of individuals at uniform prices. Table 2.1 shows the demand schedule for regular-size pizzas at a Pizza Hut restaurant. This demand schedule

FIGURE 2.2 The Diamond-Water Paradox Resolved

Equilibrium price ($/unit)

Offer pricew = f(M.U.w)

Pd eq

Pw eq

Sdiamonds

Swater

Dwater

Ddiamonds

Quantity (gallons/day)

(carats/lifetime)

2 carats 90 gallons

Asking priced = g(M.C.d)

Chapter 2: Fundamental Economic Concepts 31

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

indicates that if the price were $9.00, customers would purchase 60 per night. Note that the lower the price, the greater the quantity that will be demanded. This is the strongest form of the law of demand—if a product or service is income superior, a household will always purchase more as the relative price declines.

The Demand Function The demand schedule (or curve) specifies the relationship between prices and quantity demanded, holding constant the influence of all other factors. A demand function speci- fies all these other factors that management will often consider, including the design and packaging of products, the amount and distribution of the firm’s advertising budget, the size of the sales force, promotional expenditures, the time period of adjustment for any price changes, and taxes or subsidies. As detailed in Table 2.2, the demand function for hybrid-electric or all-electric autos can be represented as

QD = f ðP, PS, PC , Y , A, AC , N , CP , PE , TA, T=S …Þ [2.1] where QD = quantity demanded of (e.g., Toyota Prius or Chevy Volt)

P = price of the good or service (the auto)

PS = price of substitute goods or services (e.g., the popular gasoline-powered Honda Accord or Chevy Malibu)

PC = price of complementary goods or services (replacement batteries)

Y = income of consumers

A = advertising and promotion expenditures by Toyota, Honda, and General Motors (GM)

AC = competitors’ advertising and promotion expenditures

N = size of the potential target market (demographic factors)

CP = consumer tastes and preferences for a “greener” form of transportation

PE = expected future price appreciation or depreciation of hybrid autos

TA = purchase adjustment time period

T/S = taxes or subsidies on hybrid autos

The demand schedule or demand curve merely deals with the price-quantity relation- ship itself. Changes in the price (P) of the good or service will result only in movement along the demand curve, whereas changes in any of the other demand determinants in the demand function (PS, PC, Y, A, AC, N, CP, PE, and so on) shift the demand curve. This is illustrated graphically in Figure 2.3. The initial demand relationship is line DD 0. If the

TABLE 2.1 SIMPLIFIED DEMAND SCHEDULE: PIZZA HUT RESTAURANT

PRICE OF PIZZA ($/UNIT)

QUANTITY OF PIZZAS SOLD (UNITS PER TIME PERIOD)

10 50

9 60

8 70

7 80

6 90

5 100

demand function A relationship between quantity demanded and all the determinants of demand.

substitute goods Alternative products whose demand increases when the price of the focal product rises.

complementary goods Complements in consumption whose demand decreases when the price of the focal product rises.

32 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIGURE 2.3 Shifts in Demand

Price ($/unit)

Quantity (units)

Q1 Q2 Q3 Q40

P1

P2

D

D�

D1

D2

D�2

D�1

TABLE 2.2 PARTIAL LIST OF FACTORS AFFECTING DEMAND

DEMAND FACTOR EXPECTED EFFECT

Increase (decrease) in price of substitute goodsa (PS) Increase (decrease) in demand (QD)

Increase (decrease) in price of complementary goodsb (PC) Decrease (increase) in QD

Increase (decrease) in consumer income levelsc (Y) Increase (decrease) in QD

Increase (decrease) in the amount of advertising and marketing expenditures (A)

Increase (decrease) in QD

Increase (decrease) in level of advertising and marketing by competitors (AC)

Decrease (increase) in QD

Increase (decrease) in population (N) Increase (decrease) in QD

Increase (decrease) in consumer preferences for the good or service (CP)

Increase (decrease) in QD

Expected future price increases (decreases) for the good (PE) Increase (decrease) in QD

Time period of adjustment increases (decreases) (TA) Increase (decrease) in QD

Taxes (subsidies) on the good increase (decrease) (T/S) Decrease (increase) in QD

aTwo goods are substitutes if an increase (decrease) in the price of Good 1 results in an increase (decrease) in the quantity demanded of Good 2, holding other factors constant, such as the price of Good 2, other prices, income, and so on, or vice versa. For example, margarine may be viewed as a rather good substitute for butter. As the price of butter increases, more people will decrease their con- sumption of butter and increase their consumption of margarine. bGoods that are used in conjunction with each other, either in production or consumption, are called complementary goods. For example, DVDs are used in conjunction with DVD players. An increase in the price of DVD players would have the effect of decreasing the demand for DVDs, ceteris paribus. In other words, two goods are complementary if a decrease in the price of Good 1 results in an in- crease in the quantity demanded of Good 2, ceteris paribus. Similarly, two goods are complements if an increase in the price of Good 1 results in a decrease in the quantity demanded of Good 2. cThe case of inferior goods—that is, those goods that are purchased in smaller total quantities as income levels rise—will be discussed in Chapter 3.

Chapter 2: Fundamental Economic Concepts 33

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

original price were P1, quantity Q1 would be demanded. If the price declined to P2, the quantity demanded would increase to Q2. If, however, changes occurred in the other deter- minants of demand, we would expect to have a shift in the entire demand curve. If, for ex- ample, a subsidy to hybrids were enacted, the new demand curve might become D1D 01. At any price, P1, along D1D 01, a greater quantity, Q3, will be demanded than at the same price before the subsidy on the original curve DD 0. Similarly, if the prices of substitute products such as the Honda Accord or Chevy Malibu were to decline sharply, the demand curve would shift downward and to the left. At any price, P1, along the new curve D2 0D2, a smal- ler quantity, Q4, would be demanded than at the same price on either DD 0 or D1D 01.

In summary, movement along a demand curve is often referred to as a change in the quantity demanded, while holding constant the effects of factors other than price that de- termine demand. In contrast, a shift of the entire demand curve is often referred to as a change in demand and is always caused by some demand determinant other than price.

Import-Export Traded Goods In addition to the previous determinants of demand, the demand for goods traded in for- eign markets is also influenced by external factors such as exchange rate fluctuations. When Microsoft sells computer software overseas, it prefers to be paid in U.S. dollars. This is because a company like Microsoft incurs few offshore expenses beyond advertising and therefore cannot simply match payables and receivables in a foreign currency. To ac- cept euros, Japanese yen, or Australian dollars in payment for software purchase orders would introduce an exchange rate risk exposure for which Microsoft would want to be compensated in the form of higher prices on its software. Consequently, the foreign ex- ports of Microsoft are typically transacted in U.S. dollars and are therefore tied inextricably to the price of the dollar against other currencies. As the value of the dollar rises, offshore buyers must pay a larger amount of their own currency to obtain the U.S. dollars required to complete a purchase order for Microsoft’s software, and this decreases the export demand. Even in a large domestic market like the United States, companies often find that these export demand considerations are key determinants of their overall demand.

Example Exchange Rate Impacts on Demand: Cummins Engine Company Cummins Engine Company of Columbus, Indiana, is the largest independent man- ufacturer of new and replacement diesel engines for heavy trucks and for construc- tion, mining, and agricultural machinery. Volvo and Daimler-Benz are their major competitors, and 53 percent of sales occur offshore. The Cummins and Daimler- Benz large diesel truck engines sell for approximately $40,000 and €35,000, respec- tively. In the 2002 recession, Cummins suffered substantial declines in cash flow. One reason was obvious: diesel replacement engines are not needed when fewer goods are being delivered, and therefore fewer diesels are wearing out.

In addition, however, between 1999 and 2002, the value of the U.S. dollar (€ per $) increased by 30 percent from €.85/$ to €1.12/$. This meant that a $40,000 Cummins diesel engine that had sold for €34,000 in Munich in 1999 became €44,800, whereas the €35,000 Mercedes diesel alternative that had been selling for $41,176 in Detroit declined to $31,250 because of the stronger U.S. dollar. Cummins faced two unattrac- tive options, either of which would reduce its cash flow. It could either cut its profit margins and maintain unit sales, or maintain margins but have both offshore and

(Continued)

34 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Individual and Market Supply Curves What determines the planned rate of sale Qs/t? Like the demand schedule, the supply schedule is a list of prices and corresponding quantities that an individual or group of sellers desires to sell at uniform prices, holding constant the influence of all other factors. A number of these other determinants of supply that management will often need to consider are detailed in Table 2.3. The supply function can be represented as

QS = f ðP, PI , PUI , T , EE, F, RC, PE, T=S … Þ [2.2]

where Qs = quantity supplied (e.g., of domestic autos)

P = price of the autos

PI = price of inputs (e.g., sheet metal)

PUI = price of unused substitute inputs (e.g., fiberglass)

T = technological improvements (e.g., robotic welding)

EE = entry or exit of other auto sellers

F = accidental supply interruptions from fires, floods, etc.

RC = costs of regulatory compliance

PE = expected (future) changes in price

TA = adjustment time period

T/S = taxes or subsidies

TABLE 2.3 PARTIAL LIST OF FACTORS AFFECTING SUPPLY

SUPPLY FACTOR EXPECTED EFFECT AT EVERY PRICE

Increase (decrease) in the price of inputs (PI) Decrease (increase) in supply

Increase (decrease) in the price of unused substitute inputs (PUI) Decrease (increase) in supply

Technological improvements (T) Increase in supply

Entry (Exit) of other sellers (EE) Increase (decrease) in supply

Supply disruptions (F) Decrease in supply

Increase (decrease) in regulatory costs (RC) Decrease (increase) in supply

Expected future price increases (decreases) (PE) Decrease (increase) in supply

Time period of adjustment lengthens (shortens) (TA) Increase (decrease) in supply

Taxes (subsidies) (T/S) Decrease (increase) in supply

domestic sales collapse. The company chose to cut margins and maintain sales. By 2005, the dollar’s value had eroded, returning to €.85/$, and Cummins’ sales perfor- mance markedly improved. In the interim, demand for Cummins engines was adversely affected by the temporary appreciation of the U.S. dollar.

In 2009, with the U.S. dollar at a still lower value of €.64/$, the Cummins Engine Co. could barely keep up with export demand since diesels to Europe were priced at €25,600 versus Mercedes’ €32,000. Similarly, in Cleveland, St. Louis, and Atlanta, Cummins $40,000 diesels were up against $54,688 Mercedes substitutes. What a great time to be an American company competing against European manufacturers.

supply function A relationship between quantity supplied and all the determinants of supply.

Chapter 2: Fundamental Economic Concepts 35

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Again, changes in the price (P) of the good or service will result only in movement along the given supply curve, whereas changes in any of the other independent variables (PS, PC, Y, A, AC, N, CP, PE, and so on) in the function shift the supply curve. As with demand, a movement along a supply curve is referred to as a change in the quantity sup- plied, while holding constant other determinants of supply. A shift of the entire supply curve is often referred to as a change in supply and is always caused by some supply determinant other than price.

Equilibrium Market Price of Gasoline In April–July 2008, Americans woke up to a new reality about gasoline that markedly affected their driving habits as well as U.S. public policy. The price of a gallon of regular octane gasoline skyrocketed from $3.00 per gallon to $4.10 (see Figure 2.4). The previous summer, when gas prices had hovered around $3 per gallon, Americans had cut back only slightly on non-essential driving.

In the summer of 2008, with regular gasoline at $4.10 per gallon, not only summer driving vacations but urban commuting itself changed in extraordinary ways. Overall, customer de- mand by the typical two-person urban household shrank from 16 gallons per week to 11.5 gallons. As a result, for the first time in U.S. history, gasoline expenditure by U.S. households declined despite a rising price at the pump—that is, 16 gallons/week at $3 in 2007 (Q3) = $48 > 11.5 gallons per week at $4.10 in 2008 (Q3) = $47.15.

Several determinants of demand and supply were identified as possible explanations for the spike in gasoline’s equilibrium market price. First, much was written about the fact that no new refinery had been built in the United States in more than 30 years, sug- gesting that refinery capacity shortages or pipeline bottlenecks might be responsible. De- clining capacity does shift the supply curve in Figure 2.2 to the left, which would imply a higher equilibrium price. But no refinery closings or pipeline disruptions could be iden- tified that summer. And the U.S. Department of Energy found refineries command only $0.36 per gallon of the final product price of gasoline for cost recovery plus profit and

Example NAFTA and the Reduced Labor Costs of Ford Assembly Plants in Detroit The North American Free Trade Agreement (NAFTA) made it possible to buy subassemblies like axles and engine blocks from Mexican suppliers like Cifunsa, SA, without paying any import tariff when the parts arrived in the United States. Since United Auto Worker (UAW) labor in Detroit auto assembly plants also makes axle subassemblies, the Mexican labor input can be thought about as an unused substitute input from the point of view of Ford Motor Company. NAFTA in effect lowered the input cost of substitute inputs for Ford. This means fewer employers would pursue labor contracts with UAW labor in Detroit and instead shift some of their production south across the Mexican border. Less demand implies lower equilibrium wages would be offered and accepted by UAW assembly line labor. Hence, the indirect effect of NAFTA was a reduction in the input costs for UAW labor that the Ford Motor Co. did utilize. As usual, lower input cost implies a shift of the supply curve down and to the right, an increase in supply.

supply curve A relationship between price and quantity supplied, holding other determinants of supply constant.

36 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

could not therefore be responsible for the $1.10 increase in the equilibrium price between July 2007 and July 2008.

Second, retail gas station owners were accused of gouging the driving public. Higher markups at retail also would shift the supply curve for gasoline back to the left, raising the equilibrium market price. But again, retail markup and indeed all gasoline marketing were found to add only $0.28 per gallon to the $4.10 price, much less than could be re- sponsible for the $1.10 run-up in gasoline’s equilibrium market price. Third, excise taxes on gasoline (earmarked for road building and maintenance) are levied by both the federal and state governments. Gasoline taxes constitute $0.41 per gallon on average across the United States. Any new excise taxes would have shifted the supply curve leftward, result- ing in a higher equilibrium market price for gasoline. President George Bush’s Council of Economic Advisors in 2007 did explore levying an additional $1 per gallon tax on gaso- line to reduce the dependence of the United States on foreign oil, but no tax increase was ever initiated. So what was responsible for the upward spike in gasoline prices?

As we have seen, the variables in the demand and supply functions in Equations 2.1 and 2.2 determining equilibrium market price may be grouped into three broad sets of factors affecting use value, cost of production, and resource scarcity.2 Since crude oil inputs account for $2.96 of the $4.10 final product price of gasoline, resource scar- city was a likely candidate to explain the increase in gasoline prices from $3 to $4.10. Higher crude oil input prices shift the supply curve leftward, leading to higher final product prices for gasoline. Figure 2.5 shows that the previous three times crude oil input prices shot up, supply disruptions in the crude oil input market were involved (i.e., during the first Gulf War in Kuwait in 1991, during an especially effective era for the OPEC cartel 1999–2001, and during the Iraq War in 2004).

In contrast, the crude oil input price rise from $40 to $80 per barrel in 2006–2007 reflected demand-side increased usage especially by India and China. India and China are only 9 percent of the 85 million barrels per day (mbd) worldwide crude oil market but these two countries have been growing very quickly. A 2 to 3 percent additional

FIGURE 2.4 Average Gas Prices in the United States

2005

0.50

1.00

1.50

2.00

2.50

3.00

3.50

$4.00

$2.90 $2.80 $3.00

$4.10

2006 2007 2008

Source: AAA Carolinas.

2Two additional factors are speculation and government intervention in the form of taxes, subsidies, and regulations.

Chapter 2: Fundamental Economic Concepts 37

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

demand can significantly raise equilibrium prices for crude oil resources because at any point in time there is a very thin inventory (8–10 days supply) working its way through the distribution network from wells to pumps to terminals to tankers to refineries. By late 2007, crude oil input prices were rising beyond $80 per barrel. As gasoline headed toward $4.10 per gallon in the United States, $9.16 per gallon in Germany, and $8.80 per gallon in Great Britain, Western drivers substantially cut back consumption. Brazil approached $6.40 per gallon and pursued a successful energy independence campaign focused on sugar cane-based ethanol plants.

Was the $80 price in late 2007 the highest price ever in the crude oil input market prior to that time? The answer is “no.” In 1981, the equilibrium crude oil price reached $36 per barrel. Using the U.S. consumer price index (CPI), since crude oil transactions worldwide are denominated in U.S. dollars, cumulative price increases between 1981 and 2007 total 228.8 percent, so $36 × a 2.288 inflation-adjustment multiplier equals $82 in 2007, and $80/2.288 equals $35 in 1981. Consequently, the $80 crude oil price in late 2007 was in fact lower than the inflation-adjusted $36 crude price in 1981 at the height of the influence of the OPEC II oil cartel. However, in early 2008, the equilibrium price of crude continued to spike upward.

When the crude price climbed above $100, large numbers of speculators acquired long positions in the crude oil futures market betting on a further price rise. Speculative

FIGURE 2.5 Supply Disruptions and Developing Country Demand Fuel Crude Oil Price Spikes

1990

Gulf War OPEC III

Iraq War

Indian, Chinese demand

1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

50

60

70

80

90

100

110

120

130

40

30

20

10

0

O il

pr ic

e, in

c on

st an

t 20

00 U

.S .$

1990–2010 Real price per barrel, mean (standard deviation)

Mean +/– 2 Standard deviations

Source: Federal Reserve Bank, St. Louis, National Economics Trends, September 2000; FedDallas, Regional Economic Data, 2006.

38 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

demand (supply) is always motivated by the anticipation of equilibrium market prices being higher (lower) tomorrow. Those who “go long” and buy futures contracts to take delivery at prices agreed on today are betting the price will go up, and those who “sell short” and write futures contracts promising to deliver in the future at prices agreed on today are betting the other way. The net long direction of speculative trading in the first half of 2008 added to the growing market demand from India and China and drove the crude oil equilibrium price still higher, eventually reaching $146 per barrel in July 2008.

Faced with $4.10 per gallon gasoline, as ExxonMobil and Shell sought to recover their extraordinary input costs for crude, American consumers decided to vacate their SUVs, join carpools, and ride the buses and trains to work. Urban mass transit system ridership shot up 20 percent in a matter of months. Other Americans purchased fuel-efficient hy- brids like the Toyota Prius. Still others mobilized behind T. Boone Pickens’s plan to con- vert the federal trucking fleet to natural gas. Fearing an onslaught of feasible substitutes like hybrid electric cars and natural gas-powered trucks, the Saudis ramped up crude oil production from their average 8.5 mbd 1990–2006 all the way to 10.5 and 10.9 mbd in 2007 and 2008 (see Figure 2.6).

FIGURE 2.6 Saudi Arabia Crude Oil Production

1970

M ill

io ns

o f

ba rr

el /d

ay

10.9

8.5

1975 1980 1985 1990 1995 2000 2005 2010

Source: U.S. Energy Information Administration.

Chapter 2: Fundamental Economic Concepts 39

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

With U.S. demand for gasoline declining and capacity to extract and refine expand- ing, the equilibrium price of crude finally turned and began to decline. The late 2008 crude oil price reversal was caused by a combination of increasing supply fundamentals (shifting the supply curve to the right), slowing demand growth, and a speculative expec- tation that in the near term crude prices would be lower (not higher). Consequently, the supply of crude oil (and especially of highly leveraged crude oil futures contracts) mush- roomed. Angola doubled production capacity to 2.1 mbd, and Saudi capacity grew to 12.5 mbd. Saudi Arabia and Kuwait also broke ground on two giant new refining facilities.

Example Speculation Sends Crude Oil Input Price on a Roller-Coaster Ride at ExxonMobil and Shell With reversed expectations of lower crude prices in the near term, the speculative bubble in crude oil quickly burst. Despite 5 percent higher market demand over the last four months of 2008 (again primarily from China and India), the equilibrium price of crude oil plummeted more than $100 a barrel from $146 in September 2008 to a low of $40 by January 2009 (see Figure 2.7). By 2009 (Q3), the crude price stood again at $75 per barrel, and gasoline was selling for $2.74 per gallon. Although North American import demand for crude oil has been flat in recent years, OPEC members clearly believe that the spectacular 22 percent demand growth from Asian developing countries in 2000–2008 will continue. Over a two- year period, rising Asian demand, massive capacity expansions, a worldwide finan- cial boom, then collapse, and speculative buying followed by speculative selling had taken oil companies and gasoline buyers on quite a roller-coaster ride.

FIGURE 2.7 Crude Oil Price, West Texas Intermediate

1999 2001 2003 2005 2007 2009 0

25

$ pe

r ba

rr el

50

75

100

125

150

Source: Thomson Datasteam.

40 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

MARGINAL ANALYSIS Marginal analysis is one of the most useful concepts in microeconomics. Resource- allocation decisions typically are expressed in terms of the marginal equilibrium conditions that must be satisfied to attain an optimal solution. The familiar profit- maximization rule for the firm of setting output at the point where “marginal cost equals marginal revenue” is one such example. Long-term investment decisions (capital expen- ditures) also are made using marginal analysis decision rules. Only if the expected return from an investment project (that is, the marginal return to the firm) exceeds the cost of funds that must be acquired to finance the project (the marginal cost of capital), should the project be undertaken. Following this important marginal decision rule leads to the maximization of shareholder wealth.

More generally, a change in the level of an economic activity is desirable if the mar- ginal benefits exceed the marginal (that is, the incremental) costs. If we define net mar- ginal return as the difference between marginal benefits and marginal costs, then an equivalent optimality condition is that the level of the activity should be increased to the point where the net marginal return is zero.

In summary, marginal analysis instructs decision makers to determine the additional (marginal) costs and additional (marginal) benefits associated with a proposed action. Only if the marginal benefits exceed the marginal costs (that is, if net marginal benefits are positive) should the action be taken.

Total, Marginal, and Average Relationships Revenue, cost, profit, and many other economic relationships can be presented using tab- ular, graphic, and algebraic frameworks. Let us first use a tabular presentation. Suppose

Example Tenneco Shipyard Marginal Analysis Resource-allocation decisions should be made by comparing the marginal (or incremental) benefits of a change in the level of an activity with the incremental costs of the change. For example, the marginal revenue benefit derived from producing and selling one more supertanker is equal to the difference between total revenue, assuming the additional unit is not sold, and total revenue includ- ing the additional sale. Similarly, marginal cost is defined as the change in total costs that occurs from undertaking some economic activity, such as the produc- tion of an additional ship design including the opportunity costs, and therefore may not necessarily always be equal to the cash outlays alone. Perhaps the Ten- neco design team has an opportunity for higher net profit as subcontractors on Boeing projects. If so, Tenneco’s routine ship-design work should be contracted out to other shipbuilding design firms who can become a trusted subcontractor to Tenneco.

marginal analysis A basis for making various economic decisions that analyzes the additional (marginal) benefits derived from a particular decision and compares them with the additional (marginal) costs incurred.

Chapter 2: Fundamental Economic Concepts 41

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Marginal Analysis and Capital Budgeting Decisions: Sara Lee Corporation The capital budgeting decision problem facing a typical firm, such as Sara Lee Cor- poration, can be used to illustrate the application of marginal analysis decision rules. Sara Lee has the following schedule of potential investment projects (all assumed to be of equal risk) available to it:

PROJECT

INVESTMENT REQUIRED ($ MILLION)

EXPECTED RATE OF RETURN

CUMULATIVE INVESTMENT ($ MILLION)

A $25.0 27.0% $ 25.0

B 15.0 24.0 40.0

C 40.0 21.0 80.0

D 35.0 18.0 115.0

E 12.0 15.0 127.0

F 20.0 14.0 147.0

G 18.0 13.0 165.0

H 13.0 11.0 178.0

I 7.0 8.0 185.0

Sara Lee has estimated the cost of acquiring the funds needed to finance these investment projects as follows:

BLOCK OF FUNDS ($ MILLION)

COST OF CAPITAL

CUMULATIVE FUNDS RAISED ($ MILLION)

First $50.0 10.0% $ 50.0

Next 25.0 10.5 75.0

Next 40.0 11.0 115.0

Next 50.0 12.2 165.0

Next 20.0 14.5 185.0

The expected rate of return on the projects listed above can be thought of as the marginal (or incremental) return available to Sara Lee as it undertakes each addi- tional investment project. Similarly, the cost-of-capital schedule may be thought of as the incremental cost of acquiring the needed funds. Following the marginal analysis rules means that Sara Lee should invest in additional projects as long as the expected rate of return on the project exceeds the marginal cost of capital funds needed to finance the project.

Project A, which offers an expected return of 27 percent and requires an out- lay of $25 million, is acceptable because the marginal return exceeds the mar- ginal cost of capital (10.0 percent for the first $50 million of funds raised by Sara Lee). In fact, an examination of the tables indicates that projects A through G all meet the marginal analysis test because the marginal return from each of these projects exceeds the marginal cost of capital funds needed to finance these projects. In contrast, projects H and I should not be undertaken because they offer returns of 11 percent and 8 percent, respectively, compared with a marginal cost of capital of 14.5 percent for the $20 million in funds needed to finance those projects.

42 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

that the total profit πT of a firm is a function of the number of units of output produced Q, as shown in columns 1 and 2 of Table 2.4.

Marginal profit, which represents the change in total profit resulting from a one-unit increase in output, is shown in column 3 of the table. (A Δ is used to represent a “change” in some variable.) The marginal profit Δπ(Q) of any level of output Q is calcu- lated by taking the difference between the total profit at this level πT(Q) and at one unit below this level πT(Q − 1).

3 In comparing the marginal and total profit functions, we

Example Marginal Analysis of Driving a Mini Cooper versus a Chevy Volt Urban sprawl and flight to the suburbs have now resulted in the mean commuter trip in the United States rising to 33 miles one way. With the housing density in most American cities well below what would be required to support extensive light rail and subway lines, the typical household must find economical ways to get at least one worker from a suburban home to the central business district and back each day. A fuel-efficient, small commuter car like the Mini Cooper is one alterna- tive. Others have recently been proposed—the Chevy Volt and Nissan Leaf, both all-electric vehicles that are recharged at the end of each 40-mile commuting trip. Technically, the Leaf and the Volt are e-REVs, extended-range electric vehicles. Each contains a small gasoline-driven internal combustion engine that runs an electric generator, but unlike hybrids such as the Ford Fusion and Toyota Prius, these e-REVs have no mechanical connection between the gasoline engine and the drivetrain. Instead, the Chevy Volt goes 40 miles on the charge contained in 220 lithium ion (L-ion) batteries which are plugged in for a recharging cycle of 8 hours at 220 volts (or 3 hours at 110 volts) at work and at home. When the battery pack falls to a 30 percent state of charge (SOC), the gasoline engine comes on to turn the generator and maintain battery power above 25 percent SOC.

Automotive engineers calculate that each mile traveled in the Chevy Volt’s all- electric mode “burns” 0.26 kilowatt hours of electricity. So, the mean commuter trip of 33 miles requires 8.58 kWh of electricity. The price of electricity in the United States varies from a peak period in the afternoon and evening to a much cheaper off- peak period late at night, and from a low of $0.07 per kWh in Washington state to $0.12 in Rhode Island. On average, a representative nighttime rate is $0.10, and a representative daytime rate is $0.13. This means that each nighttime charge will run the household $0.86, and the comparable daytime charge downtown at work will be $1.12 for a total operating cost per day of just under $2. For 300 days of work, that’s $600 per year. In contrast, the gasoline-powered Mini Cooper gets 32 mpg, so at $3.00 per gallon, the Mini’s operating cost is approximately $6 per day or $1,800 per year. The typical commuter use of e-Rev vehicles will save $4 per day or $1,200 per year relative to popular fuel-efficient gasoline-powered cars.

At an EPA-measured 41 mpg throughout a range of driving conditions, the hybrid-electric Ford Fusion qualifies for a federal tax credit of $3,400. In contrast, at an EPA-measured 238 mpg, the Chevy Volt qualifies for a $7,500 tax credit to offset the $12,000 additional cost of the L-ion battery pack over the cost of a con- ventional battery. Because the Chevy Volt’s battery pack is expected to last 10 years, the $1,200 annual capital cost for the battery pack is equal to the $1,200 energy cost savings even without the federal tax credit.

3Web Appendix A expands upon the idea that the total profit function can be maximized by identifying the level of activity at which the marginal profit function goes to zero.

Chapter 2: Fundamental Economic Concepts 43

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

note that for increasing output levels, the marginal profit values remain positive as long as the total profit function is increasing. Only when the total profit function begins de- creasing—that is, at Q = 10 units—does the marginal profit become negative. The average profit function values πA(Q), shown in column 4 of Table 2.4, are obtained by dividing the total profit figure πT(Q) by the output level Q. In comparing the marginal and the average profit function values, we see that the average profit function πA(Q) is increasing as long as the marginal profit is greater than the average profit—that is, up to Q = 7 units. Beyond an output level of Q = 7 units, the marginal profit is less than the average profit and the average profit function values are decreasing.

By examining the total profit function πT(Q) in Table 2.4, we see that profit is maxi- mized at an output level of Q = 9 units. Given that the objective is to maximize total profit, then the optimal output decision would be to produce and sell 9 units. If the mar- ginal analysis decision rule discussed earlier in this section is used, the same (optimal) decision is obtained. Applying the rule to this problem, the firm would expand produc- tion as long as the net marginal return—that is, marginal revenue minus marginal cost (marginal profit)—is positive. From column 3 of Table 2.4, we can see that the marginal profit is positive for output levels up to Q = 9. Therefore, the marginal profit decision rule would indicate that 9 units should be produced—the same decision that was ob- tained from the total profit function.

The relationships among the total, marginal, and average profit functions and the optimal output decision also can be represented graphically. A set of continuous profit functions, analogous to those presented in Table 2.4 for discrete integer values of out- put (Q), is shown in Figure 2.8. At the break-even output level Q1, both total profits and average profits are zero. The marginal profit function, which equals the slope of the total profit function, takes on its maximum value at an output of Q2 units. This point corresponds to the inflection point. Below the inflection point, total profits are increasing at an increasing rate, and hence marginal profits are increasing. Above the inflection point, up to an output level Q4, total profits are increasing at a decreasing rate, and consequently marginal profits are decreasing. The average profit function, which represents the slope of a straight line drawn from the origin 0 to each point on

TABLE 2.4 TOTAL, MARGINAL, AND AVERAGE PROFIT RELATIONSHIPS

(1) (2) (3) (4)

NUMBER OF UNITS OF OUTPUT PER UNIT OF

TIME Q TOTAL PROFIT

πT (Q) ($)

MARGINAL PROFIT Δπ(Q) = πT (Q) − πT (Q − 1)

($/UNIT)

AVERAGE PROFIT πA(Q) = πT(Q) /Q

($/UNIT)

0 −200 0 —

1 −150 50 −150.00

2 −25 125 −12.50

3 200 225 66.67

4 475 275 118.75

5 775 300 155.00

6 1,075 300 179.17

7 1,325 250 189.29

8 1,475 150 184.38

9 1,500 25 166.67

10 1,350 −150 135.00

44 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the total profit function, takes on its maximum value at an output of Q3 units. The average profit necessarily equals the marginal profit at this point. This follows because the slope of the 0A line, which defines the average profit, is also equal to the slope of the total profit function at point A, which defines the marginal profit. Finally, total profit is maximized at an output of Q4 units where marginal profit equals 0. Beyond Q4 the total profit function is decreasing, and consequently the marginal profit function takes on negative values.

THE NET PRESENT VALUE CONCEPT When costs and benefits occur at approximately the same time, the marginal decision rule (proceed with the action if marginal benefit exceeds marginal cost) applies. But, many economic decisions require that costs be incurred immediately to capture a stream of benefits over several future time periods. In these cases, the net present value (NPV) rule replaces the marginal decision rule and provides appropriate guidance for longer-term decision makers. The NPV of an investment represents the contribution of that investment to the value of the firm and, accordingly, to shareholder wealth maximization.

FIGURE 2.8 Total, Average, and Marginal Profit Functions

Total profit ($) (πT(Q))

Inflection point

A

Maximum total profit

Total profit πT(Q)

Break-even point

0

Average profit (πA(Q)) Marginal profit (Δπ(Q)) ($/unit)

Units of output (Q)

Units of output (Q)Q1 Q2 Q3 Q4

Marginal profit Δπ(Q)

Average profit πA(Q)

Maximum average profit point

0

Maximum marginal profit point

Chapter 2: Fundamental Economic Concepts 45

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Determining the Net Present Value of an Investment To understand the NPV rule, consider the following situation. You are responsible for investing $1 million to support the retirement of several family members. Your financial advisor has suggested that you use these funds to purchase a piece of land near a proposed new highway interchange. A trustworthy state road commissioner is certain that the interchange will be built and that in one year the value of this land will increase to $1.2 million. Hence, you believe initially that this is a riskless investment. At the end of one year you plan to sell the land. You are being asked to invest $1 million today in the anticipation of receiving $1.2 million a year from today, or a profit of $200,000. You wonder whether this profit represents a sufficient return on your investment.

You feel it is important to recognize that a return of $1.2 million received one year from today must be worth less than $1.2 million today because you could invest your $1 million today to earn interest over the coming year. Therefore, to compare a dollar received in the future with a dollar in hand today, it is necessary to multiply the future dollar by a discount factor that reflects the alternative investment opportunities that are available.

Instead of investing $1 million in the land venture, you are aware that you could also invest in a one-year U.S. government bond that currently offers a return of 3 percent. The 3 percent return represents the return (the opportunity cost) forgone by investing in the land project. The 3 percent rate also can be thought of as the compensation to an investor who agrees to postpone receiving a cash return for one year. The discount factor, also called a present value interest factor (PVIF), is equal to

PVIF = 1

1 + i

where i is the compensation for postponing receipt of a cash return for one year. The present value (PV0) of an amount received one year in the future (FV1) is equal to that amount times the discount factor, or

PV0 = FV1 × ðPVIFÞ [2.3]

In the case of the land project, the present value of the promised $1.2 million expected to be received in one year is equal to

PV0 = $1:2million 1

1 + 0:03

� � = $1,165,049

If you invested $1,165,049 today to earn 3 percent for the coming year, you would have $1.2 million at the end of the year. You are clearly better off with the proposed land investment (assuming that it really is riskless like the U.S. government bond invest- ment). How much better off are you?

The answer to this question is at the heart of NPV calculations. The land investment project is worth $1,165,049 today to an investor who demands a 3 percent return on this type of investment. You, however, have been able to acquire this investment for only $1,000,000. Thus, your wealth has increased by undertaking this investment by $165,049 ($1,165,049 present value of the projected investment opportunity payoffs minus the required initial investment of $1,000,000). The NPV of this investment is $165,049. In general, the NPV of an investment is equal to

NPV = Present value of future returns − Initial outlay [2.4]

This example was simplified by assuming that the returns from the investment were received exactly one year from the date of the initial outlay. If the payoff from the land

present value The value today of a future amount of money or a series of future payments evaluated at the appropriate discount rate.

46 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

investment had been not one but two years away, the PVIF would have been 1/(1.03)2 = 0.942596, and the NPV would have been 1.2 million (.942596) – 1.0 million = $131,115. The NPV rule can be generalized to cover returns received over any number of future time periods with projected growth or decay and terminal values as salvage or disposal costs. In Appendix A at the end of the book, the present value concept is developed in more detail so that it can be applied in these more complex investment settings.

Example Changing a Lightbulb Saves $40 and May Save the Planet4

Incandescent lightbulbs replaced oil lamps for interior lighting more than 100 years ago. Thomas Edison himself improved on some basic designs running electric cur- rent through a carbonized filament in an oxygen-free vacuum tube, producing less combustion and more light. General Electric had its origins selling long-lasting tungsten filament incandescent bulbs. Today, the new compact fluorescent light (CFL) bulb uses 75 percent less electricity to heat an argon vapor that emits ultra- violet light. The UV light excites a fluorescent phosphor coating on the inside of the tube, which then emits visible light. The U.S. Department of Energy estimates that if all 105 million U.S. households replaced just one heavily used incandescent bulb with a CFL bulb yielding comparable light, the electricity saved could light 3 million homes. In addition, the energy saved would remove from the environ- ment an amount of greenhouse gases from coal-burning power plants equal to the CO2 emitted by 800,000 cars. The U.K. Department of Business, Enterprise, and Reg- ulatory Reform estimates that replacing the three most frequently used lightbulbs in U.K. households would save the electricity used by all the street lamps in Britain.

The magnitude of these energy savings is certainly staggering, but at what cost? Bought for $1.19 per bulb, 1,000-hour incandescent 75-watt bulbs cost much less to install than CFL bulbs that create the same 1,250 lumens of light, last 8,000 hours, burn only 18 to 22 watts of electricity, but cost $14. So, the lifetime cost comparison hinges on whether the extra $12.81 acquisition cost of the CFL bulb is worth the extended lifetime of energy savings. Net present value techniques are designed to answer just such questions of the time value of money (savings) that are delayed.

Table 2.5 shows the initial net investments of $14 and $1.19 per bulb, the 55 kilowatt hours (kWh) of power saved on average by the CFL bulb each year, the $0.10 per kWh representative cost of the electricity,5 and the additional $1.19 in- candescent bulb replacement every 1,000 hours (the typical U.S. household’s an- nual usage). Assuming a 6 percent discount rate, the net present value of the $5.50 annual energy savings plus the $1.19 replacement cost for incandescent bulbs avoided each year for seven years yields a net present value cost savings of $40.79, which exceeds the differential $12.81 acquisition cost for the CFL bulb by $27.98. The European Union has found this $28 net present value of the cost savings from switching to CFL bulbs (plus their CO2 abatement) so compelling that incandes- cent bulbs are no longer approved for manufacture or import into the EU. More gradual U.S. phaseout of incandescent bulbs will begin in 2012.

4Based on “DOE Launches Change a Light, Change the World Campaign” (October 3, 2007), www.energy.gov and www.energystar.gov. 5Electric rates for incremental power vary by region from $.06 per kWh in the state of Washington to $.08 in the Carolinas, to $.12 in California, New York, and across New England.

Chapter 2: Fundamental Economic Concepts 47

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Sources of Positive Net Present Value Projects What causes some projects to have a positive NPV and others to have a negative NPV? When product and factor markets are other than perfectly competitive, it is possible for a firm to earn above-normal profits (economic rents) that result in positive net present value projects. The reasons why these above-normal profits may be available arise from condi- tions that define each type of product and factor market and distinguish it from a perfectly competitive market. These reasons include the following barriers to entry and other factors:

1. Buyer preferences for established brand names 2. Ownership or control of favored distribution systems (such as exclusive auto dealer-

ships or airline hubs) 3. Patent control of superior product designs or production techniques 4. Exclusive ownership of superior natural resource deposits 5. Inability of new firms to acquire necessary factors of production (management,

labor, equipment) 6. Superior access to financial resources at lower costs (economies of scale in attracting

capital) 7. Economies of large-scale production and distribution arising from

a. Capital-intensive production processes b. High initial start-up costs

These factors can permit a firm to identify positive net present value projects for internal investment. If the barriers to entry are sufficiently high (such as a patent on key technology) so as to prevent any new competition, or if the start-up period for competitive ventures is sufficiently long, then it is possible that a project may have a positive net present value. However, in assessing the viability of such a project, the manager or analyst must consider the likely period of time when above-normal returns can be earned before new competitors emerge and force cash flows back to a more normal level. It is generally unrealistic to expect to be able to earn above-normal returns over the entire life of an investment project.

Risk and the NPV Rule The previous land investment example assumed that the investment was riskless. There- fore, the rate of return used to compute the discount factor and the net present value was the riskless rate of return available on a U.S. government bond having a one-year maturity. What if you do not believe that the construction of the new interchange is a cer- tainty, or you are not confident about of the value of the land in one year? To compensate

TABLE 2.5 LIFETIME COST SAVINGS OF COMPACT FLUORESCENT LIGHT (CFL) BULBS

t=0 t=1 t=2 t=3 t=4 t=5 t=6 t=7 t=8

(END OF PERIOD VALUES)

Incandescent −$1.19 −$1.19 −$1.19 −$1.19 −$1.19 −$1.19 −$1.19 −$1.19 0

CFL −$14.00 55 kWh × $.10 = $5.50 $5.50 $5.50 $5.50 $5.50 $5.50 $5.50 $5.50

Cost difference −$12.81 NPV (8 years of $5.50 energy savings at d=6%) = $34.15

NPV (7 years of $1.19 incandescent replacement cost at d=6%) = $6.64

NPV (Lifetime cost savings) − Cost difference

($34.15 + $6.64) $40.79 $12.81

= $27.98

48 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

for the perceived risk of this investment, you decide that you require a 15 percent rate of return on your investment. Using a 15 percent required rate of return in calculating the discount factor, the present value of the expected $1.2 million sales price of the land is $1,043,478 ($1.2 million times [1/1.15]). Thus, the NPV of this investment declines to $43,478. The increase in the perceived risk of the investment results in a dramatic $121,571 decline from $165,049 in the NPV on a $1 million investment.

A primary problem facing managers is the difficulty of evaluating the risk associated with investments and then translating that risk into a discount rate that reflects an ade- quate level of risk compensation. In the next section of this chapter, we discuss the risk concept and the factors that affect investment risk and influence the required rate of return on an investment.

MEANING AND MEASUREMENT OF RISK Risk implies a chance for some unfavorable outcome to occur—for example, the possibility that actual cash flows will be less than the expected outcome. When a range of potential outcomes is associated with a decision and the decision maker is able to assign probabilities to each of these possible outcomes, risk is said to exist. A decision is said to be risk free if the cash flow outcomes are known with certainty. A good example of a risk-free investment is U.S. Treasury securities. There is virtually no chance that the Treasury will fail to redeem these securities at maturity or that the Treasury will default on any interest payments owed. In contrast, US Airways bonds constitute a risky investment because it is possible that US Airways will default on one or more interest payments and will lack sufficient funds at ma- turity to redeem the bonds at face value. In summary, risk refers to the potential variability of outcomes from a decision. The more variable these outcomes are, the greater the risk.

Probability Distributions The probability that a particular outcome will occur is defined as the relative frequency or percentage chance of its occurrence. Probabilities may be either objectively or subjec- tively determined. An objective determination is based on past outcomes of similar events, whereas a subjective determination is merely an opinion made by an individual about the likelihood that a given event will occur. In the case of decisions that are fre- quently repeated, such as the drilling of developmental oil wells in an established oil field, reasonably good objective estimates can be made about the success of a new well. In contrast, for totally new decisions or one-of-a-kind investments, subjective estimates about the likelihood of various outcomes are necessary. The fact that many probability estimates in business are at least partially subjective does not diminish their usefulness.

Using either objective or subjective methods, the decision maker can develop a probability distribution for the possible outcomes. Table 2.6 shows the probability distribution of net cash flows for two sample investments. The lowest estimated annual

TABLE 2.6 PROBABILITY DISTRIBUTIONS OF THE ANNUAL NET CASH

FLOWS (NCF) FROM TWO INVESTMENTS

INVESTMENT I INVESTMENT II

POSSIBLE NCF PROBABILITY POSSIBLE NCF PROBABILITY

$200 0.2 $100 0.2

300 0.6 300 0.6

400 0.2 500 0.2

1.0 1.0

risk A decision-making situation in which there is variability in the possible outcomes, and the probabilities of these outcomes can be specified by the decision maker.

probability The percentage chance that a particular outcome will occur.

Chapter 2: Fundamental Economic Concepts 49

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

net cash flow (NCF) for each investment—$200 for Investment I and $100 for Investment II—represents pessimistic forecasts about the investments’ performance; the middle values— $300 and $300—could be considered normal performance levels; and the highest values— $400 and $500—are optimistic estimates.

Expected Values From this information, the expected value of each decision alternative can be calculated. The expected value is defined as the weighted average of the possible outcomes. It is the value that is expected to occur on average if the decision (such as an investment) were repeated a large number of times.

Algebraically, the expected value may be defined as

r = ∑ n

j = 1 rjpj [2.5]

where r is the expected value; rj is the outcome for the jth case, where there are n possible outcomes; and pj is the probability that the jth outcome will occur. The expected cash flows for Investments I and II are calculated in Table 2.8 using Equation 2.5. In this exam- ple, both investments have expected values of annual net cash flows equaling $300.

Example Probability Distributions and Risk: US Airways Bonds6

Consider an investor who is contemplating the purchase of US Airways bonds. That investor might assign the probabilities associated with the three possible outcomes from this investment, as shown in Table 2.7. These probabilities are interpreted to mean that a 30 percent chance exists that the bonds will not be in default over their life and will be redeemed at maturity, a 65 percent chance of interest default during the life of the bonds, and a 5 percent chance that the bonds will not be redeemed at maturity. In this example, no other outcomes are deemed possible.

6The annual report for the US Airways Corporation can be found at http://investor.usairways.com

TABLE 2.7 POSSIBLE OUTCOMES FROM INVESTING IN US

AIRWAYS BONDS

OUTCOME PROBABILITY

No default, bonds redeemed at maturity 0.30

Default on interest for one or more periods 0.65

No interest default, but bonds not redeemed at maturity 0.05

1.00

TABLE 2.8 COMPUTATION OF THE EXPECTED RETURNS FROM TWO

INVESTMENTS

INVESTMENT I INVESTMENT II

r j p j r j × pj r j p j r j × pj

$200 0.2 $ 40 $100 0.2 $ 20

300 0.6 180 300 0.6 180

400 0.2 80 500 0.2 100

Expected value: r I = $300 r II = $300

expected value The weighted average of the possible outcomes where the weights are the probabilities of the respective outcomes.

50 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Standard Deviation: An Absolute Measure of Risk The standard deviation is a statistical measure of the dispersion of a variable about its mean. It is defined as the square root of the weighted average squared deviations of in- dividual outcomes from the mean:

σ =

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∑ n

j = 1 ðrj − r jÞ2pj

s [2.6]

where σ is the standard deviation. The standard deviation can be used to measure the variability of a decision alter-

native. As such, it gives an indication of the risk involved in the alternative. The larger the standard deviation, the more variable the possible outcomes and the riskier the decision alternative. A standard deviation of zero indicates no variability and thus no risk.

Table 2.9 shows the calculation of the standard deviations for Investments I and II. These calculations show that Investment II appears to be riskier than Investment I because the expected cash flows from Investment II are more variable.

Normal Probability Distribution The possible outcomes from most investment decisions are much more numerous than in Table 2.6 but their effects can be estimated by assuming a continuous proba- bility distribution. Assuming a normal probability distribution is often correct or nearly correct, and it greatly simplifies the analysis. The normal probability distribu- tion is characterized by a symmetrical, bell-like curve. A table of the standard normal probability function (Table 1 in Appendix B at the end of this book) can be used to compute the probability of occurrence of any particular outcome. From this table, for example, it is apparent that the actual outcome should be between plus and minus 1

TABLE 2.9 COMPUTATION OF THE STANDARD DEVIATIONS FOR TWO INVESTMENTS

j r j r rj − r ðrj − rÞ2 pj ðrj − rÞ2pj Investment I 1 $200 $300 −$100 $10,000 0.2 $2,000

2 300 300 0 0 0.6 0

3 400 300 100 10,000 0.2 2,000

∑ 3

j = 1 ðrj − rÞ2pj = $4,000

σ = ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∑ n

j = 1 ðrj − rÞ2pj

r =

ffiffiffiffiffiffiffiffiffiffiffi 4,000

p = $63:25

Investment II 1 $100 $300 −$200 $40,000 0.2 $8,000

2 300 300 0 0 0.6 0

3 500 300 200 40,000 0.2 8,000

∑ 3

j = 1 ðrj − rÞ2pj = $16,000

σ = ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∑ n

j = 1 ðrj − rÞ2pj

r =

ffiffiffiffiffiffiffiffiffiffiffiffiffi 16,000

p = $126:49

standard deviation A statistical measure of the dispersion or variability of possible outcomes.

Chapter 2: Fundamental Economic Concepts 51

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

standard deviation from the expected value 68.26 percent of the time,7 between plus and minus 2 standard deviations 95.44 percent of the time, and between plus and minus 3 standard deviations 99.74 percent of the time (see Figure 2.9). So a “3 sigma event” occurs less than 1 percent of the time with a relative frequency 0.0026 (i.e., 1.0 − 0.9974), and a “9 sigma event” occurs almost never, with a relative frequency less than 0.0001. Nevertheless, such extraordinary events can and do happen (see following box on LTCM).

The number of standard deviations z that a particular value of r is from the mean r can be computed as

z = r − r σ

[2.7]

Table 1 in Appendix B and Equation 2.5 can be used to compute the probability of an annual net cash flow for Investment I being less than some value r—for example, $205. First, the number of standard deviations that $205 is from the mean must be cal- culated. Substituting the mean and the standard deviation from Tables 2.8 and 2.9 into Equation 2.7 yields

z = $205 − $300

$63:25

= −1:50

In other words, the annual cash flow value of $205 is 1.5 standard deviations below the mean. Reading from the 1.5 row in Table 1 gives a value of 0.0668, or 6.68 percent.

FIGURE 2.9 A Sample Illustration of Areas under the Normal Probability Distribution Curve

0

Standard deviations

–1σ–2σ–3σ +1σ +2σ +3σ

95.44% 99.74%

68.26%

Pr ob

ab il

it y

of o

cc ur

re nc

e

15.87%

7For example, Table 1 indicates a probability of 0.1587 of a value occurring that is greater than +1σ from the mean and a probability of 0.1587 of a value occurring that is less than −1σ from the mean. Hence the proba- bility of a value between +1σ and −1σ is 68.26 percent—that is, 1.00 − (2 × 0.1587).

52 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Thus, a 6.68 percent probability exists that Investment I will have annual net cash flows less than $205. Conversely, there is a 93.32 percent probability (1 − 0.0668) that the in- vestment will have a cash flow greater than $205.

Coefficient of Variation: A Relative Measure of Risk The standard deviation is an appropriate measure of risk when the decision alternatives being compared are approximately equal in size (that is, have similar expected values of the outcomes) and the outcomes are estimated to have symmetrical probability distributions. Because the standard deviation is an absolute measure of variability,

WHAT WENT RIGHT • WHAT WENT WRONG

Long-Term Capital Management (LTCM)8

LTCM operated from June 1993–September 1998 as a hedge fund that invested highly leveraged private capital in arbitrage trading strategies on the financial derivative markets. LTCM’s principal activity was examining interest rate derivative contracts throughout the world for evidence of very minor mispricing and then betting enormous sums on the subsequent convergence of those contracts to pre- dictable equilibrium prices. Since the mispricing might be only several cents per thousand dollars invested, LTCM often needed to risk millions or even billions on each bet to secure a nontrivial absolute dollar return. With some- times as many as 100 independent bets spread across doz- ens of different government bond markets, LTCM appeared globally diversified.

In a typical month, 60 such convergence strategies with positions in several thousand counterparty contracts would make money and another 40 strategies with a similar num- ber of counterparties would lose money. Steadily, the prof- its mounted. From approximately $1 billion net asset value (equity) in February 1994, LTCM reached $7 billion of net asset value in January 1998. LTCM then paid out $2.4 bil- lion in a one-time distribution to non-partners, which equaled a 40 percent annual compound return on their investment (ROI). Shortly thereafter, in August 1998, the remaining $4.6 billion equity shrank by 45 percent, and then one month later shrank by another 82 percent to less than $600 million. In September 1998, the hedge fund was taken over by 14 Wall Street banks who, in ex- change for inserting $3.6 billion to cover the firm’s debts, acquired 90 percent of the equity ownership. What went wrong?

One potential explanation is that such events are fully expected in an enterprise so risky that it returns a 40 percent ROI. Anticipated risk and expected return are highly positively correlated across different types of investments. However, LTCM’s annual return had a standard deviation from June 1993 to June 1998 of only

11.5 percent per year as compared to 10 percent as the average for all S&P 500 stocks. In this respect, LTCM’s return volatility was quite ordinary. Another potential ex- planation is that LTCM’s $129 billion on the June 1998 balance sheet was overwhelmed by excessive off-balance sheet assets and liabilities. Although the absolute size of the numbers is staggering (e.g., $1.2 trillion in interest rate swaps, $28 billion in foreign exchange derivatives, and $36 billion in equity derivatives), LTCM’s 9 percent ratio of on-balance sheet to off-balance sheet assets was similar to that of a typical securities firm (about 12 per- cent). Even LTCM’s high financial leverage ($129 billion assets to $4.7 billion equity = 26 to 1) was customary practice for hedge funds.

What appears to have gone wrong for LTCM was that a default of the Russian government on debt obligations in August 1998 set in motion a truly extraordinary “flight to quality.” General turmoil in the bond markets caused in- terest rate volatility to rise to a standard deviation of 36 percent when 3 percent would have been typical. LTCM was caught on the wrong side of many interest rate derivative positions for which no trade was available at any price. Although LTCM had “stress tested” their trading positions against so-called “3 sigma events” (a one-day loss of $35 million), this August–September 1998 volatility proved to be a 9 sigma event (i.e., a one- day loss of $553 million).

With massive investments highly leveraged and ex- posed to a 9 sigma event, LTCM hemorrhaged $2 billion in one month. Because liquidity risk exposure of an other- wise fully diversified portfolio was to blame, many invest- ment houses have concluded that leverage should be substantially reduced as a result of the events at LTCM.

8R. Lowenstein, When Genius Failed (New York: Random House, 2000); remarks by Dave Modest, NBER Conference, May 1999; and “Case Study: LTCM,” eRisk, (2000).

Chapter 2: Fundamental Economic Concepts 53

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

however, it is generally not suitable for comparing alternatives of differing size. In these cases the coefficient of variation provides a better measure of risk.

The coefficient of variation (v) considers relative variation and thus is well suited for use when a comparison is being made between two unequally sized decision alternatives. It is defined as the ratio of the standard deviation σ to the expected value r , or

ν = σ

r [2.8]

RISK AND REQUIRED RETURN The relationship between risk and required return on an investment can be defined as

Required return = Risk-free return + Risk premium [2.9]

The risk-free rate of return refers to the return available on an investment with no risk of default. For debt securities, no default risk means that promised interest and prin- cipal payments are guaranteed to be made. The best example of risk-free debt securities are short-term government securities, such as U.S. Treasury bills. The buyer of a U.S. government debt security always is assured of receiving the promised principal and inter- est payments because the U.S. government always can print more money. The risk-free return on T-bills equals the real rate of interest plus the expected rate of inflation. The second term in Equation 2.9 is a potential “reward” that an investor can expect to receive

Example Relative Risk Measurement: Arrow Tool Company Arrow Tool Company is considering two investments, T and S. Investment T has ex- pected annual net cash flows of $100,000 and a standard deviation of $20,000, whereas Investment S has expected annual net cash flows of $4,000 and a $2,000 standard de- viation. Intuition tells us that Investment T is less risky because its relative variation is smaller. As the coefficient of variation increases, so does the relative risk of the deci- sion alternative. The coefficients of variation for Investments T and S are computed as

Investment T:

ν = σ

r

= $20,000 $100,000

= 0:20

Investment S:

ν = σ

r

= $2,000 $4,000

= 0:5

Cash flows of Investment S have a larger coefficient of variation (0.50) than do cash flows of Investment T (0.20); therefore, even though the standard deviation is smaller, Investment S is the more risky of the two alternatives.

coefficient of variation The ratio of the standard deviation to the expected value. A relative measure of risk.

54 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

from providing capital for a risky investment. This risk premium may arise for any number of reasons. The borrower firm may default on its contractual repayment obligations (a default risk premium). The investor may have little seniority in presenting claims against a bankrupt borrower (a seniority risk premium). The investor may be un- able to sell his security interest (a liquidity risk premium as we saw in the case of LTCM), or debt repayment may occur early (a maturity risk premium). Finally, the re- turn the investor receives may simply be highly volatile, exceeding expectations during one period and plummeting below expectations during the next period. Investors gener- ally are considered to be risk averse; that is, they expect, on average, to be compensated for any and all of these risks they assume when making an investment.

Example Risk-Return Trade-Offs in Stocks, Bonds, Farmland, and Diamonds Investors require higher rates of return on debt securities based primarily on their default risk. Bond-rating agencies, such as Moody’s, Standard and Poor’s, and Fitch, provide evaluations of the default risk of many corporate bonds. Moody’s, for example, rates bonds on a 9-point scale from Aaa through C, where Aaa- rated bonds have the lowest expected default risk. As can be seen in Table 2.10, the yields on bonds increase as the risk of default increases, again reflecting the positive relationship between risk and required returns.

Table 2.10 also shows investment in diamonds has returned 3 percent whereas farmland has returned 6.5 percent, U.S. stocks have returned 10 percent, biotech stocks have returned 12.6 percent, and emerging market stocks have returned 16 percent compounded annually from 1970 to 2010. These compound annual returns mirror the return variance of diamonds (lowest), farmland, U.S. stocks, biotech stocks, and emerging market stocks (highest).

TABLE 2.10 RELATIONSHIP BETWEEN RISK AND REQUIRED RETURNS

DEBT SECURITY YIELD

U.S. Treasury bill 3.8%

U.S. Treasury bonds (25 year +) 5.06

Aaa-rated corporate bonds 6.49

Aa-rated bonds 6.93

A-rated bonds 7.18

Baa-rated corporate bonds 7.80

Other investments

Diamonds 3.0

Farmland 6.5

Stocks

All U.S. stocks 10.1

Biotech stocks 12.6

Emerging market stocks 16.0

Source: Board of Governors of the Federal Reserve System, Federal Reserve Bulletin.

Chapter 2: Fundamental Economic Concepts 55

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

SUMMARY

� Demand and supply simultaneously determine equilibrium market price. The determinants of de- mand (supply) other than price shift the demand (supply) curve. A change in price alone leads to a change in quantity demanded (supplied) without any shift in demand (supply).

� The offer price demanders are willing to pay is determined by the marginal use value of the pur- chase being considered. The asking price suppliers are willing to accept is determined by the variable cost of the product or service being supplied.

� The equilibrium price of gasoline fluctuates pri- marily because of spikes and collapses in crude oil input prices caused at various times by supply disruptions and gluts, increasing demand in devel- oping countries, and speculation.

� Changes in price result in movement along the de- mand curve, whereas changes in any of the other variables in the demand function result in shifts of the entire demand curve. Thus “changes in quan- tity demanded along” a particular demand curve result from price changes. In contrast, when one speaks of “changes in demand,” one is referring to shifts in the entire demand curve.

� Some of the factors that cause a shift in the entire demand curve are changes in the income level of consumers, the price of substitute and complemen- tary goods, the level of advertising, competitors’

advertising expenditures, population, consumer preferences, time period of adjustment, taxes or subsidies, and price expectations.

� The marginal analysis concept requires that a deci- sion maker determine the additional (marginal) costs and additional (marginal) benefits associated with a proposed action. If the marginal benefits exceed the marginal costs (that is, if the net mar- ginal benefits are positive), the action should be taken.

� The net present value of an investment is equal to the present value of expected future returns (cash flows) minus the initial outlay.

� The net present value of an investment equals the contribution of that investment to the value of the firm and, accordingly, to the wealth of share- holders. The net present value of an investment depends on the return required by investors (the firm), which, in turn, is a function of the perceived risk of the investment.

� Risk refers to the potential variability of outcomes from a decision alternative. It can be measured ei- ther by the standard deviation (an absolute mea- sure of risk) or coefficient of variation (a relative measure of risk).

� A positive relationship exists between risk and re- quired rates of return. Investments involving greater risks must offer higher expected returns.

Exercises 1. For each of the determinants of demand in Equation 2.1, identify an example illustrating the effect on the demand for hybrid gasoline-electric vehicles such as the Toyota Prius. Then do the same for each of the determinants of supply in Equation 2.2. In each instance, would equilibrium market price increase or de- crease? Consider substitutes such as plug-in hybrids, the Nissan Leaf and Chevy Volt, and complements such as gasoline and lithium ion laptop computer batteries.

2. Gasoline prices above $3 per gallon have affected what Enterprise Rental Car Co. can charge for various models of rental cars. SUVs are $37 with one-day return and subcompacts are $41 with one-day return. Why would the equilibrium price of SUVs be lower than the equilibrium price of subcompacts?

Answers to the exercises in blue can be found in Appendix D at the back

of the book.

56 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

3. The Ajax Corporation has the following set of projects available to it:

PROJECT* INVESTMENT REQUIRED

($ MILLION) EXPECTED RATE

OF RETURN

A 500 23.0%

B 75 18.0

C 50 21.0

D 125 16.0

E 300 14.0

F 150 13.0

G 250 19.0

*Note: All projects have equal risk.

Ajax can raise funds with the following marginal costs:

First $250 million 14.0%

Next 250 million 15.5

Next 100 million 16.0

Next 250 million 16.5

Next 200 million 18.0

Next 200 million 21.0

Use the marginal cost and marginal revenue concepts developed in this chapter to derive an optimal capital budget for Ajax.

4. The demand for MICHTEC’s products is related to the state of the economy. If the economy is expanding next year (an above-normal growth in GNP), the com- pany expects sales to be $90 million. If there is a recession next year (a decline in GNP), sales are expected to be $75 million. If next year is normal (a moderate growth in GNP), sales are expected to be $85 million. MICHTEC’s economists have estimated the chances that the economy will be either expanding, normal, or in a recession next year at 0.2, 0.5, and 0.3, respectively. a. Compute expected annual sales. b. Compute the standard deviation of annual sales. c. Compute the coefficient of variation of annual sales.

5. Two investments have the following expected returns (net present values) and standard deviation of returns:

PROJECT EXPECTED RETURNS STANDARD DEVIATION

A $ 50,000 $ 40,000

B $250,000 $125,000

Which one is riskier? Why? 6. The manager of the aerospace division of General Aeronautics has estimated the

price it can charge for providing satellite launch services to commercial firms. Her most optimistic estimate (a price not expected to be exceeded more than 10 per- cent of the time) is $2 million. Her most pessimistic estimate (a lower price than this one is not expected more than 10 percent of the time) is $1 million. The expected value estimate is $1.5 million. The price distribution is believed to be approximately normal.

Chapter 2: Fundamental Economic Concepts 57

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

a. What is the expected price? b. What is the standard deviation of the launch price? c. What is the probability of receiving a price less than $1.2 million?

Case Exercise REVENUE MANAGEMENT AT AMERICAN

AIRLINES9 Airlines face highly cyclical demand; American reported profitability in the strong ex- pansion of 2006–2007 but massive losses in the severe recession of 2008–2009. De- mand also fluctuates day to day. One of the ways American copes with random demand is through marginal analysis using revenue management techniques. Revenue or “yield” management (RM) is an integrated demand-management, order-booking, and capacity-planning process.

To win orders in a service industry without slashing prices requires that companies create perceived value for segmented classes of customers. Business travelers on air- lines, for example, will pay substantial premiums for last-minute responsiveness to their flight change requests. Other business travelers demand exceptional delivery re- liability and on-time performance. In contrast, most vacation excursion travelers want commodity-like service at rock-bottom prices. Although only 15–20 percent of most airlines’ seats are in the business segment, 65–75 percent of the profit contribution on a typical flight comes from this group.

The management problem is that airline capacity must be planned and allocated well in advance of customer arrivals, often before demand is fully known, yet unsold inventory perishes at the moment of departure. This same issue faces hospitals, con- sulting firms, TV stations, and printing businesses, all of whom must acquire and schedule capacity before the demands for elective surgeries, a crisis management team, TV ads, or the next week’s press run are fully known.

One approach to minimizing unsold inventory and yet capturing all last-minute high-profit business is to auction off capacity to the highest bidder. The auction for free-wheeling electricity works just that way: power companies bid at quarter ’til the hour for excess supplies that other utilities agree to deliver on the hour. However, in airlines, prices cannot be adjusted quickly as the moment of departure approaches. Instead, revenue managers employ large historical databases to predict segmented cus- tomer demand in light of current arrivals on the reservation system. They then ana- lyze the expected marginal profit from holding in reserve another seat in business class in anticipation of additional “last-minute” demand and compare that seat by seat to the alternative expected marginal profit from accepting one more advance res- ervation request from a discount traveler.

Suppose on the 9:00 A.M. Dallas to Chicago flight next Monday, 63 of American’s 170 seats have been “protected” for first class, business class, and full coach fares but only 50 have been sold; the remaining 107 seats have been authorized for sale at a discount. Three days before departure, another advance reservation request arrives in the discount class, which is presently full. Should American reallocate capacity and

9Based on Robert Cross, Revenue Management (New York: Broadway Books, 1995); and Frederick Harris and Peter Peacock, “Hold My Place Please: Yield Management Improves Capacity Allocation Guesswork,” Marketing Management (Fall 1995), pp. 34–46.

58 Part 1: Introduction

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

take on the new discount passenger? The answer depends on the marginal profit from each class and the predicted probability of excess demand (beyond 63 seats) next Monday in the business classes.

If the $721 full coach fare has a $500 marginal profit and the $155 discount fare has a $100 marginal profit, the seat in question should not be reallocated from busi- ness to discount customers unless the probability of “stocking out” in business is less than 0.20 (accounting for the likely incidence of cancellations and no-shows). There- fore, if the probability of stocking out is 0.25, the expected marginal profit from hold- ing an empty seat for another potential business customer is $125, whereas the marginal profit from selling that seat to the discount customer is only $100 with cer- tainty. Even a pay-in-advance no-refund seat request from the discount class should be refused. Every company has some viable orders that should be refused because ad- ditional capacity held in reserve for the anticipated arrival of higher profit customers is not “idle capacity” but rather a predictable revenue opportunity waiting to happen.

In this chapter, we developed the marginal analysis approach used in solving American’s seat allocation decision problem. The Appendix to Chapter 14 discusses further the application of revenue management to baseball, theatre ticketing, and hotels.

Questions 1. Make a list of some of the issues that will need to be resolved if American Air-

lines decides to routinely charge different prices to customers in the same class of service.

2. Would you expect these revenue management techniques of charging differential prices based on the target customers’ willingness to pay for change order respon- siveness, delivery reliability, schedule frequency, and so forth to be more effective in the trucking industry, the outpatient health care industry, or the hotel indus- try? Why or why not?

3. Sometimes when reservation requests by deep discount travelers are refused, de- manders take their business elsewhere; they “balk.” At other times, such deman- ders negotiate and can be “sold up” to higher fare service like United’s Economy Plus. If United experiences fewer customers balking when reservation requests for the cheapest seats are refused, should they allocate preexisting capacity to protect fewer seats (or more) for late-arriving full-fare passengers?

Chapter 2: Fundamental Economic Concepts 59

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PART2 DEMAND AND FORECASTING

ECONOMIC ANALYSIS AND DECISIONS

1. Demand Analysis and Forecasting

2. Production and Cost Analysis 3. Pricing Analysis 4. Capital Expenditure Analysis

POLITICAL AND SOCIAL ENVIRONMENT

1. Business Conditions (Trends, Cycles, and Seasonal Effects)

2. Factor Market Conditions (Capital, Labor, Land, and

Raw Materials) 3. Competitors’ Responses 4. External, Legal, and Regulatory Constraints 5. Organizational (Internal)

Constraints

Cash Flows Risk

Firm Value (Shareholders’ Wealth)

61

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

3 CHAP T E R

Demand Analysis CHAPTER PREVIEW Demand analysis serves three major managerial objectives. First, it provides the insights necessary for marketing teams to effectively manage demand. Second, it helps forecast unit sales to inform operations decisions, and third it projects the revenue portion of a firm’s cash flow stream for financial planning. This chapter develops the theory of demand and introduces the elasticity properties of the demand function. The price elasticity of demand is a measure of the sensitivity of quantity demanded to a change in one of the factors influencing demand, such as price, advertising, promotions, packaging, or income levels. We analyze consumer behavior in selecting hotels and auto rentals subject to a reimbursements budget constraint on a business trip. The household demand for gasoline over the past several years is extensively discussed. A Case Study on direct mail couponing by a Chevrolet dealership examines the determinants of price elasticity for various target markets. Web Appendix A develops the relationship between cost- of-living price indices and new product introductions.

MANAGERIAL CHALLENGE Health Care Reform and Cigarette Taxes1

When the Canadian government raised cigarette taxes enough to push the price per pack over $4, adult smoking declined by 38 percent and teenage smoking declined even more, by 61 percent. In 1997, a similar U.S. excise tax increase funded the “Tobacco settlement.” In exchange for immunity from civil liability in class action lawsuits by injured smokers, Philip Morris, Reynolds Tobacco, Liggett, and other cigarette manufacturers agreed to pay $368 billion over 25 years. The state attorneys general had sued the cigarette manufacturers to recover the additional Medicare and Medicaid costs of smoking-related ill- nesses. Under the settlement, the average U.S. price of $1.43 per pack rose by 62 cents to $2.05 (by 35 per- cent).2 Some critics of the proposal insisted at the time that the tobacco tax should be higher (perhaps as much as $1.50 higher) to deter young smokers from acquiring the habit. The stated objective for reducing teenage smoking was 30 percent in five years and 50 percent in seven years.

One important element of the debate regarding the “optimal” cigarette tax increase depends on how sensi- tive consumption is to changes in price. An excellent measure of this sensitivity is the price elasticity of de- mand, defined as the percentage change in quantity de- manded that occurs as a result of some percentage change in price. Economists have estimated the price elasticity of adult cigarette demand to be −0.4, indicat- ing that for a 10 percent increase in price, quantity de- manded can be expected to decline by 4 percent. For teenagers, however, the price elasticity is thought to be much higher—namely, −0.6—indicating that for a 10 percent increase in price, quantity demanded can be expected to decline by 6 percent. So, using price elastic- ity as a guide, the 35 percent increase in price should result in a 21 percent decline in teen smoking.

Because a 21 percent reduction is far below the stated goal of a 30 percent reduction in teenage smoking, the U.S. Congress decided in 1999 to raise the federal excise tax another 60 cents. State legislatures got involved as

62

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

well, such that by 2009 the price is approaching $4.00 and much higher in some states. In the past three years, states from Florida to New Hampshire and out to Texas and South Dakota have added $1.00 to their cigarette excise taxes. New Jersey, New York, Wisconsin, Washington, Hawaii, Rhode Island, Massachusetts, Vermont, and Arizona, for example, have imposed more than $2 a pack in state excise taxes alone, making the price of a pack $5.

In the ongoing debate over the amount of cigarette tax required for health care cost recovery, policy makers face a difficult set of trade-offs. On the one hand, if the primary

STATE EXCISE TAX RATES ON CIGARETTES (JULY 1, 2009)

STATE TAX RATE

(¢ PER PACK) RANK STATE TAX RATE

(¢ PER PACK) RANK

Alabama 42 45 Nebraska 64 38

Alaska 200 9 Nevada 80 34

Arizona 200 10 New Hampshire 178 15

Arkansas 115 27 New Jersey 270 3

California 87 32 New Mexico 91 31

Colorado 84 33 New York 275 2

Connecticut 200 11 North Carolina 35 48

Delaware 115 25 North Dakota 44 44

Florida 133 22 Ohio 125 23

Georgia 37 46 Oklahoma 103 28

Hawaii 260 4 Oregon 118 26

Idaho 57 42 Pennsylvania 135 21

Illinois 98 29 Rhode Island 346 1

Indiana 95 30 South Carolina 7 51

Iowa 136 19 South Dakota 153 17

Kansas 79 35 Tennessee 62 39

Kentucky 60 40 Texas 141 20

Louisiana 36 47 Utah 69 36

Maine 200 14 Vermont 224 7

Maryland 200 12 Virginia 30 49

Massachusetts 251 6 Washington 202 8

Michigan 200 4 West Virginia 55 43

Minnesota 150 18 Wisconsin 252 5

Mississippi 68 37 Wyoming 60 41

Missouri 17 50 Dist. of Columbia 200 13

Montana 170 16 U.S. Median 115

Source: Tax Foundation.

MANAGERIAL CHALLENGE Continued

© M ed io Im ag es /G et ty Im ag es

Chapter 3: Demand Analysis 63

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

DEMAND RELATIONSHIPS The Demand Schedule Defined The demand schedule is the simplest form of the demand relationship. It is merely a list of prices and corresponding quantities of a product or service that would be demanded over a particular time period by some individual or group of individuals at uniform prices. Table 3.1 shows the demand schedule for gasoline. This demand schedule indicates that if the price per gallon were $2.50, U.S. urban households would purchase 18 gallons per week. At a price of $3.50, the target households would purchase just 14 gallons per week. Note that the lower the price, the greater the quantity demanded. This inverse or negative relationship between price and the desired rate of purchase is generally referred to as the “law of demand” and refers to the movement down a single demand schedule.

A shift of demand schedules as illustrated in Figure 3.1 is quite different. Demand schedules shift when one of the determinants of demand discussed in Chapter 2 and listed in Equation 2.1 changes. For example, the demand schedule for gas-guzzling SUVs in 1999 reflected low gasoline prices, an important complement in the consump- tion of heavy fuel-inefficient cars and trucks. In 1999 at $2.00 per gallon gasoline, Ford Motor Co. sold 428,000 Ford Explorers for $22,000 at $4,000 profit each. By 2007, with

objective is to generate income to fund health care costs, the tax should be set such that it will maximize tax reve- nue. On the other hand, if the primary objective is to discourage smoking, a much higher tax could be justified. In either case, however, knowledge of the true price elas- ticity of demand is an essential element of this important policy decision. In this chapter, we investigate how to cal- culate and use such price elasticity metrics.

Discussion Questions

� Think back to when you were a teenager. Were you more or less price sensitive than you are now in making gasoline consumption deci- sions when you encounter an unexpected

discount or price increase? Brainstorm about why?

� What about pizza consumption decisions? Again, why?

1Based in part on “Add $2 to the Cost of a Pack of Cigarettes” and “And Even Teen Smokers May Kick the Habit,” BusinessWeek (March 15, 1993), p. 18; “Critics Question Tobacco Pact’s Effect on Teen Smoking,” Wall Street Journal (August 19, 1997), p. A20; “Major Makers of Cigarettes Raise Prices,” Wall Street Journal (August 31, 1999), p. A3; and “Politicians Are Hooked on Cigarette Taxes,” Wall Street Journal (February 20, 2002), p. A2. 2The 35 percent price increase in the United States is calculated by dividing the $0.62 tax increase by the average of the original $1.43 and the post-tax $2.05 price.

TABLE 3.1 U.S. HOUSEHOLD DEMAND FOR GASOLINE

PRICE ($/GALLON) QUANTITY PURCHASED (GALLONS PER WEEK)

$4.00 11.5

$3.50 14

$3.00 16

$2.50 18

$2.00 20

$1.50 22

$1.00 24

Source: Household Vehicles Energy Use: Latest and Trends, Energy Information Administration, U.S. Department of Energy, various issues.

MANAGERIAL CHALLENGE Continued

64 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

$3.50 per gallon gasoline, Ford Motor could only sell 127,000 Explorers despite deep dis- counts to $19,000 that sacrificed essentially all the profit.3 What had changed? Tastes and preferences were running in favor of “greener,” more sustainable transportation (like the gasoline-electric hybrid Honda Prius). But the main difference from the earlier period was that gasoline prices had risen to $3.50 per gallon. As a result, demand for gas- guzzler SUVs like the Ford Explorer shifted down to the left. By 2008, with gasoline prices rising to over $4.00 for the first time ever in the United States, the Ford Explorer sold fewer than 30,000 units and was withdrawn from the marketplace.

Constrained Utility Maximization and Consumer Behavior The concept of demand is based on the theory of consumer choice. Each consumer faces a constrained optimization problem, where the objective is to choose among the combi- nations of goods and services that maximize satisfaction or utility, subject to a constraint on the amount of funds (i.e., the household budget) available. Think of a food and hous- ing budget allowance from your employer while you are traveling on a business trip or, alternatively, a set of friends who share these expenses while rooming together. In this constrained utility-maximizing framework, economists have identified two basic reasons for the increase in quantity demanded as the result of a price reduction. These factors are known as real income and substitution effects.

Real Income Effect When the price of a good—for example, apartment housing rent—declines, the effect of this decline is that the purchasing power of the consumer has increased. This is known as the real income effect of the price change. For example, if an in- dividual normally purchases 600 square feet at $1,000 per month, a price decline to $800 per month would enable the consumer to purchase the same amount of housing for $200 less per month. This savings of $200 represents an increase in purchasing power, which may be used to purchase greater quantities of housing (as well as other income-superior goods).

Sometimes the real income effect of a price reduction is minuscule because so little of the household’s budget is expended on the good (like a salt canister purchased once every other year), but at other times the change in purchasing power is enormous.

FIGURE 3.1 Demand for SUV (Ford Explorer) as Gasoline Price Doubled

428,000 Units sold

D1999

D2007

127,000

$19,000

$22,000

D ol

la r

pr ic

e ($

)

3“Oil’s Rise to $100,” Wall Street Journal (January 3, 2008), p. A6.

Chapter 3: Demand Analysis 65

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Consider a young family who spends 40 percent of their disposable income on apart- ment housing. In general, the sign and magnitude of the real income effect of a price change has much to do with the positioning and targeting decisions in the firm’s market- ing plan—that is, at whom the product or service is targeted. French-American families, for example, spend almost twice as much of their disposable income on food (22 per- cent) as does the representative American family (12 percent). As a result, the quantity demanded of income-superior food items like veal or wines we would expect to be much more price sensitive among French-American families.

Substitution Effect When the price of a good such as movies declines, it becomes less expensive in relation to other substitute entertainment outings—for example, restau- rant meals. As a result of the price decline, the rational consumer can increase his or her satisfaction or utility by purchasing more of the good whose price has declined and less of the substitutes. This is known as the substitution effect of the price change.

For example, suppose that the prices of movie admission and snacks versus a restau- rant meal are $20 and $30, respectively. Furthermore, assume that initially a household purchases one movie outing and two restaurant meals per week for a total entertainment expenditure of $80 per week. If the price of movies declines to $16.67, some households will decide to increase their consumption to three movies per week and decrease their consumption of restaurant meal outings to one per week—which requires the same total expenditure of $80. Each such household now realizes that they must forego almost two movies per week to fund a restaurant meal outing, so they cut back on restaurant meals. Thus, we see that a decrease in the relative price of movies versus restaurant meals leads to an increase in the quantity demanded of movies.

In summary, because of the combined impact of the real income and substitution ef- fects, a decline in the price will always have a positive impact on the quantity demanded for income-superior goods and services (for which more consumption is preferred to less as purchasing power rises). In this case, both the income and substitution effects push the consumer toward an increase in quantity demanded as price falls.4 This law of demand is almost never violated.

Example Consumption Choices on a Business Trip to San Francisco On a two-week business trip to San Francisco, your employer has authorized a $1,000 travel budget for your housing and auto rentals against which you can sub- mit receipts for reimbursement. The price of housing is $100 per 100 sq. ft. for the 10 weekdays, and the price of a mid-size auto rental with parking and gas runs $100 per day. You plan to allocate $700 for housing and use the rest to rent a car for the three-day intervening weekend from Friday noon to Monday noon in order to see Yosemite National Park. While flying to San Francisco, however, you notice in the airline’s in-flight magazine a $30 per day Hertz discount coupon good throughout the two weeks you are traveling.

(Continued)

4For income-inferior goods and services, like efficiency apartments (for which less is preferred to more as pur- chasing power rises), the real income and substitution effects have opposite and partially offsetting impacts on the quantity demanded. Nevertheless, the net effect, even in the case of inferior goods, is usually that again more goods and services will be demanded as the price declines. So, the law of demand holds for most inferior goods as well.

66 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Figure 3.2 illustrates how this Hertz discount coupon expands your purchasing power substantially. Most obviously, now that you face a budget constraint reflect- ing $70 discount prices for auto rentals (the starred line in the figure), a specialist in consumption who plans to spend the entire $1,000 on auto rentals (and sleep in the car) could rent an auto each of the 14 days in San Francisco whereas earlier only 10 days was available with the initial (the solid line) budget constraint. Whether this increase in your purchasing power from Hertz’s $30 price reduction will, in fact, trigger several additional auto rental days (say, from three to five) allowing you perhaps to drive across the Golden Gate Bridge or up to the Napa Valley some weekday evening, depends on a number of factors.

First, is the Hertz discount offer for “$30 off” available on all of Hertz’s vehicles, or is it just available on subcompact econoboxes like the Ford Focus? If the latter, you might decide to spend your increased purchasing power on a nicer, bigger hotel room and actually consume fewer rental days of what you perceive to be the inferior Ford Focus. Similarly, if the discount coupon is only good for luxury car rentals, that choice may be well beyond your budget.

The second factor that determines whether the Hertz price discount stimu- lates more demand is whether switching costs are high or low. That is, are you a member of the Avis #1 Club, a loyalty program that earns points for free gifts

FIGURE 3.2 Consumption Choice on a Business Trip

Auto rental (days) 10

650 700 720

910

1000

New budget constraint

Old budget constraint

Hypothetical budget with no Δ purchasing power

H ou

si ng

( sq

ua re

f ee

t)

543 14

(Continued)

Chapter 3: Demand Analysis 67

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Up to this point we have been considering the demand for goods and services pur- chased on a short-term basis. Durable goods such as clothes, furniture, autos, appliances, and computers may be stored and repaired rather than replaced regularly. For example, an electric furnace can be fixed again and again at considerably less cost than buying a new heating system. Obsolescence in style, convenience, and prestige plays a large role in affect- ing the replacement demand of durables. Also, customers may question whether their in- come will be sufficient and steady enough to make installment loan payments or whether adequate repair services will be available over the economic working life of a discontinued model. Because these expectational factors come into play, demand for durable goods is more volatile, and its analysis is more complex than a similar analysis for nondurables.

from Avis auto rentals? Or has your employer already prepaid on a nonreimbur- sable contract with Avis based on your three-day weekend rental plan? Or is a smaller hotel room really unusable because you require the 700 sq. ft. room’s seating area separate from your bedroom in which to conduct business meet- ings? If so, your switching costs are too high for the Hertz price discount to have much of a positive effect on your quantity demanded.

At lower levels of switching costs, however, we would expect you to substitute somewhat toward Hertz rental cars and away from housing as the relative price of auto rentals falls from 100 sq. ft./rental car day to 70 sq. ft./day. Conceptually, the magnitude of the substitution effect away from housing space toward more rental car days (say, from your originally planned three to four days) depends upon the perceived closeness of the substitutes, which is another way of saying the size of the switching costs. The feasible choices within your original budget are displayed on the dotted hypothetical budget constraint in Figure 3.2 illustrating the lower $70 relative price of rental cars.

Third and finally, the size of the purchasing power effect (up or down from the choice of four rental car days) as spending power rises from the dotted hypothetical budget constraint (at $910 and $70 rental car prices) to the new actual outside bud- get constraint (at $1,000 and $70 rental car prices) depends upon the positioning of the eligible Hertz vehicles. Are they positioned like the inferior subcompact Ford Focus, like the midsize Ford Taurus you presently drive, or like the income- superior Ford Mustang convertible? If Ford Mustang convertibles are eligible for the discount, perhaps your quantity demanded rises from three to five days. The total effect of the $30 price discount coupon would then be +2 days, the sum of the ever positive substitution effect as relative price declines (from three to four days) and a reinforcing positive purchasing power effect (on from four to five days).

On the other hand, if the Hertz coupon is applicable only to subcompacts that you perceive as income inferior, maybe you do business with Hertz but actually decrease auto rental consumption from three days before the price change to two days afterward and spend your increased purchasing power on a better hotel cost- ing $860 for the 10-day stay. The total effect of the $30 price discount coupon in this unusual case would be −1 days, the sum of a positive substitution effect as relative price declines (from three to four days) plus an offsetting negative purchas- ing power effect (from four back to two days) in light of the inferior nature of the product positioning. If Hertz was attempting to clear an excess inventory of un- rented subcompacts, they should have targeted some economy-class airline custo- mers who will perceive the Ford Focus as income superior.

durable goods Goods that yield benefits to the owner over a number of future time periods.

68 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Targeting, Switching Costs, and Positioning The customer’s desired rate of purchase can be markedly affected not only by pricing but also by several other marketing decisions such as targeting the most likely customers, establishing loyalty pro- grams, and carefully positioning the product. Let’s discuss each of these in turn. Target- ing is often the subject of extensive marketing research using surveys, focus groups, and statistical analysis. Companies want to know the subtle latent inclinations and declina- tions of potential target customer groups before designing their promotions and ad cam- paigns. “Knowing thy customer” sufficiently well to pick the right target for a promising product is the first priority in marketing.

Frequent buyers in customer loyalty programs will often attempt to secure repeat pur- chase customers, thereby saving on selling expenses. Chrysler sold several minivans to the baby boom generation by offering perfectly positioned cars promoted with exaggerated trade-ins (above market value) for frequent buyers enrolled in an owner loyalty program. The Dodge Caravan and Plymouth Voyager minivans earned $6,100 profit margins on a $19,000 car, and $30,000 Chrysler Town and Country minivans earned almost $10,000. These were extraordinary profit margins in the cutthroat competitive mass market domestic auto industry of the 1990s. At the peak of the baby boom generation’s child-rearing years, Chrysler Corporation minivans were the highest volume vehicle sold in America (569,449 in 1993).6

Positioning of a product in the customer’s mind itself is quite important. As purchas- ing power rises, any goods and services that the target households perceive as inferior to some preferable substitute will likely experience declining unit sales. Marketers often therefore go to extraordinary lengths to create product images and customer associations to which the target households aspire. But this objective presents quite a challenge because aspirant good perceptions are very sensitive to culture and sociodemographic complexities.

THE PRICE ELASTICITY OF DEMAND From a decision-making perspective, any firm needs to know not only the direction but also the magnitude effects of changes in the determinants of demand. Some of these fac- tors are under the control of management, such as price, advertising, product quality,

WHAT WENT RIGHT • WHAT WENT WRONG

Chevy Volt5

The hybrid gasoline-electric Toyota Prius at $24,000 proved to be an aspirant good to 20-somethings, but every product is inferior to somebody. Yuppies, for example, re- vealed a preference for the Toyota Camry, Honda Accord, Chevy Malibu, and BMW 3 series at $28,000 to $38,000, even though the technology was not hybrid. Chevrolet hoped to capture the more green-conscious young profes- sionals with their plug-in hybrid the Chevy Volt at a planned price point of $34,000. But the 1,300 lithium-ion batteries needed to power the Volt proved to be $18,000 more expensive than Chevrolet had anticipated.

The problem is that the Volt’s original target market will likely have insufficient disposable income to purchase

at $52,000 ($34,000 + $18,000). And at that high a price point, the Lexus HS small SUV will be available in 2010 as a plug-in hybrid. Business owners with more money to spend will also likely reject the Chevy Volt in favor of the $75,000 plug-in hybrid Tesla that out-accelerates a Ferrari and is doing very well among Silicon Valley entre- preneurs. The positioning of the Chevy Volt appears problematic for attracting any sizeable customer base.

5“Briefing: The Electrification of Motoring,” The Economist (September 5, 2009), p. 75.

6“Iacocca’s Minivan: How Chrysler Succeeded in Creating One of the Most Profitable Products of the Decade,” Fortune (May 30, 1994), p. 112.

Chapter 3: Demand Analysis 69

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

and customer service. Other demand determinants, including disposable income and the prices of competitors’ products, are outside the direct control of the firm. Nevertheless, effective demand management still requires that the firm be able to measure the magni- tude of the impact of changes in these variables on quantity demanded.

Price Elasticity Defined The most commonly used measure of the responsiveness of quantity demanded or sup- plied to changes in any of the variables that influence the demand and supply functions is elasticity. In general, elasticity should be thought of as a ratio of the percentage change in quantity to the percentage change in a determinant, ceteris paribus (all other things remaining unchanged). Price elasticity of demand (ED) is therefore defined as:

ED = %ΔQ %ΔP

= ΔQ ΔP

× Base P Base Q

, ceteris paribus [3.1]

where

ΔQ = change in quantity demanded ΔP = change in price:

The final terms in Equation 3.1 show that price elasticity depends on the inverse of the slope of the demand curve ΔQ/ΔP (i.e., the partial sensitivity of demand in the target market to price changes, holding all other determinants of demand unchanged) times the

Example Pizza Hut and Ford Dealers Respond to Deficient Demand Pizza Hut anticipates a purchase frequency of 60 pizzas per night at a price of $9 and plans their operations accordingly. When fewer than the 60 customers arrive on a given evening, the Pizza Hut franchise does something very different than a Ford auto dealer might in similar circumstances. The restaurant slashes orders; fewer pizza dough balls are flattened and spun out and baked. Instead of slashing prices in the face of deficient demand, restaurants order less production and in- crease the size of their servings. Why is that?

One insight hinges on the lack of customer traffic in a restaurant that might be attracted on short notice by a given discount. The demand by Pizza Hut customers, in other words, is not very price sensitive while the final preparation stages of Pizza Hut supply are quite flexible and can be adjusted easily. In contrast, in the auto business, customer demand can be stimulated on short notice by sharp price dis- counts, while the supply schedule at the end of the model year is very inflexible. Ford Motor assembles and ships a number of cars in response to firm orders by their retail dealers and then insists on a no-returns policy. Thereafter, in the face of deficient demand (below the planned rate of sale), Ford dealerships tend to slash their asking prices to clear excess inventory.

In sum, auto dealerships adopt price discounts as their primary adjustment mechanism, while restaurants slash orders. Fundamentally, what causes the differ- ence in these two businesses? In the one case, quantity demanded is very price sen- sitive and quantity supplied is not (retail autos at the end of the model year). In the other case, demand is price insensitive, and supply is quite flexible (restaurants). This difference is characterized by the price elasticity of demand and supply.

ceteris paribus Latin for “all other things held constant.”

price elasticity of demand The ratio of the percentage change in quantity demanded to the percentage change in price, assuming that all other factors influencing demand remain unchanged. Also called own price elasticity.

70 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

price point positioning P where elasticity is calculated for Q unit sales on the demand curve. Because of the law of demand (i.e., the inverse relationship between price and quantity demanded), the sign of the own price elasticity will always be negative.

When the percentage change in price (the denominator in the first term of Equation 3.1) exceeds the percentage change in Q (the numerator), price elasticity calculates as a fraction, less than one in absolute value. This lack of demand responsiveness is described as “inelastic demand.” When the reverse holds,

|%ΔQ| > |%ΔP| → |εp| > 1

demand is described as “elastic.” Because higher price points (and lower baseline Q) re- sult in higher and higher elasticity, eventually, at high enough prices, all linear demand curves become elastic.

Example Price Elasticity at Various Price Points along a Linear Demand Curve for Gasoline To illustrate, the demand for gasoline in Table 3.1 is estimated from U.S. Con- sumer Expenditure Survey data and varies markedly by the type of household. For two-person urban households with no children, demand is very price inelastic, measuring −0.56 at lower price points like $2.50 per gallon. This means, using Equation 3.1, that if price rises by 40 percent (say, from $2.00 to $3.00), gallons consumed per week will fall by 22 percent (20 to 16 gallons):

−4 gallons=18 gallons +$1:00=$2:50

= −22% +40%

= −0:56

At still higher prices, like $3.00, elasticity has been measured at –0.75. The rea- son is that even if the incremental decline in desired rate of purchase remains ap- proximately 4 gallons for each $1 price increase, the base quantity will have fallen from 16 to approximately 12 gallons, so the percentage change in Q will now increase substantially. Similarly, the percentage change in price from another $1 increase will decline substantially because the price base is much bigger at $3.50 than at $2.50.

Figure 3.3 illustrates the price rise that occurred from January 2008 to July 2008 as gasoline spiked from $3.00 per gallon to $4.00 in many cities across the United States. Despite the peak driving season, quantity demanded collapsed from 14 gal- lons per week the previous summer (at $3 prices per gallon) to 11.5. Mass transit ridership skyrocketed, growing by 20 percent in that one summer in several U.S. cities. Discretionary Sunday drives ended. That summer of 2008, three weekend- in-a-row trips to the shore stopped. Americans decided to cut out essentially all discretionary driving. The price elasticity calculated for this $3.00 to $4.00 price range was:

−4:5 gallons=13:75 +$1:00=$3:50

= −33% +29%

= −1:14

For the first time in U.S. transportation history, the demand for gasoline was price elastic! What in previous summers had been a weekly expenditure by urban households of $48 ($3 × 16 gallons) fell to $46 ($4 × 11.5 gallons). With |%ΔQ| = |−33%| > |%ΔP| = +29%, consumer expenditure on gasoline and the total retail revenue from gasoline sales actually declined as prices shot up.

Chapter 3: Demand Analysis 71

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Arc Price Elasticity The arc price elasticity of demand is a technique for calculating price elasticity between two prices. It indicates the effect of a change in price, from P1 to P2, on the quantity demanded. The following formula is used to compute this elasticity measure:

ED =

Q2 − Q1 Q2 + Q1

2

� � P2 − P1 P2 + P1

2

� � = Q2 − Q1 P2 − P1

· P2 + P1 Q2 + Q1

= ΔQ ΔP

P2 + P1 Q2 + Q1

[3.2]

where Q1= quantity sold before a price change Q2= quantity sold after a price change P1= original price P2= price after a price change

The fraction (Q2 +Q1)/2 represents average quantity demanded in the range overwhich the price elasticity is being calculated. (P2 + P1)/2 also represents the average price over this range.

Because the slope remains constant over the entire schedule of linear demand but the value of (P2 + P1)/(Q2 + Q1) changes, price elasticity at higher prices and smaller volume is therefore larger (in absolute value) than price elasticity for the same product and same de- manders at lower price points and larger volume. Equation 3.2 can be used to compute a price that would have to be charged to achieve a particular level of sales. Consider the NBA Corporation, which had monthly basketball shoe sales of 10,000 pairs (at $100 per pair) before a price cut by its major competitor. After this competitor’s price reduction, NBA’s sales declined to 8,000 pairs a month. From past experience, NBA has estimated the price elasticity of demand to be about −2.0 in this price-quantity range. If the NBA wishes to restore its sales to 10,000 pairs a month, determine the price that must be charged.

Letting Q2 = 10,000, Q1 = 8,000, P1 = $100, and ED = −2.0, the required price, P2, may be computed using Equation 3.2:

−2:0 =

10,000 − 8,000 ð10,000 + 8,000Þ=2

P2 − $100 ðP2 + $100Þ=2

P2 = $89:50

A price cut to $89.50 would be required to restore sales to 10,000 pairs a month.

FIGURE 3.3 Retail Gasoline Price per Gallon

JAN. 2008

U SD

/g al

lo n

0

1

2

3

4

5

APR. JULY OCT. JAN. 2009 APR. JULY 2009

Source: Energy Information Administration.

72 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Point Price Elasticity The preceding formulas measure the arc elasticity of demand; that is, elasticity is com- puted over a discrete range of the demand curve or schedule. Because elasticity is nor- mally different at each price point, arc elasticity is a measure of the average elasticity over that range.

By employing some elementary calculus, the elasticity of demand at any price point along the curve may be calculated with the following expression:

ED = ∂QD ∂P

· P QD

[3.3]

where

∂QD ∂P

= the partial derivative of quantity with respect to price ðthe inverse of the slope of the demand curveÞ

QD = the quantity demanded at price P P = the price at some specific point on the demand curve

Equation 3.3 consists of two magnitudes: (1) a partial derivative effect of own price changes on the desired rate of purchase (QD/t), and (2) a price point that (along with a baseline QD) determines the percentage change.

The daily demand function for Christmas trees at sidewalk seasonal sales lots in mid- December can be used to illustrate the calculation of the point price elasticity. Suppose that demand can be written algebraically as quantity demanded per day:

QD = 45,000 − 2,500P + 2:5Y [3.4]

If one is interested in determining the point price elasticity when the price (P) is equal to $40 and per capita disposable personal income (Y) is equal to $30,000, taking the partial derivative of Equation 3.4 with respect to P yields:

∂QD ∂P

= −2,500 trees per dollar

Substituting the relevant values of P and Y into Equation 3.4 gives

QD = 45,000 − 2,500ð40Þ + 2:50ð30,000Þ = 20,000 From Equation 3.4, one obtains

ED = −2,500 trees $

$40 20,000 trees

� � = −5:0

Interpreting the Price Elasticity: The Relationship between the Price Elasticity and Revenues Once the price elasticity of demand has been calculated, it is necessary to interpret the meaning of the number obtained. Price elasticity may take on values over the range from 0 to −∞ (infinity) as indicated in Table 3.2.

When demand is unit elastic, a percentage change in price P is matched by an equal percentage change in quantity demanded QD. When demand is elastic, a percentage change in P is exceeded by the percentage change in QD. For inelastic demand, a percentage change in P results in a smaller percentage change in QD. The theoretical extremes of perfect elas- ticity and perfect inelasticity are illustrated in Figure 3.4. AAA-grade January wheat sells on the Kansas City spot market with perfectly elastic demand facing any particular grain dealer; Panel A illustrates this case. Addicted smokers have almost perfectly inelastic de- mand; their quantity demanded is fixed no matter what the price, as indicated in Panel B.

Chapter 3: Demand Analysis 73

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The price elasticity of demand indicates immediately the effect a change in price will have on the total revenue (TR) = total consumer expenditure. Table 3.3 and Figure 3.5 illustrate this connection.

FIGURE 3.4 Perfectly Elastic and Inelastic Demand Curves

D

PANEL A Perfectly elastic

|ED| = –∞

D�

Quantity demanded (units)

D

PANEL B Perfectly inelastic

|ED| = 0

D�

Quantity demanded (units)

Pr ic

e ($

/u ni

t)

Pr ic

e ($

/u ni

t)

TABLE 3.2 PRICE ELASTICITY OF DEMAND IN ABSOLUTE VALUES

RANGE DESCRIPTION

ED = 0 Perfectly inelastic

0 < |ED| < 1 Inelastic

|ED| = 1 Unit elastic

1 < |ED| < ∞ Elastic

|ED| = ∞ Perfectly elastic

TABLE 3.3 THE RELATIONSHIP BETWEEN ELASTICITY AND MARGINAL

REVENUE

PRICE, P ($/UNIT)

QUANTITY, QD (UNITS)

ELASTICITY ED

TOTAL REVENUE P · QD ($)

MARGINAL REVENUE ($/UNIT)

10 1 10

9 2 −6.33 18 8

8 3 −3.40 24 6

7 4 −2.14 28 4

6 5 −1.44 30 2

5 6 −1.00 30 0

4 7 −0.69 28 −2

3 8 −0.46 24 −4

2 9 −0.29 18 −6

1 10 −0.15 10 −8

74 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

When demand elasticity is less than 1 in absolute value (i.e., inelastic), an increase (decrease) in price will result in an increase (decrease) in (P · QD). This occurs because an inelastic demand indicates that a given percentage increase in price results in a smaller percentage decrease in quantity sold, the net effect being an increase in the total expendi- tures, P · QD. When demand is inelastic—that is, |ED| < 1—an increase in price from $2 to $3, for example, results in an increase in total revenue from $18 to $24.

In contrast, when demand is elastic—that is, |ED| > 1—a given percentage increase (decrease) in price is more than offset by a larger percentage decrease (increase) in

FIGURE 3.5 Price Elasticity over Demand Function

Quantity (units)

0

0

P2

MC

P1

|ED| > 1 (elastic)

|ED| = 1 (unitary)

|ED| < 1 (inelastic)

Q2 Q1

Q2

MR

Quantity (units)

Total revenue

D

D�

A

B

Pr ic

e, m

ar gi

na l r

ev en

ue (

$/ un

it )

To ta

l r ev

en ue

a nd

t ot

al p

ro fi

t ($

)

Profit

TRmax

� max

Q1

Chapter 3: Demand Analysis 75

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

quantity sold. An increase in price from $9 to $10 results in a reduction in total con- sumer expenditure from $18 to $10 (again, see Table 3.3).

When demand is unit elastic, a given percentage change in price is exactly offset by the same percentage change in quantity demanded, the net result being a constant total consumer expenditure. If the price is increased from $5 to $6, total revenue would re- main constant at $30, because the decrease in quantity demanded at the new price just offsets the price increase (see Table 3.3). When the price elasticity of demand |ED| is equal to 1, the total revenue function is maximized. In the example, total revenue equals $30 when price P equals either $5 or $6 and quantity demanded QD equals ei- ther 6 or 5.

As shown in Figure 3.5, when total revenue is maximized, marginal revenue equals zero. At any price higher than P2, the demand function is elastic. Hence, successive equal percentage increases in price may be expected to generate higher and higher percentage decreases in quantity demanded because the demand function is becoming increasingly elastic. Alternatively, successive equal percentage reductions in price be- low P2 may be expected to generate ever lower percentage increases in quantity de- manded because the demand function is more inelastic at lower prices. Again, then, price P2 is a pivot point for which total revenue is maximized where marginal revenue equals zero.

To summarize, a change in TR arises from two sources: a change in prices and a change in unit sales. Specifically,

ΔTR = (ΔP × Q0) + (ΔQ × P1)

Then dividing by P1 × Q0 yields the very useful expression,

%ΔTR = ðΔP=P1Þ + ðΔQ=Q0Þ %ΔTR = %ΔP + %ΔQ

[3.5]

That is, the percentage effect on sales revenue is the signed summation of the percentage change in price and in unit sales.

If Johnson & Johnson lowers the price on BAND-AID bandages 10 percent and sales revenue goes up 24 percent, we can conclude that unit sales must have risen 34 percent, because applying Equation 3.5

24% = −10% + 34%

The relationship between a product’s price elasticity of demand and the marginal rev- enue at that price point is one of the most important in managerial economics. This re- lationship can be derived by analyzing the change in revenue resulting from a price change. To start, marginal revenue is defined as the change in total revenue resulting from lowering price to make an additional unit sale. Lowering price from P1 to P2 in Figure 3.4 to increase quantity demanded from Q1 to Q2 results in a change in the initial revenue P1AQ10 to P2BQ20. The difference in these two areas is illustrated in Figure 3.5 as the two shaded rectangles. The horizontal shaded rectangle is the loss of revenue caused by the price reduction (P2 − P1) over the previous units sold Q1. The vertical shaded rectangle is the gain in revenue from selling (Q2 − Q1) additional units at the new price P2. That is, the change in total revenue from lowering the price to sell another unit can always be written as follows:

MR = ΔTR ΔQ

= P2ðQ2 − Q1Þ + ðP2 − P1ÞQ1

ðQ2 − Q1Þ [3.6]

marginal revenue The change in total revenue that results from a one-unit change in quantity demanded.

76 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

where P2(Q2 − Q1) is the vertical shaded rectangle and (P1 − P2)Q1 is the horizontal shaded rectangle. Rearranging, we have:

MR = P2 + ðP2 − P1ÞQ1 ðQ2 − Q1Þ

= P2 1 + ðP2 − P1ÞQ1 ðQ2 − Q1ÞP2

� �

MR = P2 1 + ΔPQ1 ΔQP2

� �

The ratio term is the inverse of the price elasticity at the price point P2 using the quan- tity Q1. For small price and quantity changes, this number closely approximates the arc price elasticity in Equation 3.2 between P1 and P2. Therefore, the relationship between marginal revenue and price elasticity can be expressed algebraically as follows:

MR = P 1 + 1 ED

� � [3.7]

Using this equation, one can demonstrate that when demand is unit elastic, marginal revenue is equal to zero. Substituting ED = −1 into Equation 3.7 yields:

MR = P 1 + 1 −1

� � = Pð0Þ = 0

A commission-based sales force and the management team have this same conflict; salespeople often develop ingenious hidden discounts to try to circumvent a company’s list pricing policies. Lowering the price from P1 to P2 to set |ED| = 1 will always maxi- mize sales revenue (and therefore, maximize total commissions).

The fact that total revenue is maximized (and marginal revenue is equal to zero) when |ED| = 1 can be shown with the following example: Custom-Tees, Inc., operates a

Example Content Providers Press Publishing Companies to Lower Prices Entertainment and publishing companies pay songwriters, composers, playwrights, and authors a fixed percentage of realized sales revenue as a royalty. The two groups often differ as to the preferred price and unit sales. Referring to Figure 3.5, total revenue can be increased by lowering the price any time the quantity sold is less than Q2. That is, at any price above P2 (where marginal revenue re- mains positive), the total revenue will continue to climb only if prices are lowered and additional units sold. Songwriters, composers, playwrights, patent holders, and authors often therefore press their licensing agents and publishers to lower prices whenever marginal revenue remains positive—that is, to the point where demand is unit elastic. The publisher, on the other hand, will wish to charge higher prices and sell less quantity because operating profits arise only from marginal revenue in excess of variable cost per unit. Unless variable cost is zero, the publisher always wants a positive marginal revenue and therefore a price greater than P2 (for example, P1).

Chapter 3: Demand Analysis 77

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

kiosk in Hanes Mall where it sells custom-printed T-shirts. The demand function for the shirts is

QD = 150 − 10P [3.8]

where P is the price in dollars per unit and QD is the quantity demanded in units per period.

The inverse demand curve can be rewritten in terms of P as a function of QD:

P = 15 − QD 10

[3.9]

Total revenue (TR) is equal to price times quantity sold:

TR = P · QD

= 15 − QD 10

� � QD

= 15QD − Q2D 10

Marginal revenue (MR) is equal to the first derivative of total revenue with respect to QD:

MR = dðTRÞ dQD

= 15 − QD 5

To find the value of QD where total revenue is maximized, set marginal revenue equal to zero:7

MR = 0

15 − QD 5

= 0

Q*D = 75 units

Substituting this value into Equation 3.9 yields:

P* = 15 − 75 10

= $7:50 per unit

Thus, total revenue is maximized at Q*D = 75 and P* = $7.50. Checking:

ED = ∂QD ∂P

· P QD

= ð−10Þ ð7:5Þ 75

= −1

|ED| = 1

The Importance of Elasticity-Revenue Relationships Elasticity is often the key to marketing plans. A product-line manager will attempt to maximize sales revenue by allocating a marketing expense budget among price promo- tions, advertising, retail displays, trade allowances, packaging, and direct mail, as well as in-store coupons. Knowing whether and at what magnitude demand is responsive to each of these marketing initiatives depends on careful estimates of the various demand elasticities of price, advertising, packaging, promotional displays, and so forth.

7To be certain one has found values for P and QD, where total revenue is maximized rather than minimized, check the second derivative of TR to see that it is negative. In this case d2TR/dQ2D = −1/5, so the total revenue function is maximized.

78 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

One managerial insight that this VW example illustrates is that firms should always seek to raise prices for products in the inelastic range of their demand. To lower prices in such a range would both increase costs (of producing and distributing additional out- put demanded at the lower prices) and also decrease revenue. It is far better to approach unit elasticity from below, raising prices, thereby increasing revenue while at the same time saving the production and distribution costs. In fact, astute profit-maximizing firms will carry these price increases right on into the elastic range beyond the point of maxi- mum revenue and unit elasticity (above and beyond point B at P2 and Q2 in Figure 3.5).

Starting from the other direction at zero output, a profit-maximizing firm will only lower price to increase revenue as long as the incremental change in total revenue (the MR in Figure 3.5) exceeds the positive change in total variable cost (labeled height MC).

Example VW’s Invasion of North America When Volkswagen (VW) entered the U.S. market with its no-frills automobile, the original Beetle, European excess inventory had stockpiled at ports and road ter- minals awaiting export to a new market. Consequently, VW focused for a time on any demand stimulus that would increase revenue. In the U.S. market, VW had no dealer network and initially provided sales and service only at the docks in Bayonne, New Jersey; Charleston, South Carolina; and Houston, Texas. General Motors and Ford were developing compact cars as well, so VW decided to enter the market at what appeared to be a ridiculously low promotional price of $6,500 (in 2010 U.S. dollars). GM and Ford were convinced that nothing that cheap would be perceived as a real car. Two years later, a 25 percent price increase was intro- duced. Although VW lost some potential customers at $8,100, the extra $1,600 per car on all the cars they continued to sell easily offset the revenue loss from a few lost sales that could have been made at the original $6,500 price. Compare the long horizontal shaded area in Figure 3.6 that represents increased revenue to the verti- cal shaded area of lost revenue. The price elasticity of demand was in the inelastic range of demand. By the fourth year, VW had raised the price another 20 percent to $9,600, and again revenue rose.

Finally, at $9,600 (in 2010 dollars), the extra receipts from the price increase across all remaining sales were just sufficient to offset the loss in revenue from the lost sales. At $9,600, price elasticity had reached the unit elastic price point. Volkswagen then proceeded to build a U.S. dealer network. These changes in- creased the potential size of the market in the United States and shifted the de- mand for Volkswagen products to the right. At the same price of $9,600, with a dealer network and a larger quantity base, the measured price elasticity then de- clined (again into the inelastic range), and Volkswagen was again in a position to raise price.

In 1968, 562,000 Beetles were sold at a price of $1,500 (about $11,500 today), and revenue again increased (to $843 million). Although the product remained very inexpensive for a new car, the Highway Safety Act of 1966, plus Ralph Nader’s crusade against small rear-engine cars, plus low gasoline prices caused consumers to begin losing interest in Beetles and start buying large numbers of Mustangs, Ca- maros, and a new more powerful Super Beetle. In 1969, revenue and price of the Beetle increased one last time. At $1,800, revenue was $968 million ($1,800 × 538,000 units sold). That is, demand remained in the inelastic range since a price increase from $1,500 to $1,800 had resulted in higher total revenue.

Chapter 3: Demand Analysis 79

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

That is, the profit-maximizing output will always occur in the elastic region of the firm’s demand at a price above the unit elastic price point.

Factors Affecting the Price Elasticity of Demand The market demand for bedroom furniture is extremely price elastic (−3.04), whereas the market demand for coffee (−0.16) is extremely price inelastic. As shown in Table 3.4, price elasticities vary greatly among different products and services.9 Some of the factors that account for the differing responsiveness of consumers to price changes are examined next.

FIGURE 3.6 Raising Price with Demand in the Inelastic Range (Actual Prices, Various Years)

$ 1,350

$ 1,200

$ 1,000 $ 800

ED > –1 (e.g., –0.87)

ED < –1 (e.g., –2.4)

ED = –1

ED > –1 (e.g., –0.33)

Quantity in thousands (autos)

Pr ic

e ($

/u ni

t)

D with dealer network

D without dealer network

384 538 562

$1,800

$1,500

+ + + + + –

Example Price Elasticity Estimates for Coffee Vary by Price8

A study by Huang, Siegfried, and Zardoshty on the demand for coffee confirms the relationship between price levels and the price elasticity of demand shown in Figure 3.5. They found that the price elasticity of demand ranged from −0.16 for off-peak up to −0.89 for the peak price level. Thus, coffee users are nearly nine times more sensitive to price changes at high prices than at low price levels. Because these are market-level, not firm-level, elasticity estimates, observing price elasticities less than 1.0 in absolute value does not contradict the managerial insight that is conveyed by Figure 3.6—that is, firms will always increase price until demand is no longer in the inelastic region.

8Cliff J. Huang, J.J. Siegfried, and Farangis Zardoshty, “The Demand for Coffee in the United States, 1963–1977,” Quarterly Review of Economics and Business 20, no. 2 (Summer 1980), pp. 36–50. Another more recent estimate of the demand elasticity for coffee can be found in Albert A. Okunade, “Functional Forms and Habits Effects in the U.S. Demand for Coffee,” Applied Economics (November 1992).

9RTi is one of a number of consulting firms that estimate price elasticities; access at www.rtiresearch.com

80 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE 3.4 EMPIRICAL PRICE ELASTICITIES

COMMODITY (GOOD/SERVICE)

PRICE ELASTICITY OF MARKET DEMAND

COMMODITY (GOOD/SERVICE)

PRICE ELASTICITY OF MARKET DEMAND

Alcoholic beverages Furniture −3.04m

Beer −0.84e Glassware/China −1.20c

Wine −0.55e Household appliances −0.64c

Liquor −0.50e International air transportation

Apparel United States/Europe −1.25h

Market −1.1m Canada/Europe −0.82h

Firms −4.1m Outdoor recreation −0.56f

Coffee School lunches −0.47d

Regular −0.16b Shoes −0.73c

Instant −0.36b Soybean meal −1.65j

Credit charges on bank cards −2.44l Telephones −0.10a

Dental visits Textiles

Adult males −0.65g Market −1.5m

Adult females −0.78g Firms −4.7m

Children −1.40g Tires −0.60c

Food Tobacco products −0.46c

Market −1.0n Tomatoes −2.22k

Firms −3.8n Wool −1.32i

aD. Cracknell and M. Knott, “The Measurement of Price Elasticities—The BT Experience,” International Journal of Forecasting 11 (1995), pp. 321–329. bCliff J. Huang, John J. Siegfried, and Farangis Zardoshty, “The Demand for Coffee in the United States, 1963–1977,” Quarterly Re- view of Economics and Business 20, no. 2 (Summer 1980), pp. 36–50. cH. S. Houthakker and Lester D. Taylor, Consumer Demand in the United States, 2nd ed. (Cambridge, MA: Harvard University Press, 1970). dGeorge A. Braley and P.E. Nelson Jr., “Effect of a Controlled Price Increase on School Lunch Participation: Pittsburgh, 1973,” Amer- ican Journal of Agricultural Economics (February 1975), pp. 90–96. eDale Heien and Greg Pompelli, “The Demand for Alcoholic Beverages: Economic and Demographic Effects,” Southern Economic Journal (January 1989), pp. 759–769. fRussel L. Gum and W.E. Martin, “Problems and Solutions in Estimating the Demand for and Value of Rural Outdoor Recreation,” American Journal of Agricultural Economics (November 1975), pp. 558–566. gWillard G. Manning Jr. and Charles E. Phelps, “The Demand for Dental Care,” The Bell Journal of Economics 10, no. 2 (Autumn 1979), pp. 503–525. hJ.M. Cigliano, “Price and Income Elasticities for Airline Travel: The North Atlantic Market,” Business Economics (September 1980), pp. 17–21. iC.E. Ferguson and M. Polasek, “The Elasticity of Import Demand for Raw Apparel Wool in the United States,” Econometrica 30 (1962), pp. 670–699. jH. Knipscheer, L. Hill, and B. Dixon, “Demand Elasticities for Soybean Meal in the European Community,” American Journal of Ag- ricultural Economics (May 1982), pp. 249–253. kDaniel B. Suits, “Agriculture,” in Structure of American Industry, 7th ed., ed. W. Adams (New York: Macmillan, 1986). lJ. Starvins, “Can Demand Elasticity Explain Sticky Credit Card Rates?” New England Economic Review (July/August 1996), pp. 43–54. mRichard D. Stone and D.A. Rowe, “The Durability of Consumers’ Durable Goods,” Econometrica 28 (1960), pp. 407–416. nM.D. Shapiro, “Measuring Market Power in U.S. Industry,” NBER Working Paper, No. 2212 (1987).

Chapter 3: Demand Analysis 81

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Availability and Closeness of Substitutes The most important determi- nant of the price elasticity of demand is the availability and perceived closeness of sub- stitutes. The greater the number of substitute goods within the relevant market, the more price elastic is the demand for a product because a customer can easily shift from one good to a close substitute good if the price rises.10 The price elasticity of demand for Johnson & Johnson’s BAND-AID bandages is high because numerous companies offer a nearly identical product. The closeness of substitutes is a related but different concept. Intravenous feeding systems have distant substitutes for hospital patients in shock or otherwise unable to digest food, so the price elasticity of demand for that product is low. Sunbeam white bread, however, has very close substitutes from in-store bakeries and from numerous competitors in branded breads. So, in that case, price elasticity of demand is high.

Percentage of the Consumer’s Budget The demand for relatively high-priced goods tends to be more price elastic than the demand for inexpensive items because ex- pensive items account for a greater proportion of a person’s budget. Consequently, we would expect the demand for apartment housing to be more price elastic than the de- mand for children’s toys. The greater the percentage of the budget spent on a good, the larger the purchasing power released by any given price reduction or absorbed by any given price increase. And the larger this “income effect,” the greater the price elasticity for income-superior goods.

INTERNATIONAL PERSPECTIVES

Free Trade and the Price Elasticity of Demand: Nestlé Yogurt

The 1990s were characterized by an explosion of free trade agreements among important trading partners. The Europe 1992 plan virtually eliminated trade bar- riers, and goods flowed freely and without tariffs from one European country to another. Increasing standardization of products in these markets further reduced trading barriers. In 1994, the North Ameri- can Free Trade Agreement (NAFTA) was ratified by the United States, Canada, and Mexico, and the Gen- eral Agreement on Tariffs and Trade (GATT) was implemented, leading to a worldwide reduction in tariffs and other trade barriers. In 2001, the United States launched the Doha Round of free trade talks that continue to the present.

What are the implications of these reduced trade barriers for estimates of price elasticity of demand? Free trade results in an effective increase in the number of substitute goods that are available to

consumers and businesses in any country. Conse- quently, as barriers to free trade come down, de- mand will become more price elastic for goods that historically have not been able to flow easily between countries. Nestlé’s yogurt and custard pro- ducts now travel from manufacturing sites in the British Midlands to Milan, Italy, in 17 hours, whereas the customs processing and transportation bottlenecks once required 38 hours. Similarly, iron forging of crankshafts and engine blocks for U.S. autos now occurs primarily in Mexico, and trans- missions for Detroit are often constructed as subas- semblies in Japan. The winners in this globalization process should be consumers, who will have a wider variety of products to choose from at ever more competitive prices. The losers will be those firms that cannot compete in a global market on the basis of cost, quality, and service.

10The demand for durable goods tends to be more price elastic than the demand for nondurables. This is true because of the ready availability of a relatively inexpensive substitute in many cases—i.e., the repair of a used television, car, or refrigerator, rather than buying a new one.

82 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Positioning as Income Superior How a product is positioned to the target customers has much to do with whether the release in purchasing power associated with a price discount will result in a substantial, moderate, or trivial increase (or pos- sibly even a decrease) in unit sales. In the “Cash for Clunkers” federal subsidy pro- gram in 2009, some but not all replacement vehicles were eligible for a $5,000 discount if the customer turned in an older car and bought a more fuel-efficient one. Since the program made no distinction between high-value used cars and almost- worthless junk cars, most participants turned in clunkers and received a $4,000 to near $5,000 increase in their purchasing power. The new automobiles whose demand increased substantially were income-superior full-size family sedans and hybrid SUVs, not the mid-size and economy cars that would have significantly improved fuel efficiency. This should have come as no surprise because many of the fuel- efficient cars like the Ford Fiesta, Ford Focus, and Chevy Geo are seen as income- inferior products.

Time Period of Adjustment To respond to a price decrease, potential customers must first learn about the discount and then incur the cost of adjusting their own sched- ules to complete a purchase during the sale period. Because both search and adjustment costs for consumers are higher if sale prices last only a few minutes, the demand re- sponse to price changes is diminished the shorter the time period of adjustment. Predict- able end-of-model-year promotions in the auto industry lasting throughout the month of August stimulate much more elastic demand than unannounced “Midnight Madness” sales that last only a few hours.

The long-run demand for many products also tends to be more elastic than short-run demand because of the increase in the number of effective substitutes that become avail- able over time. For example, the only available alternatives for gasoline consumption in the short run are not taking a trip or using some form of public transportation. Over time, as consumers replace their cars, they find another excellent substitute for gasoline—namely, more fuel-efficient vehicles.

THE INCOME ELASTICITY OF DEMAND Among the variables that affect demand, disposable income of the target customers is often one of the most important. Business analysts compute an income elasticity of demand analogous to the price elasticity of demand.

Income Elasticity Defined Income elasticity of demand can be expressed as

Ey = %ΔQD %ΔY

, ceteris paribus [3.10]

where

ΔQD = change in quantity demanded ΔY = change in income

Various measures of income can be used in the analysis. One commonly used mea- sure is consumer disposable income, calculated on an aggregate, household, or per capita basis.

income elasticity The ratio of the percentage change in quantity demanded to the percentage change in income, assuming that all other factors influencing demand remain unchanged.

Chapter 3: Demand Analysis 83

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Arc Income Elasticity The arc income elasticity is a technique for calculating income elasticity between two income levels. It is computed as

Ey =

Q2 − Q1 ðQ2 + Q1Þ=2 Y2 − Y1

ðY2 + Y1Þ=2 =

ΔQ ΔY

ðY1 + Y2Þ ðQ1 + Q2Þ [3.11]

where

Q2 = quantity sold after an income change Q1 = quantity sold before an income change Y2 = new level of income Y1 = original level of income

For example, assume that an increase in disposable personal income in Rhode Island from $1.00 billion to $1.10 billion is associated with an increase in boat sales in the state

Example Targeting a Direct Mail Coupon at a Ford Dealership One of the key steps for successful auto dealerships is to develop an extensive data- base on prospective and repeat purchase customers. These databases are often used to target promotional material from the manufacturer to specific local households thought to be most likely to respond to direct mail advertising. Suppose Ford decides to offer select households $5,000 off several models. To emphasize the role of positioning in demand analysis, let’s focus on the effect of these direct mail coupons on the sporty Ford Taurus four-door sedan. Suppose Taurus is currently experiencing inventory overhang with production runs that have exceeded recent sales.

To whom in the local dealer’s database should we target a “$5,000 off” coupon? One choice is a newly married couple attending community college who recently shopped at our dealership for a subcompact Ford Focus. The second choice is a young professional couple of German heritage who already own a four-year-old 85,000 mile Taurus as their commuter car along with a vintage BMW. Third choice is a French immigrant couple, both management consultants, who ride mass transit but recently test drove a Volvo sedan. The final choice is a retired couple who once owned a Taurus and recently purchased their third Ford Crown Vic, a full-size sedan.

In selecting a target customer for the direct mail coupon, the dealership’s mar- keting team assesses switching costs, positioning, and likely spending power. Al- though the newlyweds would aspire to a sporty midsize sedan like the Taurus, their anticipated uses and spending power do not match those of typical Taurus customers. The young French immigrant couple aspires already to a Volvo, imply- ing Taurus positioning is perceived as inferior. The retired couple has substantial brand loyalty to Ford but high switching cost, given their full-size uses and needs. Not so with the Germanic yuppies who already spend large amounts of disposable income on autos and may continue to see Taurus as an aspirant good.

The projected demand of the German-heritage couple is likely to be most elastic to the price change on offer. Their switching costs are low because they presently consume an almost identical substitute. They will likely find the new model Taurus an aspirant good. And the percentage of their budget spent on automobiles is larg- est of the four potential target households.

84 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

from 5,000 to 6,000 units. Determine the income elasticity over this range. Substituting the relevant data into Equation 3.11 yields

Ey =

6,000 − 5,000 ð6,000 + 5,000Þ=2 $1:10 − $1:00

ð$1:10 + $1:00Þ=2 =

1,000 $0:10

ð$2:10Þ ð11,000Þ

= 1:91

Thus, a 1 percent increase in income would be expected to result in a 1.91 percent in- crease in quantity demanded, ceteris paribus.

Point Income Elasticity The arc income elasticity measures the responsiveness of quantity demanded to changes in income levels over a range. In contrast, the point income elasticity provides a measure of this responsiveness at a specific point on the demand function. The point income elas- ticity is defined as

Ey = ∂QD ∂Y

· Y QD

[3.12]

where

Y = income QD = quantity demanded of some commodity ∂QD ∂Y

= the partial derivative of quantity with respect to income

The algebraic demand function for Christmas trees (Equation 3.4) introduced earlier in the chapter can be used to illustrate the calculation of the point income elasticity. Sup- pose one is interested in determining the point income elasticity when the price is equal to $40 and per capita personal disposable income is equal to $30,000. Taking the partial derivative of Equation 3.4 with respect to Y yields

∂QD ∂Y

= 2:50

Recall from the point price elasticity calculation described earlier in the chapter that substituting P = $40 and Y = $30,000 into Equation 3.4 gave QD equal to 20,000 units. Therefore, from Equation 3.12, one obtains

Ey = 2:50 $30,000 20,000

� � = 3:75

Thus, from an income level of $30,000, one could expect demand for Christmas trees to increase by 37.5 percent for each 10 percent increase in per capita disposable income, ceteris paribus.

Interpreting the Income Elasticity For most products, income elasticity is ex- pected to be positive; that is, Ey > 0. Such goods are referred to as income-superior goods. Those goods having a calculated income elasticity that is negative are called inferior goods. Inferior goods are purchased in smaller absolute quantities as the income of the consumer increases. Subcompact autos and such food items as canned mackerel or dried beans are frequently cited as examples of inferior goods.

Knowledge of the magnitude of the income elasticity of demand for a particular prod- uct is especially useful in forecasting unit sales of economic activity. In industries that

Chapter 3: Demand Analysis 85

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

produce goods having high income elasticities (such as new furniture), a major increase or decrease in economic activity will have a significant impact on demand. Knowledge of income elasticities is also useful in developing marketing strategies for products. For ex- ample, products having a high income elasticity can be promoted as being luxurious and stylish, whereas goods having a low income elasticity can be promoted as being economical.

Advertising Elasticity Advertising elasticity measures the responsiveness of sales to changes in advertising expenditures as measured by the ratio of the percentage change in sales to the percentage change in advertising expenditures.

Eadv = %ΔQD %ΔADV

, ceteris paribus

The higher the advertising elasticity coefficient EA, the more responsive sales are to changes in the advertising budget. An awareness of this elasticity measure may assist

Example Income Elasticities: Empirical Estimates Estimates of the income elasticity of demand have been made for a wide variety of goods and services, as shown in Table 3.5. Note that the income elasticities for goods that are often perceived as necessities (e.g., many food items and housing) are less than 1.0, whereas the income elasticities for items that are usually viewed as luxuries (e.g., European travel) are greater than 1.0.

TABLE 3.5 EMPIRICAL INCOME ELASTICITIES

COMMODITY (GOOD/SERVICE) INCOME ELASTICITY

European travel 1.91a

Apples 1.32b

Beef 1.05b

Chicken 0.28b

Dental visits

Adult males 0.61c

Adult females 0.55c

Children 0.87c

Housing (low-income renters) 0.22d

Milk 0.50a

Oranges 0.83a

Potatoes 0.15a

Tomatoes 0.24a

aJ.M. Cigliano, “Price and Income Elasticities for Airline Travel: The North Atlantic Market,” Business Economics (September 1980), pp. 17–21. bDaniel B. Suits, “Agriculture,” in Structure of American Industry, 7th ed., ed. W. Adams (New York: Macmillan, 1986). cWilland G. Manning Jr. and Charles E. Phelps, “The Demand for Dental Care,” Bell Journal of Economics 10, no. 2 (Autumn 1979), pp. 503–525. dElizabeth A. Roistacher, “Short-Run Housing Responses to Changes in Income,” American Eco- nomic Review (February 1977), pp. 381–386.

86 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

advertising or marketing managers in their determination of appropriate levels of adver- tising outlays relative to price promotions or packaging expenditures.

CROSS ELASTICITY OF DEMAND Another determinant of demand that substantially affects the demand for a product is the price of a related (substitute or complementary) product.

Cross Price Elasticity Defined The cross price elasticity of demand, Ecross, is a measure of the responsiveness of changes in the quantity demanded (QDA) of Product A to price changes for Product B (PB).

Ecross = %ΔQDA %ΔPB

, ceteris paribus [3.13]

where

ΔQDA = change in quantity demanded of Product A ΔPB = change in price of Product B

Interpreting the Cross Price Elasticity If the cross price elasticity measured between Products A and B is positive (as might be expected in our butter/margarine example or between such products as plastic wrap and aluminum foil), the two products are referred to as substitutes for each other. The higher the cross price elasticity, the closer the substitute relationship. A negative cross price elas- ticity, on the other hand, indicates that two products are complementary. For example, a significant decrease in the price of DVD players would probably result in a substantial increase in the demand for DVDs.

Antitrust and Cross Price Elasticities The number of close substitutes may be an important determinant of the degree of com- petition in a market. The fewer the number of close substitutes that exist for a product, the greater the amount of market power that is possessed by the producing or selling firm. An important issue in antitrust cases involves the appropriate definition of the rel- evant product market to be used in computing statistics of market control (e.g., market share). A case involving DuPont’s production of cellophane was concerned with this issue. Does the relevant product market include just the product cellophane or does it include the much broader flexible packaging materials market?

The Supreme Court found the cross price elasticity of demand between cellophane and other flexible packaging materials to be sufficiently high so as to exonerate DuPont from a charge of monopolizing the market. Had the relevant product been considered to be cellophane alone, DuPont would have clearly lost, because it produced 75 percent of all cellophane output and its only licensee, Sylvania, produced the rest. But when other flexible packaging materials were included in the product market definition, DuPont’s share dropped to an acceptable 18 percent level. The importance of the definition of the relevant product market and the determination of the cross price elasticity of demand among close substitute products has often been emphasized by the courts.11

cross price elasticity The ratio of the percentage change in the quantity demanded of Good A to the percentage change in the price of Good B, assuming that all other factors influencing demand remain unchanged.

11See, for example, U.S. v. Alcoa, 148 F.2d 416, 424; Times Picayune Publishing Co. v. U.S. 345 U.S. 594; Con- tinental Can Co. v. U.S., 378 U.S. 441, 489. See John E. Kwoka and Lawrence J. White, The Antitrust Revolu- tion: Economics, Competition, and Policy (New York: Oxford University Press, 1999) for a further discussion of some of the economic issues involved in antitrust laws.

Chapter 3: Demand Analysis 87

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Why Pay More for Fax Paper and Staples at Staples?12

Just what constitutes an available substitute has as much to do with the cross price elasticity of rival firm supply as it does with the cross price elasticity of demand. In a hotly contested recent merger proposal, the Federal Trade Commission (FTC) has argued that office superstores like OfficeMax, Office Depot, and Staples are a separate relevant market from other smaller office supply retailers. Office Depot had 46 percent of the $13.22 billion in 1996 superstore sales of office supplies, Sta- ples had 30 percent, and OfficeMax had the remaining 24 percent. Office Depot and Staples proposed to merge, thereby creating a combined firm with 76 percent of the market. Such mergers have been disallowed many times under the Sherman Antitrust Act’s prohibition of monopolization.

The two companies insisted, however, that their competitors included not only OfficeMax but all office supply distribution channels, including small paper goods specialty stores, department stores, discount stores like Target, warehouse clubs like Sam’s Club, office supply catalogs, and some computer retailers. This larger office supply industry is very fragmented, easy to enter (or exit), and huge—1996 sales topped $185 billion. By this latter standard, the proposed merger involved admit- tedly the largest players in the industry, but companies with only 3–5 percent mar- ket shares. Under this alternative interpretation of the relevant market, Office Depot and Staples should have been allowed to proceed with their merger.

Have superstores like Home Depot and Lowes in do-it-yourself building sup- plies, PetSmart in pet supplies, and Office Depot, OfficeMax, and Staples created a new time-saving customer shopping experience and demand pattern in towns where they are clustered? Office supply products are search goods for which custo- mers can detect quality prior to purchase and locate just the quality-price combi- nation they desire. Brand name reputations should therefore have little effect on repeat purchase shopping patterns at Office Depot and Staples. Is this case devoid of a rationale for antitrust action? Have successful entrepreneurs simply created a new segment within the traditional relevant market for office products?

The FTC undertook two sets of experiments to advise the commissioners who voted to deny the proposed merger. Prices for everything from paper clips to fax paper were sampled in 40 cities and towns where Office Depot and Staples com- peted and in other similar locations where only one of the superstores was present. The prices were significantly higher in the single superstore markets. Apparently, despite an enormous rival supply of traditional office product retailers, target cus- tomers (like secretaries responsible for securing resupply) are willing to pay more for staples at Staples.

As Walmart has demonstrated in other search good categories, shoppers will flock to a superstore despite numerous small retailers closer to the customer. So, despite the enormous preexisting supply of traditional rivals and the exceptional ease of entry (and exit) at small scale, competition for superstore retailers comes only from other superstore retailers. As a result, the Sherman Act warrants denying the proposed merger in office supply superstores. Although stores like Walmart are entitled to become “category killers” on their own sales growth, the FTC has de- cided to bar superstore mergers as a route to obtaining near-monopoly status.

12Based on “FTC Votes to Bar Staples’ Bid for Rival,” Wall Street Journal (March 11, 1997), p. A3.

88 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

An Empirical Illustration of Price, Income, and Cross Elasticities A study by Chapman, Tyrrell, and Mount examined the elasticity of energy use by resi- dential, commercial, and industrial users.13 They hypothesized that the demand for elec- tricity was determined by the price of electricity, income levels, and the price of a substitute good—natural gas.

Table 3.6 summarizes the electricity-use elasticities with respect to price, income, and the substitute product (here, natural gas) prices. As shown in the table, the price elastic- ity of demand for electricity was relatively elastic in all markets, with the highest price elasticity being in the industrial market. This is consistent with the observation that many assembly plants, foundries, and other heavy industrial users switch to self- generated power with natural gas-fired turbines when electricity prices spike. The posi- tive cross elasticity shows that electricity and natural gas are, indeed, substitute goods.

THE COMBINED EFFECT OF DEMAND ELASTICITIES When two or more of the factors that affect demand change simultaneously, one is often interested in determining their combined impact on quantity demanded. For example, suppose that a firm plans to increase the price of its product next period and anticipates that consumers’ disposable incomes will also increase next period. Other factors affecting demand, such as advertising expenditures and competitors’ prices, are expected to re- main the same in the next period. From the formula for the price elasticity (Equation 3.1), the effect on quantity demanded of a price increase would be equal to

%ΔQD = ED(%ΔP)

Similarly, from the formula for the income elasticity (Equation 3.11), the effect on quantity demanded of an increase in consumers’ incomes would be equal to

%ΔQD = Ey(%ΔY)

Each of these percentage changes (divided by 100 to put them in a decimal form) would be multiplied by current period demand (Q1) to get the respective changes in quantity demanded caused by the price and income increases. Assuming that the price and in- come effects are independent and additive, the quantity demanded next period (Q2) would be equal to current period demand (Q1) plus the changes caused by the price and income increases:

Q2 = Q1 + Q1 [ED(%ΔP)] + Q1[Ey(%ΔY)]

TABLE 3.6 ELECTRICITY-USE ELASTICITIES

PRICE ELASTICITY

INCOME ELASTICITY

CROSS ELASTICITY (GAS)

Residential market −1.3 0.3 0.15

Commercial market −1.5 0.9 0.15

Industrial market −1.7 1.1 0.15

13D. Chapman, T. Tyrrell, and T. Mount, “Electricity Demand Growth and the Energy Crisis,” Science (November 17, 1972), p. 705.

Chapter 3: Demand Analysis 89

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

or

Q2 = Q1 [1 + ED(%ΔP) + Ey(%ΔY)] [3.14]

The combined use of income and price elasticities, illustrated here for forecasting de- mand, can be generalized to include any of the elasticity concepts that were developed in the preceding sections of this chapter.

SUMMARY

� Demand relationships can be represented in the form of a schedule (table), graph, or algebraic function.

� The demand curve is downward sloping, indicating that consumers are willing to purchase more units of a good or service at lower prices (iron law of demand).

� The total effect of a price reduction is the sum of an ever present positive substitution effect and a sometimes positive, sometimes negative, and possi- bly zero purchasing power (or real income) effect.

� The magnitude of the substitution effect depends upon switching costs—that is, the perceived close- ness of substitutes. The magnitude of the purchas- ing power effect depends upon positioning of the product and the targeting of particular customers who are likely to find the product offering an aspi- rant good rather than an inferior good.

� Elasticity refers to the responsiveness of quantity demanded (or supplied) to changes in price or an- other related variable. Thus price elasticity of de- mand refers to the percentage change in quantity demanded associated with a percentage change in price, holding constant the effects of other deter- minants of demand. Demand is said to be relatively price elastic (inelastic) if a given percentage change in price results in a greater (smaller) percentage change in quantity demanded.

� When demand is unit elastic, marginal revenue equals zero and total revenue is maximized. When demand is elastic, an increase (decrease) in price will result in a decrease (increase) in total revenue. When demand is inelastic, an increase (decrease) in price will result in an increase (de- crease) in total revenue.

Example Price and Income Effects: The Seiko Company Suppose Seiko is planning to increase the price of its watches by 10 percent in the coming year. Economic forecasters expect real disposable personal income to in- crease by 6 percent during the same period. From past experience, the price elas- ticity of demand has been estimated to be approximately −1.3 and the income elasticity has been estimated at 2.0. These elasticities are assumed to remain con- stant over the range of price and income changes anticipated. Seiko currently sells 2 million watches per year. Determine the forecasted demand for next year (assum- ing that the percentage price and income effects are independent and additive). Substituting the relevant data into Equation 3.14 yields

Q2 = 2,000,000 ½1 + ð−1:3Þð:10Þ + ð2:0Þð:06Þ� = 1,980,000 units

The forecasted demand for next year is 1.98 million watches assuming that other factors that influence demand, such as advertising and competitors’ prices, remain unchanged. In this case, the positive impact of the projected increase in household income is more than offset by the decline in quantity demanded associated with a price increase.

90 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

� Prices should always be increased in the inelastic region of the firm’s demand because lower price points would result in reduced revenue, despite in- creased unit sales.

� Income elasticity of demand refers to the percent- age change in quantity demanded associated with a percentage change in income, holding constant the effects of determinants of demand.

� Cross elasticity of demand refers to the percent- age change in quantity demanded of Good A

associated with a percentage change in the price of Good B.

� The magnitude of price elasticity varies a great deal across target customers and across products be- cause of differences in (1) the number and per- ceived closeness of substitutes (also known as switching costs), (2) the percentage of the budget expended on the product, (3) the positioning of the product as income superior to that target market, and (4) the time period of adjustment.

Exercises 1. The Potomac Range Corporation manufactures a line of microwave ovens costing $500 each. Its sales have averaged about 6,000 units per month during the past year. In August, Potomac’s closest competitor, Spring City Stove Works, cut its price for a closely competitive model from $600 to $450. Potomac noticed that its sales vol- ume declined to 4,500 units per month after Spring City announced its price cut. a. What is the arc cross elasticity of demand between Potomac’s oven and the

competitive Spring City model? b. Would you say that these two firms are very close competitors? What other

factors could have influenced the observed relationship? c. If Potomac knows that the arc price elasticity of demand for its ovens is −3.0,

what price would Potomac have to charge to sell the same number of units it did before the Spring City price cut?

2. The price elasticity of demand for personal computers is estimated to be −2.2. If the price of personal computers declines by 20 percent, what will be the expected percentage increase in the quantity of computers sold?

3. The Olde Yogurt Factory has reduced the price of its popular Mmmm Sundae from $2.25 to $1.75. As a result, the firm’s daily sales of these sundaes have in- creased from 1,500/day to 1,800/day. Compute the arc price elasticity of demand over this price and consumption quantity range.

4. The subway fare in your town has just been increased from a current level of 50 cents to $1.00 per ride. As a result, the transit authority notes a decline in rider- ship of 30 percent. a. Compute the price elasticity of demand for subway rides. b. If the transit authority reduces the fare back to 50 cents, what impact would

you expect on the ridership? Why? 5. If the marginal revenue from a product is $15 and the price elasticity of demand

is −1.2, what is the price of the product? 6. The demand function for bicycles in Holland has been estimated to be

Q = 2,000 + 15Y − 5.5P

where Y is income in thousands of euros, Q is the quantity demanded in units, and P is the price per unit. When P = 150 euros and Y = 15(000) euros, deter- mine the following:

a. Price elasticity of demand b. Income elasticity of demand

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

Chapter 3: Demand Analysis 91

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

7. In an attempt to increase revenues and profits, a firm is considering a 4 percent increase in price and an 11 percent increase in advertising. If the price elasticity of demand is −1.5 and the advertising elasticity of demand is +0.6, would you expect an increase or decrease in total revenues?

8. The Stopdecay Company sells an electric toothbrush for $25. Its sales have aver- aged 8,000 units per month over the past year. Recently, its closest competitor, Decayfighter, reduced the price of its electric toothbrush from $35 to $30. As a result, Stopdecay’s sales declined by 1,500 units per month. a. What is the arc cross elasticity of demand between Stopdecay’s toothbrush

and Decayfighter’s toothbrush? What does this indicate about the relation- ship between the two products?

b. If Stopdecay knows that the arc price elasticity of demand for its toothbrush is −1.5, what price would Stopdecay have to charge to sell the same number of units as it did before the Decayfighter price cut? Assume that Decayfighter holds the price of its toothbrush constant at $30.

c. What is Stopdecay’s average monthly total revenue from the sale of electric toothbrushes before and after the price change determined in part (b)?

d. Is the result in part (c) necessarily desirable? What other factors would have to be taken into consideration?

9. The Sydney Transportation Company operates an urban bus system in New South Wales, Australia. Economic analysis performed by the firm indicates that two major factors influence the demand for its services: fare levels and downtown parking rates. Table 1 presents information available from 2005 operations. Fore- casts of future fares and hourly parking rates are presented in Table 2.

TABLE 1

AVERAGE DAILY TRANSIT RIDERS (2005)

AVERAGE DOWN- TOWN ROUND-TRIP

FARE PARKING RATE

5,000 $1.00 $1.50

TABLE 2

YEAR ROUND-TRIP FARE AVERAGE PARKING

RATES

2006 $1.00 $2.50

2007 $1.25 $2.50

Sydney’s economists supplied the following information so that the firm can esti- mate ridership. Based on past experience, the coefficient of cross elasticity be- tween bus ridership and downtown parking rates is estimated at 0.2, given a fare of $1.00 per round trip. This is not expected to change for a fare increase to $1.25. The price elasticity of demand is currently estimated at −1.1, given hourly parking rates of $1.50. It is estimated, however, that the price elasticity will change to −1.2 when parking rates increase to $2.50. Using these data, estimate the average daily ridership for 2006 and 2007.

10. The Reliable Aircraft Company manufactures small, pleasure-use aircraft. Based on past experience, sales volume appears to be affected by changes in the price

92 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

of the planes and by the state of the economy as measured by consumers’ dispos- able personal income. The following data pertaining to Reliable’s aircraft sales, selling prices, and consumers’ personal income were collected:

YEAR AIRCRAFT

SALES AVERAGE PRICE

DISPOSABLE PERSONAL INCOME (IN CONSTANT 2006 DOLLARS—

BILLIONS)

2006 525 $17,200 $610

2007 450 8,000 610

2008 400 8,000 590

a. Estimate the arc price elasticity of demand using the 2006 and 2007 data. b. Estimate the arc income elasticity of demand using the 2006 and 2007 data. c. Assume that these estimates are expected to remain stable during 2008.

Forecast 2008 sales for Reliable assuming that its aircraft prices remain con- stant at 2007 levels and that disposable personal income will increase by $40 billion. Also assume that arc income elasticity computed in (b) above is the best available estimate of income elasticity.

d. Forecast 2008 sales for Reliable given that its aircraft prices will increase by $500 from 2007 levels and that disposable personal income will increase by $40 billion. Assume that the price and income effects are independent and additive and that the arc income and price elasticities computed in parts (a) and (b) are the best available estimates of these elasticities to be used in mak- ing the forecast.

11. Federal excise taxes on gasoline vary widely across the developed world. The United States has the lowest taxes at U.S. $0.40 per gallon (or £0.07 per liter), Ca- nada has taxes of $0.60 per gallon, Japan and much of Europe is $2.00 per gallon, while Britain has the highest tax at $2.83 a gallon or £0.5 per liter. If gasoline taxes are intended to reduce the time losses from road congestion in urban envir- onments and gasoline pre-tax costs about £0.40 per liter, why might the optimal tax in Canada be 50 percent higher than in the United States? What would be an explanation for why adjacent countries would have such different estimates of the price elasticity of demand for auto driving?14

12. What conceptual determinant of auto demand price elasticity is most closely as- sociated with the differences in switching cost across the target customers in the previous case study—low switching cost for the German couple who commute in an old Taurus and high switching cost for the older couple who drive Crown Vics?

13. Illustrate the relationship between product positioning and customer targeting using the facts of the Ford dealership direct mail coupon example. Which cus- tomer is least likely and second least likely to buy a Taurus for this reason?

Case Exercise POLO GOLF SHIRT PRICING

The setting is a Ralph Lauren outlet store, and the product line is Polo golf shirts. A product manager and the General Manager for Outlet Sales are analyzing the discount to be offered at the outlet stores. Let’s work through the decision at the level of one

14Based on “Fueling Discontent,” The Economist (May 19, 2001), p. 69.

Chapter 3: Demand Analysis 93

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

color of golf shirts sold per outlet store per day. The decision being made is how low a price to select at the start of any given day to generate sales at that price throughout the day. The demand, revenue, and variable cost information is collected on the fol- lowing spreadsheet:

QUANTITY SOLD

UNIFORM PRICE

TOTAL REVENUE

MARGINAL REVENUE

VARIABLE COST

0 $50.00 $0 $0 $28

1 $48.00 $48 $48 $28

2 $46.00 $92 $44 $28

3 $45.00 $135 $43 $28

4 $44.00 $176 $41 $28

5 $42.00 $210 $34 $28

6 $40.00 $240 $30 $28

7 $38.31 $268 $28 $28

8 $36.50 $292 $24 $28

9 $34.50 $311 $19 $28

10 __________ __________ $16 $28

11 __________ __________ $13 $28

12 __________ __________ $10 $28

13 __________ __________ $7 $28

14 __________ __________ $4 $28

15 __________ __________ $0 $28

16 __________ __________ ($1) $28

17 __________ __________ ($4) $28

18 __________ __________ ($7)

Questions 1. Identify the change in total revenue (the marginal revenue) from the fourth shirt

per day. What price reduction was necessary to sell four rather than three shirts? 2. What is the change in total revenue from lowering the price to sell seven rather

than six shirts in each color each day? 3. Break out the components of the $28 marginal revenue from the seventh unit sale

at $38.31—that is, how much revenue is lost per unit sale relative to the price that would “move” six shirts per color per day?

4. Calculate the total revenue for selling 10–16 shirts per day. Calculate the reduced prices necessary to achieve each of these sales rates.

5. What number of shirt unit sales most pleases a sales clerk with sales- commission-based bonuses?

6. Would you recommend lowering price to the level required to generate 15 unit sales per day? Why or why not?

7. What is the operating profit or loss on the fifteenth shirt sold per color per day? What about the twelfth? The tenth?

8. How many shirts do you recommend selling per color per day? What then is your recommended dollar markup and markup percentage? What dollar margin and percentage margin is that?

94 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

4 CHAP T E R

Estimating Demand CHAPTER PREVIEW The preceding chapter developed the theory of demand, including the concepts of price elasticity, income elasticity, and cross price elasticity of demand. A manager who is contemplating an increase in the price of one of the firm’s products needs to know the quantitative magnitude of the impact of this increase on quantity demanded, total revenue, and profits. Is the demand elastic, inelastic, or unit elastic with respect to price over the range of the contemplated price increase? What growth of unit sales can be expected if consumer incomes increase as a result of a recovery from a severe recession?

Governments and not-for-profit institutions are also faced with similar questions. What will be the impact of an increase in mass transit fares or bridge tolls? Will automobile commuting decrease by 5, 10, or 20 percent? Will a sales tax increase boost revenue enough to cover a projected budget shortfall? This chapter discusses some of the techniques and problems associated with estimating demand.

MANAGERIAL CHALLENGE Global Warming and the Demand for Public Transportation1

There is now little scientific debate that fossil-fuel burn- ing and deforestation have resulted in a net release into Earth’s atmosphere of more CO2 in the past two de- cades than at any previous time in human history. The parts per million (ppm) concentrations of CO2 today (385 ppm) are 10 percent higher than the 350 ppm 20 years ago in 1990, which were themselves 10 percent higher than the 320 ppm 30 years earlier in 1960, which were 8 percent higher than the 295 ppm 60 years earlier in 1900. That is, the pace of CO2 accumulation has been accelerating rapidly. And it has been acknowledged for over a century that coal burning increases the green- house effect and that with 90 percent certainty (since a well-balanced study was issued by the Intergovernmen- tal Panel on Climate Change in 2007) the greenhouse effect of greater concentrations of CO2 warms the planet by trapping more solar radiation. The consequence is a melting of ice sheets at both poles, raising ocean levels and seawater temperatures, thereby altering weather patterns to no longer align with crop cycles, and sharply worsening cyclones, hurricanes, and coastal flooding.

The climate change crisis is reaching a tipping point that necessitates our immediately raising prices to reflect these social costs of carbon-based energy from coal, gas- oline, natural gas, and so forth. The European Union has already done so with carbon trading permits like the sulfur dioxide and nitrous oxide pollution permits established in 1990 by the U.S. Clean Air Act. China and the United States are the largest CO2 emitters,

© Ke ith

Br of sk y/ Ph ot od is c Gr ee n/ Ge tty

Im ag es

95

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

and Australia is 10th largest on a per capita basis. The leading proposals to introduce price effects that discour- age the use of carbon-based fuels in the United States and in Australia involve: (1) an emissions trading scheme, and (2) a carbon tax directly linked to the carbon footprint of each commercial product available for sale.

Both alternatives will raise the cost of living. Austra- lia estimates that the cost for electricity consumed by a

two-worker family of four will rise by $416 per year, for gasoline by $166 per year, and for fertilizers and truck- ing associated with food production and delivery by $136 per year. One plan under consideration is to return the $706 rise in the cost of living to families as a tax rebate. How the higher cost of driving to work and the lump sum rebate checks for $706 will affect demand for a variety of products, including mass transit, is the subject of this chapter.

$1.25

$1.50

$2.00

$1.75

$1.00

$.75

$.50

$.25

1975 1980 1985 1990 1995 2000 2005 2010

20101975 1980 1985 1990 1995 2000 2005

50¢

75¢

$1.00 $1.10

$1.25

$1.60

$1.75

$2.00 Cost of riding

Energy crisis drove up ridership before runaway inflation and a slowing economy forced PAT to raise fares again.

After eight years of no fare increase, inflation and other factors caught up. Fares were raised.

Higher fuel costs require increased operating revenue.

100

80

Ridership (in millions)

National energy crisis, gas rationing sent ridership to record high of 109.6 million in 1975.

90

60

70

Population declines, high unemployment, cheaper gas, low parking rates combined for record low ridership of 88.3 million in 1983.

72 million 66 million 70 millionN

um be

r of

t ri

ps

Exodus to the suburbs and HOV lanes caused ridership to collapse.

110

88 million

Cont.

MANAGERIAL CHALLENGE Continued

96 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Port Authority Transit (PAT) provides public trans- portation services to the residents of Allegheny County (Pittsburgh and suburbs). It operates a fleet of 1,027 buses and 55 light rail vehicles in providing 70 million transit trips per year. In 2009, PAT adopted a $250 mil- lion operating budget that Pennsylvania state law requires not to exceed fare revenues plus federal, state, and county subsidies. PAT’s cash (base) fare was in- creased from $1.60 to $1.75 in 2004, and further increased to $2.00 in 2008.

In analyzing the effects of any additional fare increase, a number of issues need to be addressed. Most importantly,

� How would the fare (price) increase affect de- mand and overall revenues?

� What other factors, besides fares, could affect demand?

Examination of the data in the preceding graphs gives some possible answers to these questions. Note that ridership declined whenever PAT raised fares (i.e., in 1971, 1976, 1980, 1982, 1992, and 2002) except that ridership increased during the mid- to late 1970s when there were gasoline shortages and in 2008 in response to much higher gas prices at the pump. Finally, note that ridership declined steeply once in the early 1980s— when there was higher unemployment and population declined in Pittsburgh itself—and a second time in the early 1990s when high occupancy vehicle (HOV) lanes reserved for carpools opened on several interstates lead- ing into Pittsburgh.

Econometric models can be used to estimate the de- terminants of mass transit demand and the associated capacity requirements. In the summer of 2008, with gas- oline at $4.01 per gallon, PAC saw a 15+ percent in- crease in mass transit ridership. If the 5 percent reduction in greenhouse gas emissions pledged by the United States at the Copenhagen Summit on Climate Change in December 2009 is to be met, a substantial number of Americans will need to stop driving to work and instead carpool or ride the bus. This chapter focuses on the techniques that are used in developing such demand predictions.

Discussion Questions

� Do you think a carbon tax or emissions trad- ing scheme such as the United States used to abate air pollution (acid rain) should be used to reduce CO2 emissions?

� If the only question is how far and how fast the cost of coal-fired electricity and hydrocarbon- based fuels like natural gas and gasoline will rise, what do you predict will be the impact on the demand for mass transit?

� Brainstorm about other greener transportation alternatives.

1Based on “The Road to Copenhagen” and “Global Pressure for Local Climate Scheme,” Sydney Times Herald (December 5, 2009), p. 1.

1900

295

320

350

385

1960 1990 2010

C O

2 (p

ar ts

p er

m ill

io n)

Largest CO2 emitters from

energy consumption, 2007

China 6.28 billion tons United States 6.01 Russia 1.67 India 1.40 Japan 1.26 Germany .84 Canada .59 United Kingdom .58 Korea .52

Source: U.S. Energy Administration, Carbon Dioxide Gas Emissions and Atmospheric Concentration 1900–2010.

MANAGERIAL CHALLENGE Continued

Chapter 4: Estimating Demand 97

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

ESTIMATING DEMAND USING MARKETING RESEARCH TECHNIQUES Before examining some of the statistical techniques that are useful in estimating demand relationships, this section looks at three different marketing research methods that can be used in analyzing demand. These techniques are consumer surveys, focus groups, and market experiments.

Consumer Surveys Consumer surveys involve questioning a sample of consumers to determine such factors as their purchase intent, willingness to pay, sensitivity to price changes, and awareness of advertising campaigns. Consumer surveys can provide a great deal of useful information to a firm. Consumer expectations about future business and credit conditions may pro- vide significant insights into the consumers’ propensity to purchase many items, espe- cially durable goods. Using a little imagination and asking less direct questions may also offer insights. If questioning reveals that consumers are unaware of price differences among several competing products, it might be concluded that at least within the current range of prices, demand may be price inelastic.

Consumer Focus Groups Another means of recording consumer responses to changes in factors affecting demand is through the use of panel data on consumer focus groups. In these situations, for ex- ample, experimental groups of consumers are given a small amount of money with which to buy certain items. The experimenter can observe the impact on actual pur- chases as price, prices of competing goods, and other variables are recorded. Then the group of consumers is closely observed discussing the choices they made and why. Of course the costs of setting up and running such a clinic are substantial, and the partici- pants may suspect that the experimenter is interested in sensitivity to prices and may respond more than otherwise would be the case.

Example Estimating Cross Elasticity: Simmons Mattress Company The Simmons Mattress Company conducted an experiment involving the relative prices of its mattresses. Two identical types of mattresses, some with the Simmons label and others with a fictitious brand name such as Sleepwell, were offered for sale at the same prices and varying price spreads to determine cross price elasticity. It was found that with identical prices, Simmons outsold the unknown brand 15 to 1; with a $5 premium over the unknown brand, Simmons’s sales were 8 times greater; and with a 25 percent premium, sales were about the same.

In controlled test markets, one must take great care to assure that sales effects are not due to unusually bad weather, competitive advertising or competitive price re- ductions, and even local strikes or large layoffs that change consumer incomes signif- icantly. The duration of market experiments is usually quite short, and the magnitude of price or advertising changes in most experiments is usually small. In spite of these limitations, direct market experimentation may prove quite useful.

focus groups A market research technique employing close observation of discussion among target consumers.

98 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Market Experiments in Test Stores Another approach that is sometimes used to garner information about the demand function is the market experiment, which examines the way consumers behave in con- trolled purchase environments. A test store may vary one or more of the determinants of demand, such as price or advertising or the presence of an NFL or NBA logo on a sweatshirt, and observe the impact on quantity demanded. This approach may be espe- cially useful in developing measures of the promotion elasticity of demand for a product.

STATISTICAL ESTIMATION OF THE DEMAND FUNCTION Effective decision making eventually requires the quantitative measurement of economic relationships. Econometrics is a collection of statistical techniques available for estimating such relationships. The principal econometric techniques used in measuring demand relationships are regression and correlation analysis. The simple (two-variable) linear regression model and the more complex cases of multiple linear regression models and nonlinear models (discussed in Appendix 4A) are presented next.

Specification of the Model The next step is to specify the form of the equation, or regression relation, that indicates the relationship between the independent variables and the dependent variable(s). Nor- mally the specific functional form of the regression relation to be estimated is chosen to depict the true demand relationships as closely as possible. Graphing such relationships often will tell whether a linear equation is most appropriate or whether logarithmic, ex- ponential, or other transformations are more appropriate. See Appendix 4A for a discus- sion of these transformations.

Example Variable Identification and Data Collection: Sherwin-Williams Company Sherwin-Williams Company is attempting to develop a demand model for its line of exterior house paints. The company’s chief economist feels that the most impor- tant variables affecting paint sales (Y) (measured in gallons) are:

1. Promotional expenditures (A) (measured in dollars). These include expendi- tures on advertising (radio, TV, and newspapers), in-store displays and liter- ature, and customer rebate programs.

2. Selling price (P) (measured in dollars per gallon). 3. Disposable income per household (M) (measured in dollars).

The chief economist decides to collect data on the variables in a sample of 10 company sales regions that are roughly equal in population.2 Data on paint sales, promotional expenditures, and selling prices were obtained from the com- pany’s marketing department. Data on disposable income (per capita) were obtained from the Bureau of Labor Statistics. The data are shown in Table 4.1.

2A sample size of 10 observations was chosen to keep the arithmetic simple. Much larger samples are used in actual ap- plications. The desired accuracy and the cost of sampling must be weighed in determining the optimal sample size.

Chapter 4: Estimating Demand 99

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Linear Model A linear demand model for Sherwin-Williams paint would be speci- fied as follows:

Q = α + β1A + β2P + β3M + e [4.1]

where α, β1, β2, and β3 are the parameters of the model and e is the error term to reflect the fact that the observed demand value will seldom equal the exact value predicted by the model. The values of the parameters are estimated using the regression techniques described later in the chapter. Demand theory implies that price (P) would have a nega- tive effect on gallons of paint sold (Q) (i.e., as the price rises, quantity demanded de- clines, holding constant all other variables) and that promotional expenditures (A) and income (M) would have a positive effect on paint sales.

The parameter estimates may be interpreted in the following manner. If we rearrange Equation 4.1 to solve for price (P), the intercept of the resulting inverse demand function identifies the maximum price that can be charged. The value of each β coefficient pro- vides an estimate of the change in quantity demanded associated with a one-unit change in the given independent variable, holding constant all other independent variables. The β coefficients are equivalent to the partial derivatives of the demand function:

β1 = ∂Q ∂A

, β2 = ∂Q ∂P

, β3 = ∂Q ∂M

[4.2]

Recall that the elasticity of linear demand with respect to price is defined as

ED = ∂Q ∂P

· P Q

[4.3]

Now substitute Equation 4.2 into this expression to yield

ED = β2 · P Q

[4.4]

Equations 4.3 and 4.4 show that price elasticity of linear demand depends upon the tar- get market’s price sensitivity (∂Q/∂P), as well as the price point positioning of the prod- uct per unit sale (P/Q).

TABLE 4.1 SHERWIN-WILLIAMS COMPANY DATA

SALES REGION

SALES (Y ) (×1,000

GALLONS)

PROMOTIONAL EXPENDITURES (A ) (×$1,000)

SELLING PRICE (P) ($/GALLON)

DISPOSABLE INCOME (M) (×$1,000)

1 160 150 15.00 19.0

2 220 160 13.50 17.5

3 140 50 16.50 14.0

4 190 190 14.50 21.0

5 130 90 17.00 15.5

6 160 60 16.00 14.5

7 200 140 13.00 21.5

8 150 110 18.00 18.0

9 210 200 12.00 18.5

10 190 100 15.50 20.0

100 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Multiplicative Exponential Model Another commonly used demand relation- ship is the multiplicative exponential model. In the Sherwin-Williams example, such a model would be specified as follows:

Q = αAβ1Pβ2Mβ3 [4.5]

This model is also popular because of its ease of estimation. For instance, Equation 4.5 may be transformed into a simple linear relationship in logarithms (adding an error term) as follows:

log Q = log α + β1log A + β2log P + β3log M + e [4.6]

and the parameters log α, β1, β2, and β3 can be easily estimated by any regression pack- age. The intuitive appeal of this multiplicative exponential functional form is based on the fact that the marginal impact of a change in price on quantity demanded is depen- dent not only on the price change, but also on all the other determinants of demand too—that is, all the elements in the marketing mix and the target consumer’s household income, and so forth.

Demand functions in the multiplicative exponential form possess the convenient fea- ture that the elasticities are constant over the range of data used in estimating the pa- rameters and are equal to the estimated values of the respective parameters. In the Sherwin-Williams data, for example, the price elasticity of demand is defined as

ED = ∂Q ∂P

· P Q

[4.7]

Differentiating Equation 4.5 with respect to price results in

∂Q ∂P

= β2αA β1Pβ2 − 1Mβ3 [4.8]

So, using Equation 4.7,

ED = β2αA β1Pβ2 − 1Mβ3

P Q

� � [4.9]

Substituting Equation 4.5 for Q in Equation 4.9, and canceling and then combining terms, yields

ED = β2

That is, multiplicative exponential demand functions have constant price and other elas- ticities. This property contrasts sharply with the elasticity of a linear demand function that changes continuously over the entire price or income range of the demand curve. However, pricing analysts at Sherwin-Williams may be able to tell us that the percentage change in quantity demanded for either a 10 percent price increase or a 10 percent price cut is a con- stant 15 percent. And the same answer may apply at rather different price points when clearance sales occur. If so, a multiplicative exponential demand model is appropriate.

A SIMPLE LINEAR REGRESSION MODEL The analysis in this section is limited to the simplest case of one independent and one dependent variable, where the form of the relationship between the two variables is linear:

Y = α + βX + e [4.10]

X is used to represent the independent variable and Y the dependent variable.3

3Capitalized letters X and Y represent the name of the random variables. Lowercase x and y represent specific values of the random variables.

Chapter 4: Estimating Demand 101

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Assumptions Underlying the Simple Linear Regression Model

Assumption 1 The value of the dependent variable Y is postulated to be a random variable, which is dependent on deterministic (i.e., nonrandom) values of the indepen- dent variable X.5

Assumption 2 A theoretical straight-line relationship (see Figure 4.1) exists between X and the expected value of Y for each of the possible values of X. This theoretical regression line

E(Y|X) = α + βX [4.11]

has a slope of β and an intercept of α. The regression coefficients α and β constitute pop- ulation parameters whose values are unknown, and we desire to estimate them.

Assumption 3 Associated with each value of X is a probability distribution, p(y|x), of the possible values of the random variable Y. When X is set equal to some value xi, the value of Y that is observed will be drawn from the p(y|xi) probability distribution and will not necessarily lie on the theoretical regression line. As illustrated in Figure 4.2, some values of y|xi are more likely than others, and the mean E(y|xi) lies on the theoret- ical regression line. If ei is defined as the deviation of the observed yi value from its the- oretical value y0i, then

yi = y0i + εi yi = α + βxi + εi

[4.12]

Example Linear, Not Exponential, Sales at Global Crossing Inc.4

In 2002, telecom network providers like Global Crossing and WorldCom were wildly optimistic about the growth of telecom traffic, because of the projected growth of the Internet. Much like the adoption of color television, the penetration of the Internet into the American household has exhibited a classic S-shaped pat- tern of exponential growth fueled by the early adopters (1994–1996), followed now by a long period of much slower, approximately linear growth in demand. Pur- chasing and installing fiber optic cable networks as though the exponential demand growth were continuing unabated led to a quick saturation of a market that would have been better specified with a linear time trend,

Q = α + β1A + β2P + β3M + β4T + e

where T is time (2006 = 0, 2007 = 1, 2008 = 2, etc.). The effect of these demand projections on Global Crossing is examined in the Managerial Challenge at the be- ginning of Chapter 5.

4Based on “Adoption Rate of Internet by Consumers Is Slowing” and “Has Growth of the Net Flattened,” Wall Street Journal (July 16, 2001), pp. B1 and B8.

5Stochastic (i.e., random) values of the right-hand-side independent variable are addressed in Appendix 4A under simultaneous equations relationships.

102 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

or, in general, the linear regression relation becomes (as illustrated in Figure 4.3)

Y = α + βX + e [4.13]

where e is a zero mean stochastic disturbance (or error) term. The disturbance term (ei) is assumed to be an independent random variable [that is,

E(eiej) = 0 for i ≠ j] that is normally distributed with an expected value equal to zero [that is, E(ei) = 0] and with a constant variance equal to σ2ε [that is, Eðε2i Þ = σ2ε for all i].

Together, Assumptions 1 and 3 imply that the N(0, σ2ε ) disturbance term is expected to be uncorrelated with the independent variables in the regression model.

Estimating the Population Regression Coefficients Once the regression model is specified, the unknown values of the population regression coefficients α and β are estimated by using the n pairs of sample observations (x1, y1), (x2, y2),… , (xn, yn). This process involves finding a sample regression line that best fits the sample of observations the analyst has gathered.

FIGURE 4.2 Conditional Probability Distribution of Dependent Variable

Observed values of

Y

X

x1 x2 xn

α + βX

yi

0

p(y xi )

Theoretical r egression

line

FIGURE 4.1 Theoretical Regression Line

y

y� = E(yi|xi)

xi

x

Theoretical regression line E(Y|X) = � + �X

i

Chapter 4: Estimating Demand 103

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The sample estimates of α and β can be designated by a and b, respectively. The esti- mated or predicted value of Y, ŷi, for a given value of X (see Figure 4.4) is

ŷi = a + bxi [4.14]

Letting ei be the deviation of the observed yi value from the estimated value ŷi, then

yi = ŷi + ei

= a + bxi + ei [4.15]

or, in general, the sample regression equation becomes

Y = a + bX + e [4.16]

Although there are several methods for determining the values of a and b (that is, finding the regression equation that provides the best fit to the series of observations),

FIGURE 4.4 Deviation of the Observations about the Sample Regression Line

yi

Y

ei

xi X

Sample regression line

Y = a + bXi

FIGURE 4.3 Deviation of the Actual Observations about the Theoretical Regression Line

yi

Y

εi

xi X

Theoretical regression line

= α βxi + εi +

E(Y X) = α βx +

104 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the best known and most widely used is the method of least squares. The objective of least-squares analysis is to find values of a and b that minimize the sum of the squares of the ei deviations. (By squaring the errors, positive and negative errors cumulate with- out canceling each other. From Equation 4.15, the value of ei is given by

ei = yi − a − bxi [4.17]

Squaring this term and summing over all n pairs of sample observations, one obtains

∑ n

i = 1 e2i = ∑

n

i = 1 ðyi − a − bxiÞ2 [4.18]

Using calculus, the values of a and b that minimize this sum of squared deviations expression are given by

b = n∑xiyi − ∑xi∑yi n∑x2i − ð∑xiÞ2

[4.19]

a = y − bx [4.20]

where x and y are the arithmetic means of X and Y, respectively (that is, x = ∑x=n and y = ∑y=n) and where the summations range over all the observations (i = 1, 2,… , n).

Example Estimating Regression Parameters: Sherwin-Williams Company (continued) Returning to the Sherwin-Williams Company example, suppose that only pro- motional expenditures are used to predict paint sales. The regression model can be calculated from the sample data presented earlier in Table 4.1. These data are reproduced here in columns 1–3 of Table 4.2 and shown graphically in Figure 4.5.

The estimated slope of the regression line is calculated as follows using Equation 4.19:

b = 10ð229,100Þ − ð1,250Þð1,750Þ

10ð180,100Þ − ð1,250Þ2

= 0:433962

Similarly, using Equation 4.20, the intercept is estimated as

a = 175 − 0:433962ð125Þ = 120:75475

Therefore, the equation for estimating paint sales (in thousands of gallons) based on promotional expenditures (in thousands of dollars) is

Y = 120.755 + 0.434X [4.21]

and is graphed in Figure 4.5. The coefficient of X (0.434) indicates that for a one- unit increase in X ($1,000 in additional promotional expenditures), expected sales (Y) will increase by 0.434 (× 1,000) = 434 gallons in a given sales region.

Chapter 4: Estimating Demand 105

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

USING THE REGRESSION EQUATION TO MAKE PREDICTIONS A regression equation can be used to make predictions concerning the value of Y, given any particular value of X. This is done by substituting the particular value of X, namely, xp, into the sample regression equation (Equation 4.14):

TABLE 4.2 WORKSHEET FOR ESTIMATION OF THE SIMPLE REGRESSION

EQUATION: SHERWIN-WILLIAMS COMPANY

SALES REGION

PROMOTIONAL EXPENDITURES

(× $1,000) SALES

(× 1,000 GAL)

(1) (2) (3) (4) (5) (6)

i xi yi xiyi xi 2 yi

2

1 150 160 24,000 22,500 25,600

2 160 220 35,200 25,600 48,400

3 50 140 7,000 2,500 19,600

4 190 190 36,100 36,100 36,100

5 90 130 11,700 8,100 16,900

6 60 160 9,600 3,600 25,600

7 140 200 28,000 19,600 40,000

8 110 150 16,500 12,100 22,500

9 200 210 42,000 40,000 44,100

10 100 190 19,000 10,000 36,100

Total 1,250 1,750 229,100 180,100 314,900

∑xi ∑yi ∑xiyi ∑xi 2 ∑yi

2

x = Σxi=n = 1; 250=10 = 125

y = Σyi=n = 1; 750=10 = 175

FIGURE 4.5 Estimated Regression Line: Sherwin-Williams Company

0

50

100

150

200

250

Y = 120.755 + 0.434X

50 100 150 200 250

X (Promotional expenditures) ($1,000)

Y (

Pa in

t sa

le s)

( 1,

00 0

ga llo

ns )

106 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

ŷ = a + bxp

where, as you recall, ŷ is the hypothesized expected value for the dependent variable from the probability distribution p(Y|X).6

Suppose one is interested in predicting Sherwin-Williams’s paint sales for a metropol- itan area with promotional expenditures equal to $185,000 (i.e., xp = 185). Substituting xp = 185 into the estimated regression equation (Equation 4.21) yields

ŷ = 120:755 + 0:434ð185Þ = 201:045

or 201,045 gallons. Caution must be exercised in using regression models for prediction when the value of the

independent variable lies outside the range of observations from which the model was esti- mated. For example, here we cannot be certain that the prediction of paint sales based on the linear regression model would be reasonable for promotional expenditures of $300,000 since $200,000 was our largest sample value. Such factors as diminishing returns and the exis- tence of saturation levels can cause relationships between economic variables to be nonlinear.

A measure of the accuracy of estimation with the regression equation can be obtained by calculating the standard deviation of the errors of prediction (also known as the standard error of the estimate). The error term ei was defined earlier in Equation 4.17 to be the difference between the observed and predicted values of the dependent variable. The standard deviation of the ei term is based on the summed squared error (SSE) ∑e

2 i

normalized by the number of observations minus two:

se =

ffiffiffiffiffiffiffiffiffiffiffi ∑e2i n − 2

s =

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∑ðyi − a − bxiÞ2

n − 2

s

or, when this expression is simplified,7

se =

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∑y2i − a∑yi − b∑xiyi

n − 2

s [4.22]

If the observations are tightly clustered about the regression line, the value of se will be small and prediction errors will tend to be small. Conversely, if the deviations ei be- tween the observed and predicted values of Y are fairly large, both se and the prediction errors will be large.

In the Sherwin-Williams Company example, substituting the relevant data from Table 4.2 into Equation 4.22 yields

se =

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 314,900 − 120:75475ð1,750Þ − 0:433962ð229,100Þ

10 − 2

r

= 22:799

or a standard error of 22,799 gallons.

6The expected value of the error term (e) is zero, as indicated earlier in Assumption 3.

standard error of the estimate The standard deviation of the error term in a linear regression model.

7This formula applies to the case of simple regression in Equation 4.16. As additional variables are added to the linear regression model, the degrees of freedom in the denominator of Equation 4.22 becomes smaller and smaller: n − 3, n − 4, n − 5, etc.

Chapter 4: Estimating Demand 107

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The standard error of the estimate (se) can be used to construct prediction intervals for y.8 An approximate 95 percent prediction interval is equal to9

ŷ ± 2se [4.23]

Returning to the Sherwin-Williams Company example, suppose we want to construct an approximate 95 percent prediction interval for paint sales in a sales region with pro- motional expenditures equal to $185,000 (i.e., xp = 185). Substituting ŷ = 201.045 and se = 22.799 into Equation 4.23 yields

201.045 ± 2(22.799)

or a prediction interval from 155.447 to 246.643 (that is, from 155,447 gallons to 246,643 gallons) for promotions over the range $50,000 to $200,000.

Inferences about the Population Regression Coefficients For repeated samples of size n, the sample estimates of α and β—that is, a and b—will tend to vary from sample to sample. In addition to prediction, often one of the purposes of regression analysis is testing whether the slope parameter β is equal to some particular value β0. One standard hypothesis is to test whether β is equal to zero.

10 In such a test the concern is with determining whether X has a significant effect on Y. If β is either zero or close to zero, then the independent variable X will be of no practical benefit in predicting or explaining the value of the dependent variable Y. When β = 0, a one-unit change in X causes Y to change by zero units, and hence X has no effect on Y.

To test hypotheses about the value of β, the sampling distribution of the statistic b must be known.11 It can be shown that b has a t-distribution with n − 2 degrees of freedom.12,13 The mean of this distribution is equal to the true underlying regression co- efficient β, and an estimate of the standard deviation can be calculated as

8An exact 5 percent prediction interval is a function of both the sample size (n) and how close xp is to x and is given by the following expression:

ŷ ± tk=2;n−sse

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 +

1 n

+ ðxp − xÞ2 Σðxi − xÞ2

s

where tk/2,n−2 is the value from the t-distribution (with n − 2 degrees of freedom) in Table 2 of the Statistical Tables (Appendix B) in the back of the book. 9For large n (n > 30), the t-distribution approximates a normal distribution, and the t-value for a 95 percent pre- diction interval approaches 1.96 or approximately 2. For most applications, the approximation methods give sat- isfactory results. 10The intercept parameter, α, is of less interest in most economic studies and will be excluded from further analysis. 11In addition to testing hypotheses about β, one can also calculate confidence intervals for β in a manner sim- ilar to Equation 4.23 using the sampling distribution of β. 12A t-test is usually used to test for the significance of individual regression parameters when the sample size is relatively small (30 or less). For larger samples, tests of statistical significance may be made using the stan- dard normal probability distribution, which the t-distribution approaches in the limit. 13Degrees of freedom are the number of observations beyond the minimum necessary to calculate a given regression coefficient or statistic. In a regression model, the number of degrees of freedom is equal to the number of observations less the number of parameters (α’s and β’s) being estimated. For example, in a simple (two-variable) regression model, a minimum of two observations is needed to calculate the slope (β) and inter- cept (α) parameters—hence the number of degrees of freedom is equal to the number of observations n – 2. Three parameters to be estimated necessitate n – 3, and so on.

108 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

sb =

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi s2e

∑x2i − ð∑xiÞ2=n

s [4.24]

where se is the standard deviation of the error terms from Equation 4.22. Suppose that we want to test the null hypothesis:

H0: β = β0

against the alternative hypothesis:

Ha: β ≠ β0

at the k = 5 percent level of significance.14 We calculate the statistic

t = b − β0 sb

[4.25]

and the decision is to reject the null hypothesis, if t is either less than −t0.25,n−2 or greater than +t0.25,n−2 where the t0.25,n−2 value is obtained from the t-distribution (with n − 2 degrees of freedom) in Table 2 (Appendix B).15 Business applications of hypothesis test- ing are well advised to keep the level of significance small (i.e., no larger than 1 percent or 5 percent). One cannot justify building a marketing plan around advertising and retail displays incurring millions of dollars of promotional expense unless the demand estima- tion yields a very high degree of confidence that promotional expenditures actually “drive” sales (i.e., β ≠ 0). There are simply too many other potentially more effective ways to spend marketing dollars.

In the Sherwin-Williams Company example, suppose that we want to test (at the k = 0.05 level of significance) whether promotional expenditures are a useful variable in predicting paint sales. In effect, we wish to perform a statistical test to determine whether the sample value—that is, b = 0.433962—is significantly different from zero. The null and alternative hypotheses are

H0: β = 0 ðno relationship between X and YÞ Ha: β ≠ 0 ðlinear relationship between X and YÞ

14The level of significance (k) used in testing hypotheses indicates the probability of making an incorrect deci- sion with the decision rule—i.e., rejecting the null hypothesis when it is true. For example, with H0: β ≥ 0, set- ting k equal to .05 (i.e., 5 percent) indicates that there is one chance in 20 that we will conclude that an effect exists when no effect is present—i.e., a 5 percent chance of “false positive” outcomes.

Medical researchers trying to identify statistically significant therapies that could save lives, and research and development (R&D) researchers trying to identify potential blockbuster products, worry more about re- ducing the risk of “false negatives”—i.e., of concluding they have discovered nothing when their research could save a life or a company. Medical and R&D researchers therefore often perform hypothesis tests with k = 0.35 (i.e., with 65 percent confidence that the null hypothesis β = 0 should be rejected). They are willing to increase from 5 to 35 percent the probability of false positives (i.e., that they will conclude that β is positive for a treat- ment or therapy and can help the patient when in fact β = 0). This level of significance is preferred not be- cause they wish to engender false hope but rather because a higher tolerance for false positives reduces the probability of “false negative” decisions that would arise if they conclude that β = 0 when in fact β is positive. Ultimately, the question of what significance level to choose must be decided by the relative cost of false posi- tives and false negatives in a given situation. 15One-tail tests can also be performed. To test H0: β ≤ β0 against Ha: β > β0, one calculates t using Equation 4.25 and rejects H0 at the k level of significance of t > tk,n−2, where tk,n−2 is obtained from the t-distribution (Table 2 of Appendix B) with n − 2 degrees of freedom. Similarly, to test H0: β ≥ β0 against Ha: β < β0, one calculates t using Equation 4.25 and rejects H0 at the k level of significance if t < −tk,n−2.

Chapter 4: Estimating Demand 109

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Because there were 10 observations in the sample used to compute the regression equa- tion, the sample statistic b will have a t-distribution with 8(= n − 2) degrees of free- dom. From the t-distribution (Table 2 of Appendix B), we obtain a value of 2.306 for t.025,8. Therefore, the decision rule is to reject H0—in other words, to conclude that β ≠ 0 and that a statistically significant relationship exists between promotional expendi- tures and paint sales—if the calculated value of t is either less than −2.306 or greater than +2.306.

Using Equation 4.24, sb is calculated as

sb =

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð22:799Þ2

180,100 − ð1,250Þ2=10

s

= 0:14763

The calculated value of t from Equation 4.25 becomes

t = 0:433962 − 0

0:14763 = 2:939

Because this value is greater than +2.306, we reject H0. Therefore, based on the sample evidence, we conclude that at the 5 percent level of significance a linear, positive relationship exists between promotional expenditures and paint sales.

Example Are Designer Jeans and Lee Jeans Complements or Substitutes? To trigger retail sales, merchants often find that promotion, display, and assort- ment are almost as important as the right price point. VF Corporation, the parent company owning Lee jeans, is considering whether to add to their portfolio a new upscale brand of jeanswear, 7 for All Mankind. One question they hope to answer with demand estimation is whether their Lee jeans are purchased alone or whether the typical customer buying Lee also buys a dress pair of designer jeans such as Guess or 7 for All Mankind. Alternatively, would 7 for All Mankind jeans canni- balize the sales of Lee jeans? In short, are these two products perceived as comple- ments or substitutes? The company collected sales, price, and socioeconomic data on the target market for 48 quarters; the current values of the variables are QLEE = 50,000, PLEE = $20, PLEVI = $20, PGUESS = $35, Disposable Income = $80,000, Local Target Market Population = 100,000. Listed below are the results from the demand estimation:

QLEE = 133,500− 1,250 PLEE + 450 PLEVI − 571:43 PGUESS − 1:25 INC + 0:50 POP

ð3:0Þ ð−9:1Þ ð4:3Þ ð−1:5Þ ð−16:4Þ ð4:97Þ R2 = 0:92 SSE = 184; 000; 000

As one can readily see, price elasticity of demand is statistically significant since the absolute value of the t-score −9.1 is larger than the 99-percent critical value of 3.55. This price elasticity is calculated as −1250 × current price/current unit sales = −1,250 × $20/50,000 = −0.5. This inelastic numerical price elasticity suggests Lee

(Continued)

110 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Correlation Coefficient A first measure of the degree of association between two variables is called the correla- tion coefficient. Given n pairs of observations from the population, (x1, y1), (x2, y2),… , (xn, yn), the sample correlation coefficient is defined as

r = ∑ðxi − xÞðyi − yÞffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ∑ðxi − xÞ2∑ðyi − yÞ2

q and, when this expression is simplified, it is calculated as

r = n∑xiyi − ∑xi∑yiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

½n∑x2i − ð∑xiÞ2�½n∑y2i − ð∑yiÞ2� q [4.26]

The value of the correlation coefficient (r) ranges from +1 for two variables with per- fect positive correlation to −1 for two variables with perfect negative correlation. In Fig- ure 4.6, Panels (a) and (b) illustrate two variables that exhibit perfect positive and negative correlation, respectively. A positive correlation coefficient indicates that high va- lues of one variable tend to be associated with high values of the other variable, whereas a negative correlation coefficient indicates just the opposite—high values of one variable tend to be associated with low values of the other variable. Very few, if any, relationships between economic variables exhibit perfect correlation. Figure 4.6 Panel (c) illustrates zero correlation—no discernible relationship exists between the observed values of the two variables.

has introduced some effective switching costs and established a brand identity that builds customer loyalty and has desensitized its customers to price increases. In- come elasticity is also statistically significant at −1.25 × $80,000/50,000 = −2.0, sug- gesting that VF should not place its product in upscale department stores in suburban malls like Macy’s and Neiman Marcus but rather should market its jeans through Dillard’s and JCPenney. As expected, the sign on the price of Levi jeans is positive and statistically significant, suggesting that Lee and Levi jeans are substi- tutes. What about the designer jeans?

If the two products go out of stores together, VF may want to display and pro- mote them together. If not, perhaps 7 for All Mankind jeans should be sold through Internet distribution channels or specialty retail stores like Barney’s of New York so as not to cannibalize Lee jeans sales at Dillard’s and JCPenney. The marketing team also collects data on the potential size (POP) of the target cus- tomer market quarter by quarter and controls for its size. Proxying the effects of changes in the price of designer jeans with historical data on Guess jeans, the de- mand estimate is −547.43. Multiplying this coefficient estimate times the mean price of $35 and dividing by 50,000 yields a cross-price elasticity estimate of −0.38. But the t-score of −1.5 reveals that the cross price elasticity estimate is sta- tistically significantly different from zero with a confidence level of only 84 percent. This finding implies that indeed Lee jeans and designer jeans are not substitutes, and VF Corporation can stop worrying about the one product cannibalizing sales of the other.

Chapter 4: Estimating Demand 111

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Sherwin-Williams Company example discussed earlier can be used to illustrate the calculation of the sample correlation coefficient. Substituting the relevant quantities from Table 4.2 into Equation 4.26, we obtain a value of

r = 10ð229,100Þ − ð1,250Þð1,750Þffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

½10ð180,100Þ − ð1,250Þ2�½10ð314,900Þ − ð1,750Þ2� q

= 0:72059 or 0:721

for the correlation between the sample observations of promotional expenditures and paint sales.

The Analysis of Variance Another convenient measure of the overall “fit” of the regression model to the sample of observations is the r-squared.

We begin by examining a typical observation (yi) in Figure 4.7. Suppose we want to predict the value of Y for a value of X equal to xi. While ignoring the regression line for

FIGURE 4.6 Correlation Coefficient

y

x

r = �1

(a) Perfect positive correlation

y

x

r = –1

(b) Perfect negative correlation

y

x

r = 0

(c) No correlation

FIGURE 4.7 Partitioning the Total Deviation

Y

xi

Sample regression line

y = a + bXˆ

yi

Unexplained error

Explained error

Total deviationi

y _

X

Sample mean

112 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the moment, what error is incurred if we use the average value of Y (that is, y) as the best estimate of Y? The graph shows that the error involved, labeled the “total deviation,” is the difference between the observed value (yi) and y . Suppose we now use the sample regres- sion line to estimate Y. The best estimate of Y, given X = xi, is ŷi. As a result of using the regression line to estimate Y, the estimation error has been reduced to the difference be- tween the observed value (yi) and ŷi. In the graph, the total deviation (yi − y) has been partitioned into two parts—the unexplained portion of the total deviation (yi − ŷi) and that portion of the total deviation explained by the regression line (ŷi − y); that is,

Total deviation = Unexplained error + Explained error ðyi − yÞ = ðyi − ŷiÞ + ðŷi − yÞ

If we decompose the total error of each observation in the sample using this proce- dure and then sum the squares of both sides of the equation, we obtain (after some alge- braic simplification):16

Total SS = Unexplained SS + Explained SS

SST = ∑e2i + SSR = SSE + SSR

∑ðyi − yÞ2 = ∑ðyi − ŷiÞ2 + ∑ðŷi − yÞ2 [4.27] We can now use this sum-of-squares analysis to illustrate a measure of the fit of the

regression line to the sample observations. The sample coefficient of determination or r-squared (r2) is equal to the ratio of the Explained SS to the Total SS:

r2 = ∑ðŷi − yÞ2 Σðyi − yÞ2

= SSR SST

[4.28]

This r2 ratio measures the proportion of the variation in the dependent variable that is explained by the regression line (the independent variable) and ranges in value from 0—when none of the variation in Y is explained by the regression—to 1—when all the variation in Y is explained by regression.

Table 4.3 shows the calculation of the Explained, Unexplained, and Total SS for the Sherwin-Williams Company example that was introduced earlier. The Explained SS is 4,491.506 and the Total SS is 8,650.000, and therefore, by Equation 4.28 the coefficient of determination is

r2 = 4,491:506 8,650:000

= 0:519

In sum, the regression model, with promotional expenditures as the sole independent variable, explains about 52 percent of the variation in paint sales in the sample. Note also that this r2 is equal to the square of the correlation coefficient, that is, r2 = 0.519 = (r)2 = (0.72059)2. For the multiple linear regression model, the F-test is used to test the hypoth- esis that all the regression coefficients are zero.

The components of the r2 can be reconfigured into an F-ratio

F = SSR

SSE=d:f : [4.29]

to test whether the estimated regression equation explains a significant proportion of the variation in the dependent variable. The decision is to reject the null hypothesis of no

16A standard convention in statistics is to let the prefix “SS” represent the “Sum of Squares” as illustrated by SSE, the sum of squared errors.

coefficient of determination A measure of the proportion of total variation in the dependent variable that is explained by the independent variable(s).

Chapter 4: Estimating Demand 113

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

relationship between X and Y (i.e., no explanatory power) at the k = 5 percent level of significance if the calculated F-ratio is greater than the F0.5,1,n−2 value obtained from the F-distribution in Table 3 of the Statistical Tables (Appendix B). Thus, forming the F-ratio we obtain

F = 4,491:506 4,158:5=8

= 8:641

The critical value of F.05,1,8 from the F-distribution (Table 3 of Appendix B) is 5.32. Therefore, we reject, at the 5 percent level of significance, the null hypothesis that there is no relationship between promotion expenditures and paint sales. In other words, we conclude that the regression model does explain a significant proportion of the variation in paint sales in the sample.

MULTIPLE LINEAR REGRESSION MODEL A linear relationship containing two or more independent variables is known as a multi- ple linear regression model. In the (completely) general multiple linear regression model, the dependent variable Y is hypothesized to be a function of m independent variables X1, X2,… , Xm, and to be of the form

Y = α + β1X1 + β2X2 + … + βmXm + e [4.30]

In the Sherwin-Williams Company example, paint sales (Y) were hypothesized to be a function of three variables—promotional expenditures (A), price (P), and household dis- posable income (M) (see Equation 4.1):

Q = α + β1A + β2P + β3M + e

TABLE 4.3 CALCULATION OF THE EXPLAINED, UNEXPLAINED, AND TOTAL SS FOR THE

SHERWIN-WILLIAMS COMPANY

i x i y i

ŷ = 120.75475 + 0.433962 xi

EXPLAINED SS ðŷi − yÞ2

UNEXPLAINED SS (yi − ŷ i )

2 TOTAL SS ðyi − yÞ2

1 150 160 185.849 117.702 668.171 225.000

2 160 220 190.189 230.696 888.696 2,025.000

3 50 140 142.453 1,059.317 6.017 1,225.000

4 190 190 203.208 795.665 174.451 225.000

5 90 130 159.811 230.696 888.696 2,025.000

6 60 160 146.792 795.665 174.451 225.000

7 140 200 181.509 42.373 341.917 625.000

8 110 150 168.491 42.373 341.917 625.000

9 200 210 207.547 1,059.317 6.017 1,225.000

10 100 190 164.151 117.702 668.171 225.000

4,491.506 4,158.504 8,650.000*

Σðŷi − yÞ2 ∑ (yi − ŷi)2 Σðyi − yÞ2

*“Total SS” differs slightly from the sum of “Explained SS” and “Unexplained SS” because of rounding.

114 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Use of Computer Programs Using matrix algebra, procedures similar to those explained for the simple linear regression model can be employed for calculating the estimated regression coefficients (α’s and β’s). A variety of computer programs can be used to perform these procedures.

The output of these programs is fairly standardized to include the estimated regres- sion coefficients, t-statistics of the individual coefficients, R2, analysis of variance, and F-test of overall significance.

Estimating the Population Regression Coefficients From the computer output in Figure 4.8, the following regression equation is obtained:

Y = 310.245 + 0.008A − 12.202P + 2.677M [4.31]

The coefficient of the P variable (−12.202) indicates that, all other things being equal, a $1.00 price increase will reduce expected sales by −12.202 × 1,000 = 12,202 gallons in a given sales region.

Using the Regression Model to Make Forecasts As in the simple linear regression model, the multiple linear regression model can be used to make point or interval predictions. Point forecasts can be made by substituting the particular values of the independent variables into the estimated regression equation.

In the Sherwin-Williams example, suppose we are interested in estimating sales in a sales region where promotional expenditures are $185,000 (i.e., A = 185), selling price is $15.00 (P), and disposable income per household is $19,500 (i.e., M = 19.5). Substituting these values into Equation 4.31 yields

ŷ = 310.245 + .008(185) − 12.202(15.00) + 2.677(19.5) = 180.897 gallons

Whether to include one, two, or all three independent variables in predicting ŷ de- pends on the mean prediction error (e.g., here 185,000 − 180,897 = 4,103) in this and subsequent out-of-sample forecasts.

The standard error of the estimate (se) from the output in Figure 4.8 can be used to con- struct prediction intervals for Y. An approximate 95 percent prediction interval is equal to

ŷ ± 2se

For a sales region with the characteristics cited in the previous paragraph (i.e., A = 185, P = $15.00, and M = 19.5), an approximate 95 percent prediction interval for paint sales is equal to

180.897 ± 2(17.417)

or from 146,063 to 215,731 gallons.

Inferences about the Population Regression Coefficients Most regression programs test whether each of the independent variables (Xs) is statisti- cally significant in explaining the dependent variable (Y). This tests the null hypothesis:

H0: βi = 0

against the alternative hypothesis:

Ha: βi ≠ 0

The decision rule is to reject the null hypothesis of no relationship between paint sales (Y) and each of the independent variables at the .05 significance level, if the

Chapter 4: Estimating Demand 115

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

respective t-value for each variable is less than −t.025,6 = −2.447 or greater than t.025,6 = +2.447. As shown in Figure 4.8, only the calculated t-value for the P variable is less than −2.447. Hence, we can conclude that only selling price (P) is statistically signifi- cant (at the .05 level) in explaining paint sales. This inference might determine that marketing plans for this type of paint should focus on price and not on the effects of promotional expenditures or the disposable income of the target households.

FIGURE 4.8 Computer Output: Sherwin-Williams Company

Dep var: SALES (Y) N: 10 Multiple R: 0.889 Multiple R squared: 0.790

Adjusted multiple R squared: 0.684 Standard error of estimate: 17.417

Variable Coefficient Std error Std coef Tolerance T P(2 tail)

CONSTANT

PROMEXP (X1)

SELLPR (X2)

DISPINC (X3)

310.245

0.008

–12.202

2.677

95.075

0.204

4.582

3.160

0.000

0.013

–0.741

0.225

. 3.263

0.038

–2.663

0.847

0.017

0.971

0.037

0.429

0.3054426

0.4529372

0.4961686

Analysis of Variance

Source Sum-of-squares DF Mean-square F-ratio P

Regression 6829.866 3 2276.622 7.505 0.019 Residual 1820.134 6 303.356

Example The Estimated Demand for New Automobiles New car registrations of recently purchased autos vary over time in predictable ways. This economic theory reasoning can identify explanatory variables to include in the empirical model of new car demand. First, any consumer durable demand increases with rising population of the target customer group. Therefore, one must either control for population size as an explanatory variable or, alternatively, divide registrations by population, thereby creating a dependent variable of new car de- mand per capita, as in Table 4.4. Second, many new car purchases are financed, so minimum cash deposit requirements (Minimum Deposit) and auto financing rates (Interest Rate) are as important as the sale price (Price) in triggering a deci- sion to purchase during one month rather than another. Third, one would expect changes in disposable income (Income) to affect a household’s decision to replace its prior car. Higher household income would be associated with increased demand for superior models. Geopolitical events like the first Gulf War with its attendant gas price spikes should also affect demand for autos since gasoline is the primary complement in consumption of automobiles. Finally, as with other fad items, the introduction of popular new models enhances subsequent purchase decisions, so higher auto sales last period should have a positive effect on additional auto sales this period. These positive lagged effects of past sales should dampen as one gets further away from the product introductions that triggered the initial surge in sales.

Table 4.4 reports the empirical results. These are multiplicative exponential models, like Equations 4.5 and 4.6. Thus, the dependent variable and each of the

(Continued)

116 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

explanatory variables are in logarithms, and therefore the parameter estimates themselves can be interpreted as elasticities—price elasticity of demand, income elasticity of demand, interest rate elasticity of demand, minimum deposit elasticity of demand, and so forth.

The researchers who performed this study found, first of all, that market demand for new car registrations per capita was price inelastic (−0.341), suggesting some sub- stantial pricing power for many models. Next, a minimum deposit elasticity of −0.105 means that a 20 percent increase in cash deposit leads to a 2.1 percent de- crease in demand (i.e., 0.2 × −0.105 = 0.021). As expected, a 50 percent increase in auto financing rates from, say, 6 percent to 10 percent, leads to a 22 percent decline in auto demand (i.e., 0.5 × −0.436 = −0.22). Autos appear to be income elastic (1.947) such that a 10 percent increase in disposable income results in a 19.5 percent rise in auto demand. Gas shortages and time waiting in queues to buy the comple- ment gasoline led to a 14 percent reduction in auto demand.17 Finally, one period lagged demand; New Autot−1 had a significant positive effect on current purchases with the coefficient being between 0 and 1, as expected.18 Overall, this model ex- plained 96 percent of the time-series variation in new car sales per capita.

17The one exception is that the 0/1 Oil Crisis Dummy variable is added to the regression model directly, without taking logs, since the logarithm of zero is equal to negative infinity and is therefore an undefined value in the regression pro- grams. As a consequence, the elasticity of demand with respect to the 0/1 event is eβ − 1 (or in this case, e−0.146 − 1 = 0.864 − 1 = −13.6%). Hence, we conclude −14 percent. 18In contrast, a coefficient greater than 1 (or less than −1) on the lagged dependent variable would imply inherently unstable dynamics of exponentially accelerating demand growth (or decay).

TABLE 4.4 OLS ESTIMATES OF THE DETERMINANTS OF THE U.K.

PER CAPITA DEMAND FOR NEW AUTOS

EXPLANATORY VARIABLES COEFFICIENT

Constant −15.217

(−5.66)a

log Price −0.341

(−2.25)b

log Minimum Deposit −0.105

(−1.78)

log Interest Rate −0.436

(−5.31)a

log Income 1.947

(10.94)a

Oil Crisis Dummy −0.146

(−4.45)a

log New Cart-1 0.404

(3.24)a

Adjusted R2 0.965

Durbin-Watson 2.11

F 91.62

N 20

Notes: t-statistics in parentheses. Hypotheses tests are one-tailed. a,bStatistical significance at the 1 percent and 5 percent levels, respectively. Source: Managerial and Decision Economics 17, January 1996, pp. 19–23.

Chapter 4: Estimating Demand 117

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Analysis of Variance Techniques similar to those described for the simple linear regression model are used to evaluate the overall explanatory power of the multiple linear regression model.

The multiple coefficient of determination (r2) is a measure of the overall “fit” of the regression model. The squared multiple R value of .790 in Figure 4.8 indicates that the three-variable regression equation explains 79 percent of the total variation in the depen- dent variable (paint sales).

The F-ratio in the computer output of Figure 4.8 is used to test the hypothesis that all the independent variables (X1, X2,… , Xm) together explain a significant proportion of the variation in the dependent variable (Y). One is using the F-value to test the null hypothesis:

H0: All βi = 0

against the alternative hypothesis:

Ha: At least one βi ≠ 0

In other words, we are testing whether at least one of the explanatory variables con- tributes information for the prediction of Y. The decision is to reject the null hypothesis at the k level of significance if the F-value from the computer output is greater than the Fk,m,n−m−1 value from the F-distribution (with m and n − m − 1 degrees of freedom). Table 3 (Appendix B) provides F-values.

In the Sherwin-Williams example, suppose we want to test whether the three inde- pendent variables explain a significant (at the .05 level) proportion of the variation in income. The decision rule is to reject the null hypothesis (no relationship) if the calcu- lated F-value is greater than F.05,3,6 = 4.76. Because F = 7.505, we reject the null hypoth- eses and conclude that the independent variables are useful in explaining paint sales with (1 − 0.019) = 98.1 percent confidence.

SUMMARY

� Empirical estimates of the demand relationships are essential if the firm is to achieve its goal of shareholder wealth maximization. Without good estimates of the demand function facing a firm, it is impossible for that firm to make profit- maximizing price and output decisions.

� Consumer surveys involve questioning a sample of consumers to determine such factors as their will- ingness to buy, their sensitivity to price changes or levels, and their awareness of promotional campaigns.

� Focus groups make use of carefully directed discus- sion among groups of consumers. The results may be influenced by significant experimental bias.

� Market experiments observe consumer behavior in real-market situations. By varying product

characteristics, price, advertising, or other factors in some markets but not in others, the effects of these variables on demand can be determined. Market experiments are very expensive.

� Statistical techniques are often found to be of great value and relatively inexpensive as a means to make empirical demand function estimates. Re- gression analysis is often used to estimate statisti- cally the demand function for a good or service.

� The linear model and the multiplicative exponen- tial model are the two most commonly used func- tional relationships in demand studies.

� In a linear demand model, the coefficient of each independent variable provides an estimate of the change in quantity demanded associated with a one-unit change in the given independent variable,

118 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

holding constant all other variables. This marginal impact is constant at all points on the demand curve. The elasticity of a linear demand model with respect to each independent variable (e.g., price elasticity and income elasticity) is not con- stant, but instead varies over the entire range of the demand curve.

� In a multiplicative exponential demand model, the marginal impact of each independent variable on quantity demanded is not constant, but instead varies over the entire range of the demand curve. However, the elasticity of a multiplicative exponen- tial demand model with respect to each indepen- dent variable is constant and is equal to the estimated value of the respective parameter.

� The objective of regression analysis is to develop a functional relationship between the dependent and independent (explanatory) variable(s). Once a functional relationship (that is, regression equa- tion) is developed, the equation can be used to make forecasts or predictions concerning the value of the dependent variable.

� The least-squares technique is used to estimate the regression coefficients. Least-squares minimizes the sum of the squares of the differences between the observed and estimated values of the depen- dent variable over the sample of observations.

� The t-test is used to test the hypothesis that a spe- cific independent variable is useful in explaining variation in the dependent variable.

� The F-test is used to test the hypothesis that all the independent variables (X1, X2, . . . , Xm) in the re- gression equation explain a significant proportion of the variation in the dependent variable.

� The coefficient of determination (r2) measures the proportion of the variation in the dependent vari- able that is explained by the regression equation (that is, the entire set of independent variables).

� The presence of association does not necessarily imply causation. Statistical tests can only establish whether or not an association exists between vari- ables. The existence of a cause-and-effect economic relationship should be inferred from economic reasoning.

Exercises 1. Consider the Sherwin-Williams Company example discussed in this chapter (see Table 4.1). Suppose one is interested in developing a simple regression model with paint sales (Y) as the dependent variable and selling price (P) as the independent variable. a. Determine the estimated regression line. b. Give an economic interpretation of the estimated intercept (a) and slope

(b) coefficients. c. Test the hypothesis (at the .05 level of significance) that there is no relation-

ship (that is, β = 0) between the variables. d. Calculate the coefficient of determination. e. Perform an analysis of variance on the regression, including an F-test of the

overall significance of the results (at the .05 level). f. Based on the regression model, determine the best estimate of paint sales in

a sales region where the selling price is $14.50. Construct an approximate 95 percent prediction interval.

g. Determine the price elasticity of demand at a selling price of $14.50.

2. The Pilot Pen Company has decided to use 15 test markets to examine the sensi- tivity of demand for its new product to various prices, as shown in the following table. Advertising effort was identical in each market. Each market had approxi- mately the same level of business activity and population. a. Using a linear regression model, estimate the demand function for Pilot’s

new pen.

Answers to the exercises in blue can be found in Appendix D at the back

of the book.

Chapter 4: Estimating Demand 119

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

b. Evaluate this model by computing the coefficient of determination and by performing a t-test of the significance of the price variable.

c. What is the price elasticity of demand at a price of 50 cents?

TEST MARKET

PRICE CHARGED

QUANTITY SOLD (THOUSANDS OF PENS)

1 50¢ 20.0

2 50¢ 21.0

3 55¢ 19.0

4 55¢ 19.5

5 60¢ 20.5

6 60¢ 19.0

7 65¢ 16.0

8 65¢ 15.0

9 70¢ 14.5

10 70¢ 15.5

11 80¢ 13.0

12 80¢ 14.0

13 90¢ 11.5

14 90¢ 11.0

15 40¢ 17.0

3. In a study of housing demand, the county assessor is interested in developing a regression model to estimate the market value (i.e., selling price) of residential property within his jurisdiction. The assessor feels that the most important vari- able affecting selling price (measured in thousands of dollars) is the size of house (measured in hundreds of square feet). He randomly selected 15 houses and mea- sured both the selling price and size, as shown in the following table.

OBSERVATION SELLING PRICE

(× $1,000) SIZE (× 100 ft2)

i Y X2

1 265.2 12.0

2 279.6 20.2

3 311.2 27.0

4 328.0 30.0

5 352.0 30.0

6 281.2 21.4

7 288.4 21.6

8 292.8 25.2

9 356.0 37.2

10 263.2 14.4

11 272.4 15.0

12 291.2 22.4

13 299.6 23.9

14 307.6 26.6

15 320.4 30.7

120 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

a. Plot the data. b. Determine the estimated regression line. Give an economic interpretation of

the estimated slope (b) coefficient. c. Determine if size is a statistically significant variable in estimating selling

price. d. Calculate the coefficient of determination. e. Perform an F-test of the overall significance of the results. f. Construct an approximate 95 percent prediction interval for the selling price

of a house having an area (size) of 15 (hundred) square feet.

4. Cascade Pharmaceuticals Company developed the following regression model, using time-series data from the past 33 quarters, for one of its nonprescription cold remedies:

Y = −1.04 + 0.24X1 − 0.27X2

where Y = quarterly sales ðin thousands of casesÞ of the cold remedy X1 = Cascade’s quarterly advertising ð× $1,000Þ for the cold remedy X2 = competitors’ advertising for similar products ð× $10,000Þ

Here is additional information concerning the regression model:

sb1 = 0:032 sb2 = 0:070 R2 = 0:64 se = 1:63 F-statistic = 31:402

Durbin-Watson (d) statistic = 0.4995 a. Which of the independent variables (if any) appears to be statistically signif-

icant (at the 0.05 level) in explaining sales of the cold remedy? b. What proportion of the total variation in sales is explained by the regression

equation? c. Perform an F-test (at the 0.05 level) of the overall explanatory power of the

model. d. What additional statistical information (if any) would you find useful in the

evaluation of this model?

5. General Cereals is using a regression model to estimate the demand for Tweetie Sweeties, a whistle-shaped, sugar-coated breakfast cereal for children. The follow- ing (multiplicative exponential) demand function is being used:

QD = 6,280P −2.15A1.05N3.70

where QD = quantity demanded, in 10 oz: boxes P = price per box, in dollars A = advertising expenditures on daytime television, in dollars N = proportion of the population under 12 years old

a. Determine the point price elasticity of demand for Tweetie Sweeties. b. Determine the advertising elasticity of demand. c. What interpretation would you give to the exponent of N?

6. The demand for haddock has been estimated as

log Q = a + b log P + c log I + d log Pm

Chapter 4: Estimating Demand 121

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

where Q = quantity of haddock sold in New England P = price per pound of haddock I = a measure of personal income in the New England region

Pm = an index of the price of meat and poultry

If b = −2.174, c = 0.461, and d = 1.909, a. Determine the price elasticity of demand. b. Determine the income elasticity of demand. c. Determine the cross price elasticity of demand. d. How would you characterize the demand for haddock? e. Suppose disposable income is expected to increase by 5 percent next year.

Assuming all other factors remain constant, forecast the percentage change in the quantity of haddock demanded next year.

7. An estimate of the demand function for household furniture produced the follow- ing results:

F = 0.0036Y1.08R0.16P−0.48 r2 = 0.996

where F = furniture expenditures per household Y = disposable personal income per household R = value of private residential construction per household P = ratio of the furniture price index to the consumer price index

a. Determine the point price and income elasticities for household furniture. b. What interpretation would you give to the exponent for R? Why do you

suppose R was included in the equation as a variable? c. If you were a supplier to the furniture manufacturer, would you have pre-

ferred to see the analysis performed in physical sales units rather than dollars of revenue? How would this change alter the interpretation of the price coef- ficient, presently estimated as −0.48?

8. Consider again the Sherwin-Williams Company example discussed in this chapter (see Table 4.1). Suppose one is interested in developing a multiple regression model with paint sales (Y) as the dependent variable and promotional expendi- tures (A) and selling price (P) as the independent variables. a. Determine the estimated regression line. b. Give an economic interpretation of the estimated slope (bs) coefficients. c. Test the hypothesis (at the 5 percent level of significance) that there is no

relationship between the dependent variable and each of the independent variables.

d. Determine the coefficient of determination. e. Perform an analysis of variance on the regression, including an F-test of the

overall significance of the results (at the 5 percent level). f. Based on the regression model, determine the best estimate of paint sales in a

sales region where promotional expenditures are $80(000) and the selling price is $12.50.

g. Determine the point promotional and price elasticities at the values of pro- motional expenditures and selling price given in part (f).

122 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

9. The county assessor (see Exercise 4) feels that the use of more independent vari- ables in the regression equation might improve the overall explanatory power of the model.

In addition to size, the assessor feels that the total number of rooms, age, and whether or not the house has an attached garage might be important variables affecting selling price. The data for the 15 randomly selected dwellings are shown in the following table. a. Using a computer regression program, determine the estimated regression

equation with the four explanatory variables shown in the following table. b. Give an economic interpretation of each of the estimated regression

coefficients. c. Which of the independent variables (if any) is statistically significant (at the

.05 level) in explaining selling price? d. What proportion of the total variation in selling price is explained by the

regression model? e. Perform an F-test (at the 5 percent significance level) of the overall explana-

tory power of the model. f. Construct an approximate 95 percent prediction interval for the selling price

of a 15-year-old house having 1,800 square feet, 7 rooms, and an attached garage.

OBSERVATION

SELLING PRICE

(× $1,000) SIZE

(×100 ft2)

TOTAL NO. OF ROOMS AGE

ATTACHED GARAGE

(NO = 0, YES = 1)

i Y X1 X2 X3 X4

1 265.2 12.0 6 17 0

2 279.6 20.2 7 18 0

3 311.2 27.0 7 17 1

4 328.0 30.0 8 18 1

5 352.0 30.0 8 15 1

6 281.2 21.4 8 20 1

7 288.4 21.6 7 8 0

8 292.8 25.2 7 15 1

9 356.0 37.2 9 31 1

10 263.2 14.4 7 8 0

11 272.4 15.0 7 17 0

12 291.2 22.4 6 9 0

13 299.6 23.9 7 20 1

14 307.6 26.6 6 23 1

15 320.4 30.7 7 23 1

Case Exercise SOFT DRINK DEMAND ESTIMATION

Demand can be estimated with experimental data, time-series data, or cross-section data. Sara Lee Corporation generates experimental data in test stores where the effect of an NFL-licensed Carolina Panthers logo on Champion sweatshirt sales can be

Chapter 4: Estimating Demand 123

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

carefully examined. Demand forecasts usually rely on time-series data. In contrast, cross-section data appear in Table 1. Soft drink consumption in cans per capita per year is related to six-pack price, income per capita, and mean temperature across the 48 contiguous states in the United States.

Questions 1. Estimate the demand for soft drinks using a multiple regression program avail-

able on your computer. 2. Interpret the coefficients and calculate the price elasticity of soft drink demand. 3. Omit price from the regression equation and observe the bias introduced into the

parameter estimate for income. 4. Now omit both price and temperature from the regression equation. Should a

marketing plan for soft drinks be designed that relocates most canned drink ma- chines into low-income neighborhoods? Why or why not?

TABLE 1 Soft Drink Demand Data (available as an Excel file on this book’s Web site)

CANS/ CAPITA/YR

6-PACK $ PRICE

INCOME $/ CAPITA

MEAN TEMP. °F

Alabama 200 2.19 13 66

Arizona 150 1.99 17 62

Arkansas 237 1.93 11 63

California 135 2.59 25 56

Colorado 121 2.29 19 52

Connecticut 118 2.49 27 50

Delaware 217 1.99 28 52

Florida 242 2.29 18 72

Georgia 295 1.89 14 64

Idaho 85 2.39 16 46

Illinois 114 2.35 24 52

Indiana 184 2.19 20 52

Iowa 104 2.21 16 50

Kansas 143 2.17 17 56

Kentucky 230 2.05 13 56

Louisiana 269 1.97 15 69

Maine 111 2.19 16 41

Maryland 217 2.11 21 54

Massachusetts 114 2.29 22 47

Michigan 108 2.25 21 47

Minnesota 108 2.31 18 41

Mississippi 248 1.98 10 65

Missouri 203 1.94 19 57

Montana 77 2.31 19 44

Nebraska 97 2.28 16 49

Nevada 166 2.19 24 48

New Hampshire 177 2.27 18 35

New Jersey 143 2.31 24 54

New Mexico 157 2.17 15 56

124 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

CANS/ CAPITA/YR

6-PACK $ PRICE

INCOME $/ CAPITA

MEAN TEMP. °F

New York 111 2.43 25 48

North Carolina 330 1.89 13 59

North Dakota 63 2.33 14 39

Ohio 165 2.21 22 51

Oklahoma 184 2.19 16 82

Oregon 68 2.25 19 51

Pennsylvania 121 2.31 20 50

Rhode Island 138 2.23 20 50

South Carolina 237 1.93 12 65

South Dakota 95 2.34 13 45

Tennessee 236 2.19 13 60

Texas 222 2.08 17 69

Utah 100 2.37 16 50

Vermont 64 2.36 16 44

Virginia 270 2.04 16 58

Washington 77 2.19 20 49

West Virginia 144 2.11 15 55

Wisconsin 97 2.38 19 46

Wyoming 102 2.31 19 46

Chapter 4: Estimating Demand 125

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

4A APPENDIX

Problems in Applying the Linear Regression Model

INTRODUCTION When the simple linear and multiple linear regression models were discussed in Chapter 4, several assumptions were made about the nature of the relationships among the variables. How can we determine if the assumptions are being violated in a given situation? How does the violation of the assumptions affect the parameter estimates and prediction accu- racy of the model? What methods (if any) exist for overcoming the difficulties caused by the inapplicability of the assumptions in a given situation?

Econometrics provides answers to some, but not all, of these questions. Some of the problems that may invalidate the simple regression results and require further analysis include the following:

1. Autocorrelation 2. Heteroscedasticity 3. Specification and measurement errors 4. Multicollinearity 5. Simultaneous equation relationships and the identification problem 6. Nonlinearities

Each of these problems is discussed in this appendix.

Autocorrelation In many economic modeling and prediction problems, empirical data are in the form of a time series—a series of observations taken on the variables at different points in time. For example, we may be interested in predicting total domestic television sales by using U.S. disposable income as the independent variable over a period of 10 to 15 years. In working with time-series data, a problem known as autocorrelation can arise.

Recall that one of the assumptions underlying the regression model (specifically, As- sumption 3) is that the disturbance term et must be an independent random variable. In other words, we assume that each successive error, et, is independent of earlier and later errors so that the regression equation produces no predictable pattern in the successive values of the disturbance term. The existence of a significant pattern in the successive values of the error term constitutes autocorrelation. Successive values of the disturbance term can exhibit either positive or negative autocorrelation. Positive or negative autocor- relation, as shown in Figure 4A.1 (a) and (b), is inferred whenever successive distur- bances tend to be followed by disturbances of the same sign.

Negative autocorrelation reflects an undershooting and overshooting process like purchases of storable consumer goods. If a household buys too much breakfast cereal

autocorrelation An econometric problem characterized by the existence of a significant pattern in the successive values of the error terms in a linear regression model.

126

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

one week, that household will probably buy less than average the next week, and again more than average the following week. Positive autocorrelation can result from cyclical and seasonal variation in economic variables. Another cause of positive autocorrelation is self-reinforcing trends in consumer purchase patterns, for example, in fashion retail- ing. If Hermes scarves are in fashion, each successive week of sales data will be further above trend than the previous week until the fad slows and the Hermes look goes out of fashion. Either positive or negative autocorrelation may also result if significant ex- planatory variables are omitted from the regression equation or if nonlinear relation- ships exist.

As a safeguard when working with time-series data, the disturbances (et values) should be examined for randomness. Statistical tests are available to check for autocorre- lation. One commonly used technique is the Durbin-Watson statistic. It is calculated as follows:

d = ∑ n

t=2 ðet − et−1Þ2

∑ n

t=1 e2t

[4A.1]

where et is the estimated error term in period t and et−1 is the error term in period t − 1. The Durbin-Watson statistic tests for first-order autocorrelation, that is, whether the er- ror in period t is dependent on the error in the preceding period t − 1. The value of d ranges from 0 to 4. If there is no first-order autocorrelation, the expected value of d is 2. Values of d less than 2 indicate the possible presence of positive autocorrelation, whereas values of d greater than 2 indicate the possible presence of negative autocorrelation.

The presence of autocorrelation leads to several undesirable consequences in the regression results. First, although the estimates of α and β will be unbiased, the least- squares procedure will misestimate the sampling variances of these estimates. (An esti- mator is unbiased if its expected value is identical to the population parameter being estimated. The computed a and b values are unbiased estimators of α and β, respec- tively, because E(a) = α and E(b) = β.) In particular, the standard error (se in Equation 4.22) will either be inflated or deflated depending on whether we have positive or neg- ative autocorrelation. As a result, the use of the t-statistic to test hypotheses about these parameters may yield incorrect conclusions about the importance of the individ- ual predictor (that is, independent) variables. In addition, the r2 and F tests are invalid under autocorrelation.

FIGURE 4A.1 Types of Autocorrelation (Numbers 1, 2, 3,… , 10 refer to successive time periods.)

Y

X

Regression line

10

9

876

5 43

2 1

(a) Positive autocorrelation

Y

X

Regression line

10

9

8

7

6

5

4

3

2

1

(b) Negative autocorrelation

Appendix 4A: Problems in Applying the Linear Regression Model 127

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Several procedures are available for dealing with autocorrelation.1 If one can determine the functional form of the dependence relationship in the successive values of the residuals, then the original variables can be transformed by a lag structure to remove this pattern. Another technique that may help to reduce autocorrelation is to include a new linear trend or time variable in the regression equation. A third procedure is to calculate the first differences in the time series of each of the variables (that is, Yt+1 − Yt, X1,t+1 −X1,t,X2,t+1 −X2,t, and so on) and then calculate the regression equation using these transformed variables. A fourth method is to include additional variables of the form X1

2

or X1X2 in the regression equation. Usually one of these procedures will yield satisfactory results consistent with the independent errors assumption.

Heteroscedasticity In developing the ordinary least-squares regression model, another of the assumptions (Assumption 3) is that the error terms have a constant variance. Departure from this assumption is known as heteroscedasticity, which is indicated whenever there is a sys- tematic relationship between the absolute magnitude of the error term and the magni- tude of one (or more) of the independent variables.

One form of heteroscedasticity is illustrated in Figure 4A.2. Savings by households is postulated to be a function of household income. In this case, it is likely that more vari- ability will be found in the savings of high-income households compared with low- income households simply because high-income households have more money available for potential savings. Another example arising frequently with cross-sectional sales data is that the error variance with large-size retail stores, divisions, or firms exceeds the error variance for smaller entities.

In many cases, this form of heteroscedasticity can be reduced or eliminated by divid- ing all the variables in the regression equation by the independent variable that is thought to be causing the heteroscedasticity. Another method for dealing with heterosce- dasticity is to take logarithms of the data. Again, this transformation alters the form of the hypothesized relationship among the variables. More advanced, generalized least- squares techniques can account for the nonuniform error variance and preserve the orig- inal, hypothesized relationship.

FIGURE 4A.2 Illustration of Heteroscedasticity

Y

X

xi x2 0

p(y xi )

1See D. Gujarati, Basic Econometrics (New York: McGraw-Hill, Inc., 2007), Chapter 12, for a much more detailed discussion of procedures for dealing with autocorrelation.

heteroscedasticity An econometric problem characterized by the lack of a uniform variance of the error terms about the regression line.

128 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Specification and Measurement Errors Specification errors can result whenever one or more significant explanatory variables are omitted from the regression equation. If the omitted variable is moderately or highly cor- related with one of the explanatory variables included in the regression equation, then the affected regressor will be estimated with bias. Omitted variable bias may lead to over- estimating or underestimating the true regression coefficients.

The direction of bias in the estimated parameters should always be diagnosed. The mis- estimated parameter for X1(b1) may be written as the sum of the true parameter (β1) plus the effect of the omitted variable j, which depends on βj and the correlation coefficient rij:

b1 = β1 + βjr1,j [4A.2]

If one knows that the likely sign of the correlation coefficient between the omitted vari- able and the included explanatory variable (r1,j) is positive, and if the hypothesized effect of the omitted variable on the dependent variable (βj) is positive, the estimated parame- ter will be positively biased. For example, omitting household income from a demand estimation of luxury car rentals is likely to result in a positive bias on the parameter on the price variable, since higher income and the price paid for a luxury car for a week are probably positively correlated and since household income itself is hypothesized to be a positive determinant of luxury car rentals. On the other hand, in the Sherwin-Williams paint demand data, the correlation coefficient between disposable income and price is −0.514 (see Figure 4A.3). Omitting DISPINC from the demand estimation in Figure 4.8 would lead to a negative bias in the estimated effect of price on sales. Although these diagnoses of the omitted variable bias can never replace a fully and correctly specified model, they do allow much more informed decision making based on incomplete data.

Sometimes a close proxy variable is available and should be substituted for the omit- ted variable. The closer the proxy, the better the estimation because proxy variables al- ways introduce some measurement error. Measurement errors in the dependent variable do not affect the validity of the assumptions underlying the regression model or the pa- rameter estimates obtained by the least-squares procedure because these errors become part of the overall residual or unexplained error. However, measurement error in the ex- planatory variables introduces a stochastic component in the Xs and may cause the values of the error term ei to be correlated with the observed values of these explanatory variables. Consequently, the assumption that the disturbance terms are independent ran- dom variables (Assumption 3) is violated, and the resulting least-squares estimates of the regression coefficients (α, β) are biased.

Simultaneous equation estimation techniques discussed in a later section are one method of dealing with stochastic explanatory variables. Measurement error can also be modeled, if the form of the error in the X variables can be specified.

FIGURE 4A.3 Correlation Coefficients: Sherwin-Williams Company

SALES Y

PROMEXP X1

SELLPR X2

DISPINC X3

SALES Y 1.000

PROMEXP X1 0.721 1.000

SELLPR X2 0.866 0.739 1.000

DISPINC X3 0.615 0.710 0.514 1.000

Appendix 4A: Problems in Applying the Linear Regression Model 129

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Multicollinearity Whenever a high degree of intercorrelation exists among some or all of the explanatory variables in the regression equation, it becomes difficult to determine the separate influ- ences of each of the explanatory variables on the dependent variable because the stan- dard deviations (sb) of their respective regression coefficients become biased upward. Whenever two or more explanatory variables are highly correlated (or collinear), the t- test is therefore no longer a reliable indicator of the statistical significance of the individ- ual explanatory variables. Under such a condition, the least-squares procedure tends to yield highly unstable estimates of the regression coefficients from one sample to the next. The presence of multicollinearity, however, does not necessarily invalidate the use of the regression equation for prediction purposes. Provided that the intercorrelation pattern among the explanatory variables persists into the future, the equation can produce reli- able forecasts of the value of the dependent variable.

A number of techniques exist for dealing with multicollinearity: taking larger samples, detrending variables, or taking logs. In the end, however, the most important point is to diagnose the presence of multicollinearity so that insignificant hypothesis tests will not be wrongly attributed to weak cause-effect relationships. For example, consider the vari- ables that were used to explain paint sales in the Sherwin-Williams example discussed earlier. The correlation coefficients between each of the variables are shown in Figure 4A.3. Note the high degree of intercorrelation (in absolute value terms) between promo- tional expenditures and selling price and between promotional expenditures and dispos- able income, indicating that the standard deviations of the estimates of these three regression coefficients may be inflated.

Simultaneous Equation Relationships and the Identification Problem Many economic relationships are characterized by simultaneous interactions. For exam- ple, recognition of simultaneous relationships is at the heart of marketing plans. The op- timal advertising expenditure for a product line like Hanes Her Way hosiery depends on sales (i.e., on the quantity Hanes expects to sell). However, sales obviously also depend on advertising; a particularly effective ad campaign that just happens to match a random swing in customer fashion will drive sales substantially upward. And this sales boost will increase spending on advertising. Sales (i.e., demand) and advertising are simultaneously determined.

In attempting to estimate the parameters of simultaneous equation relationships with single equation models, one encounters the identification problem. For example, in de- veloping demand functions from empirical data, one is faced with the simultaneous rela- tionship between the demand function and the supply function. Suppose demand can be written as a function of price (P), income (M), and a random error e1,

Qd = β1 + β2P + β3M + e1 [4A.3]

and supply can be written as a function of price, input costs (I), and another random error e2,

Qs = α1 + α2P + α3I + e2 [4A.4]

or, rearranging, supply is

P = −α1 α2

+ 1 α2

ðQs − ε2Þ − α3 α2

I [4A.5]

multicollinearity An econometric problem characterized by a high degree of intercorrelation among some or all of the explanatory variables in a regression equation.

identification problem A difficulty encountered in empirically estimating a demand function by regression analysis. This problem arises from the simultaneous relationship between two functions, such as supply and demand.

130 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Because quantity demanded will equal quantity supplied in market-clearing equilibrium (i.e., Qd = Qs), we can substitute Equation 4A.3 for Qs in 4A.5 to obtain

P = −α1 α2

+ 1 α2

ðβ1 + β2P + β3M + ε1 − ε2Þ − α3 α2

I [4A.6]

P = 1

α2 − β2 ð−α1 + β1 + β3M + ε1 − ε2 − α3IÞ: [4A.7]

Observed values of P in Equation 4A.7 are quite obviously a stochastic explanatory vari- able because they are correlated with the disturbance term in the demand function e1. An ordinary least-squares regression of Equation 4A.3 (the demand function) therefore violates Assumption 3 that disturbance terms must be independent of the Xs—that is, E(Piei) = 0. As a result, the price coefficient in Equation 4A.3 (β2) will be biased.

To see why this poses a problem, recall that the price-output combinations actually observed result from an interaction of the supply and demand curves at a point in time. This is illustrated in Figure 4A.4. If D1, D2, D3, and D4 represent the true demand curves at four different points in time and S1, S2, S3, and S4 the corresponding supply curves, one would have been seriously misled to conclude that the true demand relation- ship was depicted by DD0 and was generally inelastic, when in fact demand was quite elastic and shifting. During the four successive time periods in which price-output com- binations were observed, both the demand and supply curves had shifted. Recall from Chapter 3 that to obtain a true estimate of the actual demand curve, one must hold con- stant the effects of all other variables in the demand functions, allowing only price and quantity demanded to vary.

FIGURE 4A.4 Quantity of Computer Memory Chips Purchased (Sold) with Shifting Supply and Demand

1

2

3

4

S1

0 Units

Quantity purchased (sold)

D

D1 S2

D2

S3

D3

S4

D4 D�

Pr ic

e of

c om

pu te

r m

em or

y ch

ip s

($ /u

ni t)

Appendix 4A: Problems in Applying the Linear Regression Model 131

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Under what circumstances may valid empirical estimates of the demand curve be made? If the supply curve shifts but the demand curve remains constant, observed price- output combinations will trace out the true demand curve. This is illustrated in Figure 4A.5. If, for example, technological advances were being introduced in the production of computer memory chips during Periods 1, 2, 3, and 4, then the supply curve would shift downward and to the right, from S1 to S4, tracing out the actual demand curve.

If both curves have shifted during the time period under consideration, identifying the demand curve requires that more than just price-output data be available. In other words, other variables, such as income and advertising, which may cause a shift in the demand function, must also be included in the model. Alternative statistical estimation techniques, such as two-stage least-squares (2SLS), may be used to separate supply curve shifts from shifts in the demand curve.2

NONLINEAR REGRESSION MODELS Although the relationships among many economic variables can be satisfactorily repre- sented using a linear regression model, situations do occur in which a nonlinear model is clearly required to portray adequately the relationship. Various models are available to deal with these situations. The transformations that are discussed here include the semi- logarithmic transformation, the double-log transformation, reciprocal transformation, and polynomial transformations.

FIGURE 4A.5 Quantity of Computer Memory Chips Purchased (Sold) with Stable Demand and Shifting Supply

Pr ic

e of

c om

pu te

r m

em or

y ch

ip s

($ /u

ni t)

1

2

3 4

S1

0 Units

Quantity purchased

D �

S2

S3 S4

D

2A discussion of these alternative estimation procedures is beyond the scope of this book. The reader is referred to Gujarti op. cit., Chapter 18.

132 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Semilogarithmic Transformation Sometimes, when heteroscedasticity is suspected, the dependent variable can be esti- mated best by taking the logarithm of one or more of the independent variables. For ex- ample, cross-section regression models, which use firm size as one of the independent variables, often take the log of firm size because of the potential problems caused by in- cluding in the same equation firms of $10 million in assets with firms of $10 billion in assets.

A semilog transformation of the form

Y = a + b log Assets + cX + Dz [4A.8]

is then estimated with standard least-squares techniques.

Double-Log Transformation In Chapter 4, we saw that a multiplicative exponential model (see Equation 4.5 and Table 4.4) is often used in demand studies. A three-variable exponential regression func- tion can be represented as

Z = AVβ1Wβ2 [4A.9]

Multiplicative exponential functions such as these can be transformed to linear relation- ships by taking logarithms of both sides of the equation to yield

log Z = log A + β1 log V + β2 log W

Reciprocal Transformation Another transformation, which is useful in relationships that exhibit an asymptotic be- havior, is the reciprocal transformation. The two possible cases are shown in Figure 4A.1. In Figure 4A.6 (a) the relationship is of the form

Y = α + β

Z [4A.10]

and in Figure 4A.6 (b) it is of the form

Example Constant Elasticity Demand: Pepsi If the data on soft drinks in the Case Exercise at the end of Chapter 4 (see Table 1) represent firm-level unit sales, marketing analysts in the company (PepsiCo, Inc.) may confirm that price elasticity coefficients have been very similar at several price points in recent years. If the same results of nearly constant elasticity estimates have arisen from detailed studies of income elasticity in upper- and lower-income neighborhoods, then a demand specification like Equation 4A.9 and a double-log estimation of the data in Table 1 would be indicated. Results of such an estimation are listed below:

Log Q =1:050 − 3:196 LogPrice + 0:221 LogIncome + 1:119 LogTemp ð1:72Þ ð−4:92Þ ð1:19Þ ð4:23Þ

SSE = 0:111 R2 = 0:671

Numbers in parentheses are t-score statistics.

Appendix 4A: Problems in Applying the Linear Regression Model 133

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Y = α − β

Z [4A.11]

Defining the transformation X = 1/Z, Equations 4A.10 and 4A.11 yield the following respective simple linear regression models:

Y = α + βX + e

and Y = α − βX + e

whose parameters can be estimated by the usual least-squares procedures.

Polynomial Transformation As will be seen in Chapter 8, the cost-output function for a firm is often postulated to follow a quadratic or cubic pattern. This type of relationship can be represented by means of a polynomial function. For example, a third-degree (that is, cubic) polynomial function can be represented as

Y = α + β1Z + β2Z 2 + β3Z

3 [4A.12]

Letting X1 = Z, X2 = Z 2, X3 = Z

3, Equation 4A.12 can be transformed into the follow- ing multiple linear regression model:

Y = α + β1X1 + β2X2 + β3X3

Standard least-squares procedures can be used in estimating the parameters of this model.

FIGURE 4A.6 Reciprocal Transformations

Y = α – β Z

Y

Z (a)

α

Y = α + β Z

Y

Z (b)

α

134 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

SUMMARY

� Various methodological problems can occur when applying the single-equation linear regression model. These include autocorrelation, heteroscedasticity, specification and measurement errors, multicollinear- ity, simultaneous equation relationships, and nonlin- earities. Many of these problems can invalidate the regression results. In most cases, methods are avail- able for detecting and overcoming these problems.

� Because of the simultaneous equation relationship that exists between the demand function and the supply function in determining the market- clearing price and quantity, analysts must exercise great care when estimating and interpreting empir- ical demand functions.

Exercises 1. Suppose an appliance manufacturer is doing a regression analysis, using quarterly time-series data, of the factors affecting its sales of appliances. A regression equa- tion was estimated between appliance sales (in dollars) as the dependent variable and disposable personal income and new housing starts as the independent vari- ables. The statistical tests of the model showed large t-values for both independent variables, along with a high r2 value. However, analysis of the residuals indicated that substantial autocorrelation was present. a. What are some of the possible causes of this autocorrelation? b. How does this autocorrelation affect the conclusions concerning the signifi-

cance of the individual explanatory variables and the overall explanatory power of the regression model?

c. Given that a person uses the model for forecasting future appliance sales, how does this autocorrelation affect the accuracy of these forecasts?

d. What techniques might be used to remove this autocorrelation from the model?

2. A product manager has been reviewing selling expenses (that is, advertising, sales commissions, and so on) associated with marketing a line of household cleaning products. The manager suspects that there may be some sort of diminishing mar- ginal returns relationship between selling expenses and the resulting sales generated by these expenditures. After examining the selling expense and sales data for various regions (all regions are similar in sales potential) shown in the following table and graph, however, the manager is uncertain about the nature of the relationship.

REGION SELLING

EXPENSE ($000) SALES (100,000

UNITS) LOG (SELLING

EXPENSE) LOG

(SALES)

A 5 1 3.6990 5.0000

B 30 4.25 4.4771 5.6284

C 25 4 4.3979 5.6021

D 10 2 4.0000 5.3010

E 55 5.5 4.7404 5.7404

F 40 5 4.6021 5.6990

G 10 1.75 4.0000 5.2430

H 45 5 4.6532 5.6990

I 20 3 4.3010 5.4771

J 60 5.75 4.7782 5.7597

Appendix 4A: Problems in Applying the Linear Regression Model 135

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

a. Using the linear regression model

Y = α + βX

where Y is sales and X is selling expenses, estimate α, β, and the r2 statistic by the least-squares technique.

b. Using the exponential function model

Y = αXβ

apply the double-logarithmic transformation to obtain a linear relationship that can be estimated by the least-squares technique.

c. Applying the least-squares technique, estimate α, β, and the r2 statistic for the transformed (linear) model in part (b). (Note that the logarithms of the X and Y variables needed in the calculations are given in the table.)

d. Based on the r2 statistics calculated in parts (a) and (c), which model appears to give a better fit of the data?

e. What implications does the result in part (d) have for the possible existence of a diminishing marginal returns relationship between sales and selling ex- penses as suggested by the manager?

f. What other transformations of the variables might we try to give a better fit to the data?

3. a. Using the data in Table 4.1 for the Sherwin-Williams Company, estimate a multiplicative exponential demand model (see Equation 4.5) for paint sales.

b. Compare the results in part (a) (i.e., parameter estimates, standard errors, statistical significance) with the linear model developed in the chapter.

4. The following table presents data on sales (S), advertising (A), and price (P):

OBSERVATION SALES (S) ADVERTISING (A) PRICE (P)

1 495 900 150

2 555 1,200 180

3 465 750 135

4 675 1,350 135

5 360 600 120

6 405 600 120

7 735 1,500 150

8 435 750 150

9 570 1,050 165

10 600 1,200 150

a. Estimate the following demand models:

(i) S = α + β1A + β2P (ii) S = αAβ1Pβ2

b. Determine whether the estimated values of β1 and β2 are statistically signifi- cant (at the .05 level).

c. Based on the value of R2 and the F-ratio, which model gives the best fit?

5. The county assessor (see Exercise 9 of Chapter 4) is concerned about possible multicollinearity between the size (X1) and total number of rooms (X2) variables. Calculate the correlation coefficient between these two variables and diagnose the magnitude of the collinearity problem.

136 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

5 CHAP T E R

Business and Economic Forecasting With over 50 domestic models of cars already on sale here in the U.S., the Japanese auto industry isn’t likely to carve out a big slice of the U.S. auto market.

BusinessWeek, 1969 [Today, Toyota and Honda are No. 2 and No. 4 in U.S. car sales.]

General Motors simply does not believe that an all-electric car will be attractive enough to the driving public to warrant our continuing to develop the product.

GM executive, 1996

CHAPTER PREVIEW Forecasting demand or input costs is often quite difficult. But it is one of the pivotal concerns of all managers because the shareholder value of a firm depends on accurately forecasting these components of the expected future cash flows. Forecasts at the firm level depend not only on prices and advertising and rival response tactics but also on the growth rate of the macro economy, the level of interest rates, the rate of unemployment, the value of the dollar in foreign exchange markets, and the rate of inflation. In this chapter we discuss the strengths and weaknesses of several classes of forecasting techniques including trend analysis, smoothing techniques, barometric indicators, survey and opinion polling, and time-series econometric methods.

MANAGERIAL CHALLENGE Excess Fiber Optic Capacity at Global Crossing Inc.1

The capacity of U.S. fiber optic networks to transmit high-speed data and voice signals once outstripped tele- com demand so much that 97 percent of the installed capacity in the United States was idle “silent fiber.” Indeed, if all telecom network traffic in the United States were routed through Chicago, only one-quarter of that city’s fiber optic capacity would be in use. With all this excess capacity in the market, fiber optic network providers like Global Crossing Inc. saw their

pricing power collapse. A 1-megabyte data connection between New York and Los Angeles that in 1995 had sold for as much as $12,000 per year declined by 2001 to $3,000 and by 2002 to $1,200. As their sales revenue plummeted, over 50 telecom network companies sought bankruptcy protection from their creditors.

How did this extreme excess capacity situation de- velop? First, an innovation in signal compression tech- nology caused capacity to outpace the growth of the

137

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

market. Between 1995 and 2003, dense wave divisional multiplexing (DWDM) expanded the transmission ca- pacity of a single strand of fiber optic cable from 25,000 one-page e-mails per second to 25 million per second, a thousand-fold increase. In addition, however, telecom network providers like Global Crossing and WorldCom were wildly optimistic about the growth of telecom traf- fic, fueled by the projected growth of the Internet, so they continued to lay additional cable. UUNet, a subsid- iary of WorldCom, forecasted that Internet use would

continue to double every 100 days. This exponential, more than tenfold growth per year (1,333 percent) was an extrapolation of the sales growth UUNet actually experienced in 1995–1996.

When the U.S. Department of Commerce and the Federal Communications Commission repeated this forecast, network providers continued to buy and bury redundant fiber optic cable. Between 2001 and 2002, U.S. fiber optic total capacity grew from 8,000 gigabytes per second to 80,000, but by 2003, the growth rate of demand had slowed (from 1,333 percent) to only 40 percent per year. Global Crossing’s failure to forecast this demand slowdown proved disastrous.

As we saw in Chapter 4, the penetration of the Internet into the American household has followed a classic S-shaped pattern. Initially, the Internet experi- enced exponential growth by the early adopters, then linear growth, and eventually fractional growth. This S-shaped penetration curve is drawn in Figure 5.1. It took both color television and the Internet only eight years (1947–1955 for color TVs and 1993–2001 for di- rect Internet access) to reach approximately 60 percent of U.S. households. However, color television adoptions then hit a wall, requiring another 30 years (1955–1985)

FIGURE 5.1 Market Penetration Curves of the Color Television and the Internet

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

1940 1950 1960 1970 1980 1990 2000 2010

Color TV

Internet access

Cont.

MANAGERIAL CHALLENGE Continued

© Do n Fa rra ll/ Ph ot od is c Re d/ Ge tty

Im ag es

138 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

THE SIGNIFICANCE OF FORECASTING Accurately forecasting future business prospects is one of the most important functions of management. Sales forecasts are necessary for operationsmanagers to plan the proper future levels of production. The financial managers require estimates of not only future sales reve- nue but also disbursements and capital expenditures. Forecasts of credit conditions must also be made so that the cash needs of the firm may be met at the lowest possible cost.

Public administrators and managers of not-for-profit institutions must also forecast. City government officials, for example, forecast the level of services that will be required of their various departments during a budget period. How many police officers will be needed to handle the public-safety problems of the community? How many streets will require repair next year, and how much will this cost? What will next year’s school en- rollment be at each grade level? The hospital administrator must forecast the health care needs of the community and the amount and cost of charity patient care.

SELECTING A FORECASTING TECHNIQUE The forecasting technique used in any particular situation depends on a number of factors.

Hierarchy of Forecasts The highest level of economic aggregation that is normally forecast is that of the national economy. The usual measure of overall economic activity is gross domestic product (GDP); however, a firm may be more interested in forecasting some of the specific com- ponents of GDP. For example, a machine tool firm may be concerned about plant and equipment expenditure requirements. Retail establishments are concerned about future levels and changes in disposable personal income rather than the overall GDP estimate.

The next levels in the hierarchy of economic forecasts are the industry sales forecast, fol- lowed by individual firm sales forecasts. A simple, single firm forecast might take the industry sales estimate and relate this to the expected market share of the individual firm. Future mar- ket share might be estimated on the basis of historical market shares as well as on changes that are anticipated in marketing strategies, new products and model changes, and relative prices.

Within the firm, a hierarchy of forecasts also exists. Managers often estimate company-wide or regional dollar sales and unit sales by product line. These forecasts

to achieve 98 percent penetration. Similarly with the Internet, customers who did not adopt high-speed broadband connections early are resistant to adopt now. Consequently, demand growth will likely remain low in Europe and the United States.

Discussion Questions

� Forecasting provides very useful projections for established products and services, but newly introduced offerings have wildly differ- ent success results. Name a few products that

have exploded with exponentially increasing demand shortly after their introduction. How about products that have largely been ignored?

� Do you see any common features among es- tablished and new products that might prove useful to an economic forecaster?

1Based on “Adoption Rate of Internet by Consumers Is Slowing,” Wall Street Journal (July 16, 2001), p. B1; “Has Growth of the Net Flattened?” Wall Street Journal (July 16, 2001), p. B8; “Behind the Fiber Glut,” Wall Street Journal (July 26, 2001), p. B1; and “Innovation Outpaced the Market- place,” Wall Street Journal (September 26, 2002), p. B1.

MANAGERIAL CHALLENGE Continued

Chapter 5: Business and Economic Forecasting 139

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

are used by operations managers in planning orders for raw materials, employee- hiring needs, shipment schedules, and release-to-production decisions. In addition, marketing managers use sales forecasts to determine optimal sales force allocations, to set sales goals, and to plan promotions. The sales forecast also constitutes a crucial part of the financial manager’s forecast of the cash needs of the firm. Long-term forecasts for the economy, the industry, and the firm are used in planning long-term capital expenditures for plant and equipment and for charting the general direction of the firm.

Criteria Used to Select a Forecasting Technique Some forecasting techniques are quite simple, inexpensive to develop and use, and best suited for short-term projections, whereas others are extremely complex, require signifi- cant amounts of time to develop, and may be quite expensive. The technique used in any specific instance depends on a number of factors, including the following:

1. The cost associated with developing the forecasting model 2. The complexity of the relationships that are being forecast 3. The time period of the forecast (long-term or short-term) 4. The lead time needed to make decisions based on the forecast, and 5. The accuracy required of the forecasting model

Evaluating the Accuracy of Forecasting Models In determining the accuracy, or reliability, of a forecasting model, one is concerned with the magnitude of the differences between the observed (actual) (Y) and the forecasted values (Ŷ). Various measures of model accuracy are available. For example, in the discus- sion of regression analysis in the previous chapter, the coefficient of determination, or R2, was used as a measure of the “goodness of fit” of the predicted values from the model to the patterns in the actual data. In addition, the mean prediction error, or root mean square error (RMSE),

RMSE =

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 n

ΣðYt − Ŷ tÞ2 r

[5.1]

is often used to evaluate the accuracy of a forecasting model (where n is the number of observations). The smaller the value of the RMSE, the greater the accuracy.

WHAT WENT RIGHT • WHAT WENT WRONG

Crocs Shoes2

In 2002, a colorful foam clog that was lightweight and nearly indestructible appeared on the market. Crocs were an overnight sensation, and 100 million pairs were sold in seven years. The company forecasted double-digit sales growth for the next five years, and then did a very success- ful initial public offering that raised $200 million. The new capital was reinvested to ramp up Crocs’ manufactur- ing capacity. Then the severe worldwide recession of

2008–2009 hit, and the bottom fell out of the market. No one needed replacements for a nearly indestructible fashion fad that couldn’t help you look for a job. In one year (2007–2008), the company swung from a profit of $168 million to a loss of $185 million.

2Based on “Once-Trendy Crocs Could Be on Their Last Legs,” Washington Post (July 16, 2009), p. C2.

140 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

ALTERNATIVE FORECASTING TECHNIQUES The managerial economist may choose from a wide range of forecasting techniques. These can be classified in the following general categories:

• Deterministic trend analysis • Smoothing techniques • Barometric indicators • Survey and opinion-polling techniques • Macroeconometric models • Stochastic time-series analysis • Forecasting with input-output tables

DETERMINISTIC TREND ANALYSIS Data collected for use in forecasting the value of a particular variable may be classified into two major categories—time-series or cross-sectional data. Time-series data are de- fined as a sequence of the values of an economic variable at different points in time. Cross-sectional data are an array of the values of an economic variable observed at the same time, like the data collected in a census across many individuals in the population. No matter what type of forecasting model is being used, one must decide whether time- series or cross-sectional data are most appropriate.

Components of a Time Series In the analysis of time-series data, time in years, months, or weeks is represented on the horizontal axis, and the values of the dependent variable are on the vertical axis. The variations that are evident in the time series in Figure 5.2 can be decomposed into four components:

1. Secular trends. These are long-run trends that cause changes in an economic data series [solid line in Panel (a) of Figure 5.2]. For example, in empirical demand anal- yses, such factors as increasing population size or evolving consumer tastes may result in trend increases or decreases of a demand series over time.

2. Cyclical variations. These are major expansions and contractions in an economic series that are usually greater than a year in duration [broken line in Panel (a) of Figure 5.2]. For example, the housing industry appears to experience regular expan- sions following contractions in demand. When cyclical variations are present, regression estimates using the raw data will be distorted due to the presence of pos- itive autocorrelation. Care must then be taken to specify an appropriate lag structure to remove the autocorrelation.

3. Seasonal effects. Seasonal variations during a year tend to be more or less consistent from year to year. The data in Panel (b) of Figure 5.2 (broken line) show significant seasonal variation. For example, two-thirds of Hickory Farms’ (a retailer of holiday food gifts) annual sales occur between November and January.

4. Random fluctuations. Finally, an economic series may be influenced by random factors that are unpredictable [solid line in Panel (b) of Figure 5.2], such as hurri- canes, floods, and tornados, as well as extraordinary government actions like a wage- price freeze or a declaration of war.

time-series data A series of observations taken on an economic variable at various past points in time.

cross-sectional data Series of observations taken on different observation units (for example, households, states, or countries) at the same point in time.

secular trends Long- run changes (growth or decline) in an economic time-series variable.

cyclical variations Major expansions and contractions in an economic series that usually are longer than a year in duration.

seasonal effects Variations in a time series during a year that tend to appear regularly from year to year.

Chapter 5: Business and Economic Forecasting 141

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Some Elementary Time-Series Models The simplest time-series model states that the forecast value of the variable for the next period will be the same as the value of that variable for the present period:

Ŷt+1 = Yt [5.2]

For example, consider the sales data shown in Table 5.1 for the Buckeye Brewing Company. To forecast monthly sales, the model uses actual beer sales for March 2007 of 2,738 (000) barrels as the forecast value for April.

FIGURE 5.2 Secular, Cyclical, Seasonal, and Random Fluctuations in Time Series Data

Year 2

Cyclical variations

Secular trend

Panel (a)

1995

Time (years)

1996 1997 1998 1999 2000 2001 2002 2003 2004

Seasonal effects

J

Time (month)

Panel (b)

Random fluctuations

F M A M J J A S O N D J F M A M J J A S O N D Year 1

Sa le

s ($

) Sa

le s

($ )

142 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Where changes occur slowly and the forecast is being made for a relatively short pe- riod in the future, such a model may be quite useful. However, because Equation 5.2 re- quires knowledge of this month’s sales, the forecaster may be faced with the task of speeding up the collection of actual data. Another problem with this model is that it makes no provision for incorporating special promotions by the firm (or its competitors) that could cause large sales deviations.

Further examination of the Buckeye beer sales data in Table 5.1 indicates a slight upward trend in sales—beer sales in most months are higher than in the same month of the previous year. Second, we note that sales are somewhat seasonal—beer sales are high during the summer months and low during the winter. The tendency for recent increases to trigger further increases in beer sales may be incorporated by slightly adjust- ing Equation 5.2 to yield this equation:

Ŷt+1 = Yt + (Yt − Yt−1) [5.3]

For example, Buckeye’s sales forecast for April 2007 using this model would be

Ŷ t+1 = 2,738 + ð2,738 − 2,693Þ = 2,783ð000Þ barrels

Other forecasting models that incorporate trends and seasonal effects such as these are discussed later.

Secular Trends Long-run changes in an economic time series can follow a number of different types of trends. Three possible cases are shown in Figure 5.3. A linear trend is shown in Panel (a). Panels (b) and (c) depict nonlinear trends. In Panel (b), the economic time series follows a constant rate of growth pattern. The earnings of many corporations follow this type of trend. Panel (c) shows an economic time series that exhibits a declining rate of growth. Sales of a new product may follow this pattern. As market saturation oc- curs, the rate of growth will decline over time.

TABLE 5.1 BUCKEYE BREWING COMPANY ’S MONTHLY BEER SALES

(THOUSANDS OF BARRELS)

YEAR

MONTH 2005 2006 2007

January 2,370 2,446 2,585

February 2,100 2,520 2,693

March 2,412 2,598 2,738

April 2,376 2,533

May 3,074 3,250

June 3,695 3,446

July 3,550 3,986

August 4,172 4,222

September 3,880 3,798

October 2,931 2,941

November 2,377 2,488

December 2,983 2,878

Chapter 5: Business and Economic Forecasting 143

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Linear Trends A linear time trend may be estimated by using least-squares regres- sion analysis to provide an equation of a straight line of “best fit.” (See Chapter 4 for a further discussion of the least-squares technique.) The equation of a linear time trend is given in the general form

Ŷt = α + βt [5.4]

where Ŷt is the forecast or predicted value for period t, α is the y intercept or constant term, t is a unit of time, and β is an estimate of this trend factor.

FIGURE 5.3 Time-Series Growth Patterns

Y

Time (t)

(a) Linear trend

Y = � + �t Y

Time (t)

(b) Constant rate of growth trend

Y = Yo (1 + g) t

Y

Time (t)

Y = e�1 – �2/t

(c) Declining rate of growth trend

^

^ ^

Example Linear Trend Forecasting: Prizer Creamery Suppose one is interested in forecasting monthly ice cream sales of the Prizer Creamery for 2007. A least-squares trend line could be estimated from the ice cream sales data for the past four years (48 monthly observations), as shown in Figure 5.4. Assume that the equation of this line is calculated to be

Ŷt = 30,464 + 121.3t

where Ŷ t = predicted monthly ice cream sales in gallons in month t 30,464 = number of gallons sold when t = 0

t = time period ðmonthsÞðwhere December 2002 = 0, January 2003 = 1, February 2003 = 2, March 2003 = 3, . . .Þ

The coefficient (121.3) of t indicates that sales may be expected to increase by 121.3 gallons on the average each month. Based on this trend line and ignoring any seasonal effects, forecasted ice cream sales for August 2007 (t = 56) would be

Y56 = 30,464 + 121:3ð56Þ = 37,257 gallons

This seasonally unadjusted forecast is given by the August 2007 point (�) on the trend line in Figure 5.4. As can be seen in the graph, ice cream sales are subject to seasonal variations. Later in this section we will show how this seasonal effect can be incorporated into the forecast.

144 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Linear time trend forecasting is easy and inexpensive to do, but it is generally too simple and inflexible to be used in many forecasting circumstances. Constant rate of growth time trends is one alternative.

Constant Rate of Growth Trends The formula for the constant rate of growth forecasting model is

Ŷt = Y0(1 + g) t [5.5]

where Ŷt is the forecasted value for period t, Y0 is the initial (t = 0) value of the time series, g is the constant growth rate per period, and t is a unit of time. The predicted value of the time series in period t (Ŷt) is equal to the initial value of the series (Y0) compounded at the growth rate (g) for t periods. Because Equation 5.5 is a nonlinear relationship, the parameters cannot be estimated directly with the ordinary least- squares method. However, taking logarithms of both sides of the equation gives

log Ŷt = log Y0 + log(1 + g) · t

or Ŷ 0t = α + βt [5.6]

FIGURE 5.4 Prizer Creamery: Monthly Ice Cream Sales

200720060 J FMAMJ J A SONDJ FMAMJ J A SONDJ FMAMJ JA SONDJ FMAMJ J A SONDJ FMAMJ J A SOND

2003 12 2004 24 36 48 60 Time (t) (month)

2005

24,000

26,000

28,000

30,000

32,000

34,000

36,000

38,000

40,000

42,000

0

Historical period

Forecast period

Seasonally adjusted forecast (August 2007)

Unadjusted forecast

(August 2007)

Trend line Yt = 30,464+121.3t

Sa le

s Y

t (g

al lo

ns )

^

Chapter 5: Business and Economic Forecasting 145

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

where Ŷ 0t = log Ŷ t , α = log Y0; and β = logð1 + gÞ. Equation 5.6 is a linear relationship whose parameters can be estimated using standard linear regression techniques.

For example, suppose that annual earnings data for the Fitzgerald Company for the past 10 years have been collected and that Equation 5.6 was fitted to the data using least-squares techniques. The annual rate of growth of company earnings was estimated to be 6 percent. If the company’s earnings this year (t = 0) are $600,000, then next year’s (t = 1) forecasted earnings would be

Ŷ1 = 600,000ð1 + 0:06Þ1 = $636,000

Similarly, forecasted earnings for the year after next (t = 2) would be

Ŷ2 = 600,000ð1 + 0:06Þ2 = $674,160

Declining Rate of Growth Trends The curve depicted in Figure 5.3, panel (c) is particularly useful for representing sales penetration curves in marketing applications. Using linear regression techniques, one can specify a semilog estimating equation,

log Ŷt = β1 − β2(1/t)

and recover the β1 and β2 parameters of this nonlinear diffusion process as a new prod- uct spreads across a target population. β1 and β2 measure how quickly a new product or new technology or brand extension penetrates and then slowly (ever more slowly) satu- rates a market.

Seasonal Variations When seasonal variations are introduced into a forecasting model, its short-run predic- tive power may be improved significantly. Seasonal variations may be estimated in a number of ways.

Ratio to Trend Method One approach is the ratio to trend method. This method assumes that the trend value is multiplied by the seasonal effect.

Example Seasonally Adjusted Forecasts: Prizer Creamery (continued) Recall in the Prizer Creamery example discussed earlier that a linear trend analysis (Equation 5.4) yielded a sales forecast for August 2007 of 37,257 gallons. This esti- mate can be adjusted for seasonal effects in the following manner. Assume that over a four-year period (2003–2006) the trend model predicted the August sales patterns shown in Table 5.2 and that actual sales are as indicated. These data indi- cate that, on the average, August sales have been 7.0 percent higher than the trend value. Hence, the August 2007 sales forecast should be seasonally adjusted upward by 7.0 percent to 39,865. The seasonally adjusted forecast is shown by the point (&� ) above the trend line in Figure 5.4. If, however, the model predicted February 2007 (t = 50) sales to be 36,529, but similar data indicated February sales to be 10.8 percent below trend on the average, the forecast would be adjusted downward to 36,529(1 − 0.108) = 32,584 gallons.

(Continued)

146 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Dummy Variables Another approach for incorporating seasonal effects into the linear trend analysis model is the use of dummy variables. A dummy variable is a vari- able that normally takes on one of two values—either 0 or 1. Dummy variables, in gen- eral, are used to capture the impact of certain qualitative factors in an econometric relationship, such as sex—male-0 and female-1. This method assumes that the seasonal effects are added to the trend value. If a time series consists of quarterly data, then the following model could be used to adjust for seasonal effects:

Ŷt = α + β1t + β2 D1t + β3 D2t + β4 D3t [5.7]

where D1t = 1 for first-quarter observations and 0 otherwise, D2t = 1 for second-quarter observations and 0 otherwise, D3t = 1 for third-quarter observations and 0 otherwise, and α and β are parameters to be estimated using least-squares techniques. In this model the values of the dummy variables (D1t, D2t, D3t) for observations in the fourth quarter of each year (base period) would be equal to zero. In the estimated model, the value β2 D1t repre- sents the impact of a first-quarter observation (D1) on values of the forecast, Yt, relative to the forecast from the omitted class (4th quarter), when D2t and D3t take values of 0.

The introduction of these trend and seasonality factors into a forecasting model should significantly improve the model’s ability to predict short-run turning points in the data series, provided the historical causal factors have not changed significantly.3

The models of time-series trend forecasting discussed in this section may have sub- stantial value in many areas of business. However, such models do not seek to relate changes in a data series to the causes underlying observed values in the series. For exam- ple, the nation’s money supply series has at times proved very useful for forecasting in- flationary pressure in the economy. But narrow definitions of the nation’s money supply have gradually broadened to include bank-card lines of credit, which may have become a more important measure of household purchasing power. Inflation forecasts based on narrow money supply measures today would yield large errors between the actual and predicted inflation (see Figure 5.5).

SMOOTHING TECHNIQUES Smoothing techniques are another type of forecasting model, which assumes that a re- petitive underlying pattern can be found in the historical values of a variable that is being forecast. By taking an average of past observations, smoothing techniques attempt

TABLE 5.2 PRIZER CREAMERY ’S AUGUST ICE CREAM SALES

(AUGUST) FORECAST ACTUAL ACTUAL/ FORECAST

2002 31,434 33,600 1.0689

2003 32,890 35,600 1.0824

2004 34,346 36,400 1.0598

2005 35,801 38,200 1.0670

2006 37,257 — 0000 —00000

Sum = 4.2781

Adjustment factor = 4.2781/4 = 1.0695 (i.e., 1.07)

3See more extensive discussion of these issues in F. Diebold, Elements of Forecasting, 4th ed. (Cincinnati: South-Western College Publishing, 2007).

Chapter 5: Business and Economic Forecasting 147

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

to eliminate the distortions arising from random variation in the series. As such, smoothing techniques work best when a data series tends to change slowly from one pe- riod to the next with few turning points. Housing price forecasts would be a good appli- cation for smoothing techniques. Gasoline price forecasts would not. Smoothing techniques are cheap to develop and inexpensive to operate.

Moving Averages Moving averages are one of the simplest of the smoothing techniques. If a data series possesses a large random factor, a trend analysis forecast like those discussed in the pre- vious section will tend to generate forecasts having large errors from period to period. In an effort to minimize the effects of this randomness, a series of recent observations can be averaged to arrive at a forecast. This is the moving average method. A number

Example Dummy Variables and Seasonal Adjustments: Value-Mart Company The Value-Mart Company (a small chain of discount department stores) is inter- ested in forecasting quarterly sales for next year (2008) based on Equation 5.7. Using quarterly sales data for the past eight years (2000–2007), the following model was estimated:

Ŷt = 22.50 + 0.250t − 4.50D1t − 3.20D2t − 2.10D3t [5.8]

where Ŷ t = predicted sales ð$ millionÞ in quarter t 22:50 = quarterly sales ð$ millionÞ when t = 0

t = time period ðquarterÞ ðwhere the fourth quarter of 1999 = 0; first quarter of 2000 = 1, second quarter of 2000 = 2, . . .Þ

The coefficient of t (0.250) indicates that sales may be expected to increase by $0.250 million on the average each quarter. The coefficients of the three dummy variables (−4.50, −3.20, and −2.10) indicate the change (i.e., reduction because the coefficients are negative) in sales in Quarters 1, 2, and 3, respectively, because of seasonal effects. Based on Equation 5.8, Value-Mart’s quarterly sales forecasts for 2008 are shown in Table 5.3. On just this basis, Value-Mart would have ordered inventory for what proved to be a disastrous 2008–2009.

TABLE 5.3 VALUE-MART ’S QUARTERLY SALES FORECAST (2008)

TIME PERIOD

DUMMY VARIABLE

SALES FORECAST ($ MILLION)

Ŷ = 22.50 + 0.250 t − 4.50 D1 t − 3.20 D2 t

QUARTER t D1 t D2 t D3 t −2.10 D3 t

1 33 1 0 0 26.25

2 34 0 1 0 27.80

3 35 0 0 1 29.15

4 36 0 0 0 31.50

148 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

of observed values are chosen, their average is computed, and this average serves as a forecast for the next period. In general, a moving average may be defined as

Ŷ t+1 = Yt + Yt−1 + . . . + Yt−N+1

N [5.9]

where Ŷ t+1 = forecast value of Y for one period in the future Yt , Yt−1,Yt−N+1 = observed values of Y in periods t, t − 1, . . . ,

t − N + 1, respectively N = number of observations in the moving average

The greater the number of observations N used in the moving average, the greater the smoothing effect because each new observation receives less weight (1/N) as N increases. Hence, generally, the greater the randomness in the data series and the slower the turning point events in the data, the more preferable it is to use a rela- tively large number of past observations in developing the forecast. The most appro- priate moving average period is the choice of N that minimizes the root mean square error (Equation 5.1).

FIGURE 5.5 Actual and Expected Inflation

Forecast date

Actual

Expected

16

14

12

10

8

6

4

2

0 60 65 70 75 80 85 90 95

Percent

Note: “Expected” is the inflation forecast for the year following the forecast date; “Actual” is the actual inflation rate over that period.

Source: Federal Reserve Bank of Philadelphia, Business Review, May/June 1996.

Chapter 5: Business and Economic Forecasting 149

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Moving Average Forecasts: Walker Corporation The Walker Corporation is examining the use of various smoothing techniques to fore- cast monthly sales. The company collected sales data for 12 months (2006) as shown in Table 5.4 and Figure 5.6. One technique under consideration is a three-month moving average. Equation 5.9 can be used to generate the forecasts. The forecast for Period 4 is computed by averaging the observed values for Periods 1, 2, and 3.

Ŷ4 = Y3 + Y2 + Y1

N

= 1,925 + 1,400 + 1,950

3 = 1,758

[5.10]

Similarly, the forecast for Period 5 is computed as

Ŷ5 = Y4 + Y3 + Y2

N

= 1,960 + 1,925 + 1,400

3 = 1,762

[5.11]

Note that if one subtracts Ŷ4 from Ŷ5, the result is the change in the forecast from Ŷ4, or

ΔŶ4 = Ŷ5 − Ŷ4

= Y4 + Y3 + Y2

N −

Y3 + Y2 + Y1 N

= Y4 N

− Y1 N

[5.12]

TABLE 5.4 WALKER CORPORATION ’S THREE-MONTH MOVING

AVERAGE SALES FORECAST TABLE

SALES ($1,000) ERROR

t MONTH ACTUAL Yt FORECAST Ŷ t (Ŷ t − Ŷ t) (Ŷ t − Ŷ t) 2

1 January 2006 1,950 — — —

2 February 1,400 — — —

3 March 1,925 — — —

4 April 1,960 1,758 202 40,804

5 May 2,800 1,762 1,038 1,077,444

6 June 1,800 2,228 −428 183,184

7 July 1,600 2,187 −587 344,569

8 August 1,450 2,067 −617 380,689

9 September 2,000 1,617 383 146,689

10 October 2,250 1,683 567 321,489

11 November 1,950 1,900 50 2,500

12 December 2,650 2,067 583 339,889

13 January 2007 * 2,283 — 0000 —0000000

Sum = 2,837,257

RMSE = ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2,837,257=9

p = $561ð000Þ

(Continued)

150 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

First-Order Exponential Smoothing One criticism of moving averages as smoothing techniques is that they normally give equal weight (a weight of 1/N) to all observations used in preparing the forecast, even though intuition often indicates that the most recent observation probably contains

Adding this change to Ŷ4, the following alternative expression for Ŷ5 can be derived:

Ŷ5 = Ŷ4 + Y4 N

− Y1 N

[5.13]

or, in general,

Ŷ t+1 = Ŷ t + Yt N

− Yt−N N

[5.14]

which indicates that each moving average forecast is equal to the past forecast, Ŷt, plus the weighted effect of the most recent observation, Yt/N, minus the weighted effect of the oldest observation that has been dropped, Yt−N/N. As N becomes larger, the smoothing effect increases because the new observation, Yt, has a small impact on the moving average.

As shown in Table 5.4, Walker’s forecast for January 2007 (t = 13) is $2,283 (000). Note also that the root mean square error (RMSE) of the three-month (N) moving average period is $561(000).

FIGURE 5.6 Walker Corporation’s Three-Month Moving Average Sales Forecast Chart

1,400 1,500 1,600 1,700 1,800 1,900 2,000 2,100 2,200 2,300 2,400 2,500 2,600 2,700 2,800 2,900 3,000

^

2007

0 J F M A M J J A S O N D J F 1 2 3 4 5 6 7 8 9 10 11 12 13 14

2006 Time (t)

Three-month moving average

forecast

Y t,

Y t

Sa le

s ($

00 0)

January 2007

forecast

Chapter 5: Business and Economic Forecasting 151

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

more immediately useful information than more distant observations. Exponential smoothing is designed to overcome this objection.4

Consider the following alternative forecasting model:

Ŷt+1 = wYt + (1 − w)Ŷt [5.15]

This model weights the most recent observation by w (some value between 0 and 1 in- clusive), and the past forecast by (1 − w). A large w indicates that a heavy weight is being placed on the most recent observation.5

Using Equation 5.15, a forecast for Ŷt may also be written as

Ŷt = wYt−1 + (1 − w)(Ŷt−1) [5.16]

By substituting Equation 5.16 into 5.15, we get

Ŷt+1 = wYt + w(1 − w)Yt−1 + (1 − w) 2Ŷt−1 [5.17]

By continuing this process of substitution for past forecasts, we obtain the general equation

Ŷt+1 = wYt + w(1 − w)Yt−1 + w(1 − w) 2Yt−2 + w(1 − w)

3Yt−3 + . . . [5.18]

Equation 5.18 shows that the general formula (Equation 5.15) for an exponentially weighted moving average is a weighted average of all past observations, with the weights defined by the geometric progression:

w, (1 − w)w, (1 − w)2w, (1 − w)3w, (1 − w)4w, (1 − w)5w, . . . [5.19]

For example, a w of 2/3 would produce the following series of weights:

w = 0:667 ð1 − wÞw = 0:222 ð1 − wÞ2w = 0:074 ð1 − wÞ3w = 0:024 ð1 − wÞ4w = 0:0082 ð1 − wÞ5w = 0:0027

With a high initial value of w, heavy weight is placed on the most recent observation, and rapidly declining weights are placed on older values.

Another way of writing Equation 5.15 is

Ŷt+1 = Ŷt + w(Yt − Ŷt) [5.20]

This indicates that the new forecast is equal to the old forecast plus w times the error in the most recent forecast. A w that is close to 1 indicates a quick adjustment process for any error in the preceding forecast. Similarly, a w closer to 0 suggests a slow error cor- rection process.

It should be apparent from Equations 5.15 and 5.20 that exponential forecasting tech- niques can be very easy to use. All that is required is last period’s forecast, last period’s observation, plus a value for the weighting factor, w. The optimal weighting factor is normally determined by making successive forecasts using past data with various values of w and choosing the w that minimizes the RMSE given in Equation 5.1.

4More complex double exponential smoothing models generally give more satisfactory results than first-order exponential smoothing models when the data possess a linear trend over time. See Diebold, op. cit. 5The greater the amount of serial correlation (correlation of values from period to period), the larger will be the optimal value of w.

152 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Exponential Smoothing: Walker Corporation (continued) Consider again the Walker Corporation example discussed earlier. Suppose that the company is interested in generating sales forecasts using the first-order expo- nential smoothing technique. The results are shown in Table 5.5. To illustrate the approach, an exponential weight w of 0.5 will be used. To get the process started, one needs to make an initial forecast of the variable. This forecast might be a weighted average or some simple forecast, such as Equation 5.2:

Ŷt+1 = Yt

The latter approach will be used. Hence the forecast for Month 2 made in Month 1 would be $1,950(000) (Ŷt+1 = 1,950). The Month 3 forecast value is (using Equation 5.20)

Ŷ3 = 1,950 + 0:5ð1,400 − 1,950Þ = 1,950 − 275 = $1,675ð000Þ

Similarly, the Month 4 forecast equals

Ŷ4 = 1,675 + 0:5ð1,925 − 1,675Þ = $1,800ð000Þ

The remaining forecasts are calculated in a similar manner. As can be seen in Table 5.5, Walker’s sales forecast for January 2007 using the

first-order exponential smoothing technique is $2,322(000). Also, the root mean square error of this forecasting method (with w = 0.50) is $491(000).

TABLE 5.5 WALKER CORPORATION: FIRST-ORDER EXPONENTIAL

SMOOTHING SALES FORECAST

SALES ($1,000) ERROR

t MONTH ACTUAL Yt FORECAST Ŷ t (Yt − Ŷ t ) (Yt − Ŷ t ) 2

1 January 2006 1,950 — — —

2 February 1,400 1,950 −550 302,500

3 March 1,925 1,675 250 62,500

4 April 1,960 1,800 160 25,600

5 May 2,800 1,880 920 846,400

6 June 1,800 2,340 −540 291,600

7 July 1,600 2,070 −470 220,900

8 August 1,450 1,835 −385 148,225

9 September 2,000 1,642 358 128,164

10 October 2,250 1,821 429 184,041

11 November 1,950 2,036 −86 7,396

12 December 2,650 1,993 657 431,649

13 January 2007 — 2,322 — 000000—000000

Sum = 2,648,975

RMSE = ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2,648,975=11

p = $491ð000Þ

Chapter 5: Business and Economic Forecasting 153

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

BAROMETRIC TECHNIQUES The time-series forecasting models discussed earlier assume that future patterns in an economic time series may be predicted by projecting a repeat of past patterns, but very few economic time series exhibit consistent enough cyclical variations to make simple projection forecasting reliable. For example, Table 5.6 conveys why the prediction of a business cycle’s turning point proves to be so difficult. Although the duration of postwar U.S. business cycles averages 70 months (from peak to peak), three cycles have lasted 100 months or more, while others have been as short as 32 and even 18 months. Econ- omists, however, have long recognized that if it were possible to isolate sets of time series that exhibited a close correlation, and if one or more of these time series normally led (in a consistent manner) the time series in which the forecaster had interest, then this leading series could be used as a predictor or barometer.

Although the concept of leading or barometric forecasting is not new,6 current baro- metric forecasting is based largely on the work done at the National Bureau of Economic Research (http://www.nber.org). The barometric forecasting model developed there is used primarily to identify potential future changes in general business conditions, rather than conditions for a specific industry or firm.

Leading, Lagging, and Coincident Indicators Economic indicators may be classified as leading, coincident, or lagging indicators, de- pending on their timing relative to business cycle peaks and troughs (see Figure 5.7).

TABLE 5.6 DURATION OF U.S. BUSINESS CYCLES (IN MONTHS)

BUSINESS CYCLE‡

CONTRACTION* EXPANSION†

Oct 1945 Nov 1948 8 37 88 45

Oct 1949 July 1953 11 45 48 56

May 1954 Aug 1957 10 39 55 49

Apr 1958 Apr 1960 8 24 47 32

Feb 1961 Dec 1969 10 106 34 116

Nov 1970 Nov 1973 11 36 117 47

Mar 1975 Jan 1980 16 58 52 74

July 1980 July 1981 6 12 64 18

Nov 1982 July 1990 16 92 28 108

Mar 1991 Mar 2001 8 120 100 128

Nov 2001 Dec 2007 8 73 128 89

July 2009 19 92

Average post-war cycle 11 59 69 69

*Months from previous peak to trough.

†Months from trough to next peak.

‡Months from previous trough to next trough and months from previous peak to next peak. Source: “U.S. Business Cycle Expansions and Contractions,” National Bureau of Economic Research at www.nber.org.

6Andrew Carnegie used to count the number of smoke-belching chimneys in Pittsburgh to forecast the level of business activity and consequently the demand for steel.

154 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Conference Board, a private not-for-profit research institute in New York, has iden- tified 11 series that tend to lead the peaks and troughs of business cycles, 4 series of roughly coincident indicators of economic activity, and 7 series that tend to lag the peaks and troughs of economic activity. Table 5.7 lists the title of each series and the mean lead or lag of the series in relation to peaks and troughs of economic activity.

The rationale for the use of many of the series listed in this table is obvious. Many of these series represent commitments to future levels of economic activity. Building per- mits precede housing starts, and orders for durable goods precede their actual produc- tion. The value of any one of these indicators depends on the variability in the length of the lead (lag). Leading and lagging indicators predict the direction of future change in economic activity; they reveal little or nothing about the magnitude of the changes.

SURVEY AND OPINION-POLLING TECHNIQUES Survey and opinion-polling are other forecasting tools that may be helpful in making short-period forecasts. Business firms normally plan additions to plant and equipment well in advance of the actual expenditures; consumers plan expenditures for autos, vaca- tions, and education in advance of the actual purchase; and governments at all levels pre- pare budgets in advance of the spending.

The greatest value of survey and opinion-polling techniques is that they may help to uncover if consumer tastes are changing or if business executives begin to lose confi- dence in the economy; survey techniques may be able to uncover these trends before their impact is felt.

FIGURE 5.7 Barometric Indicators

Leading indicator

In di

ca to

r le

ve l (

va lu

e) Coincident indicator

Lagging indicator

Business cycle peak date

Business cycle trough date

Time

Chapter 5: Business and Economic Forecasting 155

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE 5.7 CYCLICAL LEADS (−) AND LAGS (+) FOR LEADING, COINCIDENT, AND LAGGING

INDICATORS (LENGTH IN MONTHS)

AT REFERENCE PEAKSa

S E R IE

S N O :

SERIES TITLE JU L Y 1 9 9 0

JU L Y 1 9 8 1

JA N : 1 9 8 0

N O V : 1 9 7 3

D E C : 1 9 6 9

A P R : 1 9 6 0

A U G : 1 9 5 7

JU L Y 1 9 5 3

N O V : 1 9 4 8

M E A N

LEADING INDICATORS

1 Average weekly hours, manufacturing −15 −7 −10 −7 −14 −11 −21 −3 −11 −11.0

5 Average weekly initial claims for unemployment insurance (inverted)b −22 0 −16 −9 −11 −12 −23 −10 −13 −12.9

8 Manufacturers’ new orders in 1987 dollars, con- sumer goods and materials −2 −2 −13 −8 −13 −13 −25 −3 −5 −9.3

32 Vendor performance, slower deliveries diffusion index +1 −3 −9 0 −4 −14 −28 −12 −7 −8.4

20 Contracts and orders for plant and equipment in 1987 dollars −7 −3 −10 −1 −11 −13 −9 −5 −7 −7.3

29 Building permits, new private housing units −21 −10 −19 −11 −10 −17 −30 −8 −13 −15.4

92 Change in manufacturers’ unfilled orders in 1987 dollars, durable goods (smoothed)c −3 −6 −13 −6 −7 −12 −19 −26 −3 −10.6

99 Change in sensitive materials prices (smoothed)c +2 −7 −7 +3 −10 −17 −17 −9 n.a. −7.8

19 Index of stock prices, 500 common stocks −1 −8 NST −10 −12 −9 −13 −6 −30 −11.1

106 Money supply M2 in 1987 dollars −7 NST −24 −10 −11 NST −16 NST −17 −14.2

83 Index of consumer expectations −18 −2 −38 −15 −10 −2 −9 −5 n.a. −12.4

910 Composite index of 11 leading indicators −18 −8 −15 −9 −11 −11 −20 −5 −7 −11.6

940 Ratio, coincident index to lagging index −4 −4 −15 −11 −9 −12 −27 −9 −10 −11.2

COINCIDENT INDICATORS

41 Employees on nonagricultural payrolls −1 0 +2 +11 +3 0 −5 −1 −2 +0.8

51 Personal income less transfer payments in 1987 dollars −3 +1 0 0 NST +1 0 −1 −1 −0.4

47 Index of industrial production +2 0 +2 0 −2 −3 −5 0 −4 −1.1

57 Manufacturing and trade sales in 1987 dollars −4 −6 −10 0 −2 −3 −6 −3 +1 −3.7

920 Composite index of 4 coincident indicators −1 +1 0 0 −2 −3 −5 0 −1 −1.2

LAGGING INDICATORS

91 Average duration of unemployment (inverted)a −13 +5 −6 −2 −2 +2 +1 +2 0 −1.4

77 Ratio, manufacturing, and trade inventories to sales in 1987 dollars +6 +15 +5 +16 +11 +9 +8 +5 +8 +9.2

62 Change in index of labor cost per unit of output, manufacturing (smoothed)c +8 +6 +5 +16 +1 +10 +6 +6 0 +6.4

109 Average prime rate charged by banks −14 +1 +3 +10 +2 +3 +4 +7 NST +2.0

101 Commercial and industrial loans outstanding in 1987 dollars 0 +14 +2 +10 +8 NST +1 −1 +3 +4.6

95 Ratio, consumer installment credit to personal income −10 NST −7 +5 NST +8 +5 +5 NST +1.0

120 Change in Consumer Price Index for services (smoothed)c +2 +2 +5 +11 +4 −6 −5 n.a. n.a. +1.9

930 Composite index of 7 lagging indicators −8 +3 +3 +13 +3 +3 +4 +5 NST +3.1

aReference peaks and troughs are the cyclical turning points in overall business activity; specific peaks and troughs are the cy- clical turning points in individual series. This table lists, for the composite indexes and their components, the leads (−) and lags (+) of the specific peaks and troughs in relation to the corresponding reference peaks and troughs. See National Bureau of Economic Research information on the selection of cyclical peaks and troughs available at http://www.nber.org. bThis series is inverted; i.e., low values are peaks and high values are troughs. cThis series is smoothed by an autoregressive-moving-average filter developed by Statistics Canada.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

AT REFERENCE TROUGHS*

S E R IE

S N O :

SERIES TITLE JU L Y 1 9 9 0

JU L Y 1 9 8 1

JA N : 1 9 8 0

N O V : 1 9 7 3

D E C : 1 9 6 9

A P R : 1 9 6 0

A U G : 1 9 5 7

JU L Y 1 9 5 3

N O V : 1 9 4 8

M E A N

LEADING INDICATORS

1 Average weekly hours, manufacturing +1 −1 0 0 −2 −2 0 −1 −6 −1.2

5 Average weekly initial claims for unemployment insurance (inverted)a 0 −2 −2 0 −1 0 0 +4 0 −0.1

8 Manufacturers’ new orders in 1987 dollars, con- sumer goods and materials 0 −1 −2 0 0 0 −2 −7 −4 −1.8

32 Vendor performance, slower deliveries diffusion index 0 −8 −2 −1 +1 −11 −4 −6 −7 −4.2

20 Contracts and orders for plant and equipment in 1987 dollars +3 +4 −2 +9 −1 +1 −1 −2 −6 +0.6

29 Building permits, new private housing units −2 −13 −3 0 −10 −2 −2 −8 −9 −5.4

92 Change in manufacturers’ unfilled orders in 1987 dollars, durable goods (smoothed)b +20 −2 −1 +1 −3 −9 −2 −5 −4 −0.6

99 Change in sensitive materials prices (smoothed)b 0 −5 0 −2 −2 −1 −4 −4 −4 −2.4

19 Index of stock prices, 500 common stocks −5 −4 NST −3 −5 −4 −4 −8 −4 −4.6

106 Money supply M2 in 1987 dollars −2 NST −2 −2 −7 NST −3 NST −15 −5.2

83 Index of consumer expectations −5 −8 −4 −1 −6 −3 +1 −6 n.a. −4.0

910 Composite index of 11 leading indicators −2 −10 −2 −1 −1 −2 −2 −4 −4 −3.1

940 Ratio, coincident index to lagging index −2 −10 −2 0 −8 −1 0 −5 0 −2.9

COINCIDENT INDICATORS

41 Employees on nonagricultural payrolls +11 0 0 +1 0 0 +1 +3 0 +1.8

51 Personal income less transfer payments in 1987 dollars +8 0 0 −1 NST −2 0 −1 −3 +0.1

47 Index of industrial production 0 +1 0 0 0 0 0 −1 0 0

57 Manufacturing and trade sales in 1987 dollars −2 +1 −1 0 0 −1 0 −5 −3 −1.2

920 Composite index of 4 coincident indicators 0 +1 0 0 0 0 0 +2 0 +0.3

LAGGING INDICATORS

91 Average duration of unemployment (inverted)a +19 +8 +6 +10 +19 +5 +6 +12 +8 +10.3

77 Ratio, manufacturing, and trade inventories to sales in 1987 dollars +36 +14 +6 +44 +27 +14 +13 +12 +9 +17.4

62 Change in index of labor cost per unit of output, manufacturing (smoothed)b +6 +10 +7 +8 +12 +7 +6 +11 +1 +9.7

109 Average prime rate charged by banks +35 +8 +1 +25 +16 +57 +4 +14 NST +17.9

101 Commercial and industrial loans outstanding in 1987 dollars +24 +11 +8 +18 +15 NST +4 +3 −1 +10.2

95 Ratio, consumer installment credit to personal income +21 0 NST +11 NST +9 +7 +6 NST +9.0

120 Change in Consumer Price Index for services (smoothed)b +18 +2 +3 +5 +27 +5 +8 n.a. n.a. +9.7

930 Composite index of 7 lagging indicators +36 +7 +3 +21 +15 +6 +4 +9 NST +9.3

Source: Business Cycle Indicators available from The Conference Board at www.conference-board.org.

TABLE 5.7 CYCLICAL LEADS (−) AND LAGS (+) FOR LEADING, COINCIDENT, AND LAGGING

INDICATORS (LENGTH IN MONTHS) (CONTINUED)

*n.a. Not available. Data needed to determine a specific turning point are not available.

NST No specific turn. No specific turning point is discernible in the data.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Forecasting Macroeconomic Activity Some of the best-known surveys available from private and governmental sources in- clude the following:

1. Plant and equipment expenditure plans—Surveys of business intentions regarding plant and equipment expenditures are conducted by McGraw-Hill, the National Industrial Conference Board, the U.S. Department of Commerce, Fortune magazine, the Securities and Exchange Commission, and a number of individual trade associa- tions. The McGraw-Hill survey, for example, is conducted twice yearly and covers all large corporations and many medium-sized firms. The survey reports plans for ex- penditures on fixed assets, as well as for expenditures on research and development. More than 50 percent of all new investment is accounted for by the McGraw-Hill survey.

The Department of Commerce–Bureau of Economic Analysis plant and equip- ment expenditures survey is conducted quarterly and published regularly in the Sur- vey of Current Business. The sample is larger and more comprehensive than that used by McGraw-Hill.

The National Industrial Conference Board surveys capital appropriations com- mitments made by the board of directors of 1,000 manufacturing firms. The survey picks up capital expenditure plans that are to be made sometime in the future and

Example Leading Indicators Change7

The Index of Leading Economic Indicators, which can be accessed at http://www .conference-board.org, is constantly under scrutiny by both private and public fore- casting agencies. When any series appears outdated or begins to generate mislead- ing signals, a replacement can often emerge from a consensus of best practices in business forecasting. Three of the series in Table 5.7 have been ranked “Poor” at predicting recessions and recoveries in the last decade by the Conference Board, a prominent trade association of major corporations that collects, analyzes, and dis- tributes business cycle data. Two of the three (i.e., manufacturers’ unfilled orders for durables and the change in sensitive materials prices) were removed from the Index and replaced by the interest rate spread between 10-year Treasury bond yields and 3-month Treasury bill yields.

The interest rate spread is an attempt to capture the effects of monetary policy on the business cycle. A long-bond yield at least 1.21 percent higher than the T-bill yield implies less than a .05 probability of recession four quarters ahead. If the Fed- eral Reserve tightens credit, however, such that short-term interest rates rise 0.82 percent above long-term rates, the probability of recession increases to 50 percent and more. At an interest rate spread of 2.40 percent, the probability of recession four quarters ahead rises to 90 percent. This new indicator of credit conditions should effectively supplement the generally poor third predictor, the M2 measure of the nation’s money supply. However, because oil price hikes have returned to a position of prominence in business planning, the sensitive materials price series may soon be restored to the Index. For descriptions of recent revisions in this in- dex, see http://www.tcb-indicators.org/.

7Based on “Makeup of Leading Indicators May Shift,” Wall Street Journal (August 11, 1996), p. A2.

158 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

for which funds have been appropriated. It is especially useful to firms that sell heavily to manufacturers and may aid in picking turning points in plant and equip- ment expenditures. This survey is published in the Survey of Current Business. You can access the Survey of Current Business on the Internet at http://www.bea.doc.gov/ bea/pubs.htm.

2. Plans for inventory changes and sales expectations—Business executives’ expectations about future sales and their intentions about changes in inventory levels are reported in surveys conducted by the U.S. Department of Commerce, McGraw-Hill, Dun and Bradstreet, and the National Association of Purchasing Agents. The National Asso- ciation of Purchasing Agents survey, for example, is conducted monthly, using a large sample of purchasing executives from a broad range of geographical locations and industrial activities in manufacturing firms.

3. Consumer expenditure plans—Consumer intentions to purchase specific products— including household appliances, automobiles, and homes—are reported by the Sur- vey Research Center at the University of Michigan (http://www.isr.umich.edu/src/) and by the Census Bureau. The Census Bureau survey, for example, is aimed at un- covering varying aspects of consumer expenditure plans, including income, holdings of liquid and nonliquid assets, the likelihood of making future durable goods pur- chases, and consumer indebtedness.

Sales Forecasting Opinion polling and survey techniques are also used on a micro level within the firm for forecasting sales. Some of the variations of opinion polling that are used include the following:

1. Sales force polling—Some firms survey their own salespeople in the field about their expectations for future sales by specific geographical area and product line. The idea is that the employees who are closest to the ultimate customers may have significant insights to the state of the future market.

2. Surveys of consumer intentions—Some firms (especially in durable goods industries) conduct their own surveys of specific consumer purchases. Consider an auto dealer who pursues a “customer for life” relationship with his or her target market. Such a dealer or a furniture company may conduct a mail survey to estimate target house- holds’ intentions of purchasing replacement autos or furniture.

ECONOMETRIC MODELS Another forecasting tool that is available to the managerial economist is econometric modeling. Econometrics is a combination of theory, statistical analysis, and mathematical model building to explain economic relationships. Econometric models may vary in their level of sophistication from the simple to the extremely complex. Econometric tech- niques for demand estimation were discussed in Chapter 4.

Advantages of Econometric Forecasting Techniques Forecasting models based on econometric methodology possess a number of significant advantages over time-series trend analysis, barometric models, and survey or opinion poll-based techniques. The most significant advantage is that they identify independent

Chapter 5: Business and Economic Forecasting 159

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

variables (such as price or advertising expenditures in a demand model) that the man- ager may be able to manipulate.

Another advantage of econometric models is that they predict not only the direction of change in an economic series, but also the magnitude of that change. This represents a substantial improvement over the trend projection models, which failed to identify turn- ing points, and the barometric models, which do not forecast the magnitudes of expected changes.

Single-Equation Models The simplest form of an econometric model is the single-equation model, as was devel- oped for explaining the demand for Sherwin-Williams house paint in Chapter 4. Once the parameters of the demand equation are estimated, the model can be used to make forecasts of demand for house paint in a given region.

Multi-Equation Models Although in many cases single-equation models may accurately specify the relationship that is being examined, frequently the interrelationships may be so complex that a

Example Single-Equation Forecasts: The Demand for Game-Day Attendance in the NFL8

Welki and Zlatoper report a model that explains the major determinants of the demand for game-day attendance at National Football League games. A forecasting model such as this might be used by a team to plan the most opportune times for special promotions and to predict demand for items sold at the stadium concession outlets. The following variables were used to estimate the model shown on the next page:

ATTENDANCE game attendance

PRICE average ticket price

INCOME real per capita income

COMPCOST price of parking at one game

HMTMRECORD season’s winning proportion of the home team prior to game day

VSTMRECORD season’s winning proportion of the visiting team prior to game day

GAME number of regular season games played by the home team

TEMP high temperature on game day

RAIN dummy variable 1 = rain, 0 = no rain

DOME dummy variable 1 = indoors, 0 = outdoors

DIVRIVAL dummy variable 1 = teams are in same division, 0 = teams are not in same division

CONRIVAL dummy variable 1 = conference game, 0 = nonconference game

NONSUNDAY dummy variable 1 = game day is not Sunday, 0 = game day is Sunday

SUNNIGHT dummy variable 1 = game moved to Sunday night for coverage on ESPN, 0 otherwise

BLACKOUT dummy variable = 1 if game is blacked out for local TV, 0 otherwise

(Continued)

160 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

system of several equations becomes necessary. This can be illustrated by examining a simple model of the national economy:

C = α1 + β1Y + e1 [5.21]

I = α2 + β2Pt−1 + e2 [5.22]

T = β3GDP + e3 [5.23]

GDP = C + I + G [5.24]

Y = GDP − T [5.25]

where C = consumption expenditures I = investment

Pt−1 = profits; lagged one period GDP = gross domestic product

T = taxes Y = national income G = government expenditures

INDEPENDENT VARIABLE

EXPECTED SIGN

ESTIMATED COEFFICIENT T -STATISTIC

INTERCEPT ? 98053.00 11.49

PRICE − −642.02 −3.08

INCOME ? −1.14 −3.12

COMPCOST − 574.94 1.34

HMTMRECORD + 16535.00 6.38

VSTMRECORD ? 2588.70 1.05

GAME ? −718.65 −3.64

TEMP ? −66.17 −1.27

RAIN − −2184.40 −1.23

DOME ? −3171.70 −1.66

DIVRIVAL + −1198.00 −0.70

CONRIVAL ? −1160.00 −0.58

NONSUNDAY + 4114.80 1.74

SUNNIGHT + 804.60 0.28

BLACKOUT − −5261.00 −3.15

These results indicate that weather conditions have little impact on the atten- dance at games. Fans appear to favor games played outdoors rather than in domed stadiums. Conference and divisional rivalries do not appear to impact demand greatly. Higher prices negatively impact attendance, but demand appears to be in- elastic at current price levels. The quality of the team, as measured by its winning percentage, has a significant positive impact on attendance. A model similar to this could be used as the basis for forecasting demand for any type of athletic event.

8A.M. Welki and T.J. Zlatoper, “U.S. Professional Football: The Demand for Game-Day Attendance in 1991,” Mana- gerial and Decision Economics (September/October 1994), pp. 489–495.

Chapter 5: Business and Economic Forecasting 161

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Equations 5.21, 5.22, and 5.23 are behavioral or structural equations, whereas Equa- tions 5.24 and 5.25 are identities or definitional equations. Once parameters of a sys- tem of equations have been estimated,9 forecasts may be generated by substituting known or estimated values for the independent variables into the system and solving for a forecast.

Complex Models of the U.S. Economy A number of complex multiple-equation econometric models of the U.S. economy have been developed and are used to forecast business activity. Information on three of these models and the forecasting techniques they employ is summarized in Table 5.8. As can be seen, some of the large econometric models still rely heavily on the judgment of their staffs of economists.

Consensus Forecasts: Blue Chip Forecaster Surveys The Federal Reserve Bank of Philadelphia (Livingston surveys) and Blue Chip Economic Indicators in Aspen, Colorado, conduct semiannual surveys of leading U.S. economists regarding their forecasts of unemployment, inflation, stock prices, and economic growth. The 50 to 60 economists who are regularly surveyed represent a cross section from large corporations and banks, labor unions, government, investment banking firms, and uni- versities. The Livingston and Blue Chip Forecaster surveys have been used by federal and

TABLE 5.8 CHARACTERISTICS OF THREE ECONOMETRIC MODELS OF THE

U.S. ECONOMY

MODEL

CHARACTERISTIC

WHARTON ECONOMETRIC FORECASTING ASSOCIATES

CHASE ECONOMETRIC ASSOCIATES

TOWNSEND- GREENSPAN

Approximate number of variables forecasted

10,000 700 800

Forecast horizon (quarters) 2 10–12 6–10

Frequency of model updates (times per year)

12 12 4

Date model forecast first regularly issued

1963 1970 1965

Forecasting techniques

(a) Econometric model 60% 70% 45%

(b) Judgment 30% 20% 45%

(c) Time-series methods — 5% —

(d) Current data analysis 10% 5% 10%

Source: S.K. McNees, “The Record of Thirteen Forecasters,” New England Economic Review, September– October 1981, pp. 5–21; and A. Bauer et al., “Transparency, Expectations, and Forecasts,” Federal Reserve Bank St. Louis Review, September/October, 2003, pp. 1–25.

9See Gujarti, op. cit.

162 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

state budget offices and many corporations to gauge the expectation of businesses regard- ing future economic growth and inflation.

As a broad-based consensus forecast, the Livingston and Blue Chip survey data tend to be more stable over time than any individual forecast. Referring again to Figure 5.5, there is evidence that economists have tended to underestimate both increases and decreases in the inflation rate.10 Figure 5.8 indicates the record of the Livingston and Blue Chip forecasts in predicting major expansions and recessions. As can be seen, econ- omists have tended to predict well the relatively moderate recessions and expansions but have not predicted sharp short recessions like those that occurred in 1974–1975, 2001, and the severe recession in 2008–2009.11

STOCHASTIC TIME-SERIES ANALYSIS Finally, consider two forecasting approaches that capitalize on the interdependencies in business data: stochastic time-series analysis and input-output analysis. Deterministic trend analysis, discussed earlier, was concerned with extrapolating deterministic past trends in the data (e.g., seasonal effects and population growth time trends). In contrast, stochastic time-series analysis attempts to remove deterministic time trends and instead model, estimate, and hopefully replicate the stochastic process generating any remaining patterns in the data across successive time periods—that is, any remaining autocorrela- tion patterns. Autocorrelation was discussed in Appendix 4A.

Consider a simple autoregressive first-order process with positive drift α,

yt = α+ βyt−1+ εt ε iid � Nð0, σ2εÞ [5.26]

where β = 1 by hypothesis and where et is a pure white noise disturbance drawn independently each period from a zero-mean, constant-variance normal probability

FIGURE 5.8 Livingston and Blue Chip Forecasts of GDP Growth

–16

4 0

–4 –8

–12

1953–54 1957–58 1960 1970 1974–75 1980–82 2001 2008–09

Note: Bars indicate actual values. Lines connecting dots reflect forecasts.

R ea

l G N

P gr

ow th

( pe

rc en

t) 16 12

8

10Based on H. Taylor, “The Livingston Surveys: A History of Hopes and Fears,” Business Review, Federal Reserve Bank of Philadelphia (May/June 1996), pp. 15–25. 11Based on K. Kliesen, “The 2001 Recession,” Federal Reserve Bank St. Louis Review (September/October 2003), pp. 23–38; and author updates.

Chapter 5: Business and Economic Forecasting 163

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

distribution (an iid, independent and identically distributed, disturbance). As illustrated in Figure 5.9 Panel (a), when α equals zero, such a series has no tendency to revert to any particular value (no “mean reversion”). In contrast, such series wander, and are inherently not forecastable, and therefore the last realization of yt is the best prediction of the next realization of the series. Similarly, when α is non-zero, the level of yt has no tendency to mean revert to any particular trend line; each innovation can result in a new trend line as illustrated in Figure 5.9 Panel (b). This is the famous “random walk” model applicable to the level of stock prices. Under the efficient markets hypothesis, a stock price like yt in Equation 5.26 is “fully informative” in the sense that it incorpo- rates all publicly available information that could possibly be useful in forecasting next period’s stock price. For business forecasters, the difficulty is that commodity prices, exchange rates, interest rates, and possibly other macroeconomic variables like real GDP and the overall price index may also exhibit these random walk properties.

Random walk variables pose several problems if one tries to make a forecast based on ordinary least-squares (OLS) regression analysis. For one thing, two random walk vari- ables with positive (negative) drift will almost certainly exhibit spurious correlation. Because each series is trending upward (downward) and not reverting to its mean, an OLS estimation on two variables generated by the process in Equation 5.26 will indicate a significant positive relationship between the variables when no causal link between the two exists. For example, even though real GDP and the overall price index of the econ- omy (the GDP price deflator) have random shocks that may be totally unrelated, and even though real growth and inflation may have unrelated structural determinants (e.g., population growth versus monetary expansions), the t score in a simple OLS regression of real GDP on the price index can easily be as high as 12.0 (i.e., 99 percent confidence in a positive relationship). This can be very misleading to the forecaster seek- ing leading indicators for use in a business plan; imagine selling your firm’s senior officers on the idea that because inflation is up, the firm can expect substantial real growth in demand next period. You might well be sent on an extended unpaid leave or even fired.

FIGURE 5.9 Random Walks Illustrated

Time

Panel (a)

Yt Yt

Time

Panel (b)

164 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Of course, not all business time series exhibit these random walk properties. For exam- ple, a firm’s profitability and earnings do mean revert in response to competitive entry and exit whenever they move substantially above or below the risk-adjusted mean for the industry, and therefore profits and earnings are not a random walk.12 Hence, it is crucial to know whether the data one is working with are, or are not, generated by a random walk process.

The second problem posed by random walks is that the level of yT after a number of periods T is

yT = y0 + ∑ T

i=0 ðα + εT−iÞ [5.27]

= y0 + Tα + Σεt [5.28]

the cumulative sum of the drift parameter plus all the white noise errors up to period T. Another way of describing this phenomenon is to say that all innovations to ran- dom walk variables result in permanent effects; the shocks just continue to cumulate and don’t wash out as the time series lengthens. Therefore, “trends” in business data have two meanings. Some trends are deterministic like the upward and downward sales trends for bathing suits in the spring and summer versus the fall and winter buy- ing seasons. Other trends, however, are stochastic; stochastic trends are the permanent effects of innovations in a random walk process like Equation 5.26. Since these ∑et do not cancel out, it is appropriate to think of them as trends too. The problem is that the variance of yT as the time series lengthens is equal to Tσ

2 e—that is, the variance of

the stock price or interest rate has no limit! This makes it quite difficult to reduce RMSE with the forecasting techniques we have seen so far. For example, even long lag structures in OLS regression models of stock price changes often have R2 as low as 0.02 to 0.05 and very large RMSEs. Again, such series have enormous variance as T grows.

Although many advanced techniques beyond the scope of this text are motivated by the random walk stochastic process,13 two simple methods we have already in- troduced at least partially address both of the previous complications. First, all ran- dom walk-like processes have very long, slowly decaying autocorrelation functions. The first-order autoregressive AR(1) random walk process with drift α in Equation 5.26 is said to be integrated of order one—written I(1)—because the coefficient on the first autoregressive lag is by hypothesis β = 1. Indeed, this particular first- order autocorrelation function never decays. Consequently, the Durbin-Watson sta- tistic introduced in Equation 4A.1 and Figure 4A.1 can be used to detect the pres- ence of severe autocorrelation in such variables. The DW statistic will definitely fall well below 2.0 for data generated by Equation 5.26—that is, below dL for positively autocorrelated series and above (4-dL) for negatively autocorrelated series generated by a process like Equation 5.26 with β = −1. So, one can use the DW statistic as a diagnostic instrument to detect the possibility of a non-mean-reverting, random walk process.

12See E. Fama and K. French, “Forecasting Profitability and Earnings,” Journal of Business (April 2000), pp. 164–175. 13One useful introduction to additional techniques in stochastic time-series analysis is F. Diebold, Elements of Forecasting, 4th ed. (Cincinnati: South-Western, 2007). For a more advanced treatment, consult W. Enders, Applied Econometric Time-Series, 3rd ed. (New York: John Wiley and Sons, 2009).

Chapter 5: Business and Economic Forecasting 165

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Moreover, the I(1) property of an AR(1) random walk implies that taking the first difference of the price or interest rate series in Equation 5.26,

Δyt = α + (β − 1)yt−1 + et [5.29]

would leave us with a process that did mean revert—that is, a process that reverted to the drift parameter α—if in fact β = 1. It is straightforward to estimate the first difference of the time series in Equation 5.29 or, more generally, to estimate a vector autoregression in first differences,

Δyt = α + ðβ − 1Þyt−1 + ∑ ∞

t=1 Δyt−1 + εt [5.30]

and conclude whether the null hypothesis β = 1 is true or false.14 If true, any series with these properties should be differenced and incorporated into the forecasting regressions as first differences, not as levels.15 If the situation as described in Equations 5.27, 5.28, and 5.29 pertains to both the dependent and an explanatory variable, the entire forecast- ing model should be specified in first differences. In that case, these two series are said to be cointegrated and will exhibit a nonspurious co-movement with one another, the pres- ence of which could prove quite important to achieving the standard forecasting objec- tive of low RMSE.

FORECASTING WITH INPUT-OUTPUT TABLES Another forecasting approach that capitalizes on the cross-sectional interdependence be- tween various intermediate product and final product industries is input-output analysis. Input-output analysis enables the forecaster to trace the effects of an increase in demand for one product to other industries. An increase in the demand for automobiles will first lead to an increase in the output of the auto industry. This in turn will lead to an in- crease in the demand for steel, glass, plastics, tires, and upholstery fabric. In addition, secondary impacts will occur as the increase in the demand for upholstery fabric, for ex- ample, requires an increase in the production of fibers used to make the fabric. The de- mand for machinery may also increase as a result of the fabric demand, and so the pattern continues. Input-output analysis permits the forecaster to trace through all these inter-industry effects that occur as a result of the initial increase in the demand for auto- mobiles. The Bureau of Economic Analysis of the U.S. Department of Commerce pro- duces a complicated set of tables specifying the interdependence among the various industries in the economy.16

14These tests can be performed with the t-statistic on the (β − 1) parameter on yt−1 but require using a modi- fied set of Dickey-Fuller critical values. See Appendix B, Table 7. 15If β = 1 is rejected, each series in question should be differenced again, and the second differences tested in exactly the same fashion. If in that case β = 1, second differences would be incorporated rather than first dif- ferences. If neither first nor second differences indicate an I(1) or I(2) series, the forecaster proceeds using the levels of the original data.

cointegrated Stochastic series with a common order of integration and exhibiting an equilibrium relationship such that they do not permanently wander away from one another.

16The most recent input-output tables may be found for 16 industry groups and 432 detailed industries in “Input-Output Tables” on the U.S. Bureau of Economic Analysis Web site at www.bea.gov.

166 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

SUMMARY

� A forecast is a prediction concerning the future value of some economic time-series.

� The choice of a forecasting technique depends on the cost of developing the forecasting model, the complexity of the relationship being forecast, the time period for the forecast, the accuracy required of the model, and the lead time required to make decisions based on the forecasting model.

� Data used in forecasting may be in the form of a time series—that is, a series of observations of a variable over a number of past time periods—or

they may be cross-sectional—that is, observations are taken at a single point in time for a sample of individuals, firms, geographic regions, communi- ties, or some other set of observable units.

� Deterministic trend-forecasting models are based on an extrapolation of past values into the future. Time-series forecasting models may be adjusted for seasonal, secular, and cyclical trends in the data. Stochastic time-series forecasting models investi- gate the randomness-generating process in the un- derlying data.

INTERNATIONAL PERSPECTIVES

Long-Term Sales Forecasting by General Motors in Overseas Markets

General Motors has an extensive forecasting system for both its North American and its overseas opera- tions that is implemented by the Corporate Product Planning and Economics Staff. The process generates short- and long-term forecasts of the U.S. vehicle market and long-term forecasts for overseas markets. A discussion of the overseas forecasting process follows.

General Motors produces forecasts for motor ve- hicle sales in nearly 60 countries. These countries vary in the number of cars per 1,000 people (car den- sity), from less than 10 to over 500. The primary factor used to explain the growth in car density is the level of and changes in income in each country. In the first step of the forecasting process, the mac- roeconomic relationship between key economic vari- ables, including income levels and motor vehicle sales, is estimated. Specifically, estimates are made of the income elasticity of demand in each country. The second step attempts to monitor changes over time in the relationships established in Step 1.

The third step consists of consultations between the Product Planning and Economics Staff and the Marketing Staff of each GM overseas operation. The objective of this phase is to identify any special fac-

tors in each country that might require a significant modification in the forecasts generated from the econometric models. For example, in the early and mid-1980s, it was felt that certain voluntary restraint policies that had been adopted by the Japanese gov- ernment would hold down demand by up to 50 per- cent, relative to the forecasts from the econometric model. When these policy barriers were removed, Japanese car sales skyrocketed up to levels predicted by the econometric model. More recently, the entry of China into the World Trade Organization has caused a shift upward in the future projected net ex- ports (exports minus imports) of cars to China.

The final step provides models of alternative fu- ture scenarios that reflect the impact of major changes in the economic environment for which full information is unavailable. For example, GM de- veloped a scenario plan for the opening of the Chi- nese market where the Buick is a very successful luxury brand; it predicted more sales by 2010 in the fast-growing markets of China and India than in all other foreign locations combined. To derive cash flow forecasts from foreign sales, GM must model (and manage) its exchange rate risk exposure, which is discussed in the next chapter on managing exports.

Chapter 5: Business and Economic Forecasting 167

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

� When a data series possesses a great deal of ran- domness, smoothing techniques, such as moving averages and exponential smoothing, may improve the forecast accuracy.

� Neither trend analysis models nor smoothing techni- ques are capable of identifying major future changes in the direction of an economic data series.

� Barometric techniques, which employ leading, lag- ging, and coincident indicators, are designed to forecast changes in the direction of a data series but are poorly suited for forecasting the magnitude of the change.

� Survey and opinion-polling techniques are often useful in forecasting such variables as business cap- ital spending and major consumer expenditure

plans and for generating product-specific or re- gional sales forecasts for a firm.

� Econometric methods seek to explain the reasons for a change in an economic data series and to use this quantitative, explanatory model to make future fore- casts. Econometric models are one of the most useful business forecasting tools, but they tend to be expen- sive to develop and maintain. Their net benefit hinges on their success in reducing root mean forecast error in out-of-sample forecasting environments.

� Trends in business data are either deterministic or stochastic. Stochastic trends introduced by random walk variables like stock prices require careful di- agnosis and special methods.

Exercises 1. The forecasting staff for the Prizer Corporation has developed a model to predict sales of its air-cushioned-ride snowmobiles. The model specifies that sales S vary jointly with disposable personal income Y and the population between ages 15 and 40, Z, and inversely with the price of the snowmobiles P. Based on past data, the best estimate of this relationship is

S = k YZ P

where k has been estimated (with past data) to equal 100.

a. If Y = $11,000, Z = $1,200, and P = $20,000, what value would you predict for S?

b. What happens if P is reduced to $17,500? c. How would you go about developing a value for k? d. What are the potential weaknesses of this model?

2. a. Fred’s Hardware and Hobby House expects its sales to increase at a constant rate of 8 percent per year over the next three years. Current sales are $100,000. Forecast sales for each of the next three years.

b. If sales in 2003 were $60,000 and they grew to $100,000 by 2007 (a four-year period), what was the actual annual compound growth rate?

c. What are some of the hazards of employing a constant rate of growth fore- casting model?

3. Metropolitan Hospital has estimated its average monthly bed needs as

N = 1,000 + 9X

where X = time period ðmonthsÞ; January 2002 = 0 N = monthly bed needs

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

168 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Assume that no new hospital additions are expected in the area in the foreseeable future. The following monthly seasonal adjustment factors have been estimated, using data from the past five years:

MONTH ADJUSTMENT FACTOR

January +5%

April −15%

July +4%

November −5%

December −25%

a. Forecast Metropolitan’s bed demand for January, April, July, November, and December 2007.

b. If the following actual and forecast values for June bed demands have been recorded, what seasonal adjustment factor would you recommend be used in making future June forecasts?

YEAR FORECAST ACTUAL

2007 1,045 1,096

2006 937 993

2005 829 897

2004 721 751

2003 613 628

2002 505 560

4. Stowe Automotive is considering an offer from Indula to build a plant making automotive parts for use in that country. In preparation for a final decision, Stowe’s economists have been hard at work constructing a basic econometric model for Indula to aid the company in predicting future levels of economic ac- tivity. Because of the cyclical nature of the automotive parts industry, forecasts of future economic activity are quite important in Stowe’s decision process.

Corporate profits (Pt−1) for all firms in Indula were about $100 billion. GDP for the nation is composed of consumption C, investment I, and government spending G. It is anticipated that Indula’s federal, state, and local governments will spend in the range of $200 billion next year. On the basis of an analysis of recent economic activity in Indula, consumption expenditures are assumed to be $100 billion plus 80 percent of national income. National income is equal to GDP minus taxes T. Taxes are estimated to be at a rate of about 30 percent of GDP. Finally, corporate investments have historically equaled $30 billion plus 90 percent of last year’s corporate profits (Pt−1).

a. Construct a five-equation econometric model of the state of Indula. There will be a consumption equation, an investment equation, a tax receipt equation, an equation representing the GDP identity, and a national income equation.

b. Assuming that all random disturbances average to zero, solve the system of equations to arrive at next year’s forecast values for C, I, T, GDP, and Y. (Hint: It is easiest to start by solving the investment equation and then working through the appropriate substitutions in the other equations.)

Chapter 5: Business and Economic Forecasting 169

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

5. A firm experienced the demand shown in the following table.

YEAR ACTUAL DEMAND

5-YEAR MOVING AVERAGE

3-YEAR MOVING AVERAGE

EXPONENTIAL SMOOTHING

(W = 0.9)

EXPONENTIAL SMOOTHING

(W = 0.3)

2000 800 xxxxx xxxxx xxxxx xxxxx

2001 925 xxxxx xxxxx — —

2002 900 xxxxx xxxxx — —

2003 1025 xxxxx — — —

2004 1150 xxxxx — — —

2005 1160 — — — —

2006 1200 — — — —

2007 1150 — — — —

2008 1270 — — — —

2009 1290 — — — —

2010 * — — — —

*Unknown future value to be forecast.

a. Fill in the table by preparing forecasts based on a five-year moving average, a three-year moving average, and exponential smoothing (with a w = 0.9 and a w = 0.3). Note: The exponential smoothing forecasts may be begun by assuming Ŷt+1 = Yt.

b. Using the forecasts from 2005 through 2009, compare the accuracy of each of the forecasting methods based on the RMSE criterion.

c. Which forecast would you have used for 2010? Why?

6. The economic analysis division of Mapco Enterprises has estimated the demand function for its line of weed trimmers as

QD = 18,000 + 0.4N − 350PM + 90PS

where N = number of new homes completed in the primary market area PM = price of the Mapco trimmer PS = price of its competitor’s Surefire trimmer

In 2010, 15,000 new homes are expected to be completed in the primary market area. Mapco plans to charge $50 for its trimmer. The Surefire trimmer is expected to sell for $55.

a. What sales are forecast for 2010 under these conditions? b. If its competitor cuts the price of the Surefire trimmer to $50, what effect will

this have on Mapco’s sales? c. What effect would a 30 percent reduction in the number of new homes

completed have on Mapco’s sales (ignore the impact of the price cut of the Surefire trimmer)?

170 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

7. The Questor Corporation has experienced the following sales pattern over a 10-year period:

YEAR SALES ($000)

1997 121

1998 130

1999 145

2000 160

2001 155

2002 179

2003 215

2004 208

2005 235

2006 262

2007 *

*Unknown future value to be forecast.

a. Compute the equation of a trend line (similar to Equation 5.4) for these sales data to forecast sales for the next year. (Let 1997 = 0, 1998 = 1, etc., for the time variable.) What does this equation forecast for sales in the year 2007?

b. Use a first-order exponential smoothing model with a w of 0.9 to forecast sales for the year 2007.

8. Bell Greenhouses has estimated its monthly demand for potting soil to be the following:

N = 400 + 4X

where N = monthly demand for bags of potting soil X = time periods in months ðMarch 2006 = 0Þ

Assume this trend factor is expected to remain stable in the foreseeable future. The following table contains the monthly seasonal adjustment factors, which have been estimated using actual sales data from the past five years:

MONTH ADJUSTMENT FACTOR

March +2%

June +15%

August +10%

December −12%

a. Forecast Bell Greenhouses’ demand for potting soil in March, June, August, and December 2007.

b. If the following table shows the forecasted and actual potting soil sales by Bell Greenhouses for April in five different years, determine the seasonal adjustment factor to be used in making an April 2008 forecast.

YEAR FORECAST ACTUAL

2007 500 515

2006 452 438

2005 404 420

2004 356 380

2003 308 320

Chapter 5: Business and Economic Forecasting 171

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

9. Savings-Mart (a chain of discount department stores) sells patio and lawn furni- ture. Sales are seasonal, with higher sales during the spring and summer quarters and lower sales during the fall and winter quarters. The company developed the following quarterly sales forecasting model:

Ŷt = 8.25 + 0.125t − 2.75D1t + 0.25D2t + 3.50D3t

where Ŷ t = predicted sales ð$millionÞ in quarter t 8:25 = quarterly sales ð$millionÞ when t = 0

t = time period ðquarterÞ where the fourth quarter of 2002 = 0, first quarter of 2003 = 1, second quarter of 2003 = 2, . . .

D1t = 1 for first-quarter observations

0 otherwise

D2t = 1 for second-quarter observations

0 otherwise

D3t = 1 for third-quarter observations

0 otherwise

Forecast Savings-Mart’s sales of patio and lawn furniture for each quarter of 2010.

10. Use the monthly series on the Consumer Price Index (all items) from the previous two years to produce a forecast of the CPI for each of the next three years. Is the precision of your forecast greater or less at 36 months ahead than at 12 months ahead? Why? Compare your answer to that of Moody’s on-line U.S. Macro Model at http://www.economy.com/.

Case Exercises CRUISE SHIP ARRIVALS IN ALASKA

The summer months bring warm weather, mega fauna (bears), and tourists to the coastal towns of Alaska. Skagway at the top of the Inland Passage was, in the nine- teenth century, the entrance to the Yukon. Today this town attracts multiple cruise ships per day; literally thousands of passengers disembark into a town of 800 for a taste of the Alaskan frontier experience between 10 a.m. and 5 p.m. Some ride steam trains into the mountains while others wander the town spending money in galleries, restaurants, and souvenir shops. The Skagway Chamber of Commerce is trying to de- cide which transportation mode in the table of visitor arrival statistics should receive the highest priority in the tourist promotions for next season.

Questions 1. Plot the raw data on arrivals for each transportation mode against time, all on the

same graph. Which mode is growing the fastest? Which the slowest? 2. Plot the logarithm of arrivals for each transportation mode against time, all on

the same graph. Which now appears to be growing the fastest? Logarithms are especially useful for comparing series with two divergent scales

since 10 percent growth always looks the same, regardless of the starting level.

n n n

172 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

When absolute levels matter, the raw data are more appropriate, but when growth rates are what’s important, log scales are better.

3. Now create an index number to represent the growth of arrivals in each transporta- tion mode by dividing the first (smallest) number in each column into the remain- ing numbers in the column. Plot these index numbers for each transportation mode against time, all in the same graph. Which is growing the fastest?

4. In attempting to formulate a model of the passenger arrival data on cruise ships over time, would a nonlinear (perhaps a multiplicative exponential) model be preferable to a linear model of cruise ship arrivals against time? What about in the case of the passenger arrivals by ferry against time?

5. Estimate the double-log (log linear) time trend model for log cruise ship arrivals against log time. Estimate a linear time trend model of cruise ship arrivals against time. Calculate the root mean square error between the predicted and actual value of cruise ship arrivals. Is the root mean square error greater for the double log time trend model or for the linear time trend model?

SKAGWAY VISITOR ARRIVAL STATISTICS

YEAR CRUISE FERRY HIGHWAY AIR

1983 48,066 25,288 72,384 3,500

1984 54,907 25,196 79,215 3,750

1985 77,623 31,522 89,542 4,000

1986 100,695 30,981 91,908 4,250

1987 119,279 30,905 70,993 4,953

1988 115,505 31,481 74,614 5,957

1989 112,692 29,997 63,789 7,233

1990 136,512 33,234 63,237 4,799

1991 141,284 33,630 64,610 4,853

1992 145,973 37,216 79,946 7,947

1993 192,549 33,650 80,709 10,092

1994 204,387 34,270 81,172 10,000

1995 256,788 33,961 87,977 17,000

1996 299,651 35,760 86,536 20,721

1997 438,305 27,659 91,849 11,466

1998 494,961 31,324 100,784 20,679

1999 525,507 31,467 92,291 15,963

Data are available as an Excel file on the book’s Web site.

Source: The Skagway News, November 16, 1999.

LUMBER PRICE FORECAST Questions 1. One of the most important variables that must be forecasted accurately to project

the cost of single-family home construction is the price of Southern pine framing lumber. Use the following data to forecast two- and four-year-ahead lumber prices. Compare the forecast accuracy of at least two alternative forecasting methods.

Chapter 5: Business and Economic Forecasting 173

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

LUMBER PRICE INDEX

2007 205.1 1985 100.0

2006 192.5 1984 101.3

2005 173.2 1983 101.2

2004 160.0 1982 93.8

2003 137.1 1981 96.4

2002 134.5 1980 95.2

2001 140.2 1979 99.0

2000 146.4 1978 90.9

1999 176.3 1977 77.9

1998 168.4 1976 67.7

1997 182.7 1975 58.3

1996 168.7 1974 60.5

1995 162.7 1973 58.3

1994 168.9 1972 47.6

1993 163.2 1971 41.9

1992 137.5 1970 37.4

1991 123.9 1969 41.3

1990 121.7 1968 37.3

1989 118.9 1967 32.9

1988 111.5 1966 33.0

1987 105.8 1965 31.6

1986 100.7 1964 31.4

Data are available as an Excel file on the book’s Web site.

Source: Forest Product Market Prices and Statistics, Annual Yearbooks

(Eugene, OR: Randon Length Productions), various issues.

174 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

6 CHAP T E R

Managing in theGlobal Economy CHAPTER PREVIEW Business plans today involve foreign supply chains, offshore manufacturing, and targeted marketing on several continents. Many U.S., German, Japanese, Taiwanese, and Korean companies engage in foreign direct investment and manufacture through subsidiaries abroad. Other companies outsource to low-wage high-quality contract manufacturing partners in places like China, Mexico, Portugal, Brazil, Indonesia, and the Caribbean. Still others buy parts and supplies or preassembled components from foreign firms. And almost all companies face competition from imports and produce an export product to sell abroad. Indeed, export markets are increasingly the primary source of sales growth for many U.S. manufacturers. The United States is the world’s largest import-export trader. Germany and China represent the second and third largest shares of world trade. Careful analysis and accurate forecasting of these international purchases and sales provide pivotal information for capacity planning, for production scheduling, and for pricing, promotion, and distribution plans in many companies.

In this chapter, we investigate how international trade in merchandise and services plus international capital flows determine long-term trends in exchange rates, which we analyze with standard demand and supply tools in the market for U.S. dollars as foreign exchange (FX). Purchasing power parity provides a way to assess these FX trends and incorporate cash flow from net export sales in business scenario planning. We then explore the reasons for and patterns of trade in the world’s economy with special attention to regional trading blocs and emergent economies, like the European Union, NAFTA, and China. The chapter closes with perspectives on the U.S. trade deficit. Our attention throughout is focused on a management approach to international trade and policy.

MANAGERIAL CHALLENGE Financial Crisis Crushes U.S. Household Consumption and Business Investment: Will Exports to China Provide the Way Out?

The gross domestic product (GDP) in the United States, one measure of aggregate demand, is a little over $14 trillion. To understand this figure, one can look from three perspectives: (1) at the comparative size of other big economies like China, Japan, and Germany; (2) at the relative size of the components of U.S. GDP; and (3) at the sheer size of a trillion. The number one trillion

(1,000,000,000,000) is 500 times larger than the cash position of a large multinational company such as Pep- siCo ($2 billion). It is 1,000 times larger than the one billion people in Mainland China, and 4,000 times larger than the annual budget of a typical college ($250 million). So, 14 trillion is a very large number indeed, and the United States is the world’s largest economy.

175

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Japan has the second highest GDP in the world at ¥455 trillion (T) Japanese yen (JYN) or about $5T U.S. dollars (USD) using the market exchange rate of JPY95/USD in late 2009.China,Germany, andFrance are theworld’s third, fourth, and fifth largest economies at $4.4T, $3.7T, and $2.8T, respectively. The Chinese case is especially interest- ing because the Chinese currency does not trade on freely fluctuating currency markets. Instead, the Chinese govern- ment has insisted on an officially sanctioned rate of ex- change (CNY6.875/USD) that is allowed to vary over a half of one percent range—a so-called “managed float.”

It is estimated by the International Monetary Fund (IMF) that if the Chinese allowed their currency to float, it would command approximately 45 percent higher value, and then only 3.875 Chinese yuan would ex- change for one U.S. dollar. Using this higher value of the CNY, China would be ranked as $7.8T in size, by far the second largest in the world. The Chinese economy has already been widely recognized as fastest growing throughout the past decade (12–15 percent annually).

How the $14 trillion U.S. GDP divides among its var- ious components of C + I + G + Net X is also insightful for thinking about managing exports. Consumption (C) is by far the largest component of U.S. GDP, accounting for about 10T of the 14T in recent years:

GDP = ð C + I + G Þ + Net X 2008 ð10:2T + 2:0T + 2:9TÞ + ð−0:7TÞ

Exports − Imports ð2T − 2:7TÞ

= $15:1T + ð−$700BÞ = $14:4 trillion

2009 ð9:98T + 1:6T + 2:9TÞ + ð−0:37TÞ ð2:03T − 2:4TÞ

= $14:5T + ð−$370BÞ = $14:1 trillion

These U.S. proportions of GDP (71 percent Con- sumption, 11 percent Investment, 21 percent Govern- ment, 10 percent exports, and 13 percent imports) would not characterize the fast-developing economies like China and India nor the export-driven economies like Korea, Japan, and Holland where investment (I) and the import-export sector are much larger. But at 23 percent of GDP, imports plus exports do employ lots of Americans, and this fact was crucial in escaping the 2008–2009 severe recession. The U.S. Federal Re- serve kept real interest rates close to zero throughout 2009 not only to stimulate business borrowing and in- vestment but also to keep the U.S. dollar’s value low, thereby stimulating U.S. exports.

In recent years, export growth contributed the ma- jority of real U.S. GDP growth: +1.5 percent in 2009 from $2 trillion to $2.03 trillion (see above equations) when real GDP declined by −2 percent from $14.4 to $14.1 trillion. Export growth contributed +0.88 percent of the +1.07 percent growth in real GDP in 2008, and +2 percent of the 0.95 percent growth in 2007. As a result, export growth offered one of the only escape routes from the severe 2008–2009 recession. Remember that fully 71 percent of U.S. GDP (domestic consump- tion C) had gone down by $220B, investment had col- lapsed by $400B, and fixing everything imaginable in the private investment sector (I) would have addressed only 11 percent of the U.S. economy. As a result, in the near term, export growth along with deficit spending through aggressive government fiscal policy (G) was chosen by the Obama Administration to stem job losses and jumpstart a recovery.

What were the symptoms of recession in the United States in 2009 and what challenges did managers face as a result? First, U.S. GDP contracted for four quarters in a row [−2.7 percent in 2008 (Q3), −5.4 percent in 2008 (Q4), −6.7 percent in 2009 (Q1), and −1 percent in 2009 (Q2)]. That has never happened since these national income statistics started being collected at the end of World War II. So, the downturn was severe and persistent. Secondly, unemployment skyrocketed above 10 percent by 2009 (Q3) when 4 to 5 percent is “natural” in a fully employed U.S. economy. That too has happened only rarely, just once since 1947 (at the depths of the 1982 recession). Finally, industrial produc- tion declined for 17 of 18 months. These business activity declines were not projected by most business planning that companies had undertaken in 2006, 2007, and 2008.

Cont.

MANAGERIAL CHALLENGE Continued

© Im ag in ec hi na

vi a AP

Im ag es

176 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

A first challenge was to understand the sources of these downturns in consumption, investment, and in- dustrial production. Thereafter, managers had to begin to consider how to respond to the timing of and pro- spects for recovery. Consumption has collapsed because the majority of American households are now stock- holders, and many have retirement account investments with U.S.-based mutual funds. American mutual fund assets lost $2.4 trillion in value during 2008. So, U.S. households suffered a massive reduction in wealth equivalent to the cash position of 1,200 PepsiCo cor- porations. Not surprisingly, durable goods consumption (autos, appliances) was again down by −7 percent in 2009 (Q2) after plummeting −40 percent in 2008. Non-durable consumption also declined −3 percent af- ter falling by −11 percent in 2008.

Investment collapsed by another −20 percent in mid- 2009 after declining by −51 percent in 2009 (Q1) and −41 percent in 2008. Inventory purchases particularly were slashed. Why? Three reasons seem apparent in the National Trends economic data from the Federal Reserve Bank St Louis: (1) demand traffic has declined to a trickle in many durables such as home appliances, suggesting that anticipated unit sales should be fore- casted at very low levels in the near term, (2) energy costs spiked in mid-2008, raising the producer price

index by 10 percent, (3) export demand from some of the United States’ largest trading partners (Japan and the EU) declined by half, reflecting a worldwide reces- sion that measured out to −2 percent of world GDP. The only bright spot in mid-2009 was China, where rapid growth continued.

As to the timing of recovery, J.P. Morgan’s global purchasing managers’ index swung sharply upward in the first quarter of 2009. After expanding in 2002– 2007 and contracting throughout 2008, the purchasing spending plans moved back to neutral. The extraordi- narily tight inventory position of many companies was surely responsible, but to whom could these supply chain officers have been planning to sell? The answer is China. Chinese GNP grew in 2009 at 9 percent, and Chinese retail sales grew at 17 percent when much of the rest of the world was slowing to a near stop.

In an important sense, China is not simply an export machine. Investment is 41 percent of GDP and con- sumption is just 36 percent, reflecting much infrastruc- ture building. In 2008–2010, steel use in China totaled 440 million, 515 million, and 540 million metric tons—40 percent larger than the tonnage of India, the United States, and the EU combined. Household de- mand is just starting to accelerate as China changes from developing to developed-country living standards.

U.S. EXPORTS TO CHINA

China’s WTO entry

$0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008

$10

$20

$30

$40

$50

$60

$70

U .S

. d ol

la rs

( bi

lli on

s)

Steel use in selected areas, in millions of metric tons

Source: World Steel Assn.

600

500

400

300

200

100

0 India U.S. EU China

’08 ’09 ’10

’08

’09 ’10

Source: World Steel Association.

Cont.

MANAGERIAL CHALLENGE Continued

Chapter 6: Managing in the Global Economy 177

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

INTRODUCTION Around the globe, the reduction of trade barriers and the opening of markets to foreign imports have increased the competitive pressure on manufacturers who once dominated their domestic industries. Tennis shoes and dress footwear once produced in large facto- ries in Britain and the United States now come from Korea, China, and Italy. Auto and auto parts manufacturing once dominated by Ford, General Motors, and American parts suppliers like Magna now occurs largely in Japan and China. In addition, Toyota owns 13 assembly and parts plants in North America, 4 in South America, 6 in Europe, 4 in India and Pakistan, and 25 across Asia. In retailing, McDonald’s operates in over

As a result, U.S. exports to China have continued to grow right through the severe worldwide recession of 2008–2009.

The ways out of the current recession are therefore probably at this juncture limited to expansionary fiscal policy and an expansion of exports. Massive deficit spending by federal governments on infrastructure and productivity-enhancing training and facilities is more attractive than several other alternatives. Federal tax re- bates are unlikely to shake households free of their new- found fascination with savings and reduced consump- tion. In 2007, federal tax rebates led households to pay down credit card debt rather than go out and replace an appliance that was wearing out. Ironically, credit card companies then flagged those households as greater credit risks and promptly cut their credit limits. Such a sequence of events of course increased the households’ desire to save for security since they could no longer expect to live on credit cards should their breadwinner suffer a layoff. John Maynard Keynes described this sit- uation as a classic liquidity trap where consumption spiraled downward despite tax policy designed to accen- tuate household liquidity.

Monetizing federal deficits by printing money is al- most always a mistake because of the risk that ap- proach poses for renewed inflation if industrial production does not quickly increase. As a result, in- dependent central bankers in Europe, Japan, and to a lesser extent the United States will resist any attempt to monetize these swollen federal deficits. More tradi- tional monetary policy of purchasing short-term fed- eral debt (e.g., T-bills) and replacing those assets circulating in the capital markets with an infusion of cash in the economy will normally (if done carefully and in moderation) not trigger inflation but will lower short-term interest rates. Since lower borrowing costs enhance the ability of businesses to fund their working

capital requirements, expansionary monetary policy can in principle be transmitted into increased real ac- tivity in the economy. Unfortunately for the United States and Japan, short-term interest rates are already near zero (0.1 to 0.4 percent), so traditional monetary policy really has no room left to stimulate these two largest economies.

Ingenious credit policy initiatives have been tried but are proving largely unsuccessful in breaking loose more bank lending. Bankers have proven stubborn about ex- tending credit to small and medium-sized businesses that survive on such loans from one stage of the business cycle or of seasonal sales to the next. Having been burned by overextending credit 2004–2007, bankers and bank regulators are overreacting by under-lending. As a result, business credit conditions have tightened precisely when companies need to begin to build inven- tories to trigger new sales from returning customers.

In this scenario, only deficit spending on G and greater exports especially to India and China hold the prospect of restarting an economic expansion. Compa- nies should poise themselves to take advantage of the extraordinary federal spending implied and to develop further their export sales to China.

Discussion Questions

� Which part of the global economy is growing? How do $1.47 trillion German exports in 2008, $1.43 trillion Chinese exports, and $1.48 tril- lion U.S. exports compare to the sizes of GDP in those countries?

� Why is the export sector so crucial to the re- covery from the 2008–2009 severe recession?

� What U.S. companies are likely to profit first from increased exports?

MANAGERIAL CHALLENGE Continued

178 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

100 foreign countries, Walmart sells $200 billion overseas, and international sales at Coca-Cola, Kellogg, Gillette, and Pampers exceed those in the United States. Boeing and Microsoft are the two largest U.S. exporters but other key U.S. exports include mo- vies, electrical equipment, heavy machinery, accounting and consulting services, and franchise retailers.

Similarly, outsourcing numerous components and subassemblies to foreign companies has become a standard supply chain management practice for U.S. manufacturers. Every three days, the wings for a new long-haul 300-seat 787 arrive from Japan at Everett, Washington, near Seattle for final assembly by Boeing. For its minivans, Chrysler may decide to cast engine blocks in Mexico, acquire electronics from Taiwan, perform machine tooling of ball bearings in Germany, and locate final assembly in Canada. Wooden furniture “made in the U.S.A.” now includes foreign components from Canada, Mexico, and the Far East equal to 38 percent of the total value added. So, when 32 per- cent of Canadian output and 25 percent of Mexican or Malaysian output flows into the United States as imports, much of that final product flow contains preassembled compo- nents owned by U.S. companies. This explains in large measure why fully one-third of U.S. corporate profits come from overseas operations. The world of business has truly become a matter of managing in the global economy.

IMPORT-EXPORT SALES AND EXCHANGE RATES As illustrated in the previous example, import/export sales and profit margins are very sensitive to changes in exchange rates. Over the period 1980–2010, foreign exchange (FX) rates were four times more volatile than interest rates and 10 times more volatile than inflation rates. Figure 6.1 shows that the £/$ and DM/$ FX rates in the 1980s, the

WHAT WENT RIGHT • WHAT WENT WRONG

Export Market Pricing at Toyota1

In February 2002, 1 U.S. dollar (USD) exchanged for ¥135. A popular performance sports car, the Toyota Celica GT-S Coupe, made in Japan and shipped to U.S. dealers, sold for $21,165, meaning that each sale realized almost ¥3 million (i.e., ¥2,857,275) in revenue. Two years later in 2004, the dollar exchanged for only ¥104. This 25 percent deprecia- tion of the USD made Japanese exports to the United States potentially much less profitable. To recover costs and earn the same profit margin back home in Japan, Toyota needed to price the 2004 GT-S Coupe at $27,474 in order to again realize ¥2,857,275 (because ¥2,857,275 ÷ (¥104/USD) = $27,474). Rather than steeply raising the sticker price of their sporty performance coupe and trying to limit the erosion of market share by emphasizing horse- power, style, or manufacturing quality, Toyota opted in- stead to slash profit margins. Specifically, a $390 price increase to $21,555 for the 2004 Celica GT-S earned just ¥2,241,720 in revenue back at headquarters in Tokyo (over ¥600,000 less than in 2002). Toyota’s minimalist approach

to price increases lowered profit margins but preserved the Celica’s foreign market share.

Different companies react in different ways to the chal- lenges presented by such severe currency fluctuations. GM and Ford tend to maintain margins and sacrifice market share whereas Toyota tends to slash margins to grow market share. In part, because of these pricing decisions, between 1985 and 2010 Toyota’s share of the North American car market rose from 5 percent to 22 percent while General Motors’ share fell from 45 percent to 21 percent.

Might one approach have been right for Toyota with 60 percent of its revenue from exports and the opposite approach right for GM? If you were advising Toyota today, would you recommend they emphasize high margins and current profitability or reduced margins and further growth of market share? What role does scale of production and selling cost savings based on market penetration play?

1Based on Ward’s Automotive World, various issues.

Chapter 6: Managing in the Global Economy 179

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

¥/$ FX rate in the 1990s, and the €/$ FX rate in the 2000s were particularly volatile. An- alyzing and forecasting the cash flow effects of such massive FX rate changes (sometimes called “FX risk”) provides crucial information for the marketing, operations, and finan- cial plans of companies like Boeing, Microsoft, and IBM as well as ChemChina, Toyota, and Volkswagen.

Foreign Exchange Risk Foreign exchange risk exposures are of three types: transaction risk exposure, translation risk exposure, and operating risk exposure.

Transaction risk exposure occurs when a purchase agreement or sales contract (a specific “transaction”) commits the company to make future payables or accept future receivables in a foreign currency. Over the time period between executing the contract and actually making or receiving the payments, the company has FX transaction risk ex- posure. Many financial derivatives like FX forward, swap, and option contracts have emerged to assist corporate treasurers in constructing hedges that lay off these transac- tion risk exposures for a modest cost (perhaps 5 percent) known and fixed in advance. Appendix 6A explains the mechanics of these financial hedges.

FIGURE 6.1 Foreign Exchange (FX) Rates: The Value of the U.S. Dollar against Several Major Currencies

0.44

210

238

128

1.82

2.94

1.75 1.43

8.30 8.28 8.28 7.63 6.83

126

0.85

1.08

1.12

1.06

0.93

0.77

0.670.64

94 109

125130

0.77

0.56

0.66 0.66

0.63

0.63

0.55 0.48

0.59

1980 1985 1990 1995

CNY/USD ¥/$

Pr ic

e of

U SD

Year

107 108

122

0.89

2000 2005 2010

2.30

2.10

1.76

1.88

GBP/USD £/$

JPY/USD ¥/$

EUR/USD �/$

DEM/USD DM/$

transaction risk exposure A change in cash flows resulting from contractual commitments to pay in or receive foreign currency.

180 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Secondly, translation risk exposure occurs when a company’s foreign assets (or lia- bilities) are affected by persistent exchange rate trends. Accordingly, the accounting books in the home country must be adjusted. A $100 million assembly plant owned by Volkswagen and located in the United States will need to be written down on the com- pany’s German balance sheet when the U.S. dollar falls from 1.12 €/$ in 2001 to 0.64 €/$ in 2008 (again, see Figure 6.1). This €48 million translation risk exposure, [(€1.12 − €0.64)/$] × $100 million, is easily offset with a balance sheet hedge. A balance sheet hedge would match the magnitude of Volkswagen’s asset loss in the euro value of its American plant to an equivalent reduction in the euro cost of VW’s liabilities in the United States (say, a decline from €112 million to €64 million cost of $100 million in pension plan commitments to American workers). The intent is to leave the net asset position of the U.S. division of Volkswagen unchanged. In general, unless the parent company becomes financially distressed such that the foreign asset write-downs threaten important collateral pledged for business loans, such balance sheet adjustments in for- eign subsidiaries due to FX risk are often just ignored.

Finally, FX fluctuations that result in substantial changes in the operating cash flow of foreign subsidiaries, like those befalling Toyota for the Celica GT-S, are examples of FX operating risk exposure. Operating risk exposures are more difficult to hedge than transaction risk exposures and more difficult to forecast than translation risk exposures. As a result, operating risk exposures necessitate more managerial attention and extensive analysis.

For one thing, the deterioration of export revenues from sales in foreign subsidiaries is just one side of the problem that a rising domestic currency poses. In addition, de- pending on the viability of global competition, operating risk exposures may entail a sub- stantial deterioration of domestic sales as well. When a home currency becomes more valuable (appreciates), competing import products become cheaper in the currency of the home market. These relationships are well illustrated (on the next two pages) by the export and domestic business of the Cummins Engine Co. of Columbus, Indiana.

INTERNATIONAL PERSPECTIVES

Collapse of Export and Domestic Sales at Cummins Engine2

U.S. manufacturer Cummins Engine Co. is the world’s leading producer of replacement diesel en- gines for trucks. Like all durable equipment makers, Cummins’s revenues are highly cyclical, declining steeply in economic downturns. If households buy fewer appliances, clothing, and furniture, less ship- ping by truck is required to deliver inventories from warehouses to restock retail shelves. Less shipping means less truck mileage, and less truck mileage means a slower replacement demand for diesel en- gines. For example, in the severe recession of 2008– 2009, Cummins’s sales fell off 29 percent, operating margins collapsed from 9.6 percent to 4.5 percent, and cash flow plummeted by 49 percent from $5.54 to $2.85 per share. As the U.S. economy improved during late 2009 and 2010, Cummins’s sales and

cash flow rose quickly (see Figure 6.2). This period exhibited one of Cummins’s normal cyclical sales and profit patterns. But not so during 1999–2001 when Cummins’s sales declined 15 percent, margins col- lapsed from 9.4 percent to 4.3 percent, and cash flow plunged from $2.58 to $0.82 per share. There was no U.S. recession during this period. What fac- tors were responsible for this earlier collapse of Cum- mins’s sales, margins, and cash flows?

With a 70 percent market share, Cummins Engine competes domestically against No. 2 Caterpillar (20 percent) and Detroit Diesel (10 percent). But Cum- mins also sells 53 percent of its replacement diesel engines in the export market competing primarily against Mercedes-Benz diesels. The euro price that a Cummins engine can sell for in Munich or

(Continued)

translation risk exposure An accounting adjustment in the home currency value of foreign assets or liabilities.

operating risk exposure A change in cash flows from foreign or domestic sales resulting from currency fluctuations.

Chapter 6: Managing in the Global Economy 181

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Rotterdam (and still recover its cost plus a small profit) is as important to Cummins’s cash flow as metal input costs or wage bargains with the United Machinists Union. A $40,000 835 cubic inch Cum- mins Series N diesel sold for approximately €34,000 in 1999 and again in 2004 across Europe. In each of those years, the exchange rate between the euro and the U.S. dollar stood at approximately 0.85—that is, €0.85 per USD. In the intervening time period, how- ever, the U.S. dollar appreciated substantially. By 2001, the value of the dollar had soared almost 27 percent from €0.85 to €1.12 (again see Figure 6.1).3

The effect of the 1999–2001 dollar appreciation on Cummins’s export sales was catastrophic. For Lufthansa Airfreight, a German trucking company, to buy a $40,000 Cummins diesel engine at the dol- lar’s peak in 2001 required (€1.12/$) × $40,000 = €44,800! No feature of the equipment had changed. No service offering had changed. No warranty had changed. The export diesels from the United States to Germany (once priced at €34,000) had simply

become €10,800 more expensive solely because the euro currency of those European buyers had become much less valuable relative to the steeply appreciating U.S. dollar. Such enormous price increases of the Cummins export product made substitute products, like a €34,000 Mercedes-Benz diesel, substantially more attractive to European buyers than before the change in the exchange rate.

Moreover, in 2001, Mercedes-Benz perceived a huge opportunity to sell its own replacement diesels into Cummins’s home territory in the United States. An import diesel made by Mercedes-Benz (which had sold in Boston, Cleveland, and Chicago for €34,000/0.85 = $40,000 in 1999 and would do so again in 2004) could cover cost and a small profit while selling for just $30,357 (€34,000/1.12) at the peak of the dollar in 2001. Not surprisingly, Mercedes-Benz diesels sold very well that year throughout the United States. Cummins was forced to bargain with its most loyal domestic customers much more so than usual and shave its margins.

FIGURE 6.2 Cummins Engine Cash Flow and Operating Margins

201020092008200720062005200420032002

Cash flow per share

Operating margin

200120001999

9.4% 7.4% $ 0.825.3% 5.6% 5.0%

8.2% 12.0% 12.6%

9.4% 9.6%

4.5%

$2.80

$5.54

$2.58

$1

$2

$3

$4

$5

D ol

la rs

p er

s ha

re

Source: Value Line Investment Survey, Ratings and Reports, October 23, 2009.

182 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

OUTSOURCING A lasting effect of all the international competitive pressure on U.S. manufacturers in these strong dollar years was a laser-like dedication to cost cutting. With just-in-time de- livery of component subassemblies, production to order, and other lean manufacturing techniques, U.S. companies sliced inventory costs, reduced scrap, and cut manufacturing cycle times enough to boost productivity 25 percent from 2001 to 2005.4 That is, the ratio of merchandise produced to input cost rose at firms like Cummins, Caterpillar, General Electric, and Boeing. One reason was better operations management techniques but another was the outsourcing of IT, less complex assembly, and raw materials han- dling tasks to places like Mexico, Korea, Malaysia, and now India and China.

An HP laptop computer is 95 percent outsourced, with components from eight coun- tries delivered just-in-time each day for assembly in a Shanghai factory and then air freighted overnight to Memphis by FedEx for shipment directly to customers the next day (see Figure 6.3). Outsourcing is not new. U.S. companies like Merck, DuPont, and IBM have outsourced for decades to Germany, France, Ireland, and now India in order to access skilled people for analytical jobs in R&D. In addition, basic manufacturing of low-skill, low-wage jobs has always moved offshore, first from Europe to America in the late nineteenth century and then in the mid-twentieth century to Canada, Mexico, Brazil, and Portugal, and more recently to Malaysia, Thailand, India, and now China.

One reason is that shipping costs from Shanghai to New York for a 40-foot container filled with 6,000 garments runs just $8,000 ($1.35 per garment) for the 30-day ocean voyage through the Panama Canal (see Figure 6.3), and only $10,000 ($1.70 per gar- ment) for the 20-day intermodal service through Long Beach harbor and then onto U.S. trucks. Given their high value/weight ratio, Hewlett-Packard laptops warrant the $50,000 cost for a Boeing 747 air freighter from Shanghai to the FedEx hub in Memphis, Tennessee, adding approximately $2 cost per laptop (see Figure 6.3). Across all types of cargo, ocean shipping adds only 3 to 4 percent to the delivered price of a product.

Offshoring does necessitate some additional cost, however. Full costs offshore include costs for careful vendor selection, intellectual property protection, and compensation for expatriate managers. Hewitt Associate LLC estimates that to transfer a director-level U.S. employee to Beijing or Shanghai requires $190,000 additional compensation and a $60,000 housing allowance relative to U.S. job sites. By one estimate, $5 per hour call- center jobs in India or $2 per hour factory jobs in China require another $12 per hour

Therefore, in response to the 27 percent steep appre- ciation of the dollar, not only did Cummins’s export sales collapse, but so did Cummins’s domestic sales (and margins). These operating risk exposures from fluctuations in the FX rates require substantial man- agement attention because they are of uncertain mag- nitude and unpredictable timing making them much more difficult to hedge away than transaction risks.

The strong U.S. dollar in 1999–2003 (and earlier in 1982–1985; see Figure 6.1) put U.S. manufacturers of traded goods, like autos, VCRs, airplanes, and diesel

engines, at a very big disadvantage. U.S. exports sharply declined, U.S. imports sharply increased, and the U.S. trade deficit (exports minus imports) skyrocketed.

2Based on Value Line Investment Survey, Part III: Ratings and Reports, vari- ous issues. 3Exchange rate percentage changes are calculated as the difference from one period to the next divided by the average exchange rate over the period. The reason for this midpoint procedure is that when the EUR/USD exchange rate returns in 2001–2004 to very nearly its original level (i.e., €0.85/$ at 1999 and 2004 in Figure 6.2), the midpoint calculation yields a downward adjustment of −27 percent, equal and opposite to the +27 percent rise in 1999–2001.

4Based on “Lean and Unseen,” The Economist (July 1, 2006), pp. 55–56; and Murray Weidenbaum, “Outsour- cing: Pros and Cons,” Business Horizons (2005), pp. 311–315.

Chapter 6: Managing in the Global Economy 183

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

to cover these additional expatriate costs of outsourcing. A pivotal question then is whether domestic factory workers in the United States are available for $14 to $17 per hour. On average, the answer is “no”; the full labor cost of production workers in 2008 U.S. manufacturing totals $24.59 per hour.5 So, outsourcing to realize lower manufactur- ing cost is going to continue.

Outsourcing to foreign contract manufacturers is, however, as much about importing competitiveness as it is about exporting jobs. Time-to-market and the capacity to inno- vate quickly have become more important in auto manufacturing, for example, than the assembly cost of the next shipment of three-speed transmissions. If Toyota and Nokia and Advanced Micro Devices (AMD) introduce major product innovations on two- year, six-month, and three-week timetables, respectively, and if those cycles have become the key to customer acceptance, then GM, Motorola, and Intel have no choice but to access the people and processes that can match that capability. If those people are Indian software engineers, Mexican foundry employees, and Chinese assembly workers, then the best hope for American workers is to sell to those developing economies in Asia and Europe the International Harvester tractors, GE appliances, Boeing aircraft, Cisco

FIGURE 6.3 Outsourcing Shipping Costs and Component Sources for HP Personal Computer

Shanghai-Long Beach-N Y

20 days, $10,000

30 days, $8,000

4

6

7

3

2 5

8

Japan United States

0 25 50 75

Shangha i-Memphis

2 days

, $50,000 1

Shanghai-Panama Canal-NY

Import-Export Trade (% GDP)

1

2

5

2 3 4

6

5

1

1 8 6

6 4 2

1 7

Hard-Disk Drive Battery Memory Chips Liquid-Crystal Display Microprocessors Graphics Processors

U.S. , China , Singapore , Japan China S. Korea , Taiwan , U.S. , Germany S. Korea , Taiwan , Japan , China U.S. U.S. , Canada , Taiwan

Country-of-Origin of Major Components Assembled in Shanghai

2008 USD Cost for Ocean Freight versus Airfreight from Shanghai to FedEx Distribution Hub

China Russia India Brazil EU 15

Source: The Economist, August 9, 2008, p. 64; Wall Street Journal, June 9, 2005, p. B1; World Trade Organization; and Thomson Datastream.

5U.S. Department of Labor, Bureau of Labor Statistics, International Comparisons of Hourly Compensation Costs in Manufacturing, 2007 (March 26, 2009), p. 6.

184 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

network servers, as well as the management consulting, banking, and legal services those lower-wage economies require for this stage of their development.

As the U.S. dollar weakened persistently against the euro throughout 2003–2008 (again see Figure 6.1), American exporters experienced a turnaround and profited from booming European sales. Seven percent of IBM’s reported 11 percent sales growth in 2003–2004 was attributed to currency fluctuations. Similarly, Colgate-Palmolive and Microsoft, respectively, reported two-thirds of a 20 percent sales increase in Europe and nine-tenths of a 12 percent sales increase in Europe, the Middle East, and Africa were attributable to the lower in-country prices emergent from a weaker U.S. dollar.6

CHINA TRADE BLOSSOMS An amazing story is unfolding in the People’s Republic of China (PRC) along the Yellow, Yangtze, and Pearl River systems and in the port megacities of Shanghai, Guangzhou, Dalian, and Zhuhai. Three decades ago, China accounted for an insignificant 0.6 percent of world trade despite a highly educated workforce and an immense population of 1.3 billion. China’s share of world trade has increased more than tenfold, growing in the past five years from 4.2 percent in 2003 to 8 percent in 2008.7 Eight percent of $32 trillion in import + export trade worldwide is $2.5 trillion, equivalent to 57 percent of China’s $4.33 trillion economy (officially measured at nominal FX rates in 2008 as third largest behind the United States’ $14 trillion and Japan’s $5 trillion). China has generated over 10 percent GDP growth per annum, 1995–2008. The trade share of China with the United States has risen from 10.7 percent of all U.S. imports in 2003 to 15.9 percent in 2008, only marginally less than the United States’ largest trading partner—Canada with a 16 percent share.8 China’s economic real growth is unprece- dented both in its longevity and its double digit magnitude.

China’s largest trading partners are Japan, Taiwan, the United States, South Korea, and Hong Kong. The principal categories of exports are clothing to Hong Kong, machinery, toys, furniture, shoes, and clothing to the United States, textiles to Japan, and telecommu- nications equipment to Germany. In many cases Chinese factories are assembling compo- nents in facilities that have emerged as joint venture investments of the PRC’s China International Trust and Investment Corp. with Taiwanese, Japanese, European, and U.S. multinationals. A good example is HP’s laptop computer assembled in Shanghai by Quanta Computer, a Taiwanese partner in HP’s supply chain (see Figure 6.3). With wage rates of $1 to $2 per hour for assembly line work, China has quickly become the world’s largest producer of televisions, computers, toys, bicycles, steel, and wood furniture. Chinese import flows reflect the goods and services needed to sustain such astounding growth— that is, electrical machinery and aircraft from the United States, steel from South Korea, autos and chemicals from Japan, rubber, iron ore, ships, and cement from Australia.

From joint ventures in auto parts with Ford Motor to aircraft construction with McDonnell-Douglas, foreign companies historically practiced caution in staging their co-investment in China.9 An unbridled enthusiasm warranted by the growth opportunity of a lifetime was balanced against threats to their intellectual property. Two events in 2000 and 2001 set a different story in motion. In 2001, China joined the World Trade Organization (WTO) and agreed to drop some restrictions on foreign investment and

6“Dollar’s Dive Helps U.S. Companies,” Wall Street Journal (April 21, 2003), p. A2. 7Jinghai Zheng et al., “Potential Output in a Rapidly Growing Economy: The Case of China,” Federal Reserve Bank of St. Louis Review (July/August 2009), p. 333. 8Federal Reserve Bank of St. Louis, National Economic Trends (July 15, 2009), p. 18. 9See James McGregor, One Billion Customers: Lessons from the Front Lines of Doing Business in China (New York: Wall Street Journal Books, 2005).

Chapter 6: Managing in the Global Economy 185

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

to abide by the WTO standards for protecting patents and copyright. The year before, the U.S. Congress normalized trade relations by extending to China permanent most- favored-nation status, which removed numerous U.S. tariffs. As a direct result, 57,000 investment projects have been started with U.S. and other foreign direct investment part- ners since then. Chinese auto parts output has grown the past five years from a ¥315 billion to a ¥960 billion yuan business, about 35 percent of which is exported (to Ford Motor and other partners).10 Overall, Chinese exports in 2007 grew by 27 percent, and Chinese imports grew by 20 percent to average over the 2001–2007 period 6.4 percent of world imports. Only the United States at 15.5 percent and Germany at 7.4 percent im- ported more merchandise.

China Today China today is a study in contrasts not unlike other fast-growing nations throughout his- tory. For more than a decade from 1995–2007, China’s GDP grew at double-digit nomi- nal rates of 12 to 15 percent.11 With GDP doubling every 5 to 6 years and population growth of less than 1 percent (about 10 million on a 1.2 billion base), the standard of living in the Eastern coastal provinces of China is advancing very rapidly. What acceler- ated Chinese development to make possible this spectacular and almost unprecedented record of economic growth?

If China were simply an export platform for multinational corporations seeking cost savings in assembling manufactured goods, the WTO answer would suffice. And our dis- cussion in the previous section on Hewlett-Packard’s laptop computer assembly in Shanghai is instructive. But considering China an export machine like Korea is mislead- ing. Over the past 10 years, net exports accounted for only about 10 percent of Chinese real GDP growth.12 Instead, a recent Booz-Allen survey found that China was being re- lied upon for product development as well as procurement, more so than Canada, Europe, Latin America, or India.13 In addition, some of China’s giant state-owned companies like ChemChina are globalizing and competing like private-sector firms against other multi- nationals.14 And in 2008 some of China’s state-owned enterprises (SOEs) like PetroChina, State Grid Corporation, and China Mobile bid successfully to run offshore utilities in the Philippines and acquired Paktel mobile operations in Pakistan.

In addition, another important part of the answer to the unprecedented Chinese growth is the liberalization of property rights. Starting in 1978, China’s enterprising owners of start-ups were granted entitlement to residual cash flows from their small businesses. In 1980, farmers were given 99-year use rights and residual appropriation rights to 3 acres. Ownership rights remained with the State but these “resource holders” became entitled to long-term leasing of their land assets to small business (especially small factory) owners. By 1998, still more extensive ownership rights were extended to urban property.15 The Chinese can now buy and sell and develop for investment not only resident houses and apartments but attached commercial spaces.

The middle class created by these new property rights arrangements is expanding rap- idly. One can drive in eastern coastal regions of China along the motorway southwest of

10Chinese Car-Parts Makers Expand,” Wall Street Journal (June 12, 2009), p. B2. 11In 2008 (Q4), Chinese GDP growth slowed to 9% and further to 7% in 2009 (Q1) but Q2 returned to 13%. Of course other major economies were slumping into negative “growth” in 2008–2009. See “China’s Recov- ery,” The Economist (July 18, 2009), pp. 37–38. 12Rebalancing the World Economy: China,” The Economist (August 1, 2009), pp. 65–66. 13Based on Booz-Allen Offshoring Business Network Survey, Forbes (September 3, 2007), p. 56. 14“Special Report: China’s Business Landscape,” The McKinsey Quarterly 2008, No. 3, pp. 1–6. 15See “A Survey of China,” The Economist (March 25, 2006).

186 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Shanghai and see nothing but brick McMansions kilometer after kilometer all the way to Hangzhou 150 km away. The pattern repeats itself in an eerily repetitive way. Each farmer with 3 acres leases an acre to an aspiring small factory owner. One half acre con- tains the factory owner’s McMansion; the other half acre contains the factory. Another half acre is the farmer’s McMansion, and perhaps a third McMansion and small factory all sit on 2.5 of the 3 acres. A few garden plots survive on the remainder. Next door exactly the same pattern unfolds. The three-story homes are all brick and concrete as building materials and construction labor are very cheap. Take any exit off the motorway and more such 3-acre enclaves unfold in every direction. Occasionally these densely packed McMansion patterns are punctuated by a large 10-acre factory complex or a new town development with office buildings, shops, and schools.

The Shanghai region is clearly the most developed in China. Several eastern coastal China provinces have always profited from expanded trade relative to inland provinces. And the Special Economic Zones of Xiamen in Fujian Province across the Taiwan Straits and Guangzhou in Guangdong Province across from Hong Kong have similar geo- graphic, cultural, and government-mandated economic advantages. Nevertheless, Shang- hai City in Jiangsu Province and adjacent Hangzhou, capital of Zhejiang Province, are something special. On the approaches to Shanghai or Hangzhou, one can find planned communities with living standards equivalent to the outlying suburban “new towns” like Reston near Washington, D.C., and Sun City near Phoenix, Arizona. The Shanghai re- gion looks like this in many directions while more rural provinces in western China re- main largely untouched by modern life.

Hangzhou was the largest city in the world in the thirteenth and first half of the four- teenth centuries. The Lower Song dynasty made the city its capital from 1123 up to the Mongol Invasion in 1276. Hangzhou was connected to Beijing 1,260 km away in central China by a Grand Canal of China completed in 609. Shanghai and Hangzhou are at the northern and southern reaches of the Yangtze River delta. Hangzhou is a thriving me- tropolis of green spaces and light industry the size of Atlanta with 4 million residents in the metropolitan area. It ranks as China’s 18th largest city, much smaller than Shang- hai (19 million), Beijing in central China (12 million), Guangzhou (10 million), and Shenzhen (9 million).

Massive infrastructure investment is another striking impression in the Shanghai- Hangzhou corridor. Train stations, airports, bridges, and tunnels appear to have been built anew in the past five years. Central planning decisions by Party officials streamline the creation of large infrastructure right-of-ways. For example, the avenue from down- town to the Hangzhou airport is lined with high-tech companies like HP, Sony, and Sie- mens. Clearing previous land uses and accumulating parcels for such infrastructure developments happen with a speed and certainty in China that belies the difficulty of reaching agreement on such projects in Germany, Japan, or the United States. Constitu- tional challenge to publicly mandated land use decisions is unknown in China.

Another sharp contrast with Europe and the Americas exists in foreign exchange poli- cies where China maintains a so-called managed float of foreign currency exchanged at only officially authorized FX rates. This situation contrasts sharply with the United States.

THE MARKET FOR U.S. DOLLARS AS FOREIGN EXCHANGE To explain the determination of FX rates in countries with freely fluctuating exchange rates, we now address demand, supply, and market equilibrium in the currency markets. Since American manufacturers, like Cummins Engine, incur many of their expenses at

Chapter 6: Managing in the Global Economy 187

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

domestic manufacturing sites in the United States, American manufacturers tend to require that export purchase orders be made payable in U.S. dollars. This receivables pol- icy requires that Munich buyers of Cummins diesels transact simultaneously in the for- eign currency and diesel markets. To buy a Cummins diesel, Munich customers (or their financial intermediaries) will supply euros and demand dollars to secure the currency required for the dollar-denominated purchase order and payment draft awaited by the Cummins Engine shipping department. This additional demand for the dollar and the concurrent additional supply of euros drive the price of the dollar higher than it other- wise would have been. Thus, the equilibrium price of the dollar as foreign exchange (in euros per dollar on the vertical axis in Figure 6.4) rises. In general, any such unantici- pated increase in export sales results in just such an appreciation of the domestic cur- rency (here the USD).

Similarly, any unanticipated decrease in export sales results in a depreciation of the domestic currency. For example, in 2001–2005, collapse of Boeing Aircraft export sales in competition with Europe’s Airbus contributed to the dollar’s collapse against the euro over that same period (again see Figure 6.1). But this downward FX trend of the dollar assisted Cummins Engine as well as Boeing in stabilizing its sales and cash flows both at home and abroad. With the dollar worth fewer euros, the dollar cost of Ameri- can imports from Mercedes-Benz and Airbus became more expensive, while American exports priced in euros by Cummins and Boeing sales reps across Europe became cheaper. This automatic self-correcting adjustment of flexible exchange rates in response to trade flow imbalances is one of the primary arguments for adopting a freely fluctuat- ing exchange rate policy.

FIGURE 6.4 The Market for U.S. Dollars as Foreign Exchange (Depreciation of the Dollar, 2001–2008)

Pr ic

e of

U .S

. d ol

la r

(� p

er $

)

D0

S2001

S2003

Quantity ($)

1.12

0.93

0

0.77 0.64

S2005 S2008

188 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Import/Export Flows and Transaction Demand for a Currency To examine these effects more closely, let’s turn the argument around and trace the cur- rency flows when Americans demand imported goods. The Mazda and BMW dealers would have some inventory stock on hand, but suppose an unexpectedly large number of baby boomers wish to recapture their youth by purchasing sporty BMW convertibles. Just as Cummins Engine prefers to be paid in U.S. dollars, so too BMW wishes to be paid in euros. Therefore, BMW purchase orders must be accompanied by euro cash pay- ments. How is that accomplished? First, the local BMW dealer in Charlotte requests a wire transfer from Bank of America (BOA). BOA debits the dollar account of the dealer, and then it authorizes payment from the euro cash balances of BOA and presents a wire transfer for an equivalent sum to the Munich branch of Deutsche Bank for deposit in the BMW account. Both import buyer and foreign seller have done business in their home currencies and exchanged a handsome new car. And the merchandise trade account of the U.S. balance of payments would show one additional import transaction valued at the BMW convertible’s purchase price.

If BOA failed to anticipate the import transaction and euro wire transfer request, BOA’s foreign currency portfolio would now be out of balance. Euro balances must be restored to support future anticipated import-export transactions. BOA therefore goes (electronically) into the interbank foreign currency markets and demands euros. Al- though the American bank might pay with any currency in excess supply in its foreign currency portfolio that day, it would normally pay in U.S. dollars. In particular, if no other unanticipated import or export flows (and no unanticipated capital flows) have oc- curred, BOA would pay in U.S. dollars. Therefore, unanticipated demand by Americans for German imports both raises the demand for euros and (as the flip side of that same transaction) increases the supply of U.S. dollars in the foreign currency markets.

The Equilibrium Price of the U.S. Dollar In particular, in the market for U.S. dollars as foreign exchange (see Figure 6.4), the sup- ply curve shifts to the right. This shift of market supply represents BOA and many other correspondent banks supporting import transactions by selling dollars to acquire other foreign currencies. The equilibrium price on the y-axis of Figure 6.4 is the price of the dollar expressed in amounts of foreign currency—for example, British pounds per USD, Chinese yuan per USD, Japanese yen per USD, or euros per USD. As the supply of U.S. dollars increased S2001 to S2008, the equilibrium price of the dollar declined continuously from €1.12 to €0.93, €0.77 and €0.64.

For example, in order for the American imports of Airbus airplanes to increase for 2002–2005, the supply of U.S. dollars in the foreign currency market had to increase. Thus, the spectacular dollar appreciation of the previous three years (1999–2001) slowed and the dollar began to depreciate (see Figure 6.1). Again, American consumers and companies needed to acquire euros to purchase Franco-German imports. U.S. financial intermediaries supplied dollars in the market for dollars as foreign exchange to acquire the foreign currencies their local American customers (Delta, United, Continental) re- quested in support of these foreign import transactions.

Speculative Demand, Government Transfers, and Coordinated Intervention The U.S. dollar depreciation of 2001–2008 reflects several factors besides transaction de- mand. FX rates also depend upon speculative demand, government transfers, and central

Chapter 6: Managing in the Global Economy 189

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

bank interventions. Speculative demand is very volatile. Transfers can involve either debt repayment (reducing the supply of a currency as a debtor nation takes money out of cir- culation and returns it to the lender nation’s treasury) or foreign aid (increasing the sup- ply of a currency). Government interventions can be coordinated across several central banks or uncoordinated and can be sterilized or unsterilized. Sterilized interventions in- volve offsetting transactions in the relevant government bond market. For example, the Federal Reserve might sell dollars in the foreign currency markets, but then turn around and acquire dollars by selling an equal dollar volume of T-bonds to Japanese or Chinese investors, leaving the supply of dollars in international exchange essentially unchanged.

What is the proportionate weight of each of these factors in determining the equilib- rium value of a currency? An important first perspective is that only one out of five for- eign exchange transactions supports an import or export trade flow; the other four support international capital flows. In 2008, for example, the average daily volume of for- eign currency transactions in the 43 largest foreign currency spot markets had a dollar value of $1.005 trillion. The average daily volume of world exports was $43 billion. So the dollar volume of foreign currency flows outstrips the dollar volume of foreign trade flows by 23 to 1. Daily FX rate fluctuations therefore reflect not import-export trade flows but rather international capital flows, much of which is speculative and transitory.

Because of the sheer volume of transaction, intervention in the foreign currency mar- kets by any one central bank therefore has almost no chance of affecting the equilibrium value of a currency. Take the Bank of Japan, for instance. In June 2009, the central bank of Japan had official reserves of foreign currencies equal (at existing exchange rates) to $988 billion.16 By comparison, China had $1.4 trillion, the European Central Bank had $49 billion, while the U.S. Federal Reserve had $42 billion in foreign currency reserves.

Suppose the Bank of Japan decided to try to initiate a depreciation of the Japanese yen (JYN) to improve the competitiveness of their export sector. Investing one quarter of their entire reserves or $247 billion, the Bank of Japan’s intervention would be easily over- whelmed by the sheer enormity of the $1 trillion mobile capital awash daily in the interna- tional currency markets. Indeed, the official reserves of all the central banks total only approximately $3 trillion, equal to only three day’s worth of foreign currency transactions.

It therefore takes a coordinated intervention by several large central banks with deep reserves to have any real chance of permanently affecting a currency’s value. One such coordinated intervention took place as a result of the Plaza Accords in 1985 when the G-7 nations (the United States, Britain, Japan, Germany, France, Italy, and Canada) all agreed to a sustained sale of U.S. dollars over 1986 and 1987. Figure 6.1 shows that this coordinated intervention was effective in lowering dollar FX rates against the GBP, the JYN, and the DEM.

Short-Term Exchange Rate Fluctuations The transaction demand determinants of long-term quarterly or annual trends in ex- change rates are fundamentally different from the determinants of day-to-day exchange rate fluctuations. Short-term exchange rate movements from week to week, day to day, or even hour to hour are determined by arbitrage activity in the international capital markets and by speculative demand. Sometimes current events set speculators off in sup- port of demanding and holding a currency (a long position), and sometimes the reverse (a short position). Behaviorally, each speculator tries to guess what the others will do, and much herd-like stampeding behavior based on volatile investor expectations often ensues.

sterilized interventions Central bank transactions in the foreign exchange market accompanied by equal offsetting transactions in the government bond market, in an attempt to alter short-term interest rates without affecting the exchange rate.

16International Monetary Fund, International Reserves and Foreign Liquidity, various issues.

190 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Arbitrage is the act of buying real assets, commodities, stocks, bonds, loans, or even televisions, iPods, and sandals cheaply and selling them at higher prices later. Arbitrage activity is triggered by temporary violations of arbitrage equilibrium conditions, which equalize, for example, real rates of return on 90-day government bonds (adjusted for any differences in default risk). When such conditions do not hold, opportunities for ar- bitrage profits exist, and arbitrage activity will appear quickly, proceed at huge volume, and continue until the relevant arbitrage equilibrium conditions are reestablished. Again, the sheer magnitude of the floodtide of $1 trillion per day of international currency flows quickly closes (within hours or even minutes) the window of arbitrage profit opportunity in FX trading. If the buying and selling prices and terms of delivery can be arranged si- multaneously, then the transaction is one of pure arbitrage. If the second transaction is delayed, we often call the activity speculation.

DETERMINANTS OF LONG-RUN TRENDS IN EXCHANGE RATES Long-run trends in FX rates are quite different. Understanding the forces that set in mo- tion the exchange-rate-induced swings in sales and profit margins that constitute operat- ing risk exposure is crucial for effectively managing export business. And since domestic business today is almost universally subject to effective and intense import competition, the same holds true for largely domestic businesses. The quarter-to-quarter or year- to-year trends in FX rates depend on three factors: real growth rates, real interest rates, and anticipated cost inflation rates. We now discuss each of these determinants in turn.

The Role of Real Growth Rates As we have seen, a primary determinant of the year-to-year exchange rate fluctuations in Figure 6.1 is the net direction of trade flows. Unanticipated increases in imports lower a local currency’s value, whereas unanticipated increases in exports raise a local currency’s value. The stimulus underlying such trade flow imbalances may be either business cycle based or productivity based. In an expansion, domestic consumption (including import consumption) increases, causing exports from a country’s trading partners to increase; in a domestic contraction, import consumption decreases, causing exports from the trading partners to decline.

During the years 2002–2006, U.S. gross domestic product (GDP) grew at 1.6 percent, 2.7 percent, 4.2 percent, 3.5 percent, and 2.8 percent in real terms (that is, adjusted for inflation) (see Table 6.1). The United States had an accelerating economy. In those same years, euro area real GDP, on the other hand, exhibited anemic growth at 0.9 percent, 0.5 percent, 1.8 percent, 1.4 percent, and finally 3.0 percent. Although Canadian growth rates were comparable to the United States, Mexico and Japan also had slower growth than the United States (again see Table 6.1). Among the five largest U.S. export trading partners, only China was growing faster than the United States (at 8.3 percent, 9.5 per- cent, 9.5 percent, 9.0 percent, and 11.6 percent). These growth-rate trends led to in- creased import flows into the United States of goods like autos, textiles, furniture, and consumer electronics, while causing decreased exports from the United States of goods like computer software, PCs, grains, films, aircraft, professional services, medical devices, and diesel engines to everywhere except China.

As U.S. net exports fell (declining exports minus rising imports) in 2002–2007, the dollar depreciated, so year after year throughout this period, foreign buyers had to acquire fewer dollars to complete purchase transactions with U.S. companies like Microsoft, IBM, Archer Daniels Midland, Boeing, McKinsey, and Cummins Engine. In the market

arbitrage Buying cheap and selling elsewhere for an immediate profit.

speculation Buying cheap and selling later for a delayed profit (or loss).

Chapter 6: Managing in the Global Economy 191

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example European Slowdown Decreases DuPont Exports17

“Beggar thy neighbor” became the rallying cry for trade policy in the Mercantilist period from 1500–1750, during which punitive tariffs and other forms of trade protectionism isolated city and provincial economies. Today, however, rather than attempting self-sufficiency, most nations are better off recognizing mutual interde- pendence, encouraging both import and export activities, and specializing in accor- dance with comparative advantage. In that freer trade environment, growing neighbors are the best neighbors.

In the late 1990s, an economic slowdown in Europe undercut the export sales of American manufacturers and U.S. multinationals. After several prior years of steady increases, import demand by households and companies flattened in Ger- many and France, and it actually declined in Italy and the United Kingdom. Not only Cummins Engine with 28 percent that year, but also Sun Microsystems (36 percent), and DuPont (39 percent), and even Wrigley Chewing Gum (41 percent) and McDonald’s (37 percent) realized a large proportion of their revenue in Eur- ope. Across all S&P 500 U.S. companies, 21 percent of 1997 sales revenue arose from European sales. As a result, the U.S. industrial sector was particularly hard hit by the European slowdown; DuPont chemicals shipments from the United States to Europe, for example, declined by 20 percent on a year-to-year basis.

17“Weak Growth in Europe Threatens U.S. Exports,” Wall Street Journal (March 11, 1999), p. A1.

TABLE 6.1 TRANSACTION DETERMINANTS OF LONG-TERM EXCHANGE RATE TRENDS

UNITED STATES GERMANY/EURO

AREA JAPAN CHINA

REAL GDP

REAL R PPI

REAL GDP

REAL R PPI

REAL GDP

REAL R PPI

REAL GDP

REAL R PPI

1995 2.7 3.1 1.9 2.0 3.2 2.0 1.9 1.3 −0.6 10.5 −6.7 14.1

1996 3.7 2.5 2.7 1.0 1.4 1.0 2.6 0.5 −1.4 9.6 0.7 6.8

1997 4.5 3.3 0.4 1.9 1.4 1.8 1.4 −1.1 0.2 8.8 5.8 2.0

1998 4.2 4.0 −0.8 1.8 2.5 0.5 −1.9 0.0 0.3 7.8 5.5 −0.9

1999 4.4 3.1 1.8 2.9 1.9 3.9 −0.1 0.5 −0.9 7.1 4.6 −0.8

2000 3.7 3.1 3.8 4.0 2.3 7.1 2.9 0.9 −0.2 8.0 2.9 2.4

2001 0.8 0.9 1.9 2.0 1.9 2.0 0.4 0.8 −0.7 7.5 2.7 2.8

2002 1.6 0.1 −1.3 1.0 1.0 −0.1 0.1 1.0 −0.9 8.3 3.5 1.3

2003 2.7 −1.1 3.2 0.8 0.3 1.4 1.8 0.34 −0.5 9.5 1.5 3.1

2004 4.2 −1.1 3.6 1.8 0.0 2.3 2.3 0.03 1.4 9.5 −0.7 7.5

2005 3.5 0.1 4.9 1.4 0.0 4.1 2.6 0.33 2.0 9.0 1.5 4.6

2006 2.8 3.0 3.2 3.0 0.9 2.0 2.1 0.0 −0.9 11.6 −1.8 3.6

2007 2.0 3.4 2.7 2.6 1.2 2.3 2.4 0.6 −0.7 13.0 −1.5 4.2

2008 1.1 −0.8 2.2 0.7 1.3 2.3 −0.7 −0.7 −0.9 9.0 −3.1 4.9

Notes: Real GDP refers to the growth rate of gross domestic product adjusted for price changes by the GDP deflator. Real r refers to the short-term interest rate on government debt minus the annual percentage change in consumer prices. PPI is the annual percentage change in producer prices.

Source: Federal Reserve Bank of St. Louis, International Economic Trends, July 2009; European Central Bank, Statistics Pocketbook, 2009.

192 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

for USDs as foreign exchange depicted in Figure 6.5, the decline of net exports from the United States decreased the demand for dollars; D2001 shifted down to D2002 and on to D2005 and D2008. At the same time, increased purchases of foreign imports by Americans increased the supply of dollars. That is, in the market for dollars as foreign exchange, S2001 shifted right to S2005 and on to S2009. These shifts led to a depreciation of the U.S. dollar over the period 2001–2008. In sum, the decrease in exports and increase in imports led to a fall in the price of the dollar against the euro from €1.12 per dollar in 2001 to €0.64 per dollar in 2008 and against the yen from ¥125/$ in 2002 to ¥94/$ in 2009. The collapse of these €/$ and ¥/$ FX rates are depicted at the far right of Figure 6.1.

Continuing decay in the dollar’s value threatens to derail the competitiveness of Chi- nese and Japanese exports and depreciate a massive balance sheet asset held by the cen- tral banks of China and Japan. In particular, in 2009 the Chinese (at $1.6 trillion) and Japanese (at $550 billion) alone held over $2 trillion in official U.S. dollar reserves. As a result, in 2008–2009 these two and a few other developing country central banks in Asia regularly purchased hundreds of billions of U.S. Treasury securities in an attempt to bol- ster the USD’s value.18

FIGURE 6.5 The Market for U.S. Dollars as Foreign Exchange (Depreciation against the Yen, 2007–2009, and the Euro, 2001–2008)

Pr ic

e of

U .S

. d ol

la r

(y en

/$ )

an d

eu ro

/$

D2008

D2005

D2002

D2001

S2005

S2009

S2001

Quantity ($)

(125) 1.12

0.64 (94)

0

(108)

5 4

2

3

1

18By accepting U.S. T-bill promissory notes and T-bonds in exchange for official dollar reserves, the Chinese, Japanese, and Singaporean central banks effectively reduce the supply of U.S. dollars in circulation. As in Figures 6.4 and 6.5, reduced supply of USDs implies a higher equilibrium price. See “The Dollar U-Turn?” Federal Reserve Bank at St. Louis, International Economic Trends (February 2006), p. 1.

Chapter 6: Managing in the Global Economy 193

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Role of Real Interest Rates The second factor determining long-run trends in exchange rates is comparable interest rates adjusted for inflation. The higher the real rate of interest in an economy, the greater the demand for the financial assets offered by that economy. If a Japanese, German, or Swiss investor can earn higher returns (for equivalent risk) from U.S. Treasury bills than from Euromoney or Japanese government bills, foreign owners of capital will move quickly to rebalance their portfolios to incorporate more U.S. assets. Since the New York Federal Reserve Bank auctioning off new T-bills; J.P. Morgan underwriting a new issue of DuPont bonds; Merrill-Lynch selling T-bills, T-bonds, and DuPont bonds in the secondary (resale) market; and the New York Stock Exchange settlements department all require payment in U.S. dollars, the foreign investor who desires U.S. financial assets must first acquire U.S. dollars to complete his or her purchase transactions. So, a higher real interest rate in the United States (relative to European, Japanese, and British rates) implies international capital inflow into the United States and an increased demand for and appreciation of U.S. dollars.

What really matters in triggering these international capital flows is an investor’s ex- pectation of the domestic value of the foreign interest earned when capital invested over- seas is redeemed and the interest paid is converted back from the foreign to the investor’s home currency. Nominal interest rates minus consumer inflation rates in the foreign country approximate this post-redemption return (see the column labeled real r in Table 6.1). In mid-1999, U.S. three-month and six-month T-bills yielded on average 5.3 percent with a 2.2 percent inflation forecast, returned therefore 3.1 percent after ad- justing for inflation. A year later in mid-2000, the real rate of return remained 3.1 per- cent in the United States (i.e., 6.5 percent − 3.4 percent). In the euro area, short-term interest rates also rose over this period from 3.0 percent to 4.4 percent, but so did antic- ipated inflation, growing from 1.1 percent in mid-1999 to 2.1 percent in mid-2000. Con- sequently, the real rate of short-term interest across Europe increased from 1.9 percent to 2.3 percent.

With 120 basis points favoring investment in U.S. assets in 1999 ([.031 − .019] × 100), 150 basis points in 1998, and 190 basis points in 1997, foreign capital literally flooded into T-bills and other U.S. short-term money instruments. For example, between 1996 and 2000, European companies and investment funds like BPA Amoco, British Telecom, BASF, Bayer, and UBS-Warburg invested $650 billion in foreign direct investment (FDI) in the United States. This 4-year figure exceeds by half the entire FDI by Europe in the United States over the past 50 years. To reflect such an enormous international capital flow, the demand for the dollar in Figure 6.5 increases, causing the steep dollar apprecia- tion of 1999, 2000, and 2001 shown in Figure 6.1.

In 2002, however, following the attack of September 11, 2001, in the midst of a three- quarter U.S. recession, real interest rates in the United States fell to 0.1 percent, the low- est rate in 40 years, and then continued lower to −1.07 percent in 2003 and −1.11 per- cent in 2004. Predictably, based on this real interest rate factor, the dollar would then collapse against the euro, and it did (see Figure 6.1). In 2008–2009 real interest rates in the United States have again gone negative (see Table 6.1), and the dollar is at its lowest level ever against the euro and other major currencies.

The Role of Expected Inflation Inflationary expectations provide an important third determinant of long-term trends in exchange rates. Suppose you were entering into a long-term contract to replace the diesel engines installed in a fleet of trucks over the next three to five years. Would you be in- clined to approach and enter into negotiation with Cummins Engine, where recent

194 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

material costs have been low, cost-saving productivity increases have been substantial, and union bargaining pressure may be declining? Or would you approach a substitute supplier like Mercedes-Benz, where all those factors are reversed, suggesting the strong possibility of an upsurging inflationary trend in input costs that underlies the price you could negotiate for a German diesel engine over the next several quarters?

Cost inflation is usually compared across economies by examining an index of pro- ducer prices. From 2003 to 2008, the percentage change in producer prices in the United States was 3.2, 3.6, 4.9, 3.7, 2.7, and 2.2 percent (as shown in Table 6.1) whereas in the euro area, producer prices increased by less—that is, 1.4, 2.3, 4.1, 2.0, 2.0, and 2.3 per- cent. Clearly, the lower price in a long-term fixed price contract for replacement diesels would not be available from Cummins, the company in a country experiencing higher cost-push inflation. Consequently, export sales on U.S. traded goods like Cummins die- sels would decrease.19

In Figure 6.5, the D2003, D2005, and D2008 demands for the U.S. dollar all shift down- ward, and the dollar depreciates still further. Indeed, the relative purchasing power parity (PPP) hypothesis (discussed in the next section) holds that goods arbitrage on products like diesel engines will continue until the €/$ exchange rate adjusts downward sufficiently to reflect entirely the inflation differential. That is, the cost inflation between the United States and Europe in 2003 of 1.8 percent (3.2 percent − 1.4 percent) differential in the producer price index favoring Europe should, according to PPP, depreciate the value of the dollar by approximately 1.8 percent. And so too for every year since, until finally in 2008, the two inflation rates were approximately equal. We discuss purchasing power parity in the next section.

PURCHASING POWER PARITY When there are no significant transportation costs, legal impediments, or cultural bar- riers associated with moving goods or services between markets, then the price of each product should be the same from one international market to another. This conclusion is known as the law of one price. When the different markets represent different countries, the law of one price says that prices will be the same in each country after making the appropriate conversion from one currency to another. Alternatively, one can say that ex- change rates between two currencies will equal the ratio of the price indexes between the countries. In international finance and trade, this relationship is known as the absolute version of purchasing power parity.

Absolute purchasing power parity implies that differentially higher core price inflation in one location than another (which results in a doubling of U.S. prices for traded goods like autos, aircraft, and iPods) will ultimately result in a 50 percent depreciation of the U.S. currency. For example, if after a sustained period of U.S. price inflation, one needs $200 to buy a book that before the inflation cost $100, and if Japanese publishers will continue to print and sell that same book (in English) for an unchanged price of ¥10,000, the book will be produced in Japan and exported to the United States, driving down the dollar. How so? Such imports by Americans will necessitate a demand for Jap- anese yen to accomplish the purchase, and the supply of dollars to support these import transactions will continue to grow until the exchange rate reflecting the price of the U.S.

19Eventually, if producer cost inflation differentials between the United States and Germany persist, import- export companies that specialize in capital equipment transactions between the United States and Germany will join the surging demand for American products and buy replacement diesels cheaply in the United States for resale at a profit in Germany. This goods arbitrage activity, discussed in the next section, limits the extent to which such cost differentials can persist for long periods of time.

Chapter 6: Managing in the Global Economy 195

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

dollar declines from ¥10,000/$100 = ¥100/$1 all the way down to ¥10,000/$200 = ¥50/$. At that point, the two prices for acquiring the book will again be the same in the two economies, adjusting the book prices in the different currencies with the new FX rate. In short, goods arbitragers in global supply chains will prevent the delivered prices in the United States and Asia from remaining different for very long.

PPP Offers a Better Yardstick of Comparative Growth Purchasing power parity helps answer the difficult question of just how big the Chinese economy has become relative to Japan or the United States. To relate the ¥475T Japanese economy to the U.S. economy, one simply divides by the market exchange rate—JPY95 per USD. So, in 2009 the Japanese economy is equivalent to a $5T economy. But the same procedure is not available for China because the CNY’s “managed float” is far from a market-based equilibrium foreign exchange (FX) rate.

To address this issue of finding a commensurate yardstick to use in measuring Chi- nese GDP, one alternative is the FX rate implied by purchasing power parity (PPP). Ab- solute purchasing power parity hypothesizes that for traded goods like sandals, autos, iPods, and iron ore and given sufficient time for adjustment,

Pricelocal currency A = Pricelocal currency B × PPP Implied FXA for B.

For example, at an FX rate implied by PPP, Apple iPods sell for identical prices in the United States and United Kingdom. So, an Apple iPod Silver Classic with 120 gigabytes of memory selling on Amazon in the United States for $225 and on Amazon U.K. for £164 implies a PPP FX rate of $1.37/£:

$225US = £164 × PPP Implied FX$for£ $225US = £164 × $1:37=£:

Again, $1.37/£ is not a nominal FX rate (i.e., the British pound is commanding $1.69 this morning in New York trading). Nor is an Apple iPod the perfect traded good. Apple strives with great diligence to segment their market across countries, as do other branded product manufacturers. And there are many other qualifications of PPP that we investigate in the next section. Nevertheless, averaging a thousand such calculations of FX rates implied by purchasing power parity for U.S.-China trade, the

Example Birkenstocks for Sale Cheap! Birkenstock sandals provide a good illustration of a traded good for which pur- chasing power parity should apply. In August 2008, Birkenstocks sold for €57 in Rome and €60 in Paris. Similarly, using the nominal exchange rate of £0.6/$, London and New York prices were nearly identical at £40 and $70. The difference is too small to cover the shipping and generate an arbitrage profit. Were this not the case, enterprising arbitragers would buy cheap in one location and sell at a dis- count in the other, causing prices to converge. Birkenstock has authorized dealers in many parts of the world, and sells at a variety of price points through those dealers. Preventing counterfeit knockoffs between Singapore, where the pricing is equivalent to USD76, and Sydney, where pricing is equivalent to USD98, presents a constant struggle. Nevertheless, PPP predicts such arbitrage activity across many goods would put downward pressure on the currently very strong Australian dollar.

196 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

International Monetary Fund (IMF) calculates an implied PPP FX rate for China of 3.8 Chinese yuan to the U.S. dollar, or USD.263/CNY value of the yuan.20 In other words, the official “managed float” exchange rate of CNY6.83/USD or USD.146/CNY under- values the yuan by almost half.

The PPP implied size of the Chinese economy is therefore much bigger than $4.33T. It is closer to $7.8T, substantially larger than Japan’s $5T economy (mistakenly thought to be second largest in the world). At these PPP-adjusted FX rates, the United States is 21 percent and China is 12 percent of world GDP, and the IMF figures the United States contributed 14 percent while China contributed 30 percent to global GDP growth in re- cent years. Chinese steel production reached 510 million tons in 2009 while Europe pro- duced 110 million tons, and the United States produced just 60 million.21 China’s retail sales to domestic consumers grew at 14 percent per annum from 2004–2007 and at a 17 percent pace from 2008–2009 when the rest of the world’s consumer economy was slowing abruptly.

One reason was that China introduced the largest fiscal stimulus worldwide (13 per- cent of GDP) and sharply relaxed credit conditions in late 2008 and early 2009.22 But another reason is that the Chinese economy is on the brink of developed-country living standards as an expectation for perhaps half the population located in the eastern, coastal provinces. That would not be so significant but for the fact that China contains One Bil- lion Customers (the title of James McGregor’s widely cited 2005 book on doing business in China). Perhaps 150 million people (equal to half the population of the United States) have become “middle class” with PPP-adjusted Chinese incomes over $20,000. These Chinese households are pursuing consumption, both domestic and imported, 10-fold greater than a decade ago. In aggregate, Chinese consumption is presently 35 percent of GDP compared to 50–60 percent in most other Asian countries and 70 percent in the West. As this gap narrows, the sales opportunities for Western products adapted to the Chinese market are going to be extraordinary.

Relative Purchasing Power Parity A less restrictive form of the law of one price is known as relative purchasing power parity. The relative PPP hypothesis states that in comparison to a period when exchange rates be- tween two countries are in equilibrium, changes in the differential rates of inflation between two countries will be offset by equal, but opposite, changes in the future spot exchange rate. For example, if prices in the United States rise by 4 percent per year and prices in Europe rise by 6 percent per year, then relative PPP asserts that the euro will weaken relative to the U.S. dollar by approximately 2 percent.

The exact relative purchasing power parity relationship is

Relative PPP : S1 S0

� � =

1 + πh 1 + πf

� � [6.1]

where S1 is the expected future spot rate at time period 1, S0 is the current spot rate, πh is the expected home country (U.S.) inflation rate, and πf is the expected foreign inflation rate.

Using the previous example, if U.S. prices are expected to rise by 4 percent over the coming year, prices in Europe are expected to rise by 6 percent during the same time,

20International Monetary Fund, World Economic Outlook Database (April 2009). 21“China Takes a Hard Look at Its Steel Industry,” Wall Street Journal (October 13, 2009), p. B1. 22“Rebalancing the World Economy: China,” The Economist (August 1, 2009), pp. 65–66.

relative purchasing power parity A relationship between differential inflation rates and long-term trends in exchange rates.

Chapter 6: Managing in the Global Economy 197

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

and the current spot exchange rate (S0) is $0.60/€, then the expected spot rate for the euro in one year (S1) will be

S1=$0:60 = ð1 + 0:04Þ=ð1 + 0:06Þ S1 = $0:5887=€

= $0:60ð1 − 0:0189Þ=€ The higher European inflation rate can be expected to result in a decline in the future spot value of the euro relative to the dollar by 1.89 percent.23

The dashed lines in Figure 6.6 indicate purchasing power parity between the strongest European currency that formed the euro and the U.S. dollar, accounting for cumulative inflation in the United States and Germany from 1973–2000. Over this period, the con- sumer price index in Europe rose from 67.1 to 137.2 (a 104 percent increase), and the consumer price index in the United States rose from 49.3 to 166.7 (a 238 percent in- crease). Starting from a 1973 exchange rate of DM2.81/$, the predicted equilibrium ex- change rate in 2000 implied by the hypothesis of relative purchasing power parity would be (DM2.81/$1.00) × (204/338) = 1.70, close to the actual 1997 exchange rate of DM1.76/USD but far below the 2001 exchange rate of DM2.30/USD. Therefore, referring again to Figure 6.6, the dollar was substantially above its purchasing power parity level in 1984–1986 and again in 1999–2003. In each case, a subsequent sharp decline in value against European currencies ensued. Between 2001 and 2009, for example, the dollar has lost 54 percent of its value against the euro (see Figures 6.1 and 6.7).

Qualifications of PPP Purchasing power parity calculations can be very sensitive to the starting point for the analysis. In 2000, the ¥/$ exchange rate averaged ¥107/$, whereas in 2002 the average value of the dollar fell to ¥130/$ (see Figure 6.1). Since 2000, the United States has in- flated by 22 percent, and Japan has deflated by 2 percent. The 2000 starting point for a relative PPP calculation implies a 2009 exchange rate predicted by PPP of ¥86/$ (quite close to the actual value of ¥89/$), while the 2002 starting point implies a predicted 2009

FIGURE 6.6 Purchasing Power Parity (DM/$, 1973–2001)

2.67

1.82

2.94

2.17

1.75 1.43

1.88

1973 1980 1985 1990 1995

Purchasing power parity

Year

Pr ic

e of

U .S

. d ol

la r

(D M

/$ )

2000

2.30

2.10

23Several other parity conditions in international finance are discussed in R.C. Moyer, W. Kretlow, and J. McGuigan, Contemporary Financial Management, 10th ed. (Cincinnati: Cengage/South-Western, 2008), Chapter 21.

198 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

exchange rate of ¥110/$. Clearly, the difference is nontrivial, and many such applications of the PPP hypothesis will hinge on which year the analyst chooses to start. Figure 6.6 highlights this qualification by displaying a rather wide band of exchange rates implied by the PPP hypothesis.

Purchasing power parity has several other qualifications as well. For the full PPP ad- justments to take place in exchange rates as domestic prices inflate, the traded goods must be nearly identical in quality and use in the two economies. Cross-cultural differ- ences (e.g., the Islamic aversion to Western clothing for women or the American procliv- ity for pickup trucks and SUVs) can short-circuit these adjustments. In addition, both economies must have similar trade policies. If Europe has much higher agricultural sub- sidies and trade barriers to foreign crops than the United States, that policy may prevent the trade flows and subsequent exchange rate adjustments hypothesized by PPP. Similar qualifications apply to differences in value-added taxes and other sales taxes across econ- omies. Despite these caveats, purchasing power parity has proven to be a useful bench- mark for assessing trends in currency values.

No one would ever execute a currency arbitrage trade based on the predictions of the purchasing power parity hypothesis. Currency arbitrage is triggered by unanticipated events that generate very temporary profit opportunities lasting only several hours or a few days. The trade flows predicted by PPP in response to inflation differentials, on the other hand, are a much longer-term process requiring several quarters or even years. Companies with a substantial proportion of their sales abroad must identify these

WHAT WENT RIGHT • WHAT WENT WRONG

GM, Toyota, and the Celica GT-S Coupe24

As we saw at the start of the chapter, Toyota tends to pre- serve unit sales by slicing yen profit margins when their domestic currency strengthens. In contrast, GM sometimes even raises export prices when the dollar strengthens in order to preserve dollar profit margins on their overseas sales. From 1980 to 1984, for example, GM raised Opel prices 48 percent as the dollar strengthened 47 percent. The depreciated revenue from foreign currency sales trans- actions in Europe was almost perfectly offset by 48 percent list price increases. Of course, this implied a skyrocketing 95 percent (48 percent + 47 percent) higher price in Frank- furt, Munich, and Koln. The Opel Division of General Motors found it difficult to explain to potential German customers why asking prices should essentially double in so short a period.

Why did GM and Toyota have such different pricing and markup policies? One might suspect that costs were higher in the United States during 1980–1988. In fact, the unit labor costs in manufacturing were lower in the United States than in Japan over this period. Perhaps Toyota sought greater sales volume to realize scale econo- mies or take advantage of learning curve reductions in unit cost as cumulative volume increased. Total quality initia- tives in Japanese manufacturing have realized much- heralded cost savings on the assembly plant floor as more

vehicles pass quality inspections without requiring rework. Nissan’s plant in Smyrna, Tennessee, for example, is her- alded as the most efficient assembly line in America, with worker productivity almost 35 percent higher than the av- erage GM plant. Finally, Toyota, Honda, and Nissan cer- tainly are export-driven companies. Between 1985 and 1988, Nissan generated 50 percent of its sales abroad (fully 45 percent in the United States alone).

General Motors, in contrast, has 72 percent domestic sales, 12 percent export sales, and 16 percent sales from overseas production divisions such as Opel in Europe and Holden in Australia. Consequently, GM does not fo- cus marketing and operations planning on export sales or overseas sales. Nevertheless, every company should always analyze its import market competition. By offsetting ap- proximately half of any unfavorable exchange rate moves (involving an appreciation of the yen) by reducing export margins, all the Japanese auto manufacturers have grown at GM’s expense. From a high of 45 percent market share, GM has shrunk below 20 percent, and a still growing Toyota at 19 percent is poised to become the world’s larg- est car company.

24Based on “General Motors and the Price of the Dollar,” Harvard Business School Publishing, 1987.

Chapter 6: Managing in the Global Economy 199

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

longer-term trends in exchange rates to build into their three-year business plans, and purchasing power parity proves useful for just that purpose. For example, a realization that the dollar was well above purchasing power parity levels in 1999–2003 should have influenced production and pricing policies at Cummins Engine at the start of the twenty- first century. In the past five years (2004–2009), Cummins sales and cash flows have tri- pled because of the decade-long decline in the USD.

Being attuned to the international business environment of exchange rate trends does not allow fine tuning, but it does allow better medium-term planning of production vol- ume, proactive pricing, target markets, and segmented distribution channels, all of which may offer profit advantages. Some companies make these considerations a focus of their business plans and prosper in international markets; others are less successful.

The Appropriate Use of PPP: An Overview What, then, is the appropriate role of purchasing power parity in managing international business? First, one should be aware that PPP is a very long-run proposition. Arbitrage activities are motivated whenever the price of auto tires, biscuits, or DVDs in one econ- omy departs markedly from the price of similar goods in another economy, not far dis- tant. In such circumstances, entrepreneurs emerge to buy cheap in one location and sell dear in the other, but this goods arbitrage activity takes time. Goods arbitrage requires logistical infrastructure like freight terminals, distribution networks, reliable retail rela- tionships, and effective transnational marketing campaigns. Until all these matters can be resolved, international markets can remain somewhat segmented, thereby preventing the complete convergence of prices of identical products between the United States and Japan, between EU countries, between the United States and the United Kingdom, and even between the United States and Canada. Unlike arbitrage in financial markets, the goods arbitrage underlying PPP may take months, years, or even decades. As a result, the variance of prices for like goods across countries is larger (often 10 times larger) than the variance of prices through time within an economy.25

In the short run, therefore, when the prices of traded goods in the United States di- verge from those in the United Kingdom, nominal exchange rates may not respond to fully offset the price differentials, as predicted by PPP. Instead, price stickiness in traded goods combined with the lemming-like behavior of herds of FX speculators leads to more volatile nominal exchange rates than would characterize a fully adjusting price re- gime. In particular, nominal exchange rates may overshoot or undershoot their equilib- rium levels in adjusting to demand or monetary shocks. To avoid this problem, many analysts perform cross-border comparisons of price levels or trade statistics using pur- chasing power parity estimates for the prior 15-year period.

For example, if one observes in June 2004 that 10 feet of 0.019-gauge gutter pipe sells in do-it-yourself stores in the United States for $3.36 and that the value of the British pound is $1.80/£ (i.e., £0.55/$), it might be very misleading to calculate that identically priced gutter pipe should sell for £1.84 in the United Kingdom (i.e., $3.36 × £0.55/$). Even if the British are paying £2.10, the apparent arbitrage profit of (£2.10 − £1.84) × $1.80/£ = $0.47 per 10 feet may not be available. Consequently, rushing out to organize the distribution and export of U.S. gutter pipe for sale in United Kingdom every time the exchange rate overshoots or undershoots does not make sense. Instead, one should mul- tiply the $3.36 U.S. price by a £0.63/$ purchasing power parity value of the pound over

25C. Engel and J.H. Rogers, “How Wide Is the Border?” American Economic Review, 86 (1996), pp. 1112–1125; and “Goods Prices and Exchange Rates,” Journal of Economic Literature, 35 (1997), pp. 1243–1272.

200 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

1990–2004 (see Figure 6.1). This would imply that gutter pipe in British stores selling for as much as £2.12 would not be overpriced relative to the United States. And, in light of the confidence bounds around PPP conditions, one might consider a British price of £2.12 + or – 10 percent (£1.91–£2.33) consistent with the absence of arbitrage opportu- nities in gutter pipe.

Big Mac Index of Purchasing Power Parity26

Big Mac hamburgers sold in 120 countries around the world are as close to identical as the McDonald’s parent corporation can make them, yet individual country managers have complete discretion in setting the price. In 1974, the Big Mac sold for approxi- mately $1.00 in the United States and DM2.50 in Germany. In that same year, 2.50 Ger- man marks exchanged for one U.S. dollar in the foreign currency markets. The Big Mac index of PPP for a hamburger (i.e., DM2.50/$1.00) just equaled the actual exchange rate.

By 1996, the cost of a Big Mac in downtown Munich had risen to $4.90, slightly higher than implied by the 93 percent cumulative German inflation between 1974 and 1996 (i.e., 1.93 × DM2.50 = DM4.82). In 1996 in Atlanta, on the other hand, the Big Mac sold for $2.36, considerably lower than implied by the 219 percent cumulative U.S. inflation in 1974–1996 (i.e., 3.19 × $1.00 = $3.19). Consequently, the Big Mac index of purchasing power parity (i.e., DM4.90/$2.36 = 2.08) at the end of 1996 implied that the U.S. dollar should exchange for DM2.08. With the actual exchange rate in 1996 at DM1.76/$1 (see Figure 6.1), PPP implied that the dollar was substantially undervalued. Sustained dollar appreciation against the mark in 1997–2001 from DM1.76/$ to DM2.30/$ suggests that “burger economics” has merit.

The Big Mac relationships are not a perfect application of the relative PPP hypothesis for several reasons: (1) because the 17 percent European Union VAT tax exceeds the U.S. sales tax, (2) because downtown land rents and utilities in Munich substantially ex- ceed those in Atlanta, and (3) because the degree of fast food industry competition facing a Munich McDonald’s is lower than that facing an Atlanta McDonald’s. Also, of course, a prepared Big Mac cannot be purchased in Munich for effective resale in Atlanta; goods arbitrage with perishable commodities is infeasible. Nevertheless, such PPP measures can help evaluate trends in currency value.

When the dollar reached DM2.10/$ in 1999, the Big Mac index of PPP predicted that any additional increases in dollar value would constitute overshooting. Sure enough, as conveyed in Figure 6.1, no sooner had the U.S. dollar risen to DM2.30/$ and €1.12/$ in 2001, then it collapsed 34 percent during 2002–2004 against European currencies (i.e., to €0.80/$). Even the trade-weighted index value of the dollar declined steeply (see Figure 6.7), despite the fact that Canada, Mexico, Japan, and China are now more prominent than Germany, France, and the other European countries in U.S. trade.

Trade-Weighted Exchange Rate Index Figure 6.7 shows the value of the U.S. dollar against the currencies of the United States’ largest trading partners from 1981–2009. This trade-weighted exchange rate, sometimes called the effective exchange rate (EER), calculates the weighted average value of the dol- lar against 19 currencies, where the weights are determined by the volume of import plus

26Based on “McCurrencies: Where’s the Beef?” The Economist (April 27, 1996), p. 110; “Big MacCurrencies,” The Economist (April 9, 1994), p. 88; and “Big MacCurrencies,” The Economist (April 6, 2005).

Chapter 6: Managing in the Global Economy 201

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

export trade between the United States and each of its trading partners. The EER for the United States is therefore defined as

EER$t = ∑ i eI$itwit [6.2]

where wit is the relative proportion of total import and export trade from and to country i in time period t (e.g., 0.049 for the United Kingdom in 2000 but only 0.035 in 2008) and where eI$it is the exchange rate index for the dollar in period t against the currency of country i. For Britain, eI$£08 is calculated as the £/$ exchange rate in 2008 divided by the £/$ exchange in the base year, multiplied by 100—that is, (£0.60/$ ÷ £0.40/$) × 100 = 1.50 × 100 = 150.

From 1995–2001, the trade-weighted U.S. dollar appreciated substantially. All three factors determining long-run trends in exchange rates were involved. Real growth rates in the United States declined over this period relative to several of the United States’ largest trading partners. Real interest rates on U.S. T-bills were high and rising relative to those same trading partners. And finally, cost inflation in the U.S. producer price in- dex was at a post-World War II low relative to the United States’ largest trading part- ners. So capital flowed into the United States, and U.S. export trade rose dramatically.

As it had been for their Japanese and European competitors in the 1980s, export trade provided the engine of growth for many U.S. companies in the 1990s and in 2005–2008. Panel (a) of Figure 6.8 shows that from the 1970s to the late 2000s, the share of exports relative to U.S. GDP grew from 5 percent to 13 percent. By 1994, one-fifth of real U.S. GDP growth was attributable to exports. Between 1994 and 1997, that contribution of exports to GDP growth doubled [see Figure 6.8, Panel (b)]. Between 1998 and 2002, the U.S. export sector decayed, primarily because the value of the dollar had appreciated above purchasing power parity (see Figure 6.6). However, at its highest recent peak values of ¥134 in mid-1998, DM2.30 in mid-2000, and €1.12 in 2001, the U.S. dollar was still well below its spectacular 1985 peak of ¥238 and DM2.94. This historical

FIGURE 6.7 Trade-Weighted Effective Exchange Rate Index, U.S. Dollar (1981–2009)

100

80

60

120

140

160

Index, March 1973 = 100, shaded areas indicate recessions in the United States

Purchasing power parity

989796959493929190898887868584838281 Year

99 00 01 02 03 04 05 06 07 08 09

Source: National Economic Trends, Federal Reserve Bank of St. Louis, quarterly.

202 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

perspective can prove useful, because it took only a 34 percent depreciation of the dollar in 2001–2004 to restart the U.S. export sector. Figure 6.8, Panel (b) shows that in 2004– 2009, exports have steadily increased to exceed 13 percent of GDP and are again contrib- uting the majority of U.S. real GDP growth.

Example Are Dirty Dishes Mobile? Dixan, Joy, Dawn, and Generic Patti Scala from Scala S.p.A. Sometimes purchasing power parity (PPP) fails to characterize the price of identi- cal items for sale in different currencies because some complementary factors in consumption or production are immobile. Unique land is one immobile factor; dirty dishes are another. In July 2001, after a 35 percent two-year appreciation of the U.S. dollar to 2,284 lire and 1.18 euros, Tuscan villagers found themselves serv- ing mountains of primi patti (first dishes) of various pastas to throngs of U.S. tour- ists. As a result, dirty dishes piled up high in many Italian sinks.

As elsewhere, dirty dishes in Italy eventually require patti scala (dish soap), and Scala S.p.A. from Castrocielo, France, supplies lemon-scented generic Patti Scala throughout the grocery stores of Tuscany. Generic Patti Scala at 1,900 lire com- petes directly against Dixan, a popular branded dish soap at 2,600 lire, both in reg- ular size 750-ml plastic containers. What is extraordinary at first glance about these Italian prices is that in July 2001, the 1,900 lire and 2,600 lire equated to $0.84 and $1.14, respectively, much less than the price of equivalent dish soaps in U.S. stores (e.g., Harris Teeter sells 750 ml of generic dish soap for $1.80; Joy, Palmolive, and Dawn sell for $1.99, $2.49, and $2.79 in roughly comparable sizes (see Table 6.2). Does this imply that Joy, Palmolive, and Dawn should anticipate massive erosion of their U.S. market share because of an invasion of lower-priced Dixan imports from Italy?

The answer is definitely no, for three reasons. First, of course, customers of Joy, Palmolive, and Dawn in the United States do not recognize the Dixan brand. A massive advertising and promotional campaign by Dixan would be necessary to

(Continued)

TABLE 6.2 COMPETING DISH SOAP PRODUCTS IN THE UNITED

STATES AND ITALY

BRANDED PRODUCTS JUNE 2001 PRICE

UNIT VOLUME

(ml)

Joy $1.99 740

Palmolive $2.49 739

Dawn $2.79 740

Dixan (June 2001) $1.14 (2,600 lire) + 30% shipping = $1.45 750

Dixan (June 1999) $1.56 (2,600 lire) + 30% shipping = $2.03 750

GENERIC PRODUCTS

Harris Teeter Dish Soap $1.80 828

Patti Scala (June 2001) $0.84 (1,900 lire) + 30% shipping = $1.09 750

Patti Scala (June 1999) $1.15 (1,900 lire) + 30% shipping = $1.48 750

Chapter 6: Managing in the Global Economy 203

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

INTERNATIONAL TRADE: A MANAGERIAL PERSPECTIVE

Shares of World Trade and Regional Trading Blocs The United States is both the second largest exporter and the largest importer in the world’s economy. Thus, the United States generates the largest share of the $32.5 trillion in bilateral world trade (10.6 percent), comprising 8 percent of all exports and 13 percent of all imports. The next nine countries with the largest shares of world trade in 2008 were Germany (8.2 percent), China (7.9 percent), Japan (5 percent), France (4 percent), the Netherlands (4 percent), Italy (3 percent), United Kingdom (3 percent), Belgium (3 per- cent), and Canada (2.7 percent). Although these largest trading nations are predominantly

overcome the brand name barriers to entry in the U.S. market. International arbi- trage in goods and full attainment of PPP is partially prevented by the customer- switching costs imposed by these brand names. Most branded products compete in at least partially segmented domestic markets.

Second, even in the absence of branded products, PPP may fail to hold in the generic dish soap market because dirty dishes are immobile. Although the chemical ingredients, lemon scents, and hand softeners in Scala S.p.A.’s Patti Scala are virtu- ally identical to those in Harris Teeter Dish Soap, and despite the availability of surplus Patti Scala at 1,900 lire in Italy, the immobile dirty dish complements in consumption are located in the United States. One must therefore incorporate a transportation cost into these price comparisons. A delivered price of Patti Scala in the United States would be 1,900/2,284 lire = $0.84, plus perhaps 30 percent shipping cost—that is, $1.09. Transportation costs explain a substantial portion of the observed price differential between identical products priced in different currencies.

Still, one should ask, why do the prices of generic dish soap products shipped into another currency’s domestic market in the summer of 2001 differ by as much as $1.09 to $1.80? The answer lies in recognizing that purchasing power parity is a hypothesis about long-term price dynamics. Exchange rates often overshoot/under- shoot their equilibrium levels, and consequently the ratio of retail prices in one economy to the exchange rate-adjusted retail prices in another economy should be computed over several years. For example, two years earlier the Italian lira was much stronger than in the summer 2001; specifically, in June 1999, the U.S. dollar exchanged for just 1,665 lire, versus 2,248 lire in June 2001. With the suggested retail prices of Patti Scala and Dixan the same in 1999 as in 2001, Patti Scala at (1,900/1,665 lire = ) $1.14 + 30 percent transportation cost = $1.47 for 750 ml (i.e., $0.0020 per ml). This unit price in Italy is closely aligned with Harris Teeter’s unit price of $0.0022-per-ml for 828 ml of generic dish soap at $1.80. Furthermore, Dixan at (2,600/1,665 lire =) $1.56 + 30 percent transportation cost = $2.03 for 750 ml is nearly identical to Joy’s $1.99 price for 740 ml of branded dish soap.

In conclusion, successful brand name campaigns, the immobility of comple- ments, and temporary exchange rate overshooting/undershooting may create sig- nificant gaps between the final product prices of like products sold in different currencies. Nevertheless, a judicious use of PPP calculations can be very revealing.

204 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Western-developed economies, the next six largest trading nations are predominantly rap- idly developing Asian economies: South Korea (2.6 percent), Hong Kong (2.3 percent), Russia (2.3 percent), Spain (2 percent), Singapore (2 percent), Mexico (1.9 percent), Tai- wan (1.5 percent), and India (1.4 percent). Altogether, the World Trade Organization (WTO) includes 149 nations who have agreed to share trade statistics, coordinate the lib- eralization of trade policy (i.e., the opening of markets), and cooperatively resolve their trade disputes in accordance with WTO rules and procedures.

Figure 6.9 shows that over the past two decades the import and export trade flows to and from the United States have grown to equal fully 31 percent of GDP; American exports are 13 percent of GDP or about $1.8 trillion worth of goods and services, of which 53 per- cent now goes to emerging markets. Some nations are much more export driven. In 2009, Germany and China both exported nearly $1.5 trillion but that represents a whopping 42 percent of German and 28 percent of Chinese GDP. British exports are 27 percent, Mex- ican exports are 35 percent, Canadian exports are 43 percent, South Korean exports are 53 percent, Malaysian exports are 93 percent, and Belgian exports are 177 percent of GDP!

Most nations continue to protect with tariffs and other trade barriers some infant or politically sensitive industries. France, for example, remains a largely agricultural polity and therefore lowers its agricultural subsidies only after great hand-wringing and ex- tended periods of tough negotiations with its European neighbors. The WTO’s Doha round of tariff-reduction negotiations was suspended in mid-2006 primarily because the EU refused to give up its €47 billion common agricultural policy (CAP). France receives €10.6 billion of these CAP subsidies, 80 percent of which go not to rustic French family farms but rather to large French landholders for intensive industrial farming. The United States is hardly blameless as agricultural subsidies increased from 19 percent to 22 per- cent of total agricultural value, while the EU reduced agricultural subsidies from 47 per- cent to 35 percent over the period 1993–2003. Ethanol, sugar, and corn price support programs for small family farmers are embedded in American populist politics.

Today, the United States, the European Union (EU), and Canada have some of the lowest tariffs equivalent to 3 percent to 4 percent on imports. Mexico and India have

FIGURE 6.8 The Growth of the Export Sector in the U.S. Economy

1997 98

Growth of export sector in U.S. economy

Export contribution Real GDP growth20

16

12

6

4 1968 72 76 80 84 88 92 96 00 04 06 08 99 00

4

3

2

1

0

Panel (b)

Pe rc

en t

Pe rc

en t

Percent of GDP, 1968–2009

Panel (a)

03 0401 0502 07 08 0906

Source: U.S. Department of Commerce, Bureau of Economic Analysis.

Chapter 6: Managing in the Global Economy 205

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

some of the highest, equivalent to 12 percent to 15 percent. Fortunately, regional trading blocs like the EU and the North American Free Trade Agreement (NAFTA) have been highly successful in removing trade barriers, negotiating multilateral reductions in tariffs, and promoting free trade as a mechanism of peaceful competition between nations.

Across the world economy, six such regional trading blocs have emerged (see Figure 6.10). In South America, Argentina, Brazil, Paraguay, and Uruguay have formed one trading block (MERCOSUR) whose import-export merchandise trade doubled from 1996 to 2006 and is now approaching $600 billion, 80 percent of which is between the member countries (see Figure 6.11). The Andean group (Peru, Colombia, Bolivia, and Ecuador) has formed another trading bloc; both are attempts to mirror the NAFTA free trade area of Canada, the United States, and Mexico.27 The Brazilian ($1.6 trillion), Canadian ($1.4 trillion) and Mexican ($1.2 trillion) economies are comparable in size, about one-tenth the size of the U.S. economy. Brazil, Mexico, and Argentina have large industrial bases protected by 14–16 percent import tariffs, whereas Paraguay, Bolivia, and Guatemala are commodity-based economies employing 6–8 percent import tariffs. The United States and Canada have 3.5 percent average import tariffs.

Trade disputes between Brazil and the United States have arisen over Brazilian steel, sugar, frozen orange juice, and ethanol exports to the United States. In 2009, Brazil

FIGURE 6.9 U.S. Imports and Exports as a Percentage of GDP Pe

rc en

t of

w or

ld t

ra de

( %

)

25

U.S. imports + exports as a percentage of GDP

1985 86 87 88 89 90 91 92 93 94 95 96 97 98 99 0300 04 0501 02 08 0906 07

28

31

22

20

18

Source: U.S. Department of Commerce, Bureau of Economic Analysis.

27Seven Southeast Asian (ASEAN) and 16 trans-Pacific economies including Japan and Mexico (APEC) have also formed trading blocs.

206 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

achieved oil independence through an extensive sugar cane-based ethanol industry pro- ducing over 5 billion gallons a year. Despite being the largest ethanol producer in the world with 6.5 billion gallons produced, the United States has effectively barred competi- tion by imposing a protective (“infant industry”) tariff on Brazilian ethanol. Other pro- ducers of ethanol are the EU with 570 million, China with 486 million, and Canada with 211 million gallons produced annually.28

Comparative Advantage and Free Trade Within a regional trading bloc like EU, NAFTA, MERCOSUR, or APEC, each member can improve its economic growth by specializing in accordance with comparative advan- tage and then engaging in free trade. Intuitively, low-wage countries like Spain, Mexico, Puerto Rico, China, and Thailand enjoy a cost advantage in the manufacture of labor- intensive goods such as clothes and the provision of labor-intensive services like sewing or coupon claims processing. Suppose one of these economies also enjoys a cost advan- tage in more capital-intensive manufacturing like auto assembly. One of the powerful in- sights of international microeconomics is that in such circumstances, the low-cost economy should not produce both goods, but rather it should specialize in that produc- tion for which it has the lower relative cost, while buying the other product from its higher-cost trading partner. Let’s see how this law of comparative advantage in bilateral trade reaches such an apparently odd conclusion.

FIGURE 6.10 Regional Trading Zones (Percentage of World Trade, 2008)

Australia Brunei Canada Chile China Hong Kong Indonesia Japan Malaysia

1.2 0.1 2.7 0.4 7.9 2.3 0.8 4.8 1.1

Mexico New Zealand P.N.G. Philippines Singapore South Korea Taiwan Thailand U.S.

1.9 0.3 0.0 0.4 2.0 2.6 1.5 1.1

10.6

APEC

Brunei Indonesia Malaysia Philippines Singapore Thailand Vietnam

0.1 0.8 1.1 0.4 2.0 1.1 0.2

ASEAN

Argentina Brazil Paraguay Uruguay Bolivia Chile

0.4 1.2 0.1 0.1 0.3 0.3

MERCOSUR

Austria Belgium/Lux. Britain Denmark Finland France Germany Greece Ireland Italy Netherlands Portugal Spain Sweden

1.1 2.9 3.4 0.7 0.6 4.0 8.2 0.3 0.8 3.4 3.7 0.5 2.1 1.1

EU

Canada Mexico U.S.

2.7 1.9

10.6

NAFTA

Source: World Trade Organization; “Special Report on America’s Economy,” The Economist, April 3, 2010.

28“ADM Makes Ethanol Push into Brazil with Venture,” Wall Street Journal (November 5, 2008), p. B1.

law of comparative advantage A principle defending free trade and specialization in accordance with lower relative cost.

Chapter 6: Managing in the Global Economy 207

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Consider the bilateral trade between the United States and Japan in automobile car- buretors and computer memory chips. Suppose the cost of production of carburetors in Japan is ¥10,000 compared to $120 in the United States. At an exchange rate of ¥100/$, the Japanese cost-covering price of $100 is lower than the U.S. cost-covering price of $120. Suppose, in addition, that memory chips cost ¥8,000 in Japan compared to $300 in the United States. Again, the price of the Japanese product (i.e., $80) is lower than the price of the U.S. product. Japan is said to enjoy an absolute cost advantage in the manufacture of both products. However, Japan is 83 percent (i.e., $100/$120) as expen- sive in producing carburetors as the United States while being only 27 percent (i.e., $80/ $300) as expensive in producing memory chips. Japan is said to have a comparative ad- vantage in memory chips and should specialize in the manufacture of that product.

The gains from specialization in accordance with comparative advantage and subse- quent trade are best demonstrated using the real terms of trade. Real terms of trade identify what amounts of labor effort, material, and other resources are required to pro- duce a product in one economy relative to another. In Japan, the manufacture of mem- ory chips requires the sacrifice of resources capable of manufacturing 0.8 carburetors (see Table 6.3), whereas in the United States the manufacture of a memory chip requires the sacrifice of 2.5 carburetors. That is, Japan’s relative cost of memory chips (in terms of

FIGURE 6.11 Gross Domestic Product, Exports as a Percentage of GDP, and Average Import Tariffs for MERCOSUR Countries

Bolivia

Chile

Uruguay $92 bn, 58%, 12.3%

Argentina $328 bn, 45%, 13.5%

Paraguay $16 bn, 111%, 8%

Brazil $1.61 trn, 26%, 14.3%

French Guiana

Surinam

Guyana

Venezuela

Colombia

Ecuador

Peru

absolute cost advantage A comparison of nominal costs in two locations, companies, or economies.

real terms of trade Comparison of relative costs of production across economies.

208 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

carburetor production that must be forgone) is less than a third as great as the relative cost of memory chips in the United States. On the other hand, U.S. carburetor produc- tion requires the resources associated with only 0.4 U.S. memory chips, while Japanese carburetor production requires the sacrifice of 1.25 Japanese memory chips. The U.S. rel- ative cost of carburetors is much lower than that of the Japanese. Said another way, the Japanese are particularly productive in using resources to manufacture memory chips, and the United States is particularly productive in using similar resources to produce carburetors. Each country has a comparative advantage: the Japanese in producing mem- ory chips and the United States in producing carburetors.

Assess what happens to the total goods produced if each economy specializes in pro- duction in accordance with comparative advantage and then trades to diversify its con- sumption. Assume that the United States and Japan produced one unit of each product initially, that labor is immobile, that no scale economies are present, and that the quality of both carburetors and both memory chips is identical. If the Japanese cease production of carburetors and specialize in the production of memory chips, they increase memory chip production to 2.25 chips (see Table 6.3). Similarly, if the United States ceases pro- duction of memory chips and specializes in the production of carburetors, it increases carburetor production to 3.5 carburetors. In these circumstances, the United States could offer Japan 1.5 carburetors for a memory chip, and both parties would end up unambig- uously better off. The United States would enjoy a residual domestic production after trade of 2.0 carburetors plus the import of one memory chip. And the Japanese would enjoy a residual domestic production after trade of 1.25 memory chips plus the import of 1.5 carburetors. As demonstrated in Table 6.3, each economy would have replaced all the products they initially produced, plus each would enjoy additional amounts of both goods—that is, unambiguous gains from trade.

Import Controls and Protective Tariffs Some nations reject free-trade policies and instead attempt to restrain the purchase of foreign imports in order to expand the production of their domestic industries by

TABLE 6.3 REAL TERMS OF TRADE AND COMPARATIVE ADVANTAGE

ABSOLUTE COST, U.S. ABSOLUTE COST, JAPAN

Automobile carburetors $120 ¥10,000

Computer memory chips $300 ¥8,000

RELATIVE COST, U.S. RELATIVE COST, JAPAN

Automobile carburetors $120/$300 = 0.4 Chips ¥10K/¥8K = 1.25 Chips

Computer memory chips $300/$120 = 2.5 Carbs ¥8K/¥10K = 0.8 Carbs

GAINS FROM TRADE, U.S. GAINS FROM TRADE, JAPAN

Initial goods 1.0 Carb + 1.0 Chip 1.0 Carb + 1.0 Chip

After specialization:

Carburetors produced (1.0 + 2.5) Carbs 0

Memory chips produced 0 (1.0 + 1.25) Chips

Trade +1.0 Chip +1.5 Carb

−1.5 Carb −1.0 Chip

Net goods 2.0 Carbs + 1.0 Chip 1.5 Carbs + 1.25 Chips

Chapter 6: Managing in the Global Economy 209

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

artificially raising the price of substitute foreign products using protective tariffs. Figure 6.12 shows that Hong Kong, the United States, and Australia have some of the lowest protective tariffs while Mexico and India have some of the highest. To further bar im- ports, some nations also impose direct import controls like a maximum allowable quota of a specific type of foreign import. Once in the 1990s, to preserve American manufacturing jobs, the U.S. Congress imposed so-called voluntary import restraints (VIRs) on Japanese autos. In response, Toyota and Honda built assembly plants all over the United States, and laid-off GM, Ford, and Chrysler workers went to work build- ing Camrys and Accords. The same U.S. policy has now been proposed against Chinese- manufactured textiles.

National income is typically reduced by such import barriers. One reason why is be- cause trading partners who experience export growth often purchase more American- manufactured imports. As the Japanese export-driven economy accelerated in the past decade, Japanese households increased import consumption of fashion merchandise from Seven for Mankind, Apple, Gap, and Prada by 40 percent. Similarly, in this decade the most popular upscale car for a burgeoning class of new business owners in Shanghai is a Buick LeSabre. In 2010, GM is planning to export the Cadillac CTS to China and brand it as a Buick. A second reason for higher national income when import barriers are avoided arises because import controls lead inevitably to a reduction in the demand of a nation’s currency in the foreign exchange market. For example, when some Ameri- can households are prohibited from completing import purchases of Japanese- manufactured Toyotas and Hondas they would otherwise have bought, those households fail to request the Japanese yen they would have needed to accomplish the import pur- chases. As we studied earlier in the chapter, this reduced demand for the Japanese yen

FIGURE 6.12 Trade-Weighted Tariffs, 2008

Hong Kong

0

Percent

3 6 9 12 15

Turkey

United States

Australia

Canada

EU

Japan

China

Russia

South Korea

Brazil

Mexico

India

Source: World Bank database, 2009.

210 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

results in a higher FX value of the USD against the JPY, which in turn retards the U.S. export sector, worsening the trade balance (exports minus imports), thereby decreasing U.S. national income. Better to resist import controls, while pressing our newest big trad- ing partner China to replace their managed float with a market-determined flexible ex- change rate.

In short, free trade and open markets offer the prospect of higher national income. The World Bank estimates that developing countries with open economies grow by 4.5 percent per year, whereas those with import controls and protective tariffs grow by only 0.7 percent per year. Rich country comparisons also favored free trade: 2.3 percent to, again, 0.7 percent. In the 1990s, this gap widened still further. Those developing coun- tries whose import plus export trade as a percentage of GDP ranked in the top 50 per- centiles of developing countries had GDP growth per person of 5 percent. Those in the bottom 50 percentiles saw GDP growth per person actually shrink by 1 percent.29

Clearly, globalization and trade enhance prosperity, even in the lesser developed countries.

There are several valid arguments for trade restrictions (import quotas or tariffs): (1) to protect infant industries until they reach minimum efficient scale, (2) to offset govern- ment subsidies provided to foreign competitors with one’s own countervailing duties, and (3) to impose antidumping sanctions for foreign goods sold below their domestic cost. Chilean salmon, Argentine honey, and Brazilian frozen orange juice concentrate and slab steel have all been subject to U.S. antidumping duties in recent years. The Brazilians, in particular, claim that their export prices simply reflect a countervailing price reduction relative to their domestic cost in order to offset the enormous $30 billion in farm subsi- dies that the United States makes available to its agricultural sector. The United States does spend $20,000 per full-time farmer to subsidize agricultural products (third behind Switzerland’s $27,000 and Japan’s $23,000).

In April 2004, The WTO’s court agreed that cotton subsidies in the United States and sugar subsidies in EU protectorate Caribbean nations violated international trade rules. The WTO has undertaken as part of its Doha round of trade talks to begin sorting out these claims and counterclaims on agricultural subsidies and steel duties. Whether the United States will ultimately be sustained in imposing antidumping sanctions on $8 bil- lion of steel or whether Brazil will be sustained in their countervailing duties to offset the massive U.S. agricultural subsidies remains to be seen.

The Case for Strategic Trade Policy Although the logic of free trade has dominated academic debate since 1750, and the twentieth century saw the repeal of many import controls and tariffs, a few exceptions are worth noting. The WTO has very effectively spearheaded the negotiation of mutual trade liberalization policies. However, unilateral reduction of tariffs when trading part- ners stubbornly refuse to relax import controls or open their domestic markets seldom makes sense. The United States has found it necessary to threaten tariffs on Japanese consumer electronics, for example, in order to negotiate successfully the opening of Jap- anese markets to U.S. cellular phones and computer chips. This threat bargaining and negotiated mutual reduction of trade barriers illustrates the concept of “strategic trade policy.”

In the spring of 1999, continued EU import controls on U.S.-based Dole and Chiquita bananas from Central America led the United States to impose WTO-sanctioned duties on $180 million of European products from Louis Vuitton plastic handbags and British

29“Globalization, Growth and Poverty,” World Bank Report (December 2001).

Chapter 6: Managing in the Global Economy 211

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example President Bush’s 2002 Steel Tariffs and President Obama’s 2009 Tire Tariffs: Justified Sanctions or Hypocritical Protectionism?30

Despite its heavy weight and high transportation cost, steel is sourced from facto- ries around the globe. At $350/ton, sheet steel from Brazil and South Korea can be transported to the United States for $70/ton shipping cost and still sell $55 below the $475/ton cost of the big integrated hot-rolled steel producers in the United States like Bethlehem and Republic Steel (see Table 6.4). As a result, 15 of 17 such integrated U.S. producers who convert iron ore in blast furnaces using 160,000 United Steel Workers (USW) are operating under bankruptcy protection from their creditors.

Only AK Steel, U.S. Steel, Nucor, and other mini-mill producers are profitable. Nucor employs one-third as many workers as the integrated mills, using less ex- pensive electric arc furnaces. In addition, because Nucor starts with scrap metal, not iron ore, their principal input cost is highly correlated with final steel sheet and steel slab prices, a highly effective internal hedge. For example, when slab steel prices sank from $300 to $200 recently, the price of Nucor’s scrap steel input plummeted to all-time lows as well.

Yet, the USW is a powerful political lobby, and U.S. International Trade Com- mission hearings plus WTO antidumping rules and dispute resolution procedures provided the vehicle for President Ford and the second President Bush to consider imposing tariff sanctions to protect the domestic steel industry. In March 2002, Bush decided to impose quotas on Brazilian steel and 30 percent tariffs on steel slab and steel sheet from several other nations for three years. Citing credit, land, and energy subsidies, President Obama imposed a similar punitive tariff of 35 per- cent on light-truck and auto tires from China in 2009. Predictably, such trade

(Continued)

TABLE 6.4 SHEET STEEL, $ COST PER TON

U.S. integrated makers Japanese integrated German integrated

Brazil

U.S. Mini-mills

South Korea

Former Soviet republics

March 1998 April 2001

$250 300 350 400 450 500

Source: World Steel Dynamics, 2002.

internal hedge A balance sheet offset or foreign payables offset to fluctuations in foreign receipts attributable to exchange rate risk.

212 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

cashmere sweaters to French Roquefort cheese and foie gras. The British and French for- mer colonies in the Caribbean like St. Lucia enjoy approximately $150 million in banana farm profits as a result of the EU’s quota limiting the importing of U.S. bananas into Europe. But the cost to EU consumers has been estimated at $2 billion in higher-priced fruit. Despite this apparently lopsided cost-benefit analysis, it took strategic trade policy by the United States to induce the EU to relax its banana import controls.

Increasing Returns Another motivation for strategic trade policy arises in markets where domestic producers encounter increasing returns. Suppose the Boeing Corp. and Airbus find that learning curve advantages in airframe manufacturing offer a 1 percent reduction in variable cost for every market share point above 30 percent. A firm with a 40 percent (or 50 percent)

restrictions increase the cost of motor vehicles and steel construction in the United States. For example, Precision Technologies of Houston estimated it would raise their costs $3 million on $57 million in annual sales of high-stress steel tubing for the drilling industry.

30Based on “Rust Never Sleeps,” The Economist (March 9, 2002), p. 61; “U.S. Companies Cry Foul,” Wall Street Jour- nal (March 19, 2002), p. A2; “U.S. Protectionism Imperils Free Trade Talks with Latins,” Wall Street Journal (March 20, 2002), p. A15; “Brazil Claims Victory over E.U.,” Wall Street Journal (August 20, 2004), p. A1; and “Airbus Ruling Fuels Critics,” Wall Street Journal (September 8, 2009), p. B4.

Example Intel’s Chip Market Access Improves in Japan31

In the mid-1980s, Japanese semiconductor producers grabbed much of the world market in the dynamic random access memory chips (DRAMs) used in personal computers. Through a combination of very aggressive pricing and an almost closed domestic market, Japanese semiconductors achieved massive economies of scale and unit costs far below those in the United States. When the retail price of PCs declined steeply in 1986, the American and European manufacturers like Intel ceased production of basic memory chips.

That same year, an international dumping complaint filed at the WTO proved that many Japanese chips had been sold in the United States and Europe below cost. The U.S. Department of Commerce then imposed punitive tariffs on Japanese chips, laptop computers, and televisions. To avoid these higher tariffs, the Japanese agreed to open access to their domestic market and set a 20 percent target for for- eign memory chip sales in Japan. Intel Corp., Siemens, and other producers ex- panded, and the Japanese producers Sharp and Toshiba cut production. When the chip access accords expired in 1995, the Japanese, U.S., and German semicon- ductor manufacturers formed joint ventures to co-design and jointly manufacture subsequent generations of flash memory chips. In this instance, strategic trade pol- icy opened markets and expanded the national income of all the participating nations.

31Based on “America Chips Away at Japan,” The Economist (March 27, 1993); and “Foreign Chip Sales Up in Japan,” Financial Times (December 16, 1994).

Chapter 6: Managing in the Global Economy 213

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

market share of the world output of smaller wide-body aircraft (like the Boeing 737) will experience variable costs only 90 percent (or 80 percent) as great as smaller competitors. These circumstances are very rare indeed in the industrial sector of the economy, since they imply that diminishing returns in production are more than offset by learning curve advantages at higher output. However, where such circumstances exist, the United States, Europe, Japan, and now China use industrial policies to jump-start the preemptive devel- opment of dominant companies using public subsidies to research and development.

Network Externalities The information economy has a higher incidence than the industrial economy of these increasing returns phenomena. Frequently, in the information economy context, cost re- ductions at larger market share are associated with externalities in the installation of a network or the adoption of a technical standard. As the installed base of Windows soft- ware expands, Microsoft finds it increasingly less difficult to convince new customers to adopt their product. Computer users find it much easier to exchange documents and ex- plain new applications if their coworkers and customers employ the same operating sys- tem. As a result, the marketing cost to secure the next adoption from a marginal buyer actually declines, the larger the market share that Windows attains. The same thing holds for Apple’s operating system. The greater the installed base, the more software the independent programmers will write for use on a Mac. The more software available for use on a Mac, the lower the variable cost to Apple of successfully marketing the next Apple laptop.

FREE TRADE AREAS: THE EUROPEAN UNION AND NAFTA Free trade and increased specialization in accordance with comparative advantage has the relatively low-wage Spaniards and Portuguese assembling high value-added German components for BMWs and Blaupunkt radios. Hungarian factories turn out components for Polish plants assembling buses that are sold across Europe. Reduced trade barriers at

Example Microsoft and Apple: A Role for Protective Tariffs? Should strategic trade policy in the United States protect a company with the pos- sibility of achieving increasing returns? Microsoft and Apple pose a good case in point. Without import controls and protective tariffs, Microsoft successfully achieved a larger dollar volume of export trade than any other American company. Had the U.S. competitor been Apple alone, a serious question would have arisen. Could Apple have secured a 30 percent-plus market share in the absence of import controls and protective tariffs against a foreign Microsoft?

Is it an appropriate role of the government to pursue such cost advantages for one domestic company or an alliance of domestic companies? Or should govern- ment pursue the consumer interest of lowest prices wherever that product is pro- duced? These are the questions hotly debated in strategic trade policy today. The case exercise on reciprocal protectionism at the end of Chapter 13 looks at these strategic trade policy questions in the context of tactical choices of Boeing and Airbus.

214 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

borders have cut transportation time. The English Channel ferry now unloads in 15 min- utes rather than the previous 1.5 hours, and yogurt from Nestlé’s subsidiary in Birming- ham, England, now speeds across Europe to target customers in Milan in 11 hours rather than the previous 38. Reduced intra-European tariffs on foodstuffs, beer, wine, and autos have markedly reduced the cost of living. Although employee-paid social security taxes differ widely from 22 percent of earnings in France to 15 percent in Germany and just 3 percent in Spain, once wide differences in value-added taxes have been reconciled at uniform 17 percent rates in most cases.

Few pan-European marketing plans exist. The peak penetration of TV into Spanish households (20 percent viewership) occurs at 2 to 4 in the afternoon. Only 8 percent of Spanish households are tuned in from 6 to 8 in the evening when the 22 percent peak British audience is “watching the telly.” The Spanish consider packaged pet food a luxury and purchase yogurt through the pharmacy. Milanese brag about overpaying for a Sony television, while Munich shoppers search for days to find 5 percent discounts on fashion clothing or appliances. In short, segmented markets characterize both the original EU-12 and the much enlarged EU-27.

This is no more easily demonstrated than by examining the price variation for stan- dard goods across the EU-15 enlarged by 12 new members in 2004–2007. In Table 6.5, at the origin of the Common Market in 1992, Coca-Cola and Heinz Ketchup were approx- imately twice as expensive in Denmark and Spain as in Belgium and the United

TABLE 6.5 PRICE DIFFERENTIALS IN EUROPE

LARGEST MEAN DIFFERENCE BY COUNTRY HIGH LOW

1992 Coca-Cola (1.5 L) Ecu 0.69 (Belgium) Ecu 1.45 (Denmark)

Heinz Ketchup Ecu 0.86 (UK) Ecu 1.92 (Spain)

Clothes Washer Ecu 407 (UK) Ecu 565 (Italy)

Portable TV DM434 (Germany) DM560 (Italy)

VCR DM1383 (Germany) DM1873 (Spain)

1998 Big Mac Ecu 1.75 (Spain) Ecu 2.10 (Belgium)

Ford Mondeo DM32,000 (Spain) DM48,000 (Germany)

2004 Compact Disk €13.50 (France) €21.80 (Ireland)

Pampers €6.75 (Hungary) €21.00 (Denmark)

Big Mac Meal €2.80 (Estonia) €8.80 (Norway)

Coca-Cola €0.65 (Lithuania) €2.50 (France)

Movie Ticket €4.00 (Lithuania) €15.00 (Britain)

Milk (1 G) €2.20 (Czech Rep.) €4.86 (Norway)

% STANDARD DEVIATION ACROSS EUROPE, EURO-11 PRICE INDEX, PRETAX

1998 Household Insurance 51%

Coca-Cola by the glass 29%

Local Phone Service 25%

Yogurt 20%

Gasoline 14%

Levi 501 Jeans 10%

VW Golf 5%

Source: Financial Times, The Economist, various issues.

Chapter 6: Managing in the Global Economy 215

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Kingdom. Twelve years later in 2004, food products like Coca-Cola and milk remained 100 percent more expensive in Norway than in Hungary or the Czech Republic. Appli- ances and consumer electronics can be easily bought in one location and sold in another. Hence in 1992, clothes washers and portable TVs differed in price less than 20 percent.

As one would expect, VW Golf cars and other easily arbitraged goods like Levi 500 jeans exhibit the smallest price variance across the Union, and immobile services like in- surance and perishable food like soft drinks by the glass and yogurt exhibit the highest price variance.

Optimal Currency Areas In 1999, eleven European currencies were replaced by the euro in a single currency union. But one of the original E-12 (Britain) decided to opt out, as have Sweden and Denmark subsequently. Today, none of the new Eastern European members of the much enlarged E-27 free trade area has been invited to join the euro’s currency union. Why is that?

How far and wide a single currency should be adopted as the official monetary unit across a free trade area depends on a complex mix of economic, social, and political fac- tors. The benefits of a single currency are the avoidance of exchange rate risk on intra- regional trade and the associated hedging costs plus a massive reduction in FX conversion costs. Companies such as the German pharmaceutical firm Hoechst AG estimates that it saves over €6 million annually in covered forward contract costs by not needing to lay off risk exposure on European currencies. Hedging does cost on average 5 percent of the value at risk, and the total value of world trade equals $32 trillion per year, so conceiv- ably $1.5 trillion in hedging costs are at issue worldwide. In addition, just the mere act of exchanging foreign currency, which once required 1 in every 200 full-time employees in Europe, is now totally avoided.

The boundary of an optimal currency area hinges on three factors: the magnitude of intraregional trade, the mobility of labor, and the correlation of macroeconomic shocks between countries within the proposed currency union.32

Intraregional Trade Every nation of the European Union trades more with other EU members than with the rest of the world, and in Belgium, the Netherlands, and Ireland trade makes up a major- ity of their GDP. Italy, Germany, and France do 55 to 60 percent of their total trade with other European countries, and Spain does 70 percent. Ireland, Portugal, and Benelux do close to 80 percent. Even Britain trades more with their regional trading bloc partners than with the rest of the world combined.

Mobility of Labor With monetary policy constrained by the need for credibility in fighting inflation, and with fiscal policy constrained by rigid guidelines for membership in the currency union, labor mobility must adjust rapidly to stabilize swings in unemployment caused by local- ized market conditions. For example, if Italy is in a slump, Germany and France are growing slowly, and Ireland and Portugal are booming, a common monetary policy com- bined with little fiscal autonomy requires that labor move quickly from Continental

free trade area A group of nations that have agreed to reduce tariffs and other trade barriers.

value at risk The notional value of a transaction exposed to appreciation or depreciation because of exchange rate risk.

32See O. Blanchard and J. Wolfers, “The Role of Shocks and Institutions in the Rise of European Unemploy- ment,” Economic Journal (March 2000); R. Mundell, “Threat to Prosperity” Wall Street Journal (March 30, 2000), p. A30; and M. Chriszt, “Perspectives on a North American Currency Union,” Fed Atlanta Economic Review no. 4 (2000), pp. 29–36.

216 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Europe to the fringes of the EU. Although such mobility occurs easily in the United States where a family can find schools, houses, and language in one state very similar to a state they leave, the same cannot be said in Europe. Relocation costs in the United States of $25,000 compare to an estimated $75,000 in Europe.

Cultural differences are enormous from one corner of the EU to another. While Eu- ropean professionals have trained in foreign capitals and accepted diversity in manner- isms and cultural practices for centuries, working-class Europeans tend to be more culturally prejudicial. Thus, even those who would take the initiative to cross national and cultural borders in pursuit of transitory jobs typically will experience less than full acceptance as “guest workers.” Expansion of the EU free trade area into Eastern Europe has compounded this problem of the labor mobility required to justify a currency union.

Correlated Macroeconomic Shocks European countries differ in the strength of their unions, payroll taxes, minimum wages, layoff restrictions, and unemployment insurance. Europe’s many separate labor markets cause common macroeconomic shocks to result in a growing dispersion of natural rates of unemployment from 4 percent in Ireland to 20 percent in Spain. Output per capita differs by 100 percent between wealthy Milan, Munich, and the Rhineland versus poor Greece, Southern Italy, Portugal, and Eastern Europe. Poland is about half as wealthy as the EU-15. Manufacturing labor compensation varies spectacularly from $42/hour in Denmark and $38 in Germany, halving to $21 in Spain and $18 in Greece, and then halving again to $8/hour in Portugal, Hungary, and the Czech Republic and $6/hour in Poland. The breadth of these labor market rates suggests not only an extraordinary im- mobility of labor but also a heterogeneity of macroeconomic forces that gives rise to uncorrelated shocks.

LARGEST U.S. TRADING PARTNERS: THE ROLE OF NAFTA Canada, not Japan, is by far the largest trading partner of the United States with almost twice the share (20.4 percent) of American goods exported there than anywhere else in the world’s economy (see Figure 6.13). U.S. exports to Canada include everything from merchandise, like Microsoft’s software and transmissions, carburetors, and axles for as- sembly at automated Chrysler minivan plants in Ontario, to professional services like

Example If Europe, What about One Currency for NAFTA? Should the Mexican peso, the Canadian dollar, and the U.S. dollar be replaced by a single currency for NAFTA? Seventy-nine percent of Canadian trade and 88 per- cent of Mexican trade involves the United States, and exports plus imports account for 70 percent of GDP in Canada and 58 percent in Mexico. Nevertheless, a one- size-fits-all monetary policy and little fiscal autonomy may well be as inappropriate for North America as it is for the E-27. Although Canadian macroeconomic shocks and responses are highly correlated with the United States, Mexico is a petroleum exporter with business cycles more similar to Venezuela than to the United States. For the same reason that Britain opted out of a common currency, so should Mex- ico. And the immigration and labor mobility issues between the United States and Mexico echo similar issues in Europe.

Chapter 6: Managing in the Global Economy 217

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

strategic management consulting by McKinsey & Co. Canada is also the largest source for U.S. imports (16 percent), with natural resources and finished goods manufacturing leading the list. Only slightly less, China provides 15.95 percent of U.S. imports—espe- cially computers and their parts, telecom equipment, toys and video games, clothing, fur- niture, iron and steel, shoes—and purchases 5.4 percent of U.S. exports—especially electrical machinery, power turbines, passenger aircraft, and scrap metal. Mexico con- sumes and assembles an 11.8 percent share of U.S. exports (up 50 percent since 1996) and provides 10.2 percent of U.S. imports. Mexico supplies large quantities of auto parts, steel, and oil to the United States. From 1993–2003, following the passage of NAFTA, Mexican tariffs declined from 40 percent to 16 percent, and exports to the United States increased from 67 percent to 88 percent of total Mexican exports.

NAFTA also reduced the non-tariff trade barriers for American companies like General Motors and Walmart to sell in Mexico. Walmart now operates 520 retail stores across Mexico. U.S.-owned manufacturing and processing plants have long em- ployed semiskilled labor at $3/hour in Mexico. This trade activity mirrors the labor- intensive assembly that German manufacturing companies perform in Portugal, Hun- gary, and the Czech Republic at $8/hour and Poland and Brazil at $6/hour. For exam- ple, with untariffed import of subassemblies, NAFTA lowered Mexican production costs for heavy equipment assembler Freightliner by $2,500 per truck between 1992 and 1998, enough to warrant opening a second maquiladora assembly plant in Mexico.

NAFTA’s million and a half new Mexican jobs have caused wage rates to begin to rise from $1.60 to $2.92/hour in the past five years. Consequently, some of the assem- bly line, call center, and data processing jobs have now moved to still lower wage areas in India and China, where manufacturing labor costs are less than $2/hour. Recall that manufacturing labor costs in the United States and Canada are $25 and $29/hour, respectively.

FIGURE 6.13 Largest U.S. Trading Partners

Goods Export Shares, 2008 Goods Import Shares, 2008

Mexico 11.84%

China 5.46%

Japan 5.10%

Germany 4.27%

Canada 20.45%

UK 4.20%

Other OECD 17.72%

France 2.26%

All Other 28.71%

China 15.95%

Japan 6.58%

Germany 4.60%

Canada 16.03%

Other OECD 12.08%

France 2.08%

All Other 29.71%

UK 2.77%

Mexico 10.20%

Source: National Economic Trends, Federal Reserve Bank of St. Louis.

maquiladora A foreign- owned assembly plant in Mexico that imports and assembles duty- free components for export and allows owners to pay duty only on the “value added.”

218 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Japanese goods like Toyota and Honda autos, Sony consumer electronics, Canon copiers, and Fuji film constitute 8 percent of total U.S. imports, and the Japanese ab- sorb 6 percent of U.S. exports, primarily aircraft, chemicals, computers, timber, corn, and coal. Germany is the fifth largest trading partner of the United States. Germany exports principally motor vehicles and parts (e.g., Mercedes-Benz diesel engines), spe- cialized machinery, and chemicals to the United States; Germany imports aircraft, computers, motor vehicles and parts (e.g., Cummins engines), and scientific equipment from the United States.

A Comparison of the EU and NAFTA Between the EU and NAFTA regional trading blocs, the EU has the larger share of world trade (35 percent compared to 29 percent for the NAFTA in 2008). Recall, however, much of the EU trade (even prior to 1992) was always with other Western European countries inside the regional trading bloc. Table 6.6 shows this was also true of U.S.-Canada trade, but not true for trade with Mexico. Twenty years pre-NAFTA, Mex- ico purchased only 4.4 percent of U.S. exports. After the reduction in trade barriers as- sociated with NAFTA, Mexican growth boomed, and only then in 1998–2003 did Mexico become the second largest purchaser of U.S. exports (with a 11 percent share today).

Another important contrast between the EU and NAFTA is that social security pro- grams impose a heavy burden on manufacturing competitiveness in Europe. French,

Example Household Iron Manufacturer in Mexico Becomes Major Engine Block Supplier to Detroit: Cifunsa SA33

Since its passage in 1994, NAFTA has made Mexico into a leading outsourcing lo- cation for the worldwide auto parts industry. Duty-free access to the United States for auto subassemblies like transmissions and wiring harnesses, plus a growing sec- tor of skilled nonunionized workers, induced GM, Ford, DaimlerChrysler, and Volkswagen, as well as Mexican firms like San Luis Corp. and Grupo Industrial Saltillo SA, to invest $18 billion in auto plants and equipment between 1994 and 2000.

Cifunsa SA is a Grupo subsidiary that specialized immediately after World War II in the manufacture of metal castings for household appliances, especially hand irons. Today, Cifunsa has converted its aluminum and steel casting expertise to the production of engine blocks. Indeed, Cifunsa has the dominant position as a sup- plier of engine monoblocks to the North American car companies. Other Mexican metal casters play a major role in supplying heavy axles and coil springs for trucks and SUVs. Many windshields installed on U.S.-assembled cars and trucks also come from Mexico. Although some of this import-export trade is motivated by lower wage rates at Mexican than at comparable U.S. parts suppliers, another fac- tor is the desire of the American and European car companies to decrease their dependence on unionized plants. A 2009 strike at GM’s Delphi subsidiary lasted almost six weeks.

33Based on “Mexico Is Becoming Auto-Making Hot Spot,” Wall Street Journal (June 23, 1998); and “Mexico Becomes a Leader in Car Parts,” Wall Street Journal (March 30, 1999), p. A21.

Chapter 6: Managing in the Global Economy 219

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Swedish, Italian, and Czech social security contributions now add over 20 percent to wage costs relative to 5 percent to 10 percent in Japan, Korea, Canada, and the United States. Five weeks of paid vacation in Germany is standard, and the Germans spend 8.4 percent of GDP on pension payments. Compare two weeks of paid vacation in the United States and the 5 percent of GDP Americans spend on pension payments.

Opting out of these EU social programs has allowed the British economy to match U.S. total labor costs. As a consequence, although wages for time worked in the U.S., British, and German manufacturing sectors are very similar (approximately $15 per hour), labor costs for holiday and leave pay add $5.70 an hour in Germany and only $1.03 an hour in the United States. When pensions and health care coverage are in- cluded, the total labor cost in Germany in 2009 rises to $38/hour versus $26/hour in the United States. In addition, French labor law makes it nearly impossible to lay off and furlough workers. Consequently, few entrepreneurial businesses in France proceed beyond very small sole proprietorships.

What all this demonstrates is that the institutional arrangements in the country sur- rounding a company are as important to its ultimate competitive success as the business plan, the quality of management decisions, and the commitment of dedicated employees. The enhanced competitive pressure arising from free trade and the opening of markets has served to accentuate the disadvantages of inefficient institutional arrangements. Rather than struggle against regulations that raise costs, global supply chain managers just take their business elsewhere.

Gray Markets, Knockoffs, and Parallel Importing The prices charged for identical goods varied widely across Europe both before and after the formation of the Common Market (again, see Table 6.5). In 1998, a Ford Mondeo cost 50 percent more in Germany than in Spain. To lower overall consumer prices and to improve competitiveness throughout the Union, the European Commission (EC) has often adopted policies that encourage price competition. Goods arbitragers who want to buy Black and Decker power tools in Spain and sell them in Germany, or buy Kawasaki motorcycles in Holland and sell them in Britain, are encouraged to do so. Volkswagen was fined €15 million for refusing to supply Northern Italian VW dealers who sold cars

TABLE 6.6 DESTINATION OF U.S. GOODS EXPORTS

1970–1975 1998–2003 2008

COUNTRY SHARE (%) COUNTRY SHARE (%) COUNTRY SHARE (%)

Canada 21.4 Canada 24.0 Canada 19.0

Japan 10.2 Mexico 13.5 Mexico 11.0

Germany 5.4 Japan 9.4 China 6.0

United Kingdom 4.9 United Kingdom 5.2 Japan 5.0

Mexico 4.4 Germany 3.9 United Kingdom 4.5

Netherlands 3.9 Korea 3.8 Germany 3.0

France 3.1 Taiwan 3.2 Korea 3.0

Italy 2.9 Netherlands 2.9 Netherlands 2.0

Brazil 2.7 France 2.8 Brazil 2.0

Belgium-Luxembourg 2.3 China 2.4 Taiwan 2.0

Source: Federal Reserve Bank of St. Louis, U.S. Department of Commerce.

220 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

to large numbers of Munich weekenders traveling across the Alps for the Verona opera (and the inexpensive German car prices available in Italy). The EC also has eliminated any contractual link between product sales and after-market service; any government- certified repair shop can purchase parts to perform VW, Nikon, or Sony maintenance and service. The problem, of course, is that such gray markets may lead to counterfeit sales and substandard service passed off as branded sales and authorized service.

As the world’s largest exporter in the $182 billion movie industry and the $90 billion computer software industry, the United States has threatened retaliation unless major trading partners aggressively punish violators of international copyright and trademark protection. Japan has agreed by prohibiting Microsoft computer software sales in viola- tion of the manufacturer-authorized distribution agreements. However, at the very same time, Japan’s high court allowed the parallel importing of gray-market Steinway pianos and some copyrighted music.

Example EU Ban on Some Parallel Imports Pleases European but not U.S. and Japanese Manufacturers34

Manufacturers often seek to maintain different prices in different franchise territo- ries for identical branded goods like Levi jeans, Nike shoes, Microsoft Windows, or Sony DVDs. The European Court of Justice (ECJ) has ruled that copyright and trademark protection for Silhouette sunglasses, an Austrian export product, was infringed by an Austrian retailer who purchased the sunglasses at deep discount in Bulgaria and reimported the product for sale in Austria at prices below those suggested by the manufacturer. Sourcing a product cheaply wherever it can be pur- chased around the world and then backshipping at discounted prices into the high- valued markets is a commonplace occurrence for many trading companies. The policy question is whether the discounted product can be effectively distinguished by customers from knockoff counterfeit products and whether the brand name reputation of the manufacturer is thereby diminished.

Prior EU rulings had allowed such parallel imports, which occur any time a foreign export product is purchased in one EU country for resale in another EU country. For example, Tesco, a British retailer, purchases Levi jeans and Nike shoes overseas and offers them for sale at a discount in Britain where the Levi- and Nike- authorized distribution channels have much higher price points. Similarly, Merck pharmaceuticals produced in Germany and sold at substantial discount in Spain are backshipped into Germany by discount German retailers without prohibition (Merck v. Primecrown and Beecham v. Europharm, 1995). What was new about the Silhouette case was that Silhouette was itself an EU manufacturer. The ECJ de- cided to extend to European brand-name products (like Silhouette sunglasses) in- tellectual property protection not extended to foreign brand-name products. The purchase in Bulgaria for resale in Austria of an Austrian-manufactured product at prices well below the authorized retail price in Austria was prohibited. So, parallel importing in Europe has been somewhat limited by the Silhouette ruling.

34Based on “Set-Back for Parallel Imports,” BBC World Service (July 16, 1998); “Parallel Imports,” Financial Times (May 20, 1996); “Music Market Indicators,” The Economist (May 15, 1999); D. Wilkinson, “Breaking the Chain: Par- allel Imports and the Missing Link,” European Intellectual Property Review (1997); and “Prozac’s Maker Confronts China over Knockoffs,” Wall Street Journal (March 25, 1998), p. B9.

parallel imports The purchase of a foreign export product in one country to resell as an unauthorized import in another country.

Chapter 6: Managing in the Global Economy 221

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The price impact of a policy prohibiting parallel imports can be enormous, however. In the early 2000s, Australia carefully protected intellectual property by aggressively prosecuting gray-market sellers of music CDs. Cheap imitations and counterfeit substitutes were rare in Sydney or Melbourne, but the result was that popular music CDs sold for $6.33 more in Australia than in other Far Eastern economies. The United Kingdom and China have, on that basis, chosen the opposite policy for selected products. China permits the copying of patented medicines. Eli Lilly’s Prozac, an antidepressant, sells for $1.73 per capsule, but a chemically identical knockoff from Shanghai Zhong Qi Pharmaceutical and Jiangsu Changzhov Pharmaceutical sells for $1.02 per capsule under the brand name You Ke. Similarly, the British obtain almost 10 percent of their pharmaceuticals and over 30 percent of their wine, liquor, and beer through parallel importing of products not authorized for retail sale in the United Kingdom.

PERSPECTIVES ON THE U.S. TRADE DEFICIT Figure 6.14 shows that the U.S. international trade deficit has finally begun to shrink. In the past three years, the sum of U.S. export–import trade (the trade deficit) has fallen by half from −$758 billion (6 percent of U.S. GDP in 2006) to −$336 billion (3 percent of GDP in 2009). And well it should. U.S. exports have surged and imports have become prohibitively expensive as the U.S. dollar decayed by fully 43 percent in 2001–2009 (see Figure 6.7).

International trade outflow is just one component of a country’s balance of payments with the rest of the world. Persistent U.S. trade deficits are offset by massive capital in- flows including asset sales (such as British Petroleum’s $55 billion purchase of Amoco)

WHAT WENT RIGHT • WHAT WENT WRONG

Ford Motor Co. and Exide Batteries: Are Country Managers Here to Stay?35

As export market policies on parallel importing change, companies like Ford Motor, Procter and Gamble, and Ex- ide Batteries wrestle with the question of whether to orga- nize worldwide operations by product line or by country. That is, should operations and marketing decisions be con- trolled by global business units for Tide detergent, Pam- pers diapers, and Crest toothpaste, or should country managers in Spain, Germany, and China call the shots on input contracts, manufacturing standards, assembly loca- tion, and the all-important pricing and promotion?

By developing global product lines, Ford Motor claimed $5 billion in savings when it eliminated overlapping plants, standardized suppliers, realized volume discounts on com- ponents, and brought new products to market faster. A consolidated worldwide design team and centralized manufacturing authority saved money. However, Ford Europe’s market share slipped from 13 percent to 8.8 per-

cent because untargeted marketing and inflexible pricing became divorced from local market conditions.

Exide pursued the same path by organizing global busi- ness units around its automotive, industrial, and network telecom batteries. Major plants were spread out as far apart as China, Brazil, and Germany because subassembly com- ponents from these far-flung sources often cut labor costs tenfold and shortened lead times from three months to five weeks. Yet, Exide found that some regional sales teams continued to excel in ways global sales teams could not, and so the North American industrial battery operation has again become a separate division with pricing and pro- motion authority. As a result, relationship marketing be- tween Exide–North America and Ford headquarters in Detroit secured a major new account for Exide.

35“Place vs. Product: It’s Tough to Choose a Management Model,” Wall Street Journal (June 21, 2001), p. A1; and “The World as a Single Machine,” The Economist (June 20, 1998), pp. 3–18.

222 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

and debt issuance of U.S. Treasury bonds and bills to foreign lenders, especially the Japanese and Chinese. Figure 6.14 shows that only once in the past 35 years has the United States generated a trade surplus. Instead, the services trade surplus (e.g., McKin- sey & Co. exporting management consulting and Halliburton exporting oil field develop- ment services) is typically overwhelmed by a massive deficit in merchandise trade. By 2006 and 2007, the trade deficit of merchandise plus services trade had ballooned to −$758 billion and −$700 billion, $252 billion of which was with China itself.

Several factors contributed to this enormous 6 percent of U.S. GDP trade deficit rela- tive to the 2.7 percent 25-year average exhibited across Figure 6.14. First, U.S. manufac- turers of merchandise such as Bali bras and capital equipment such as Ford autos increasingly outsource the production of components and subassemblies to lower-wage Caribbean, Mexican, and Chinese partners and subsidiaries. These intermediate goods show up in the trade statistics as imports when they return for final assembly to the United States, even if they are produced by foreign subsidiaries owned by U.S. firms. Sec- ond, the price of crude oil skyrocketed in July 2007 to $148, three times its normal price level of $40 to $50 per barrel. As a consequence, for several months in 2007 the United States was importing from OPEC nations approximately $2 billion per day in oil. Third, throughout the decade of the 2000s, the U.S. dollar was reversing a massive 31 percent appreciation 1995–2001 against the currencies of almost all U.S. trading partners (see Figure 6.7).

Ultimately, as we have seen in this chapter, such trends in exchange rates have a significant effect on net export trade flows and vice versa. With the USD depreciating 43 percent since 2001 and pushing well below purchasing power parity (again see Fig- ure 6.7), U.S. exports were bound to recover. One example is Cummins Engine, which experienced a 26 percent increase in sales in 2006–2008 despite the worst domestic recession in decades. When combined with the experience of other export power- houses across U.S. manufacturing, a merchandise trade deficit of only −$464 billion ($270 billion with China alone) added to a +$128 billion trade surplus in services led to a much reduced −$336 billion annualized trade deficit by mid-2009 (just 3 percent of U.S. GDP).

FIGURE 6.14 The U.S. Trade Balance (Exports–Imports) as a Percentage of GDP

1975 1980 1985 1990 1995 2005 20102000

1

0

–1

–2

–3

–4

–5

–6

Percent

Source: Federal Reserve Bank of St. Louis Review, various issues.

Chapter 6: Managing in the Global Economy 223

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

SUMMARY

� Export sales are very sensitive to changes in ex- change rates. Exports become more expensive (cheaper) in the foreign currencies of the import- ing countries when the domestic (i.e., home) cur- rency of the manufacturer strengthens (weakens).

� Outsourcing to lower-wage manufacturing plants is a centuries-old phenomenon. Outsourcing often imports innovation and access to highly skilled an- alytical capabilities in testing and design as well as exporting lower-skilled jobs. Containerized ocean shipping is cheap, and yet, outsourcing costs must include increased costs of vendor selection, quality control, intellectual property insurance, and expatriate management compensation.

� The Chinese $4.3 trillion economy is understated by nominal exchange rates to be smaller than the $5 trillion Japanese economy. A more accurate measure would be $7.8 trillion.

� For more than a decade, China has grown at 12 to 15 percent with liberalized private property rights, assembly for multinational outsourcing, and a booming retail domestic sector. In the aftermath of the global financial crisis, China is the engine for growth worldwide as major trading partners export to its burgeoning middle class.

� Major currencies are traded in the foreign ex- change markets; there are markets for U.S. dollars as foreign exchange, British pounds as foreign ex- change, German euros as foreign exchange, and so forth. Demand and supply in these markets reflect the speculative and transactions demands of inves- tors, import–export dealers, corporations, financial institutions, the International Monetary Fund, cen- tral bankers, and governments throughout the global economy.

� Companies often demand payment and offer their best fixed-price quotes in their domestic currency because of transaction risk exposure and operating risk exposure to exchange rate fluctuations. Alter- natively, such companies can manage the risk of exchange rate fluctuations themselves by setting up internal or financial hedges involving forward, option, or currency swap contracts.

� Internal hedges may be either balance sheet hedges addressing translation risk or operating hedges

matching anticipated foreign sales receipts with an- ticipated expenses in that same foreign currency. Financial hedges often address transaction risk ex- posure by setting up positions in financial deriva- tive contracts to offset cash flow losses from currency fluctuations. Such hedging costs about 5 percent of the value at risk.

� Foreign buyers (or their financial intermediaries) usually must acquire euros to execute a purchase from Mercedes-Benz, U.S. dollars to execute a pur- chase from General Motors, or yen to execute a purchase from Toyota. Each buyer in these interna- tional sales transactions usually supplies its own do- mestic currency. Additional imports by Americans of Japanese automobiles would normally therefore result in an increased demand for the yen and an increased supply of dollars in the foreign currency markets, that is, a dollar depreciation.

� Long-term trends in exchange rates are determined by transaction demand, government transfer pay- ments, and central bank or IMF interventions.

� Three transaction demand factors are real (inflation-adjusted) growth rates, real (inflation- adjusted) interest rates, and expected cost inflation. The lower the expected cost inflation, the lower the real growth rate, and the higher the real rate of interest in one economy relative to another, the higher the exports, the lower the import demand, and the higher the demand for financial instru- ments from that economy. All three determinants imply an increased demand or decreased supply of the domestic currency, that is, a currency appreciation.

� Consumer price inflation serves as a good predic- tor of the combined effect of all three transaction demand factors on post-redemption returns to for- eign asset holders. Projected changes in consumer inflation therefore directly affect international cap- ital flows that can easily overwhelm the effect of trade flows on exchange rates.

� The relative strength of a currency is often mea- sured as an effective exchange rate index, a weighted average of the exchange rates against ma- jor trading partners, with the weights determined by the volume of import plus export trade.

224 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

� Free trade increases the economic growth of both industrialized and developing nations. Tariffs, du- ties, and import quotas sometimes play a role in strategic trade policy to force multilateral reduction of tariffs, open markets, or secure increasing returns.

� Trade restrictions (quotas or tariffs) may be war- ranted under special circumstances to protect in- fant industries, to offset government subsidies with countervailing duties, or to impose antidumping sanctions on foreign imports sold at a price below their domestic cost.

� Speculative demand especially influences short- term changes in exchange rates. Since the total dol- lar volume of foreign currency trading worldwide is $1.5 trillion per day, these short-term fluctua- tions can be quite volatile.

� International capital flows and the flow of tradable goods across nations respond to arbitrage opportu- nities. Arbitrage trading ceases when parity condi- tions are met. One such condition is relative purchasing power parity.

� Relative purchasing power parity (PPP) hypothesizes that a doubling of consumer prices in one economy will lead to trade flows that cut in half the value of the currency. Over long periods of time and on an approximate basis, exchange rates do appear related to differential rates of inflation across economies. PPP serves a useful benchmark role in assessing long-term trends in exchange rates.

� The European Union (EU) and the North Ameri- can Free Trade Agreement (NAFTA) are two of several large trading blocs that have organized to

open markets to free trade. The EU is the largest producer of world output with very dissimilar economies that have reduced trade barriers and specialized in accordance with comparative advan- tage. Marketing across the EU must address clus- ters of very different consumers.

� Whether a nation should join a (single) currency union depends on: (1) the magnitude of intra- regional trade, (2) the mobility of labor, and (3) the correlation of macroeconomic shocks.

� The United States is both the largest single-nation ex- porter and the largest importer in the world economy. The largest trading partner of the United States is China, followed by Canada, Mexico, Japan, the United Kingdom, and Germany. The United States’ share of world export trade (11 percent) has grown in recent years, along with that of Germany and China.

� The trade flows of the United States are often in deficit (i.e., imports exceed exports); the last time there was a trade surplus in the United States was during the recession of 1981–1982. The balance- of-trade deficit of the United States is offset by international capital flows into the United States. The balance-of-payments accounts reflect this ac- counting identity.

� The U.S. 2009 trade deficit was generated by $464 billion more merchandise imported into the United States than exported. Services generated a $128 bil- lion trade surplus. In recent years, these trade defi- cits have been approximately 3 percent of a $14 trillion gross domestic product in the United States, half the level of earlier in the decade.

Exercises 1. If the U.S dollar depreciates 20 percent, how does this affect the export and do- mestic sales of a U.S. manufacturer? Explain.

2. If the U.S. dollar were to appreciate substantially, what steps could a domestic manufacturer like Cummins Engine Co. of Columbus, Indiana, take in advance to reduce the effect of the exchange rate fluctuation on company profitability?

3. After an unanticipated dollar appreciation has occurred, what would you recom- mend a company like Cummins Engine do with its strong domestic currency?

4. What is the difference between transaction demand, speculative demand, and au- tonomous transactions by central banks, the World Bank, and the IMF in the for- eign exchange markets? Which of these factors determines the long-term quarterly trends in exchange rates?

5. Would increased cost inflation in the United States relative to its major trading partners likely increase or decrease the value of the U.S. dollar? Why?

Answers to the exercises in blue can be found in Appendix D at the back

of the book

Chapter 6: Managing in the Global Economy 225

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

6. If the domestic prices for traded goods rose 50 percent over 10 years in Japan and 100 percent over those same 10 years in the United States, what would happen to the yen/dollar exchange rate? Why?

7. If Boeing’s dollar aircraft prices increase 20 percent and the yen/dollar exchange rate declines 15 percent, what effective price increase is facing Japan Air Lines for the purchase of a Boeing 747? Would Boeing’s margin likely rise or fall if the yen then depreciated and competitor prices were unchanged? Why?

8. Unit labor costs in Germany approach $30 per hour, whereas in Britain unit labor costs are only $17 per hour. Why would such a large difference persist between two members of the EU free trade area?

9. If unit labor costs in Spain and Portugal rise, but unit labor costs in Germany decline and other producer prices remain unchanged, what effect should these factors, by themselves, have on export trade, and why?

10. What three factors determine whether two economies with separate fiscal and monetary authorities should form a currency union? Give an illustration of each factor using NAFTA economies.

11. Would yogurt or Prada handbags have wider price variation in a free trade area like the European Union? Why?

12. If core price inflation has grown at a compound growth rate of 2 percent per an- num in the United States and 0.04 percent in Japan for the past eight years, what exchange rate represents PPP today if the two currencies eight years ago in 2002 were in parity and exchanged at the rate of ¥120/$?

Case Exercises PREDICTING THE LONG-TERM TRENDS IN

VALUE OF THE U.S. DOLLAR AND EURO Analyze the data on inflation rates, interest rates, and growth rate forecasts in Table 6.1 to determine what the likely near-term trend movement of the U.S. dollar will be. Analyze how each of the previously mentioned factors will affect the euro-dollar rate.

ELABORATE THE DEBATE ON NAFTA PROS:

1. trade between Mexico, Canada, and U.S. has tripled to $1 trillion, 2. food exports to Mexico have grown at 8 percent annually, 3. Mexico has increased its exports and foreign direct investment, 4. U.S. manufacturing has become more cost competitive using components and

subassemblies from Canada and Mexico,

CONS: 1. 100,000s of U.S. jobs have shifted below the border, 2. wage concessions have been forced upon those U.S. workers who remain

employed in industries that outsource to Mexico, 3. the United States has run a persistent trade deficit with Mexico and Canada,

226 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

6A APPENDIX

Foreign Exchange Risk Management To reduce the potentially wide swings in cash flows and net assets resulting from cur- rency fluctuations, companies either set up internal hedges or employ financial deriva- tives to create financial hedges. Internal hedges may be either operating hedges or balance sheet hedges. Operating hedges reduce operating cash flow risk exposure by matching anticipated foreign sales receipts against projected foreign operating expenses in that same currency. A few companies like Nestlé and Unilever have so many foreign operations and so many global brands (Nestlé Crunch, Carnation, Perrier, Kit Kat, Lipton Tea, Dove Soap, Wishbone, Bird’s Eye, Obsession) that their operating hedges circumvent the need for further risk management.

In contrast, balance sheet hedges address primarily translation risk exposure by matching assets and liabilities in various countries and their respective currencies. Fewer than 25 percent of all U.S., Asian, and U.K. companies consider translation risk important.

Financial hedges in a manufacturing or service company reduce transaction risk ex- posure by setting up positions in financial derivative contracts to offset cash flow losses from currency fluctuations. Over 93 percent of U.S., Asian, and U.K. companies em- ploy forward contracts to manage transaction risk exposure. Goldman Sachs estimates that a financial hedge for $100 million of risk exposure costs about $5.2 million, ap- proximately 5 percent of the value at risk. As the dollar appreciated steeply against the euro in 2000, Coca-Cola established a covered hedge for their euro net cash flow risk exposure that year. Goodyear, in contrast, decided that these hedging costs were prohibitive and ended up earning just $68 million profit on their European operations rather than the $92 million ($97 million million gross profit – $5 million hedging cost) that they would have earned with a fully hedged position.

By establishing a short position in the foreign currency forward or options markets, a company can hedge the domestic cash flow from their export sales receipts. For example, suppose Cummins Engine had sales contracts with their German dealers for future

Example Internal Hedge from BMW Operations on I-85 in South Carolina The North American subsidiary of BMW now accepts purchase orders accompa- nied by payment in U.S. dollars and uses those same dollar receipts to cover BMW marketing and plant expenses in the United States. BMW has built a Spartanburg, South Carolina, plant to assemble their popular Z4 model sports car. This massive facility along the major East Coast Interstate I-85 has tens of millions of dollars of labor and local materiel expenses. Such offsetting expense flows payable in U.S. dollars is a way of hedging the operating risk exposure of BMW receivables in U.S. dollars. With an internal hedge provided by a German ball-bearing plant, Cummins Engine could accomplish the same thing.

227

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

delivery in 2010 of €5 million of diesel engines. Cummins has risk exposure to a decline in the value of these export sales receipts. So, to lay off this exchange rate risk, in 2009 Cummins sells a euro forward contract in the foreign exchange derivative markets to es- tablish a hedge. Cummins’s transaction is described as a covered hedge because Cum- mins anticipates euro receivables (from their German dealers) just equal to the amount of their short forward position. That is, their contract sales receipts “cover” their obliga- tion to deliver as a seller of a euro forward contracts.

Besides setting up internal hedges or short forward positions to achieve a perfect financial hedge, Cummins Engine and BMW could also have entered into a currency-swap contract to exchange their anticipated future streams of dollar and euro cash flows from export sales. Cummins would swap a prespecified amount of Cummins’s anticipated euro sales receipts in Germany for a prespecified amount of their anticipated dollar receipts from auto sales in North America. However, these swap contract alternatives to demanding payment in their domestic (home) currency impose some transaction fees on BMW and Cummins Engine. Therefore, BMW generally will offer its best fixed price for an export transaction on a pur- chase order payable in euros. And for the same reason, the best fixed price from Cummins Engine generally will be available only on a purchase order payable only in U.S. dollars.

INTERNATIONAL PERSPECTIVES

Toyota and Honda Buy U.S. Assembly Capacity 1

To exempt their cars and trucks from U.S. tariffs, and to improve the delivery time and reliability for their most popular models, Toyota and Honda have each purchased four assembly plants in North America. When a manufacturer’s home currency is strong, foreign direct investment in overseas plant and equipment is especially attractive. From 1985 to 1993, the yen rocketed from ¥238 per U.S. dollar to ¥94 per U.S. dollar (see Figure 6.1), as Honda and Toyota employed the strengthening yen to acquire their U.S. manufacturing capacity. A $1 billion U.S. assembly plant that drew down balance sheet cash by ¥94 billion in 1993 would have drawn down fully

¥238 billion cash 10 years earlier (or would have in- troduced a ¥238 billion debt liability on the balance sheet). These reduced liabilities in acquiring new fixed assets provided a balance sheet hedge offsetting the ever lower yen receipts in 1980–1995 from the U.S. sale of another Camry or Accord. Nevertheless, the U.S. assembly plants were clearly a response to U.S. protectionist trade policy and not motivated by the balance sheet hedges.

1Based on “Japanese Carmakers Plan Major Expansion of American Capacity,” Wall Street Journal (September 24, 1997), p. A1; and “Detroit Is Getting Sideswiped by the Yen,” BusinessWeek (November 11, 1996), p. 54.

Example Cummins Engine Goes Short Selling forward contracts in 2009 at a prespecified forward price (say, $1.50/€) that agreed to deliver €5 million at a future settlement date in 2010 would make money for Cummins Engine if the dollar appreciated and the euro receipts from foreign sales declined in dollar value. For example, at $1.30/€ in 2010, Cummins is entitled to receive $1.50/€ for European currency that it could purchase in the spot market in 2010 for $1.30/€. Therefore, Cummins “cancels” its forward position and can collect from the futures market settlements process a gain of $0.20/€ × €5 million = $1,000,000. This cash flow would be just sufficient to cover their $0.20 per euro loss in value on the €5 million in 2010 sales receipts from their German dealers. As intended, these two cash flows from a covered hedge just offset each other; the perfect hedge lays off the FX risk.

228 Part 2: Demand and Forecasting

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PART3 PRODUCTION AND COST

ECONOMIC ANALYSIS AND DECISIONS

1. Demand Analysis 2. Production and Cost Analysis 3. Product, Pricing, and Output

Decisions 4. Capital Expenditure Analysis

ECONOMIC, POLITICAL, AND SOCIAL ENVIRONMENT

1. Business Conditions (Trends, Cycles, and Seasonal Effects)

2. Factor Market Conditions (Capital, Labor, Land, and Raw Materials)

3. Competitors’ Reactions and Tactical Response

4. Organizational Architecture and Regulatory Constraints

Cash Flows Risk

Firm Value (Shareholders’ Wealth)

229

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

7 CHAP T E R

Production Economics CHAPTER PREVIEW Managers are required to make resource allocation de- cisions about production operations, marketing, financing, and personnel. Al- though these decisions are interrelated, it is useful to discuss each of them separately. Production decisions determine the types and amounts of inputs— such as land, labor, raw and processed materials, factories, machinery, equipment, and managerial talent—to be used in the production of a desired quantity of out- put. The production manager’s objective is to minimize cost for a given output or, in other circumstances, to maximize output for a given input budget. First, we an- alyze the choice of a single variable input with fixed input prices. Later, we analyze the optimal multi-input combination and introduce the concept of returns to scale. Appendix 7A considers the production manager’s dual problem of constrained cost minimization and constrained output maximization. Appendix 7B examines the production economics of renewable and exhaustible natural resources.

MANAGERIAL CHALLENGE Green Power Initiatives Examined: What Went Wrong in California’s Deregulation of Electricity?1

Electric power plants entail huge capital investments. Pollution abatement technology in large coal-fired plants and redundant safety devices in nuclear power plants require almost a billion dollars of extra capital. Pacific Gas and Electric’s (PG&E) Diablo Canyon nu- clear power plant near Santa Barbara, California, cost $5.8 billion. The British plan to install 5,000 offshore wind turbines at a price of $20 billion to replace half of the 58 gigawatts of power (77 percent of total capac- ity) the United Kingdom obtains today from natural gas and coal. Why spend so much to secure these greener technologies?

One prominent reason is the €14 ($20) pollution al- lowance per ton of coal that has emerged from the car- bon dioxide emissions trading scheme (ETS) introduced in 2005 by the European Union to combat the damag- ing effects of global warming. A second reason to pursue greener power is the $5 per megawatt hour (MWh) carbon tax on electricity enacted in 1991 by Sweden, Finland, Denmark, Norway, the Netherlands, and

more recently by Ireland, France, British Columbia, and Boulder, Colorado. The carbon tax on conventional sources of electricity in Boulder is estimated at $21 per year for the average residence, $94 for the average com- mercial site, and $9,600 for the average industrial site.

Finally, greener power technologies have much lower variable costs than smaller-scale natural gas-powered and fuel oil-powered electricity generating plants. The operating cost for coal itself is only $25/MWh versus $35/MWh for natural gas and fuel oil-fired power plants or $65 for diesel-fired plants (see the step function graph in the exhibit on the next page). The operating cost for nuclear and wind power is even lower—only $4/MWh. A trade-off between investing in high-fixed-cost plants with lower and stable variable costs versus low- fixed-cost plants requiring higher variable costs came into focus during the electricity deregulation crisis in California.

California implemented legislation to decouple elec- tricity generation and distribution, allowing large retail

230

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

and industrial customers to purchase electricity from long-distance suppliers in places like the state of Washington where hydropower is plentiful. As a result, two California utilities, PG&E and Southern Cal Edison, scaled back their generating plant expansion plans and began to meet peak demand by purchasing 25 percent of their power in wholesale spot markets. It has been estimated that as much as 55 percent of the variation in peak-hour daily wholesale prices is attributable to the small-scale diesel-fired independent generating plants that fire up to meet the last 5 percent of peak demand (see the graph).

The monthly average wholesale price of electricity in California shot up from $25–$50/MWh in the 1990s to $200+/MWh in the 2000s. California Edison and PG&E were restrained by the public utility commission from

passing through to their retail customers almost $11 billion in higher wholesale costs. With the onslaught of rolling blackouts in 2001, the California Public Utility

Dpeak

Dintermediate9:30–11 A.M. and P.M.

11:30 A.M. to 9:30 P.M.

11 P.M. to 9:30 A.M.D off-peak

Variable cost of electricity generation, pre-tax, various energy sources

O pe

ra ti

ng c

os t/

M W

h ($

)

$2 $4

$25

$35

$65

Capacity

Coal-fired

Nuclear Hydro

Natural gas and fuel oil-fired

Diesel-fired

Spot market purchasing

S

20% 40% 80% 95%70%

MANAGERIAL CHALLENGE Continued

© Do n Fa rra ll/ Ph ot od is c Gr ee n/ Ge tty

Im ag es

Chapter 7: Production Economics 231

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The economic theory of production consists of a conceptual framework to assist managers in deciding how to combine most efficiently the various inputs needed to produce the desired output (product or service), given the existing tech- nology. This technology consists of available production processes, equipment, labor and management skills, as well as information-processing capabilities. Pro- duction analysis is often applied by managers involved in assigning costs to the various feasible output levels and in communicating with plant engineers the operations plans of the company.

THE PRODUCTION FUNCTION The theory of production centers around the concept of a production function. A production function relates the maximum quantity of output that can be produced from given amounts of various inputs for a given technology. It can be expressed in the form of a mathematical model, schedule (table), or graph. A change in technology, such as the introduction of more automated equipment or the substitution of skilled for un- skilled workers, results in a new production function. The production of most outputs (goods and services) requires the use of large numbers of inputs. The production of

Commission approved a 46 percent increase in retail electricity rates, but PG&E had already declared bank- ruptcy. Such blackouts and brownouts are now pro- jected in Britain where outdated capacity is being retired and demand is growing faster than the power utilities can build new capacity.

One possible solution is to charge electricity customers a variable rate per MWh to reflect the differential cost at different hours of the day, days of the week, and seasons of the year. As the variable cost of electricity rises and falls throughout the day along the supply schedule in the ex- hibit, the asking price of electricity rises and falls accord- ingly. France has long applied such differential pricing to electricity. A second new approach is to deploy ex- tremely small-scale diesel generators or natural gas-fired microturbine generators in factories and commercial es- tablishments. Operating costs of the microturbines are $70–$120/MWh, much higher than utility-supplied elec- tricity. Yet, the capital cost that must be recovered is less than 1/1000th of traditional power plants. La Quinta Mo- tels, for example, saved $20,000 in one year with a micro- turbine at one of its Dallas properties. Puget Sound Energy of Seattle has begun installing microturbines in substations near growing neighborhoods that are fore- casted to be sources of peak demand on capacity. In this chapter, we will study the dilemma of whether to substi- tute higher cost variable inputs for fixed inputs entailing substantial capital investment.

Discussion Questions

� What are the cost tradeoff issues between coal and natural gas as a source of fuel for electricity?

� Among the contentious questions at the Co- penhagen Climate Change conference in 2009 was exactly how the developed economies like Germany, France, Britain, and the United States, which have increasingly adopted nuclear and other green technologies, will financially induce the rapidly developing nations of Brazil, India, and China to forgo the cheap coal that so damaged the planet for the past 130 years. Make a proposal you think would work.

� If you were asked to pay three times as much for electricity use in the early evening as early in the morning, would you get up early to do laundry before going to work or school?

1Based on “The Lessons Learned” and “Think Small,” Wall Street Journal (September 17, 2001), pp. R4, R13, R15, and R17; “Are Californians Starved for Energy?” Wall Street Journal (September 16, 2002), p. A1; “How to Do Deregulation Right,” BusinessWeek (March 26, 2001), p. 112; and “The Looming Energy Crunch,” The Economist (August 8, 2009), p. 49.

production function A mathematical model, schedule (table), or graph that relates the maximum feasible quantity of output that can be produced from given amounts of various inputs.

inputs A resource or factor of production, such as a raw material, labor skill, or piece of equipment that is employed in a production process.

MANAGERIAL CHALLENGE Continued

232 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

gasoline, for example, requires the use of many different labor skills (roughnecks, chemical engineers, refinery maintenance workers), raw materials (crude oil, chemical additives, heat), and types of equipment (boilers, distillation columns, cracking cham- bers). Also, production processes often result in joint outputs. For example, petroleum refining results in jet fuel, propane, butane, gasoline, kerosene, lubricant oil, tar, and asphalt.

Letting L and K represent the quantities of two inputs (labor L and capital K) used in producing a quantity Q of output, a production function can be represented in the form of a mathematical model, such as

Q = αLβ1Kβ2 [7.1]

where α, β1, and β2 are constants. This particular multiplicative exponential model is known as the Cobb-Douglas production function and is examined in more detail later in the chapter. Production functions also can be expressed in the form of a schedule (or table), as illustrated in the following ore-mining example.

Example An Illustrative Production Function: Deep Creek Mining Company The Deep Creek Mining Company uses capital (mining equipment) and labor (workers) to mine uranium ore. Various sizes of ore-mining equipment, as mea- sured by its brake horsepower (bhp) rating, are available to the company. The amount of ore mined during a given period (Q) is a function only of the number of workers assigned to the crew (L) operating a given type of equipment (K). The data in Table 7.1 show the amount of ore produced (measured in tons) when vari- ous sizes of crews are used to operate the equipment.

A two-input, one-output production function at Deep Creek Mining can also be represented graphically as a three-dimensional production surface, where the height of the square column associated with each input combination in Figure 7.1 indicates the amount of iron ore output produced.

TABLE 7.1 TOTAL OUTPUT TABLE—DEEP CREEK MINING COMPANY

CAPITAL INPUT K (BHP, BRAKE HORSEPOWER)

250 500 750 1,000 1,250 1,500 1,750 2,000

LABOR INPUT L (NUMBER OF WORKERS)

1 1 3 6 10 16 16 16 13

2 2 6 16 24 29 29 44 44

3 4 16 29 44 55 55 55 50

4 6 29 44 55 58 60 60 55

5 16 43 55 60 61 62 62 60

6 29 55 60 62 63 63 63 62

7 44 58 62 63 64 64 64 64

8 50 60 62 63 64 65 65 65

9 55 59 61 63 64 65 66 66

10 52 56 59 62 64 65 66 67

Cobb-Douglas production function A particular type of mathematical model, known as a multiplicative exponential function, used to represent the relationship between the inputs and the output.

Chapter 7: Production Economics 233

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Fixed and Variable Inputs In deciding how to combine the various inputs (L and K) to produce the desired output, inputs are usually classified as being either fixed or variable. A fixed input is defined as one required in the production process but whose quantity employed in the process is constant over a given period of time regardless of the quantity of output produced. The costs of a fixed input must be incurred regardless of whether the production process is operated at a high or a low rate of output. A variable input is defined as one whose quantity employed in the process changes, depending on the desired quantity of output to be produced.

The short run corresponds to the period of time in which one (or more) of the inputs is fixed. To increase output, then, the firm must employ more of the variable input(s) with the given quantity of fixed input(s). Thus, for example, with an auto assembly plant of fixed size and capacity, the firm can increase output only by employing more labor, such as by scheduling additional shifts.

FIGURE 7.1 The Production Function—Deep Creek Mining Company

62 64

62 64

64 6364

63 596162

56560 58

55

43

50

44

13

5255

50

44

29

16

55

55 55

55

60

60 60

62 62

61

63 63

63

67 6666

6665 65

64

65 65

65 64

64 63

62

60

62

60 58

55

6

29

44

29 29

24

16 16

16

10

16

6 4

16

6

3 2 1

44 44

29

brake

Capital input K ( horsepower)

Lab or i

npu t L

(nu mb

er o f wo

rker s)

250 500

7501 ,0001

,2501 ,5001

,7502 ,000

1 2

3 4

5 6

7 8

9 10

55

short run The period of time in which one (or more) of the resources employed in a production process is fixed or incapable of being varied.

234 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

As the time period under consideration (the planning horizon) is lengthened, how- ever, more of the fixed inputs become variable. Over a planning horizon of about six months, most firms can acquire or build additional plant capacity and order more manufacturing equipment. In lengthening the planning horizon, a point is eventually reached where all inputs are variable. This period of time is called the long run.

In the short run, because some of the inputs are fixed, only a subset of the total pos- sible input combinations is available to the firm. By contrast, in the long run, all possible input combinations are available.

PRODUCTION FUNCTIONS WITH ONE VARIABLE INPUT Suppose, in the Deep Creek Mining Company example of the previous section, that the amount of capital input K (the size of mining equipment) employed in the production process is a fixed factor. Specifically, suppose that the firm owns or leases a piece of min- ing equipment having a 750-bhp rating. Depending on the amount of labor input L used to operate the 750-bhp equipment, varying quantities of output will be obtained, as shown in the 750 column of Table 7.1 and again in the Q column of Table 7.2.

Marginal and Average Product Functions Once the total product function is given (in tabular, graphic, or algebraic form), the mar- ginal and average product functions can be derived. The marginal product is defined as the incremental change in total output ΔQ that can be produced by the use of one more unit of the variable input ΔL, while K remains fixed. The marginal product is defined as2

MPL = ΔQ ΔL

or ∂Q ∂L

[7.2]

for discrete and continuous changes, respectively.

TABLE 7.2 TOTAL PRODUCT, MARGINAL PRODUCT, AVERAGE PRODUCT, AND

ELASTICITY—DEEP CREEK MINING COMPANY (CAPITAL INPUT, BHP = 750)

LABOR INPUT (NUMBER OF WORKERS)

TOTAL PROD- UCT TPL (= Q) (TONS OF ORE)

MARGINAL PROD- UCT OF LABOR, MPL (ΔQ ÷ ΔL )

AVERAGE PROD- UCT OF LABOR,

APL (Q ÷ L)

PRODUCTION ELASTICITY, EL

(MPL ÷ APL)

0 0 — — —

1 6 + 6 6 1.0

2 16 +10 8 1.25

3 29 +13 9.67 1.34

4 44 +15 11 1.36

5 55 +11 11 1.0

6 60 + 5 10 0.50

7 62 + 2 8.86 0.23

8 62 0 7.75 0.0

9 61 − 1 6.78 −0.15

10 59 − 2 5.90 −0.34

long run The period of time in which all the resources employed in a production process can be varied.

marginal product The incremental change in total output that can be obtained from the use of one more unit of an input in the production process (while holding constant all other inputs).

2Strictly speaking, the ratio ΔQ/ΔL represents the incremental product rather than the marginal product. For simplicity, we continue to use the term marginal throughout the text, even though this and similar ratios are calculated on an incremental basis.

Chapter 7: Production Economics 235

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The marginal product of labor in the ore-mining example is shown in the third column of Table 7.2 and as MPL in Figure 7.2.

The average product is defined as the ratio of total output to the amount of the vari- able input used in producing the output. The average product of labor is

APL = Q L

[7.3]

The average product of labor for the Deep Creek ore-mining example is shown in the fourth column of Table 7.2.

The Law of Diminishing Marginal Returns The tabular production function just discussed illustrates the production law of dimin- ishing marginal returns. Initially, the assignment of more workers to the crew operating the mining equipment allows greater labor specialization in the use of the equipment. As a result, the marginal output of each worker added to the crew at first increases, and total output increases at an increasing rate. Thus, as listed in Table 7.2 and graphed in Figure 7.2, the addition of a second worker to the crew results in 10 additional tons of output; the addition of a third worker results in 13 additional tons of output; and the addition of a fourth worker yields 15 additional tons.

A point is eventually reached, however, where the marginal increase in output for each worker added to the crew begins to decline. This decrease in output occurs because only a limited number of ways exist to achieve greater labor specialization and

FIGURE 7.2 Total Product, Marginal Product of Labor, and Average Product of Labor—Deep Creek Mining Company

TP

APL

MPL 0

–2 Labor input L (number of workers)

8

16

24

32

40

48

56

64

T ot

al o

ut pu

t Q

( to

ns o

f or

e)

1 2 3 4 5 6 7 8 9 10

average product The ratio of total output to the amount of the variable input used in producing the output.

236 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

because each additional worker introduces crowding effects. Thus, the addition of a fifth worker to the crew yields a marginal increase in output of 11 additional tons, compared with the marginal increase of 15 additional tons for the fourth worker. Similarly, the ad- ditions of the sixth and seventh workers to the crew yield successively smaller increases of 5 and 2 tons, respectively. With enough additional workers, the marginal product of labor may become zero or even negative. Some work is just more difficult to accomplish when superfluous personnel are present.

Increasing Returns with Network Effects The law of diminishing marginal returns is not a mathematical theorem, but rather an empirical assertion that has been observed in almost every production process as the amount of the variable input increases. A noteworthy exception occurs, however, with network effects. The greater the installed base of a network product, such as Microsoft Office and Outlook, the larger the number of compatible network connections and therefore the more possible value for a new customer. Consequently, as the soft- ware’s installed base increases, Microsoft’s promotions and other selling efforts to acquire new customers become increasingly more productive.

A manufacturer’s product line costs usually now include marketing and distribution activities as well as the labor and material direct costs of standard production and assem- bly. The reason is that, like service firms, many manufacturers today compete on customer inquiry systems, change order responsiveness, delivery reliability, and techno- logical updates, not just on delivery times and warranty repairs. Qualifying for and actu- ally winning a customer order often requires quality characteristics and support services beyond the physical unit of production. For example, Ford Motor wants all its manufacturing suppliers to meet the ISO 9000 manufacturing quality standards for con- tinuous improvement processes. Walmart requires that its fashion clothing suppliers deliver shipments just in time (JIT) for planned departure from Walmart distribution centers. Disney World gift shops choose manufacturers who can alter their production schedules on short notice in order to provide much greater change-order responsiveness than traditional make-to-order manufacturing of Mickey Mouse coffee mugs.

WHAT WENT RIGHT • WHAT WENT WRONG

Factory Bottlenecks at a Boeing Assembly Plant3

The Boeing Company assembles wide-body aircraft (747s, 767s, and 777s) at its 4.3-million-square-foot Everett, Washington, assembly plant, the largest building in the world. Fifteen railcars a day deliver parts that are directed to five assembly lines by overhead cranes cruising on 31 miles of networked track. The variable inputs in this production process are millions of parts and thousands of skilled assemblers.

As Boeing ramped up production from 244 aircraft deliv- eries in 1995 to 560 in 1999, the Everett plant went to three shifts of 6,000, 4,000, and 1,500 workers, and twice as many parts poured in. But crowding effects descended on the Ever- ett plant. Although final assembly of an aircraft body contin- ued to take its normal 21 workdays, overtime was required to maintain this roll-out schedule, in large part because of lost,

defective, and reworked parts. At times, piles of redundant parts would appear on the shop floor, while at other times, shortages of seats and electronic gear caused delays. As a result, work-in-progress inventory skyrocketed and work orders got out of sequence. By late 1997, overtime was run- ning almost a billion dollars over budget, and the production operations at Everett were “hopelessly snarled.”

To resolve the problem, in 1999 Boeing scrapped its anti- quated parts-tracking system and adopted lean production techniques. It cut parts order sizes and outsourced subassem- bly work away from bottlenecked points on the final assembly line. By 2001, a continuous stream of fewer parts arrived at autonomous worker cells just in time, as required to complete a smoothly flowing assembly process for 527 aircraft deliveries.

3“Boeing’s Secret,” BusinessWeek (May 20, 2002), pp. 113–115; “Gaining Altitude,” Barron’s (April 29, 2002), pp. 21–25; and Everett plant tours.

network effects An exception to the law of diminishing marginal returns that occurs when the installed base of a network product makes the efforts to acquire new customers increasingly more productive.

Chapter 7: Production Economics 237

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

These network-based relationships are depicted in Figure 7.3. From 0 percent to 30 percent market share, the selling efforts required to achieve each additional share point have a diminishing returns effect on the probability of adoption by the next poten- tial user (note the reduced slope of the sales penetration curve). Consequently, additional share points are increasingly more and more expensive over this range. But when the number of other users of a network-based device reaches a 30 to 40 percent share, the next 40 to 50 share points are cheaper and cheaper to promote.

Beyond the 30 percent inflection point, each additional share point of users leads to an increasing probability of adoption by another user—hence a decrease in the market- ing expense required to secure another unit sale (note the increased slope of the sales penetration curve in this middle range). These network-based effects of compatibility with other users reflect increased value to the potential adopter. As Sony Blu-ray achieved more than 30 percent acceptance in the marketplace, it effectively became an industry standard and further share points were cheaper, not more expensive. Even- tually, beyond a market share of 80 to 90 percent, securing the final adopters becomes increasingly expensive because selling efforts have again become subject to diminishing returns.

Example Increasing Returns at Sony Blu-ray and Microsoft Windows At times, additional marketing and distribution activities can lead to increasing returns and declining marginal costs. For example, securing the adoption of an industry standard favorable to one’s own product (e.g., Sony’s high-definition Blu-ray digital video standard) involves promotional and other selling efforts, which grow more productive the more widely the product is adopted by consumers. The more Blu-ray DVDs there are, the more TV networks and independent studios produce programs and movies with this technology. And the more Blu-ray pro- grams and movies that become available, the easier and cheaper it was for Sony in 2006–2007 to sell Blu-ray players initially priced from $800 to $497 to even more customers. In February 2008, when Warner Brothers withdrew their support for the competing Toshiba high-definition DVD standard, Sony and partner Matsushita Electric had the $24 billion home high-def market to themselves. Blu- ray players were priced at $388. In late 2009, Korean electronics giant Samsung emerged to contest for market share with prices slashed as low as $80 for basic models up to $221 for Internet streaming video models using Blu-ray.

A similar reason for increasing returns at Microsoft is that the more adoptions Microsoft Windows secures, the more Windows-compatible applications will be in- troduced by independent software programmers. And the more applications intro- duced, the greater the chance will be for further adoptions. Normal sales penetration or saturation curves (like the MPL curve in Figure 7.2) exhibit initially increasing marginal returns to promotional expenses followed by eventually dimin- ishing marginal returns. However, with the adoption of new industry standards or a network technology, increasing returns can persist. Once Sony HDTV achieved a 30 percent adoption, increasing returns in marketing its product offering intro- duced a disruptive technology that displaced other competitors. Netscape’s once- dominant Internet search engine experienced exactly this sort of displacement by Microsoft’s Windows, which bundled Internet Explorer (IE) for no additional charge. IE then grew to a 92 percent market share in Internet browsers.

238 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Producing Information Services under Increasing Returns It is insightful to compare the production economics of old-economy companies that produce things to new-economy companies that produce information. Things, when sold, the seller ceases to own. Information, when sold, the seller can sell again (at least until information spillovers overwhelm the target market). Things must be replicated through expensive manufacturing processes, whereas information is replicable at almost zero incremental cost. Things exist in one location. Information can exist simultaneously in many locations. The production and marketing of things are subject to eventually di- minishing returns. The marketing (and maybe the production) of information is subject to increasing returns. That is, the more people who use my information, the more likely it is that another person will want to acquire it (for any given marketing cost), or, said another way, the cheaper it is to secure another sale. Things often involve economies of scale in production. Information is produced by small companies at comparably low costs. Things focus a business on supply-side thinking and the high costs of distribution. Information products focus a business on demand-side thinking and have almost no costs of distribution. By getting the next customer to adopt, one can set in motion a “vir- tuous circle” of higher customer value, lower overhead costs, and lower prices and costs for the next customer. Hence, Microsoft evolved to dominate an information-oriented business like computer network software. Chapter 11 discusses increasing returns as a source of dominant firm market power.

The Relationship between Total, Marginal, and Average Product Figure 7.4 illustrates a production function total value added or total product (TP) with a single variable input to highlight the relationships among the TP, AP, and MP concepts.

FIGURE 7.3 Increasing Returns with Network Effects

Pr ob

ab ili

ty o

f ad

op ti

on b

y ne

xt t

ar ge

t cu

st om

er 1.0

4 8 16 30 85 Time path of market share (%)

Sales penetration curve

Chapter 7: Production Economics 239

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

In the first region labeled “Increasing returns,” the TP function is increasing at an in- creasing rate. Because the marginal value added or marginal product (MP) curve mea- sures the slope of the TP curve (MP = ∂Q/∂L), the MP curve is increasing up to L1. In the region labeled “Decreasing returns,” the TP function is increasing at a decreasing rate, and the MP curve is decreasing up to L3. In the region labeled “Negative returns,” the TP function is decreasing, and the MP curve continues decreasing, becoming negative beyond L3. An inflection point occurs at L1. Next, if a line is drawn from the origin 0 to any point on the TP curve, the slope of this line, Q/L, measures the average value added or average product (AP). Hence, we see that the AP curve reaches a maximum at a point where the average and the marginal products are equated.4

FIGURE 7.4 The Relationships between Total, Average, and Marginal Product Curves

Increasing returns Decreasing returns

0 L1 L2

Negative returns

TPEp = 0Ep = 1

Stage III Ep < 0

Stage II 0 < Ep < 1

Stage I Ep > 1

Point of maximum marginal returns

Q3

AP

MPInput L (units)

L3

L1 L2 L3

A ve

ra ge

o ut

pu t,

m ar

gi na

l ou

tp ut

( un

it s

of o

ut pu

t pe

r un

it o

f in

pu t)

T ot

al o

ut pu

t Q

( un

it s)

1 2 3 S U B

Gains from specialization

Region of diminishing returns

4Note also that the marginal product MP equals the average product AP at L2, because the marginal product MP is equal to the slope of the TP curve (MP = ∂Q/∂L), and at L2 the average product AP is also equal to the slope of the TP curve.

Consider, for example, the following analogies: A baseball player’s batting average for the season is 0.250 or a college student’s grade point average is 3.0. If that player has an excellent night at bat (his marginal performance) and goes 4 for 4 (1.000), or the student achieves a 4.0 in the current semester, then his season average or GPA will be pulled up. On the other hand, if he goes hitless or fails everything, this poor marginal performance will pull down his season average. If he goes 1 for 4 (or 3.0 in semester grades), this marginal performance will have no im- pact on his season average (marginal performance equals average performance). Hence, the MP curve will always intersect with the AP curve when it is at a maximum. As we will see in the next chapter, a firm’s marginal cost curve always intersects the average cost curve at its minimum point, for the same reason.

240 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

To take another example, in order to ensure that dredge spoil once removed does not flow back into harbors and navigable waterways, the U.S. Army Corps of Engineers will pay concrete block manufacturers for every cubic yard of the muddy slurry the manufac- turers actually use in their production process. Too much dredge spoil, in combination with concrete mix and sand, results in more unusable cracked blocks leaving the kilns. However, with a negative price on the input, manufacturers employ dredge spoil into the range of Stage III production. Such exceptions prove the general rule that optimal production with a single variable input and positive input prices necessitates restricting input choices to Stage II.

Example Three Stages of Production on the Camaro Assembly Line In analyzing the production function, economists have identified three different stages of production based on the relationships among the TP, AP, and MP functions. Stage I is defined as the range of L over which the average product (AP) is increasing; it occurs from the origin (0) up to L2 (perhaps from 0 up to 3 auto assembly line workers plus a substitute at job station #44 on the Camaro assembly line) and represents the region of net gains from specialization. Stage II corresponds to the range of L from the point at which the average product is at a maximum (L2) to the point where the marginal product (MP) declines to zero (L4). Note the endpoint of Stage II thus corresponds to the point of maxi- mum output on the TP curve. Here, additional employees may constitute “go- fers” who secure parts that are running low or soft drinks and candy for the crew. Stage III encompasses the range of L over which the total product is de- clining or, equivalently, the marginal product is negative. Stage III thus corre- sponds to all values of L greater than (i.e., to the right of) L3, beyond which crowding effects overwhelm any output attributable to the additional workers, perhaps apprentices.

To determine the optimal quantity of labor input L to employ, first note that the rational producer would not operate the production process over the excessive range of values of input L contained in Stage III, because the marginal product of input L is negative beyond L4. Even if the variable input were free, the rational producer would not wish to proceed into Stage III where the apprentice workers are a source of net crowding effects and total output actually declines. By the same token, no manager whose productivity per worker is rising due to the gains from specialization (i.e., AP increasing in Stage I) should stop adding workers as long as the incremental cost for additional substitute workers remains constant at a rate below the marginal value added.

In general, then, how much of the variable input to employ over the remaining range of potentially optimal input choice (Stage II) depends on the level of the incremental variable input costs. If labor costs are high, as in a United Auto Workers’ assembly plant, production may proceed just a short distance into Stage II in hiring labor. If labor costs are lower in a nonunionized plant, for example, labor hiring may proceed well across Stage II to include relatively low-level pro- ductivity workers, like the second “gofer” at L3. Of course, some input costs are subsidized (e.g., job training programs), and others may be effectively negative such that revenues actually increase even beyond L4 then more apprentice inputs are used.

Chapter 7: Production Economics 241

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

DETERMINING THE OPTIMAL USE OF THE VARIABLE INPUT With one of the inputs (K) fixed in the short run, the producer must determine the optimal quantity of the variable input (L) to employ in the production process. Such a determination requires the introduction into the analysis of output prices and labor costs. Therefore, the analysis begins by defining marginal revenue product and marginal factor cost.

Marginal Revenue Product Marginal revenue product (MRPL) is defined as the amount that an additional unit of the variable input adds to total revenue, or

MRPL = ΔTR ΔL

[7.4]

where ΔTR is the change in total revenue associated with the given change (ΔL) in the variable input, and MRPL is equal to the marginal product of L (MPL) times the marginal revenue (MRQ) resulting from the increase in output obtained:

MRPL = MPL · MRQ [7.5]

Consider again the Deep Creek Mining Company example (Table 7.2) of the previous section where K (capital) is fixed at 750 bhp. Suppose that the firm can sell all the ore it can produce at a price of $10 per ton; for example, in a perfectly competitive market, the firm would realize a constant marginal revenue equal to the going market equilibrium price. The marginal revenue product of labor (MRPL) is computed using Equation 7.5 and is shown in Table 7.3.5

Sometimes in practice this concept is referred to as the marginal value added—that is, the amount by which potential sales revenue is increased as a result of employing an ad- ditional unit of variable input to increase output. In Europe, for example, rather than taxing the final retail sales value of goods, instead each level of production from raw ma- terial to finished goods distribution is taxed on its marginal value added at each stage of production.

Marginal Factor Cost Marginal factor cost (MFCL) is defined as the amount that an additional unit of the variable input adds to total cost, or

MFCL = ΔTC ΔL

[7.6]

where ΔTC is the change in cost associated with the given change (ΔL) in the variable input.

In the ore-mining example, suppose that the firm can employ as much labor (L) as it needs by paying the workers $50 per shift (CL). In other words, the labor market is as- sumed to be perfectly competitive. Under these conditions, the marginal factor cost (MFCL) is equal to CL, or $50 per worker. It is constant regardless of the level of opera- tion of the mine (see the last column of Table 7.3).

marginal revenue product (MRPL) The amount that an additional unit of the variable production input adds to total revenue. Also known as marginal value added.

5Input levels in Stage III (MPL < 0) have been eliminated from consideration.

marginal factor cost (MFCL) The amount that an additional unit of the variable input adds to total cost.

242 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Optimal Input Level Given the marginal revenue product and marginal factor cost, we can compute the opti- mal amount of the variable input to use in the production process. Recall from the dis- cussion of marginal analysis in Chapter 2 that an economic activity should be expanded as long as the marginal benefits exceed the marginal costs. For the short-run production decision, the optimal level of the variable input occurs where

MRPL = MFCL [7.7]

As can be seen in Table 7.3, the optimal input is L = 6 workers because MRPL = MFCL = $50 at this point. At fewer than six workers, MRPL > MFCL and the addition of more labor (workers) to the production process will increase revenues more than it will increase costs. Beyond six workers, the opposite is true—costs will increase more than revenues.

PRODUCTION FUNCTIONS WITH MULTIPLE VARIABLE INPUTS Using the Deep Creek Mining Company example, suppose now that both capital (mea- sured by the maximum brake horsepower [bhp] rating of the equipment) and labor (measured by the number of workers) are variable inputs to the ore-mining process. The firm can choose to operate the production process using any of the capital-labor combinations shown previously in Table 7.1.

Production Isoquants A production function with two variable inputs can be represented graphically by a set of two-dimensional production isoquants. A production isoquant is either a geometric curve or an algebraic function representing all the various combinations of the two in- puts that can be used in producing a given level of output. In the Deep Creek example, a

TABLE 7.3 MARGINAL REVENUE PRODUCT AND MARGINAL FACTOR

COST—DEEP CREEK MINING COMPANY

LABOR INPUT L (NUMBER

OF WORKERS)

TOTAL PRODUCT Q = (TPL) (TONS OF

ORE)

MARGINAL PRODUCT OF LABOR MPL (TONS

PER WORKER)

TOTAL REVENUE TR = P · Q

($)

MARGINAL REVENUE

MRQ = ΔTR ΔQ

($/TON)

MARGINAL REVENUE PRODUCT MRPL =

MPL · MRQ ($/WORKER)

MARGINAL FACTOR

COST MFCL ($/WORKER)

0 0 — 0 — — —

1 6 6 60 10 60 50

2 16 10 160 10 100 50

3 29 13 290 10 130 50

4 44 15 440 10 150 50

5 55 11 550 10 110 50

6* 60 5 600 10 &50 &50 7 62 2 620 10 20 50

8 62 0 620 10 0 50

production isoquant An algebraic function or a geometric curve representing all the various combinations of two inputs that can be used in producing a given level of output.

Chapter 7: Production Economics 243

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

production isoquant shows all the alternative ways in which the number of workers and various sizes of mining equipment can be combined to produce any desired level of out- put (tons of ore). Several of the production isoquants for the ore-mining example are shown in Figure 7.5. For example, an output of 6 tons can be produced using any of three different labor-capital combinations: one worker and 750-bhp equipment, two workers and 500-bhp equipment, or four workers and 250-bhp equipment. Similarly, as seen in the graph, an output of 62 tons can be produced using any one of five different labor-capital combinations.

Although each isoquant indicates how quantities of the two inputs may be substi- tuted for one another, these choices are normally limited for two reasons: first, some input combinations in Figure 7.5 employ an excessive quantity of one input. Just as more than eight workers result in negative marginal returns in choosing a single vari- able input for Deep Creek Mining (see Figure 7.2), so too here with 750-bhp machin- ery crowding effects introduced by the presence of an eighth worker would actually reduce output. Similarly, more than 1,500-bhp machinery would result in negative marginal returns to capital equipment with only five workers. Because all such ineffi- cient mixes of capital and labor increase the input requirements (and therefore costs) without increasing output, they should be eliminated from consideration in making in- put substitution choices.

Second, input substitution choices are also limited by the technology of production, which often involves machinery that is not divisible. Although one can find smaller and larger mining equipment, not every brake horsepower machine listed on the Y axis of Figure 7.5 will be available. The industrial engineering of mining operations often requires that we select from three or four possible fixed proportions production processes involving a particular size drilling machine and a requisite size labor force to run it.

FIGURE 7.5 Production Isoquants—Deep Creek Mining Company

Q = 62

0 1 2 3 4 5 6 7 8

Labor input L (number of workers)

250

9 10

500

1,000

1,250

1,500

1,750

2,000

Q = 55Q = 29Q = 6

C ap

it al

in pu

t K

( br

ak e

ho rs

ep ow

er )

750

244 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Marginal Rate of Technical Substitution In addition to indicating the quantity of output that can be produced with any of the var- ious input combinations that lie on the isoquant curve, the isoquant also indicates the rate at which one input may be substituted for another input in producing the given quantity of output. Suppose one considers the meaning of a shift from point A to point B on the isoquant labeled “Q = 29” in Figure 7.7. At point A, three workers and a 750-bhp machine are being used to produce 29 tons of output, whereas at point B, four workers and a 500-bhp machine are being used to produce the same amount of output. The first input

Example Just What Exactly Is a Refinery, and Why Won’t Anyone Build One?6

New petroleum refineries are under way in Kuwait and in Saudi Arabia that will be capable of processing 600,000 barrels of crude per day (bbl/d) and 450,000 bbl/d, respectively. However, no petroleum refinery has been built in the United States in more than 30 years. Why not? To answer this question requires knowing a little about what a refinery is and what it does.

In essence, refineries are enormous chemical plants that begin by superheating various grades of crude oil in large vessels and then pass the vapors that boil off through fractional distillation columns where they phase change back into various liquids as the distillates cool (see Figure 7.6). Crude oil contains literally hundreds of hydrocarbons, and the distillates run the gamut from lubricants and grease that “cool” to a liquid at a whopping 450°C near the bottom of the distillation column to propane at the top. Jet fuel, diesel, and some gasoline liquefy at 250°C with the help of a catalytic cracking process in a separate converter, and farther up the col- umn naphtha distillate yields gasoline after passing through a reformer. Kerosene, butane, and polyethylene (the basic building block for plastic) also distill out.

Refining is a classic variable proportions production process. The chemical cracking process of breaking long chains of hydrocarbons can use more or less pressure, more or less heat, more or less high-quality but expensive light sweet crude or sulfurous heavy crude. From one 42-gallon barrel of crude, the input mix can be optimized to achieve about 20 gallons of gasoline and 10 gallons of diesel and heating oil. Much of the equipment involved is 10 stories high and expensive. The fixed-cost investment today for a major refinery totals $2 billion. This cost must be recovered from just 22 percent of the final product price of gasoline, whereas crude itself commands 54 percent of the final product price. Profit margins in refining come to just a few cents per gallon, much like the thin margins in gaso- line retailing. Oil exploration and development are much more profitable than re- fining in part to compensate for the extraordinary capital investment risks involved.

PERCENT OF FINAL PRODUCT PRICE OF GASOLINE BY STAGE OF PRODUCTION (2006)

Exploration and development and extraction of crude oil 54%

Federal and state excise taxes 16

Refining 22

Distribution and retail 8

6Based on “Working Knowledge: Oil Refineries,” Scientific American (June 2006), pp. 88–89.

Chapter 7: Production Economics 245

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

mix is capital-intensive like Chrysler’s highly robotic Ontario minivan plant; the second is more labor intensive, like hand-tooled auto assembly. In moving from input mix A to input mix B, we substituted one additional unit of labor for 250 units of capital. The rate at which capital has been replaced with labor in producing the given output is equal to 250/L or 250 units of capital per unit of labor. The rate at which one input may be

FIGURE 7.6 Crude Oil Is Made Into Different Fuels from Distillation/Cracking/Reformer Processes

A refinery’s most important processes

Products made from a barrel of crude oil (42 gal)

Other products

Liquefied petroleum gas (LPG)

Heavy fuel oil

Jet fuel

Diesel fuel & heating oil

Gasoline

7.6 1.7 1.1

4

10

19.6 gal

LPG

Gasoline

Jet fuel

Diesel fuel

LPG Gasoline

Motor gasoline Jet fuel Diesel fuel

Gasoline Vapors LPG

Naphtha

Kerosene

Diesel distillate

Medium- weight gas oil Heavy gas oil

Residuum Industrial fuel

Asphalt base

End products

Reformer

Alkylation unit

Coker

Cracking units

Distillation column

Source: Energy Information Administration, U.S. Department of Energy.

246 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

substituted for another input in the production process, while total output remains constant, is known as the marginal rate of technical substitution (MRTS).

MRTS is given by the slope of the curve relating K to L—that is, the slope of the iso- quant. The slope of the AB segment of the isoquant in Figure 7.7 is equal to the ratio of AC to CB. Algebraically, AC = K1 − K2 and CB = L1 − L2; therefore, the slope is equal to (K1 – K2) ÷ (L1 – L2). Because the slope is negative and one wishes to express the substi- tution rate as a positive quantity, a negative sign is attached to the slope:

MRTS = − K1 − K2 L1 − L2

= − ΔK ΔL

[7.8]

In the Deep Creek Mining Company example, ΔL = 3 − 4 = −1, ΔK = 750 − 500 = 250. Substituting these values into Equation 7.8 yields

MRTS = − 250 −1

= 250

Therefore, along Q = 29 between input combinations A and B, 250 bhp substituted for one worker.

It can be shown that the MRTS is equal to the ratio of the marginal products of L and K by using the definition of the marginal product (Equation 7.2). This definition yields ΔL = ΔQ/MPL and ΔK = ΔQ/MPK. Substituting these expressions into Equation 7.8 (and dropping the minus sign) yields

MRTS = ΔQ=MPK ΔQ=MPL

MRTS = MPL MPK

[7.9]

FIGURE 7.7 The Production Isoquant Curve—Deep Creek Mining Company

10 2 3 4 5 6 7 8

Labor input L (number of workers)

250

9 10

500

750

1,000

1,250

1,500

1,750

2,000

Q = 29

C

A

B

K1

K2

L1 L2

C ap

it al

in pu

t K

( br

ak e

ho rs

ep ow

er )

Captial-intensiveC

Labor-intensiveD

marginal rate of technical substitution (MRTS) The rate at which one input may be substituted for another input in producing a given quantity of output.

Chapter 7: Production Economics 247

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

DETERMINING THE OPTIMAL COMBINATION OF INPUTS As shown in the previous section, a given level of output can be produced using any of a large number of possible combinations of two inputs. The firm needs to determine which combination will minimize the total costs for producing the desired output.

Isocost Lines The total cost of each possible input combination is a function of the market prices of these inputs. Assuming that the inputs are supplied in perfectly elastic fashion in com- petitive markets, the per-unit price of each input will be constant, regardless of the amount of the input that is purchased. Letting CL and CK be the per-unit prices of inputs L and K, respectively, the total cost (C) of any given input combination is

C = CLL + CKK [7.10]

Example Isocost Determination: Deep Creek Mining Company (continued) In the Deep Creek Mining Company example discussed earlier, suppose that the cost per worker is $50 per period (CL) and that mining equipment can be leased at a price of $0.20 per brake horsepower per period (CK). The total cost per period of using L workers and equipment having K brake horsepower to produce a given amount of output is

C = 50L + 0.20K [7.11]

From this relationship, it can be seen that the mining of 55 tons of ore per pe- riod using five workers and equipment having 750 bhp would cost 50(5) + 0.20(750) = $400. However, this combination is not the only mixture of workers and equipment that costs $400. Any combination of inputs satisfying the equation

$400 = 50L + 0.20K

would cost $400. Solving this equation for K yields

K = $400 0:20

− 50 0:20

L

= $2,000 − 250L

Thus, the combinations L = 1 and K = 1,750; L = 2 and K = 1,500; L = 3 and K = 1,250 (plus many other combinations) all cost $400.

The combinations of inputs costing $400 can be represented as the isocost line in Figure 7.8 labeled “C = $400.” An isocost line exists for every possible total cost C. Solving Equation 7.11 for K gives the equation of each isocost line in Figure 7.8. Note that only the y-intercept C/0.20 changes as one moves from one isocost line to another.

K= C

0:20 −250L [7.12]

That is, all the isocost lines are parallel, each one having a slope of −250.

248 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Once the isoquants and isocosts are specified, it is possible to solve for the optimum combination of inputs. The production decision problem can be formulated in two dif- ferent ways, depending on the manner in which the production objective or goal is stated. One can solve for the combination of inputs that either

1. Minimizes total cost subject to a given constraint on output 2. Maximizes output subject to a given total cost constraint

Constrained cost minimization in Option 1 is the dual problem to the constrained output maximization problem in Option 2.

Minimizing Cost Subject to an Output Constraint Consider first the problem in which the director of operations desires to release to pro- duction a number of orders for at least Q(2) units of output. As shown in Figure 7.9, this constraint requires that the solution be in the feasible region containing the input com- binations that lie either on the Q(2) isoquant or on isoquants that fall above and to the right having larger output values (the shaded area). The total cost of producing the re- quired output is minimized by finding the input combinations within this region that lie on the lowest cost isocost line. Combination D on the C(2) isocost line satisfies this con- dition. Combinations E and F, which also lie on the Q(2) isoquant, yield higher total costs because they fall on the C(3) isocost line. Thus, the use of L1 units of input L and K1 units of input K will yield a (constrained) minimum cost solution of C

(2) dollars. At the optimal input combination, the slope of the given isoquant must equal the

slope of the C(2) lowest isocost line. As in the previous section, the slope of an isoquant is equal to dK/dL and

− dK dL

= MRTS = MPL MPk

[7.13]

FIGURE 7.8 Isocost Lines—Deep Creek Mining Company

C

A

B

D

10 2 3 4 5 6 7 8

Labor input L (number of workers)

9 10

500

750

1,000

1,250

1,500

1,750

2,000

C = $200

C = $300

C = $400

C = $500

C ap

it al

in pu

t K

( br

ak e

ho rs

ep ow

er )

250

Chapter 7: Production Economics 249

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Taking the derivative of the isocost equation (Equation 7.12), the slope of the isocost line is given by

dK dL

= − CL CK

[7.14]

Multiplying Equation 7.14 by (−1) and setting the result equal to Equation 7.13 yields

− dK dL

= − − CL CK

� �

= MPL MPK

Thus, the following equilibrium condition, the “equimarginal criterion,”

MPL MPK

= CL CK

or, equivalently, MPL CL

= MPK CK

[7.15]

must be satisfied in order for an input combination to be an optimal solution to the problem of minimizing cost subject to an output constraint. Equation 7.15 indicates that the marginal product per dollar input cost of one factor must be equal to the mar- ginal product per dollar input cost of the other factor.

Note in Figure 7.10 that maximizing output subject to a feasible region demarcated by the Q(2) cost constraint yields exactly the same (L1, K1) optimal input combination that satisfies the equimarginal criterion.

A FIXED PROPORTIONS OPTIMAL PRODUCTION PROCESS The previous section analyzed the least-cost combination of divisible inputs in variable proportions production, where one input substituted continuously for another. However, Deep Creek Mining’s production choices involve indivisible capital equipment, such as one small or one large mining drill and a predetermined number of workers to run the

FIGURE 7.9 Cost Minimization Subject to an Output Constraint

K1

Input L (units)

L1

Q(1) Q(2)

C(1) C(2) C(3)

E

D F

Feasible region

In pu

t K

( un

it s)

250 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

chosen equipment. Similarly, an auto fender stamping machine in an assembly plant must be used in fixed proportion to labor and sheet metal supplies. And three hours of setup, maintenance, and cleaning may be required to support a five-hour printing press run. Three additional hours of work by maintenance personnel would be required for a second press run, and a third shift of maintenance workers would be required for 24-hour print- ing operations. Although a higher output rate can be achieved by scaling up all the inputs, each of these production processes is one of fixed, not variable, proportions.

Linear programming techniques are available to determine the least-cost process for fixed proportions production. The Deep Creek Mining Company example can be used to illustrate the graphical approach to finding such a solution.

Production Processes and Process Rays A production process can be defined as one in which the inputs are combined in fixed proportion to obtain the output. By this definition, a production process can be repre- sented graphically as a ray through the origin having a slope equal to the ratio of the number of units of the respective resources required to produce one unit of output. Three production process rays for Deep Creek Mining are shown in Figure 7.11. Along Process Ray M1, the inputs are combined in the ratio of two workers to a 1,250-bhp dril- ling machine. Hence, Ray M1 has a slope of 625 bhp per mine worker.

FIGURE 7.10 Output Maximization Subject to a Cost Constraint

Feasible region

In pu

t K

( un

it s)

K1

Input L (units)

L1

Q(1) Q(2)

Q(3)

C(1) C(2) C(3)

B

A

C

Example Cost Minimization: Deep Creek Mining Company (continued) Suppose one is interested in finding the combination of labor input and capital equipment that minimizes the cost of producing at least 29 tons of ore. Assume that the isocost lines are the ones defined by Equation 7.11 and graphed in Figure 7.8 earlier in this section. Figure 7.11 combines several isoquants and isocost lines for the ore-mining problem. The shaded area in the graph represents the set of feasible input combinations, that is, those labor and capital production processes that yield at least Q = 29 tons of output. Processes M2 and M3 minimize the cost of producing 29 tons at $300. M1 imposes higher costs of $350.

production process A fixed-proportions production relationship.

Chapter 7: Production Economics 251

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Operating multiple production processes like M1, M2, and M3 can offer a firm flexibil- ity in dealing with unusual orders, interruptions in the availability of resources, or bind- ing resource constraints. However, not all fixed-proportions production processes are equally efficient. The firm will prefer to use one or two production processes exclusively if they offer the advantage of substantial cost savings. Mine 1 employs process M1 to produce 29 tons with two workers and a 1,250-bhp drilling machine at a total cost of 50(2) + 0.20(1,250) = $350 or $350/29 = $12.07 per ton. Mine 2 uses a more labor- intensive process (M2) with three workers and a smaller 750-bhp machine and incurs a lower total cost of $300. Mine 2 is the benchmark operation for Deep Creek in that this M2 process produces 29 tons at minimum cost—specifically, $300/29 = $10.34 per ton.

MEASURING THE EFFICIENCY OF A PRODUCTION PROCESS Mine 1 with production process M1 is said to be allocatively inefficient because it has chosen the wrong input mix; the mine has allocated its input budget incorrectly. Its 1,250-bhp machine is too large for the number of workers hired and the output desired. By producing 29 tons of output for $350 relative to the lowest cost benchmark at $300, process M1 exhibits only $300/$350 = 85.7 percent allocative efficiency.

In addition to allocative inefficiency involving the incorrect input mix, a production operation can exhibit technical inefficiency. For example, the industrial engineering indi- cated by the production isoquants in Figure 7.12 suggests that the process M3 also should be capable of producing 29 tons. The “C = $300” isocost line is tangent to the boundary of the feasible region (i.e., the “Q = 29” isoquant) at not only three workers and a 750-bhp machine (M2), but also at four workers and a 500-bhp machine (M3). In principle, both production processes yield the desired 29 tons of ore at a minimum total cost of $300 and will thereby satisfy the condition in Equation 7.15.

FIGURE 7.11 A Fixed Proportions Production Decision—Deep Creek Mining Company

0 1 2 3 4 5 6 7 8

Labor input L (number of workers)

250

9 10

500

750

1,000

1,250

1,500

1,750

2,000

C =250

C =300

C =350

Q =16

Q =29 Q=44

C ap

it al

in pu

t K

( br

ak e

ho rs

ep ow

er )

Feasible region

M3

M2

M1

D

C

A

B

allocative efficiency A measure of how closely production achieves the least-cost input mix or process, given the desired level of output.

252 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

However, suppose Mine 3 has been unable to achieve more than 27 tons of output. Although it has adopted a least-cost process, Mine 3 would then be characterized as tech- nically inefficient. In particular, Mine 3 exhibits only 27 tons/29 tons = 93 percent technical efficiency by comparison to the benchmark plant. Despite adopting the least- cost process, Mine 3’s 93 percent technical efficiency may be inadequate. Benchmark plants often do substantially better, with many processes meeting 98 percent and 99 percent of their production goals.

Overall production efficiency is defined as the product of technical, scale, and alloca- tive efficiency. If a 100 percent scale-efficient plant has 93 percent technical efficiency and 85.7 percent allocative efficiency, then its overall production efficiency is 0.93 × 0.857 = 0.797, or 79.7 percent. Your job as an operations manager might be to decide which least-cost process Mine 1 in Figure 7.11 should now adopt. Because M2 and M3 are both allocatively efficient for 29 tons of output, but process M3 experienced technical inefficiency problems resulting in an inability to realize its maximum potential output, Process M2 would be preferred.

RETURNS TO SCALE An increase in the scale of production consists of a proportionate increase in all inputs simultaneously. The proportionate increase in output that results from the given propor- tionate increase in all the inputs is defined as the physical returns to scale. Suppose, in the Deep Creek Mining Company example, one is interested in determining the effect on the number of tons of ore produced (output) of a 1.50 factor increase in the scale of production from a given labor-capital combination of four workers and equipment having 500 bhp. A 1.50 factor increase in the scale of production would constitute a labor-capital combination of 4 × 1.5 = 6 workers and equipment having 500 × 1.5 = 750 bhp. From Table 7.1, note that the labor-capital combination of four workers and 500 bhp yields 29 tons of output, whereas the combination of six workers and 750 bhp yields 60 tons of output. Output increased by the ratio of 60/29 = 2.07. Thus, a 1.50 factor increase in input use has resulted in more than a 1.50 factor output increase (specifically, 2.07).

Example GM’s A-Frame Supplier Achieves 99.998 Percent Technical Efficiency Continuous quality improvement initiatives frequently raise the standard of excel- lence for which technically inefficient plants must strive. Just-in-time delivery sys- tems, for example, accentuate the need for high reliability to produce on time as promised with near zero defects. One A-frame supplier to General Motors assem- bly plants reduced defective parts to five per million (i.e., 0.002 of 1 percent) and agreed to pay a $4,000 per minute “chargeback” for any late deliveries that cause GM assembly line delays. That figure represents the $80,000 per hour direct cost of the 2,000 manufacturing employees on side-by-side assembly lines, plus $120,000 for the time-and-a-half overtime labor to catch up the 70 vehicles of lost produc- tion, plus $26,000 shipping delay costs, plus $14,000 for utilities equal to $240,000 per hour. Under such enormous penalty costs, an auto component supplier to GM must constantly monitor and proactively solve production problems before they arise in order to ensure near 100 percent technical efficiency.

technical efficiency A measure of how closely production achieves maximum potential output given the input mix or process.

overall production efficiency A measure of technical and allocative efficiency.

returns to scale The proportionate increase in output that results from a given proportionate increase in all the inputs employed in the production process.

Chapter 7: Production Economics 253

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Measuring Returns to Scale An increase in the scale of production can be represented graphically in a two- dimensional isoquant map, as is shown in Figure 7.12. Increasing the scale of production by a factor of λ = 2 from the combination of 101 units of input L and 1001 units of input K to 20 units of input L and 200 units of K results in an increase in the quantity of out- put from Q(1) to Q(2). Three possible relationships that can exist between the increase in inputs and the increase in outputs are as follows:

1. Increasing returns to scale: Output increases by more than λ; that is, Q(2) > λQ(1). 2. Decreasing returns to scale: Output increases by less than λ; that is, Q(2) < λQ(1). 3. Constant returns to scale: Output increases by exactly λ; that is, Q(2) = λQ(1).

Figure 7.12 illustrates three different production functions that exhibit these three types of returns to scale. In Panel (a), showing increasing returns to scale, doubling input L from 10 to 20 units and input K from 100 to 200 units yields more than double the

Example Technical and Allocative Efficiency in Commercial Banks at BB&T7

Wave after wave of bank merger activity may be motivated by the potential for sub- stantial improvements in operating efficiency. Combining loan officers, facilities, and deposits of various kinds, the representative commercial bank in the United States “produces” only 63 percent of the loan value not in default (so-called “current status loans” or “performing loans”) that the most efficient benchmark banks produce. In contrast, natural gas-fired power plants average 93 percent overall efficiency. The problem (and opportunity for improvement) in commercial banks is twofold. First, some banks adopt inefficient processes such as allowing the bor- rower to pick one senior loan officer who will review and approve or disapprove the loan application versus two anonymous loan officers assigned by the bank. Linear programming studies show that allocative efficiency in U.S. commercial banking averages only 81 percent, meaning that the least-cost process is 19 per- cent cheaper. Best practices benchmarking implies one bank may need to imitate another bank’s borrower-screening or loan-monitoring processes.

When several banks do manage to adopt identical least-cost processes, yet one produces more current-status loans or larger performing loans than the others, the maximum feasible potential output in that type of institution can be identified. Technical efficiency then measures the observed bank output divided by the maxi- mum potential output of the benchmark bank with identical processes. The smaller a bank’s loan value not in default, the lower the technical efficiency. The represen- tative commercial bank in the United States is only 78 percent technically efficient.

Bank takeovers, buyouts, and mergers often result in a concerted effort to im- prove allocative and technical efficiency. Afterward, the so-called “bank [in]effi- ciency ratio” of non-interest operating expenses (e.g., headcount) to net interest plus fee income often declines substantially. As a result, capitalized value often then rises enough to allow an acquirer like Branch Banking and Trust (BB&T) to recover a merger premium of up to 20–30 percent paid for Pittsburgh National in excess of the takeover target bank’s premerger value.

7Based on D. Wheelock and P. Wilson, “Evaluating the Efficiency of Commercial Banks,” St. Louis Federal Reserve Review (July/August 1995), pp. 39–52; and A. Kleit and D. Terrell, “Measuring Potential Efficiency Gains from Dereg- ulation of Electricity Generation,” Review of Economics and Statistics (August 2001), pp. 523–550.

254 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

amount of output (i.e., an increase from 100 to 250). In Panel (b), showing decreasing returns to scale, a similar doubling of two inputs, L and K, yields less than double the amount of output (i.e., an increase from 10,000 to 15,000). Finally, in Panel (c), showing constant returns to scale, a similar doubling of inputs L and K yields exactly double the amount of output (i.e., an increase from 1,000 to 2,000).

Increasing and Decreasing Returns to Scale Many firm-level production functions are characterized by first increasing and then de- creasing returns to scale. A number of industrial engineering arguments have been pre- sented to justify this inconsistency. One major argument given for initially increasing returns is the opportunity for specialization in the use of capital and labor. Equipment that is more efficient in performing a limited set of tasks can be substituted for less- efficient all-purpose equipment. Similarly, the efficiency of workers in performing a small number of related tasks is greater than that of less highly skilled, but more versatile, workers. Decreasing returns to scale thereafter often arises from the increasingly complex problems of coordination and control faced by management as the scale of production is increased. For example, managers may be limited in their ability to transmit and receive status reports over a wider and wider span of control.

The Cobb-Douglas Production Function A somewhat simpler case is the Cobb-Douglas production function that has returns to scale determined by the sum of the parameters (β1 + β2) in the equation:

Q = αLβ1Kβ2 [7.16]

If β1 + β2 is less than, equal to, or greater than 1, the Cobb-Douglas production function will exhibit decreasing, constant, or increasing returns, respectively.

The multiplicative exponential Cobb-Douglas function can be estimated as a linear regression relation by taking the logarithm of Equation 7.16 to obtain

log Q = log α + β1 log L + β2 log K [7.17]

Thus, once the parameters of the Cobb-Douglas model are estimated, the sum of the ex- ponents of the labor (β1) and capital (β2) variables can be used to test for the presence of increasing, constant, or decreasing returns to scale.

FIGURE 7.12 Production Isoquants Exhibiting Increasing, Decreasing, and Constant Returns to Scale

100

Input L1 (units)

Q(2) = 250

10 20

200

300

Q(1) = 100

30

(a) Increasing returns to scale

Input L2 (units)

Q(2) = 15,000

Q(1) = 10,000

(b) Decreasing returns to scale

Input L3 (units)

Input K3 (units)

Q(2) = 2,000

Q(1) = 1,000

(c) Constant returns to scale

Input K2 (units)

Input K1 (units)

100

200

300

100

200

300

10 20 30 10 20 30

Chapter 7: Production Economics 255

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Empirical Studies of the Cobb-Douglas Production Function in Manufacturing In their original study, Cobb-Douglas fitted a production function of the form in Equa- tion 7.16 to indices of production Q, labor L, and capital K over time in the U.S. manufacturing sector. Q was an index of physical volume of manufacturing; L was an index of the average number of employed wage earners only (i.e., salaried employees, of- ficials, and working proprietors were excluded); and K was an index of the value of plants, buildings, tools, and machinery reduced to dollars of constant purchasing power. With the sum of the exponents restricted to one (constant returns to scale), the following function was obtained:

Q = 1.01L.75K.25 [7.18]

In later studies, Cobb-Douglas made several modifications that altered their results somewhat. These modifications included revisions in the output and labor indices, re- moving the secular trend from each index by expressing each yearly index value as a percentage of its overall trend value, and dropping the assumption of constant returns to scale. With these modifications, the estimated production function for the manufacturing sector was

Q = 0.84L.63 K.30 [7.19]

A 10 percent increase in labor input results in about a 6 percent increase in output, and a 10 percent increase in capital input results in approximately a 3 percent increase in output. Also, the resulting sum of the exponents of the labor and capital variables is slightly less than 1, which indicates the presence of decreasing returns to scale in the broadly defined manufacturing sector.

A Cross-Sectional Analysis of U.S. Manufacturing Industries Cross-sectional data have also been used to estimate Cobb-Douglas production functions for 18 U.S. manufacturing industries. Using aggregate data on plants located within each state, John Moroney estimated the following three-variable model:

Q = αLβ1p L β2 n K

β3 [7.20]

where Q is the value added by the production plants, Lp is production worker work hours, Ln is nonproduction work years,

8 and K is gross book values of depreciable and depletable assets.9 The results for several of the industries are shown in Table 7.4. The sum of the exponents (β1 + β2 + β3) ranged from a low of 0.947 for petroleum to a high of 1.109 for furniture. In 13 of the 18 industries studied, the statistical tests showed that the sum of the exponents was not significantly different from 1.0. This evidence supports the hy- pothesis that most manufacturing industries exhibit constant returns to scale.

8Nonproduction workers are management and other staff personnel. 9“Book values” of assets are the historic values of these assets as they appear on the balance sheet of the firm. Book values may differ significantly from current replacement values and hence may overstate or understate the actual amount of capital employed in the firm.

256 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE 7.4 PRODUCTION ELASTICITIES FOR SEVERAL INDUSTRIES

INDUSTRY

CAPITAL ELASTICITY*

b1

PRODUCTION WORKER ELASTICITY

b2

NONPRODUCTION WORKER ELASTICITY

b3

SUM OF ELASTICITIES b1 + b2 + b3

Food and beverages .555 .439 .076 1.070*

(.121) (.128) (.037) (.021)

Textiles .121 .549 .335 1.004

(.173) (.216) (.086) (.024)

Furniture .205 .802 .103 1.109*

(.153) (.186) (.079) (.051)

Petroleum .308 .546 .093 .947

(.112) (.222) (.168) (.045)

Stone, clay, etc. .632 .032 .366 1.029

(.105) (.224) (.201) (.045)

Primary metals .371 .077 .509 .958

(.103) (.188) (.164) (.035)

Number in parentheses below each elasticity coefficient is the standard error.

*Significantly greater than 1.0 at the 0.05 level (one-tailed test).

Source: John R. Moroney, “Cobb-Douglas Production Functions and Returns to Scale in U.S. Manufacturing Industry,” Western Economic Journal 6, no. 1 (December 1967), Table 1, p. 46.

Example Moneyball: A Production Function for Major League Baseball10

Team sports such as major league baseball are similar to other enterprises in that they attempt to provide a product (team victories) by employing various skills of team members. In acquiring team members through trades, the free agent market, and minor leagues/colleges, the owner is faced with various input mix trade-offs. For example, a baseball team owner may have to decide whether to trade a starting pitcher to obtain a power hitter or whether to sign a free agent relief pitcher in ex- change for a frequent base stealer. These decisions are all made in the context of a baseball production function, subject to various constraints (e.g., budgetary limits and penalties, league rules on recruiting and transfers, etc.). Michael Lewis’s 2004 book Moneyball argued that most major league baseball teams in the United States were not implementing the most efficient production mix of players, and then ex- plained how the Oakland Athletics had uncovered this production inefficiency and subsequently outperformed their competitors using only a minimal salary budget.

In an attempt to quantify the factors that contribute to winning, a Cobb- Douglas production function for baseball was developed using data from 26 major league baseball teams. Output (Q) was measured by team victories. Inputs (X1, X2, X3, etc.) from five different categories were included in the model:

• Hitting. This factor involves two different subskills: hitting frequency, as mea- sured by the team batting average, and hitting with power, as measured by the team’s home runs. Slugging percentage, which accounts for the additional con- tribution of doubles and triples to runs scored, has proven even more effective

(Continued)

Chapter 7: Production Economics 257

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

in predicting wins. Moneyball showed that the role of walks was being ignored and argued for an on-base percentage rather than slugging percentage. The A’s led the American League in walks in 1999 and 2001 and were second or third in 2000, 2002 and 2004.

• Running. One measure of speed is a team’s stolen base total. Oakland’s Ricky Henderson had a record-setting 130 stolen bases in one season!

• Defense. This factor also involves two subskills: catching those chances that the player is able to reach, as measured by fielding percentage, and catching difficult chances that many players would not be able to reach, as measured by total chances accepted. Because these two variables are highly correlated with each other (i.e., multicollinear), separate regressions were run with each variable.

• Pitching. The most obvious measure of the pitching factor is the team’s earned run average (ERA). However, ERA depends not only on pitching skill, but also on the team’s defensive skills. A better measure of pure pitching skills is the strikeouts-to-walks ratio for the pitching staff.

• Coaching. Teams often change managers when they are performing unsatisfac- torily, so this factor is thought to be important. However, the ability of a man- ager (coach) is difficult to measure. Two different measures are used in this study—the manager’s lifetime won-lost percentage and number of years spent managing in the major leagues. Separate regressions are run with each variable.

Finally, a dummy variable (NL = 0, AL = 1) was used to control for any differ- ences between leagues, such as the designated hitter rule.

The results of four regressions are shown in Table 7.5. Several conclusions can be drawn from these results:

TABLE 7.5 EMPIRICAL ESTIMATES OF BASEBALL PRODUCTION

FUNCTIONS

VARIABLE EQUATION

1 EQUATION

2 EQUATION

3 EQUATION

4

Constant .017 .018 .010 .008

League dummy −.002 −.003 .004 .003

Batting average 2.017* 1.986* 1.969* 1.927*

Home runs .229* .299* .208* .215*

Stolen bases .119* .120* .110* .112*

Strikeouts/walks .343* .355* .324* .334*

Total fielding chances 1.235 1.200

Fielding percentage 5.62 5.96

Manager W/L percentage −.003 −.004

Manager years −.004 −.002

R 2 (coef. of

determination) .789 .790 .773 .774

*Statistically significant at the 0.05 level.

(Continued)

258 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

SUMMARY

� A production function is a schedule, graph, or mathematical model relating the maximum quan- tity of output that can be produced from various quantities of inputs.

� For a production function with one variable input, the marginal product is defined as the incremental change in total output that can be produced by the use of one more unit of the variable input in the production process.

� For a production function with one variable input, the average product is defined as the ratio of total output to the amount of the variable input used in producing the output.

� The law of diminishing marginal returns states that, with all other productive factors held constant, the use of increasing amounts of the variable factor in the production process beyond some point will re- sult in diminishing marginal increases in total out- put. Increasing returns can arise with network effects especially involving information economy goods and industry standards.

� In the short run, with one of the productive factors fixed, the optimal output level (and optimal level of

the variable input) occurs where marginal revenue product equals marginal factor cost. Marginal rev- enue product is defined as the amount that an ad- ditional unit of the variable input adds to total revenue. Marginal factor cost is defined as the amount that an additional unit of the variable in- put adds to total cost.

� A production isoquant is either a geometric curve or algebraic function representing all the various combinations of inputs that can be used in produc- ing a given level of output.

� The marginal rate of technical substitution is the rate at which one input may be substituted for another input in the production process, while total output remains constant. It is equal to the ratio of the marginal products of the two inputs.

� In the long run, with both inputs being variable, minimizing cost subject to an output constraint (or maximizing output subject to a cost constraint) requires that the production process be operated at the point where the marginal product per dollar input cost of each factor is equal.

1. Hitting average contributes almost six times as much as pitching to a team’s success. This finding tends to contradict conventional wisdom, which says that pitching and defense win championships.

2. Home runs contribute about twice as much as stolen bases to a team’s success. 3. Coaching skills are not significant in any of the regression equations. 4. Defensive skills are not significant in any of the regression equations. 5. Finally, the sums of the statistically significant variables in each of the four

equations range from 2.588 to 2.709. Because these are all much greater than 1.0, the baseball production functions examined all exhibit increasing returns to scale. Better hitting and on-base and running skills yield more than propor- tional increases in games won. Because of the link between winning and atten- dance, ballplayers with these characteristics also yield greater revenues, especially to teams that contend for division and league championships.

10Based on Charles E. Zech, “An Empirical Estimation of a Production Function: The Case of Major League Baseball,” The American Economist 25, no. 2 (Fall 1981), pp. 19–23; Michael Lewis, Moneyball (New York: Norton, 2004); John Hakes and Ray Sauer,”An Economic Evaluation of the Moneyball Hypothesis,” Journal of Economic Perspectives (Summer 2006), pp. 173–185; and “The Real Most Valuable Players,” Wall Street Journal Online (2007).

Chapter 7: Production Economics 259

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

� The degree of technical efficiency of a production process is the ratio of observed output to the max- imum potentially feasible output for that process, given the same inputs.

� The degree of allocative efficiency of a production process is the ratio of total cost for producing a given output level with the least-cost process to the observed total cost of producing that output.

� Physical returns to scale is defined as the propor- tionate increase in the output of a production

process that results from a given proportionate in- crease in all the inputs.

� The Cobb-Douglas production function, which is used extensively in empirical studies, is a multipli- cative exponential function in which output is a (nonlinear) increasing function of each of the in- puts, with the sum of the exponential parameters indicating the returns to scale.

Exercises 1. In the Deep Creek Mining Company example described in this chapter (Table 7.1), suppose again that labor is the variable input and capital is the fixed input. Specifi- cally, assume that the firm owns a piece of equipment having a 500-bhp rating. a. Complete the following table:

LABOR INPUT L (NO. OF

WORKERS)

TOTAL PRODUCT TPL (= Q)

MARGINAL PRODUCT

MPL

AVERAGE PRODUCT

APL

1 _____ _____ _____

2 _____ _____ _____

3 _____ _____ _____

4 _____ _____ _____

5 _____ _____ _____

6 _____ _____ _____

7 _____ _____ _____

8 _____ _____ _____

9 _____ _____ _____

10 _____ _____ _____

b. Plot the (i) total product, (ii) marginal product, and (iii) average product functions.

c. Determine the boundaries of the three stages of production.

2. From your knowledge of the relationships among the various production func- tions, complete the following table:

VARIABLE INPUT

L

TOTAL PRODUCT TPL (= Q)

AVERAGE PRODUCT

APL

MARGINAL PRODUCT

MPL

0 0 — —

1 _____ _____ 8

2 28 _____ _____

3 _____ 18 _____

4 _____ _____ 26

5 _____ 20 _____

6 108 _____ _____

7 _____ _____ −10

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

260 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

3. The amount of fish caught per week on a trawler is a function of the crew size assigned to operate the boat. Based on past data, the following production sched- ule was developed:

CREW SIZE (NUMBER OF WORKERS)

AMOUNT OF FISH CAUGHT PER WEEK (HUNDREDS OF LBS)

2 3

3 6

4 11

5 19

6 24

7 28

8 31

9 33

10 34

11 34

12 33

a. Over what ranges of workers are there (i) increasing, (ii) constant, (iii) de- creasing, and (iv) negative returns?

b. How large a crew should be used if the trawler owner is interested in maxi- mizing the total amount of fish caught?

c. How large a crew should be used if the trawler owner is interested in maxi- mizing the average amount of fish caught per person?

4. Consider Exercise 3 again. Suppose the owner of the trawler can sell all the fish caught for $75 per 100 pounds and can hire as many crew members as desired by paying them $150 per week. Assuming that the owner of the trawler is interested in maximizing profits, determine the optimal crew size.

5. Consider the following short-run production function (where L = variable input, Q = output):

Q = 6L3 − 0.4L3

a. Determine the marginal product function (MPL). b. Determine the average product function (APL). c. Find the value of L that maximizes Q. d. Find the value of L at which the marginal product function takes on its

maximum value. e. Find the value of L at which the average product function takes on its max-

imum value.

6. Consider the following short-run production function (where L = variable input, Q = output):

Q = 10L − 0.5L2

Suppose that output can be sold for $10 per unit. Also assume that the firm can obtain as much of the variable input (L) as it needs at $20 per unit. a. Determine the marginal revenue product function. b. Determine the marginal factor cost function. c. Determine the optimal value of L, given that the objective is to maximize

profits.

Chapter 7: Production Economics 261

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

7. Suppose that a firm’s production function is given by the following relationship:

Q = 2:5 ffiffiffiffiffiffi LK

p        ði:e:, Q = 2:5L:5K :5Þ

where Q = output L = labor input K = capital input

a. Determine the percentage increase in output if labor input is increased by 10 percent (assuming that capital input is held constant).

b. Determine the percentage increase in output if capital input is increased by 25 percent (assuming that labor input is held constant).

c. Determine the percentage increase in output if both labor and capital are increased by 20 percent.

8. Based on the production function parameter estimates reported in Table 7.4: a. Which industry (or industries) appears to exhibit decreasing returns to scale?

(Ignore the issue of statistical significance.) b. Which industry comes closest to exhibiting constant returns to scale? c. In which industry will a given percentage increase in capital result in the

largest percentage increase in output? d. In what industry will a given percentage increase in production workers re-

sult in the largest percentage increase in output?

9. Consider the following Cobb-Douglas production function for the bus transporta- tion system in a particular city:

Q = αLβ1Fβ2Kβ3

where L = labor input in worker hours F = fuel input in gallons K = capital input in number of buses Q = output measured in millions of bus miles

Suppose that the parameters (α, β1, β2, and β3) of this model were estimated using annual data for the past 25 years. The following results were obtained:

α = 0.0012 β1 = 0.45 β2 = 0.20 β3 = 0.30

a. Determine the (i) labor, (ii) fuel, and (iii) capital input production elasticities. b. Suppose that labor input (worker hours) is increased by 2 percent next year

(with the other inputs held constant). Determine the approximate percentage change in output.

c. Suppose that capital input (number of buses) is decreased by 3 percent next year (when certain older buses are taken out of service). Assuming that the other inputs are held constant, determine the approximate percentage change in output.

d. What type of returns to scale appears to characterize this bus transportation system? (Ignore the issue of statistical significance.)

e. Discuss some of the methodological and measurement problems one might encounter in using time-series data to estimate the parameters of this model.

262 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

10. Extension of the Cobb-Douglas Production Function—The Cobb-Douglas produc- tion function (Equation 7.16) can be shown to be a special case of a larger class of linear homogeneous production functions having the following mathematical form:11

Q = γ½∂K−ρ + ð1 − ∂ÞL−ρ�−ν=ρ

where γ is an efficiency parameter that shows the output resulting from given quantities of inputs; ∂ is a distribution parameter (0 ≤ ∂ ≤ 1) that indicates the division of factor income between capital and labor; ρ is a substitution parameter that is a measure of substitutability of capital for labor (or vice versa) in the pro- duction process; and ν is a scale parameter (ν > 0) that indicates the type of re- turns to scale (increasing, constant, or decreasing). Show that when ν = 1, this function exhibits constant returns to scale. [Hint: Increase capital K and labor L each by a factor of λ, or K* = (λ)K and L* = (λ)L, and show that output Q also increases by a factor of λ, or Q* = (λ)(Q).]

11. Lobo Lighting Corporation currently employs 100 unskilled laborers, 80 factory technicians, 30 skilled machinists, and 40 skilled electricians. Lobo feels that the marginal product of the last unskilled laborer is 400 lights per week, the marginal product of the last factory technician is 450 lights per week, the marginal product of the last skilled machinist is 550 lights per week, and the marginal product of the last skilled electrician is 600 lights per week. Unskilled laborers earn $400 per week, factory technicians earn $500 per week, machinists earn $700 per week, and electricians earn $750 per week.

Is Lobo using the lowest cost combination of workers to produce its targeted output? If not, what recommendations can you make to assist the company?

Case Exercise THE PRODUCTION FUNCTION

FOR WILSON COMPANY Economists at the Wilson Company are interested in developing a production func- tion for fertilizer plants. They collected data on 15 different plants that produce fertil- izer (see the following table).

Questions 1. Estimate the Cobb-Douglas production function Q = αLβ1Kβ2 , where Q = output;

L = labor input; K = capital input; and α, β1, and β2 are the parameters to be estimated.

2. Test whether the coefficients of capital and labor are statistically significant. 3. Determine the percentage of the variation in output that is “explained” by the

regression equation.

11See R. G. Chambers, Applied Production Analysis (Cambridge: Cambridge University Press, 1988).

Chapter 7: Production Economics 263

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

4. Determine the labor and capital estimated parameters, and give an economic in- terpretation of each value.

5. Determine whether this production function exhibits increasing, decreasing, or constant returns to scale. (Ignore the issue of statistical significance.)

PLANT OUTPUT

(000 TONS) CAPITAL ($000)

LABOR (000 WORKER

HOURS)

1 605.3 18,891 700.2

2 566.1 19,201 651.8

3 647.1 20,655 822.9

4 523.7 15,082 650.3

5 712.3 20,300 859.0

6 487.5 16,079 613.0

7 761.6 24,194 851.3

8 442.5 11,504 655.4

9 821.1 25,970 900.6

10 397.8 10,127 550.4

11 896.7 25,622 842.2

12 359.3 12,477 540.5

13 979.1 24,002 949.4

14 331.7 8,042 575.7

15 1064.9 23,972 925.8

264 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

7A APPENDIX

Maximization of Production Output Subject to a Cost Constraint, Advanced Material

Using graphical analysis, we illustrated in the chapter that the following condition (Equation 7.15)

MPL CL

= MPK CK

must be satisfied in determining the combination of inputs (L and K) that minimizes total cost subject to an output constraint. It turns out that the same result arises in max- imizing output subject to a cost constraint, the mathematical dual of the earlier con- strained minimization problem.

Given the production function to identify potential output possibilities,

Q = f(L, K) [7A.1]

and the cost constraint

C = CLL + CKK [7A.2]

we define an artificial variable λ (lambda) and form the Lagrangian function

LQ = Q − λ(CLL + CKK − C) [7A.3]

Differentiating LQ with respect to L, K, and λ and setting the (partial) derivatives equal to zero (condition for a maximum) yields

∂LQ ∂L

= ∂f ðL, KÞ

∂L − λCL = 0 [7A.4]

∂LQ ∂K

= ∂f ðL, KÞ

∂K − λCK − 0 [7A.5]

∂LQ ∂λ

= CLL + CKK − C = 0 [7A.6]

Recognizing that ∂f ðL, KÞ∂L = MPL and ∂f ðL, KÞ

∂K = MPK , solving Equations 7A.4 and 7A.5 for λ yields

λ = MPL CL

[7A.7]

λ = MPK CK

[7A.8]

265

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Setting Equations 7A.7 and 7A.8 equal to each other gives the optimality condition

MPL CL

= MPK CK

[7A.9]

Exercise 1. The output (Q) of a production process is a function of two inputs (L and K) and is given by the following relationship:

Q = 0.50LK − 0.10L2 − 0.05K2

The per-unit prices of inputs L and K are $20 and $25, respectively. The firm is interested in maximizing output subject to a cost constraint of $500. a. Formulate the Lagrangian function:

LQ = Q − λ(CLL + CKK − C)

b. Take the partial derivatives of LQ with respect to L, K, and λ, and set them equal to zero.

c. Solve the set of simultaneous equations in Part (b) for the optimal values of L, K, and λ.

d. Based on your answers to Part (c), how many units of L and K should be used by the firm? What is the total output of this combination?

e. Give an economic interpretation of the λ value determined in Part (c). f. Check to see whether the optimality condition (Equation 7A.9) is satisfied

for the solution you obtained.

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

266 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

7B APPENDIX

Production Economics of RenewableandExhaustibleNatural Resources, Advanced Material

Natural resource inputs pivotally affect the success of major industries in both devel- oped and developing economies. Timber used for home construction, coal and natural gas converted to electric power, and petroleum converted to gasoline are examples where a company’s profitability hinges on natural resource inputs. Natural resources are often subdivided into two categories: renewable and exhaustible resources. Renew- able resources like clean air, clean water, grazing land, timber, and fisheries often pres- ent common property and externality problems that must be analyzed and managed differently from the frequently immobile and therefore usually privatized exhaustible resources such as seabed manganese, coal fields, subterranean crude oil, and natural gas deposits. Public regulation is therefore very involved in decision making about re- newables. For example, migratory ocean fisheries of bluefin tuna are being overhar- vested such that their catch is declining at an alarming rate (see Figure 7B.1). Several bilateral trade negotiations and the United Nations have focused on this problem and proposed regulatory solutions.

RENEWABLE RESOURCES1 Fundamentally, all renewable resources are capital assets that must be analyzed with a dynamic stock-flow model where time (time to harvest and time to extinction) plays an explicit role. The optimal decision-making question with renewables is when to produce (that is, to harvest an oyster, net a fish, or cut down a tree) and when to let the resource grow another year to yield an even greater harvest later. In the following equation, the flow of harvested resources, h, is determined by harvest- ing effort, E, and by the remaining resource stock (capital stock), S—for example, the population size of the fishery or the cumulative board feet of the forest that remain unharvested:

h = f(E,S) [7B.1]

where diminishing positive marginal returns to harvesting effort are assumed (∂f/∂E) > 0, ∂2f/∂2 E < 0. Larger stocks left in the fishery, the forest, or the oyster bed imply a higher flow rate from harvesting for any given effort (∂f/∂S > 0) and therefore a lower harvest- ing cost per unit output.

1This section relies upon the excellent survey by Gardner Brown, “Renewable Resource Management and Use without Markets,” Journal of Economic Literature 38 (December 2000), pp. 875–914.

267

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

One way to grow a renewable resource is just to do less harvesting. More generally, the net rate of growth of a resource stock per unit of ΔS/Δt is the difference between the planned flow rate of harvest h and the biological growth function g(S).

ΔS Δt

= −h + gðSÞ [7B.2]

Just as it is quite natural to assume diminishing returns relative to harvesting effort, so, too, limitations of habitat, space, or nutrients introduce diminishing returns as a species grows—that is, ∂2g/∂2S < 0. This latter assumption is reflected in the accelerating decay as a species collapses (see Figure 7B.1 on tons of bluefish tuna caught) or, analogously, the cumulative board feet of lumber in a timber harvest. The eventually diminishing slope of this growth function ΔS/Δt measures the amount of additional fish caught or, analogously, lumber that comes from waiting a unit of time Δt rather than harvesting the resource immediately.

The potential revenue from waiting to harvest and selling the extra resources later, expressed as a percentage growth rate on the value of the timber or fishery stock (PΔS/ Δt)/PS, has a pivotal relation to the resource owner’s inflation-adjusted, risk-adjusted dis- count rate, r. The discount rate is the opportunity cost of waiting—that is, the interest that could be earned on the owner’s funds if they were invested at comparable risk. In the following Equation 7B.3 and in the lower panel of Figure 7B.2, we express this per- centage as a rate of return on investment (ROI) in harvesting. In particular, when the percentage growth of the resource stock is greater than the owner’s discount rate (at, for example, Point A in Figure 7B.2), the owner should harvest more (moving down along ROI toward D and up the yield hill in the top panel of Figure 7B.2 toward h *W=O). When the percentage growth of the resource stock is less than the discount rate,

FIGURE 7B.1 Collapse of Bluefin Tuna Stocks as Evidenced by Declining Total Catch from Increased Effort

1970 1975 1980 1985 1990 1995 2005

10,000

20,000

30,000

40,000

50,000

60,000

70,000

80,000

90,000

M et

ri c

to ns

Source: United.com, October 2009.

268 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the rate of harvest should be decreased as at point B in Figure 7B.2. The optimal capital stock equates the two at h *W=O and point D:

ROI = PΔS=t PS

= ΔS=Δt

S

� � = r [7B.3]

In general, positive discount rates imply that the optimal capital stock (the species population size) is smaller than would be associated with the maximum sustainable harvest, sometimes called the maximum sustainable yield (MSY) (at the peak of the top panel in Figure 7B.2). The reason is that the capital stock SMSY that yields a maxi- mum harvest hMSY drives the ROI in harvesting to zero (at point C). Any positive dis- count rate would exceed ROI at C and thereby imply that the resource owner should reduce the timber or fishery stock. For example, Figure 7B.2 illustrates that SW/O, which equates ROI and r, is smaller than SMSY.

However, it is important to reiterate that the rate of growth of S—that is, g(S)—may not be a constant but rather may be a positive function of S itself. Because of this faster biological replenishment of a fishery as its overall size increases, the ROI from harvesting

FIGURE 7B.2 Renewable Resource Growth, Maximum Sustainable Yield, and Optimal Capital Stock

hMSY

R at

e of

h ar

ve st

( lb

s/ ti

m e)

D is

co un

t ra

te (

% )

r

hW/O

SW/O

Maximum sustainable yield

SMSY

ROI withwaitingROI

A

B D E

C

h*

*

W

SW*

SW/O SMSY Resource capital stock

(lbs, board ft)

SW*

Opportunity cost of waiting

Resource capital stock (lbs, board ft)

w/o waiting

maximum sustainable yield (MSY) The largest production harvest that can be produced by the resource stock as a perpetuity.

Appendix 7B: Economics of Renewable and Exhaustible Natural Resources 269

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

would become larger than it would otherwise be. That is, ROIW/O, which represents the biological process without this extra boost from increasing size, would then shift to the right to ROIW. Consequently, the optimal capital stock increases to S *W. Note that the optimal rate of harvest h *W could then be smaller than the optimal rate of harvest h *W=O without this effect. This result is consistent with the intuition that waiting to harvest later will generate a larger and larger resource base from which to harvest (depending upon the particulars of the biological regeneration of the renewable population).

EXHAUSTIBLE NATURAL RESOURCES Some natural resources like coal, crude oil, natural gas, and diamonds are formed over tens of thousands of years. Although limited and fixed in this geological sense, more in- tense exploration and development can often locate additional resources. Nevertheless, eventually the supply of low-sulfur coal or sweet light crude oil or high-quality natural gas will be exhausted. Long before this happens, we usually discover the possibility of a replacement natural resource or a synthetic substitute. As we pointed out in Chapter 2, jojoba bean oil turned out to be a good natural substitute for sperm whale lubricant used with high-friction machinery like jet aircraft engines. Similarly, synthetic diamonds have substituted for natural diamonds in many industrial applications.

The analysis of exhaustible natural resources can be distinguished from renewable re- sources in two important ways: First, the net growth rate in Equation 7B.2 for renewable resources reduces to simply −h in the case of exhaustible resources, because the growth

Example Oyster Seedbed Replenishment on Chesapeake Bay Take, for example, the oyster harvesting decision. Oysters in the larval stage grow and multiply quickly under the right conditions but require shell or shell-like sur- faces (cultch) on which to attach. One excellent cultch alternative is the clean shell generated as a by-product from oyster packing plants. Unfortunately, overharvest- ing has destroyed the private incentive to replenish the oyster beds with clean cultch. Since deep-water oyster beds are common property, no individual harvester can fully appropriate the returns from replenishing the deep-water beds with clean cultch. Instead, each harvester has an incentive to reap the maximum yield in a downward spiraling race to exhaustion of the resource. This is the so-called “trag- edy of the commons.”

To prevent overharvesting and inadequate replenishment from destroying the oyster industry, Virginia and Maryland have regulated catch sizes and methods, and both states have provided a public subsidy to replenish the beds with clean shell cultch. Maine lobstermen abide by similar catch restrictions and adopt maintenance quotas largely through a voluntary association—an economic club— that co-manages the shared natural resource.

In other cases, like timber stands and wildlife management of game, the seedling and maturation of game process can be enhanced by replenishing in a different manner. Thinning the forest by harvesting large older trees and periodically re- moving the undergrowth allows the best immature trees and seedlings room and sunlight to grow. The same thing is true in many larger game animal species. Hence, the optimal rate of harvest can actually increase with more proactive for- estry and game management.

270 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

rate of the stock itself is exactly zero. Second, the optimal capital stock of a renewable resource does not vary when its price changes. This is easily seen in Equation 7B.3 where the prices, P, in both the numerator and denominator of the return on investment from waiting cancel each other out. That is, any price change that is unrelated to the size of the capital stock has no effect—none whatsoever—on the efficient rate of harvest or the optimal capital stock of renewable natural resources. In contrast, just the opposite is true for an exhaustible resource. Since there is no way for the natural resource to regenerate itself, the only reason to hold onto and not harvest coal or crude oil or natural gas today is if you, the owner, believe that the price is going to rise in the near future. Price changes and price change expectations are therefore the key to exhaustible resource decisions.

Let’s begin an analysis of this resource extraction problem by defining consensus ex- pectations for future prices in time period T(PT) as

PT = P0(1+r) T [7B.4]

where r is the real rate of interest (more precisely, the inflation-adjusted, risk-adjusted discount rate for our resource extraction decision). Dividing each unit of time t into n subperiods, the compound growth version of these consensus price expectations may be written as2

PT = P0½limð1+r=nÞnT � = P0erT n → ∞

[7B.5]

As before, we can express the harvest-now-or-wait decision in terms of the opportu- nity cost of waiting (the real rate of interest r) relative to the percentage rate of growth of resource prices:3

ΔPT=ΔT PT

= rP0erT

PT [7B.6]

which reduces, using Equation 7B.5, to

ΔPT=ΔT PT

= r [7B.7]

Equation 7B.7 states that as long as the expected price increase (say, 8 percent) exceeds the interest rate (say, 4 percent), one should leave the coal, oil, and natural gas in the ground and harvest later. If interest rates rise above this percentage growth rate of the exhaustible resource prices, the resource should be extracted and sold now.

Rearranging Equation 7B.5 to solve for the current resource price we obtain

P0 = PT/e −rT [7B.8]

which has some interesting interpretations as more than a continuous time formula for calculating the present value of future prices. First, T can be interpreted as the time to exhaustion of the resource at the current rate of use. Hence, new coal field or oil field discoveries that lead to an increase in the proven reserves of coal, oil, or natural gas raise T, which according to Equation 7B.8, must result in a lower current market price P0. Similarly, more energy dependence and a faster rate of use lowers T and raises the cur- rent market price P0. Finally, since optimal use leads to persistent appreciation of the coal or oil or gas at the rate r, we eventually expect synthetic substitutes to emerge as

2The number e is 2.7183. . . , the base of the natural logarithms. 3 This expression is based on the calculus result that de

rT

dT = re rT .

Appendix 7B: Economics of Renewable and Exhaustible Natural Resources 271

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

these natural resources near exhaustion. That is, at a sufficiently high price of crude oil and the resulting high price for gasoline (say, $4.00 per gallon in July 2008), the return for R&D investment to develop alternative fuel sources and products climbs steeply. Suf- ficient R&D investment often leads to technological breakthroughs like hybrid electric cars that then lower PT directly, and through Equation 7B.8, lower the current resource price P0 as well.

So, high prices for exhaustible resources are inevitable, eventually. The goods news is, however, that these high prices often set in motion the discovery of substitutes that pre- vent the actual exhaustion of the coal, oil, or natural gas resource. The trick for an astute exhaustible resource owner is to withhold the resource long enough to create upward pressure on prices that are just below the price level that would set off the discovery and adoption of substitutes. Only in this way can a premature price collapse be pre- vented prior to the planned time of exhaustion of the resource stock.

Example Saudi Arabian Oil Minister Plays a Waiting Game4

No resource owner on Earth has more natural resources than the Royal King- dom of Saud. The proven oil reserves in Saudi Arabia are extensive enough to last, at the current rates of extraction, for almost another 66.5 years (see Figure 7B.3, A). In particular, Saudi Arabia extracts about 10 million barrels per day, or 4 billion barrels per year against proven reserves of 264 billion barrels, roughly 25 percent of the entire world’s supply. In contrast, the United States, which ex- tracts about 6.5 million barrels per day, has proven reserves only one-tenth as big—just 28 billion barrels. This is less than 3 percent of the world’s supply. Without additional exploration and development, the United States will run out of oil in a scant 12 years. Saudi Arabia therefore has an objective shared with American consumers: both want the rate of price increase of crude oil (and of gasoline) to remain smaller than the rate preferred by an Oklahoma or Texas oilman. U.S. oil interests are continuously exploring and developing new deposits of crude oil and natural gas, but that doesn’t alter the fact that their exhaustible resources will run out sooner rather than later. Therefore, their incentive is to urge policies that will raise the price quickly.

As a result, President Ronald Reagan sent Vice President George H.W. Bush to Saudi Arabia in 1986 to urge the Saudis to raise the price of crude oil. The Reagan and Bush Sr. presidencies never developed synthetic oil from abundant U.S. re- serves of shale oil, as proposed by President Jimmy Carter. In addition, the William Clinton and George W. Bush presidencies never adopted a national energy policy that would support replacing gasoline internal combustion engines with hydrogen fuel cell-powered or all-electric vehicles. As a result, the United States lacks the in- frastructure of hydrogen pump stations that would be required to implement hy- drogen fuel cells for autos. In recent years, the Saudis have massively increased production from approximately 3 billion barrels per year in 2002 to almost 4 bil- lion barrels per year in 2008 (compare Figure 7B.3, A and B). Gasoline today re- mains less expensive (about 80 percent less) than the full cost of producing and distributing hydrogen fuel for powering cars, just as the Saudis intended.

(Continued)

272 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Whenever crude prices rose in 2002–2009 above the cost-covering price of corn-based ethanol ($70 per barrel in 2008), Saudi Arabia increased production to slow the rate of the price increase. This was as true in the distant past as in the current decade. Saudi Arabia’s initial share of OPEC’s output at the formation of the cartel in 1973 was 32 percent. Saudi Arabia continuously increased its rate of extraction until, at peak crude oil prices in 1981, it was producing almost 47 per- cent of OPEC’s total output—9.6 million of the industry’s 20 million barrels of oil per day. Again, these policies were designed to discourage the development of sub- stitutes by slowing down the rate of price increases of crude oil. Any other ex- haustible resource owner with nearly a 100-year supply of proven reserves (like OPEC members Iran, Iraq, Kuwait, Venezuela, and United Arab Emirates (UAE) as listed in Figure 7B.3A) would want the same thing.

4Based on estimates from the International Energy Agency; OPEC, Annual Statistical Bulletin (2002–2009); and “Why the U.S. Is Still Hooked on Oil Imports,” Wall Street Journal (March 18, 2003), p. A1.

FIGURE 7B.3A Years of Proven Oil Reserves Remaining to Exhaustion (at Current Rates of Extraction)

0 50 100 150 Years Remaining

Angola

China

Qatar

Canada

United States

Nigeria

Kazakhstan

Libya

Russia

UAE

Venezuela

Kuwait

Iran

Iraq

Saudi Arabia

December 2008, Billion Barrels of Reserves

66.566.5

86.5

100+

99.6

100+

89.7

21.8

64.6

70.0

45.6

12.4

24.1

54.1

11.1

19.7

264.1

Source: The Economist, June 13, 2009, p. 101.

Appendix 7B: Economics of Renewable and Exhaustible Natural Resources 273

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Exercises 1. Using Figure 7B.3, contrast the annual rate of oil extraction of the United States and of Saudi Arabia in 2002 and 2008. Explain why one country’s output in- creased by 33 percent and the other’s was flat.

2. With interest rates at historic lows in the United States, what is the effect on the optimal rate of extraction for a Texas oilfield owner? Explain the intuition that supports your answer.

3. Explain the concept of maximum sustainable yield for a fishery. Is maximum sus- tainable yield required to preserve biodiversity? Is it the most efficient rate of har- vest for a renewable natural resource?

FIGURE 7B.3B

0 20 40 60 80 100 120 140 160 180 200 220 240 260 Years Remaining

Algeria Brazil Kazakhstan Azerbaijan Canada Oman Angola Indonesia Britain India Yemen Australia Argentina

Norway Qatar China Nigeria Mexico Libya United States Russia Venezuela Iran Kuwait

Iraq UAE

Saudi Arabia

December 2002, Billion Barrels of Reserves

85 100+ 100+ 100+

67 64 19 11 57 22 31 20 56 8 18 18 28 64 9 16 20 10 6 18 24 14 10

Source: The Economist, June 29, 2002, p. 102; BP Statistical Review of World Energy, 2002.

274 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

8 CHAP T E R

Cost Analysis CHAPTER PREVIEW Economic cost refers to the cost of attracting a resource from its next best alternative use (the opportunity cost concept). Managers seeking to make the most efficient use of resources to maximize value must be concerned with both short-run and long-run opportunity costs. Short-run cost- output relationships help managers to plan for the most profitable level of output, given the capital resources that are immediately available. Long-run cost-output relationships involve attracting additional capital to expand or contract the plant size and change the scale of operations. Achieving minimum efficient scale is often the key to a successful operations strategy.

MANAGERIAL CHALLENGE US Airways Cost Structure1

US Airways Corporation (USAir) was formed through the merger of several diverse regional airlines, including Allegheny, Mohawk, Lake Central, Pacific Southwest, Piedmont Airlines, and America West. Although these mergers led to a “national” competitor in the airline industry, USAir’s market strength remained in the Northeast and far West. USAir has a major or dominant presence in Charlotte, Philadelphia, Washington, New York, Boston, and Phoenix.

Unlike the much more successful Southwest Airlines, which flies only one type of plane, the Boeing 737, US Airways possesses a diverse fleet of aircraft. This diver- sity results in higher costs of maintenance and crew training and a much more complex crew scheduling problem. In addition, USAir is burdened with restrictive work rules that increase labor costs, which account for 40 percent of total airline operating costs. The net result is that USAir’s cost per available seat mile is close to the highest in the industry at 10.89 cents compared to 11.62 cents for Delta, 10.86 cents for Northwest, 9.80 cents for American Airlines, 10.56 cents for Continental, 7.7 cents for Southwest, and 6.74 cents for JetBlue. In 2005, USAir merged with and tried to adopt the practices of America West, a low-cost discounter with Phoenix and Las Vegas hubs and a 7.68 cent operating cost per avail- able seat mile.

Because of the traditionally weak competition in its Northeast market stronghold, USAir has the highest av- erage passenger fare received for each revenue passenger mile flown: 18.8 cents, compared with 14.21 cents for United, 15.18 cents for American, 12.76 cents for Con- tinental, and 12.37 cents for Southwest. The combina- tion of high fares and high costs per available seat mile invited competition. Continental Airlines, after emerg- ing from bankruptcy in 1999 with a new, lower cost structure, announced a major restructuring of its East Coast route system to compete head on with USAir in much of its core business area. In addition, Southwest has entered some of USAir’s traditional markets (partic- ularly Baltimore/Washington).

In May 2008, American Airlines announced that it would immediately begin charging $25 per bag on all AA flights, not for extra luggage but for the first bag! Crude oil had doubled from $70 to $130 per barrel in the previous 12 months, and jet fuel prices had acceler- ated even faster. AA’s new baggage policy applied to all ticketed passengers except first class and business class. On top of incremental airline charges for sandwiches and snacks introduced the previous year, this new announce- ment stunned the travel public. Previously, only a few deep-discount U.S. carriers with very limited route struc- tures such as People Express had charged separately for

275

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

THE MEANING AND MEASUREMENT OF COST In its most elementary form, cost simply refers to the sacrifice incurred whenever an ex- change or transformation of resources takes place. This association between forgone oppor- tunities and economic cost applies in all circumstances. However, the appropriate manner to measure costs is a function of the purpose for which the cost information is to be used.

Accounting versus Economic Costs Accountants have been primarily concerned with identifying highly stable and predict- able costs for financial reporting purposes. As a result, they define and measure cost by

both food and baggage service. Since American Airlines and many other major carriers had belittled that policy as part of their overall marketing campaign against deep discounters, AA executives faced a dilemma.

Jet fuel surcharges (introduced in the winter of 2008) had recovered the year-over-year average variable cost increase for jet fuel expenses, but incremental costs

remained uncovered. A quick back-of-the-envelope cal- culation outlines the problem.

If total variable costs for a 500-mile flight on a 180- seat 737-800 rise from $22,000 in 2007 Q2 to $36,000 in 2008 Q2 because of $14,000 of additional fuel costs, then it would not be unexpected that a set of competi- tively priced carriers would seek to recover $14,000/ 180 = $78 per seat in jet fuel surcharges. The $78 in- crease in average variable cost was added to the prices for each fare class. For example, the $188 Super Saver airfare restricted to 14-day advance purchase and Satur- day night stay overs would go up to $266. Class M air- fares requiring 7-day advance purchase but no Saturday stay overs would rise from $289 to $367. Full coach economy airfares without purchase restrictions would rise from $419 to $497, and so on.

Similarly, by 2008, the incremental jet fuel cost for luggage transported 500 miles had risen to approxi- mately $1 per pound. A first suitcase was traveling free under the prior baggage policy as long as it weighed under 42 pounds. The representative suitcase on AA was found to weigh 25.4 pounds. Therefore, the new $25 baggage fee for the first bag on American Airlines just covers its incremental fuel cost.

Discussion Questions

� Critique the above argument for charging $25 per bag from each of the following perspec- tives: (a) demand, (b) cost, (c) public relations, and (d) business strategy.

1Based on “UAL Latest Cost Cuts,” Wall Street Journal (May 12, 2005), p. A10.

MANAGERIAL CHALLENGE Continued

© Ja m es

La ur itz /D ig ita lV

is io n/ Ge tty

Im ag es

276 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the known certain historical outlay of funds. Thus, the price paid for commodity or ser- vice inputs, expressed in dollars, is one measure of the accounting cost. Similarly, the interest paid to bondholders or lending institutions is used to measure the accounting cost of funds to the borrower.

Economists, on the other hand, have been mainly concerned with measuring costs for decision-making purposes. The objective is to determine the present and future costs of resources associated with various alternative courses of action. Such an objective requires a consideration of the opportunities forgone (or sacrificed) whenever a resource is used in a given course of action. So, although both the accounting cost and the economic cost of a product will include such explicit costs as labor, raw materials, supplies, rent, inter- est, and utilities, economists will also include the implicit opportunity costs of time and capital that the owner-manager has invested in the enterprise. The opportunity cost of the owner’s time is measured by the most attractive salary offer that the owner could have received by applying his or her talents, skills, and experience in the management of a similar (but second-best) business owned by someone else. Similarly, the opportu- nity cost of the capital is measured by the profit or return that could have been received if the owner had chosen to employ capital in his or her second-best (alternative) invest- ment of comparable risk.

Economic profit is defined as the difference between total revenues and these total economic costs, implicit opportunity costs as well as explicit outlays:

Economic profit = Total revenues − Explicit costs − Implicit costs [8.1]

When one recognizes that such first-best and second-best uses change over time, it becomes clear that the historical outlay of funds to obtain a resource at an earlier date (the accounting cost basis) may not be the appropriate measure of opportunity cost in a decision problem today. For example, consider the following three cases of a substantive distinction between economic cost and accounting cost.

Three Contrasts between Accounting and Economic Costs

Depreciation Cost Measurement The production of a good or service typically requires the use of licenses and plant and equipment. As these capital assets are used, their service life is expended, and the assets wear out or become obsolete. Depreciation is a loss of asset value. If the Phillips Tool Company owns a machine that has a current market value of $8,000 and is expected to have a value of $6,800 after one more year of use, then the opportunity cost of using the machine for one year (the economist’s mea- sure of depreciation cost) is $8,000 – $6,800 = $1,200. Assuming that 2,000 units of output were produced during the year, the depreciation cost would be $1,200 ÷ 2,000 units = $.60 per unit.

Unfortunately, it is often difficult, if not impossible, to determine the exact service life of a capital asset and the future changes in its market value.2 Some assets are unique (patents); others are not traded in liquid resale markets (plants); and still others are rendered obsolete with little predictability (computers). To overcome these measurement problems with economic depreciation cost, accountants have adopted certain procedures for allocating a portion of the acquisition cost of an asset to each

opportunity costs The value of a resource in its next best alternative use. Opportunity cost represents the return or compensation that must be forgone as the result of the decision to employ the resource in a given economic activity.

capital assets A durable input that depreciates with use, time, and obsolescence.

2This concept of the future cost of the partially consumed asset is termed the replacement cost of the asset rather than the historical acquisition cost of the asset.

Chapter 8: Cost Analysis 277

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Opportunity Costs at Bentley Clothing Store Robert Bentley owns and operates the Bentley Clothing Store. A traditional income statement for the business is shown in Panel (a) of Table 8.1. The mortgage on the store has been paid, and therefore no interest expenses are shown on the income statement. Also, the building has been fully depreciated and thus no depreciation charges are shown. From an accounting standpoint and from the perspective of the Internal Revenue Service, Bentley is earning a positive accounting profit of $190,000 (before taxes).

However, consider the store’s profitability from an economic standpoint. As in- dicated earlier in the chapter, implicit costs include the opportunity costs of time and capital that the entrepreneur has invested in the firm. Suppose that Bentley

(Continued)

TABLE 8.1 PROFITABILITY OF BENTLEY CLOTHING STORE

(a) Accounting Income Statement

Net sales $650,000

Less: Cost of goods sold 250,000

Gross profit 400,000

Less: Expenses

Employee compensation* 150,000

Advertising 30,000

Utilities and maintenance 20,000

Miscellaneous 10,000

Total 210,000

Net profit before taxes $190,000

(b) Economic Profit Statement

Total revenues $650,000

Less: Explicit costs

Cost of goods sold 250,000

Employee compensation* 150,000

Advertising 30,000

Utilities and maintenance 20,000

Miscellaneous 10,000

Total 460,000

Accounting profit before taxes 190,000

Less: Implicit costs

Salary (manager) 130,000

Rent on building 88,000

Total 218,000

Economic profit (or loss) before taxes ($28,000)

*Employee compensation does not include any salary to Robert Bentley.

278 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

accounting time period, and in turn to each unit of output that is produced within that time period. This allocation is typically done by one of several arbitrary methods of assigning a portion of the historical cost to each year of the service life. If the ma- chine is purchased by Phillips for $10,000 and is expected to have a 10-year life and no salvage value, the straight-line method of depreciation ($10,000 ÷ 10 = $1,000) would calculate the depreciation cost of this asset each year. Assuming that 2,000 units of output are produced in a given year, then $1,000 ÷ 2,000 = $0.50 would be allocated to the cost of each unit produced by Phillips. Note from this example that the calculated accounting depreciation cost does not equal the economic depreciation cost actually incurred, if in fact the market value of the machine drops to $6,800 after one year.

Inventory Valuation Whenever materials are stored in inventory for a period of time before being used in the production process, the accounting and economic costs may differ if the market price of these materials has changed from the original purchase price. The accounting cost is equal to the actual acquisition cost, whereas the economic cost is equal to the current replacement cost. As the following example illustrates, the use of the acquisition cost can lead to incorrect production decisions.

could go to work as a clothing department manager for a large department or spe- cialty store chain and receive a salary of $130,000 per year. Also assume that Bent- ley could rent his building to another merchant for $88,000 (net) per year. Under these conditions, as shown in Panel (b) of Table 8.1, Bentley is earning a negative economic profit (–$28,000 before taxes). By renting his store to another merchant and going to work as manager of a different store, he could make $28,000 more than he is currently earning from his clothing store business. Thus, accounting profits, which do not include opportunity costs, are not always a valid indication of the economic profitability (or loss) of an enterprise.

Example Inventory Valuation at Westside Plumbing and Heating Westside Plumbing and Heating Company is offered a contract for $100,000 to provide the plumbing for a new building. The labor and equipment costs are cal- culated to be $60,000 for fulfilling the contract. Westside has the materials in in- ventory to complete the job. The materials originally cost the firm $50,000; however, prices have since declined and the materials could now be purchased for $37,500. Material prices are not expected to increase in the near future and hence no gains can be anticipated from holding the materials in inventory. The question is: Should Westside accept the contract? An analysis of the contract under both methods for measuring the cost of the materials is shown in Table 8.2. Assuming that the materials are valued at the acquisition cost, the firm should not accept the contract because an apparent loss of $10,000 would result. By using the replace- ment cost as the value of the materials, however, the contract should be accepted, because a profit of $2,500 would result.

(Continued)

Chapter 8: Cost Analysis 279

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Sunk Cost of Underutilized Facilities The Dunbar Manufacturing Company recently discontinued a product line and was left with 50,000 square feet of unneeded warehouse space. The company rents the entire warehouse (200,000 square feet) from the owner for $1 million per year (i.e., $5 per square foot) under a long-term (10-year) lease agreement. A nearby company that is expanding its operations offered to rent the 50,000 square feet of unneeded space for 1 year for $125,000 (i.e., $2.50 per square foot). Should Dunbar accept the offer to rent the unused space, assuming that no other higher offers for the warehouse space are expected?

One could argue that Dunbar should reject the offer because the additional rent (rev- enue) of $2.50 per square foot is less than the lease payment (cost) of $5 per square foot. Such reasoning, however, will lead to an incorrect decision. The lease payment ($5 per square foot) represents a sunk cost that must be paid regardless of whether the other company rents the unneeded warehouse space. As shown in Table 8.3, renting the un- needed warehouse space reduces the net cost of the warehouse from $1 million to $875,000, a savings of $125,000 per year to Dunbar. The relevant comparison is between

To see which method is correct, examine the income statement of Westside at the end of the accounting period. If the contract is not accepted, then at the end of the accounting period the firm will have to reduce the cost of its inventory by $12,500 ($50,000 – $37,500) to reflect the lower market value of this unused inven- tory. The firm will thus incur a loss of $12,500. If the contract is accepted, then the company will make a profit of $2,500 on the contract, but will also incur a loss of $12,500 on the materials used in completing the contract. The firm will thus incur a net loss of only $10,000. Hence, acceptance of the contract results in a smaller overall loss to Westside than does rejection of the contract. For decision-making purposes, replacement cost is the appropriate measure of the cost of materials in inventory, and Westside should accept the contract.

TABLE 8.2 EFFECT OF INVENTORY VALUATION METHODS ON

MEASURED PROFIT—WESTSIDE PLUMBING

AND HEATING COMPANY

ACQUISITION COST REPLACEMENT COST

Value of contract $100,000 $100,000

Costs

Labor, equipment 60,000 60,000

Materials 50,000 37,500

110,000 97,500

Profit (or loss) ($10,000) $ 2,500

TABLE 8.3 WAREHOUSE RENTAL DECISION—DUNBAR MANUFACTURING

COMPANY

DECISION

DO NOT RENT RENT

Total lease payment $1,000,000 $1,000,000

Less: Rent received on unused space — 125,000

Net cost of warehouse to Dunbar Manufacturing Company $1,000,000 $ 875,000

sunk cost A cost incurred regardless of the alternative action chosen in a decision- making problem.

280 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the incremental revenue ($125,000) and the incremental costs ($0 in this case). Thus, sunk costs (such as the lease payment of $5 per square foot in this example) should not be considered relevant costs because such costs are unavoidable, independent of the course of action chosen.

Conclusions

1. Costs can be measured in different ways, depending on the purpose for which the cost figures are to be used.

2. The costs appropriate for financial reporting purposes are not always appropriate for decision-making purposes. Typically, changes and modifications have to be made to reflect the opportunity costs of the various alternative actions that can be chosen in a given decision problem. The relevant cost in economic decision making is the op- portunity cost of the resources rather than the historical outlay of funds required to obtain the resources.

3. Sunk costs, which are incurred regardless of the alternative action chosen, should seldom be considered in making operating decisions.

SHORT-RUN COST FUNCTIONS In addition to measuring the costs of producing a given quantity of output, economists are also concerned with determining the behavior of costs when output is varied over a range of possible values. The relationship between cost and output is expressed in terms of a cost function: a schedule, graph, or mathematical relationship showing the mini- mum achievable cost of producing various quantities of output.

The discussion in Chapter 7 concerning the inputs used in the production process distin- guished between fixed and variable inputs. A fixed input was defined as an input that is re- quired in the production process, but whose quantity used in the process is constant over a given period of time regardless of the level of output produced. Short-run questions relate to a situation in which one or more of the inputs to the production process are fixed. Long-run questions relate to a situation in which all inputs are variable; that is, no restrictions are im- posed on the amount of a resource that can be employed in the production process. The length of time required to vary all the inputs can be as long as a decade (e.g., in shipbuilding). In other cases, the long-runmay be just a few weeks (in the 7–Eleven convenience store business).

The total cost of producing a given quantity of output is equal to the sum of the costs of each of the inputs used in the production process. In discussing short-run cost functions, it is useful to classify costs as either fixed or variable costs. Fixed costs represent the costs of all the inputs to the production process that are fixed or constant over the short run. Vari- able costs consist of the costs of all the variable inputs to the production process. Whereas variable costs may not change in direct proportion to the quantity of output produced, they will increase (or decrease) in some manner as output is increased (or decreased).

Average and Marginal Cost Functions Once the total cost function is determined, one can then derive the average and marginal cost functions. The average fixed cost AFC, average variable cost AVC, and average total cost ATC are equal to the respective fixed, variable, and total costs divided by the quan- tity of output produced:

AFC = FC Q

[8.2]

AVC = VC Q

[8.3]

cost function A mathematical model, schedule, or graph that shows the cost (such as total, average, or marginal cost) of producing various quantities of output.

fixed costs The costs of inputs to the production process that are constant over the short run.

variable costs The costs of the variable inputs to the production process.

Chapter 8: Cost Analysis 281

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

ATC = TC Q

[8.5]

Also,

ATC = AFC + AVC [8.6]

Example Short-Run Cost Functions: Deep Creek Mining Company To illustrate the nature of short-run costs and show how the short-run cost function can be derived from the production function for the firm, consider again the Deep Creek Mining Company example that was discussed in Chapter 7. It was assumed that two inputs, capital and labor, are required to produce or mine ore. Various-sized pieces of capital equipment, as measured by their brake horsepower rating K, are available to mine the ore. Each of these pieces of equipment can be operated with various-sized labor crews L. The amount of output (tons of ore) that can be produced in a given period with each capital-labor input combi- nation is shown again in Table 8.4. It was also assumed that the rental cost of using the mining equipment per period is $0.20 per brake horsepower and that the cost of each worker (labor) employed per period is $50. This yielded the following total cost equation for any given combination of labor L and capital K (Equation 7.21):

C = 50L + 0:20k

Suppose that Deep Creek has signed a lease agreeing to rent, for the next year, a 750-brake-horsepower piece of mining equipment (capital). During the ensuing year (the short run), the amount of capital that the company can employ in the ore-mining process is fixed at 750-brake horsepower. Therefore, for each period a fixed cost of $0.20 × 750 = $150 will be incurred, regardless of the quantity of ore that is produced. The firm must operate the production process at one of the capital-labor combinations shown in the third column of Table 8.4. Output can be increased (decreased) by employing more (less) labor in combination with the given 750-brake-horsepower capital equipment. Labor is thus a variable input to the production process.

The short-run cost functions for Deep Creek are shown in Table 8.5. The var- ious possible output levels Q and the associated capital-labor input combinations L and K are obtained from Table 8.4. The short-run variable cost VC is equal to $50 times the number of workers (L) employed in the mining process. The short-run fixed cost FC is equal to the rental cost of the 750-brake-horsepower equipment ($150). The total cost in the short run is the sum of the fixed and variable costs:

TC = FC + VC [8.4]

In Figure 8.1 the three curves from the data given in Table 8.5 are plotted. Note that the TC curve has an identical shape to that of VC, being shifted upward by the FC of $150.

282 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Marginal cost is defined as the incremental increase in total cost that results from a one-unit increase in output, and is calculated as3

TABLE 8.4 PRODUCTION FUNCTION—DEEP CREEK MINING COMPANY

CAPITAL INPUT K (BRAKE HORSEPOWER)

250 500 750 1,000 1,250 1,500 1,750 2,000

LABOR INPUT L (NUMBER OF WORKERS)

1 1 3 6 10 16 16 16 13

2 2 6 16 24 29 29 44 44

3 4 16 29 44 55 55 55 50

4 6 29 44 55 58 60 60 55

5 16 43 55 60 61 62 62 60

6 29 55 60 62 63 63 63 62

7 44 58 62 63 64 64 64 64

8 50 60 62 63 64 65 65 65

9 55 59 61 63 64 65 66 66

10 52 56 59 62 64 65 66 67

FIGURE 8.1 Short-Run Variable, Fixed, and Total Cost Functions—Deep Creek Mining Company

TC

0 10

Output Q (tons)

50

C os

t ($

)

20 30 40 50 60 70

100

150

200

250

300

350

400

450

500

VC

FC

marginal cost The incremental increase in total cost that results from a one-unit increase in output. 3Technically, the ratio ΔTC/ΔQ represents the incremental cost associated with a discrete change in output by

more than one unit rather than the marginal cost associated with one additional unit of output.

Chapter 8: Cost Analysis 283

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

MC = ΔTC ΔQ

= ΔVC ΔQ

[8.7]

or, in the case of a continuous TC function, as

MC = dðTCÞ dQ

[8.8]

= dðVCÞ dQ

[8.9]

The average and marginal costs for Deep Creek that were calculated in Table 8.5 are plotted in the graph shown in Figure 8.2. Except for the AFC curve, which is continually declining, note that all other average and marginal cost curves are U-shaped.

The Deep Creek example illustrated the derivation of the various cost functions when the cost data are given in the form of a schedule (tabular data). Consider another exam- ple where the cost information is represented in the form of an algebraic function. Sup- pose fixed costs for the Manchester Company are equal to $100, and the company’s variable costs are given by the following relationship (where Q = output):

VC = 60Q − 3Q2 + 0:10Q3 [8.10]

Given this information, one can derive the total cost function using Equation 8.2:

TC = 100 + 60Q − 3Q2 + 0:10Q3

Next, AFC, AVC, and ATC can be found using Equations 8.3, 8.4, and 8.5, respectively, as follows:

TABLE 8.5 SHORT-RUN COST FUNCTIONS—DEEP CREEK MINING COMPANY

OUTPUT VARIABLE

COST FIXED COST TOTAL COST

AVERAGE FIXED COST

AVERAGE VARIABLE

COST

AVERAGE TOTAL COST

MARGINAL COST

Q

LABOR INPUT

L VC = $50 · L

CAPITAL INPUT

K FC = $150 TC =

FC + VC AFC = FC

Q AVC = VC

Q ATC = TC

Q MC = TC

Q

0 0 $0 750 $150 $150 — — — —

6 1 50 750 150 200 $25.00 $8.33 $33.33 50 6

= $8.33

16 2 100 750 150 250 9.38 6.25 15.63 50 10

= 5.00

29 3 150 750 150 300 5.17 5.17 10.34 50 13

= 3.85

44 4 200 750 150 350 3.41 4.55 7.95 50 15

= 3.33

55 5 250 750 150 400 2.73 4.55 7.27 50 11

= 4.55

60 6 300 750 150 450 2.50 5.00 7.50 50 5

= 10.00

62 7 350 750 150 500 2.42 5.65 8.06 50 2

= 25.00

284 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

AFC = 100 Q

AVC = 60 − 3Q + 0:10Q2

ATC = 100 Q

+ 60 − 3Q + 0:10Q2

Finally, Manchester’s marginal cost function can be obtained by differentiating the vari- able cost function (Equation 8.10) with respect to Q:

MC = dðVCÞ dQ

= 60 − 6Q + 0:30Q2

The average total cost curve in Figure 8.2, which is equal to the sum of the vertical heights of the average fixed and average variable cost curves, initially declines and subse- quently begins rising beyond a particular level of output. At Q = 55 the average total cost curve is at its minimum value. As discussed in the previous chapter, specialization in the use of the variable inputs initially results in increasing returns and declining marginal costs and average variable costs. Eventually, however, the gains from specialization are overwhelmed by crowding effects, diminishing marginal returns set in, and then mar- ginal and average variable costs begin increasing as shown in Figure 8.3. This reasoning is used to explain the U-shaped pattern of the short-run ATC, AVC, and MC curves in Figures 8.2 and 8.3 and indeed in all short-run average cost structures.

FIGURE 8.2 Short-Run Average and Marginal Cost Functions—Deep Creek Mining Company

C os

t pe

r un

it (

$/ to

n)

4

8

12

16

20

24

28

32

36

40

Output Q (tons)

0 10 20 30 40 50 60 70

ATC

AVC

AFC

MC

Chapter 8: Cost Analysis 285

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

LONG-RUN COST FUNCTIONS Over the long-run planning horizon, using the available production methods and tech- nology, the firm can choose the plant size, types and sizes of equipment, labor skills, and raw materials that, when combined, yield the lowest cost of producing the desired amount of output. Once the optimum combination of inputs is chosen to produce the desired level of output at least cost, some of these inputs (plant and equipment) become fixed in the short run. If demand increases unexpectedly and the firm wishes to produce not Q1, as planned, but rather Q2 as shown in Figure 8.3, it may have little choice but to lay on additional variable inputs such as overtime labor and expedite the rush-order de- livery of supplies to meet its production goals. Not surprisingly, such arrangements are expensive, and short-run average cost will temporarily rise to B at C 02.

Should this demand persist, a larger fixed input investment in plant and equipment is warranted. Then, unit cost can be reduced from C 02 to C2. Associated with the larger fixed input investment is another short-run average cost function SAC2. Several of these other short-run average cost functions (SRAC3, SRAC4) are shown in Figure 8.3. The long-run average cost function consists of the lower boundary or envelope of all these short-run curves. No other combination of inputs exists for producing each level of out- put Q at an average cost below the cost that is indicated by the LRAC curve.

Optimal Capacity Utilization: Three Concepts To assess capacity utilization, assume the firm has been producing Q1 units of output using a plant of size “1,” having a short-run average cost curve of SAC1. The average cost of producing Q1 units is therefore C1, and Q1 is the optimal output for the plant size represented by SAC1. Optimal output for a given plant size is a short-run concept of capacity utilization.

Suppose that the firm now wishes to expand output to Q2. What will the average cost be of producing this higher volume of output? In the short run, as we saw earlier, the average cost would be C 02. However, in the long run, it would be possible for the firm to build a plant of size “2,” having a short-run average cost curve of SAC2. With this larger plant, the average cost of producing Q2 units of output would be only C2. Thus,

FIGURE 8.3 Long-Run and Short-Run Average Cost Functions

Output Q (units)

Q1

C2

C�2

C1

C3

Q2 Q3 Q4

A ve

ra ge

c os

t C

( $/

un it

)

SRAC1

SRAC2 SRAC3

SRAC4

LRAC

A B

optimal output for a given plant size Output rate that results in lowest average total cost for a given plant size.

286 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

because the firm has more options available to it in the long run, average total cost of any given output generally can be reduced. SAC2 represents the optimal plant size for a given output rate Q2. Should demand then collapse, even a company like Toyota may be stuck with excess manufacturing capacity and find it wishes to cut output back to the original level Q1, despite the much higher unit costs at point A. This is exactly what hap- pened in 2009 when U.S. Toyota sales fell by one-third on average, but some models like the Toyota Tundra plunged by 61 percent.

However, as the business cycle recovers, if the firm can execute a marketing plan to sell still more output, a still more efficient allocation of resources is available. Only when optimal output increases to Q3, where the firm will build the universally least- cost optimal plant size represented by SAC3, will further opportunities for cost reduc- tion cease. This concept of optimal capacity utilization applies to the long run, given the technology in place at this plant. Short-run average total cost with underutilization of capacity at Point A or overutilization of capacity at Point B in Figure 8.3 is always higher than the minimum average total cost in the long run (LRAC) fundamentally be- cause the production manager can vary plant and equipment in the long run, matching capacity to his or her output requirements.

ECONOMIES AND DISECONOMIES OF SCALE The long-run average total cost LRAC function is hypothesized to decline as the flow rate of throughput rises over the lower range of operations scale and is hypothesized to remain flat or rise over the higher range of scale. Declining long-run average total cost reflects internal economies of scale at one of three levels: the product level, the multi- product plant level, or the firm level of operations.

Product-Level Internal Economies of Scale A number of different sources of declining cost are associated with producing one product (say, PCs) at a higher rate of throughput per day. Special-purpose equipment, which is more efficient in performing a limited set of operations, can be substituted for less efficient general-purpose equipment.

Example The Average Cost per Kilowatt Hour in Underutilized Power Plants4

Under pressure from regulators, the electric power industry opened its customer distribution systems to freewheeling electricity. A factory in Ohio can choose to buy contract electricity from Michigan, New York, or Virginia power companies. With the new competition, the price of electricity is certain to decline, and con- sumption will increase. However, excess capacity is present in much of the power industry today, and higher-cost electric utilities will soon find themselves priced out of the market. As more efficient power plants are constructed, some estimates show the price of electricity falling by 1.8 cents using conventional coal-fired steam turbine technology and by as much as 3.0 cents using nuclear and other technolo- gies. These savings imply an $18 to $30 reduction per month in the residential electricity bill for a customer who switches to the lower-cost firms.

4Based on M. Maloney, R. McCormick, and R. Sauer, Customer Choice, Consumer Value: An Analysis of Retail Com- petition in America’s Electric Industry (Washington, DC: Citizens for a Sound Economy, 1996).

optimal plant size for a given output rate Plant size that results in lowest average total cost for a given output.

optimal plant size Plant size that achieves minimum long-run average total cost.

internal economies of scale Declining long- run average costs as the rate of output for a product, plant, or firm is increased.

Chapter 8: Cost Analysis 287

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Likewise, the production process can be broken down into a series of smaller tasks, and workers can be assigned to the tasks for which they are most qualified. Workers are then able to acquire additional proficiency through higher repetition of the tasks to which they are assigned.

In manufacturing, a related phenomenon called the learning curve effect has often been observed whereby the amount of labor input required to produce another unit of output decreases as the cumulative volume of output rises (e.g., during long production runs of 767 airframes at Boeing). The learning curve principle was first applied in air- frame manufacturing, shipbuilding, and appliance manufacturing. Learning curve effects and volume discounts in purchasing inputs (so-called external economies of scale) are easily distinguished from internal economies of scale because they depend upon cumula- tive volume of output no matter how small the production throughput rate per time pe- riod. As such, they may well be associated with no change in the scale of operations whatsoever; the firm simply buys large volumes of input at one time.

The learning curve relationship is usually expressed as a constant percentage by which the amount of an input (or cost) per unit of output is reduced each time production is doubled. For example, consider a production process in which labor input and costs fol- low an 80 percent learning curve. Assume that the first unit requires labor costs of $1,000 to produce. Based on the learning curve relationship, the second unit costs $1,000 × 0.80 = $800, the fourth unit costs $800 × 0.80 = $640, the eighth unit costs $640 × 0.80 = $512, the sixteenth unit costs $512 × 0.80 = $409.60, and so on.

This learning curve relationship plotted in Figure 8.4 can be expressed algebraically as follows:

C = aQb [8.11]

where C is the input cost of the Qth unit of output, Q is consecutive units of output produced, a is the theoretical (or actual) input cost of the first unit of output, and b is the rate of reduction in input cost per unit of output. Because the learning curve is downward sloping, the value of b is normally negative. Taking logarithms of both sides of Equation 8.11 yields

log C = log a + b log Q [8.12]

Example Mass Customization and the Learning Curve5

Mass customization is a new trend in operations management designed to stan- dardize at least some of the production processes associated with fulfilling custom orders. Lee Jeans’ customers can choose their own back pocket stitching and the number of prior stone washings at a mall kiosk, but then Lee actually assembles the custom order from stockpiles of subassemblies produced with long production runs. This learning curve effect of cumulative volume in reducing unit cost arises from increased familiarization with the tasks by workers and supervisors, improve- ments in work methods and the flow of work, and the need for fewer skilled work- ers as the tasks become more repetitive. Raw material costs per unit may also be subject to the learning curve effect if less scrap and waste occur as workers become more familiar with the production process.

5An excellent survey on mass customization is found in M. Agrawal, T. V. Kumaresh, and G. A. Mercer, “The False Promise of Mass Customization,” McKinsey Quarterly (November 3, 2001). See also “A Long March,” The Economist, (July 14, 2001), pp. 63–65.

learning curve effect Declining unit cost runs attributable to greater cumulative volume.

volume discount Reduced variable cost attributable to larger purchase orders.

288 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Regression analysis can then be used to estimate the parameters b and log a in order to forecast costs at various cumulative volumes.

The Percentage of Learning The percentage of learning, which is defined as the proportion by which an input (or its associated cost) is reduced when output is doubled, can be estimated as follows:

L = C2 C1

× 100% [8.13]

where C1 is the input (or cost) for the Q1 unit of output and C2 is the cost for the Q2 = 2Q1 unit of output.

FIGURE 8.4 Learning Curve: Arithmetic Scale

0

Q (unit number)

200

C os

t/ un

it (

$/ un

it )

400

600

800

1,000

4 8 12 16 20 24

300

500

700

900

Cumulative volume

Example Percentage of Learning: Emerson Corporation Emerson Corporation makes landing gear for commercial aircraft. To illustrate the calculation of the percentage of learning, suppose Emerson’s labor costs for the Q1 = 50th unit of output are C1 = $659.98, and labor costs for the 2Q1 = 100th unit of output are C2 = $540.84. Substituting these values into Equation 8.13 yields

L = $540:84 $659:98

× 100%

= 81:9%

The percentage of learning for labor costs in the production of these landing gear units is thus approximately 82 percent—indicating that labor costs decline by about 18 percent each time output is doubled.

Chapter 8: Cost Analysis 289

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Plant-Level Internal Economies of Scale Sources of scale economies at the plant level include capital investment, overhead, and required reserves of maintenance parts and personnel. With respect to capital investment, capital costs tend to increase less than proportionately with the productive capacity of a plant, particularly in process-type industries. For example, consider oil pipeline operations. A pipeline with twice the radius of a smaller pipeline can be constructed for perhaps as little as twice the capital investment, yet have four times the capacity (i.e., π(2r)2 = 4πr2 versus πr2).

Another source of plant-level scale economies is overhead costs, which include such administrative costs as management salaries and paperwork documentation associated with regulatory compliance. Overhead costs can be spread over a higher volume of throughput in a larger plant or facility, thus reducing average costs per unit.

Firm-Level Internal Economies of Scale In addition to product-level and plant-level economies of scale, other scale economies are associated with the overall size

Example IBM and Intel Fabricate Monster Silicon Wafers6

The semiconductor industry produces thin wafers of silicon, and then in super-clean “wafer fab” factories etches them with electrical circuit lines 1/1000 the width of a human hair. Such intermediate products are then sliced and diced into the tiny memory chips in our PCs and PDAs. One 12-inch wafer yields enough memory chips to store 5,000 sets of the Encyclopaedia Britannica. Until recently, the standard wafer measured about 8 inches across and cost $5,500 to produce in a $1.4 billion wafer fab. New monster wafers are almost 12 inches across, cost $8,000 to produce in $2 billion wafer fabs, but their 125 percent greater surface area yields 575 chips per wafer as opposed to only 240 chips from the standard 8-inch wafer. That is, for a 45 percent increase in cost, the monster wafers yield 140 percent more chips than the standard wafer. As a result, the unit cost of a chip has fallen to $14 from $23.

6Based on “Chips on Monster Wafers,” BusinessWeek (November 4, 2002), pp. 112–126.

Example Refuse Collection and Disposal in Orange County Private for-profit trash collectors in California have demonstrated the scale econo- mies of landfills. The environmental safety issues at a landfill require enormous investment in environmental impact studies, lining the site, monitoring for seepage and leeching of toxins, and scientific follow-up studies. Spreading these overhead costs across a larger output volume has led an Orange County company to seek refuse as far away as the northern suburbs of San Diego, almost an hour down the California coast. The trucks from Orange County pass right by several munici- pal landfills en route. However, the fees charged for dumping at these intermediate sites are much higher. Apparently, the variable transportation costs of hauling a ton of trash prove to be less than the higher start-up costs and environmental monitoring costs per ton at smaller-scale landfills. By state law, all of these munic- ipalities must charge a dumping fee that covers their fully allocated cost, so re- duced long-run average total cost provides a substantial price advantage for the large-scale Orange County site.

290 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

of the multiplant firm. One possible source of firm-level scale economies is in distribu- tion. For example, multiplant operations may permit a larger firm to maintain geograph- ically dispersed plants. Delivery costs are often lower for a geographically dispersed operation compared with a single (larger) plant.

Another possible source of scale economies to the firm is in raising capital funds. Be- cause flotation costs increase less than proportionately with the size of the security (stock or bond) issue, average flotation costs per dollar of funds raised is smaller for larger firms. Similar scale economies also exist in marketing and sales promotion. These scale economies can take such forms as (1) quantity discounts in securing advertising media space and time, or (2) the ability of the large firm to spread the fixed costs of advertising each period over greater throughput. In addition, the large firm may be able to achieve a relatively greater degree of brand recognition and brand loyalty for any given level of sales promotion expenditures.

Diseconomies of Scale Rising long-run average costs at higher rates of throughput are attributed to diseconomies of scale. A primary source of diseconomies of scale associated with an individual produc- tion plant is transportation costs. Another possible source of plant diseconomies is labor requirements; higher wage rates or costly worker recruiting and relocation programs may be required to attract the necessary personnel. Finally, large-scale plants are often inflexible operations designed for long production runs of one product, based often on forecasts of what the target market wanted in the past.

Diseconomies of scale at the firm level result from problems of coordination and con- trol encountered by management as the scale of operations is increased. First, the size of management staffs and their associated salary costs may rise more than proportionately as the scale of the firm is increased. Also, less direct and observable costs may occur, such as the losses arising from delayed or faulty decisions and weakened or distorted managerial incentives. Contemporary examples of these problems include General Mo- tors and Motorola.

Example Economies of Scale: Superscale Money-Center versus Community Banks The number of large money-center banks—banks that are 1,000 times larger than the typical community bank in smaller towns and cities—is growing in the United States. Shaffer and David examined the economies of scale in these superscale banks, which range in size from $2.5 billion to $120.6 billion in assets.7 Their re- sults indicate that long-run average costs decline for those superscale banks be- tween $2.5 billion and $37 billion in assets. The superscale banks have come to dominate credit card issuance, corporate lending, and custodial asset management perhaps because of the massive information technology investments required in these businesses. In contrast, community banks, thrift institutions, and credit unions dominate retail banking, where diseconomies of scale (rising long-run aver- age costs) set in beyond $50 million in assets.8

7S. Shaffer and E. David, “Economies of Superscale in Commercial Banking,” Applied Economics 23 (1991), pp. 283–293. 8T. Gilligan, M. Smirlock, and W. Marshall, “Scale and Scope Economies in the Multi-Product Banking Firm,” Journal of Monetary Economics 13 (1984), pp. 393–405.

diseconomies of scale Rising long-run average costs as the level of output is increased.

Chapter 8: Cost Analysis 291

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Flexibility and Operating Efficiency: Ford Motor Company’s Flat Rock Plant9

Ford Motor Company spent an estimated $200 million in the early 1970s to con- struct a massive plant in Flat Rock, Michigan, to build cast iron engine blocks. The plant produced exclusively V8 blocks on five ultra-high-speed assembly lines at the rate of 8,000 engine blocks per day or 2 million per year. In the 1980s, however, Ford executives decided to close the Flat Rock plant and move production to an older Cleveland engine block plant. Ford’s Cleveland plant had 10 smaller and slower production lines; the Cleveland plant was clearly the less efficient of the two factories. However, Ford executives realized it would cost less to convert Cleve- land’s smaller production lines to the new high-performance six- and four-cylinder engines that had become popular.

When Flat Rock was designed, Ford could count on long V8 production runs of perhaps one million units of its most popular Ford Mustang model. But in the 1990s, variety in the product line became the key business strategy. By 1998, Ford’s top-selling Explorer (383,852 units), Taurus (357,162 units), and Escort (283,898 units) reflected the fragmented auto marketplace. Only the F-series pickup (746,111 units) warranted Flat Rock’s massive scale. As George Booth, iron- operations manager for the casting division at Ford, explained, “Flat Rock was built to make a few parts at very high volumes. But the plant turned out to be very in- flexible for conversion to making new types and different sizes of engine blocks. Sometimes you really can be too big.”

In 2007, the F-series pickup was still the most popular selling vehicle in the United States, but the need for huge V8-engine manufacturing scale capable of producing 2 million engines continued to decline as annual sales of the popular pickup fell off to 538,910.

9Based on articles in The Economist (January 11, 2001), p. 58; AI (February 1998); and Wall Street Journal (September 16, 1981, and February 12, 2007).

INTERNATIONAL PERSPECTIVES

How Japanese Companies Deal with the Problems of Size10

Many large, successful U.S. corporations, such as General Electric, Hewlett-Packard, Sara Lee, and Johnson & Johnson, are attempting to deal with the problems of excessive size by decentralizing their operations. These companies do so by setting up in- dependent business units, each with its own profit- and-loss responsibility, thereby giving managers more flexibility and freedom in decision making.

Like their counterparts in the United States, Japanese corporations are often collections of hun- dreds of individual companies. For example, Mat- sushita Electrical Industrial Company consists of

161 consolidated units. Another example is Hitachi, Ltd., which is composed of 660 companies, with the stock of 27 of these companies being publicly traded. James Abegglen, an expert on Japanese man- agement, has observed that “As something new comes along … it gets moved out to a subsidiary so the elephant does not roll over and smother it. If all goes well, it becomes a successful company on its own. If not, it gets pulled back in.”

10Based on an article entitled “Is Your Company Too Big?” BusinessWeek (March 17, 1989), pp. 84–94.

292 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Overall Effects of Scale Economies and Diseconomies For some industries, such as textile and furniture manufacturing, long-run average costs for the firm remain constant over a wide range of output once scale economies are exhausted. In such cases, many plant sizes are consistent with least-cost produc- tion, as shown in Figure 8.5. In other industries (e.g., engine block casting), long-run average costs rise at large scale. The possible presence of both economies and

FIGURE 8.5 Long-Run Average Cost Function and Scale Economies

Output (Q)

L on

g- ru

n av

er ag

e co

st

LRAC

Minimum efficient

scale

Maximum efficient

scale

Example Aluminum-Intensive Vehicle Lowers the Minimum Efficient Scale at Ford11

For decades, the largest fixed cost on the auto assembly line has been a $30 million body-stamping machine. This massive piece of capital equipment bends sheet metal into hoods, trunks, and fenders and hydraulically presses steel plate into floor pan and door pillar shapes. Because a body-stamping machine has a physical working life of 600,000 vehicles, it has been the source of substantial scale econo- mies in the production of most auto models. Only the top-selling Ford Focus (902,008), F-series pickup (869,001), VW Golf (795,835), Opel Astra (770,003), and Chevy pickup (644,084) had sufficient sales volume in 1999 to fully depreciate a body-stamping machine within the model year. Most moderately successful auto models sell fewer than 100,000 units per year. Therefore, a six-year period is re- quired to physically “wear out” a body-stamping machine as it makes repetitive presses for a typical model.

Should Ford change body shapes and structural components every two to three years to keep their model “current”? Or should they forgo the body style changes and fully depreciate their body-stamping machines over a six-year period or lon- ger? The former decision necessitates scrapping a machine with substantial physi- cal working life remaining and recovering the capital equipment investment with a much higher unit cost per vehicle.

(Continued)

Chapter 8: Cost Analysis 293

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

One common approach to achieving sufficient scale to wear out a stamping ma- chine has been to export the product beyond limited domestic markets and sell the same model under different names in different countries (e.g., Ford Focus/Fiesta and VW Golf/Bora/Vento). Another approach has been to consolidate companies across several continents (DaimlerChrysler-Mitsubishi, Ford-Volvo-Mazda, GM- Opel-Fiat-Isuzu-Suzuki, and Renault-Nissan) to get access for domestic models into foreign markets.

A third approach has been to avoid this classic economy of scale issue with alu- minum space-frame production or with a greater use of thermoplastics. The alumi- num space-frame automobile Ford is designing (or the A2 that Audi has already brought to market) is not even half as heavy as conventional steel-and- sheet-metal cars of today. In addition to phenomenal increases in gas mileage and markedly reduced CO2 emissions, aluminum-intensive vehicles will change the scale economies of auto assembly dramatically. Aluminum space-frame and ther- moplastic components are cast, forged, and extruded into different thicknesses de- pending upon where strength is needed. Neither requires a body-stamping machine.

Although an aluminum space-frame vehicle is 10 percent more costly on aver- age than the typical steel-and-sheet-metal vehicle, the minimum efficient scale of an aluminum-intensive auto assembly process is only 50,000 cars. As illustrated in Figure 8.6, a marketing plan for smaller volume niche products such as the Ford Fusion, Chevy Volt and Audi A2 can achieve minimum efficient scale at Point A with these new aluminum production techniques. Previously with steel- and-sheet-metal automobiles, production runs at this reduced scale would have re- sulted in unit costs at Point B, more than twice as high as those of a 300,000-unit vehicle such as the VW Passat; compare Point C in Figure 8.6. Cost discrepancies this great can seldom be overcome no matter how popular the design. But, by moving to thermoplastic and aluminum components, niche vehicles can remain cost competitive.

11Based on “Aluminum Cars,” The Economist (April 15, 2000), p. 89; Consumer Reports (April 1997), p. 26; “The Global Gambles of GM,” The Economist (June 24, 2000), p. 67; and “Daimler-Chrysler Merger,” Wall Street Journal (May 8, 1998), p. A10.

FIGURE 8.6 Minimum Efficient Scale in Autos

L on

g- ru

n av

er ag

e co

st (

$)

50,000 300,000 Annual output

Aluminum auto

Steel auto

A

B

C

294 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

diseconomies of scale leads to the hypothesized long-run average cost function for a typical manufacturing firm being U-shaped with a flat middle area. Up to some mini- mum efficient scale (MES), that is, the smallest scale at which minimum long-run av- erage total costs are attained, economies of scale are present. In most industries, it is possible to increase the size of the firm beyond this MES without incurring disecon- omies of scale. Over this extended middle-scale range, average costs per unit are rela- tively constant. However, expansion beyond the maximum efficient scale eventually will result in problems of inflexibility, lack of managerial coordination, and rising long-run average total costs.

SUMMARY

� Cost is defined as the sacrifice incurred whenever an exchange or transformation of resources takes place.

� Different approaches are used in measuring costs, depending on the purposes for which the informa- tion is to be used. For financial reporting pur- poses, the historical outlay of funds is usually the appropriate measure of cost, whereas for decision- making purposes, it is often appropriate to measure cost in terms of the opportunities forgone or sacrificed.

� A cost function is a schedule, graph, or mathemati- cal relationship showing the minimum achievable cost (such as total, average, or marginal cost) of producing various quantities of output.

� Short-run total costs are equal to the sum of fixed and variable costs.

� Marginal cost is defined as the incremental increase in total cost that results from a one-unit increase in output.

� The short-run average variable and marginal cost functions of economic theory are hypothesized to be U-shaped, first falling and then rising as output is increased. Falling short-run unit costs are attrib- uted to the gains available from specialization in the use of capital and labor. Rising short-run unit

costs are attributed to diminishing returns in production.

� The theoretical long-run average cost function is often found to be L-shaped due to the frequent presence of scale economies and frequent absence of scale diseconomies. Economies of scale are attrib- uted primarily to specialization and other features of the production process or the factor markets, whereas diseconomies of scale are attributed pri- marily to problems of coordination and inflexibil- ity in large-scale organizations.

� Volume discounts in purchasing inputs and learn- ing curve effects, both of which result from a larger cumulative volume of output, can be distinguished from scale effects, which depend on the firm’s rate of production throughput per time period. Learn- ing curve advantages often, therefore, arise in small-scale plants able to make long production runs.

� Minimum efficient scale is achieved by a rate of output sufficient to reduce long-run average total cost to the minimum possible level. Smaller rates of output imply smaller plant sizes to reduce unit cost, albeit to higher levels than would be possible if a firm’s business plan could support minimum efficient scale production.

Exercises 1. US Airways owns a piece of land near the Pittsburgh International Airport. The land originally cost US Airways $375,000. The airline is considering building a new training center on this land. US Airways determined that the proposal to build the new facility is acceptable if the original cost of the land is used in the

Answers to the exercises in blue can be found in Appendix D at the back

of the book.

minimum efficient scale (MES) The smallest scale at which minimum costs per unit are attained.

Chapter 8: Cost Analysis 295

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

analysis, but the proposal does not meet the airline’s project acceptance criteria if the land cost is above $850,000. A developer recently offered US Airways $2.5 million for the land. Should US Airways build the training facility at this location?

2. Howard Bowen is a large-scale cotton farmer. The land and machinery he owns has a current market value of $4 million. Bowen owes his local bank $3 million. Last year Bowen sold $5 million worth of cotton. His variable operating costs were $4.5 million; accounting depreciation was $40,000, although the actual de- cline in value of Bowen’s machinery was $60,000 last year. Bowen paid himself a salary of $50,000, which is not considered part of his variable operating costs. In- terest on his bank loan was $400,000. If Bowen worked for another farmer or a local manufacturer, his annual income would be about $30,000. Bowen can invest any funds that would be derived, if the farm were sold, to earn 10 percent annu- ally. (Ignore taxes.) a. Compute Bowen’s accounting profits. b. Compute Bowen’s economic profits.

3. Mary Graham worked as a real estate agent for Piedmont Properties for 15 years. Her annual income is approximately $100,000 per year. Mary is considering es- tablishing her own real estate agency. She expects to generate revenues during the first year of $2 million. Salaries paid to her employees are expected to total $1.5 million. Operating expenses (i.e., rent, supplies, utility services) are expected to total $250,000. To begin the business, Mary must borrow $500,000 from her bank at an interest rate of 15 percent. Equipment will cost Mary $50,000. At the end of one year, the value of this equipment will be $30,000, even though the depreciation expense for tax purposes is only $5,000 during the first year. a. Determine the (pre-tax) accounting profit for this venture. b. Determine the (pre-tax) economic profit for this venture. c. Which of the costs for this firm are explicit and which are implicit?

4. From your knowledge of the relationships among the various cost functions, com- plete the following table.

Q TC FC VC ATC AFC AVC MC

0 125 _____ _____ _____ _____ _____ _____

10 _____ _____ _____ _____ _____ _____ 5

20 _____ _____ _____ 10.50 _____ _____ _____

30 _____ _____ 110 _____ _____ _____ _____

40 255 _____ _____ _____ _____ _____ _____

50 _____ _____ _____ _____ _____ 3 _____

60 _____ _____ _____ _____ _____ _____ 3

70 _____ _____ _____ 5 _____ _____ _____

80 _____ _____ 295 _____ _____ _____ _____

5. A manufacturing plant has a potential production capacity of 1,000 units per month (capacity can be increased by 10 percent if subcontractors are employed). The plant is normally operated at about 80 percent of capacity. Operating the plant above this level significantly increases variable costs per unit because of the need to pay the skilled workers higher overtime wage rates. For output levels up to

296 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

80 percent of capacity, variable cost per unit is $100. Above 80 percent and up to 90 percent, variable costs on this additional output increase by 10 percent. When output is above 90 percent and up to 100 percent of capacity, the additional units cost an additional 25 percent over the unit variable costs for outputs up to 80 per- cent of capacity. For production above 100 percent and up to 110 percent of capac- ity, extensive subcontracting work is used and the unit variable costs of these additional units are 50 percent above those at output levels up to 80 percent of capacity. At 80 percent of capacity, the plant’s fixed costs per unit are $50. Total fixed costs are not expected to change within the production range under consid- eration. Based on the preceding information, complete the following table.

Q TC FC VC ATC AFC AVC MC

500 _____ _____ _____ _____ _____ _____ _____

600 _____ _____ _____ _____ _____ _____ _____

700 _____ _____ _____ _____ _____ _____ _____

800 _____ _____ _____ _____ _____ _____ _____

900 _____ _____ _____ _____ _____ _____ _____

1,000 _____ _____ _____ _____ _____ _____ _____

1,100 _____ _____ _____ _____ _____ _____ _____

6. The Blair Company’s three assembly plants are located in California, Georgia, and New Jersey. Previously, the company purchased a major subassembly, which be- comes part of the final product, from an outside firm. Blair has decided to manu- facture the subassemblies within the company and must now consider whether to rent one centrally located facility (e.g., in Missouri, where all the subassemblies would be manufactured) or to rent three separate facilities, each located near one of the assembly plants, where each facility would manufacture only the subassem- blies needed for the nearby assembly plant. A single, centrally located facility, with a production capacity of 18,000 units per year, would have fixed costs of $900,000 per year and a variable cost of $250 per unit. Three separate decentralized facili- ties, with production capacities of 8,000, 6,000, and 4,000 units per year, would have fixed costs of $475,000, $425,000, and $400,000, respectively, and variable costs per unit of only $225 per unit, owing primarily to the reduction in shipping costs. The current production rates at the three assembly plants are 6,000, 4,500, and 3,000 units, respectively. a. Assuming that the current production rates are maintained at the three as-

sembly plants, which alternative should management select? b. If demand for the final product were to increase to production capacity,

which alternative would be more attractive? c. What additional information would be useful before making a decision?

7. Kitchen Helper Company decides to produce and sell food blenders and is consid- ering three different types of production facilities (“plants”). Plant A is a labor- intensive facility, employing relatively little specialized capital equipment. Plant B is a semiautomated facility that would employ less labor than A but would also have higher capital equipment costs. Plant C is a completely automated facility using much more high-cost, high-technology capital equipment and even less la- bor than B. Information about the operating costs and production capacities of these three different types of plants is shown in the following table.

Chapter 8: Cost Analysis 297

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PLANT TYPE

A B C

Unit variable costs

Materials $3.50 $3.25 $3.00

Labor 4.50 3.25 2.00

Overhead 1.00 1.50 2.00

Total $9.00 $8.00 $7.00

Annual fixed costs

Depreciation $ 60,000 $100,000 $200,000

Capital 30,000 50,000 100,000

Overhead 60,000 100,000 150,000

Total $150,000 $250,000 $450,000

Annual capacity 75,000 150,000 350,000

a. Determine the average total cost schedules for each plant type for annual outputs of 25,000, 50,000, 75,000, …, 350,000. For output levels beyond the capacity of a given plant, assume that multiple plants of the same type are built. For example, to produce 200,000 units with Plant A, three of these plants would be built.

b. Based on the cost schedules calculated in part (a), construct the long-run average total cost schedule for the production of blenders.

8. The ARA Railroad owns a piece of land along one of its right-of-ways. The land originally cost ARA $100,000. ARA is considering building a new maintenance facility on this land. ARA determined that the proposal to build the new facility is acceptable if the original cost of the land is used in the analysis, but the pro- posal does not meet the railroad’s project acceptance criteria if the land cost is above $500,000. An investor has recently offered ARA $1 million for the land. Should ARA build the maintenance facility at this location?

9. The Emerson Corporation, a manufacturer of airplane landing gear equipment, is trying to develop a learning curve model to help forecast labor costs for successive units of one of its products. From past data, the firm knows that labor costs of the 25th, 75th, and 125th units were $800, $600, and $500, respectively. Using the learning curve equation for these labor costs, log C = 3.30755 − 0.28724 log Q, calculate the estimated cost of the 200th unit of output. What is the percentage of learning at Emerson?

Case Exercise COST ANALYSIS

The Leisure Products (LP) Company manufactures lawn and patio furniture. Most of its output is sold to do-it-yourself warehouse stores (e.g., Lowe’s Home Improvement) and to retail hardware and department store chains (e.g., True Value and JCPenney), who then distribute the products under their respective brand names. LP is not in- volved in direct retail sales. Last year the firm had sales of $35 million.

One of LP’s divisions manufactures folding (aluminum and vinyl) chairs. Sales of the chairs are highly seasonal, with 80 percent of the sales volume concentrated in

298 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the January–June period. Production is normally concentrated in the September–May period. Approximately 75 percent of the hourly workforce (unskilled and semiskilled workers) is laid off (or takes paid vacation time) during the June–August period of reduced output. The remainder of the workforce, consisting of salaried plant manage- ment (line managers and supervisors), maintenance, and clerical staff, are retained during this slow period. Maintenance personnel, for example, perform major over- hauls of the machinery during the slow summer period.

LP planned to produce and sell 500,000 of these chairs during the coming year at a projected selling price of $7.15 per chair. The cost per unit was estimated as follows:

Direct labor $2.25

Materials 2.30

Plant overhead* 1.15

Administrative and selling expense 0.80

TOTAL $6.50

*These costs are allocated to each unit of output based on the projected annual production of 500,000 chairs.

A 10 percent markup ($0.65) was added to the cost per unit in arriving at the firm’s selling price of $7.15 (plus shipping).

In May, LP received an inquiry from Southeast Department Stores concerning the possible purchase of folding chairs for delivery in August. Southeast indicated that they would place an order for 30,000 chairs if the price did not exceed $5.50 each (plus shipping). The chairs could be produced during the slow period using the firm’s existing equipment and workforce. No overtime wages would have to be paid to the workforce in fulfilling the order. Adequate materials were on hand (or could be purchased at prevailing market prices) to complete the order.

LP management was considering whether to accept the order. The firm’s chief accountant felt that the firm should not accept the order because the price per chair was less than the total cost and contributed nothing to the firm’s profits. The firm’s chief economist argued that the firm should accept the order if the incremental reve- nue would exceed the incremental cost.

The following cost accounting definitions may be helpful in making this decision:

• Direct labor: Labor costs incurred in converting the raw material into the finished product.

• Material: Raw materials that enter into and become part of the final product. • Plant overhead: All costs other than direct labor and materials that are associated

with the product, including wages and salaries paid to employees who do not work directly on the product but whose services are related to the production process (line managers, maintenance, and janitorial personnel); heat; light; power; supplies; depreciation; taxes; and insurance on the assets employed in the production process.

• Selling and distribution costs: Costs incurred in making sales (e.g., billing and salespeople’s compensation), storing the product, and shipping the product to the customer. (In this case the customer pays all shipping costs.)

• Administrative costs: Items not listed in the preceding categories, including gen- eral and executive office costs, research, development, engineering costs, and mis- cellaneous items.

Chapter 8: Cost Analysis 299

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Questions 1. Calculate the incremental, or marginal, cost per chair to LP of accepting the order

from Southeast. 2. What assumptions did you make in calculating the incremental cost in Question

1? What additional information would be helpful in making these calculations? 3. Based on your answers to Questions 1 and 2, should LP accept the Southeast

order? 4. What additional considerations might lead LP to reject the order?

300 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

8A APPENDIX

Long-Run Costs with a Cobb-Douglas Production Function, Advanced Material

In Chapter 8, the shape of the firm’s long-run unit cost structure (the LRAC) was shown to depend on whether economies or diseconomies of scale were present. A priori hypoth- eses about the shape of the firm’s LRAC can be derived, before examining cost data, by postulating a production function for the firm. Consider the Cobb-Douglas production function

Q = αLβ1kβ2 [8A.1]

where L is the amount of labor, K is the amount of capital used in producing Q units of output, and α, β1, and β2 are constants. The total cost of employing L units of labor and K units of capital in a production process is equal to

C = CLL + CKK [8A.2]

where CL and CK are the per-unit prices of labor and capital, respectively. Using Lagrangian multiplier techniques, one can determine the total cost (C) of producing any level of output.

The objective is to minimize the total cost (C) of producing a given level of output Q = Q0. We begin by forming the Lagrangian function

LC = C + λðQ − Q0Þ [8A.3]

= CLL + CKK + λðαLβ1Kβ2 − Q0Þ [8A.4] Differentiating LC with respect to L, K, and λ and setting these derivatives equal to zero yields:

∂LC ∂L

= CL + λðβ1αLβ1−1Kβ2Þ = 0 [8A.5]

∂LC ∂K

= CK + λðβ2αLβ1Kβ2−1Þ = 0 [8A.6]

∂LC ∂λ

= αLβ1Kβ2 − Q0 = 0 [8A.7]

301

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Solving these equations yields the following cost-minimizing values of L and K:

L* = Q0 α

� �1=ðβ1+β2Þ β1CK β2CL

� �β2=ðβ1+β2Þ [8A.8]

K* = Q0 α

� �1=ðβ1+β2Þ β1CL β2CK

� �β1=ðβ1+β2Þ [8A.9]

Substituting Equation 8A.8 for L and Equation 8A.9 for K in Equation 8A.2 and doing some algebraic operations gives the total cost (C) of producing any level of output (Q):

C = Cβ1=ðβ1+β2ÞL C β2=ðβ1+β2Þ K

β2 β1

� �β1=ðβ1+β2Þ +

β2 β1

� �−β2=ðβ1+β2Þ" # Q α

� �1=ðβ1+β2Þ

This cost equation indicates that total costs are a function of the output level (Q), the per-unit costs of labor (CL) and capital (CK), and the parameters (α, β1, and β2) of the Cobb-Douglas production function.

Several examples can be used to illustrate the implied shapes of the firm’s LRAC. In the following examples, assume that α = 4.0 and that the per-unit costs of labor (CL) and capital (CK) are $2 and $8, respectively.

Constant Returns β1 = 0.50, β2 = 0.50 (Because β1 + β2 = 1.0, this equation is an example of constant re- turns to scale.)

LRTC = ð2Þ0:50ð8Þ0:50½1 + 1� Q 4:0

� �1 = 2:0Q

LRAC = LTC Q

= 2:0Q Q

= 2:0

These LRTC and LRAC functions are graphed in Panel (a) of Figure 8A.1. Note that when the Cobb-Douglas production function exhibits constant returns to scale, total costs in- crease linearly with output, and average total costs are constant, independent of output.

Decreasing Returns β1 = 0.25, β2 = 0.25 (Because β1 + β2 < 1.0, this equation is an example of decreasing returns to scale.)

LRTC = ð2Þ0:50ð8Þ0:50½1 + 1� Q 4:0

� �2 = 0:50Q2

LRAC = 0:50Q

These cost functions are graphed in Panel (b) of Figure 8A.1. Note that when the Cobb- Douglas production function exhibits decreasing returns to scale, total costs increase more than proportionately with output, and average total costs rise as output increases.

302 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Increasing Returns β1 = 1.0, β2 = 1.0 (Because β1 + β2 > 1.0, this equation is an example of increasing returns to scale.)

LRTC = ð2Þ0:50ð8Þ0:50½1 + 1� Q 4:0

� �0:50 = 4:0Q0:50

LRAC = 4:0 Q0:50

These cost functions are graphed in Panel (c) of Figure 8A.1. Note that when the Cobb- Douglas production function exhibits increasing returns to scale, total costs increase less than proportionately with output and average total costs fall as output increases.

FIGURE 8A.1 Long-Run Average Cost and Long-Run Total Cost Functions for a Cobb-Douglas Production Function

(a) β1 = 0.50, β2 = 0.50 Constant returns to scale

$ $

$ $

1.0 2.0 3.0 4.0 5.0

10 20 30 40 50

(b) β1 = 0.25, β2 = 0.25 Decreasing returns to scale

1.0 2.0 3.0 4.0 5.0

10 20 30 40 50

(c) β1 = 1.0, β2 = 1.0 Increasing returns to scale

LRAC LRTC

LRAC LRTC

LRAC LRTC

$

5.0 4.0 3.0 2.0 1.0

0

$

50 40 30 20 10

021 3 4 5 6 7 8 9 10 Q

10 Q

10 Q

21 3 4 5 6 7 8 9 10 Q

10 Q

10 Q

0 21 3 4 5 6 7 8 90 21 3 4 5 6 7 8 9

0 21 3 4 5 6 7 8 9 0 21 3 4 5 6 7 8 9

Appendix 8A: Long-Run Costs with a Cobb-Douglas Production Function, Advanced Material 303

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Exercises 1. Determine how many units of labor (L*) and capital (K*) are required to produce five units of output (Q0) for the production function given in the a. Constant returns example. b. Decreasing returns example. c. Increasing returns example.

2. Recompute your answers to Exercise 1, assuming that the per-unit cost of labor increases from CL = $2 to C0L = $4. How has the increase in the labor rate affected the optimal proportions of labor and capital used in the production process?

3. Use the data in Table 8.4 and a multiple regression analysis program on your computer to estimate a Cobb-Douglas production function of the form shown in Equation 8A.1. Do you observe increasing, decreasing, or constant returns to scale?

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

304 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

9 CHAP T E R

Applications of Cost Theory CHAPTER PREVIEW This chapter examines some of the techniques that have been developed for estimating the cost functions of production processes in actual firms. In the short run, knowledge of the firm’s cost function is essential when deciding whether to accept an additional order, perhaps at less than “full cost”; whether to schedule overtime for workers; or whether to temporarily suspend operations but not close the plant. In the long run, knowledge of cost- function relationships will determine the capital investments to make, the production technology to adopt, the markets to enter, and the new products to introduce. The first part of the chapter examines various techniques for empirically estimating short-run and long-run cost functions. The second part of the chapter deals with break-even and contribution analysis—an application of cost theory that is useful in examining the profitability of a firm’s operations.

MANAGERIAL CHALLENGE How Exactly Have Computerization and Information Technology Lowered Costs at Chevron, Timken, and Merck?1

Computerization and robotics have made output per worker higher and therefore lowered unit labor cost when it comes to processing insurance claims, redeem- ing coupons, or screening job resumes. Personal com- puters have decreased manyfold the time and talent required to perform routine work done previously with paper forms and time-consuming repetitive human tasks. However, not every business uses large numbers of PCs. How have computerization and information technology raised productivity and lowered cost so widely across other industries?

One key seems to be enhanced analytical and re- search and development (R&D) capability provided by computers and information technology (IT) systems. Chevron Corporation once spent anywhere from $2 to $4 million each to drill 10 to 12 exploratory wells before finding oil. Today, Chevron finds oil once in every five wells. The reason for the cost savings is a new technol- ogy that allows Chevron to display three-dimensional

graphs of the likely oil and gas deposits in potential oil fields. New fast parallel processors allow more calculation-intensive 3-D simulation modeling. Using only seismic data as inputs, Chevron can now model how the oil and gas deposits will shift and flow as a known field is pumped out. This allows a much more accurate location of secondary wells. As a result, overall production costs declined 16 percent industry-wide since 1991.

Timken, a $4-billion ball-bearing manufacturer, has also used digital 3D modeling to reconfigure pro- duction processes and implement small production runs of high-profit-margin products. Timken’s newest facility in North Carolina is a so-called flexible manufacturing system where order taking, limited cus- tomization of design, production scheduling, and the actual factory itself are all IT enabled and networked. Networked machine tools make it possible to build to order against precise specifications deliverable within

305

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

ESTIMATING COST FUNCTIONS To make optimal pricing and production decisions, the firm must have knowledge of the shape and characteristics of its short-run cost function. A cost function is a sched- ule, graph, or mathematical relationship showing the total, average, or marginal cost of producing various quantities of output. To decide whether to accept or refuse an order offered at some particular price, the firm must identify exactly what variable cost and direct fixed costs the order entails. A capability to estimate the short-run cost function is therefore crucial. In contrast, the long-run cost function is associated with the longer-term planning period in which all the inputs to the production pro- cess are variable and no restrictions are placed on the amount of an input that can be employed in the production process. Consequently, all costs, including indirect fixed costs such as headquarters facility costs, are avoidable and therefore relevant to cost estimates.

four hours rather than stockpile enormous inventories of subassemblies or insist that customers wait six to eight weeks, as was the practice before IT. Nissan re- cently estimated that $3,600 in the final price of an auto is tied up in inventory expense. The build- to-order system could save the auto industry as much as $50 billion per year out of its $80 billion inventory cost.

Pharmaceutical R&D has experienced a similar ben- efit from computerization. Drug industry basic research always starts with biochemical or biogenetic modeling of the disease mechanism. In the past, once a mecha- nism for Hodgkin’s disease or pancreatic cancer became well understood, researchers at Merck or Pfizer experi- mented on known active compounds one by one in time-consuming chemical trials. Successful therapies

emerged only after human trials on the promising com- pounds showed efficacy with few side effects. Total time to introduction of a new pharmaceutical was often lon- ger than a decade and entailed $1.5 billion in investments.

Today, the first stage of the basic research process remains much the same, but the second stage of drib- bling chemicals into a petri dish has ended. Instead, machines controlled and automated by microchips per- form thousands of reactions at once and tally the re- sults. Human researchers then take the most likely reagents and perform much more promising experi- ments that culminate in human trials. The total time to discovery has been cut by more than two-thirds, and all attendant costs have declined sharply.

Discussion Questions

� Name a business that you believe has experi- enced declining costs attributable to comput- erization. Were variable costs reduced? What fixed costs increase was involved? Does it seem clear that average total cost went down? Explain.

1Based on “The Innovators: The Rocket under the High-Tech Boom,” Wall Street Journal (March 30, 1999); “Mass Customization,” The Economist (July 14, 2001), pp. 64–67; and “The Flexible Factory,” BusinessWeek (May 5, 2003), pp. 90–101.

MANAGERIAL CHALLENGE Continued

© Ph ot oL in k/ Ph ot od is c/ Ge tty

Im ag es

306 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Issues in Cost Definition and Measurement Recall that economic cost is represented by the value of opportunities forgone, whereas accounting cost is measured by the outlays that are incurred. Some companies, such as Deep Creek Mining, record the cost of their own output (the crude oil, coal, or gas) shipped downstream to their refining-and-processing operations as expenses at the world market price on the day of shipment (i.e., at their opportunity cost). Other companies account for these same resources as their out-of-pocket expenses. If extraction costs of the company being studied are low (e.g., with Kentucky coal or West Texas intermediate crude or Persian Gulf oil), these two cost methods will diverge because the equilibrium market price is always determined by the considerably higher cost of the marginal pro- ducer (e.g., an oil platform in the North Sea).

Similar problems arise in measuring variable costs (i.e., costs that vary with output). Some companies employ only direct accounting costs, including materials, supplies, di- rect labor costs, and any direct fixed costs avoidable by refusing the batch order in ques- tion. Direct costs exclude all overhead and any other fixed cost that must be allocated (so-called indirect fixed costs). For batch decisions about whether to accept or refuse an order for a proposed charter air flight, a special production run, or a customer’s pro- posed change order, these estimates of variable plus direct fixed costs are needed. For other questions, such as offering a customized design, however, some indirect accounting cost for the IT system that allows customized design would be an appropriate inclusion in the cost data.

Several other cost measurement issues arise with depreciation. Conceptually, depre- ciation can be divided into two components: time depreciation represents the decline in value of an asset associated with the passage of time, and use depreciation represents the decline in value associated with use. For example, annual body style changes in the automobile industry or technical progress in speed and memory of personal computers renders products and production processes obsolete. Note that such time depreciation is completely independent of the rate of output at which the asset is actually operated. Because only use depreciation varies with the rate of output, only use depreciation is relevant in determining the shape of the cost-output relationship. However, accounting data on depreciation seldom break out use depreciation costs separately. Instead, the depreciation of the value of an asset over its life cycle is usually determined by arbi- trary tax regulations. Finally, capital asset values (and their associated depreciation costs) are often stated in terms of historical costs rather than in terms of replacement costs. In periods of rapidly increasing price levels, this approach will tend to understate true economic depreciation costs. These limitations need to be kept in mind when in- terpreting the cost-output relationship for a firm with numerous capital assets, such as an airline.

Controlling for Other Variables In addition to being a function of the output level of the firm, cost is a function of other factors, such as output mix, the length of production runs, employee absenteeism and turnover, production methods, input costs, and managerial efficiency.

To isolate the cost-output relationship itself, one must control for these other influ- ences by:

• Deflating or detrending the cost data. Whenever wage rates or raw material prices change significantly over the period of analysis, one can deflate the cost data to re- flect these changes in factor prices. Provided suitable price indices are available or

Chapter 9: Applications of Cost Theory 307

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

can be constructed, costs incurred at different points in time can be restated as inflation-adjusted real costs.2

• Using multiple regression analysis. Suppose a firm believes that costs should decline gradually over time as a result of innovative worker suggestions. One way to incor- porate this effect into the cost equation would be to include a time trend t as an additional explanatory variable:

C = f (Q, t) [9.1]

Other possible control variables include the number of product lines, the number of cus- tomer segments, and the number of distribution channels.

The Form of the Empirical Cost-Output Relationship The total cost function in the short run (SRTC), as hypothesized in economic theory, is an S-shaped curve that can be represented by a cubic relationship:

SRTC = a + bQ + cQ2 + dQ3 [9.2]

The familiar U-shaped marginal and average cost functions then can be derived from this relationship. The associated marginal cost function is

MC = dðSRTCÞ

dQ = b + 2cQ + 3dQ2 [9.3]

The average total cost function is

ATC = SRTC Q

= a Q

+ b + cQ + dQ2 [9.4]

The cubic total cost function and its associated marginal and average total cost func- tions are shown in Figure 9.1(a). If the results of a regression analysis indicate that the cubic term (Q3) is not statistically significant, then short-run total cost can be repre- sented by a quadratic relationship:

SRTC = a + bQ + cQ2 [9.5]

as illustrated in Figure 9.1(b). In this quadratic case, total costs increase at an increasing rate throughout the typical operating range of output levels. The associated marginal and average cost functions are

MC = dðSRTCÞ

dQ = b + 2cQ [9.6]

ATC = SRTC Q

= a Q

+ b + cQ [9.7]

As can be seen from Equation 9.6, this quadratic total cost relationship implies that mar- ginal costs increase linearly as the output level is increased.

2Two assumptions are implicit in this approach: No substitution takes place between the inputs as prices change, and changes in the output level have no influence on the prices of the inputs. For more automated plants that incorporate only maintenance personnel, plant engineers, and material supplies, these assumptions fit the reality of the production process quite well.

308 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIGURE 9.1 Polynomial Cost-Output Relationships

ATC = + b + cQa Q

SRTC = a + bQ + cQ2

MC = b + 2cQATC = + b + cQ + dQ2a Q

SRTC = a + bQ + cQ2 + dQ3

MC = b + 2cQ + 3dQ2

C os

t ($

)

Q (a) (b)

C os

t ($

/u ni

t)

Cubic total cost function

MC

C os

t ($

)

Q

C os

t ($

/u ni

t)

Quadratic total cost function

MC

ATC ATC

SRTC SRTC

WHAT WENT RIGHT • WHAT WENT WRONG

Boeing: The Rising Marginal Cost of Wide-Bodies3

Boeing and Airbus provide all the wide-bodied jets the world needs. Boeing 747s, 767s, and 777s typically have a 70 percent share of the worldwide market, but Airbus ac- cepted a majority of the new orders in 1994–1995 and doubled its output rate, especially on smaller planes, from 126 to 232 planes per year. Some analysts think Boeing should have given up even more of the order flow. Why?

One reason is that until recently, incremental orders at Boeing necessitated redrawing and duplicating the thou- sands of engineering diagrams that determine how 200,000 employees assemble any particular customer’s plane. Rather than doing mass customization from common platforms, Boeing assembles one plane at a time with new drawings for each $150 million wide-body ordered. Eventually, incre- mental variable costs must rise as designers and shop floors get congested with new instructions and diagrams.

With backorders running to almost 1,000 planes compa- nywide in the mid-1990s, Boeing boosted production from 180 to 560 commercial jets per year. At the final assembly plant for Boeing wide-bodies in Everett, Washington, just north of Seattle, throughput was increased from 15 planes

per month to 21 planes per month (i.e., by 40 percent). To increase production rates, Boeing needed to split bottle- necked assembly stations into parallel processes, which en- tailed the hiring of additional assembly workers andmassive overtime. Boeing also increased the production rate of final assembly by contracting out more subassemblies. Splitting bottlenecked assembly stations or contracting out subassem- blies substantially increases Boeing’s variable costs.

In the late 1990s, wide-body prices did not rise because of intense competitive pressure from Airbus, but Boeing’s marginal costs certainly did. As a result, for a while in the late 1990s, every wide-body plane delivered had a price less than its marginal cost (i.e., a negative gross profit margin). Of course, eventually such orders must be refused. In 2000, Boeing did slow the production throughput rate at Everett back to 15 wide-bodies per month in order to return to profitability. Today, the well-equipped 747-400 aircraft earns as much as $45 million in operating profits above its variable cost.

3Based on “Boeing’s Trouble,” Wall Street Journal (December 16, 1998), p. A23; and Everett, Washington, site visit.

Chapter 9: Applications of Cost Theory 309

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

But rising, not constant, marginal cost is characteristic of many manufacturing environments. On the other hand, some information services companies, such as IBM Global Services or network-based software companies such as Microsoft, may at times experience declining marginal costs.

Statistical Estimation of Short-Run Cost Functions Short-run cost functions have been estimated for firms in a large number of different industries—for example, food processing, furniture, railways, gas, coal, electricity, ho- siery, steel, and cement.

Statistical Estimation of Long-Run Cost Functions Long-run costs can be estimated over a substantial period of time in a single plant (time-series data) or with multiple plants operating at different rates of output (cross-sectional data). The use of cross-sectional data assumes that each firm is using its fixed plant and equipment and variable inputs to accomplish min LRAC production for that plant size along the envelop of SRAC curves we studied in Chapter 8.

The use of time-series data assumes that input prices, the production technology, and the products offered for sale remain unchanged. Both methods, therefore, require heroic assumptions, but cross-sectional data are more prevalent in estimating long-run cost functions.

Example Short-Run Cost Function for Multi-Product Food Processing In a study of a British food processing firm, Johnston constructed individual cost functions for 14 different products and an overall cost function for the firm.4

Weekly data for nine months were obtained on the physical production of each type of product and total direct costs of each product (subdivided into the four categories of materials, labor, packing, and freight). Indirect costs (such as sala- ries, indirect labor, factory charges, and laboratory expenses) remained fairly constant over the time period studied and were excluded from the analysis. A price index for each category of direct costs for each product was obtained from government sources and used to deflate all four sets of input costs, yielding a weekly deflated direct cost for each product. For the individual products, out- put was measured by physical production (quantity). For the firm as a whole, an index of aggregate output was constructed by weighting the quantities of each product by its selling price and summing over all products produced each period.

For the 14 different products and for the overall firm, the linear cost function gave an excellent fit between direct cost and output. Therefore, Johnston concluded that total direct costs were a linear function of output, and marginal costs were constant over the observed ranges of output.

4See Jack Johnston, Statistical Cost Analysis (New York: McGraw-Hill, 1960).

310 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Determining the Optimal Scale of an Operation The size at which a company should attempt to establish its operations depends on the extent of the scale economies and the extent of the market. Some firms can operate at minimum unit cost using a small scale. Consider a licensed street vendor of leather coats. Each additional sale entails variable costs for the coat, a few minutes of direct labor effort to answer potential customers’ questions, and some small allocated cost asso- ciated with the step-van or other vehicle where the inventory is stored and hauled from one street sale location to another. Ninety-nine percent of the operating cost is the vari- able cost of an additional leather coat per additional sale. Long-run average cost will be essentially flat, constant at approximately the wholesale cost of a leather coat. As a result, in street vending, a small-scale operation will be just as efficient as a large-scale operation.

Example Short-Run Cost Functions: Electricity Generation5

Another study by Johnston of the costs of electric power generation in Great Britain developed short-run cost functions for a sample of 17 different firms from annual cost-output data on each firm. To satisfy the basic conditions underlying the short-run cost function, only those firms whose capital equipment remained constant in size over the period were included in the sample. The output variable was measured in kilowatt-hours (kWh). The cost variable was defined as the “working costs of generation” and included: (1) fuel; (2) salaries and wages; and (3) repairs and maintenance, oil, water, and stores. This definition of cost does not correspond exactly with variable costs as long as maintenance is scheduled so as to just offset wear and tear from use. Each of the three cost categories was de- flated using an appropriate price index. A cubic polynomial function with an addi- tional linear time trend variable was fitted to each of 17 sets of cost-output observations.

The results of this study did not support the existence of a nonlinear cubic or quadratic cost function. The cubic term, Q3, was not statistically significant in any of the regressions, and the quadratic term, Q2, was statistically significant in only 5 of the 17 cost equations. A typical linear total cost function is given by

C = 18.3 + 0.889Q − 0.639T

where C = variable costs of generation, Q = annual output (millions of kilowatt- hours), and T = time (years). The equation “explained” 97.4 percent of the varia- tion in the cost variable.

The results of the two Johnston studies are similar to those found in many other cost studies—namely, that short-run total costs tend to increase linearly over the ranges of output for which cost-output data are available. In other words, short- run average costs tend to decline and marginal costs tend to be constant over the “typical” or “normal” operating range of the firm. At higher rates of output, we would expect to see rising marginal cost, but, of course, this circumstance is exactly what firms try to avoid. Recall Boeing’s experience in producing too many 747s per month.

5Ibid., pp. 44–63.

Chapter 9: Applications of Cost Theory 311

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Long-Run Cost Functions: Electricity Generation6

In a cross-sectional study of U.S. electric utility companies, Christensen and Greene used a logarithmic model to test for the presence of economies and dis- economies of scale. The long-run average cost curve (LRAC) using data on 114 firms is shown in Figure 9.2. The bar below the graph indicates the number of firms in each interval. Below 19.8 billion kWh (left arrow in graph), significant economies of scale were found to exist. The 97 firms in this range accounted for 48.7 percent of the total output. Between 19.8 and 67.1 billion kWh (right arrow in the graph), no significant economies of scale were present. The 16 firms in this range accounted for 44.6 percent of the total output. Above 67.1 billion kWh, diseconomies of scale (one firm and 6.7 percent of total output) were found.

6L. R. Christensen and W. H. Greene, “Economies of Scale in U.S. Electric Power Generation,” Journal of Political Economy 84:4 (August 1976).

FIGURE 9.2 Average Cost Function for U.S. Electric Utility Firms

0 10

Output (billion kWh)

20 30 40 50 60 70

6.615

6.458

6.300

6.143

5.985

5.828

5.670

5.513

5.355

5.198

5.040

4.883

4.725

5 15 25 35 45 55 65 75

Size distribution of firms

57

Significant economies

of scale

No significant economies or

diseconomies of scale

Significant diseconomies

of scale

21 12 7 4 5 4 1 1 1 1

A ve

ra ge

c os

t ($

/1 ,0

00 k

W h)

LRAC

312 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

In contrast, hydroelectric power plants have few variable costs of any kind. Instead, essentially all the costs are fixed costs associated with buying the land that will be flooded, constructing the dam, and purchasing the huge electrical generator equipment. Thereafter, the only variable inputs required are a few lubricants and maintenance workers. Conse- quently, a hydroelectric power plant has long-run average total costs that decline continu- ously as the company spreads its enormous fixed cost over additional sales by supplying power to more and more households. Similarly, electric distribution lines (the high-tension power grids and neighborhood electrical conduits) are a high-fixed-cost and low- variable-cost operation. In the electrical utility industry, large-scale operations therefore incur lower unit cost than small-scale operations, as demonstrated in Figure 9.2.

“Freewheeling” in the electrical utility industry has similar effects. When industrial and commercial electricity buyers (e.g., a large assembly plant or hospital) were allowed in January 2003 to contract freely with low-cost power suppliers elsewhere in the state or even several states away, the local public utility experienced “stranded costs.” That is, the high initial fixed costs of constructing dams, power plants, and distribution lines were left behind as sales volume declined and local customers opted out. If the costs

Example Scale Economies in the Traditional Cable Industry: Time-Warner7

Telephone landlines and traditional cable TV businesses have cost characteristics similar to electric utilities. Once the wires have been put in place, the incremental cost of extending TV or telephone service to another household is small. The ex- tent of the scale economies in such industries may warrant licensing only one cable company or one local telephone service provider. In fact, municipalities have his- torically issued an exclusive service contract to such public utilities. The rationale was that one firm could service the whole market at much lower cost than several firms dividing the market and failing therefore to realize all of the available scale economies.

However, remember that the optimal scale of operation of any facility, even a declining cost facility, is limited by the extent of the market. The expansion of the cable TV market has always been limited by the availability of videocassette recorders, DVD players, and services such as NetFlix because they are inexpensive, convenient entertainment substitutes. As a result, the potential scale economies suggested by industrial engineering studies of cable TV operations have never been fully realized.

In addition, both telephone and cable TV companies are now facing new wire- less alternative technologies. Satellite-based digital television and cell phones have cut deeply into the market once reserved exclusively for monopoly licensed com- munications companies. As a result, the average unit cost in these cable-based businesses increased from B to A as volume declined (see Figure 9.3). Conse- quently, the price required to break even has necessarily risen. Of course, that sets in motion a vicious circle; the higher the cost-covering price, the more custo- mers the cable TV and telephone companies lose to wireless alternatives.

7See W. Emmons and R. Prager, “The Effects of Market Structure in the U.S. Cable Television Industry,” Rand Jour- nal of Economics 28:4 (Winter 1997), pp. 732–750.

Chapter 9: Applications of Cost Theory 313

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

involved had been mostly variable, the local power utilities could have simply cut costs and operated profitably at smaller scale. Unfortunately, however, the costs are mostly fixed and unavoidable, so unit costs will unavoidably rise as the number of customers served declines. Consequently, the advantages of additional competition for lowering prices to consumers are projected to be almost completely offset by the rise in unit costs caused by reduced scale.8

Economies of Scale versus Economies of Scope Economies of scope occur whenever inputs can be shared in the production of different products. For example, in the airline industry, the cost of transporting both passengers and freight on a single airplane is less than the cost of using two airplanes to transport passengers and freight separately. Similarly, commercial banks that manage both credit card-based unsecured consumer loans and deeded property-secured mortgage loans can provide each activity at lower cost than they could be offered separately. These cost sav- ings occur independent of the scale of operations; hence they are distinguished from economies of scale.

Engineering Cost Techniques Engineering cost techniques provide an alternative way to estimate long-run cost func- tions without using accounting cost data. Using production data, the engineering

FIGURE 9.3 Fixed Costs Stranded by Freewheeling Electricity and Satellite-Based TV Signals

C os

t ($

/k W

h or

$ /h

ou se

ho ld

/m on

th )

DBeforeDAfter

LRAC

Quantity (kWh or households)

B

A

8M. Maloney and R. McCormick, Customer Choice, Consumer Value (Washington, DC: Citizens for a Sound Economy Foundation, 1996).

economies of scope Economies that exist whenever the cost of producing two (or more) products jointly by one plant or firm is less than the cost of producing these products separately by different plants or firms.

engineering cost techniques A method of estimating cost functions by deriving the least-cost combination of labor, capital equipment, and raw materials required to produce various levels of output, using only industrial engineering information.

314 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

approach attempts to determine the least-cost combination of labor, capital equipment, and raw materials required to produce various levels of output. Engineering methods offer a number of advantages over statistical methods in examining economies of scale. First, it is generally much easier with the engineering approach to hold constant such factors as input prices, product mix, and product efficiency, allowing one to isolate the effects on costs of changes in output. Second, use of the engineering method avoids some of the cost-allocation and depreciation problems encountered when using accounting data.

Example Economies of Scope in the Banking Industry A number of empirical studies have attempted to estimate economies of scale and scope in the banking industry, which includes commercial banks, savings and loan associations, and credit unions. A survey article by Jeffrey Clark compiled the re- sults of 13 of these studies.9 Possible sources of production economies in financial institutions include the following:

� Specialized labor. A larger depository institution may be able to employ more specialized labor (e.g., computer programmers, cash managers, investment spe- cialists, and loan officers) in producing its services. If the expertise of these workers results in the processing of a higher volume of deposit and loan ac- counts per unit of labor, then larger institutions will experience lower per-unit labor costs compared with smaller institutions.

� Computer and telecommunications technology. Once the large setup, or fixed, costs are incurred, computer and electronic funds transfer systems can be used to process additional transactions at small additional costs per transaction. Spreading the fixed costs over a higher volume of transactions may permit the larger firm to achieve lower average total costs.

� Information. Credit information about loan applicants must be gathered and analyzed before lending decisions are made. However, once gathered, this credit information can be reused, usually at little additional cost, in making decisions about lending to the institution’s customers. For example, credit information gathered in making mortgage loans can also be used in making automobile and other personal loans. Thus, larger financial institutions, which offer a wide array of different types of credit, may realize economies of scope in infor- mation gathering. That is, the cost of mortgage and auto installment lending done jointly is lower than the total cost of both when each is done separately.

All the studies reviewed by Clark employed a logarithmic cost function. The fol- lowing conclusions were derived:

� Some evidence indicates economies of scope between consumer and mortgage lending.

� Significant overall (i.e., firm-specific) economies of scale occur only at relatively low levels of output (less than $100 million in deposits). Beyond that point, most studies found an L-shaped long-run average cost curve where average to- tal cost falls steeply at low levels of output and then flattens out and becomes horizontal. In this respect, banking LRAC closely mirrors the shape of the LRAC in representative manufacturing.

9Jeffrey A. Clark, “Economies of Scale and Scope at Depository Financial Institutions: A Review of the Literature,” Federal Reserve Bank of Kansas City, Economic Review (September/October 1988), pp. 16–33.

Chapter 9: Applications of Cost Theory 315

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example The Survivor Technique: Steel Production The survivor technique has been used to examine the long-run cost functions in steel ingot production by open-hearth or Bessemer processes. Based on the data in Table 9.1, Stigler developed the sleigh-shaped long-run average cost function for steel ingot production shown in Figure 9.4. Because of the declining percen- tages at the lowest levels of output and at extremely high levels of output, Stigler concluded that both were relatively inefficient size classes. The intermediate size classes (from 2.5 to 27.5 percent of industry capacity) represented the range of op- timum size because these size classes grew or held their shares of capacity. Stigler also applied the survivor technique to the automobile industry and found an L-shaped average cost curve, indicating no evidence of diseconomies of scale at large levels of output.

TABLE 9.1 DISTRIBUTION OF STEEL INGOT CAPACITY BY RELATIVE SIZE OF COMPANY

COMPANY SIZE (PER- CENTAGE OF TOTAL INDUSTRY CAPACITY)

PERCENTAGE OF INDUSTRY CAPACITY NUMBER OF COMPANIES

1930 1938 1951 1930 1938 1951

Under ½ 7.16 6.11 4.65 39 29 22

½ to 1 5.94 5.08 5.37 9 7 7

1 to 2½ 13.17 8.30 9.07 9 6 6

2½ to 5 10.64 16.59 22.21 3 4 5

5 to 10 11.18 14.03 8.12 2 2 1

10 to 25 13.24 13.99 16.10 1 1 1

25 and over 38.67 35.91 34.50 1 1 1

Source: George J. Stigler, “The Economies of Scale,” Journal of Law and Economics (October 1958). Reprinted by permission.

FIGURE 9.4 Long-Run Average Costs of Steel Ingot Production

Percentage of industry capacity

LRAC

5 10 15 20 25 30

A ve

ra ge

c os

t

316 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

In a study designed to isolate the various sources of scale economies within a plant, Haldi and Whitcomb collected data on the cost of individual units of equipment, the initial investment in plant and equipment, and on operating costs. They noted that “in many basic industries such as petroleum refining, primarymetals, and electric power, economies of scale are found in very large plant sizes (often the largest built or contemplated).”10 Few (if any) firms were observed operating beyond these minimum efficient scale plant sizes.

The Survivor Technique It is also possible to detect the presence of scale economies or diseconomies without hav- ing access to any cost data. The survivor technique involves classifying the firms in an industry by size and calculating the share of industry output coming from each size class over time.11 If the share decreases over time, then this size class is presumed to be rela- tively inefficient and to have higher average costs. Conversely, an increasing share indi- cates that the size class is relatively efficient and has lower average costs. The rationale for this approach is that competition will tend to eliminate those firms whose size is rel- atively inefficient, allowing only those size firms with lower average costs to survive.

Despite its appeal, the survivor technique does have one serious limitation. Because the technique does not use actual cost data in the analysis, it cannot assess the magnitude of the cost differentials between firms of varying size and efficiency.

A Cautionary Tale One final note of caution: The concept of average total costs (ATC) per unit of output (i.e., so-called unit costs), so prominent in our recent discussion of scale economies, is seldom useful for managerial decision making. Indeed, making output or pricing decisions based on ATC is dead wrong. AVC and marginal cost determine optimal shutdown, optimal out- put, and optimal price decisions. Managers in prominent companies like British Telephone have been fired over this mistake when they included headquarters expense and other corpo- rate overhead in a pricing decision for an incremental new account. So, get in the habit of avoiding the use of unit costs in your decision problem reasoning. Reserve unit costs for de- scribing, debating, and planning issues related to scale economies and diseconomies alone.

BREAK-EVEN ANALYSIS Many of the planning activities that take place within a firm are based on anticipated levels of output. The study of the interrelationships among a firm’s sales, costs, and op- erating profit at various anticipated output levels is known as break-even analysis.

Break-even analysis is based on the revenue-output and cost-output functions of mi- croeconomic theory. These functions are shown together in Figure 9.5. Total revenue is equal to the number of units of output sold multiplied by the price per unit. Assuming that the firm can sell additional units of output only by lowering the price, the total rev- enue curve TR will be concave (inverted U shaped), as indicated in Figure 9.5.

The difference between total revenue and total cost at any level of output represents the total profit that will be obtained. In Figure 9.5, total profit TP at any output level is given by the vertical distance between the total revenue TR and total cost TC curves. A break-even situation (zero profit) occurs whenever total revenue equals total cost. Below an output level of Q1, losses will be incurred because TR < TC. Between Q1 and Q3,

10J. Haldi and D. Whitcomb, “Economies of Scale in Industrial Plants,” Journal of Political Economy 75, no. 1 (August 1967), pp. 373–385.

survivor technique A method of estimating cost functions from the shares of industry output coming from each size class over time. Size classes whose shares of industry output are increasing (decreasing) over time are presumed to be relatively efficient (inefficient) and have lower (higher) average costs.

11G. J. Stigler, The Organization of Industry (Homewood, IL: Richard D. Irwin 1968), Chapter 7. For other examples of the use of the survivor technique, see H. E. Ted Frech and Paul B. Ginsburg, “Optimal Scale in Medical Practice: A Survivor Analysis,” Journal of Business (January 1974), pp. 23–26.

break-even analysis A technique used to examine the relationship among a firm’s sales, costs, and operating profits at various levels of output.

Chapter 9: Applications of Cost Theory 317

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

profits will be obtained because TR > TC. At output levels above Q3, losses will occur again because TR < TC. Total profits are maximized within the range of Q1 to Q3; the vertical distance between the TR and TC curves is greatest at an output level of Q2.

We now discuss both a graphical and an algebraic method of solving break-even problems.

Graphical Method Constant selling price per unit and a constant variable cost per unit yield the linear TR and TC functions illustrated in Figure 9.6, which shows a basic linear break-even chart. Total cost is computed as the sum of the firm’s fixed costs F, which are independent of

FIGURE 9.5 Generalized Break-Even Analysis

Q2

Break-even point Profit

Loss

LossTC

Q1 = Lower break-even output Q2 = Maximum profit output Q3 = Upper break-even output

TP

Q1

Loss

0

Cost Revenue

Profit ($)

Output QLoss Q3

Break-even point

TR

Profit

Example Boeing 777 Exceeds Break-Even Sales Volume12

Boeing and Airbus, for example, are constantly calculating and recalculating their break-even sales volumes as unanticipated development costs arise on their new planes. The new double-decked jumbo jet, the Airbus 380, has $11.7 billion in development cost requiring 259 planes at undiscounted prices to break even. Advance orders have only secured 160, much less than the break-even amount. Although Airbus has sold more total planes than Boeing in recent years, Boeing has dominated the wide-bodied submarket for larger jets with a 70 percent market share. For example, by 2006 Boeing had secured 155 orders for its 777 long-haul jet whereas Airbus had orders for only 15 of its competing Airbus 340s. Break-even appears far off on the 340s as well.

12“Testing Times,” The Economist (April 1, 2006), p. 56.

318 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the output level, and the variable costs, which increase at a constant rate of VC per unit of output. Operating profit is equal to the difference between total revenues (TR) and total (operating) costs (TC).

The break-even point occurs at point Qb in Figure 9.5, where the total revenue and the total cost functions intersect. If a firm’s output level is below this break-even point (i.e., if TR < TC), it incurs operating losses. If the firm’s output level is above this break- even point (if TR > TC), it realizes operating profits.

Algebraic Method To determine a firm’s break-even point algebraically, one must set the total revenue and total (operating) cost functions equal to each other and solve the resulting equation for the break- even volume. Total revenue is equal to the selling price per unit times the output quantity:

TR = P × Q [9.8] Total (operating) cost is equal to fixed plus variable costs, where the variable cost is the product of the variable cost per unit times the output quantity:

TC = F + (V × Q) [9.9] Setting the total revenue and total cost expressions equal to each other and substituting the break-even output Qb for Q results in

TR = TC

or PQb = F + VQb [9.10]

Finally, solving Equation 9.10 for the break-even output Qb yields 13

FIGURE 9.6 Linear Break-Even Analysis Chart

R ev

en ue

, c os

t ($

)

TC

0

Output Q (units)

Qb

TR

F

Q1 Q2

Operating Loss

Break-even point

Relevant range

Positive Operating Profit

P

V

13Break-even analysis also can be performed in terms of dollar sales rather than units of output. The break- even dollar sales volume Sb can be determined by the following expression:

Sb = F

1 − V=P

where V/P is the variable cost ratio (calculated as variable cost per dollar of sales).

Chapter 9: Applications of Cost Theory 319

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PQb − VQb = F ðP − VÞQb = F

Qb = F

P − V [9.11]

The difference between the selling price per unit and the variable cost per unit, P – V, is referred to as the contribution margin. It measures how much each unit of output contributes to meeting fixed costs and operating profits. Thus, the break-even output is equal to the fixed cost divided by the contribution margin.

Because a firm’s break-even output is dependent on a number of variables—in partic- ular, the price per unit, variable (operating) costs per unit, and fixed costs—the firm may wish to analyze the effects of changes in any of the variables on the break-even output. For example, it may wish to consider either of the following:

1. Change the selling price. 2. Substitute fixed costs for variable costs.

Example Break-Even Analysis: Allegan Manufacturing Company Assume that Allegan manufactures one product, which it sells for $250 per unit (P). Variable costs (V) are $150 per unit. The firm’s fixed costs (F) are $1 million. Substituting these figures into Equation 9.11 yields the following break-even output:

Qb = $1,000,000 $250 − $150

= 10,000 units

Allegan’s break-even output can also be determined graphically, as shown in Figure 9.7.

Another illustration would be to use break-even analysis to approve or reject a batch sale promotion. Suppose that in the previous example, the $1 million is a trade rebate to elicit better shelf location for Allegan’s product. If the estimated effect of this promotion is additional sales of 9,000 units, which is less than the break-even output, the change in total contributions will fall below the $1 million promotion cost (i.e., [$250 − $150] × 9,000 < $1,000,000). Therefore, the promo- tion plan should be rejected.

Example Break-Even Analysis: Allegan Manufacturing Company (continued) Assume that Allegan increased the selling price per unit P 0 by $25 to $275. Substi- tuting this figure into Equation 9.11 gives a new break-even output:

Q0b = $1,000,000 $275 − $150

= 8,000 units

(Continued)

contribution margin The difference between price and variable cost per unit.

320 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIGURE 9.7 Linear Break-Even Analysis Chart for the Allegan Manufacturing Company

0

Output Q (units)

1,000,000

10,000

2,000,000

3,000,000

4,000,000

5,000,000

2,000 6,000 14,000 18,000

Slope of TC = V

F

= $250/unit Slope of TR = P

= $150/unit

R ev

en ue

, c os

t ($

)

TR

TC

This outcome can also be seen in Figure 9.8, in which an increase in the price per unit increases the slope of the total revenue function TR/Q and reduces the break-even output.

Rather than increasing the selling price per unit, Allegan’s management may decide to substitute fixed costs for variable costs in some aspect of the com- pany’s operations. For example, as labor wage rates increase over time, many firms seek to reduce operating costs through automation, which in effect repre- sents the substitution of fixed-cost capital equipment for variable-cost labor. Suppose Allegan determines that it can reduce labor costs by $25 per unit by leasing $100,000 of additional equipment. Under these conditions, the firm’s new level of fixed costs F 0 would be $1,000,000 + $100,000 = $1,100,000. Vari- able costs per unit V 0 would be $150 − $25 = $125. Substituting P = $250 per unit, V 0 = $125 per unit, and F 0 = $1,100,000 into Equation 9.11 yields a new break-even output:

Q0b = $1,000,000 $250 − $125

= 8,800 units

Graphically, the effect of this change in cost fixity of the operations is to raise the intercept on the vertical axis, decrease the slope of the total (operating) cost func- tion TC 0, and reduce the break-even output.

Chapter 9: Applications of Cost Theory 321

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIGURE 9.8 Linear Break-Even Analysis Chart for the Allegan Manufacturing Company Showing the Effects of a Price Increase

TC

0

Output Q (units)

1,000,000

10,000

2,000,000

3,000,000

4,000,000

5,000,000

2,000 6,000 14,000 18,000

TR

Q�

TR�

b

R ev

en ue

, c os

t ($

)

= $275/unit Slope of TR� = P�

Example Fixed Costs and Production Capacity at General Motors14

In an industry with 17 million unit sales annually, GM admitted in March 2002 that it needed to reduce automobile production capacity by 1 million cars per year to match its current sales of 5 million cars. It represented the second time in its 100-year history (1988 being the earlier event) that the company had signifi- cantly shrunk its capacity. As part of its decision to reduce its size, GM planned to close 10 of its U.S. automobile assembly lines.

In the past, GM alternated between (1) building all the cars it could produce and then using costly clearance sales to attract buyers, and (2) reducing output by running plants below capacity through a slowdown in the pace of the assembly line or elimination of an entire shift. The new strategy called for the company to use 100 percent of its American automobile production capacity five days a week with two shifts per day. If automobile demand increased above this capacity level, third- shift operations would be used to boost production. Ford had been following this strategy for some time.

In effect, GM and Ford were trading off lower fixed costs over the entire busi- ness cycle against (the possibility of) having to incur higher variable costs (e.g., use of higher cost overtime and third-shift operations) during periods of strong de- mand. As a consequence, GM’s break-even output point declined sharply.

14Jacob M. Schlesinger, “GM to Reduce Capacity to Match Its Sales,” Wall Street Journal (April 25, 1988), p. 2; Lawrence Ingrassia and Joseph B. White, “GM Plans to Close 21 More Factories, Cut 74,000 Jobs, Slash Capital Spend- ing,” Wall Street Journal (December 19, 1991), p. A3; and “A Duo of Dunces,” The Economist (March 9, 2002), p. 63.

322 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Doing a Break-Even versus a Contribution Analysis A break-even analysis assumes that all types of costs except the narrowly defined incre- mental variable cost (V) of additional unit sales are avoidable and asks the question of whether sufficient unit sales are available at the contribution margin (P – V) to cover all these relevant costs. If so, they allow the firm to earn a net profit. These questions normally arise at entry and exit decision points where a firm can avoid essentially all its costs if the firm decides to stay out or get out of a business. Contribution analysis, in contrast, applies to questions such as whether to adopt an advertising campaign, in- troduce a new product, shut down a plant temporarily, or close a division. What distin- guishes these contribution analysis questions is that many fixed costs remain unavoidable and are therefore irrelevant to the decision (indirect fixed costs), while other fixed costs will be newly committed as a result of the decision (direct fixed costs) and therefore could be avoided by refusing to go ahead with the proposal.

More generally, contribution analysis always asks whether enough additional revenue arises from the ad campaign, the new product, or the projected sales of the plant or division to cover the direct fixed plus variable costs. That is, contribution analysis cal- culates whether sufficient gross operating profits result from the incremental sales (ΔQ) attributable to the ad, the new product, or the promotion to offset the proposed increase in fixed cost. In other words, are the total contributions to cover fixed cost increased by an amount greater than the increase in direct fixed cost avoidable by the decision?

ðP − VÞ ΔQ > Δ Total Fixed Cost > Δ Indirect Fixed Cost + ΔDirect Fixed Cost > 0 + ΔDirect Fixed Cost

[9.12]

Such decisions are not break-even decisions because they ignore (abstract from) the indirect fixed costs that, by definition, cannot be avoided by rejecting the ad campaign or new product introduction proposal or by closing the plant temporarily. For example, headquarters facility cost and other corporate overhead are indirect fixed costs that can- not be avoided by any of these decisions. So, corporate overhead is not a relevant cost in making these decisions and is therefore ignored in the contribution analysis done to sup- port making such decisions.

In contrast, corporate overhead is prominent in the preceding examples of break-even analysis done to decide how or whether to enter a new business in the first place. Busi- ness certification, licensing, or franchise fees would be a good example of this concept of corporate overhead. The case exercise on charter airline operating decisions at the end of this chapter illustrates the use of contribution analysis as distinguished from break-even analysis.

Some Limitations of Break-Even and Contribution Analysis Break-even analysis has a number of limitations that arise from the assumptions made in constructing the model and developing the relevant data.

Composition of Operating Costs In doing break-even analysis, one assumes that costs can be classified as either fixed or variable. In fact, some costs are partly fixed and partly variable (e.g., utility bills). Furthermore, some fixed costs increase in a stepwise manner as output is increased; they are semivariable. For example, machinery mainte- nance is scheduled after 10 hours or 10 days or 10 weeks of use. These direct fixed costs must be considered variable if a batch production decision entails this much use.

contribution analysis A comparison of the additional operating profits to the direct fixed costs attributable to a decision.

Chapter 9: Applications of Cost Theory 323

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Multiple Products The break-even model also assumes that a firm is producing and selling either a single product or a constant mix of different products. In many cases the product mix changes over time, and problems can arise in allocating fixed costs among the various products.

Uncertainty Still another assumption of break-even analysis is that the selling price and variable cost per unit, as well as fixed costs, are known at each level of output. In practice, these parameters are subject to uncertainty. Thus, the usefulness of the results of break-even analysis depends on the accuracy of the estimates of the future selling price and variable cost.

Inconsistency of Planning Horizon Finally, break-even analysis is normally performed for a planning period of one year or less; however, the benefits received from some costs may not be realized until subsequent periods. For example, research and development costs incurred during a specific period may not result in new products for several years. For break-even analysis to be a dependable decision-making tool, a firm’s operating costs must be matched with resulting revenues for the planning period under consideration.

Operating Leverage Operating leverage involves the use of assets that have fixed costs. A firm uses operating leverage in the hope of earning returns in excess of the fixed costs of the assets, thereby increasing the returns to the owners of the firm. A firm’s degree of operating leverage (DOL) is defined as the multiplier effect resulting from the firm’s use of fixed operating costs. More specifically, DOL can be computed as the percentage change in earnings be- fore interest and taxes (EBIT) resulting from a given percentage change in sales (output):

DOL at Q = Percentage change in EBIT Percentage change in Sales

Example Taco Bell Chihuahua Drives Sales Consider the Taco Bell ad campaign with the cute little dog that was designed to pulse twenty-five 15-second spot commercials over several weeks. The ad agency quoted a cost of $750,000 per spot to secure prime-time network television reach- ing 176 million households. To decide whether to buy this ad campaign, we need to know just two things: (1) the incremental sales that demand analysis suggests will be stimulated by this campaign and (2) the contribution margin in dollars. Suppose the incremental sales are estimated at 2,100 Taco Bell meals per day for 90 days across 48 states, totaling 9,072,000 meals. If $7.99 is the average price per realized unit sale and variable costs are $5.00, should Taco Bell go ahead with the ad? The answer is yes, because when we apply Equation 9.12,

ð$7:99 − $5:00Þ 9,072,200 > 0 + ð25 × $750,000Þ $27,125,280 > $18,750,000

we see that Taco Bell would increase its operating profit by $8.4 million and make further contributions toward covering fixed cost and profit if it authorized the pro- posed ad campaign.

operating leverage The use of assets having fixed costs (e.g., depreciation) in an effort to increase expected returns.

degree of operating leverage (DOL) The percentage change in a firm’s earnings before interest and taxes (EBIT) resulting from a given percentage change in sales or output.

324 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

This relationship can be rewritten as follows:

DOL at Q =

ΔEBIT EBIT

ΔSales Sales

[9.13]

where ΔEBIT and ΔSales are the changes in the firm’s EBIT and Sales, respectively. Because a firm’s DOL differs at each sales level, it is necessary to indicate the sales

point, Q, at which operating leverage is measured. The degree of operating leverage is analogous to the elasticity of demand concept (e.g., price and income elasticities) because it relates percentage changes in one variable (EBIT) to percentage changes in another variable (sales). Equation 9.13 requires the use of two different values of sales and EBIT. Another equation (derived from Equation 9.13) that can be used to compute a firm’s DOL more easily is

DOL at Q = Sales − Variable costs

EBIT [9.14]

The variables defined in the previous section on break-even analysis can also be used to develop a formula for determining a firm’s DOL at any given output level. Because sales are equivalent to TR (or P × Q), variable cost is equal to V × Q, and EBIT is equal to total revenue (TR) less total (operating) cost, or (P × Q) – F – (V × Q), these values can be substituted into Equation 9.14 to obtain the following:

DOL at Q = ðP · QÞ − ðV · QÞ

ðP · QÞ − F − ðV · QÞ or

DOL at Q = ðP −VÞQ

ðP −VÞQ − F [9.15]

Example Operating Leverage: Allegan Manufacturing Company (continued) In the earlier discussion of break-even analysis for the Allegan Manufacturing Com- pany, the parameters of the break-even model were determined as P = $250/unit, V = $150/unit, and F = $1,000,000. Substituting these values into Equation 9.15 along with the respective output (Q) values yields the DOL values shown in Table 9.2. For example, a DOL of 6.00 at an output level of 12,000 units indicates that from a base output level of 12,000 units EBIT will increase by 6.00 percent for each 1 percent increase in output.

Note that Allegan’s DOL is largest (in absolute value terms) when the firm is operating near the break-even point (where Q = Qb = 10,000 units). Note also that the firm’s DOL is negative below the break-even output level. A negative DOL indicates the percentage reduction in operating losses that occurs as the result of a 1 percent increase in output. For example, the DOL of −1.50 at an output level of 6,000 units indicates that from a base output level of 6,000 units the firm’s op- erating losses will be reduced by 1.5 percent for each 1 percent increase in output.

(Continued)

Chapter 9: Applications of Cost Theory 325

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Business Risk Business risk refers to the inherent variability or uncertainty of a firm’s EBIT. It is a function of several factors, one of which is the firm’s DOL. The DOL is a measure of how sensitive a firm’s EBIT is to changes in sales. The greater a firm’s DOL, the larger the change in EBIT will be for a given change in sales. Thus, all other things being equal, the higher a firm’s DOL, the greater the degree of business risk.

Other factors can also affect a firm’s business risk, including the variability or uncer- tainty of sales. A firm with high fixed costs and stable sales will have a high DOL, but it will also have stable EBIT and, therefore, low business risk. Public utilities and pipeline transportation companies are examples of firms having these operating characteristics.

Another factor that may affect a firm’s business risk is uncertainty concerning selling prices and variable costs. A firm having a low DOL can still have high business risk if selling prices and variable costs are subject to considerable variability over time. A cattle feedlot illustrates these characteristics of low DOL but high business risk; both grain costs and the selling price of beef at times fluctuate wildly.

In summary, a firm’s DOL is only one of several factors that determine the firm’s business risk.

Break-Even Analysis and Risk Assessment The break-even unit sales figure can also be used to assess the business risk to which a firm is exposed. If one forecasts the mean unit sales for some future period of time, the standard deviation of the distribution of unit sales, and makes an assumption about how

A firm’s DOL is a function of the nature of the production process. If the firm employs large amounts of equipment in its operations, it tends to have relatively high fixed operating costs and relatively low variable operating costs. Such a cost structure yields a high DOL, which results in large operating profits (positive EBIT) if sales are high and large operating losses (negative EBIT) if sales are depressed.

TABLE 9.2 DOL AT VARIOUS OUTPUT LEVELS FOR ALLEGAN

MANUFACTURING COMPANY

OUTPUT DEGREE OF OPERATING LEVERAGE

Q (DOL)

0 0

2,000 −0.25

4,000 −0.67

6,000 −1.50

8,000 −4.00

10,000 (undefined) Break-even level

12,000 +6.00

14,000 +3.50

16,000 +2.67

18,000 +2.25

20,000 +2.00

business risk The inherent variability or uncertainty of a firm’s operating earnings (earnings before interest and taxes).

326 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

actual sales are distributed, one can compute the probability that the firm will have operating losses, meaning it will sell fewer units than the break-even level.

The probability of having operating losses (selling fewer than Qb units) can be com- puted using the following equation and the standard normal probability distribution as

z = Qb − Q σQ

[9.16]

where the probability values are from Table 1 in Appendix B, Q is the expected unit sales, σQ is the standard deviation of unit sales, and Qb is (as defined earlier) the break- even unit sales. The probability of operating profits (selling more than Qb units) is equal to 1 minus the probability of operating losses.

Example Business Risk Assessment: Allegan Manufacturing Company (continued) For the Allegan Manufacturing Company discussed earlier, suppose that expected sales are 15,000 units with a standard deviation of 4,000 units. Recall that the break-even volume was 10,000 units. Substituting Qb = 10,000, Q = 15,000, and σQ = 4,000 into Equation 9.16 yields

z = 10,000 − 15,000

4,000

= −1:25

In other words, the break-even sales level of 10,000 units is 1.25 standard deviations below the mean. From Table 1 in Appendix B, the probability associated with −1.25 standard deviations is 0.1056 or 10.56 percent. Thus, Allegan faces a 10.56 percent chance that it will incur operating losses and an 89.44 percent chance (100 – 10.56 percent chance of losses) that it will record operating profits from selling more than the break-even number of units of output.

SUMMARY

� In estimating the behavior of short-run and long- run cost functions for firms, the primary method- ological problems are (1) differences in the manner in which economists and accountants define and measure costs and (2) accounting for other vari- ables (in addition to the output level) that influence costs.

� Many statistical studies of short-run cost-output relationships suggest that total costs increase line- arly (or quadratically) with output, implying con- stant (or rising) marginal costs over the observed ranges of output.

� Many statistical studies of long-run cost-output re- lationships indicate that long-run cost functions are L-shaped. Economies of scale (declining aver- age costs) occur at low levels of output. Thereafter, long-run average costs remain relatively constant over large ranges of output. Diseconomies of scale are observed in only a few cases, probably because few firms can survive with costs attributable to ex- cessive scale.

� Engineering cost techniques are an alternative ap- proach to statistical methods in estimating long- run cost functions. With this approach, knowledge

Chapter 9: Applications of Cost Theory 327

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

of production facilities and technology is used to determine the least-cost combination of labor, cap- ital equipment, and raw materials required to pro- duce various levels of output.

� The survivor technique is a method of determining the optimum size of firms within an industry by classifying them by size and then calculating the share of industry output coming from each size class over time. Size classes whose share of industry output is increasing over time are considered to be more efficient and to have lower average costs.

� Break-even analysis is used to examine the rela- tionship among a firm’s revenues, costs, and oper- ating profits (EBIT) at various output levels. Frequently the analyst constructs a break-even chart based on linear cost-output and revenue- output relationships to determine the operating characteristics of a firm over a limited output range.

� The break-even point is defined as the output level at which total revenues equal total costs of opera- tions. In the linear break-even model, the break-

even point is found by dividing fixed costs by the difference between price and variable cost per unit, the contribution margin.

� Contribution analysis is used to examine operating profitability when some fixed costs (indirect fixed costs) cannot be avoided and other direct fixed costs can be avoided by a decision. Decisions on advertising, new product introduction, shutdown, and downsizing are often made by doing a contri- bution analysis.

� Operating leverage occurs when a firm uses assets having fixed operating costs. The degree of operat- ing leverage (DOL) measures the percentage change in a firm’s EBIT resulting from a 1 percent change in sales (or units of output). As a firm’s fixed operating costs rise, its DOL increases.

� Business risk refers to the variability of a firm’s EBIT. It is a function of several factors, including the firm’s DOL and the variability of sales. All other things being equal, the higher a firm’s DOL, the greater is its business risk.

Exercises 1. A study of 86 savings and loan associations in six northwestern states yielded the following cost function:15

C ¼ 2:38 − :006153Q + :000005359Q2 + 19:2X1 ð2:84Þ ð2:37Þ ð2:63Þ ð2:69Þ

where C = average operating expense ratio, expressed as a percentage and defined as total operating expense ($ million) divided by total assets ($ million) times 100 percent

Q = output; measured by total assets ð$millionÞ X1 = ratio of the number of branches to total assets ð$millionÞ

Note: The number in parentheses below each coefficient is its respective t-statistic.

a. Which variable(s) is(are) statistically significant in explaining variations in the average operating expense ratio?

b. What type of cost-output relationship (e.g., linear, quadratic, or cubic) is suggested by these statistical results?

c. Based on these results, what can we conclude about the existence of econo- mies or diseconomies of scale in savings and loan associations in the Northwest?

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

15Holton Wilson, “A Note on Scale Economies in the Savings and Loan Industry,” Business Economics (January 1981), pp. 45–49.

328 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

2. Referring to Exercise 1 again: a. Holding constant the effects of branching (X1), determine the level of total

assets that minimizes the average operating expense ratio. b. Determine the average operating expense ratio for a savings and loan asso-

ciation with the level of total assets determined in Part (a) and 1 branch. Same question for 10 branches.

3. A study of the costs of electricity generation for a sample of 56 British firms in 1946–1947 yielded the following long-run cost function:16

AVC = 1.24 + .0033Q + .0000029Q2 − .000046QZ − .026Z + .00018Z2

where AVC = average variable cost (i.e., working costs of generation), measured in pence per kilowatt-hour (kWh). (A pence was a British monetary unit equal, at that time, to 2 cents U.S.)

Q = output; measured in millions of kWh per year Z = plant size; measured in thousands of kilowatts

a. Determine the long-run variable cost function for electricity generation. b. Determine the long-run marginal cost function for electricity generation. c. Holding plant size constant at 150,000 kilowatts, determine the short-run

average variable cost and marginal cost functions for electricity generation. d. For a plant size equal to 150,000 kilowatts, determine the output level that

minimizes short-run average variable costs. e. Determine the short-run average variable cost and marginal cost at the out-

put level obtained in Part (d).

4. Assuming that all other factors remain unchanged, determine how a firm’s break- even point is affected by each of the following: a. The firm finds it necessary to reduce the price per unit because of increased

foreign competition. b. The firm’s direct labor costs are increased as the result of a new labor

contract. c. The Occupational Safety and Health Administration (OSHA) requires the

firm to install new ventilating equipment in its plant. (Assume that this action has no effect on worker productivity.)

5. Cool-Aire Corporation manufactures a line of room air conditioners. Its break- even sales level is 33,000 units. Sales are approximately normally distributed. Expected sales next year are 40,000 units with a standard deviation of 4,000 units. a. Determine the probability that Cool-Aire will incur an operating loss. b. Determine the probability that Cool-Aire will operate above its break-even

point.

6. McKee Corporation has annual fixed costs of $12 million. Its variable cost ratio is .60. a. Determine the company’s break-even dollar sales volume. b. Determine the dollar sales volume required to earn a target profit of

$3 million.

16Johnston, Statistical Cost Analysis, Chapter 4, op. cit.

Chapter 9: Applications of Cost Theory 329

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Case Exercises COST FUNCTIONS

The following cost-output data were obtained as part of a study of the economies of scale in operating a charter high school in Wisconsin:17

STUDENTS IN AVERAGE DAILY ATTENDANCE

MIDPOINT OF VALUES IN COLUMN A

OPERATING EXPENDITURE PER STUDENT

NUMBER OF SCHOOLS IN

SAMPLE

(A) (B) (C) (D)

143–200 171 $531.9 6

201–300 250 480.8 12

301–400 350 446.3 19

401–500 450 426.9 17

501–600 550 442.6 14

601–700 650 413.1 13

701–900 800 374.3 9

901–1,100 1,000 433.2 6

1,101–1,600 1,350 407.3 6

1,601–2,400 2,000 405.6 7

Questions 1. Plot the data in columns B and C in an output (enrollment-) cost graph and

sketch a smooth curve that would appear to provide a good fit to the data. 2. Based on the scatter diagram in Question 1, what kind of mathematical relation-

ship would appear to exist between enrollment and operating expenditures per student? In other words, do operating expenditures per student appear to (i) be constant (and independent of enrollment), (ii) follow a linear relationship as en- rollment increases, or (iii) follow some sort of nonlinear U-shape (possibly qua- dratic) relationship as enrollment increases?

As part of this study, the following cost function was developed:

C = f(Q, X1, X2, X3, X4, X5)

where C = operating expenditures per student in average daily attendance ðmeasured in dollarsÞ

Q = enrollment ðnumber of students in average daily attendanceÞ X1 = average teacher salary X2 = number of credit units ð“courses”Þ offered X3 = average number of courses taught per teacher X4 = change in enrollment between 1957 and 1960 X5 = percentage of classrooms built after 1950

Variables X1, X2, and X3 were considered measures of teacher qualifications, breadth of curriculum, and the degree of specialization in instruction, respec- tively. Variable X4 measured changes in demand for school services that could cause some lagging adjustments in cost. Variable X5 was used to reflect any

17John Riew, “Economies of Scale in High School Operation,” Review of Economics and Statistics 48:3 (August 1966), pp. 280–287.

330 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

differentials in the costs of maintenance and operation due to the varying ages of school properties. Statistical data on 109 selected high schools yielded the follow- ing regression equation:

C = 10:31 − :402Q + :00012Q2 + :107X1 + :985X2 + 15:62X3 + :613X4 − :102X5 ð6:4Þ* ð5:2Þ* ð8:2Þ* ð:15Þ ð1:3Þ ð3:2Þ* ð:93Þ

r2= :557

Notes: The numbers in parentheses are the t-scores of each of the respective (b) coefficients. An asterisk (*) indicates that the result is statistically significant at the 0.01 level.

3. What type of cost-output relationship (linear, quadratic, cubic) is suggested by these statistical results?

4. What variables (other than enrollment) would appear to be most important in explaining variations in operating expenditures per student?

5. Holding constant the effects of the other variables (X1 through X5), determine the enrollment level (Q) at which average operating expenditures per student are minimized. (Hint: Find the value of Q that minimizes the (∂C/∂Q function.)

6. Again, holding constant the effects of the other variables, use the ∂C/∂Q function to determine, for a school with 500 students, the reduction in per-student operat- ing expenditures that will occur as the result of adding one more student.

7. Again, holding the other variables constant, what would be the saving in per- student operating expenditures of an increase in enrollment from 500 to 1,000?

8. Based on the results of this study, what can we conclude about the existence of economies or diseconomies in operating a public high school?

CHARTER AIRLINE OPERATING DECISIONS Firm-specific demand in the scheduled airline industry is segmented by customer class and is highly uncertain so that an order may not lead to realized revenue and a unit sale. Airlines respond to this dynamic, highly competitive environment by tracking reservations at preannounced fares and reassigning capacity to the various market segments (“buckets”) as business travelers, vacationers, and convention groups book the flights above or below expected levels several days and even weeks before sched- uled departure. This systems management process combining marketing, operations, and finance is referred to as revenue management or yield management and is dis- cussed in Chapter 14.

The charter airline business, on the other hand, is much less complicated because capacity requirements are known far in advance, and all confirmed orders lead to re- alized revenue. We consider the following three decisions for a charter airline: (1) the entry/exit break-even decision, (2) the operate/shut down decision to fly/not fly a charter that has been proposed, and (3) the output decision as to how many incre- mental seats to sell if the airline decides to operate the charter flight.

Suppose the following costs for a 10-hour round-trip flight apply to the time frame and expenses of an unscheduled 5-hour charter flight from Baltimore to Las Vegas (and return the next day) on a seven-year-old Boeing 737-800 with 120 occupied

Chapter 9: Applications of Cost Theory 331

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

seats.18 Some costs listed in the table have been aggregated up to the flight level from a seat-level decision where they are incurred. Others have been allocated down to the flight level from an entry/exit or maintain ownership company-level decision. Still other costs vary with the go/no go flight-level decision itself. Your job is to analyze each cost item and figure out the “behavior of cost”—that is, with which decision each cost varies.

Fuel and landing fees $5,200

Quarterly airframe maintenance re: FAA certificate 1,000

Unscheduled engine maintenance per 10 flight hours 1,200

Pro rata time depreciation for 7th year of airframe 7,200

Flight pay for pilots per round-trip flight 4,200

Long-term hangar facility lease 6,600

Annual aircraft engine operating lease 7,100

Base salaries of headquarters personnel 2,000

Food service with seat-by-seat purchase and JIT delivery at each departure 2,400

Airport ground crew baggage handling for two flight arrivals 450

Questions 1. What are the variable costs for the decision to send one more person aboard a

charter flight that is already 80 percent booked? 2. In making an entry/exit decision, if competitive pressure is projected to force the

price down to $300, what is the break-even unit sales volume this company should have projected as part of its business plan before entering this market and should reconsider each time it considers leaving (exiting) this business altogether?

3. Identify the indirect fixed costs of the charter service for a particular one of many such charters this month.

4. If one were trying to decide whether to operate (fly) or not fly an unscheduled round-trip charter flight, what would be the total direct fixed costs and variable costs of the flight?

5. Charter contracts are negotiable, and charter carriers receive many contract offers that do not promise $300 prices or 80-percent-full planes. Should the airline accept a charter flight proposal from a group that offers to guarantee the sale of 90 seats at $250? Why or why not?

6. What are the total contributions of the charter flight with 90 seats at $250 per seat?

7. What are the net income losses for this two-day period if the airline refuses the 90-seat charter, stays in business, but temporarily shuts down? What are the net income losses if it decides to operate and fly the charter that has been proposed?

8. What is the segment-level contribution of a separate group that is willing to join the 90-seat-at-$250-per-seat charter on the same plane and same departure, but only wishes to pay $50 per seat for 10 seats?

9. Should you accept their offer? What problems do you anticipate if both charter groups are placed on the 737?

18The aerodynamics of the plane and its fuel efficiency do change as the number of seats occupied falls below 180, but you may ignore this effect.

332 Part 3: Production and Cost

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PART4 PRICING AND OUTPUT DECISIONS:

STRATEGY AND TACTICS

I n the previous chapters, we developed the theories and modeling techniques useful in ana-lyzing demand, production, and cost relationships in a firm. In this part of the book, we con-sider the profit-maximizing price-output decisions, especially as they relate to the firm’s strategic choices in competitive markets (Chapter 10). Asymmetric information conditions in a so-called lemons market as well as ideal full information exchanges are discussed. Chapters 11 and 12 consider price and output determination in dominant-firm monopoly and oligopoly mar- kets. Chapter 13 presents a game-theory framework for analyzing rival response tactics.

The final chapter in Part 4, Chapter 14, examines value-based (not cost-based) differential pric- ing in theory and practice, and Appendix 14A presents the concept of revenue management. Web Appendix E addresses specialized pricing problems including pricing for the multiproduct firm, pricing of joint products, and transfer pricing.

ECONOMIC ANALYSIS AND DECISIONS

1. Demand Analysis and Forecasting

2. Production and Cost Analysis 3. Pricing Analysis 4. Capital Expenditure Analysis

ECONOMIC, POLITICAL, AND SOCIAL ENVIRONMENT

1. Business Conditions (Trends, Cycles, and Seasonal Effects)

2. Factor Market Conditions (Capital, Labor, Land, and Raw Materials)

3. Competitors’ Responses 4. External, Legal, and Regula-

tory Constraints 5. Organizational (Internal)

Constraints

Cash Flows Risk

Firm Value (Shareholders’ Wealth)

333

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

10 CHAP T E R

Prices,Output,andStrategy:Pure andMonopolistic Competition CHAPTER PREVIEW Stockholder wealth-maximizing managers seek a pricing and output strategy that will maximize the present value of the future profit stream to the firm. The determination of the wealth-maximizing strategy depends on the production capacity, cost levels, demand characteristics, and the potential for immediate and longer-term competition. In this chapter, we provide an introduction to competitive strategic analysis and discuss Michael Porter’s Five Forces strategic framework. Thereafter, we distinguish pure competition with detailed analyses of the home contractor industry and monopolistic competition with detailed analyses of advertising expenditures in ready-to-eat cereals. In a “lemons market,” the implications of asymmetrically informed sellers, the rational hesitation of buyers to pay full price, and the resulting problem of adverse selection are also discussed.

MANAGERIAL CHALLENGE Resurrecting Apple1

Apple Computer revolutionized personal computing operating systems by introducing a graphical user inter- face (GUI) with their Macintosh in 1983. The GUI was reverse engineered and quickly imitated by Microsoft whose PC (personal computer) operating system Win- dows captured a 92 percent market share by 1997. Windows-equipped IBM and then Compaq, Dell, and Hewlett-Packard (HP) came to dominate the PC busi- ness. Apple retained PC market leadership only in the education, graphics design, and publishing sectors. Be- cause 55 percent of all PC and operating systems sales are in corporations, 33 percent are in the home, 7 per- cent in government, and only 5 percent are in educa- tion, Apple’s market share of U.S. PC sales slipped from 9.4 percent in 1993 to 2.6 percent in 1997.

Today, the assembly of personal computers is outsourced to a wide variety of Chinese and Taiwan supply chain partners operating at massive scale. Dell

Computers and HP, for example, assemble overseas whatever components the buyer wants, and they then deliver “direct to the customer” through FedEx hubs the next day. With few outsourced components, high overhead, and extensive R&D costs, Apple’s least expen- sive product offering is $1,700, whereas “comparable” HP machines sell at $1,100, and Dell’s PCs are as low as $600.

Apple initially sold primarily through retail outlets like Computertree, but to target the consumer sector, Apple has recently launched dozens of company- owned Apple stores. In addition, Apple adopted a closed (proprietary, unlicensable) operating system architec- ture. This approach sacrifices the huge installed base of Microsoft customers who attract independent soft- ware vendors to write Windows applications programs. Without compatibility to this Wintel-installed base, Apple’s offering stagnated.

334

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

INTRODUCTION To remain competitive, many companies today commit themselves to continuous im- provement processes and episodes of strategic planning. Competitive strategic analysis provides a framework for thinking proactively about threats to a firm’s business model, about new business opportunities, and about the future reconfigurations of the firm’s re- sources, capabilities, and core competencies.

Figure 10.1 displays the components of a business model in the context of a firm’s prerequisite knowledge and strategic decisions. All successful business models begin by identifying target markets—that is, what businesses one wants to enter and stay in. Phys- ical assets, human resources, and intellectual property (like patents and licenses) some- times limit the firm’s capabilities, but business models are as unbounded as the ingenuity of entrepreneurial managers in finding ways to identify new opportunities. Next, all suc- cessful business models lay out a value proposition grounded in customer expectations of perceived value and then identify what part of the value chain leading to end products the firm plans to create. Business models always must clarify how and when revenue will be

In 1999, Steve Jobs regained the leadership of Apple, intent on restoring the brand image of the once highly innovative company. The introduction of the spiffy iMac PC made a good start, allowing Apple’s market share in the U.S. personal computer market to climb back to 9.6 percent in 2010. Jobs also oversaw Apple’s effort to reinvent itself by introducing the iPod digital music player. This time, Apple was ready with layer upon layer of enhanced capabilities for each new genera- tion of its resurgent products. The competitive advantages of the iPod are process-based, rather than product-based, and rely upon cumulative capabilities with its iTunes Music Store and in partnerships with Disney Inc. and record labels. Apple maintains a 73 percent share of the $9 billion digital music industry.

Discussion Questions

� What prices to charge for iMacs and iPods re- mains a central issue for Apple management.

� Have you visited an Apple Store? Did the in- store experience enhance your perceived value for an Apple product?

� On what basis would you justify paying a price premium for an Apple laptop? What about an Apple iPod?

1Based on Apple Computer 1992, 1995 (A), 1996, and 1997, Harvard Busi- ness School Publishing; “The Road Ahead,” Wall Street Journal (June 28, 2000), p. A3; “Is Apple Losing Its Sheen?” Wall Street Journal, (June 28, 2004), p. B1; “Just What Apple Needs: Intel,” BusinessWeek (January 9, 2006), p. 1; “The Best Performers, BusinessWeek (March 23, 2006), p. 1; and Apple Inc., 2008, Harvard Business School Case Publishing.

1993 2000

Apple’s share of PC shipments

2005 2010�95

10%

2

4

6

8

0

core competencies Technology-based expertise or knowledge on which a company can focus its strategy.

MANAGERIAL CHALLENGE Continued

© Ju no ph ot o/ fS to p/ Ge tty

Im ag es

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 335

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

realized and analyze the sensitivity of gross and net margins to various possible changes in the firm’s cost structure. In specifying the required investments, business models also assess the potential for creating value in network relationships with complementary busi- nesses and in joint ventures and alliances. Finally, all successful business models develop a competitive strategy.

COMPETITIVE STRATEGY The essence of competitive strategy is threefold: resource-based capabilities, business processes, and adaptive innovation.2 First, competitive strategy analyzes how the firm can secure differential access to key resources like patents or distribution channels. From humble beginnings as an Internet bookseller that contracted out its warehousing and book delivery service, Amazon managed to become the preferred fulfillment agent for Internet sales in general. That is, the book seller Amazon acquired enough regular customers searching for CDs, office products, tools, and toys that companies like Toys “R”Us adopted Amazon as their Internet sales channel. Second, competitive strategy de- signs business processes that are difficult to imitate and capable of creating unique value for the target customers. For example, the high-frequency point-to-point streamlined op- erations processes of Southwest Airlines prove very difficult for hub-and-spoke airlines to imitate, and, as a result, in 2005 Southwest had a market capitalization equal to that of all the major U.S. carriers combined.

Similarly, at one point in their respective corporate histories, both Dell and Compaq had $12 billion in net sales and approximately $1 billion in net income in 1998. But Compaq’s business model required $6 billion in net operating assets (i.e., inventories plus net plant and equipment plus working capital) to earn $1 billion, while Dell’s required only $2 billion. How could Dell produce the same net income with one-third as much plant and equipment, inventories, and working capital as Compaq? The answer is that Dell created a direct-to-the-customer sales process; Dell builds to order with

FIGURE 10.1 The Strategy Process

1. Target market

2. Value proposition

3. Role in value chain

4. Revenue sources

5. Margins defined

6. Network value

7. Investment required

8. Competitive strategy

Components of a Business Model

1. Customers

2. Competitors

3. Market conditions

4. Capital raising

5. Resource availability

6. Socio-political constraints

Prerequisite Knowledge

1. Products

2. Prices

3. Marketing plans

4. Supply chains

5. Distribution channels

6. Projected cash flows to lenders and equity owners

Decisions

Source: Adapted from H. Chesbrough, Open Innovation (Cambridge, MA: Harvard University Press, 2003).

2This section is based onH. Chesbrough,Open Innovation (Boston: Harvard Business School Press, 2003), pp. 73–83.

336 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

subassembly components bought just in time from outside contractors, and it realizes cash from a sale within 48 hours. These value-creating business processes generated 50 percent ($1B/$2B) return on investment at Dell, whereas the comparable ROI at Compaq was just 16 percent ($1B/$6B).5

Finally, competitive strategy provides a road map for sustaining a firm’s profitability, principally through innovation. As industries emerge, evolve, and morph into other product spaces (e.g., think of Polaroid to digital cameras, calculators to spreadsheets, and mobile phones to smart phones), firms must anticipate these changes and plan how they will sustain their positioning in the industry, and ultimately migrate their busi- ness to new industries. IBM, the dominant mainframe leasing company in the 1970s, has reinvented itself twice—first in the 1980s as a PC manufacturer, and a second time in the 1990s and 2000s as a systems solution provider for a “smarter planet.” In contrast, some firms like Xerox or Kodak become entrenched in outdated competitive strategic positions.

Generic Types of Strategies6

Strategic thinking initially focuses on industry analysis—that is, identifying industries in which it would be attractive to do business. Michael Porter’s Five Forces model (dis- cussed later) illustrates this approach. Soon thereafter, however, business strategists want to conduct competitor analysis to learn more about how firms can sustain their rel- ative profitability in a group of related firms. Efforts to answer these questions are often

WHAT WENT RIGHT • WHAT WENT WRONG

Xerox3

Xerox invented the chemical paper copier and thereafter realized phenomenal 15 percent compound growth rates throughout the 1960s and early 1970s. When their initial patents expired, Xerox was ready with a plain paper copier (PPC) that established a first-mover technology advantage, but ultimately the company failed to receive any broad pat- ent extension.4 Xerox’s target market was large corporations and government installations who valued high-quality, high-volume leased machines with an enormous variety of capabilities and full-service maintenance contracts, even though supplies and usage fees were expensive.

Unable to compete on product capabilities, Japanese competitors Canon and Ricoh realized that tremendous market potential lay in smaller businesses where afford- ability per copy was a major value proposition issue. Instal- lation and service were outsourced to highly competitive independent dealer networks, and the smaller-volume copy machine itself was sold at very low initial cost with self-service replacement cartridges being the principal source of profitability.

As with later events at Apple, Xerox insisted on closed architecture software and built all of its copier components in-house rather than pursuing partnerships that could re- duce cost and trigger a larger installed base of machines. Competitors pursued just the opposite open architecture and partnership strategy to achieve network effects and drive down costs.

Between 1975 and 1985, Xerox copier sales doubled from $4 billion to $8.4 billion while those of Canon grew 25-fold from $87 million to $2.2 billion. During this “lost decade,” Xerox’s market share fell to 40 percent worldwide, and Canon and Ricoh both became $2 billion firms in a copier business that Xerox had totally dominated only 15 years earlier. Failure to adapt its once-dominant business model had doomed Xerox to nearly second-rank status.

3Based on Chesbrough, op. cit.; and on C. Bartlett and S. Ghoshal, Trans- national Management (Boston: Irwin-McGraw-Hill, 1995), Case 4–1. 4In fact, because of Xerox’s 93 percent monopoly of the copier industry in 1971, the U.S. Federal Trade Commission forced Xerox in 1975 to license its PPC technology at low royalty rates.

5Return on invested capital is defined as net income divided by net operating assets (i.e., net plant and equip- ment plus inventories plus net accounts receivable). 6This section is based in part on C. De Kluyver and J. Pearce, Strategy: A View from the Top (Upper Saddle River, NJ: Prentice-Hall, 2003).

industry analysis Assessment of the strengths and weaknesses of a set of competitors or line of business.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 337

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

described as strategic positioning. Finally, strategists try to isolate what core competencies any particular firm possesses as a result of its resource-based capabilities in order to iden- tify sustainable competitive advantages vis-à-vis their competitors in a relevant market.

Product Differentiation Strategy Profitability clearly depends on the ability to create sustainable competitive advantages. Any one of three generic types of strategies may suffice. A firm may establish a product differentiation strategy, a lowest-delivered-cost strategy, or an information technology (IT) strategy. Product differentiation strategy usually involves competing on capabili- ties, brand naming, or product endorsements. Xerox in copiers and Kodak in photo pa- per and chemicals for film development compete on product capabilities. Coca-Cola is by far the world’s most widely recognized brand. Marlboro, Gillette, P&G’s Pampers, Nestlé, Nescafe, and Kellogg’s each has nearly 50 percent shares. All of these branded products command a price premium worldwide simply because of the product image and lifestyle associated with their successful branding. Other differentiated products like Air Jordan compete on the basis of celebrity endorsements.

Which of the three generic types of strategies (differentiation, cost savings, or IT) will be most effective for a particular company depends in part on a firm’s choice of compet- itive scope—that is, on the number and type of product lines and market segments, the number of geographic locations, and the network of horizontally and vertically integrated businesses in which the company decides to invest. For example, the most profitable clothing retailer in the United States, Gap, once undertook to expand its competitive scope by opening a new chain of retail clothing stores. Old Navy’s bargain-priced khakis, jeans, and sweaters immediately began cannibalizing sales at its mid-priced parent. Even fashion-conscious teens could see little reason to pay $16.50 for a Gap-emblazoned T-shirt when Old Navy’s branding offered style and a nearly identical product for $12.50. The configuration of a firm’s resource capabilities, its business opportunities rel- ative to its rivals, and a detailed knowledge of its customers intertwine to determine the preferred competitive scope.

Example Rawlings Sporting Goods Waves Off the Swoosh Sign7

Even though $200 million Rawlings competes against heavily branded Nike with annual sales of $14 billion, Rawlings baseball gloves are extremely profitable. The key is that they are used by more than 50 percent of major leaguers, such as St. Louis Cardinal Albert Pujol and Yankees shortstop Derek Jeeter. These super- stars receive $20,000 for licensing their autographs to Rawlings for engraving on Little League gloves. But player after player can talk about a feature of Rawlings’ equipment that keeps them coming back year after year. Rawlings is very attentive to this feedback and will lengthen the webbing or stiffen the fingers on a new model in just a few weeks to please their celebrity endorsers. Quick adaptation to the vagaries of the consumer marketplace is a requisite part of any product differ- entiation strategy.

7Based on “I’ve Got It,” Wall Street Journal (April 1, 2002), p. A1.

sustainable competitive advantages Difficult to imitate features of a company’s processes or products.

product differentiation strategy A business- level strategy that relies upon differences in products or processes affecting perceived customer value.

338 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Cost-Based Strategy Competitive scope decisions are especially pivotal for cost-based strategy. A firm like Southwest Airlines with a focused cost strategy must limit its business plan to focus nar- rowly on point-to-point, medium-distance, nonstop routes.

In contrast, Dell Computers’ cost leadership strategy allows it to address a wide scope of PC product lines at prices that make its competitors wish to exit the market, as IBM did in 1999. Gateway was also unable to keep pace with Dell’s cost-cutting and by 2006 found itself at 5.3 percent market share versus a 10.6 percent peak in 1999.

Information Technology Strategy Finally, firms can seek their sustainable competitive advantage among relevant market rivals by pursuing an information technology strategy. In addition to assisting in the recovery of stolen vehicles, satellite-based GPS has allowed Allstate Insurance to confirm that certain cars on a family policy are not being driven to work, while other less expen- sive cars are being exposed to the driving hazards of commuting. This allows Allstate to cut some insurance rates and win more business from their competitors. The e-commerce strategy of Southland Corporation’s 7-Eleven convenience stores in 6,000 locations across Japan (see the Example on the next page) provide another good example.

In conclusion, a company’s strategy can result in higher profits if the company con- figures its resource-based capabilities, business processes, and adaptive innovations in such as way as to obtain a sustainable competitive advantage. Whether cost-based

Example Think Small to Grow Big: Southwest Airlines Southwest adopted operations processes for ticket sales, boarding, plane turn- around, crew scheduling, flight frequency, maintenance, and jet fuel hedging that deliver exceptionally reduced operating costs to target customers in their price-sensitive market niche. Anything that works against this cost-based strategy must be jettisoned from the business plan. Southwest has clearly accomplished its goal. As air travel plummeted in the months following the September 11, 2001, attacks on the World Trade Center, only Southwest had a break-even that was low enough to continue to make money. Southwest can cover all of its costs at 64 percent load factors (unit sales/seat capacity). American Airlines, United, Delta, and US Airways often operate well below their break-even points of 75 to 84 percent.

Much has been made of the difference in labor cost—that is, that Southwest has labor costs covered by 36 percent of sales dollars while United, American, and US Airways have labor costs covered by 48 percent of sales dollars, but the $0.07 gap between United’s $0.12 cost per revenue passenger mile (rpm) and Jet Blue’s $0.05 and Southwest’s $0.06 cost per rpm reflects process differences. Booz, Allen, Ha- milton found that only 15 percent of the operating cost difference between full- service and low-cost carriers was labor cost. Rather, the largest source of cost difference was process differences in check-in, boarding, reservations, crew sched- uling, and maintenance, These processes make possible the famed 15-minute turn- around time at Southwest.

cost-based strategy A business-level strategy that relies upon low-cost operations, marketing, or distribution.

information technology strategy A business- level strategy that relies on IT capabilities.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 339

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Dell’s Cost Leadership in PC Assembly8

Dell sells over the phone and over the Internet direct to the consumer and then assembles and delivers mass-customized PCs usually within 48 hours. In contrast, Compaq’s large dealer network requires 35 days to convert a sale into realized cash. Even rival mail-order company Gateway takes 16 days. Having no dealer network and realizing cash quickly might be processes to imitate, but Michael Dell pushed the just-in-time approach down his supply chain. Every company that builds criti- cal components for Dell must warehouse within 15 minutes’ travel time of a Dell factory. Consequently, Dell does not even order components until the customer commits to a purchase. Even after developing Internet distribution channels, Com- paq’s slower production process requires that subassembly components sit on the shelves for months. Less inventory at Dell means tying up less working capital, and less working capital means lower cost.

Between 1990 and 1998, Dell drove PC prices down, and as a result even Dell’s gross margin (the difference between net sales revenue and the direct costs of goods sold as a percentage of net sales revenue) also fell, from 33 percent to 23 percent. However, Dell’s net profit margin actually increased from 8 to 11 percent over this period. How could this happen? Dell’s selling, general, and administrative expense (SG&A) declined from 21 percent of sales to 9 percent. Again, less overhead means lower cost, and lower cost can mean higher profitability, even in an era of steeply falling prices.

The overall effect of this cost leadership strategy on market share, profits, and capitalized value has been stunning. Between 1996 and 2001, Dell’s market share in PC shipments grew from 7 percent to 24 percent. Dell’s net income increased ten- fold from $260 million in 1996 to $2.3 billion in 2001. And Dell’s market capitali- zation grew from $6 billion to $70 billion, which was the fastest growing valuation among NYSE-listed companies in several of those years.

8Based on “The New Economy Is Stronger Than You Think,” Harvard Business Review (November/December 1999), pp. 104–105; Chesbrough, op. cit., p. 55; and “How Dell Fine Tunes Its Pricing,”Wall Street Journal (June 8, 2001), p. A1.

Example The E-Commerce of Lunch at 7-Elevens in Japan9

Japanese office workers put in very long hours, often arriving at 8:00 A.M. and stay- ing well into the evening. In the midst of this long day, most take a break to go out on the street and pick up lunch. Boxed lunches, rice balls, and sandwiches are the routine offerings, but the fashion-conscious Japanese want to be seen eating what’s “in.” This situation makes an excellent opportunity for Southland Corporation’s 7-Eleven stores, which is the biggest retailer in Japan and twice as profitable as the country’s second-largest retailer, the clothing outlet Fast Retailing. Half of 7-Eleven’s sales revenue comes from these lunch items. The key to 7-Eleven Japan’s success has been electronic commerce and its information technology strategy.

7-Eleven Japan collects sales information by proprietary satellite communication networks from 8,500 locations three times a day. Like other retailers, 7-Eleven Japan uses the data for merchandising studies to improve its product packaging and

(Continued)

340 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

strategy, product differentiation strategy, or information technology strategy provides the most effective route to competitive advantage depends in large part on the firm’s strate- gic focus. IT-based strategy is especially conducive to broad target market initiatives. In addition to using IT for merchandizing lunch items, 7-Eleven Japan “drives” customer traffic to its convenience stores by allowing Internet buyers to pick up their Web pur- chases and pay at the 7-Eleven counter. Is 7-Eleven Japan a convenience store, an Inter- net fulfillment agent like Amazon, or a warehouse and distribution company? In some sense, 7-Eleven Japan is all of these. Unlike Southwest Airlines’ cost-focused strategy, 7-Eleven Japan has a much broader IT-based strategy that conveys a competitive advan- tage across several relevant markets.

The Relevant Market Concept A relevant market is a group of firms that interact with each other in a buyer-seller rela- tionship. Relevant markets often have both spatial and product characteristics. For exam- ple, the market for Microsoft’s Windows operating system is worldwide, whereas the market for Minneapolis-origin air travel is confined to suppliers in the upper Midwest. Similarly, the market for large, prime-rate commercial loans includes large banks and corporations from all areas of the United States, whereas the market for bagged cement is confined to a 250-mile radius around the plant.

The market structure within these relevant markets varies tremendously. The four largest producers of breakfast cereals control 86 percent of the total U.S. output—a concentrated market. In contrast, the market for brick and concrete block is fragmented—with the larg- est four firms accounting for only 8 percent of the total U.S. output. Recently, the share of the total U.S. output produced by the largest four firms in the women’s hosiery industry has consolidated, growing from 32 percent to 58 percent. These differences in market

shelf placements with laboratory-like experiments in matched-pair stores through- out the country. But there is more, much more. 7-Eleven Japan has built systems to analyze the entire data inflow in just 20 minutes. Specifically, 7-Eleven forecasts what to prepare for the lunch crowd downtown today based on what sells this morning and what sold yesterday evening in suburban locations. As customers be- come more fickle, product fashion cycles in sandwiches are shortening from seven weeks to, in some cases, as little time as 10 days. 7-Eleven Japan forecasts the de- mand daily on an item-by-item, store-by-store basis.

Of course, such short-term demand forecasting would be useless if food prepa- ration were a production-to-stock process with many weeks of lead time required. Instead, supply chain management practices are closely monitored and adapted continuously with electronic commerce tools. Delivery trucks carry bar code read- ers that upload instantaneously to headquarters databases. Orders for a particular sandwich at a particular store are placed before 10:00 A.M., processed through the supply chain to all component input companies in less than seven minutes, and delivered by 4:00 P.M. for the next day’s sales. Most customers praise the extraordi- nary freshness, quality ingredients, and minimal incidence of out-of-stock items. All this competitive advantage over rival grocers and noodle shops has led to con- sistent price premiums for 7-Eleven’s in-house brand.

9Based on “Over the Counter Commerce,” The Economist, (May 26, 2001), pp. 77–78.

relevant market A group of firms belonging to the same strategic group of competitors.

concentrated market A relevant market with a majority of total sales occurring in the largest four firms.

fragmented A relevant market whose market shares are uniformly small.

consolidated A relevant market whose number of firms has declined through acquisition, merger, and buyouts.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 341

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

structures and changes in market structure over time have important implications for the determination of price levels, price stability, and the likelihood of sustained profitability in these relevant markets.

PORTER’S FIVE FORCES STRATEGIC FRAMEWORK Michael Porter10 developed a conceptual framework for identifying the threats to profit- ability from five forces of competition in a relevant market. Figure 10.2 displays Porter’s Five Forces: the threat of substitutes, the threat of entry, the power of buyers, the power of suppliers, and the intensity of rivalry. Today, a sixth force is often added—the threat of a disruptive technology—such as digital file sharing for the recorded music industry or video on-demand over Web-enabled TVs for the video rental industry.

The Threat of Substitutes First, an incumbent’s profitability is determined by the threat of substitutes. Is the product generic, like AAA-grade January wheat, two-bedroom apartments, and office

FIGURE 10.2 Porter’s Five Forces Strategic Model

• Industry concentration • Tactical focus • Switching costs • Exit barriers • Cost fixity or perishability • Industry growth rate • Speed of adjustment

• Unique suppliers • Number of potential suppliers • Supply shortage/surplus • Vertical requirements contracting • Potential for forward integration

• Buyer concentration or volume purchase • Industry overcapacity • Homogeneity of buyers • Potential for backward integration • Outside alternatives • Network effects • Industry standards

Threat of substitute products and services

Threat of new entrants

Bargaining power of suppliers

Bargaining power of buyers

Level of competition in industry

• Value-price gap for functionally related products • Branded vs. generic • Network effects

• High capital requirements • Economies of scale • Absolute cost advantages • High switching costs • Lack of access to distribution channels • Objective product differentiation • Public policy constraints

Substitutes and complements

Potential entrants

Intensity of rivalry

Supplier power

Buyer power

Sustainable industry

profitability

Source: Adapted from M. Porter, Competitive Strategy (Cambridge, MA: The Free Press, 1998).

10Michael Porter, Competitive Strategy (Cambridge, MA: The Free Press, 1998). See also Cynthia Porter and Michael Porter, eds., Strategy: Seeking and Securing Competitive Advantage (Cambridge, MA: Harvard Business School Publishing, 1992).

342 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

supplies, or is it branded, like Gillette razors, Pepsi-Cola, and Campbell’s soup? The more brand loyalty, the less the threat of substitutes and the higher the incumbent’s sus- tainable profitability will be. Also, the more distant the substitutes outside the relevant market, the less price responsive will be demand, and the larger will be the optimal markups and profit margins. As video conferencing equipment improves, the margins in business air travel will decline. A video conferencing projector and sound system now leases for just $279 per month. Similarly, flavored and unflavored bottled water and other noncarbonated beverages such as juice, tea, and sports drinks are growing as much as eight times faster than U.S. soda sales. This trend will tend to erode the loyalty of Pepsi and Coke drinkers. If so, profitability will decline.

Network effects are available to enhance profitability if companies can find complementors—that is, independent firms who enhance the customer value associated with using the primary firm’s product, thereby raising profitability. For example, Micro- soft Windows has obtained such a lock-in on PC customers that independent software providers (ISPs) write highly valued applications for Windows for which Microsoft pays nothing. Similarly, Apple’s iPod attracts ISPs who enhance the customer value, and thereby support the high price point, and alter the positioning of iPod.

The closeness or distance of substitutes often hinges not only on consumer percep- tions created by advertising, but also on segmentation of the customers into separate dis- tribution channels. Pantyhose distributed through convenience stores have few substitutes at 9:00 P.M. the night before a business trip, many fewer than pantyhose sold through department store distribution channels. Consequently, the threat of substitutes is reduced, and the sustainable profit margin on convenience store pantyhose is higher. Similarly, one-stop service and nonstop service in airlines are different products with dif- ferent functionality. United’s one-stop service from Chicago provides a distant substitute for Minneapolis-origin air travelers. Consequently, Northwest Airlines enjoys high mar- gins on nonstop service from Minneapolis.

The Threat of Entry A second force determining the likely profitability of an industry or product line is the threat of potential entrants. The higher the barriers to entry, the more profitable an in- cumbent will be. Barriers to entry can arise from several factors. First, consider high cap- ital costs. The bottling and distribution business in the soft drink industry necessitates a $50 million investment. Although a good business plan with secure collateral will always attract loanable funds, unsecured loans become difficult to finance at this size. Fewer po- tential entrants with the necessary capital implies a lesser threat of entry and higher in- cumbent profitability.

Second, economies of scale and absolute cost advantages can provide another barrier to entry. An absolute cost advantage arises with proprietary IT technology that lowers a company’s cost (e.g., at 7-Eleven Japan). In the traditional cable TV industry, the huge infrastructure cost of laying wire throughout the community deterred multiple entrants. The first mover had a tremendous scale economy in spreading fixed cost across a large customer base. Of course, wireless technology for satellite-based TV may soon lower this barrier, and then numerous suppliers of TV content will exhibit similar unit cost. These new threats of entry imply lower industry profitability in cable TV.

Third, if customers are brand loyal, the costs of inducing a customer to switch to a new entrant’s product may pose a substantial barrier to entry. Year after year, hundreds of millions of dollars of cumulative advertising in the cereal industry maintains the pull- ing power of the Tony the Tiger Frosted Flakes brand. Unadvertised cereals go unno- ticed. To take another example, hotel corporations raise the switching costs for their

complementors Independent firms that enhance the focal firm’s value proposition.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 343

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

regular customers when they issue frequent-stayer giveaways. Committing room capacity to promotional giveaways raises barriers to entry. A new entrant therefore has a higher cost associated with becoming an effective entry threat in these markets.

Access to distribution channels is another potential barrier that has implications for the profitability of incumbents. The shelf space in grocery stores is very limited; all the slots are filled. A new entrant would therefore have to offer huge trade promotions (i.e., free display racks or slot-in allowances) to induce grocery store chains to displace one of their current suppliers. A related barrier to entry has emerged in the satellite tele- vision industry where Direct TV and EchoStar control essentially all the channel slots on satellites capable of reaching the entire U.S. audience. Government regulatory agencies

Example The Relevant Market for Web Browsers: Microsoft’s Internet Explorer11

One of the recurring antitrust policy questions is the definition of the relevant market for computer software. In 1996, Netscape’s user-friendly and pioneering product Navigator had an 82 percent share of the Internet browser market. But during 1996–1999, Microsoft’s Internet Explorer made swift inroads. Bundling Explorer with its widely adopted Windows operating system, Microsoft marketed an integrated software package preinstalled on PCs. Microsoft quoted higher prices for Windows alone than for Windows with Internet Explorer and threatened PC assemblers like Compaq and Gateway with removal of their Windows license unless they mounted Explorer as a desktop icon. Because most PC customers do want Windows preinstalled on their machines, Explorer penetrated deep into the browser market very quickly. By the start of 2000, some estimates showed Explorer’s market share as high as 59 percent.

If the relevant market for these products is an integrated PC operating system (OS), then Microsoft has simply incorporated new Web browser and media player technology into an already dominant Windows OS product. An analogy might be the interlock between an automobile’s ignition and steering system to deter auto theft. If, on the other hand, Internet browsers (or more recently, media players) are a separate relevant market, like stereo equipment for automobiles, then Micro- soft should not be entitled to employ anticompetitive practices like refusals to deal to extend their dominance of PC operating systems into this new software market.

Microsoft’s spectacular growth in sales of Windows 98 was not the issue. Win- ning a near monopoly of 85 percent market share in the previously fragmented OS software industry indicated a superior product, a great business plan, and good management. But allowing Microsoft to extend that market power into a new line of business using tactics that would be ineffective and self-defeating in the absence of the dominant market share in the original business is just what the antitrust laws were intended to prevent. Twenty state attorneys general in the United States and the European antitrust authorities have pursued this line of reasoning. The European Union (EU) insisted on multiple versions of Windows with (and with- out) Media Player stripped out and fined Microsoft $624 million in March 2004. Appeals were exhausted in 2009, and Microsoft paid the fine plus interest.

11Based on “U.S. Sues Microsoft over PC Browser” and “Personal Technology,” Wall Street Journal (October 21 and 30, 1997); “Microsoft’s Browser: A Bundle of Trouble,” The Economist (October 25, 1997); and U.S. News and World Report, Business and Technology (December 15, 1997).

344 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

also can approve or deny access to distribution channels. For example, the Food and Drug Administration (FDA) approves prescription drugs for certain therapeutic uses but not for others. The FDA also approves or denies exceptions to the Orphan Drug Act that gives firms patent-like exclusive selling rights when public policy pressure warrants doing so. Biogen’s highest sales product, Avonex (a weekly injection for multiple sclerosis pa- tients), received a license under the Orphan Drug Act. Other similarly situated firms have been denied approval; such a barrier to entry may prove insurmountable.

Preexisting competitors in related product lines provide a substantial threat of entry as well; see the following example.

Example Potential Entry at Office Depot/Staples12

In 1997, Office Depot (a $6 billion company) and Staples (a $4 billion company) proposed to merge. Their combined sales in the $13 billion office supply superstore industry totaled 76 percent. From another perspective, their potential competitors included not only OfficeMax but all small paper goods specialty stores, department stores, discount stores such as Kmart, warehouse clubs like Sam’s Club, office sup- ply catalogs, and some computer retailers. This much larger office supply industry is very fragmented, easy to enter, and huge ($185 billion). Using this latter stan- dard, the combined market share of Staples and Office Depot was only 6 percent.

The profit margins of Office Depot, OfficeMax, and Staples are significantly higher when only one office supply superstore locates in a town. This would sug- gest that the small-scale office suppliers offer little threat of entry into the super- store market. The exceptional ease of entry (and exit) at a small scale moderates the markups and profit margins of incumbent specialty retailers like stationery stores, but not office supply superstores. High capital requirement and scale econ- omies in warehousing and distribution appear responsible for the barriers to entry in the office supply superstore market.

12Based on “FTC Rejects Staples’ Settlement Offer,” Wall Street Journal (April 7, 1997), p. A3; and J. Baker, “Econo- metric Analysis in FTC v. Staples,” Journal of Public Policy and Marketing, 18 no. 1, (Spring 1999), pp. 11–21.

Example Eli Lilly Poses a Threat of Potential Entry for AstraZeneca13

In 2000, AstraZeneca’s cancer treatment Nolvadex became the first drug ever ap- proved for reducing the risk of breast cancer in currently healthy women. Eli Lilly markets a pharmaceutical product, Evista, long approved by the FDA for the treat- ment of osteoporosis. Preliminary tests have suggested a therapeutic potential for Evista in the prevention of breast cancer. Lilly promptly released an Evista study in which the incidence of developing cancer over a three-year period was reduced 55 percent in 10,575 women with high-risk factors for developing breast cancer. AstraZeneca sued to stop and undoubtedly slowed Lilly’s marketing efforts, but the real barrier to entry would come if the FDA denies the use of Evista for breast cancer treatment. Without such a denial, AstraZeneca’s Nolvadex faces a formida- ble direct competitor from a preexisting supplier in an adjacent relevant market.

13“Zeneca Sues Eli Lilly over Evista Promotion,” Wall Street Journal (February 26, 1999), p. B6.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 345

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Finally, a barrier to entry may be posed by product differentiation. Objective product differentiation is subject to reverse engineering, violations of intellectual property, and offshore imitation even of patented technology like the shutter in a Kodak digital camera. In contrast, subjective perceived product differentiation based on customer perceptions of lifestyle images and product positioning (e.g., Pepsi-Cola) can erect effective barriers to entry that allow incumbent firms to better survive competitive attack. In sum, the higher any of these barriers to entry, the lower the threat of potential entrants and the higher the sustainable industry profitability will be.

The Power of Buyers and Suppliers The profitability of incumbents is determined in part by the bargaining power of buyers and suppliers. Buyers may be highly concentrated, like Boeing and Airbus in the pur- chase of large aircraft engines, or extremely fragmented, like the restaurants that are cus- tomers of wholesale grocery companies. If industry capacity approximately equals or exceeds demand, concentrated buyers can force price concessions that reduce an incum- bent’s profitability. On the other hand, fragmented buyers have little bargaining power unless excess capacity and inventory overhang persist.

Unique suppliers may also reduce industry profitability. The Coca-Cola Co. estab- lishes exclusive franchise arrangements with independent bottlers. No other supplier can provide the secret ingredients in the concentrate syrup. Bottler profitability is there- fore rather low. In contrast, Coke’s own suppliers are numerous; many potential sugar and flavoring manufacturers would like to win the Coca-Cola account, and the syrup in- puts are non-unique commodities. These factors raise the likely profitability of the con- centrate manufacturers because of the lack of power among their suppliers.

Supply shortages, stockouts, and a backorder production environment can alter the relative power of buyers and suppliers in the value chain. One of the few levers a sup- plier has against huge category-killer retailers like Toys“R”Us to prevent their expropri- ating all the net value is to refuse to guarantee on-time delivery for triple orders of popular products. A deeply discounted wholesale price should never receive 100 percent delivery reliability.

Finally, buyers and suppliers will have more bargaining power and reduce firm prof- itability when they possess more outside alternatives and can credibly threaten to verti- cally integrate into the industry. HMOs can negotiate very low fees from primary care physicians precisely because the HMO has so many outside alternatives. Buyers who

Example Objective versus Perceived Product Differentiation: Xerox Shielded from competition by patents on its landmark dry paper copier, Xerox en- joyed a virtual monopoly and 15 percent compound earnings growth through the 1960s and early 1970s. During this period, its research lab in Palo Alto, California, spun off one breakthrough device after another. One year it was the graphical user interface that Apple later brought to market as a user-friendly PC. In 1979, Xerox scientists and engineers developed the Ethernet, the first local area network for connecting computers and printers. Yet, Xerox was able to commercialize almost none of these R&D successes. As a result, Japanese copier companies like Canon and Ikon reverse engineered the Xerox product, imitated its processes, and ulti- mately developed better and cheaper copiers.

346 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

control the setting of industry standards can also negotiate substantial reductions in pricing and profitability from manufacturers who may then be in a position to capture network effects. Companies favored by having their product specs adopted as an industry standard often experience increasing returns to their marketing expenditures.

The Intensity of Rivalrous Tactics In the global economy, few companies can establish and maintain dominance in any- thing beyond niche markets. Reverse engineering of products, imitation of advertising images, and offshore production at low cost imply that General Motors (GM) cannot hope to rid itself of Ford, and Coca-Cola cannot hope truly to defeat Pepsi. Instead, to sustain profitability in such a setting, companies must avoid intense rivalries and elicit passive, more cooperative responses from close competitors. The intensity of the ri- valry in an industry depends on several factors: industry concentration, the tactical fo- cus of competition, switching costs, the presence of exit barriers, the industry growth rate, and the ratio of fixed to total cost (termed the cost fixity) in the typical cost structure.

Exactly what firms and what products offer close substitutes for potential customers in the relevant market determine the degree of industry concentration. One measure of industry concentration is the sum of the market shares of the four largest or eight largest firms in an industry. The larger the market shares and the smaller the number of com- petitors, the more interdependence each firm will perceive, and the less intense the ri- valry. The ready-to-eat cereal industry has more intense rivalry than the soft drink industry, in part because Kellogg’s (37 percent), General Mills (25 percent), Post (15 per- cent), and Quaker Oats (8 percent) together comprise 85 percent of the market. When two firms enjoy 60–90 percent of industry shipments (e.g., Pepsi and Coke), their trans- parent interdependence can lead to reduced intensity of rivalry if the firms tacitly col- lude. Similarly, because Titleist and Spalding dominate the golf ball market, the rivalrous intensity is less than in the fragmented golf club business.

Sustainable profitability is increased by tactics that focus on non-price rather than price competition. Airlines are more profitable when they can avoid price wars and focus their competition for passengers on service quality—for example, delivery reliability, change-order responsiveness, and schedule convenience. But trunk route airlines between major U.S. cities provide generic transportation with nearly identical service quality and departure frequency. Consequently, fare wars are frequent, and the resulting profitability of trunk airline routes is very low. In contrast, long-standing rivals Coca-Cola and Pepsi have never discounted their cola concentrates. This absence of “gain-share discounting” and a diminished focus on price competition tactics in general increases the profitability of the concentrate business. Airlines tried to control gain-share discounting by introduc- ing “frequent flyer” programs to increase the customers’ switching cost from one compet- itor to another. This idea to reduce the intensity of rivalry worked well for a time, until business travelers joined essentially all the rival frequent flyer programs.

Sometimes price versus non-price competition simply reflects the lack of product dif- ferentiation available in commodity-like markets (e.g., in selling cement). However, the incidence of price competition is also determined in part by the cost structure prevalent in the industry. Where fixed costs as a percentage of total costs are high, margins will tend to be larger. If so, firms are tempted to fight tooth and nail for incremental custo- mers because every additional unit sale represents a substantial contribution to covering the fixed costs. All other things being the same, gain-share discounting will therefore tend to increase the greater the fixed cost is. For example, gross margins in the airline industry reflect the enormous fixed costs for aircraft leases and terminal facilities, often

cost fixity A measure of fixed to total cost that is correlated with gross profit margins.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 347

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

reaching 80 percent. Consider the following break-even sales change analysis for an air- line that seeks to increase its total contributions by lowering its prices 10 percent:

ðP0 − MCÞQ0 < ð0:9 P0 − MCÞQ1 < ð0:9 P0 − MCÞðQ0 + ΔQÞ

[10.1]

where revenue minus variable cost (MC times Q) is the total contribution. If discounting is to succeed in raising total contributions, the change in sales ΔQ must be great enough to more than offset the 10 percent decline in revenue per unit sale. Rearranging Equation 10.1 and dividing by P0 yields

ðP0 − MCÞQ0 P0

< P0 − MCÞ

P0 − 0:1

P0 P0

� � ðQ0 + ΔQÞ

ðPCMÞ Q0 < ðPCM − 0:1ÞðQ0 + ΔQÞ

where PCM is the price-cost margin, often referred to as the contribution margin. That is,

Example Price Competition at the Soda Fountain: PepsiCo Inc.14

Soft drinks are marketed through several distribution channels at different prices. The channels of distribution include independent beverage resellers, vending ma- chine companies, and company-owned bottlers supplying supermarkets, conve- nience stores, and vending machines, which account for 31 percent, 12 percent, and 11 percent, respectively, of all soft drink sales. Shelf slots in the store channels are full, and bottlers compete on stocking services and retailer rebates for prime shelf space and vending machine locations in an attempt to grow their brands. With roughly the same percent market shares in the stores, the Coca-Cola- and PepsiCo-owned bottlers attempt to avoid head-to-head price competition, which would simply lower profits for both firms, and instead seek predictable patterns of company-sponsored once-every-other-week discounts. Where independent bever- age resellers have established a practice of persistent gain-share discounting, the Coca-Cola Company and PepsiCo have often attempted to purchase the franchises and replace them with company-owned bottlers. Vending operations are very high- margin businesses, and PepsiCo and Coca-Cola increasingly service vending machines directly from their company-owned bottlers. To date, little price compe- tition has emerged in the vending channel, in part because independents must purchase from exclusive franchise bottlers in their areas.

Price competition is heating up, however, in the fountain drink side of the busi- ness. As more and more families eat more and more meals outside the household, the fountain drink channel accounted for 37 percent of total sales. Coca-Cola has long dominated the fountain drink business. At restaurants and soda shops in 2000, Coke enjoyed a 59 percent share to Pepsi’s 23 percent. Recently, PepsiCo declared an intent to vigorously pursue fountain drink sales through discount pricing tactics if necessary. This development threatens continuing profitability in this important channel of the soft drink industry.

14Based on “Cola Wars Continue,” Harvard Business School Case Publishing (1994); “Pepsi Hopes to Tap Coke’s Fountain Sales,” USA Today (November 6, 1997), p. 3B; and “Antitrust Suit Focuses on Bottlers’ Pricing and Sales Practices,” Wall Street Journal (January 20, 1999), p. B7.

break-even sales change analysis A calculation of the percentage increase in unit sales required to justify a price discount, given the gross margin.

348 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PCM ðPCM − 0:1Þ <

ðQ0 + ΔQÞ Q0

PCM ðPCM − 0:1Þ < 1 +

ΔQ Q0

[10.2]

Using Equation 10.2, an 80 percent price-cost margin implies that a sales increase of only 15 percent is all that one requires to warrant cutting prices by 10 percent. Here’s how one reaches that conclusion:

0:8 ½0:8 − 0:1� < 1 +

ΔQ Q0

1:14 < 1 + ΔQ Q0

1:14 < 1 + 0:15

In contrast, in paperback book publishing, a price-cost margin of 12 percent implies sales must increase by better than 500 percent in order to warrant a 10 percent price cut—that is, 0.12/0.02 < 1 + 5.0+. Because a marketing plan that creates a 15 percent sales increase from a 10 percent price cut is much more feasible than one that creates a 500 percent sales increase from a 10 percent price cut, the airline industry is more likely to focus on pricing competition than the paperback book publishing industry.

Example Contribution Margins at Hanes Discourage Discounting First-quality white cotton T-shirts and briefs have long been the mainstay of the Hanes Corporation. Selling these “blanks” to other companies that perform value- added finishing, dyeing, embroidering, or custom stitching, Hanes captures only the initial stages in the value chain. At a wholesale price of $1.25 and with $0.85 direct cost of goods sold, the gross margin for Hanes briefs of $0.40 must recover the fixed costs plus the distribution-and-selling expenses to earn a profit. With a $0.15 commission per unit sale as a selling expense, the contribution margin in dollars (CM) is $0.25, and the contribution margin percentage (PCM) is $0.25/$1.25 = 20%.

Because of very price-elastic demand, price discounted by as little as 15 per- cent can double unit sales. However, with contribution margins (PCM) as low as 20 percent, the additional sales triggered by the discount are much less at- tractive than one might think. Break-even sales change analysis using Equation 10.2 confirms that a doubling of sales volume is less than the incremental sales change required to restore total contributions to the levels earned before the price cut:

PCM/(PCM – ΔP) = 0.20/(0.20 – 0.15) = 4.0 = 1 + 3.0+

The interpretation here is that unit sales must increase by 300 percent (1 + 300%ΔQ) in order to restore total contributions to their preexisting level. That is, the price reduction must more than quadruple unit sales in order to raise total con- tributions (operating profit). The data displayed in Table 10.1 demonstrate this conclusion in a spreadsheet format.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 349

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Barriers to exit increase the intensity of rivalry in a tight oligopoly. If remote plants specific to a particular line of products (e.g., aluminum smelting plants) are non- redeployable, the tactics will be more aggressive because no competitor can fully recover its sunk cost should margins collapse. In addition to capital equipment, non-redeployable assets can include product-specific display racks (L’eggs); product-specific showrooms (Ethan Allen); and intangible assets that prove difficult to carve up and package for re- sale (unpatented trade secrets and basic research). Trucking companies, on the other hand, own very redeployable assets—that is, trucks and warehouses. If a trucking com- pany attacks its rivals, encounters aggressive retaliation, and then fails and must liquidate its assets, the owners can hope to receive nearly the full value of the economic working life remaining in their trucks and warehouses. As a result, competitive tactics in the trucking industry are not as effective in threatening rivals, so competitive rivalrous inten- sity is lower and profitability is higher.

Finally, industry demand growth can influence the intensity of rivalry. When sales to established customers are increasing and new customers are appearing in the market, ri- val firms are often content to maintain market share and realize high profitability. When demand growth declines, competitive tactics sharpen in many industries, especially if ca- pacity planning has failed to anticipate the decline. Furniture companies discount steeply when housing demand slows. Airline prices and profits declined sharply when demand for air travel leveled off unexpectedly after the Gulf War. Between 1965 and 1975, soft drink consumption in the United States grew by 49 percent. Again, between 1975 and 1985, demand growth was 53 percent. However, from 1985 to 1995, U.S. demand grew by only 24 percent. Sales in the United States flattened out by 1992; annual consumption had reached a plateau of approximately 50 gallons per person (i.e., a gallon per week). Porter’s model predicts that flat soft drink demand would lead to more intense rivalry and lower profitability at PepsiCo Inc. and Coca-Cola Co. Coca-Cola has deflected many competitive initiatives to its fast-growing international division in an attempt to reduce the growing likelihood of intense rivalry with PepsiCo here in the United States.

TABLE 10.1 HANES SALES VOLUME REQUIRED TO MAINTAIN OPERAT-

ING PROFIT WITH A 15 PERCENT PRICE CUT

GIVEN DATA

WITH 15%

PRICE CUT

DOUBLE SALES

VOLUME

TRIPLE SALES

VOLUME

QUA- DRUPLE SALES

VOLUME

Price 1.25 1.0625 2.125 3.1875 4.25

DCGS (VC only) −0.85 −0.85 −1.70 −2.55 −3.40

Commission −0.15 −0.15 −0.30 −0.45 −0.60

CM 0.25 0.0625 0.125 0.1875 0.25

Example Intensity of Rivalry at US Airways15

The Charlotte hub of US Airways is a very concentrated terminal facility; US Air- ways has over 92 percent of the flights. Thus, US Airways’ market share is compa- rable to Microsoft’s dominance of the operating system business with Windows. However, high indirect fixed costs for aircraft leases and facilities imply high mar- gins that make it very tempting for airlines to attract incremental customers

(Continued)

barriers to exit Economic losses resulting from non- redeployable assets or contractual constraints upon business termination.

350 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Finally, the speed of adjustment of rivalrous actions and reactions matters. Recall that if incumbents are slow to respond to tactical initiatives of hit-and-run entrants, then profitability may be driven down to the break-even levels in so-called contestable markets. In contrast, if incumbents are easily provoked and exhibit fast adjustment speeds, then profitability is often more sustainable.

The Myth of Market Share In summary, the key to profitability in many businesses is to design a strategy that re- duces the threat of substitutes, the power of buyers and suppliers, and the threat of entry. Then, firms must adopt tactics and elicit tactical responses from their rivals so that the profit potential in their effective business strategy is not eroded away. This often means forsaking gain-share discounting and other aggressive tactics that would spiral the indus- try into price wars. Price premiums reflecting true customer value are very difficult to win back once buyers have grown accustomed to a pattern of deep discount rivalry be- tween the competitors or predictably-timed clearance sales. Airlines and department store retailers are painfully aware of these tactical mistakes.

More generally, discounting and excessive promotions designed to grab market share are seldom a source of long-term profitability and often result in lower capitalized value. The soft-drink bottler 7-Up doubled and tripled its market share in the late 1970s largely through discounting. But profits declined, and the company was eventually acquired by Cadbury Schweppes. Hon Industries makes twice the return on investment of Steelcase in the office equipment market even though Hon is one-third of Steelcase’s size. Boeing was much more profitable allowing a slight majority of wide-bodied orders to go to government-subsidized Airbus rather than tie up their own assembly-line operations with hundreds of additional orders triggered by the low prices.

After the initial penetration of a new product or new technology into a relevant mar- ket, market share should never become an end in itself. Increasing market share is the means to achieve scale economies and learning-curve-based cost advantages. But addi- tional share points at any cost almost always mean a reduction in profits, not the reverse.

through price discounting. In contrast, Windows is seldom, if ever, discounted. Also, exit barriers are high in airlines but rather low in computer software, where massive sunk-cost expenses for research and development create largely patentable trade secrets that can be easily resold. Finally, industry demand growth is low in airlines but extremely high in computer software. Consequently, in one-stop flights from Charlotte, US Airways is subject to intense rivalry but Microsoft Windows is not.

Frequent price competition, high exit barriers, and flat growth all imply tremen- dous rivalrous intensity in the airline industry and downward competitive pressure on US Airways’ profit margins. The opposite is true in Microsoft’s OS software business. Windows software is seldom discounted and remains extremely profit- able. In short, airlines have industry characteristics that force nearly competitive performance on even dominant firms, whereas a dominant firm in computer oper- ating systems faces less intense rivalry.

15Based on “Mergers, Monopolies, and the Soaring Cost of Flying,” The Margin (March/April 1990), p. 19; and “Flying to Charlotte Is Easy,” Wall Street Journal (June 14, 1995), p. S1.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 351

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

A CONTINUUM OF MARKET STRUCTURES The relationship between individual firms and the relevant market as a whole is referred to as the industry’s market structure and depends upon:

1. The number and relative size of firms in the industry. 2. The similarity of the products sold by the firms of the industry; that is, the degree of

product differentiation. 3. The extent to which decision making by individual firms is independent, not inter-

dependent or collusive. 4. The conditions of entry and exit.

Four specific market structures are often distinguished: pure competition, monopoly, monopolistic competition, and oligopoly. We discuss each in turn.

Pure Monopolistic Oligopoly Monopoly Competition Competition

Pure Competition The pure competition industry model has the following characteristics:

1. A large number of buyers and sellers, each of which buys or sells such a small pro- portion of the total industry output that a single buyer’s or seller’s actions cannot have a perceptible impact on the market price.

2. A homogeneous product produced by each firm; that is, no product differentiation, as with licensed taxi cab services or AAA-grade January wheat.

3. Complete knowledge of all relevant market information by all firms, each of which acts totally independently, such as the 117 home builders of standardized two- bedroom subdivision homes in a large city.

4. Free entry and exit from the market—that is, minimal barriers to entry and exit.

The single firm in a purely competitive industry is, in essence, a price taker. Because the products of each producer are almost perfect substitutes for the products of every other producer, the single firm in pure competition can do nothing but offer its entire output at the going market price. As a result, the individual firm’s demand curve ap- proaches perfect elasticity at the market price. It can sell nothing at a higher price be- cause all buyers will rationally shift to other sellers. If the firm sells at a price slightly below the long-run market price, it will lose money.

For example, Figure 10.3 indicates the nature of the industry and firm demand curves under pure competition in tract home building. Line DD´ represents the total industry or market demand curve for tract houses and S´S is the market supply curve. At price $175,000, the market price, a total of QDI houses will be demanded by the sum of all firms in the industry. Line dd´ represents the demand curve facing each individual firm. The individual firm sells its entire output, QDF , at the market price $175,000. By definition, the quantity QDF represents only a small fraction of the total industry demand of QDI.

Why get involved in industries where revenues per sale ($175,000 in Figure 10.3) are just sufficient to cover fully allocated unit costs of $175,000? The reason is that these sales at a “razor-thin margin” are the ticket to the occasional windfalls when demand increases and price rises enough to generate excess profits (for a few months in the tract home business, a few weeks in the wildcatter oil business, a few hours in the AAA Kan- sas City corn business, or a few minutes in the T-bond resale market). Note that the

pure competition A market structure characterized by a large number of buyers and sellers of a homogeneous (nondifferentiated) product. Entry and exit from the industry is costless, or nearly so. Information is freely available to all market participants, and there is no collusion among firms in the industry.

352 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

timing and magnitude of these windfalls are not predictable. Otherwise the real estate development land, oil leases, and grain silos would rise in value, and the expected excess profit would again reduce to a razor-thin margin above break-even conditions. Also, re- member that at competitive equilibrium the business owner-manager is getting a salary or other return as great as could be received in his or her next best activity. In short, this is not the business environment where venture capital and entrepreneurial returns of 40 percent on invested capital occur regularly, but it does provide perhaps a 12 percent return with good managerial skills and cost controls. More importantly, these razor-thin margins are interrupted occasionally when windfall profits of as much as $25,000 on a tract home, $20 per barrel on crude oil, $1.50 per bushel on corn, or $5,000 per $1 million T-bill erupt for a short time.

Contestable Markets A contestable market is an extreme case of purely competi- tive markets. In this market structure, break-even performance often occurs with just a handful of firms, perhaps only one. The reason is that entry and exit are free and cost- less. Consequently, the mere threat of “hit-and-run” entry is sufficient to drive prices down to the zero-profit, full cost-covering level. Incumbents in such markets are often slower to react than the hit-and-run firms that impose all this competitive pressure. An example is the bond markets where financial arbitrage by hedge funds triggers enormous bets (perhaps tens of billions of dollars) that any government bond or bill prices that have gotten out of line will converge back to their equilibrium levels. Similarly, airlines might seem to be a contestable market; aircraft would seem to be the ultimate mobile asset, but landing slots are not, and incumbents react quickly and aggressively to hit- and-run entrants in these markets.

Monopoly The monopoly model at the other extreme of the market structure spectrum from pure competition is characterized as follows:

1. Only one firm producing some specific product line (in a specified market area), like an exclusive cable TV franchise.

FIGURE 10.3 Pure Competition in the Tract Home Building Industry

Quantity (thousands of houses/year)

QDI

D S

S� D�

QDF

d d�

Quantity (dozens of houses/year)

FirmIndustry

0 0

Pr ic

e

Pr ic

e

$175,000$175,000

ATC

monopoly A market structure characterized by one firm producing a highly differentiated product in a market with significant barriers to entry.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 353

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

2. Low cross-price elasticity of demand between the monopolist’s product and any other product; that is, no close substitute products.

3. No interdependence with other competitors because the firm is a monopolist in its relevant market.

4. Substantial barriers to entry that prevent competition from entering the industry. These barriers may include any of the following: a. Absolute cost advantages of the established firm, resulting from economies in se-

curing inputs or from patented production techniques. b. Product differentiation advantages, resulting from consumer loyalty to established

products. c. Scale economies, which increase the difficulty for potential entrant firms of fi-

nancing an efficient-sized plant or building up a sufficient sales volume to achieve lowest unit costs in such a plant.

d. Large capital requirements, exceeding the financial resources of potential entrants. e. Legal exclusion of potential competitors, as is the case for public utilities, and for

those companies with patents and exclusive licensing arrangements. f. Trade secrets not available to potential competitors.

By definition, the demand curve of the individual monopoly firm is identical with the industry demand curve, because the firm is the entire relevant market. As we will see in Chapter 11, the identity between the firm and industry demand curves allows decision mak- ing for the monopolist to be a relatively simple matter, compared to the complexity of rival- rous tactics with few close competitors in tight oligopoly groups, discussed in Chapter 12.

Monopolistic Competition E. H. Chamberlin and Joan Robinson coined the term monopolistic competition to de- scribe industries with characteristics both of competitive markets (i.e., many firms) and of monopoly (i.e., product differentiation). The market structure of monopolistic compe- tition is characterized as follows:

1. A few dominant firms and a large number of competitive fringe firms. 2. Dominant firms selling products that are differentiated in some manner: real, per-

ceived, or just imagined. 3. Independent decision making by individual firms. 4. Ease of entry and exit from the market as a whole but very substantial barriers to

effective entry among the leading brands.

By far the most important distinguishing characteristic of monopolistic competition is that the outputs of each firm are differentiated in some way from those of every other firm. In other words, the cross-price elasticity of demand between the products of indi- vidual firms is much lower than in purely competitive markets—that is, among tract home builders, oil wildcatters, AAA January wheat suppliers, or T-bill resellers. Product differentiation may be based on exclusive features (Disney World), trademarks (Nike’s swoosh), trade names (BlackBerry), packaging (L’eggs hosiery), quality (Coach hand- bags), design (Apple iPod), color and style (Swatch watches), or the conditions of sale (Dooney & Bourke). These conditions may include such factors as credit terms, location of the seller, congeniality of sales personnel, after-sale service, warranties, and so on.

Because each firm produces a differentiated product, it is difficult to define an indus- try demand curve in monopolistic competition. Thus, rather than well-defined industries, one tends to get something of a continuum of products. Generally, it is rather easy to identify groups of differentiated products that fall in the same industry, like light beers, after-shave colognes, or perfumes.

monopolistic competition A market structure very much like pure competition, with the major distinction being the existence of a differentiated product.

354 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Oligopoly The oligopoly market structure describes a market having a few closely related firms. The number of firms is so small that actions by an individual firm in the industry with respect to price, output, product style or quality, terms of sale, and so on, have a percep- tible impact on the sales of other firms in the industry. In other words, oligopoly is dis- tinguished by a noticeable degree of interdependence among firms in the industry. The products or services that are produced by oligopolists may be homogeneous—as in the cases of air travel, 40-foot steel I-beams, aluminum, and cement—or they may be differentiated—as in the cases of soft drinks, luxury automobiles, and cruise ships.

Although the degree of product differentiation is an important factor in shaping an oli- gopolist’s demand curve, the degree of interdependence of firms in the industry is of even greater significance. Primarily because of this interdependence, defining a single firm’s de- mand curve is complicated. The relationship between price and output for a single firm is determined not only by consumer preferences, product substitutability, and level of adver- tising, but also by the responses that other competitors may make to a price change by the firm. A full discussion of rival response expectations will be deferred until Chapter 12.

PRICE-OUTPUT DETERMINATION UNDER PURE COMPETITION As discussed before, the individual firm in a purely competitive industry is effectively a price taker because the products of every producer are perfect substitutes for the pro- ducts of every other producer. This leads to the familiar horizontal or perfectly elastic demand curve of the purely competitive firm. Although we rarely find instances where all the conditions for pure competition are met, securities exchanges and the commodity markets approach these conditions. For instance, the individual wheat farmer or T-bill reseller has little choice but to accept the going market price.

Short Run A firm in a purely competitive industry may either make transitory profits (in excess of normal returns to capital and entrepreneurial labor) or operate at a temporary loss in the short run.

In pure competition, the firm must sell at the market price (p1 or p2), and its demand curve is represented by a horizontal line (D1 or D2) at the market price, as shown in Figure 10.4. In the purely competitive case, marginal revenue MR is equal to price P, because the sale of each additional unit increases total revenue by the price of that unit (which remains constant at all levels of output). For instance, if

P = $8/unit

then Total revenue = TR = P · Q

= 8Q

Marginal revenue is defined as the change in total revenue resulting from the sale of one additional unit, or the derivative of total revenue with respect to Q:

MR = dTR dQ

= $8=unit

and marginal revenue equals price.

oligopoly A market structure in which the number of firms is so small that the actions of any one firm are likely to have noticeable impacts on the performance of other firms in the industry.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 355

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The profit-maximizing firm will produce at that level of output where marginal reve- nue equals marginal cost. Beyond that point, the production and sale of one additional unit would add more to total cost than to total revenue (MC > MR), and hence total profit (TR – TC) would decline. Up to the point where MC = MR, the production and sale of one more unit would increase total revenue more than total cost (MR > MC), and total profit would increase as an additional unit is produced and sold. Producing at the point where marginal revenue MR equals marginal cost MC is equivalent to maximizing the total profit function.16

The individual firm’s supply function in Figure 10.4 is equal to that portion of the MC curve from point J to point I. At any price level below point J, the firm would shut down because it would not even be covering its variable costs (i.e., P < AVC). Temporary shutdown would result in limiting the losses to fixed costs alone.

Returning to Figure 10.4, if price P = p1, the firm would produce the level of output Q1, where MC = MR (profits are maximized or losses minimized). In this case the firm would incur a loss per unit equal to the difference between average total cost ATC and average revenue or price. This is represented by the height BA in Figure 10.4. The total loss incurred by the firm at Q1 level of output and price p1 equals the rectangle p1CBA. This may be conceptually thought of as the loss per unit (BA) times the number of units

FIGURE 10.4 The Firm in Pure Competition: The Short Run

p2

Q1

Output Q (units)

ATC1MC

D1 = AR1 = MR1

Q2

I

F

EB A

HJ

G

C p1

AVC1

D2 = AR2 = MR2

0

Pr ic

e an

d co

st (

$/ un

it )

16This can be proven as follows:

π = TR − TC

dπ dQ

= dTR dQ

− dTC dQ

= MR − MC = 0

or MR = MC when profits are maximized.

Check for profit maximization by taking the second derivative of π with respect to Q, or d2π dQ2

. If it is less than zero, then π is maximized.

356 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

produced and sold (Q1). At price p1 losses are minimized, because average variable costs AVC have been covered and a contribution remains to cover part of the fixed costs (AH per unit times Q1 units). If the firm did not produce, it would incur losses equal to the entire amount of fixed costs (BH per unit times Q1 units). Hence we may conclude that in the short run a firm will produce and sell at that level of out- put where MR = MC, as long as the variable costs of production are being covered (P > AVC).

If price were p2, the firm would produce Q2 units and make a profit per unit of EF, or a total profit represented by the rectangle FEGp2. The supply curve of the competitive firm is therefore often identified as the marginal cost schedule above minimum AVC. Industry supply is the horizontal summation of these firm supply curves.

Example Profit Maximization under Pure Competition (Short Run): Adobe Corporation This example illustrates the profit-maximization conditions for a firm operating in a purely competitive market environment in the short run. Assume Adobe Corpo- ration faces the following total revenue and total cost functions:

Total revenue: TR = 8Q Total cost: TC = Q2 + 4Q + 2

Marginal revenue and marginal cost are defined as the first derivative of total revenue and total cost, or

Marginal revenue: MR = dTR dQ

= $8=unit

Marginal cost: MC = dTR dQ

= 2Q + 4

Total profit equals total revenue minus total cost:

Total profit: ðπÞ = TR − TC = 8Q − ðQ2 + 4Q + 2Þ = −Q2 + 4Q − 2

To maximize total profit, we take the derivative of π with respect to quantity, set it equal to zero, and solve for the profit-maximizing level of Q. (It is also necessary to check the second derivative to be certain we have found a maximum, not a min- imum!)17

dπ dQ

= −2Q + 4 = 0

Q* = 2 units

Because MR = $8/unit and MC = 2Q + 4 = [2(2) + 4] = $8/unit, when total profit is maximized, note that we are merely setting MC = MR.

17The check for profit maximization goes as follows:

d2π dQ2

= −2

Because the second derivative is negative, we know we have found a maximum value for the profit function.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 357

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Long Run In the long run, all inputs are free to vary. Hence, no conceptual distinction exists be- tween fixed and variable costs. Under long-run conditions in purely competitive markets, average cost will tend to be just equal to price, and all excessive profits will be eliminated (see Point A where p1 = AC1 in Figure 10.6). If not, and if, for example, a price above p1 exceeds average total costs, like P1́ generating temporary quasi-profits, then more firms will enter, the industry supply will increase (as illustrated by the parallel shift outward to the right of the ∑SRSFIRM along market demand D

2 MKT in Figure 10.6), and market price

will again be driven down toward the equilibrium, zero-profit level p1. In addition, as more firms bid for the available factors of production (say, skilled la-

bor or natural resources like crude oil), the cost of these factors will tend to rise. In that event, the entire cost structure of MC1 and AC1 will rise to reflect the higher input costs along an upward-sloping input supply schedule like that for crude oil in Figure 10.5(b). This higher input cost results in a shift up of the firm’s cost structure to AC2

Example Gasoline Price Rises to Record Levels Reflecting a Spike in Crude Oil Input Costs18

Throughout 2006, 2007, and early 2008, the price of gasoline in the United States galloped upward to reach $4 per gallon. Why did it happen? Competitive pressure at retail prevents gas station gouging of retail customers. Excise taxes average only $0.40 across the United States and have been largely unchanged for two decades. Refinery and pipeline bottlenecks were partially to blame after Gulf Coast hurri- canes Katrina and Rita. But the principal source of the run up in gasoline prices was a spectacular increase in crude oil input prices.

Figure 10.5, Panel (a), shows that six times in the past 30 years, crude oil prices have risen steeply. In each prior case, supply disruptions due to armed conflicts in the Middle East or cartel restrictions of output were responsible. In 1973 and 1999–2000, the OPEC I and OPEC III oil cartels successfully enforced reduced out- put quotas on members, thereby restraining supply and driving the market price of crude oil higher. In 1978, 1980, and 1990, three military conflicts massively re- stricted the supply of crude oil leaving the Persian Gulf. In 2004–2008, however, not supply but demand factors are involved. Demand growth in India, China, and the United States in 2004–2008 drove scarce input prices right up the rising mar- ginal cost schedule for crude oil supply exhibited in Figure 10.5, Panel (b).

Oil in the Persian Gulf region is cheapest to find, develop, and extract at a unit cost of $3 per barrel. In contrast, Venezuelan and Russian oil breaks even at $9 per barrel, West Texas oil at $13 per barrel, and the North Sea fields necessitate off- shore rigs and expensive extraction technology that generate $20-per-barrel average total cost. The full delivered costs of Alaskan North Slope oil runs $30 per barrel. These oil field production firms and their associated output trace out a traditional upward-sloping long-run supply curve (here a step function) for the crude oil industry—again see Figure 10.5, Panel (b).

By mid-year 2006, at crude oil prices between $70 and $80, Missouri and Iowa farmers were joining co-ops created to build and operate $65 million corn-fed ethanol plants. The Brazilians have been hugely successful with sugar cane-fed eth- anol plants—so successful that in 2008 Brazil declared energy independence from foreign oil.

18Based on “Special Report: The Oil Industry,” The Economist (April 22, 2006), pp. 55–73.

358 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

(see Figure 10.6) and imposes a two-way squeeze on excess profit. Such a scenario is re- ferred to as an external diseconomy of scale. External scale diseconomies are distin- guished from internal scale economies and diseconomies in that the latter reflect unit cost changes as the rate of output increases, assuming no change in input prices, whereas the former reflect the bidding up of input prices as the industry expands in response to an increase in market demand.

FIGURE 10.5 Crude Oil Prices and Costs

OPECI

Arab Oil

Embargo

OPEC II

OPEC III First Gulf

War

Iran-Iraq War

Iranian Revolution

Nominal Price Real Price (2009 dollars)

Quantity (million barrels per day)

Persian Gulf

U ni

t co

st (

20 06

U SD

) Pr

ic e

(2 00

6 U

SD )

$60

$50

$40

$30

$20

$13

$3

$9

D1

D2

D3

Shale Oil

U.S. Ethanol

LRSOIL

Venezuela/Nigeria/Mexico/Russia

Brazilian Ethanol/Canadian Tar Sands

West Texas/Canada

Panel (a) The Price of Crude Oil, Nominal and Inflation-Adjusted

Panel (b) The Rising Marginal Cost of Crude Oil and Close Substitutes

Alaska

North Sea

1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 $0

$10

$20

$30

$40

$50

$60

$70

$80

$90

external diseconomy of scale An increase in unit costs reflecting higher input prices.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 359

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Under a constant input price assumption, the long-run industry supply curve LRSIND in Figure 10.6 would be flat, a so-called constant-cost industry like timber harvesting. However, with the rising input prices for crude oil depicted in Figure 10.5, Panel (b), the long-run supply curve LRSIND for the downstream final product gasoline rises to the right, signifying an increasing-cost industry, as depicted in Figure 10.6. (It is quite possible to have downward-sloping long-run supply curves. A decreasing-cost industry occurred in the 1980s in calculators and again in the 1990s in PCs because computer chip inputs be- came less expensive as the personal computer market expanded, as shown in Figure 10.7.)

The net result is that in the long-run equilibrium, all purely competitive firms will tend to have identical costs, and prices will tend to equal average total costs (i.e., the av- erage total cost curve AC will be tangent to the horizontal price line p2). Thus, we may say that at the long-run profit-maximizing level of output under pure competition, equi- librium will be achieved at a point where P = MR = MC = AC. In long-run equilibrium, each competitive firm is producing at its most efficient (that is, its lowest unit cost) level of output and just breaking even.

FIGURE 10.6 Long-Run Equilibrium under Pure Competition (in an Increasing Cost Industry)

p2

Q2Q1

B

B�

A� A

Output Q (units)

MCSRSFIRM

LRSIND

ΣSRSFIRM

AC2

AC1

D1MKT

D2MKT

D = MR = AR

0

Pr ic

e an

d co

st (

$/ un

it )

Panel (b)Panel (a)

Industry LevelFirm Level

QUASI- PROFITS

p1

p1′

Q1′

Example Copper Price Rise by 400 Percent Contributes to Housing Bubble Home prices across the United States rose to unsustainable heights in 2006–2008. Part of the reason was demand-pull bid price pressure from lower interest rates on mortgages than ever seen in post-war U.S. markets. But another reason was cost- push asking price pressure from spiking commodity prices. A 2,100-square-foot home incorporates 440 pounds of copper plumbing, sheathing, and wiring. Be- tween 2003 and 2007, copper rose in price 400 percent.

360 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PRICE-OUTPUT DETERMINATION UNDER MONOPOLISTIC COMPETITION Monopolistic competition is a market structure with a relatively large number of firms, each selling a product that is differentiated in some manner from the products of its fringe competitors, and with substantial barriers to entry into the group of leading firms.

Product differentiation may be based on special product characteristics, trademarks, packaging, quality perceptions, distinctive product design, or conditions surrounding the sale, such as location of the seller, warranties, and credit terms. The demand curve for any one firm is expected to have a negative slope and be extremely elastic because of the large number of close substitutes. The firm in monopolistic competition has some limited discretion over price (as distinguished from the firm in pure competition) be- cause of customer loyalties arising from real or perceived product differences. Profit

FIGURE 10.7 The Computer Price Index and U.S. Final Sales of Personal Computers

0

0

100

100

200

300

400

200

300

400

500

600

700

0

20

40

60

80

100

120

160

140

Pr ic

e in

de x,

1 99

6 =

10 0

Pr ic

e in

de x,

1 99

6 =

10 0

B ill

io ns

o f

do lla

rs

Total sales (right axis)

Sales to business

(right axis)

1987

50 60 70 80 90 100 110 120

Industry output

1988 1989 1990 1991 1992 1993 1994 1995 1997 1998 1999 2000 2001

LRSIND

Price index (left axis)

1996

Computer Supply Curve in the Long-Run

Source: St. Louis Federal Reserve Bank, National Economic Trends (May 2001).

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 361

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

maximization (or loss minimization) again occurs when the firm produces at that level of output and charges that price at which marginal revenue equals marginal cost.

Short Run Just as in the case of pure competition, a monopolistically competitive firm may or may not generate a profit in the short run. For example, consider a demand curve such as D´D´ in Figure 10.8, with marginal revenue equal to MR´. Such a firm will set its prices where MR´ = MC, resulting in price P3 and output Q3. The firm will earn a profit of EC dollars per unit of output. However, the low barriers to entry in a monopolistically com- petitive industry will not permit these short-run profits to be earned for long. As new firms enter the industry, industry supply will increase, causing the equilibrium price to fall. This is reflected in a downward movement in the demand curve facing any individual firm.

Long Run With relatively free entry and exit into the competitive fringe, average costs and a firm’s demand function will be driven toward tangency at a point such as A in Figure 10.8. At this price, P1, and output, Q1, marginal cost is equal to marginal revenue. Hence a firm selling perfume or beer is producing at its optimal level of output. Any price lower or higher than P1 will result in a loss to the firm because average costs will exceed price.

Because the monopolistic competitor produces at a level of output where average costs are still declining (between Points A and B in Figure 10.8), monopolistically com- petitive firms produce with “excess” capacity. Of course, this argument overlooks the

WHAT WENT RIGHT • WHAT WENT WRONG

The Dynamics of Competition at Amazon.com19

On-line retailing started very slowly in clothing and other search goods that buyers want to “touch and feel,” but it has excelled in one experience good—namely, books. Amazon stocks less than 1,000 bestsellers but displays and provides reviews on 2.5 million popular titles. Using Ingram Book Group, the world’s largest book wholesaler, Amazon is able to ship most selections in one to three days. Sales doubled each half year and in 2004 topped $2 billion. Nevertheless, Amazon.com shares declined in value.

One difficulty for Amazon.com is that Internet retailing is a classic example of a business with low barriers to entry and exit. As soon as Amazon’s business systems for dis- play, order taking, shipping, and payments stabilized, since profits were present, one expected substantial entry activ- ity. For example, Barnes and Noble entered into an exclu- sive contract with America Online to pitch electronic book sales to AOL’s 8.5 million subscribers. Borders then quickly announced plans to enter electronic retailing. And many specialist booksellers of Civil War books, jet plane books, history books, auto books, and so forth have flooded onto the Internet search engines. Even Amazon’s wholesale supplier Ingram Book Group has entered the

fray; for $2,500, Ingram support services will set up a Web site on behalf of any new book retailer.

Amazon.com responded by offering customized notifi- cation and book discussion services to add value for readers with special interests. The information revolution has made relationship marketing to established customers a pivotal element in securing repeat purchases. Nevertheless, the nu- merous open opportunities for fast, easy, and cheap entry likely will erode the profits in electronic book retailing. A competitive rate of return on time, talent, and investment in online retailing might today amount to only 7 percent.

The imperfect consumer information, limited time for comparison shopping, and brand loyalty that retailers have depended upon are disappearing with Internet search en- gines, and retailing’s traditionally slim profit margins are quickly becoming hairline thin or nonexistent. Online re- tailing may increasingly perform like tract home building— that is, purely competitively.

19Based on “Web Browsing,” The Economist (March 29, 1997), p. 71; “In Search of the Perfect Market: A Survey of Electronic Commerce,” The Economist (May 10, 1997); “The Net: A Market Too Perfect for Profits,” BusinessWeek (May 11, 1998), p. 20; “Comparison Shopping Is the Web’s Virtue—Unless You’re a Seller,” Wall Street Journal (July 23, 1998), p. A1; and Value Line, Ratings and Reports, various issues.

362 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

extent to which idle capacity may be a source of product differentiation. Idle capacity means a firm such as Blockbuster can operate with high delivery reliability and change order responsiveness, which can be very important to renters of popular films and that warrants a price premium relative to competitive fringe airlines.

SELLING AND PROMOTIONAL EXPENSES In addition to varying price and quality characteristics of their products, firms may also vary the amount of their advertising and other promotional expenses in their search for profits. This kind of promotional activity generates two distinct types of benefits. First, demand for the general product group may be shifted upward to the right as a result of the individual firm and industry advertising activities. The greater the number of firms in an industry, the more diffused will be the effects of a general demand-increasing adver- tising campaign by any one firm. In contrast, a monopolist such as an electric utility, or a highly concentrated oligopoly such as computer operating systems, will be more in- clined to undertake an advertising campaign.

The second, more widespread incentive for advertising is the desire to shift the de- mand function of a particular firm at the expense of other firms offering similar prod- ucts. This strategy will be pursued both by oligopolists like Philip Morris and General Mills and by firms in more monopolistically competitive industries like Anheuser- Busch, Miller, and Coors.

Determining the Optimal Level of Selling and Promotional Outlays Selling and promotional expenses, often collectively referred to as advertising, are one of the most important tools of non-price competition.

To illustrate the effects of advertising expenditures and to determine the optimal sell- ing expenses of a firm, consider the case where price and product characteristics already have been determined, and all retailers are selling at the manufacturer’s suggested retail price.

FIGURE 10.8 Long-Run Equilibrium in Monopolistic Competition

P3

Q3

Output Q (units)

D�

MC

Q2

B A

E

C

AC

Q1

D

MR�

MR

D�

D

0

P1, C1 P2, C2

Pr ic

e an

d co

st (

$/ un

it )

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 363

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Long-Run Price and Output Determination: Blockbuster, Inc. The market for DVD rentals in Charlotte, North Carolina, can best be described as monopolistically competitive. The demand for DVD rentals is estimated to be

P = 10 – 0.004Q

where Q is the number of weekly DVD rentals. The long-run average cost function for Blockbuster is estimated to be

LRAC = 8 – 0.006Q + 0.000002Q2

Blockbuster’s managers want to know the profit-maximizing price and output le- vels, and the level of expected total profits at these price and output levels.

First, compute total revenue (TR) as

TR = P · Q = 10Q – 0.004Q2

Next, compute marginal revenue (MR) by taking the first derivative of TR:

MR = dTR dQ

= 10 − 0:008Q

Compute total cost (TC) by multiplying LRAC by Q:

TC = LRAC · Q = 8Q – 0.006Q2 + 0.000002Q3

Compute marginal cost (MC) by taking the first derivative of TC:

MC = dTC dQ

= 8 − 0:012Q + 0:000006Q2

Next, set MR = MC

10 − 0:008Q = 8 − 0:012Q + 0:000006Q2

0:000006Q2 − 0:004Q − 2 = 0

Use the quadratic formula to solve for Q. Q* is equal to 1,000.20 At this quan- tity, price is equal to

P* = 10 − 0:004 ð1,000Þ = 10 − 4 = $6

Total profit is equal to the difference between TR and TC, or

π = TR − TC

= 10Q − 0:004Q2 − ½8Q − 0:006Q2 + 0:000002Q3� = 10ð1,000Þ − 0:004ð1,000Þ2 − ½8ð1,000Þ − 0:006ð1,000Þ2 + 0:000002ð1,000Þ3� = $2,000

The MR and MC at these price and output levels are $2. The fact that Blockbuster expects to earn a profit of $2,000 suggests that the

firm can anticipate additional competition, resulting in price cutting that will ulti- mately eliminate this profit amount.21

(Continued)

364 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The determination of the optimal advertising outlay is a straightforward application of the marginal decision-making rules followed by profit-maximizing firms. Define MR to be the change in total revenue received from a one-unit increase in output (and the sale of that output). For fixed-price settings, MR just equals the price, P. Define MC to be the change in total costs of producing and distributing (but not of advertising) an addi- tional unit of output. The marginal profit or contribution margin from an additional unit of output is (from Chapter 9):

Contribution Margin (PCM) = P – MC [10.3]

The marginal cost of advertising (MCA) associated with the sale of an additional unit of output is defined as the change in advertising expenditures (ΔAk) where k is the unit cost of an advertising message, A, or

MCA = ΔAk ΔQ

[10.4]

The optimal level of advertising outlays is the level of advertising where the marginal profit contribution (PCM) is equal to the marginal cost of advertising, or

PCM = MCA [10.5]

As long as a firm receives a greater contribution margin than the MCA it incurs to sell an additional unit of output, the advertising outlay should be made. If pCM is less than MCA, the advertising outlay should not be made and the level of advertising should be reduced until PCM = MCA. This marginal analysis also applies to other types of non- price competition like after-sale service and product replacement guarantees.

20The solution of the quadratic formula, aQ2 + bQ + c = 0, is

Q = −b ±

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi b2 − 4ac

p

2a =

−ð−:004Þ ± ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð−:004Þ2 − 4ð0:000006Þð1 − 2Þ

q 2ð0:000006Þ

= 1,000 or − 333:33

Only the positive solution is feasible. 21Recall that the TC function includes a “normal” level of profit. Hence, this $2,000 represents an economic rent above a normal profit level.

Example Optimal Advertising at Parkway Ford The marginal profit contribution from selling Ford automobiles at Parkway Ford averages $1,000 across the various models it sells. Parkway estimates that it will have to incur $550 of additional promotional expenses per vehicle to increase its sales per day. Should the outlay for promotions be made?

Because PCM > MCA (i.e., $1,000 > $550), Parkway’s operating profit will be increased by $450 if it incurs an additional $550 of promotional expenses. Parkway should continue to make additional promotional outlays (which are likely to be less and less effective at triggering additional sales per day) up to the point where the marginal cost of advertising equals the expected (marginal profit) contribution margin.

If Parkway were then to find that MCA was greater than PCM, they should cut back on promotional outlays until the contribution margin rose enough to again equate PCM = MCA.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 365

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Optimal Advertising Intensity Optimal expenditure on demand-increasing costs like promotions, couponing, direct mail, and media advertising can be compared across firms. For example, the total contri- butions from incremental sales relative to the advertising cost of beer ads can be com- pared to the total contributions relative to the advertising cost of cereal ads. Advertising is often placed in five media (network TV, local TV, radio, newspapers, and magazines). The “reach” of a TV ad is measured as audience thousands per minute of advertising message; reach is directly related to the advertising message’s cost (k). A manager should fully fund in her marketing budget any ad campaign for which

(P – MC) (ΔQ/ΔA) > k [10.6]

where (P – MC) is the contribution margin and (ΔQ/ΔA) is the increase in demand (i.e., a shift outward in demand) attributable to the advertising.22

Expanding Equation 10.6 identifies the two determinants of the optimal advertising expenditure per dollar sales or “advertising intensity.” Ak/PQ is determined by the gross margin (P – MC)/P and by the advertising elasticity of demand Ea:

Ak PQ

= ðP − MCÞ

P A Q ðΔQ=ΔAÞ [10.7]

Ak PQ

= ðP − MCÞ

P Ea [10.8]

Both factors are important. With high margins (near 70 percent) and very effective ads, Kellogg’s spends 30 percent of every dollar of sales revenue on cereal advertising. In con- trast, the jewelry industry has 92 percent margins, the highest of all four-digit industries, but Zales’s advertising inserts in the weekend paper simply do not trigger many jewelry sales. Because the advertising elasticity in jewelry is so low, a company like Zales spends less than 10 percent of its sales revenue on advertising. Campbell’s Soup has relatively high advertising elasticity of demand given its strong brand name, but the margins on

Example Ford and P&G Tie Ad Agency Pay to Sales Historically, ad agencies have earned more income each time their clients buy an- other expensive 30-second slot on network TV (or other media), whatever the per- formance of the ad in generating incremental sales. More recently, Ford and Procter & Gamble, two of the world’s biggest advertisers, announced that hence- forth all agency billings would need to be performance-based. These incentive pay- ment plans include a fixed fee for designing ad campaigns plus incentive pay based on the incremental sales traceable to the ad. The idea is to encourage agencies to search for database marketing, Internet, and event sponsorships that far exceed the marginal media buy in advertising productivity, ΔQ/ΔA.

22Sometimes the price points at which the product can be sold change after a successful ad campaign. If so, the appropriate valuation of the incremental sales in Equation 10.6 is the new contribution margin.

366 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

canned goods are very low (less than 5 percent); consequently, Campbell’s Soup spends just one-tenth of what Kellogg’s spends on advertising as a percentage of sales revenue— just 3 percent of sales revenue.

The Net Value of Advertising Although advertising can raise entry barriers and maintain market power of dominant firms, the economics of information argues that by giving consumers information, adver- tising can reduce the prices paid. The discovery of price information may be costly and time consuming in the absence of price advertising. For example, Benham24 found the price of eyeglasses to be substantially lower in states that permitted price advertising than in those that prohibited such advertising. Also, because advertising creates brand awareness (both for good and inferior brands), advertisers who misrepresent their prod- uct will not be successful in generating repeat business.

Example Optimal Advertising Intensity at Kellogg’s and General Mills23

The ready-to-eat (RTE) cereal industry spends 55 percent of its sales revenue on marketing and promotion—30 percent on advertising alone. In part, this resource allocation decision reflects the fact that cereal demand is very sensitive to successful ad campaigns like Kellogg’s Tony the Tiger or General Mills’ Wheaties, the Break- fast of Champions. In addition, however, RTE cereal margins are among the high- est of any four-digit industry. Kellogg’s Raisin Bran sells for $4.49 and has a direct fixed plus variable manufacturing cost of $1.63. That calculates as a (4.49 – 1.63)/4.49 = 70 percent gross margin. Frosted Flakes’ margin is 72 percent, and Fruit Loops’ margin is 68 percent. These margins reflect brand loyalties built up over many years of advertising investments. In the highly concentrated RTE cereal industry, Quaker Oats (8 percent), Post (15 percent), General Mills (25 percent), and Kellogg’s (37 percent) control 85 percent of the market.

Until recently, advertising and retail displays were the predominant forms of competition in cereals. Like Coca-Cola and PepsiCo, the dominant RTE cereal companies had concluded that price discounting would be mutually ruinous and ultimately ineffective. Therefore, each company decided independently to refrain from discounting prices to attempt to gain market share. However, in June 1996, 20 percent price cuts swept through the industry, in part in response to the growth of private-label cereals (e.g., Kroger Raisin Bran) that had collectively grabbed close to 10 percent of the market. Margins on some leading brand-name products fell from 70s to 50 percent, with ingredients (15 percent), packaging (10 percent), wages (10 percent), and distribution (15 percent) accounting for the rest of the selling price.

23Based on “Cereals,” Winston-Salem Journal (March 8, 1995), p. A1; and “Denial in Battle Creek,” Forbes (October 7, 1996), pp. 44–46.

24Lee Benham, “The Effect of Advertising on the Price of Eyeglasses,” Journal of Law and Economics (October 1972), pp. 337–352.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 367

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

COMPETITIVE MARKETS UNDER ASYMMETRIC INFORMATION In competitive markets for T-shirts, crude oil, auto rentals, and delivered pizza, both buyers and sellers have full knowledge of the capabilities and after-sale performance of the standard products. Equilibrium price just covers the supplier’s cost of production for a product of known reliable quality. If suppliers were to charge more, rival offers and entry would quickly erode their sales. If suppliers were to charge less, they could not afford to stay in business. This has been the message so far of this chapter—in com- petitive markets under ideal information conditions, you get what you pay for. Such markets differ enormously from competitive markets under asymmetric information, which are sometimes called lemons markets. One prominent example of asymmetric in- formation in a lemons market is used automobiles, in which the true quality of mechan- ical repairs, or other features, often is known only to the seller. Other goods sold under asymmetric information include house paint, mail-order computer components, and common cold remedies.

In a lemons market, the buyers discount all unverifiable claims by the sellers, who mar- ket only lower-quality products at the reduced prices buyers are willing to offer. This dis- appearance of higher quality products from the marketplace illustrates the concept of adverse selection—that is, the lower-quality products are selected in and the higher- quality products are adversely selected out. To resolve the marketing problems posed by adverse selection requires credible commitment mechanisms such as warranties, brand- name reputations, collateral, or price premiums for reliable repeat-purchase transactions.

Incomplete versus Asymmetric Information One distinction that can sharpen our understanding of these complicating factors in com- petitive exchange is that between asymmetric information and incomplete information. Incomplete information is associated with uncertainty, and uncertainty is pervasive. Prac- tically all exchanges, whether for products, financial claims, or labor services, are con- ducted under conditions of uncertainty. On the one hand, decision makers often face uncertainty as to the effect of random disturbances on the outcome of their actions. This uncertainty typically leads to insurance markets. On the other hand, decision makers are sometimes uncertain as to the payoffs or even types of choices they face. This condition typically leads to intentionally incomplete contracting.

Asymmetric information exchange, in contrast, refers to situations in which either the buyer or the seller possesses information that the other party cannot verify or to which the other party does not have access. For example, mail-order suppliers of com- puter components or personal sellers of used cars often have an informationally advan- taged position relative to the buyers. The sellers know the machine’s capabilities, deficiencies, and most probable failure rate, but these are difficult matters for the buyer to assess from reading magazine ads or kicking the tires. And the typical 90-day war- ranty does nothing to alter this information asymmetry. Both buyer and seller face un- certainty against which they may choose to insure, but one has more information or better information than the other.

Search Goods versus Experience Goods In services, retailing, and many manufacturing industries, buyers generally search the market to identify low-price suppliers. Sometimes this search is accomplished by asking for recommendations from recent purchasers, by scouring the catalogs and ads, or by visiting showrooms and sales floors. In selecting a supplier, many customers are also

lemons markets Asymmetric information exchange leads to the low-quality products and services driving out the higher- quality products and services.

incomplete information Uncertain knowledge of payoffs, choices, and so forth.

asymmetric information Unequal, dissimilar knowledge.

368 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

intensely interested in multiple dimensions of product and service quality, including product design, durability, image, conformance to specifications, order delay, delivery re- liability, change-order responsiveness, and after-sale service. Customers often spend as much time and effort searching the market for the desired quality mix as they do search- ing for lowest price. Retailers and service providers understand this and often offer many quality combinations at various prices to trigger a purchase of these search goods. Con- sider, for example, the many price-quality alternatives available in clothing, sporting goods, furniture stores, and hotel chains.

On the other hand, some products and services have important quality dimensions that cannot be observed at the point of purchase. Consider, again, used cars and other resale machinery, nonprescription remedies for the common cold, house paint, and mail-order computer components. The quality of these items can be detected only through experience in using the products. Hence, products and services of this type are termed experience goods and are distinguished from search goods.

Ultimately, the problem with experience goods in competitive market exchange is the unverifiability of asymmetric information. The seller knows how to detect the difference between high- and low-quality products (e.g., between lemons and cream puffs in the used-car market), but cannot credibly relay this information to buyers, at least not in chance encounters between strangers. Fraudulent sellers will claim high quality when it is absent, and realizing this, buyers rationally discount all such information. Because of the private, impacted nature of the product quality information, the seller’s claims and omissions can never be verified without experiencing for oneself the reliability of the auto, the efficacy of the common cold remedy, the durability of the house paint, or the capability of the computer component.

All of this is not to say that the buyers of experience goods are without recourse or that the sellers are without ingenuity as to how to market their products. Warranties and investments in reputations provide mechanisms whereby the sellers of house paint and computer components can credibly commit to delivering a high-quality product. The essential point is that in the absence of these bonding or hostage mechanisms, the experience-good buyer will rationally disbelieve the seller’s claims. Consequently, the honest seller of truly high-quality experience goods will find little market for his or her higher-cost, higher-priced product. The “bad apples drive out the good” in many experience-good markets.

Adverse Selection and the Notorious Firm Suppose customers recognize that unverifiable private information about experience- good quality is present, yet knowledge of any fraudulent high-price sale of low-quality products spreads almost instantaneously throughout the marketplace. Is this extreme reputational effect sufficient to restore the exchange of high-quality/high-price experience goods? Or, can the notorious firm continue to defraud customers here and elsewhere? The answer depends on the conditions of entry and exit discussed earlier in this chapter, but not in the way you might expect.

Consider the cost structure and profits of such a notorious firm, depicted in Figure 10.9. If offered the low price Pl, the firm operates in competitive equilibrium at Q1, where the price just covers the marginal cost and average total cost (SAClow quality) for Q1 units of the low-quality product. Alternatively, if offered the high price Ph, the firm can either competi- tively supply Q1 of the high-quality experience good and again just break even against the higher costs of SAChigh quality,

25 or the firm can deliver a low-quality experience good at Q2

search goods Products and services whose quality can be detected through market search.

experience goods Products and services whose quality is undetectable when purchased.

25The minimum cost output for the plant configuration and cost structure associated with high quality could shift right or left, but to simplify, assume that the SAC just increases vertically from Point C to Point A.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 369

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

and continue to incur the lower costs of SAClow quality. The third alternative entails an ex- pansion of output along MClow quality in response to the price rise and generates profits. That is, the incremental output (Q2 – Q1) earns incremental profit equal to the difference between Ph andMClow quality—namely, the shaded area ABC (labeled bold E)—and in addi- tion, the original output Q1 earns a fraudulent rent of area GACF (labeled bold D). Although the supplier observes his own cost directly and therefore detects the availability of D + E, the problem for the experience-good buyer is that in terms of point-of-sale infor- mation, high-price transactions at Point B on MClow quality and at Point A on MChigh quality are indistinguishable. Both types of products have an asking price of Ph, and only the seller observes the output rate Q1 versus Q2.

Of course, the supplier is not indifferent between the two alternatives. The high- quality transaction offers a cash flow from operations just sufficient to cover capital costs and break even at Point A, whereas the fraudulent transaction (a low-quality product at a high price at Point B) offers a net profit for at least one period. Table 10.2 depicts this interaction between experience-good buyers and a potentially fraudulent firm as a payoff matrix. The seller can produce either low or high quality, and the buyer can offer either low or high prices. The row player (the seller) gets the below-diagonal payoffs in each cell, and the column player (the buyer) gets the above-diagonal payoffs in each cell. The buyer prefers to cover the high cost of high-quality products (in the northwest cell) rather than pay less and only cover the lower cost of low-quality products (in the southeast cell). However, the buyer is worst off when the seller fails to deliver a high- quality product for which the buyer has paid a high price (in the southwest cell). The buyer also recognizes that getting more than she pays for (in the northeast cell) would impose losses on the seller who would prefer to break even with a low-price/low-quality transaction in the southeast cell.

Each player in this business game attempts to predict the other’s behavior and re- spond accordingly. Knowing that the seller prefers profits to breaking even at high prices

FIGURE 10.9 Low-Quality Experience Goods Emerge from Competitive Markets

Q2

A

D

P h

G

Q1

P l

F

E

MChigh quality

MClow quality SAChigh quality

Dh = Ph

Dl = Pl

SAClow quality

C

B

Units0

Pr ic

e an

d co

st (

$/ un

it )

370 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

and that the seller prefers breaking even to losses at low prices, the buyer predicts that low-quality product will be forthcoming irrespective of the price offered. Therefore, the buyer makes only low-price offers. Only those who wish to be repeatedly defrauded offer to pay high prices for one-shot transactions with strangers offering experience goods.

This reasoning motivates adverse selection by the rational seller in an experience-good market. Because sellers can anticipate only low-price offers from buyers, the sellers never produce high-quality products—that is, the market for experience goods will be incom- plete in that not all product qualities will be available for sale. Anticipating that buyers will radically discount their unverifiable high-quality “cream puffs,” individual sellers of used cars choose to place only low-quality “lemons” on the market. The “cream puffs” often are given away to relatives. Similarly, jewelers in vacation locations, anticipating that out-of-town buyers will suspect uncertified spectacular gemstones are fakes, choose to sell only lower-quality gemstones. And unbranded mail-order computer components are in- evitably of lower quality. Adverse selection always causes competitive markets with asym- metric information to be incomplete. Again, the bad apples drive out the good.

Insuring and Lending under Asymmetric Information: Another Lemons Market This same adverse selection reasoning applies beyond experience-good product markets whenever asymmetric information is prominent. Consider the transaction between a bank loan officer and a new commercial borrower, or between an insurance company and a new auto insurance policyholder. Through an application and interview process and with access to various databases and credit references, the lender or insurer attempts to uncover the private, impacted information about the applicant’s credit or driving his- tory. Nevertheless, just as in the case of claims made by the itinerant seller of an experi- ence good, verification remains a problem. The applicant has an incentive to omit facts that would tend to result in loan or insurance denial (e.g., prior business failures or un- reported accidents), and knowing this, the lender may offer only higher-rate loans and the insurer higher-rate policies.

The problem is that higher-rate loans and expensive insurance policies tend to affect the composition of the applicant pool, resulting in adverse selection. Some honest, well- intentioned borrowers and good-risk insurance applicants will now drop out of the ap- plicant pool because of concern about their inability to pay principal and interest and insurance premiums on time as promised. But other applicants who never intended to repay (or drive carefully), or more problematically, those who will try less hard to avoid default or accidents, are undeterred by the higher rates. The asymmetric information and

TABLE 10.2 EXPERIENCE-GOOD PAYOFF MATRIX

Buyer

Offer High Price Offer Low Price

Better Best

Break even Loss ( D)

Worst Worse

Profit (D E ) Break even

High Quality

Seller

Low Quality

Note: Column-player payoffs are above diagonal. Row-player payoffs are below diagonal.

adverse selection A limited choice of lower-quality alternatives attributable to asymmetric information.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 371

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

higher rates have adversely selected out precisely those borrowers and drivers the lender and auto insurance company wanted to attract to their loan portfolio and insurance risk pool. Recognizing this problem, the creditors and insurers offer a restricted and incom- plete set of loan and insurance contracts. Credit rationing that excludes large segments of the population of potential borrowers and state-mandated protection against uninsured motorists are reflections of the adverse selection problem resulting from asymmetric in- formation in these commercial lending and auto insurance markets.

SOLUTIONS TO THE ADVERSE SELECTION PROBLEM In both theory and practice, there are two approaches to eliciting the exchange of high- quality experience goods, commercial loans to new borrowers, or auto insurance policies to new residents. The first involves regulatory agencies such as the Federal Trade Com- mission, the Food and Drug Administration, and the Consumer Product Safety Commis- sion. These agencies can attempt to set quotas (e.g., on minimum product durability, on minimum lending in “red-lined” underprivileged communities, or on minimum auto li- ability insurance coverage). They may also impose restrictions (e.g., on the sale of un- tested pharmaceuticals), enforce product safety standards (e.g., on the flammability of children’s sleepwear), and monitor truth-in-advertising laws. We discuss public regula- tion at greater length in Chapter 16.

Mutual Reliance: Hostage Mechanisms Support Asymmetric Information Exchange A second, quite different approach involves self-enforcing private solution mechanisms where each party relies on the other. Such reliance relationships often involve the ex- change of some sort of hostage, such as a reputational asset, an escrow account, or a surety bond. In general, hostage or bonding mechanisms are necessary to induce unregulated asymmetric information exchange. For this second approach to the adverse selection prob- lem to succeed, buyers must be convinced that fraud is more costly to the seller than the cost of delivering the promised product quality. Then, and only then, will the customers pay for the seller’s additional expected costs attributable to the higher-quality products.

One simple illustration of the use of a hostage mechanism to support asymmetric in- formation exchange is a product warranty, perhaps for an auto tire. Tires are an experi- ence good in that blowout protection and tread wear life are product qualities not detectable at the point of purchase. Only by driving many thousands of miles and ran- domly encountering many road hazards can the buyer ascertain these tire qualities di- rectly. However, if a tread wear replacement warranty and a tire blowout warranty make the sellers conspicuously worse off should they fail to deliver high-quality tires, then buyers can rely on that manufacturer’s product claims. As a consequence, buyers will be willing to offer higher prices for the unverifiably higher-quality product.

Hostage mechanisms can be either self-enforcing or enforced by third parties. Like warranties, a seller’s representations about after-sale service and product replacement guarantees are ultimately contractual agreements that will be enforced by the courts. However, other hostage mechanisms require no third-party enforcement. Suppose Du- Pont’s industrial chemicals division reveals to potential new customers the names and addresses of several satisfied current customers. This practice of providing references is not only to assist potential buyers in gauging the quality of the product or service for sale but also to deliver an irretrievable hostage. Once new customers have the easy ability to contact regular customers and blow the whistle on product malfunctions or

reliance relationships Long-term, mutually beneficial agreements, often informal.

hostage or bonding mechanisms A procedure for establishing trust by pledging valuable property contingent on your nonperformance of an agreement.

372 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

misrepresentations, the seller has an enhanced incentive to deliver high quality to both sets of buyers. Connecting all suppliers and customers in a real-time information system is a natural extension of this familiar practice of providing references. The total quality movement’s (TQM’s) ISO 9000 standards recommend that companies insist on just such information links to their suppliers’ other customers.

Brand-Name Reputations as Hostages A marketing mechanism that supports asymmetric information exchange is a brand- name reputation such as Sony Trinitron Wega digital televisions, Apple Macintosh com- puters, Pepperidge Farm snacks, and Toyota Lexus automobiles. Branding requires a substantial investment over extended periods of time. Moreover, brand names are capital assets that provide future net cash flows from repeat-purchase customers as long as the brand reputation holds up. To defraud customers by delivering less quality than the brand reputation promised would destroy the capitalized market value of the brand name. Buyers anticipate that value-maximizing managers will not intentionally destroy brand-name capital. Brand names therefore deliver a hostage, providing assurances to buyers that the seller will not misrepresent the quality of an experience good.

Ultimately, brand-name capital provides such a hostage because the disreputation ef- fects on the brand name that result from delivering fraudulent product quality cannot be separated from the salable brand asset. Successful brands can be extended to sell other products; Nestlé’s original hot chocolate brand can be extended to sell cereal-based candy bars, and Oreo cookies can be extended to sell ice cream. But the product failure of Texas Instruments (TI) personal computers means that now the TI brand name can- not be easily extended to other consumer electronic products. All the potential buyers have to figure out is whether the seller would be worse off sacrificing the value of the brand name but economizing on production expenses rather than simply incurring the

Example Credible Product Replacement Claims: Dooney & Bourke The women’s handbag market has a wide selection of brand names, prices, and qualities. Leather products have several search-good characteristics in that one can touch and feel the material in order to assess the fineness or coarseness of the grain, the evenness of the tanning process, the suppleness of the leather, and so forth. In these respects, one can search for just that quality for which one is willing to pay. However, the susceptibility to discoloring with age or exposure to the elements and the quality of the stitching are much harder to detect at the point of purchase. As a result, some aspects of handbag purchase are an experience-good exchange. Therefore, one wonders how the wide variety of prices and qualities can be sustained. Dooney & Bourke resolved this question by offering an almost preposterous replacement guarantee. Like Revo sunglasses, Dooney & Bourke offered to replace any handbag for the life of the customer. Because each state attorney general will assist any customer in enforcing this promise, the commitment was credible, and the replacement guarantee provides a hostage that supports high-price, high-quality exchanges. In particular, custo- mers can easily discern that Dooney & Bourke is better off producing an excep- tionally high-quality handbag to deliver at the first transaction rather than an unlimited series of replacements.

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 373

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

extra expense to produce a high-quality product while retaining the brand value. A brand-name asset such as Pepperidge Farm may suggest one answer, whereas Joe’s Garage suggests another.

If brand-name assets could be sold independently of their reputations (or disreputa- tions), then this hostage mechanism would cease to support experience-good exchange. Assets that can be redeployed at the grantor’s wish are not hostages in this reliance contracting sense. The implication is that easy entry and exit, which worked to ensure break-even prices just sufficient to cover costs in the normal competitive markets, may have undesirable consequences here in asymmetric information experience-good markets.

Price Premiums with Non-Redeployable Assets27

Recall that if sellers are offered prices that just cover high-quality cost, sellers of experi- ence goods prefer the profit from defrauding customers by delivering low-quality prod- ucts. But suppose buyers offered reliable sellers a continuing price premium above the cost of high-quality products. At Phh in Figure 10.10, the non-notorious firm produces

Example Customers for Life at Sewell Cadillac26

The most profitable luxury automobile dealership in the United States is operated in Dallas, Texas, by Carl Sewell. Several decades ago, Mr. Sewell realized that the critical success factor in his business was establishing repeat-purchase transactions with regular customers. Many potential buyers shop for lowest price in the new automobile market, sometimes with no more inconvenience than fingertip brows- ing of the Internet. And because the alternatives are many, and the information on posted prices is great, many dealerships spend several hundred dollars per car on personal selling costs with little prospect of repeat business. Carl Sewell decided instead to expend similarly large amounts attracting “customers for life.” He began by making the apparently preposterous claim that he would dispatch Sewell Cadil- lac emergency roadside service to any Sewell Cadillac customer experiencing car trouble anywhere in the state of Texas. To economize on the need for such trips, Sewell developed an extensive dealer-based maintenance schedule and instituted one of the first total quality management (TQM) programs in his service department.

Because these policies introduced new process-based competitive advantages, they were difficult for other dealers to imitate. These process innovations cost plenty, but the word-of-mouth reputation effects every time the dealership deliv- ered on its promise spread the name and quality image of Sewell Cadillac across North Texas. Soon customers were driving in from surrounding cities for the priv- ilege of doing high-margin business with Carl Sewell. And even more importantly, these same customers came back time and time again with very little additional selling cost to the dealership.

26See Carl Sewell and Paul B. Brown, Customers for Life (New York: Simon & Schuster, 1992).

27See B. Klein and K. Leffler, “The Role of Market Forces in Assuring Contractual Performance,” Journal of Political Economy 89, no. 4 (1981), pp. 615–641.

374 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Q´1 high-quality product and earns a continuous stream of profits (IJAG + JKA), labeled T + U. This perpetuity may now exceed (in present value) the notorious firm’s one- time-only fraudulent rent from production at Q´2—namely, D + T, plus incremental profit E + U + V. That is,

(T + U)/d > [(D + T) + (E + U + V)]/(1 + d) [10.9]

where d is an appropriate discount rate (e.g., the firm’s weighted average cost of capital, perhaps 12 percent). By Equation 10.9, lower discount rates or faster rising marginal cost (i.e., a smaller incremental profit from the expansion of output, shaded area V in Figure 10.10) decreases the likelihood of fraudulent behavior. If reliable delivery of a high- quality product does in fact earn long-term net profit in excess of the one-time-only profit from fraud, sellers will offer both low- and high-quality products at Pl and Phh, respectively, and some buyers will purchase in each market.

However, transitory profits alone do not allow an escape from adverse selection. Be- cause profits attract entry in competitive markets, the price premiums will erode, and notorious firm behavior will then return. What is missing is a mechanism to dissipate the rent from the price premiums. If the sellers invest the high-quality price premiums in firm-specific assets, such as L’eggs retail displays for convenience stores or Ethan Allen’s interiors for their showrooms, then new entrants will encounter a higher entry barrier than previously. Such barriers cause potential entrants to perceive much lower potential net profit and therefore deter entry. L’eggs or Ethan Allen’s operating profits in excess of the production cost can then persist, and high-quality, high-price experience goods can survive in the marketplace.

The rent-dissipating investments must not be in generic retail sites easily redeploy- able to the next tenant or capital equipment easily redeployable to the next manufac- turer (e.g., corporate jet aircraft). If that were the case, hit-and-run entry would recur

FIGURE 10.10 High-Quality Experience Goods Earn a Price Premium

Q2

A

D

Ph G

C

Q1

Pl

I

E

MChigh

MClow

SAChigh

SAClow

Dh

Dl

B

Q1�

Phh JK M

U VT

Units0

Dhh

Pr ic

e an

d co

st (

$/ un

it )

Q2�

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 375

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

each time high-quality prices rose above cost. New entrants would just move in on the business for a short time period and then sell off their assets in thick resale markets when profits eroded. Then, competitive equilibrium would again induce adverse selec- tion in experience-good markets. Instead, the investment that dissipates the operating profit from high-quality products must be sunk cost investment in non-redeployable assets.

Non-redeployable assets are assets whose liquidation value in second-best use is low. Usually this occurs when the assets depend on a firm-specific input such as a L’eggs or Ethan Allen brand name. Without the brand name, no firm has a use for the egg-shaped retail racks designed for L’eggs original packaging or the lavish Ethan Allen showrooms. Many such non-redeployable assets have high value in their first best use. The difference between value in first best use and liquidation value is a measure of the asset specificity. Highly specific assets make the best hostages to convince customers that asymmetric in- formation transactions will be nonfraudulent.

In summary, asymmetric information causes competitive markets for experience goods to differ rather markedly from the competitive markets for search goods. Long- run equilibrium for high-quality experience goods requires revenues in excess of total unit cost. These profits are invested by reliable sellers of experience goods in highly spe- cific assets. Potentially notorious firms with redeployable assets attract only customers seeking low-price/low-quality experience goods. In experience-good markets, you get what you pay for when reputations matter or when other hostage mechanisms establish the seller’s credibility.

Example Hostage Exchange with Efficient Uncut Diamond Sorting at De Beers28

Another illustration of experience-good exchange is block booking by the De Beers diamond cartel, which controls over 80 percent of the uncut wholesale diamond business. De Beers offers groupings of diamonds of various grades to approved wholesale buyers. Because buyers are not allowed to cull the less-valuable stones, the quality of the diamonds in any given grouping is unverifiable at the point of purchase—hence, the term sights. If these arrangements were one time only, no buyer would purchase high-price sights or agree to the culling restrictions. But be- cause block booking economizes on the duplicatory assessments of rejected stones that would otherwise result, De Beers can consistently offer its sights at net costs below the value at which the diamonds grade out. Buyers therefore have a reason for purchasing high-quality experience goods from De Beers. If a competitor of- fered no culling restrictions and lower prices, the diamond merchants would care- fully weigh the additional cost of sorting the diamonds themselves against the price premiums at De Beers and might well decide to continue doing business with De Beers. Knowing this, very few potential competitors ever enter the uncut diamond wholesale business to challenge De Beers despite its high markups and margins. De Beers’ reputation for passing on its cost savings in diamond sorting to buyers is the hostage that brings buyers back time and time again.

28Based on R. Kenney and B. Klein, “The Economics of Block Booking,” Journal of Law and Economics 26 (1983), pp. 497–540.

non-redeployable assets Assets whose value in second-best use is near zero.

asset specificity The difference in value between first-best and second-best use.

376 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

SUMMARY

� Competitive strategy entails an analysis of the firm’s resource-based capabilities, the design of business processes that can secure sustainable com- petitive advantage, and the development of a road map for innovation.

� Types of strategic thinking include industry analy- sis, competitor analysis, strategic positioning, and identification of core competencies derived from resource-based capabilities.

� Sustainable competitive advantage may arise from product differentiation strategy (product capabili- ties, branding, and endorsements), from focused cost or cost leadership strategy, or from informa- tion technology strategy.

� The choice of competitive strategy should be con- gruent with the breadth or narrowness of the firm’s strategic focus.

� A successful competitive strategy includes an ongo- ing process of reinvention and reconfiguration of capabilities and business models.

� A relevant market is a group of economic agents that interact with each other in a buyer-seller rela- tionship. Relevant markets often have both spatial and product characteristics.

� The Five Forces model of business strategy identi- fies threat of substitutes, threat of entry, power of buyers, power of suppliers, and the intensity of ri- valry as the determinants of sustainable incumbent profitability in a particular industry.

� The threat of substitutes depends upon the number and closeness of substitutes as determined by the product development, advertising, brand-naming, and segmentation strategies of preexisting com- petitors. Complements in consumption can be an enormous source of network effects, raising sus- tainable profitability.

� The threat of entry depends upon the height of barriers to potential entrants including capital re- quirements, economies of scale, absolute cost ad- vantages, switching costs, access to distribution channels, and trade secrets and other difficult- to-imitate forms of product differentiation.

� The bargaining power of buyers and suppliers de- pends upon their number, their size distribution, the relationship between industry capacity and in-

dustry demand, the uniqueness of the inputs, the potential for forward and backward integration, the ability of the buyers to influence the setting of an industry standard, and the extent to which each party to the bargain has outside alternatives.

� The intensity of rivalry depends upon the number and size distribution of sellers in the relevant mar- ket, the relative frequency of price versus non-price competition, switching costs, the proportion of fixed to total cost, the barriers to exit, the growth rate of industry demand, and the incumbent’s speed of adjustment.

� The demand for a good or service is defined as the various quantities of that good or service that con- sumers are willing and able to purchase during a particular period of time at all possible prices. The supply of a good or service is defined as the quan- tities that sellers are willing to make available to purchasers at all possible prices during a particular period of time.

� In general, a profit-maximizing firm will desire to operate at that level of output where marginal cost equals marginal revenue.

� In a purely competitive market structure, the firm will operate in the short run as long as price is greater than average variable cost.

� In a purely competitive market structure, the ten- dency is toward a long-run equilibrium condition in which firms earn just normal profits, price is equal to marginal cost and average total cost, and average total cost is minimized.

� In a monopolistically competitive industry, a large number of firms sell a differentiated product. In practice, few market structures can be best ana- lyzed in the context of the monopolistic competi- tion model. Most actual market structures have greater similarities to the purely competitive mar- ket model or the oligopolistic market model.

� Advertising expenditures are optimal from a profit-maximization perspective if they are carried to the point where the marginal profit contribution from an additional unit of output is equal to the marginal cost of advertising. The optimal level of advertising intensity (the advertising expenditure per sales dollar) varies across products and

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 377

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

industries; it is determined by the marginal profit contribution from incremental sales and by the ad- vertising elasticity of demand.

� Exchange under incomplete information and un- der asymmetric information differs. Incomplete in- formation refers to the uncertainty that is pervasive in practically all transactions and motivates insur- ance markets. Asymmetric information, on the other hand, refers to private information one party possesses that the other party cannot indepen- dently verify.

� Asymmetric information in experience-good mar- kets leads to adverse selection whereby high- price/high-quality products are driven from the market by low-quality products whose low quality is indistinguishable at the point of sale. Buyers in such lemons markets refuse to offer prices high enough to cover the cost of high quality because under competitive conditions suppliers will predict- ably commit fraud, and then perhaps move on to conduct business with unsuspecting customers un- der other product or company names.

� To escape adverse selection and elicit high-quality experience goods necessitates either intrusive and expensive regulation or some sort of bonding mechanism to induce self-enforcing reliance rela- tionships between buyers and sellers. Warranties, independent appraisals, leases with a high residual, collateral, irrevocable money-back guarantees, con- tingent payments, and brand names all provide assurance to buyers that the seller will not misrep- resent the product quality. Hostage mechanisms support asymmetric information exchange.

� Another way to escape adverse selection is for buyers to offer price premiums and repeat- purchase transactions to firms that resist fraudu- lently selling low-quality experience goods for high prices. These profits are invested by reliable sellers in non-redeployable, highly specific assets. Potentially notorious firms with redeployable assets continue to attract only customers seeking low- price/low-quality products. Under asymmetric in- formation, at best you get what you pay for, never more.

Exercises 1. The profitability of the leading cola syrup manufacturers PepsiCo and Coca-Cola and of the bottlers in the cola business is very different. PepsiCo and Coca-Cola enjoy an 81 percent operating profit as a percentage of sales; bottlers experience only a 15 percent operating profit as a percentage of sales. Perform a Porter’s Five Forces analysis that explains why one type of business is potentially so profitable relative to the other.

2. Television channel operating profits vary from as high as 45 to 55 percent at MTV and Nickelodeon down to 12 to 18 percent at NBC and ABC. Provide a Porter Five Forces analysis of each type of network. Why is MTV so profitable relative to the major networks?

3. The costs of producing steel have declined substantially from building a conven- tional hot-rolled steel mill down to the new minimill technology that requires only scrap metal, an electric furnace, and 300 workers rather than iron ore raw materials, enormous blast furnaces, rolling mills, reheating furnaces, and thou- sands of workers. What effect on the potential industry profitability would Por- ter’s Five Forces framework suggest this new technology had? Why?

4. Ethanol is again viewed as one part of a solution to the problem of shortages of petroleum products. Ethanol is made from a blend of gasoline and alcohol derived from corn or sugar cane. What would you expect the impact of this program to be on the price of corn, soybeans, and wheat?

5. Why invest capital in purely competitive industries with equilibrium margins that are razor thin and entrants that erode quasi profits? Suppose volume is not excep- tionally large, why then?

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

378 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

6. Assume that a firm in a perfectly competitive industry has the following total cost schedule:

OUTPUT (UNITS) TOTAL COST ($)

10 $110

15 150

20 180

25 225

30 300

35 385

40 480

a. Calculate a marginal cost and an average cost schedule for the firm. b. If the prevailing market price is $17 per unit, how many units will be produced

and sold? What are profits per unit? What are total profits? c. Is the industry in long-run equilibrium at this price?

7. Royersford Knitting Mills, Ltd., sells a line of women’s knit underwear. The firm now sells about 20,000 pairs a year at an average price of $10 each. Fixed costs amount to $60,000, and total variable costs equal $120,000. The production de- partment has estimated that a 10 percent increase in output would not affect fixed costs but would reduce average variable cost by 40 cents.

The marketing department advocates a price reduction of 5 percent to increase sales, total revenues, and profits. The arc elasticity of demand with respect to prices is estimated at −2. a. Evaluate the impact of the proposal to cut prices on (i) total revenue, (ii) total

cost, and (iii) total profits. b. If average variable costs are assumed to remain constant over a 10 percent in-

crease in output, evaluate the effects of the proposed price cut on total profits.

8. The Poster Bed Company believes that its industry can best be classified as monopolistically competitive. An analysis of the demand for its canopy bed has resulted in the following estimated demand function for the bed:

P = 1760 – 12Q

The cost analysis department has estimated the total cost function for the poster bed as

TC = 1 3 Q3 − 15Q2 + 5Q + 24,000

a. Calculate the level of output that should be produced to maximize short-run profits.

b. What price should be charged? c. Compute total profits at this price-output level. d. Compute the point price elasticity of demand at the profit-maximizing level of

output. e. What level of fixed costs is the firm experiencing on its bed production? f. What is the impact of a $5,000 increase in the level of fixed costs on the price

charged, output produced, and profit generated?

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 379

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

9. Jordan Enterprises has estimated the contribution margin (P – MC)/P for its Air Express model of basketball shoes to be 40 percent. Based on market research and past experience, Jordan estimates the following relationship between the sales for Air Express and advertising/promotional outlays:

ADVERTISING/PROMOTIONAL OUTLAYS SALES REVENUE

$500,000 $4,000,000

600,000 4,500,000

700,000 4,900,000

800,000 5,200,000

900,000 5,450,000

1,000,000 5,600,000

a. What is the marginal revenue from an additional dollar spent on advertising if the firm is currently spending $1,000,000 on advertising?

b. What level of advertising would you recommend to Jordan’s management?

10. Which of the following products and services are likely to encounter adverse se- lection problems: golf shirts at traveling pro tournaments, certified gemstones from Tiffany’s, graduation gift travel packages, or mail-order auto parts? Why or why not?

11. If notorious firm behavior (i.e., defrauding a buyer of high-priced experience goods by delivering low quality) becomes known throughout the marketplace only with a lag of three periods, profits on high-quality transactions remain the same, and interest rates rise slightly, are customers more likely or less likely to agree to pay high prices for an experience good? Explain.

Case Exercises BLOCKBUSTER, NETFLIX, AND REDBOX

COMPETE FOR MOVIE RENTALS29 Charging $17.99 a month for an unlimited number of movie rentals (three at one time), Netflix revolutionized the movie rental business with a one-day mailing service for DVDs and acquired 12 million subscribers and $1.5 billion in revenue. However, Blockbuster, the video rental giant from the $5.5 billion bricks-and-mortar movie rental business, decided to enter the mail-in delivery and online-DVD rental busi- nesses. Blockbuster drove prices down to $14.99, attracting 2 million subscribers. Net- flix responded with a cut-rate service of one movie at a time for $9.99 per month, which drove the net profit right out of the business.

Use Porter’s Five Forces model to answer the following questions:

29“Movies to Go,” The Economist (July 9, 2005), p. 57; and “Blockbuster Plots a Remake,” Wall Street Jour- nal (February 24, 2010), p. B1.

380 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Questions 1. Does easy access to distribution channels at grocery stores for Redbox’s 22,000

vending machines indicate high or low entry threat in the movie rental business? Why? Why might McDonald’s be an even better distribution channel than grocery stores?

2. What economies of scale were available to serve as a barrier to entry in Block- buster’s bricks-and-mortar movie rental business? Did NetFlix face a cost advantage or disadvantage?

3. Who are Blockbuster’s suppliers? Are they in a position to appropriate much of the value in the value chain? Why or why not?

4. What factors determine the intensity of rivalry in any industry? Is the intensity of rivalry in the movie rental industry high or low? Why?

5. Porter’s Five Forces model is sometimes extended to Six Forces of Competition to include the threat to profitability imposed by disruptive technology. What dis- ruptive technology has threatened the bricks-and-mortar and mail-in movie rental business?

SAVING SONY MUSIC Explore the crisis that Internet file sharing of copyrighted music recordings has caused for Vivendi Universal, Sony Music, EMI, and AOL Time Warner Music, who together formerly supplied 70 percent of the global music industry.

Questions 1. How would the Internet firms Napster, Kazaa, and Apple’s iTunes Music Store be

reflected in a Porter Five Forces industry analysis? 2. Why was the Internet a disruptive technology for Sony Music? 3. What should be Sony Music’s competitive strategy in response to this crisis? In-

clude a discussion of resource-based capabilities, business opportunities, and a road map of future innovation.

4. Is your competitive strategy for Sony Music a product differentiation strategy, a low-cost strategy, or an information technology strategy? What is your strategic focus?

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 381

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

11 CHAP T E R

Price and Output Determination: Monopoly and Dominant Firms CHAPTER PREVIEW In this chapter we analyze how firms that operate in monopoly or near-monopoly markets make output and optimal pricing decisions. In such markets the dominant firm does not have to accept the market price as a given. These firms base their price-cost markups on other factors such as the de- mand projections at various price points, indicative of the target customers’ price elasticity. In this chapter we identify the reasons for single-firm dominance and ana- lyze the components of the contribution margin and the gross margin for such firms. We introduce spreadsheet, graphical, and algebraic methods to calculate profit- maximizing price and output decisions. In addition, we look at these decisions for regulated industries: electric power, natural gas distribution and transmission, and broadcast communications. Deregulation continues to be a topic of debate, and it is important that any policy changes be consistent with microeconomic principles.

MANAGERIAL CHALLENGE Dominant Microprocessor Company Intel Adapts to Next Trend1

With continuous innovation, ever-faster, more-powerful chip designs, and a business plan riveted on supplying the $200 billion personal computer (PC) industry, Intel Cor- poration dominates the high-end market in microproces- sors. After being forced out of the dynamic random access memory (DRAM) chip business by Japanese rivals in 1986, Intel reinvented itself as the lead supplier of micro- processors for PCs. Intel has an 85 percentmarket share in the microprocessor chips for laptops and 75 percent mar- ket share for desktops. In addition, Intel sells 90 percent of the chip sets that control the flow of data from the micro- processor to the display screens, modems, and graphical user interfaces. Its market dominance provides it with enormous economies of scale in production and increas- ing returns on its marketing expenditures, which allow it to beat out its smaller rivals. The result is high markups and margins; for example, Intel has at times earned 25 percent net profit margins on its microprocessors.

Because intellectual property is the company’s most im- portant asset, Intel protects its proprietary trade secrets about chip design and manufacture by using tight nondis- closure agreements with its customers. Some Intel chip purchasers found, however, that Intel withholds vital infor- mation about technical specifications required to fully in- tegrate the chips into new products unless it is given access to its customers’ new technologies. Intergraph, a maker of high-end workstations for media applications, alleged, for example, that Intel withheld information about subtle bugs in some Intel chips until Intergraph agreed to license its graphical user interface technology to the chip supplier.

Intel’s high-end chips are designed to run Microsoft’s complex software for PCs. In 2007, 261 million units were shipped on a 2.1 billion installed base of PCs. The market for digital telephones, handheld computers, video-game players, and set-top control boxes for digital televisions may be even bigger than the PC market. Such devices

382

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

MONOPOLY DEFINED Monopoly is defined as a market structure with significant barriers to entry in which a single firm produces a highly differentiated product. Without any close substitutes for the product, the demand curve for a monopolist is often an entire relevant market demand. Just as purely competitive market structures (e.g., for AAA January wheat in Kansas City) are rare, so too are pure monopoly markets rare.

SOURCES OF MARKET POWER FOR A MONOPOLIST Monopolists or near-monopoly dominant firms enjoy several sources of market power. First, a firm may possess a patent or copyright that prevents other firms from producing

Example The Mickey Mouse Monopoly: Disney When it began, Disneyland in Anaheim, California, was unique. Other theme parks that were later developed, such as Six Flags, reduced Disney’s monopoly power. In an attempt to restore its near-monopoly position, Disney created Disney World in Orlando, Florida, but Universal Studios, SeaWorld, and other attractions through- out the Orlando area quickly offered additional theme park experiences. Were they a complement or a substitute to Disney World? Negative cross price elasticity of demand evidence suggests that they are complementary relationships. Seventy per- cent of Disney World’s business is repeat business; more variety inside or outside the park means more frequent returns for longer vacations. Because it anticipated these complement relationships, Disney long ago became a major development property owner throughout the Orlando area.

require inexpensive flash memory chips that quickly pro- cess data. Samsung and Advanced Micro Devices (AMD) are the leaders in this new chip segment. To break into this business, Former Intel president Andy Grove says Intel must prepare to sell lower-end chip products for under $40, despite the fact that Intel’s chips previously sold for $87 to $200.

Discussion Questions

� How large must the market share be before a firm dominates a relevant market?

� Make a list of reasons how a single firm comes to dominate some markets.

� What are some of the differences between op- erating profits and net cash flow to equity shareholders?

� Name some other firms that you suspect earn higher-than-normal profit margins, and brain- storm about why they do so.

1Based on “Hand-Held Combat,” Wall Street Journal (February 12, 1998), p. A1; “Showdown Looms over Chip Giant’s Power,” Wall Street Journal (June 8, 1998), p. B1; “Intel’s Surge,” Wall Street Journal (July 20, 2005), p. B1; “Intel Outside,” The Economist (May 27, 2006), pp. 59–63; and Apple Inc., Harvard Business School Case Publishing (2008).

MANAGERIAL CHALLENGE Continued ©

Ta yl or

S. Ke nn ed y/ N at io na lG

eo gr ap hi c/ Ge tty

Im ag es

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 383

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the same product. For example, Pharmacia & Upjohn, Inc. has a patent on the product Rogaine, the hair growth stimulator for balding men.

Second, a firm may control critical resources. De Beers Consolidated Mines, Ltd. owns or controls most of the diamond production in South Africa and often obtains exclusive marketing agreements with other major diamond-producing countries, including the for- mer Soviet Union. This control of raw materials enabled De Beers to maintain high world prices for cut diamonds for nearly three-quarters of a century.

A third source of monopoly power may be a government-authorized franchise. In most U.S. cities, one firm is chosen to provide exclusive cable TV services to the commu- nity. The same type of monopoly power occurs when a government agency such as the FCC adopts an industry standard that favors one company over another.

Monopoly power also happens in natural monopolies because of significant economies of scale over a wide range of output. The first entrant firm will enjoy declining long-run average costs. Under these circumstances, one supplier of the good or service is able to produce the output more cheaply than can a group of smaller competitors. These so- called natural monopolies are usually closely regulated by government agencies to restrict the profits of the monopolist.

Increasing Returns from Network Effects Finally, increasing returns in network-based businesses can be a source of monopoly mar- ket power. When Microsoft managed to achieve a critical level of adoption for its Win- dows graphical user interface (GUI), the amount of marketing and promotional expenditure required to secure the next adoption actually began to fall.

Marketing and promotions are generally subject to diminishing returns, as depicted in Figure 11.1. From 0 to 30 percent market share, the marketing required to achieve each additional share point has a diminishing effect on the probability of adoption by the next potential user (note the reduced slope of the sales penetration curve). Consequently, addi- tional share points become more and more expensive over this range. When the number of

Example Impermanent Control of a Denver Airport Hub: United Airlines The market power that comes from the control of critical resources is often tempo- rary. After the deregulation of airlines, some major carriers developed “fortress” hubs. US Airways, United, Delta, and Northwest control most of the gates at their hubs in Charlotte, Denver, Atlanta, and Minneapolis, respectively. These dominant carriers created barriers to entry by signing long-term leases with airport authori- ties. Local customer loyalty then supported 20 to 27 percent price premiums based on the delivery reliability, change order responsiveness, and nonstop scheduling convenience at these hubs.

By the mid-1990s, low-cost airlines threatened to break into the hubs of these market leaders. Delta encountered strong challenges from Kiwi and AirTran, whose presence in Atlanta caused margins on competing routes to dwindle. Fron- tier Airlines sparked discount and drive-in traffic at United’s Denver hub. Conse- quently, United fares to and from Denver declined. US Airways may still control the majority of the departures from Charlotte, but its high fares caused the city council to approach Southwest Airlines and AirTran about possible entry.

sales penetration curve An S-shaped curve relating current market share to the probability of adoption by the next target customer, reflecting the presence of increasing returns.

384 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

other users of a network-based device reached a 30 percent share, the next 50 or so share points became cheaper to promote. That is, beyond the 30 percent inflection point, each additional share point of users connected to Windows increased the probability that an- other user would adopt. Therefore, the marketing expense required to secure another unit sale decreased. (Note the increased slope of the sales penetration curve in the middle por- tion of Figure 11.1.) Then beyond 85 percent, diminishing returns again set in.

These network-based effects of compatibility with other users increase the value to the potential adopter. The same thing occurs as more independent software vendors (ISVs) write applications for an operating system like Windows that has effectively become an industry standard by achieving more than 30 percent acceptance in the marketplace. The inflection points in the sales penetration curve make it likely that Microsoft will achieve an 85 percent monopoly control of the operating system market. Whatever customer re- lationships preexisted, once Microsoft achieved a 30 percent share, its increasing returns in marketing caused a network effect that displaced other competitors. Microsoft’s share then grew to 92 percent. Netscape’s Internet search engine experienced similar displace- ment by Microsoft’s Internet Explorer when Microsoft achieved a 30 percent-plus market share by bundling Internet Explorer with Windows. In effect, it gave away the search engine for free to reach the range of increasing returns on the sales penetration curve for OS software.

Even with increasing returns set off by network effects, monopoly seldom results for three reasons. First, a higher price point for innovative new products can offset the cost savings from increasing returns of a competitor. This has been Apple’s approach to combatting Microsft dominance on the operating systems of Dell and Hewlett- Packard PCs. Apple’s gross margin exceeded 32 percent for 2005–2008, whereas Dell and HP averaged 18 percent and 25 percent, respectively. Second, network effects tend to occur in technology-based industries that have experienced falling input prices.

FIGURE 11.1 How the Adoption of a Technology Leads to Increasing Returns

Pr ob

ab ili

ty o

f ad

op ti

on b

y ne

xt t

ar ge

t cu

st om

er 1.0

9 13 16 30 85 92 Time path of market share (%)

Sales penetration curve

Microsoft Windows GUIApple GUI

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 385

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Figure 11.2 shows that between 1997 and 2009, the cost per megahertz for silicon com- puter chips fell from $2.00 to $0.25, hard drive storage device cost per megabyte fell from $0.40 to $0.03, and the cost per month for a T1 high-speed data transmission line fell from $475 to $300. During the same period, Corning fiber-optic cable became essentially free to anyone who would install it. In short, as these input suppliers grew to serve the expanding product markets in computer equipment and telecom devices, they encountered new productivity from learning curves and innovative design break- throughs that drove down their costs. Because flash memory chips and telecom equip- ment markets tend to be highly competitive, the cost savings of input suppliers such as AMD and Corning get passed along to the final product producers, including Apple, PC-assembler Dell, cell phone manufacturer Nokia, and router manufacturer Cisco. Consequently, generally lower costs for all inputs offset in large part the dominant firm advantage from increasing returns in promotion and selling expenses for compa- nies such as Microsoft.

Third, technology products whose primary value lies in their intellectual property (e.g., computer software, pharmaceuticals, and telecom networks) have revenue sources that are dependent on renewals of governmental licensures and product standards. Un- like autos or steel, once R&D costs have been recouped, the marginal cost of additional copies of the software, additional doses of the medicine, or additional users on the wire- less system are close to zero; that means every single unit sold thereafter is close to pure profit. Competitor firms who have incurred the up-front fixed costs but not succeeded in

Example What Went Right at Microsoft but Wrong at Apple Computer2

Throughout much of its history, Apple Computer, discussed in the Managerial Chal- lenge at the beginning of Chapter 10, hovered at 7 to 10 percent market share in the U.S. personal computer market. Twice in its early history, Apple reached double- digit share points (16 percent in 1986 and 13 percent in 1993). Apple never did come close to achieving the inflection point (depicted at 30 percent in Figure 11.1). Apple therefore pursued increasing returns by attempting to become an industry standard in several personal computer submarkets such as the desktop publishing, journalism, media-based advertising, and entertainment industries.

In addition, despite fiercely defending its graphical user interface (GUI) code for almost two decades with patent applications and trade secret infringement suits, finally in 1998–1999 Apple reversed its course and began licensing and alliance agreements with both Microsoft and IBM. Compatibility with other operating sys- tems had been easy to achieve, but widespread adoption of Mac programming code by independent software vendors (ISVs) had not. Consequently, to obtain a critical mass of adoptions that would trigger ISVs to begin writing software applications for the Mac, Apple reversed its company policy on the closed architecture of its GUI. The GUI code at Apple was clearly technically superior to the early genera- tion Windows products. However, the technically superior product lost out to the product that first reached increasing returns—namely, Microsoft Windows GUI running on non-Apple PCs.

2Based on “Netscape to Woo Microsoft’s Customers,” Reuters (October 1, 1998); W. Brian Arthur, “Increasing Returns and the New World of Business,” Harvard Business Review (July–August 1996); “Sorting Out the Deal,” U.S. News and World Report (August 18, 1997), p. 20; and Apple Inc., Harvard Business School Case Publishing (2008).

386 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIGURE 11.2 How Declining Component Costs Led to Falling Product Prices in the Computer and Telecom Industries

Silicon chips Hard drive storage

T1 line

$2.50

$2.00

$1.50

$1.00

$0.50 $0.25

1997 1998 1999 2000 2001 2002 2009

2009

$0.50

$0.40

$0.30

$0.20

$0.10

$0.03

1998 1999 2000 2001 2002

$500

$475

$450

$425

$400

$300

1998 1999 2000 20022001

C os

t ($

p er

M H

z)

C os

t ($

p er

m gb

)

C os

t ($

p er

m on

th )

2009

Source: “A Spoonful of Poison,” Wired (March 2002), p. 57; and price quotes.

WHAT WENT RIGHT • WHAT WENT WRONG

Pilot Error at Palm3

PalmPilot, the once-dominant product in handheld com- puters, demonstrates how fragile is the position of even an industry leader with increasing returns to promotional spending in a technology business. Despite having 80 per- cent of the handheld operating system market and despite producing 60 percent of the handheld hardware at its peak in 2000, Palm, Inc., lost market share to rivals. Palm grew so fast (165 percent year-over-year sales in- creases) that it gave little attention to operational issues such as managing the supply of inputs and forecasting demand. In 2001, it mistimed the announcement of its m500 product upgrades, which were delayed by supply

chain bottlenecks, and Palm’s customers stopped buying older models. Handspring, Sony, Hewlett-Packard, Micro- soft’s Pocket PC, and the popular BlackBerry drove prices lower and offered newer product features. Almost overnight, excess Palm IV and V inventories piled up on shelves, and inquiries about Apple iPods and Nokia’s handheld de- vices shot way up. Customers were awaiting the new model, and Palm was forced to take a $300 million write-down on their inventory losses. The stock price fell from $25 to $2 a share.

3Based on “How Palm Tumbled,” Wall Street Journal (September 7, 2001), p. A1.

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 387

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

reaching the inflection point of increasing returns will rationally spend enormous sums seeking to recoup these rents through the political process and in the courts. For example, Netscape and Sun Microsystems succeeded during Microsoft’s long antitrust trial of 1997–2002 in restricting their competitor. U.S. courts ordered restrictions on Microsoft’s installation agreements for Windows and prohibited Microsoft’s refusal to deal with Windows licensees who install Netscape’s competing Web browser software. And Genentech’s first commercial success was a multiple sclerosis drug that avoided direct challenge to a broad Schering-Plough Corporation patent by employing a special FDA rule. Similarly, Xerox was forced by antitrust authorities in the United States to license its wet paper copier technology at low royalty fees.

How do firms attempt to get around the inflection point of Figure 11.1 and achieve increasing returns? Free trials for a limited period of use is one approach. Another is giving the technology away if it can be bundled with other revenue-generating product offerings. Microsoft gave away Internet Explorer (IE) for free without being charged with predatory pricing (IE’s variable cost was $0.004; that is, it rounded to zero). Another approach is to undertake consolidation mergers and acquisitions; this strategy drove IBM’s acquisition of a host of smaller software companies, such as Lotus, and Oracle en- gaged in a hostile takeover of PeopleSoft. Some companies such as Sun Microsystems also provide JAVA and Linux programming subsidies to independent software vendors whose applications will provide network effects as complements to Sun’s JAVA-based OS. Finally, having a product adopted as an industry standard leads to increasing returns. Sony achieved this network effect with its Blu-Ray HDTV standard.

PRICE AND OUTPUT DETERMINATION FOR A MONOPOLIST Spreadsheet Approach

Example Profit versus Revenue Maximization for Polo Golf Shirts Table 11.1 shows the demand projections for daily sales of Polo golf shirts at a Ralph Lauren outlet store. For each style and color, the uniform price shown in column 2 is expected to elicit the number of unit sales per day listed in the first column. The third column displays the total revenue, and the fourth column shows the incremental revenue from lowering price to sell another unit—that is, the marginal revenue. For example, a uniform price reduction from $42 to $40 is required to increase unit sales from five to six shirts per day. Hence the marginal revenue at six shirts is calculated as (6 × $40 = $240) − (5 × $42 = 210)—that is, MR = $30.

Sales floor personnel are typically paid a salary plus a sales commission based on the total sales revenue they sell. Such an employee wants the price at the outlet store to continue dropping as long as total sales revenue rises—that is, as long as the MR remains positive up to and including 14 shirts/day at $25.79. Any fewer shirts and total revenue (P × Q = $361) would decline, reducing the sales team’s commission-based earnings. The store manager and Ralph Lauren, the parent com- pany, have other motives, however. These decision makers are concerned that

(Continued)

388 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Graphical Approach Figure 11.3 shows the price-output decision for a profit-maximizing monopolist. Just as in pure competition, profit is maximized at the price and output combination, where MC = MR. This point corresponds to a price of P1, output of Q1, and total profits equal to BC profit per unit times Q1 units. For a negatively sloping demand curve, the MR

the 14th shirt imposes an unit operating loss of –$24. That is, column 5 lists the variable cost incurred when another shirt is produced, distributed, and sold per- haps at a liquidator’s Web site or an outlet store. When the price in column 4 ($4 on the 14th shirt per day) falls below the variable cost in column 5 ($28), unit operating losses ensue. This same thing is true of the 13th, 12th, 11th shirt, and so forth.

Not until the outlet store raises its price and increases marginal revenue back to $28 will operating losses be eliminated. At this price and output ($38.31 and 7 shirts), the difference between total revenue ($268) and total variable costs ($28 × 7) is maximized at $72 per day. The store manager will be charged with the objective of pursuing these maximum operating profits and finding mechan- isms to motivate the sales personnel even though they would prefer maximum rev- enue of $361 at a $25.79 price point, despite operating losses of $39 per day.

TABLE 11.1 RALPH LAUREN POLO GOLF SHIRTS (PER COLOR, PER STORE, PER DAY)

QUANTITY SOLD

UNIFORM PRICE

TOTAL REVENUE

MARGINAL REVENUE

VARIABLE COST

UNIT OPERATING

PROFIT CUMULATIVE

PROFIT

0 $50.00 $0.00 $0.00 $28.00 $0.00 $0.00

1 $48.00 $48.00 $48.00 $28.00 $20.00 $20.00

2 $46.00 $92.00 $44.00 $28.00 $16.00 $36.00

3 $45.00 $135.00 $43.00 $28.00 $15.00 $51.00

4 $44.00 $176.00 $41.00 $28.00 $13.00 $64.00

5 $42.00 $210.00 $34.00 $28.00 $6.00 $70.00

6 $40.00 $240.00 $30.00 $28.00 $2.00 $72.00

7 $38.31 $268.17 $28.00 $28.00 $0.00 $72.00

8 $36.50 $292.00 $24.00 $28.00 ($4.00) $68.00

9 $34.50 $311.00 $19.00 $28.00 ($9.00) $59.00

10 $32.70 $327.00 $16.00 $28.00 ($12.00) $47.00

11 $30.91 $340.00 $13.00 $28.00 ($15.00) $32.00

12 $29.17 $350.00 $10.00 $28.00 ($18.00) $14.00

13 $27.46 $357.00 $7.00 $28.00 ($21.00) ($15.00)

14 $25.79 $361.00 $4.00 $28.00 ($24.00) ($39.00)

15 $24.07 $361.00 $0.00 $28.00 ($28.00) ($67.00)

16 $22.50 $360.00 ($1.00) $28.00 ($29.00) ($96.00)

17 $20.82 $354.00 ($4.00) $28.00 ($32.00) ($128.00)

18 $19.28 $347.00 ($7.00) $28.00 ($35.00) ($163.00)

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 389

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

function is not the same as the demand function. In fact, for any linear, negatively slop- ing demand function, the marginal revenue function will have the same intercept on the P axis as the demand function and a slope that is twice as steep as that of the demand curve. If, for example, the demand curve were of the form

P = a − bQ

then

Total revenue = TR = P · Q = aQ − bQ2

and

MR = dTR dQ

= a − 2bQ

The slope of the demand curve is −b, and the slope of the MR function is −2b.

Algebraic Approach

FIGURE 11.3 The Price and Output Determination of a Pure Monopoly

Q1

Output Q (units)

ACMC

Q2

B

A

CP1

D� DMR

D�

D

0

Pr ic

e an

d co

st (

$/ un

it )

Example Profit Maximization for a Theme Park Restaurant Assume a manager is faced with the following demand curve for lunch meals in a unique theme park restaurant:

Q = 400 − 20P

(Continued)

390 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Importance of the Price Elasticity of Demand Recall from Chapter 3 that marginal revenue (MR), the incremental change in total rev- enue arising from one more unit sale, can be expressed in terms of price (P) and the price elasticity of demand (ED), or

and the short-run total variable cost function is

TC = 5Q + Q2

50

To maximize profits, the theme park manager would produce and sell enough lunches such that MC = MR, and charge the corresponding uniform price:

MC = dTC dQ

= 5 + Q 25

MR may be found by rewriting the demand curve in terms of Q:

P = −Q 20

+ 20

and then multiplying by Q to find TR:

TR = P · Q

=− Q2

20 + 20Q

MR= dTR dQ

= − Q 10

+ 20

Setting MR = MC yields

− Q* 10

+ 20 = 5 + Q* 25

Q* = 107 units

Substituting Q* back into the demand equation, we may solve for P*:

P* = −107 20

+ 20

= $14:65=unit

Hence the profit-maximizing monopolist would produce 107 meals and charge a price of $14.65 each, which would yield a profit of

π* = TR − TC

= ðP* · Q*Þ − 5Q* + Q* 2

50

� �

= 14:65ð107Þ − 5ð107Þ + ð107Þ 2

50

� � = $803:57

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 391

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

MR = P 1 + 1 ED

� � [11.1]

Equating MR with MC (as shown in Figure 11.3) yields the profit-maximizing rela- tionship in terms of price and price elasticity, or

MC = P 1 + 1 ED

� � [11.2]

Hence, noncompetitive price will be greater than marginal cost. For example, if price elasticity ED = –2.0, price will equal

MC = P 1 + 1 −2

� � MC = Pð0:5Þ P = 2MC

Note from Equation 11.2 that a monopolist will never operate in the area of the de- mand curve where demand is price inelastic (i.e., |ED| < 1). If the absolute value of price elasticity is less than 1(|ED| < 1), then the reciprocal of price elasticity (1/ED)

will be less than –1, and marginal revenue P 1 + 1ED

� �h i will be negative. In Figure 11.3,

the inelastic range of output is output beyond level Q2. A negative marginal revenue means that total revenue can be increased by reducing output (through an increase in price). But we know that reducing output must also reduce total costs, resulting in an increase in profit. A firm will continue to raise prices (and reduce output) as long as the price elasticity of demand is in the inelastic range. Therefore, for a monopolist, the price-output combination that maximizes profits must occur where |ED| ≥ 1.

Equation 11.2 also demonstrates that the more elastic the demand (suggesting the ex- istence of better substitutes), the lower the price (relative to marginal cost) that any given firm can charge. This relationship can be illustrated with the following example.

Example Price Elasticity and Price Levels for Monopolists Consider a monopolist with the following total cost function:

TC = 10 + 5Q

The marginal cost (MC) function is

MC = dTC/dQ = 5

The price elasticity of demand has been estimated to be –2.0. Setting MC = MR (where MR is expressed as in Equation 11.1) results in the following price rule for this profit-maximizing monopolist:

MC = $5 = Pð1 + 1=−2:0Þ = MR P = 5=ð0:5Þ = $10=unit

If, however, demand is more price elastic, such as ED = –4.0, a profit- maximizing monopolist would set the price at

P = $5/(0.75) = $6.67/unit

392 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

THE OPTIMAL MARKUP, CONTRIBUTION MARGIN, AND CONTRIBUTION MARGIN PERCENTAGE Sometimes it is useful and convenient to express these relationships among optimal price, price elasticity, and variable cost as a markup percentage or contribution margin percent- age. Using Equation 11.2 to solve for optimal price yields (with MC = variable cost)

P = ED

ðED + 1Þ MC = ð1 + %Mark-upÞMC [11.3]

where the multiplier term ahead of MC is 1.0 plus the percentage markup.4 For example, the case of ED = –3 is a product with a −3/(−3 + 1) = 1.5 multiplier on MC—i.e., a 50 percent markup. The optimal profit-maximizing price recovers the marginal cost and then marks up MC another 50 percent. If MC = $6, this item would sell for 1.5 × $6 = $9 and the profit-maximizing markup is $3, or 50 percent more than the marginal cost.

The difference between price and marginal cost (i.e., the absolute dollar size of the markup) is often referred to as the contribution margin. With the incremental variable cost already covered, these additional dollars are available to contribute to covering fixed cost and earning a profit. They are expressed as a percentage of the total price. In the previous example, the $3 markup above and beyond the $6 marginal cost represents a 33 percent contribution to fixed cost and profit, that is, a 33 percent contribution margin on the $9 item. Using Equation 11.3 and ED = –3,

ðP − MCÞ P

= 1:5 MC − 1:0 MC

1:5 MC Contribution Margin% = 0:5=1:5 = 33%

To summarize, an elasticity of –3.0 implies that the profit-maximizing markup is 50 percent, and that 50 percent markup implies a 33 percent contribution margin. Price elasticity information therefore carries implications for the marketing plan. Combining the contribution margin percentage (33%) with incremental variable cost information in- dicates what dollar markups and product prices to announce.

One takeaway is that the more elastic the demand function for a monopolist’s output, the lower the price that will be charged, ceteris paribus. In the extreme, consider the case of a firm in pure competition with a perfectly elastic (horizontal) demand curve. In this case the price elasticity of demand approaches –∞; hence, 1 divided by the price elasticity approaches 0 and marginal revenue in Equation 11.1 becomes equal to price. Thus, the profit-maximizing rule in Equation 11.2 becomes “Set price equal to marginal cost,” and the profit-maximizing markup in Equation 11.3 is zero (i.e., the marginal cost multiplier equals just 1.0). Of course, this conclusion is the same price-cost solution developed in Chapter 10 in the discussion of price determination under pure competition.

So, the question remains: how does a noncompetitive firm establish a strategy to sus- tain higher contribution margins such as Chanel No. 5’s 91 percent when Ole Musk achieves only 8 percent? The key ideas are laid out in the Strategy Map shown in Figure 11.4. We will illustrate with Natureview Farms (NVF) Yogurt, a Vermont-based green producer of dairy products. All effective business plans begin with a value proposition for the target customers. As the U.S. population became more environmentally

4The symbol MC may be understood to refer to the accountant’s narrow definition of variable costs, operating costs that vary with the least aggregated unit sale in the business plan.

value proposition A statement of the specific source(s) of perceived value, the value driver(s), for customers in a target market.

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 393

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

conscious, Natureview Farms identified a younger, better-educated yogurt buyer who perceived value from higher-quality ingredients with longer shelf life than was typical of natural and organic ingredients. Despite the absence of chemical preservatives, NVF’s yogurt remains fresh for 50 days rather than 20. This additional functionality in combi- nation with higher-quality ingredients reliably exceeding customer expectations for fresh texture and taste warranted an enhanced price premium. But to create financial value from these customer value drivers, NVF found it necessary to boost unit sales growth and increase asset utilization by moving from the natural foods stores into Whole Foods and other specialty supermarkets. Handling the distribution channel issues with robust operations management processes and effective marketing communications proved critical to sustaining a high profit margin.

Components of the Gross Profit Margin Gross profit margin (or just “gross margin”) is a term often used in manufacturing busi- nesses to refer to the profit margin after direct fixed costs as well as variable manufactur- ing costs are subtracted. For example, in a carpet plant, the gross margin on each product line would be the plant’s wholesale revenue minus the sum of input costs plus machinery setup costs for the product’s production runs involving that type of carpet.

Example Markups and Contribution Margins on Chanel No. 5, Ole Musk, and Whitman’s Sampler Consider three products available at the typical drugstore counter: Chanel No. 5, Whitman’s Sampler, and the private label (store brand) fragrance Ole Musk. Chanel has a loyal following of regular buyers and a price elasticity of –1.1. Whitman’s has some rather close substitutes but substantial name recognition and packaging familiarity; its price elasticity measures –1.86. Finally, customers perceive many close substitutes for the generic fragrance Ole Musk, whose price elasticity measures –12.0.

Table 11.2 shows the optimal prices, markups, and contribution margins for these three products. Using Equation 11.3, the multiplier on MC for Chanel No. 5 is –1.1/(–1.1 + 1) = 11.0, and the optimal markup is therefore 1,000 percent (i.e., 10 times the incremental variable cost of the essences and the bottle). Because optimal price is 11.0 MC, the contribution margin percentage on Chanel No. 5 cal- culates as 10.0 MC/11.0 MC = 91%. Whitman’s Sampler has a multiplier of –1.86/(–1.86 + 1) = 2.16, an optimal markup of 116 percent and a contribution margin of 1.16 MC/2.16 MC = 54%. In contrast, Ole Musk with the greatest price elasticity has a multiplier of –12/(–12 + 1) = 1.09, a markup of 9 percent, and a contribution margin percentage of 0.09 MC/1.09 MC = 8%.

TABLE 11.2 OPTIMAL PRICES, MARKUPS, AND MARGINS

ED PRICE CONTRIBUTION

MARGIN MARKUP % CONTRIBUTION

MARGIN %

Chanel No. 5 ($85/oz.) −1.1 11.00 MC 10.00 MC 1000% 91%

Whitman’s Sampler ($8/lb.) −1.86 2.16 MC 1.16 MC 116% 54%

Ole Musk ($6/4 oz.) −12.0 1.09 MC 0.09 MC 9% 8%

gross profit margin Revenue minus the sum of variable cost plus direct fixed cost, also known as direct costs of goods sold in manufacturing.

394 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

A manufacturer’s income statement identifies variable manufacturing costs plus direct fixed manufacturing cost as the “direct cost of goods sold” (DCGS). Thus, the gross margin is revenue minus direct cost of goods sold.5

Gross profit margins differ across industries and across firms within the same industry for a variety of reasons. First, some industries are more capital intensive than others. Aircraft manufacturing with its large assembly plants is much more capi- tal intensive than software manufacturing. Boeing wide-body airframes have 72 per- cent gross profit margins, not because they are particularly profitable, but because airframes have high fixed costs for the capital investment tied up in large assembly plants. The first component of the gross profit margin percentage, then, is capital costs per sales dollar.

Second, differences in gross margins reflect differences in advertising, promotion, and selling costs. Leading brands in the ready-to-eat cereal industry have 70 percent gross mar- gins, but half of that price-cost differential (35 percent of every sales dollar) is spent on ad- vertising and promotion. The automobile industry also spends hundreds of millions of dollars on advertising, but that amounts to only 9 percent per sales dollar. The second com- ponent of the gross profit margin percentage is the advertising and selling expenses per sales dollar.

Third, differences in gross margins arise because of differential overhead in some businesses. The pharmaceutical industry has high gross margins, in large part because of the enormous expenditures on research and development to find new drugs. To con- duct business in that product line, other pharmaceutical firms then incur patent fees and licensing costs, which raise their overhead costs and prices. Overhead costs also may dif- fer if headquarters salaries and other general administrative expenses are high in certain firms but not others.

Finally, after accounting for any differences in capital costs, selling expenses, and overhead, the remaining differences in gross margins do reflect differential profitability.

FIGURE 11.4 Value Creation in the Strategy Map for Natureview Farms Yogurt

FINANCIAL VALUE

CUSTOMER

IMAGE

VALUE BRAND +

PROPOSITION COMMUNICA-

TIONS

INTERNAL PROCESS VALUE

QUALITY + AVAILABILITY + SELECTION + RELIABILITY + FUNCTIONALITY

ATTRIBUTES RELATIONSHIPS

LOWER INCREASE ENHANCE BOOST UNIT COSTS ASSET UTILIZATION PRICE PREMIUMS UNIT SALES

+

COST STRUCTURE

+

REVENUE MODEL

POST-SALE SERVICES

+ PARTNER- SHIPS

OPERATIONS MANAGEMENT

REGULATORY INITIATIVES

CULTURE OF INNOVATION

CUSTOMER SERVICE

Source: Based on “Strategy Map,” Harvard Business Review (February 2004).

5The gross margin definition can be applied to retail firms but not to service firms whose direct cost of goods sold is undefined by accountants. In services, the contribution margin definition of unit profit is prevalent, and activity-based costing (ABC) determines which b costs are variable to a product line or an account.

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 395

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Monopolists and Capacity Investments Because monopolists do not face the discipline of strong competition, they tend to install excess capacity or, alternatively, fail to install enough capacity. A monopolist that wants to restrain entry of new competitors into the industry may install excess capacity in or- der to threaten to flood the market with supply and lower prices, which makes entry less attractive. Even in regulated monopolies such as electric utility companies, considerable evidence shows that regulation often provides incentives for a firm to overinvest or un- derinvest in generating capacity. Because utilities are regulated so that they have an op- portunity to earn a “fair” rate of return on their assets, if the allowed return is greater (less) than the firm’s true cost of capital, the company will be motivated to overinvest (underinvest) in new plant and equipment.

Limit Pricing Maximizing short-run profits by setting marginal revenue equal to marginal cost in order to yield an optimal output of Q1 and an optimal price of P1 may not necessarily maxi- mize the long-run profits (or shareholder wealth) of the firm. By keeping prices high and earning monopoly profits, the dominant firm encourages potential competitors to com- mit R&D or advertising resources in an effort to obtain a share of these profits. Instead of charging the short-run profit-maximizing price, the monopolist firm may decide to engage in limit pricing, where it charges a lower price, such as PL in Figure 11.5, in order to discourage entry into the industry by potential rivals. With a limit-pricing strategy, the firm forgoes some of its short-run monopoly profits in order to maintain its monopoly position in the long run. The limit price, such as PL in Figure 11.5, was set below the minimum point on a potential competitor’s average total cost curve (ACpc). The appro- priate limit price is a function of many different factors.7

Example Components of the Gross Margin at Kellogg Co.6

As we noted in Chapter 10, Kellogg’s profit margin of 70 percent on Raisin Bran ([$4.49 Price – $1.63 DCGS]/$4.49) reflects brand loyalties built up over many years by massive and continuous advertising investments. On the leading brands, Kellogg spends 30 percent of each sales dollar on advertising, and adds another 5 percent on couponing, slot-in shelf space allowances, rebates, and other promo- tional expenses. Capital costs entail approximately 22 percent per sales dollar. Ex- penditures on headquarters, general administrative, R&D, and all other overhead total 8 percent. That leaves a net profit margin of about 5 percent.

Successful restaurants have almost twice the operating profit of convenience stores on food items sold (60 percent vs. 32 percent), and much of that differential (perhaps 25 percent) reflects net profit. Not so in Kellogg’s business where, as we have seen, most of its 70 percent gross margin goes to recover advertising, capital equipment, and other fixed costs, and only perhaps 5 percent reflects net profit. The much higher net profit in a successful restaurant is a reward for bearing a high risk of failure. The chance of long-term success in restaurants is really quite low; three out of five lose money.

6Based on “Cereals,” Winston-Salem Journal (March 8, 1995), p. A1; and “Denial in Battle Creek,” Forbes (October 7, 1996), pp. 44–46.

7The limit-pricing model illustrates the importance of potential competition as a control device on existing firms. See D. Carlton and J. Perloff, Modern Industrial Organization, 3rd ed. (New York: HarperCollins, 1999), Chapter 10, for an expanded discussion of the limit-pricing concept.

396 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The effect of the two different pricing strategies on the dominant firm’s profit stream is illustrated in Figure 11.6. By charging the (higher) short-run profit-maximizing price, the firm’s profits are likely to decline over time at a faster rate, as in Panel (a), than by charging a limit price as shown in Panel (b). The firm should engage in limit pricing if the present value of the profit stream from the limit-pricing strategy exceeds the present value of the profit stream associated with the short-run profit-maximization rule of MR = MC. Such a decision is more likely the higher the discount rate is. Choosing a high discount rate will place relatively higher weight on near-term profits in the calcula- tion of present discounted value and relatively lower weight on profits that occur further into the future. A high discount rate is justified when the firm’s long-term pricing policy, and hence profits, are subject to a high degree of risk or uncertainty. The higher the risk, the higher is the appropriate discount rate.

REGULATED MONOPOLIES Several important industries in the United States operate as regulated monopolies. In broad terms, the regulated monopoly sector of the U.S. economy includes public utilities such as electric power companies, natural gas companies, and communications companies. In the past, many of the transportation industries (airlines, trucking, rail- roads) also were regulated closely, but these industries have been substantially deregu- lated over the past 10 to 25 years.

Electric Power Companies Electric power is made available to the consumer through a production process charac- terized by three distinct stages. First, the power is generated in generating plants. Next, in the transmission stage, the power is transmitted at high voltage from the generating site to the locality where it is used. Finally, in the distribution stage, the power is distrib- uted to the individual users. The complete process may take place as part of the opera- tions of a single firm, or the producing firm may sell power at wholesale rates to a

FIGURE 11.5 Limit-Pricing Strategy

Q1

Quantity (units)

AC

MC

PL

D MR

Ppc

ACmin

QL

ACpc

0

Pr ic

e an

d co

st (

$/ un

it )

public utilities A group of firms, mostly in the electric power, natural gas, and communications industries, that are closely regulated by one or more government agencies. The agencies control entry into the business, set prices, establish product quality standards, and influence the total profits that may be earned by the firms.

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 397

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIGURE 11.6 The Effect of Pricing Strategies on Profit Streams as a Patent Expires

Period

1 2 3 4 5 …

(a) Short-run profit-maximization strategy

Period

1 2 3 4 5 …

(b) Limit-pricing strategy

Pr of

it (

$) Pr

of it

( $)

Example Using Limit Pricing to Hamper the Sales of Generic Drugs8

Patent protection is the key to financial success in the pharmaceutical industry. The typical patented drug emerges from tests on 250 chemical compounds, re- quires 15 years of research and FDA approval processes, and accumulates total costs of entry averaging $350 million.

Capoten is Bristol-Myers Squibb’s (BMS) hypertension drug for use in re- ducing heart-attack risk. Rather than limit pricing, BMS maintained Capoten’s 57-cents-per-pill price right to the end of its 20-year patent protection in February 1996. Competition from generics selling for 3 cents per pill was swift and disas- trously effective. In early 1996, BMS introduced its own generic product that can- nibalized sales of the branded product still further. By the fourth quarter of 1996, Capoten sales had collapsed to $25 million from $146 million the year before. In order to restore profitability, BMS and other leading pharmaceutical companies are merging in order to take advantage of economies of scale in R&D through follow- on drugs with improved efficacy or reduced side effects.

In contrast, Eli Lilly and Schering-Plough chose limit pricing and advertising for their leading medications, the antidepressant Prozac and the allergy treatment

(Continued)

398 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

second enterprise that carries out the distribution function. In the latter case, the distri- bution firm often is a department within the local municipal government or a consumer cooperative.

Investor-owned electric power companies are subject to regulation at several levels. In- tegrated firms that carry out all three stages of production are usually regulated by state public utility commissions. These commissions set the rates to be charged to the final con- sumers. The firms normally receive exclusive rights to serve individual localities through franchises granted by local governing bodies. As a consequence of their franchises, electric power companies have well-defined markets within which they are the sole provider of output. Finally, the Federal Energy Regulatory Commission (FERC) has the authority to set rates on power that crosses state lines and on wholesale power sales. Some states are continuing to partially or totally deregulate the power production and transmission ele- ments of this industry. The California crisis with deregulated electricity raises questions about the desirability of fully deregulated competition at the retail (distribution) level.9

Natural Gas Companies The highly regulated natural gas industry also is a three-stage process. The first stage is the production of the gas in the field. Transportation to the consuming locality through pipelines is the second stage. Distribution to the final user makes up the third stage. The FERC historically set the field price of natural gas, but regulation at the wellhead has been effectively phased out. Today, the FERC oversees the interstate transportation of gas by approving pipeline routes and by controlling the wholesale rates charged by

Claritin. Prozac lost patent protection in 2001, as did Claritin in 2003. One reason Schering-Plough chose a different (limit) pricing strategy is that Claritin had no improved follow-on drug available when the FDA demoted the prescription-only product to over-the-counter status at an identical dosage. As a consequence, $100-per-month-per-patient revenue was projected to decline to $9 if short-run profit maximizing prices continued. With a gross profit margin of 79 percent, Schering-Plough was facing a monumental loss of $2.1 billion in operating profits on $2.7 billion in Claritin sales. In such circumstances, smaller margins and a slower decline of market share could achieve higher profitability over a longer period.

In general, new biotechnologies have speeded up the emergence of imitation pharmaceuticals. Indeed, the first hypertension drug, Inderal, enjoyed almost a decade of pure monopoly sales before losing its exclusivity to Capoten in 1978. Prozac, on the other hand, met competition from imitators within four years of its 1988 introduction. And Recombinate, a breakthrough drug for hemophiliacs newly patented in 1992, encountered copycat products by 1994. Tactics such as limit pricing become all the more important in the presence of such quick and relatively easy imitation by fast-second competitors.

8Based on “Too Clever by Half,” The Economist (September 20, 1997), p. 68; “Time’s Up,” Wall Street Journal (August 12, 1997), p. A1; “Industry Merger Wave Heads to Europe,” Wall Street Journal (November 12, 1999), p. A15; and “Wearing Off: Schering-Plough Faces a Future without Claritin,” Wall Street Journal (March 22, 2002), p. A1.

9See M. Maloney, R. McCormick, and R. Sauer, Consumer Choice, Consumer Value: An Analysis of Retail Competition in America’s Electric Utility Industry (Washington, DC: Citizens for a Sound Economy, 1996); “Electric Utility Deregulation Sparks Controversy,” Harvard Business Review (May/June 1996); and A. Faruqui and K. Eakin, eds., Pricing in Competitive Electricity Markets (Boston: Kluwer, 2000).

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 399

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

pipeline companies to distribution firms. The distribution function may be carried out by a private firm or by a municipal government agency. In either event, the rates charged to final users also are subject to regulatory control

Communications Companies In the communications industry, the most important activities are radio, cable, television, and telephone service that are regulated by the Federal Communications Commission (FCC). Local service in the intrastate markets, which may be provided either by one of the former Bell System companies or by one of the so-called local telephone indepen- dents, is regulated by state commissions. Radio station ownership continues to become more concentrated; perhaps 70 percent of the stations in the top 100 markets are now controlled by two companies.

THE ECONOMIC RATIONALE FOR REGULATION As described in the preceding section, regulated industries furnish services that are criti- cal to the functioning of the economy. What are the justifications for imposing economic regulation on certain industries?

WHAT WENT RIGHT • WHAT WENT WRONG

The Public Service Company of New Mexico The Public Service Company of New Mexico (PNM) pro- vides electric power service (generation and distribution) and natural gas distribution services to most of New Mex- ico’s population. This monopoly position is regulated by the Public Service Commission of the State of New Mexico and, to a lesser extent, by the Federal Energy Regulatory Commission. These commissions determine the rates charged to various classes of customers in order to allow a “fair return” on the capital invested.

PNM’s experience in the 1990s suggests the complexity of the problems involved with rate-of-return regulation. PNM earned a 4.9 percent return on common equity dur- ing 1992, 8.0 percent on common equity during 1995, and 7.5 percent on common equity between 1997 and 1999. The industry average return on equity was 11–12 percent, according to Value Line. PNM earned extraordinarily low returns even though PNM is authorized by its regulatory commission to charge rates consistent with its earning a return of 12.5 percent on common equity. Why has this monopoly supplier of utility services (and many other util- ity companies) been unable to earn its authorized return?

PNM has experienced high growth in the demand for its services as the Sunbelt prospered and industry grew in the region. Faced with rapid growth in demand and in- creasing costs for its traditional fuel, natural gas, PNM’s managers examined a number of alternatives, including

purchasing power from nearby utilities, building large coal-fired plants close to New Mexico’s abundant coal re- sources, and building nuclear power plants. Given the de- sirability of having a diverse mix of fuel sources as a natural hedge against rising natural gas costs, PNM ulti- mately decided to participate with other regional utilities in the construction of several large coal-fired plants in the Four Corners region of northwest New Mexico, to build additional coal-fired plants of its own and to participate with other utilities in the construction of a five-unit nu- clear power plant called Palo Verde.

The first problem PNM faced was that their load growth did notmaterialize as expected. Then, the state ofNewMexico required that expensive pollution control devices, called scrubbers, be installed at the coal plants being constructed, thereby dramatically increasing their construction cost. Finally, the Palo Verde nuclear project was plagued by cost overruns, delays, and extensive and costly safety modifica- tions. When the construction program was completed, PNM found itself with capacity nearly 80 percent in excess of peak demand (a 20 percent reserve margin is more normal).

The regulatory process facing utilities does not ensure that a company will earn its authorized returns. Conse- quently, the New Mexico state regulatory commission re- fused to permit PNM to recover the costs of its excess capacity. Even in the absence of regulation, PNM would probably have been unable to fully recover the costs of this excess capacity.

400 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Natural Monopoly Argument Firms operating in the regulated sector are often natural monopolies in which a single supplier tends to emerge because of a production process characterized by massive econ- omies of scale. In other words, as all inputs are increased by a given percentage, the av- erage total cost of a unit of output decreases. Consequently, the long-run unit cost of output declines throughout the relevant range of output. This situation is illustrated in Figure 11.7 for a firm in long-run stable equilibrium.

Suppose that the market demand curve for output is represented by the curve DD in Figure 11.7. The socially optimal level of output would then be Q*; at that level of out- put, price would be well below the average total cost per unit AC* but equal to short-run and long-run marginal cost. A single producer is able to realize economies of scale that are unavailable to firms in the presence of competition. From a social perspective, com- petition would result in inefficiency in the form of higher costs such as unit cost (ACC) for the competitive firm than the unit cost (ACM) for the monopoly firm that is six times as large. The argument follows that production relations like those in Figure 11.7 will lead to the emergence of a single supplier. Competing firms will realize that their costs decrease as output expands. As a consequence, they will have an incentive to cut prices as long as MR exceeds LRMC to increase sales volume and spread the fixed cost. During this period, prices will be below average cost, resulting in losses for the producing firms. Unable to sustain such losses, the weaker firms gradually leave the industry, until only a single producer remains. Thus, competitive forces contribute to the emergence of the natural monopoly.

If a monopolistic position were to exist in the absence of regulation, the monopolist would maximize profit by equating marginal revenue and marginal cost at an output QM,

FIGURE 11.7 The Price-Output Determination of a Natural Monopoly

Q*

Quantity (units)

SRMC

AC*

D

D

SRAC

LRAC

LRMC

0

Pr ic

e an

d co

st (

$/ un

it )

QMQM/6

MR

P*=MC*

PM

ACM

ACC

Required subsidy or lump sum access fee

natural monopoly An industry in which maximum economic efficiency is obtained when the firm produces, distributes, and transmits all of the commodity or service produced in that industry. The production of natural monopolists is typically characterized by increasing returns to scale throughout the relevant range of output.

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 401

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

leading to a higher price PM and lower output. Thus, intervention through regulation is required to achieve the benefits of the most efficient organization of production. In its simplest form, this is the explanation of regulation based on the existence of natural monopolies.

Figure 11.7 illustrates one problem stemming from a genuine natural monopoly. Sup- pose that a regulatory agency succeeds in establishing the socially optimal price for out- put, P*. As the cost curves indicate, this price would lead to losses for the producing firm, because price would be below the average total cost AC*. This is obviously an un- sustainable result. In this situation the regulating agency normally sets prices at average cost to make sure revenues are sufficient to cover all costs. The most efficient way to realize revenue, however, is to charge a per-unit price equal to LRMC(P*) and collect the shaded deficit area in Figure 11.7 as a lump sum access fee, perhaps divided equally among one’s customers. Alternatively, with time-of-day metering, the lump sum access fees can depend on when the customer uses power—higher lump sum access fees charged at peak periods such as 4:00 P.M. to 8:00 P.M.

SUMMARY

� Monopoly is a market structure with significant barriers to entry in which one firm produces a dif- ferentiated product.

� In a pure monopoly market structure, firms will generally produce a lower level of output and charge a higher price than would exist in a more competitive market structure. This conclusion assumes no significant economies of scale that might make a monopolist more efficient than a large group of smaller firms.

� The primary sources of monopoly power include patents and copyrights, control of critical re- sources, government “franchise” grants, economies of scale, and increasing returns in networks of users of compatible complementary products.

� Increasing returns from network effects are limited by input cost reductions among competitors, by innovative new product introductions, and by lobbying efforts.

� Monopolists will produce at that level of output where MR = MC if their goal is to maximize short-run profits.

� The price charged by a profit-maximizing monop- olist will be in that portion of the demand function where demand is elastic (or unit elastic). The greater the elasticity of demand facing a monopo- list, the lower will be its price relative to marginal cost, ceteris paribus.

� Contribution margins are defined as revenue minus incremental variable cost, or revenue minus marginal cost when only one unit is sold.

� Contribution margins and markups are inversely related to the price elasticity of demand.

� Financial value derives from lower unit cost and better asset utilization in the cost structure as well as higher price premiums and more unit sales in the revenue model.

� A customer value proposition derives from the attribute, relationship, and image value drivers for a target customer market.

� Internal process value derives from operations management processes, customer service, innova- tion, and regulatory initiatives.

� Gross margins are defined as revenue minus direct costs of goods sold, and serve to recover capital costs, selling costs, and overhead as well as earn profits.

� Limit pricing—pricing a product below the short- run profit-maximizing level—is a strategy used by some monopolists to discourage rivals from enter- ing an industry.

� Public utilities are firms, mostly in the electric power, natural gas pipeline, and communications industries, that are closely regulated with respect to entry into the business, prices, service quality, and total profits.

402 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

� The rationales for public utility regulation are many. The natural monopoly argument is applied in cases where a product is characterized by increasing re- turns to scale. The one large firm can theoretically furnish the good or service at a lower cost than a group of smaller competitive firms. Regulators then set utility rates to prevent monopoly price gouging, ideally allowing the regulated firm to earn a return on investment just equal to its cost of capital.

� Price discrimination by utilities is often economi- cally desirable on the basis of cost justifications and demand justifications.

� Peak-load pricing is designed to charge customers a greater amount for the services they use during periods of greater demand. Long-distance phone services typically have been priced on a peak-load basis.

Exercises 1. Information Resources, Inc. (IRI), collects data on consumer packaged goods at 32,000 scanner checkout counters and in panel surveys of 70,000 households. IRI records indicate that department store-brand pantyhose sell for a gross margin of 43 percent and a contribution margin of 29 percent, and the store inventory turns over 14 times per year.

a. What expenses explain the difference between 43 percent and 29 percent? b. What percentage change in unit sales is required to increase total contribu-

tions if price is cut by 10 percent? c. Compare store-brand pantyhose with the products in Table 11.1. Why

should Whitman’s Sampler sell for a contribution margin of 54 percent when pantyhose sell for 29 percent?

2. Ajax Cleaning Products is a medium-sized firm operating in an industry domi- nated by one large firm—Tile King. Ajax produces a multiheaded tunnel wall scrubber that is similar to a model produced by Tile King. Ajax decides to charge the same price as Tile King to avoid the possibility of a price war. The price charged by Tile King is $20,000.

Ajax has the following short-run cost curve:

TC = 800,000 − 5,000Q + 100Q2

a. Compute the marginal cost curve for Ajax. b. Given Ajax’s pricing strategy, what is the marginal revenue function for

Ajax? c. Compute the profit-maximizing level of output for Ajax. d. Compute Ajax’s total dollar profits.

3. The Lumins Lamp Company, a producer of old-style oil lamps, estimated the following demand function for its product:

Q = 120,000 − 10,000P

where Q is the quantity demanded per year and P is the price per lamp. The firm’s fixed costs are $12,000 and variable costs are $1.50 per lamp.

a. Write an equation for the total revenue (TR) function in terms of Q. b. Specify the marginal revenue function. c. Write an equation for the total cost (TC) function in terms of Q. d. Specify the marginal cost function.

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 403

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

e. Write an equation for total profits (π) in terms of Q. At what level of output (Q) are total profits maximized? What price will be charged? What are total profits at this output level?

f. Check your answers in Part (e) by equating the marginal revenue and mar- ginal cost functions, determined in Parts (b) and (d), and solving for Q.

g. What model of market pricing behavior has been assumed in this problem?

4. Unique Creations holds a monopoly position in the production and sale of mag- nometers. The cost function facing Unique is estimated to be

TC = $100,000 + 20Q

a. What is the marginal cost for Unique? b. If the price elasticity of demand for Unique is currently –1.5, what price

should Unique charge? c. What is the marginal revenue at the price computed in Part (b)? d. If a competitor develops a substitute for the magnometer and the price

elasticity increases to –3.0, what price should Unique charge?

5. Exotic Metals, Inc., a leading manufacturer of beryllium, which is used in many electronic products, estimates the following demand schedule for its product:

PRICE ($/POUND)

QUANTITY (POUNDS/PERIOD)

$25 0

18 1,000

16 2,000

14 3,000

12 4,000

10 5,000

8 6,000

6 7,000

4 8,000

2 9,000

Fixed costs of manufacturing beryllium are $14,000 per period. The firm’s variable cost schedule is as follows:

OUTPUT (POUNDS/PERIOD)

VARIABLE COST (PER POUND)

0 $ 0

1,000 10.00

2,000 8.50

3,000 7.33

4,000 6.25

5,000 5.40

6,000 5.00

7,000 5.14

8,000 5.88

9,000 7.00

404 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

a. Find the total revenue and marginal revenue schedules for the firm. b. Determine the average total cost and marginal cost schedules for the firm. c. What are Exotic Metals’ profit-maximizing price and output levels for the

production and sale of beryllium? d. What is Exotic’s profit (or loss) at the solution determined in Part (c)? e. Suppose that the federal government announces it will sell beryllium, from

its extensive wartime stockpile, to anyone who wants it at $6 per pound. How does this affect the solution determined in Part (c)? What is Exotic Metals’ profit (or loss) under these conditions?

6. Wyandotte Chemical Company sells various chemicals to the automobile indus- try. Wyandotte currently sells 30,000 gallons of polyol per year at an average price of $15 per gallon. Fixed costs of manufacturing polyol are $90,000 per year and total variable costs equal $180,000. The operations research department has esti- mated that a 15 percent increase in output would not affect fixed costs but would reduce average variable costs by 60 cents per gallon. The marketing department has estimated the arc elasticity of demand for polyol to be –2.0.

a. How much would Wyandotte have to reduce the price of polyol to achieve a 15 percent increase in the quantity sold?

b. Evaluate the impact of such a price cut on (i) total revenue, (ii) total costs, and (iii) total profits.

7. Tennis Products, Inc., produces three models of high-quality tennis rackets. The following table contains recent information on the sales, costs, and profitability of the three models:

MODEL

AVERAGE QUANTITY

SOLD (UNITS/ MONTH)

CURRENT PRICE

TOTAL REVENUE

VARIABLE COST PER

UNIT

CONTRIBUTION MARGIN PER

UNIT CONTRIBUTION

MARGIN*

A 15,000 $30 $ 450,000 $15.00 $15 $225,000

B 5,000 35 175,000 18.00 17 85,000

C 10,000 45 450,000 20.00 25 250,000

Total $1,075,000 $560,000

*Contribution to fixed costs and profits.

The company is considering lowering the price of Model A to $27 in an effort to increase the number of units sold. Based on the results of price changes that have been instituted in the past, Tennis Products’ chief economist estimates the arc price elasticity of demand to be –2.5. Furthermore, she estimates the arc cross elasticity of demand between Model A and Model B to be approximately 0.5 and between Model A and Model C to be approximately 0.2. Variable costs per unit are not expected to change over the anticipated changes in volume.

a. Evaluate the impact of the price cut on the (i) total revenue and (ii) contri- bution margin of Model A. Based on this analysis, should the firm lower the price of Model A?

b. Evaluate the impact of the price cut on the (i) total revenue and (ii) contri- bution margin for the entire line of tennis rackets. Based on this analysis, should the firm lower the price of Model A?

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 405

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

8. The Public Service Company of the Southwest is regulated by an elected state utility commission. The firm has total assets of $500,000. The demand function for its services has been estimated as

P = $250 − $0.15Q

The firm faces the following total cost function:

TC = $25,000 + $10Q

(The total cost function does not include the firm’s cost of capital.) a. In an unregulated environment, what price would this firm charge, what

output would be produced, what would total profits be, and what rate of return would the firm earn on its asset base?

b. The firm has proposed charging a price of $100 for each unit of output. If this price is charged, what will be the total profits and the rate of return earned on the firm’s asset base?

c. The commission has ordered the firm to charge a price that will provide the firm with no more than a 10 percent return on its assets. What price should the firm charge, what output will be produced, and what dollar level of profits will be earned?

9. The Odessa Independent Phone Company (OIPC) is currently engaged in a rate case that will set rates for its Midland-Odessa area customer base. OIPC has total assets of $20 million. The Texas Public Utility Commission has determined that an 11 percent return on assets is fair. OIPC has estimated its annual demand function as follows:

P = 3,514 − 0.08Q

Its total cost function (not including the cost of capital) is

TC = 2,300,000 + 130Q

a. OIPC has proposed a rate of $250 per year for each customer. If this rate is approved, what return on assets will OIPC earn?

b. What rate can OIPC charge if the commission wants to limit the return on assets to 11 percent?

c. What problem of utility regulation does this exercise illustrate?

Case Exercise DIFFERENTIAL PRICING OF

PHARMACEUTICALS: THE HIV/AIDS CRISIS10 The HIV/AIDS crisis has been called the worst pandemic since the fourteenth- century’s Black Plague. The first incident of HIV/AIDS was discovered by the U.S. Centers for Disease Control in 1981. Over the next three decades, 60 million people

10E. Berndt, “Pharmaceuticals in U.S. Health Care: Determinants of Quality and Price,” and M. Kremer, “Pharmaceuticals and the Developing World,” Journal of Economic Perspectives (Fall 2002), pp. 45–90.

406 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

have become infected and 25 million have died. Most HIV/AIDS cases are reported in the developing world, where 95 percent of those with HIV live today. Beyond social welfare and humanitarian concerns, as a result of globalization and the fastest grow- ing international business opportunities in China and India, AIDS is now everybody’s business. Because the pharmaceutical industry especially relies upon governmental au- thority to approve formularies for reimbursement, to protect its monopoly patent rights, and to prevent importation of unauthorized, unlicensed imitation medicines, the question of how to price AIDS drugs is a public issue.

Although no one has yet developed a cure for HIV, a number of companies have patented drugs that inhibit either the virus’s ability to replicate or its ability to enter host cells. Without further drug discovery, however, the best that can be done at present once a person contracts HIV is to partially and temporarily suppress the virus, thus delaying progression of the infection. The drugs that suppress HIV are called antiretrovirals, and the first, known as Retrovir (also known by its generic name zidovudine or AZT), was introduced in 1987 by Burroughs Wellcome (now GlaxoSmithKline) and was the only approved therapy available to treat HIV until 1991. Since then, several new antiretrovirals have been developed by large pharmaceu- tical companies such as Abbott Labs, Bristol-Myers Squibb, Merck, Roche, and smal- ler biotech companies such as Agouron, Gilead Sciences, Triangle Pharmaceuticals, and Trimeris. Largely as a result of these drugs, the rate of increase of AIDS-related diseases (e.g., opportunistic infections) dramatically slowed in the United States from 1992–1995 and actually decreased in 1996 for the first time.

Yet, even in the early days of antiretroviral drug development, HIV/AIDS drug pricing was a serious and contentious issue. The core problem is the fact that most HIV/AIDS cases are outside what the United Nations classifies as “rich countries” such as the United States. North America registered about 1.4 million cases of indivi- duals living with HIV/AIDS and fewer than 25,000 deaths due to AIDS in 2008, but the comparative numbers for sub-Saharan Africa were 22 million cases and more than 1.9 million deaths. Similarly, the U.S. adult infection rate was estimated at slightly less than one half of a percent in 2008 versus over 5 percent in sub-Saharan Africa, where GDPs per capita often are less than $1,000 versus $30,000 in the United States. Com- pounding the problem is the fact that many new AIDS drugs, especially those designed to attack the growth in drug-resistant HIV, grow ever more expensive. Trimeris and Roche introduced Fuzeon in early 2003, for example, at a wholesale price of €20,245 per annum, at least three times the price of any existing HIV/AIDS drug.

The pricing decision reflects the fiscal realities of their expensive R&D-intensive business model against enlarged, global, corporate social responsibilities. A nation- state-specific pricing policy across global markets has resulted in a tenfold differential between the highest priced market, the United States, and the price charged in the poorest countries. Glaxo and Roche management teams face many serious business ethics issues in this highly charged environment. Is such a tenfold price differential sustainable? How does one manage the resulting problem of parallel importing—that is, the unauthorized reimportation of export drugs bought at lower price points else- where in the world? Will abrogation of the intellectual property rights of the drug companies in the developing world threaten intellectual property protection at home? Will a public affairs backlash in high-priced markets force drug price dis- counts? If so, how can the massive R&D investment required for ongoing drug dis- covery and development be recovered? Are these companies facing such a public

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 407

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

relations disaster that their corporate brand equity could be radically affected? What are big pharmaceutical companies’ corporate responsibilities in a public health crisis? Should Glaxo (or Roche) go it alone, or instead pursue collaborative strategies with other big pharmaceutical rivals?

Questions 1. Is the monopoly on patented pharmaceuticals warranted? What barrier to entry

prevents the re-importation into the United States of pharmaceuticals sold at lower prices abroad (say, in Canada)?

2. The contribution margin percentage on pharmaceuticals exceeds the 55 percent to 70 percent margins on ready-to-eat cereals. Identify three reasons why phar- maceutical margins are higher.

3. Suggest an approach to the big pharmaceutical company problem of differential pricing in the United States, Western Europe, and Japan versus the less-developed world.

408 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

12 CHAP T E R

PriceandOutputDetermination: Oligopoly CHAPTER PREVIEW The previous two chapters analyzed price and output decisions of firms that competed in markets with either a large number of sellers (i.e., pure competition and monopolistic competition) or essentially no other sellers (i.e., monopoly). In pure competition, the firm made its price and output decisions independently of the decisions of other firms because no single firm was large enough to affect the market price. Similarly, the monopoly firm did not need to consider the pricing actions of rival firms, because it had no competitors. This chapter, in contrast, examines price and output decisions by firms in oligopoly market structures with a small number of competitors, where each firm’s decisions are likely to evoke a response from one (or more) of these rivals. To maximize shareholder wealth, each oligopoly firm must take into account these rival responses in its own decision making. In the next chapter, game theory analysis is introduced to help you predict how your rivals will respond.

MANAGERIAL CHALLENGE Are Nokia’s Margins on Cell Phones Collapsing?1

From a stodgy Finnish industrial conglomerate selling everything from rubber boots and wire cable to toilet pa- per and televisions, Nokia transformed itself into a relent- lessly focused technology company. When Sweden’s telecommunications-equipment giant Ericsson devel- oped a cellular network across Scandinavia in the 1980s, Nokia provided the wireless but bulky radio telephones. Nokia recognized this situation as a strategic opportunity, so it spun off other business in the 1990s and focused its attention on the enormous market potential of a digital cell phone. Nokia grew from 22 percent market share in 1985 (half of Motorola’s 45 percent) to overtake the market leader in 1998. By 2008, 39 percent of the $79 billion in cell phone sales worldwide belonged to Nokia,

relative to Motorola’s 14 percent. Other major suppliers include Samsung (14 percent), Sony-Ericsson (9 percent), and LG Electronics (7 percent). The 1.6 billion cell phones in service exceed the number of land lines. With huge scale economies and a snazzy branded product, No- kia’s cell phone margins, at 23 percent, outstrip Motoro- la’s paltry 3 percent, but several reasons suggest Nokia’s margins may not be sustainable.

First, Third Generation (3G) high-speed worldwide wireless networks have rewritten the telecommunications landscape. Nokia’s European partners went deeply into debt to pay $125 billion for 3G licenses and spent another $100 billion for 3G network equipment. Nokia, NEC, and Panasonic first offered innovative 3G-ready cell phones

409

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

with a built-in digital camera. But 3G technology allows the introduction of much enhanced wireless Web prod- ucts into the handset marketplace: handheld computers by Dell; pocket audiovisual terminals by Apple’s iPhone, Palm, and BlackBerry; and game consoles from Sony- Ericsson.

Second, these enhanced mobile handsets create value principally through their software applications provided

by third-party independent software vendors (ISVs). And the ISVs will want their share of the gross margins that have made Nokia so profitable. The power of these ISV suppliers was virtually nonexistent in the voice-only mo- bile phone business, but now the success of a smart phone is all about hundreds of touch screen applications.

Third, Europe is nearly saturated with wireless hand- sets, which have achieved 83 percent penetration. Al- though Chinese cell phone demand is soaring, local vendors control distribution in those markets, and price wars are commonplace. In North America, Nokia is only now starting to form alliances with the service pro- viders like AT&T and Verizon Wireless, while Motorola dominates the market with 35 percent market share and Nokia is fourth (at 10 percent) behind Samsung (18 percent) and LG Electronics (16 percent). Consequently, Nokia demand is projected to grow at only 7 percent over the next few years, rather than the 12–20 percent of the past.

Fourth, the threat of entry is real. Japanese consumer electronics manufacturers NEC and Panasonic first adapted the 3G technology into the wireless Internet devices. Moreover, Apple iPhones are the rage across North America. High-end camera phones from Sharp and Samsung are the hottest sellers across Asia where new design features, color screens, and slide bars change as often as the whims of a fad-driven marketplace. Cycle times for new products are down to six months versus two years at the start of the millennium.

TOP TEN MOBILE COMPANIES

2006 WORLDWIDE SUBSCRIBERS (MILLIONS)COUNTRY

China Mobile China 291

Vodafone Britain 199

China Unicom China 142

Telefónica Spain 126

América Móvil Mexico 117

Deutsche Telekom Germany 99

France Télécom France 98

AT&T United States 64

Telenor Norway 68

Telecom Italia Italy 60

Source: Cowen & Co.

MANAGERIAL CHALLENGE Continued

© Cr ea ta s/ Cr ea ta s Im ag es /J up ite r Im ag es

410 Part 4: Pricing and Output Decisions: Strategy and Tactics

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

OLIGOPOLISTIC MARKET STRUCTURES An oligopoly is characterized by a relatively small number of firms offering a similar product or service. The product or service may be branded, as in soft drinks, cereals, and athletic shoes, or unbranded, as in crude oil, aluminum, and cement. The main dis- tinction of oligopoly is that the number of firms is small enough that actions by any in- dividual firm on price, output, product style or quality, introduction of new models, and terms of sale have a perceptible impact on the sales of other firms in the industry. In this easily recognizable interdependence, each firm knows that any new move, such as intro- ducing a price cut or launching a large promotional campaign, is likely to evoke a coun- termove from its rivals.

In all oligopoly markets, rival response expectations are therefore the key to firm-level analysis. If rival firms are expected to match price increases and price cuts as in airlines, a share-of-the-market demand curve may adequately illustrate the sales response to the pricing initiatives of one firm (such as Southwest Airlines, 20 percent share-of-the-market demand); see Figure 12.1, Panels (a) and (b). In other markets, if rival firms are slow to match price increases and cuts, oligopolists can discount to gain share and will lose share in response to price hikes. In still other markets such as I-beam steel, rivals match price cuts but ignore price increases. Consequently, Nucor faces a much more price elastic de- mand above the going equilibrium price than the share-of-the-market demand below that price. These asymmetric rival response expectations lead to kinked oligopoly firm demand schedules discussed later in the chapter and illustrated in Figure 12.1, Panel (c).

Oligopoly in the United States: Relative Market Shares Much of U.S. industry is best classified as oligopolistic in structure with a wide range of industry configurations. At one extreme are dominant single firms in the markets for shaving razors, hand calculators, game consoles, cigarettes, digital printers, beer, athletic shoes, and smart phones, where Gillette (80 percent), TI (78 percent), Nintendo (65 per- cent), Altria (67 percent), Hewlett-Packard (49 percent), Anheuser-Busch (55 percent), Nike (43 percent), and Nokia (41 percent) are all several times larger than their next largest competitors (see Table 12.1).

Discussion Questions

� Should Nokia invest heavily in product design modifications to satisfy the requirements of vendors like AT&T and Verizon Wireless?

� Should Nokia remain a premium cell phone supplier or instead focus on the projected growth in the low end of the camera phone market, especially in China and Latin America where China Mobile and China Unicom have 334 million subscribers and América Móvil has 117 in Mexico compared to Vodaphone’s

199 million in Britain, Telefŏnica’s 126 million in Spain, France Télécom’s and Deutsche Tel- ekcom’s 98 million each in Germany and France, and AT&T’s 64 million in North America?

� What else could Nokia do to grow its future cash flows?

1Based on “Nokia: A Finnish Tale,” The Economist (October 14, 2000), pp. 83–85; “Cellphone Squeeze Play,” Wall Street Journal (November 17, 2005), p. B1; “Special Report: Mobile Telephones,” The Economist (May 1, 2004), pp. 71–76; and “Nokia Moves to Regain Market Share,” Wall Street Journal (May 27, 2008), pp. B1–2.

MANAGERIAL CHALLENGE Continued

Chapter 12: Price and Output Determination: Oligopoly 411

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

In crackers, appliances, biotech, soft drinks, batteries, aircraft, and search engines on the Internet, not one but two firms dominate (see Table 12.1). In snack foods, Nabisco’s 45 percent market share and Keebler’s 22 percent overshadow Pepperidge Farms’ 7 per- cent. Similarly, Coke and Pepsi dominate the soft drink market, Sears and Lowe’s Home Improvement dominate the home appliances market, Boeing and Airbus dominate the wide-bodied aircraft market, and Duracell and Energizer dominate the battery market. These duopoly pairs of dominant firms often study complex tactical scenarios of moves and probable countermoves against each other. In still other cases, three firms circle war- ily, planning their tactical initiatives and retreats: tires (Goodyear 28 percent, Michelin 23 percent, Bridgestone/Firestone 21 percent); tea (Lipton 37 percent, Arizona 26 per- cent, Nestea 16 percent); textbooks (Pearson 27 percent, Cengage 22 percent, McGraw- Hill 13 percent); cereals (Kellogg 30 percent, General Mills 30 percent, Post 13 percent);

FIGURE 12.1 Rival Response Expectations Determine Firm Demand

Pr ic

e ($

)

D 20%

0.2 Q m Q mOutput

Panel (a) Panel (b) Panel (c)

Peq

Southwest Airlines share-of-market

Airlines

D

Output

Market

D

Output

Nucor

Steel I-Beams

Example Hewlett-Packard’s Dominance in Printers2

At 49 percent market share, sales of HP printers are five times larger than their closest rivals Xerox and Lexmark, which have 10 percent and 8 percent of the mar- ket, respectively. The HP business plan for printers calls for a razor-and-blades ap- proach of relatively inexpensive machines followed by a long period of selling lucrative ink and toner replacement cartridges. In 2001, HP printers and supplies made $410 million in operating profits on $5 billion in sales revenue. This repre- sents two-thirds of HP’s $647 million overall profit on only one-tenth of the $49 billion in sales. Despite vicious price wars for market share in the sub-$100 and sub-$200 segments, the printer business has clearly been a cash cow for HP. On the horizon for high-end products, HP plans to launch a digital printing press to replace the plates and film required today for commercial offset printing. In the mass market targeted by Lexmark, Canon, and Epson, penetration of PCs into American households and businesses has reached a plateau, but printing volume (and therefore demand for HP supplies) may continue to grow because of the printing of digital photographs and Web pages.

2Based on “HP Sees Room for Growth in Printer Market,” Wall Street Journal (June 28, 2001), p. B10.

412 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE 12.1 LARGEST U.S. MARKET SHARES IN OLIGOPOLISTIC INDUSTRIES

DOMINANT SINGLE FIRMS

Razors and Blades (2004)

Gillette 80%

Schick 17

Bic 3

Hand Calculators (2009)

Texas Instruments 78%

Casio 14

Game Consoles (2008)

Nintendo 65%

Sony 20

Microsoft 18

Tobacco (2004)

Altria 67%

R.J. Reynolds 13

Lorillard 11

Brown and Williamson 9

Digital Printers (2001)

Hewlett-Packard 49%

Xerox 10

Lexmark 8

U.S. Beer (2009)

Anheuser-Busch 55%

Miller 17

Coors 9

Athletic Shoes (2008)

Nike 43%

Adidas 15

Reebok 10

Smart Phones (2008)

Nokia 41%

RIM 20

Apple 14

Palm 4

DUOPOLY FIRMS

Crackers (1998)

Nabisco 45%

Keebler 22

Pepperidge Farm 7

Appliances (2008)

Sears 32%

Lowe’s 20

Home Depot 9

Best Buy 7

Biotech (2008)

Amgen 21%

Roche 20

Johnson & Johnson 8

Soft Drinks (2003)

Coca-Cola 44%

PepsiCo 32

Cadbury Schweppes 16

Batteries (2005)

Duracell 43%

Energizer 33

Rayovac 11

Wide-Body Aircraft (2006)

Boeing 50%

Airbus 49

TRIOPOLY FIRMS

Tires (2003)

Goodyear/Sumitomo 28%

Michelin 23

Bridgestone/Firestone 21

Tea (2007)

Lipton 37%

Arizona 26

Nestea 16

Textbooks (2002)

Pearson 27%

Thomson 22

McGraw-Hill 13

Rental Cars (2007)

Avis/Budget 30%

Hertz 28

Enterprise 27

LESS CONCENTRATED

Confectionary (2009)

Mars 35%

Cadbury 24

Nestle 18

Hershey 12

Kraft 11

U.S. Autos (2009)

General Motors 22%

Toyota 20

Ford 17

Honda 15

Nissan 10

Trucks (2001)

Freightliner 30%

International 17

Mack 13

Peterbuilt 12

Kenworth 11

Volvo Truck 10

Music Recording (2001)

Universal/Polygram 23%

Sony 15

EMI 13

Warner 12

BMG 8

U.S. Laptop Computers (2008)

Dell 26%

HP 25

Apple 12

Toshiba 11

Acer 10

Cell Phones (2007)

Nokia 35%

Motorola 22

Samsung 14

Sony/Ericsson 9

LG 8

U.S. Wireless (2009)

Verizon Wireless 34%

AT&T Wireless 31

Sprint/Nextel 19

T-Mobile 13

Pharmaceuticals (2008)

Pfizer/Wyeth 26%

GlaxoSmithKline 16

Novartis 15

Merck/Schering 15

Roche 15

Source: Industry Surveys, Net Advantage Database, Standard & Poor’s; and Market Share Reports, Gale Research, annual issues.

Chapter 12: Price and Output Determination: Oligopoly 413

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

rental cars (Avis 30 percent, Hertz 28 percent, Enterprise 27 percent); and candy and gum (Mars 35 percent, Cadbury 24 percent, and Nestle 18 percent).

The market share distributions in Table 12.1 are seldom static. Instead, the dynamics of the share distribution often tell important insights. In cereals, General Mills’ new product introductions continued to take share points from Kellogg during 1993–1999. But it was Post/Nabisco that was the big loser to private-label discount cereals (e.g., Kroger Raisin Bran) and their contract supplier Ralston/Purina (see Table 12.2). In wide-bodied aircraft, Boeing ceded market share to Airbus, and lowered its final assem- bly rate of production, removing bottlenecks and making itself much more profitable.

Over the 14-year period 1992–2005, the rank order of the leading airlines was largely unchanged, but every one of the major hub-and-spoke carriers lost two to three share points to the point-to-point discounters Southwest and America West (see the data in Table 12.2). Although no airline has a dominant market share nationally, a number of airlines have dominant positions at various airports around the country. For example, American has a 65 percent share at Dallas/Fort Worth, Northwest has an 84 percent share at Minneapolis/St. Paul, and US Airways has a 93 percent share at Charlotte.

Finally, Table 12.1 shows several industries in which the share distributions are less con- centrated but where the strong interdependencies between leading firms remain prominent in each firm’s business planning. Sales in the U.S. auto and truck markets are dispersed

TABLE 12.2 MARKET SHARE DISTRIBUTIONS OVER TIME IN AIRLINES, CEREALS, AND

WIDE-BODIED AIRCRAFT

AIRLINES CEREALS WIDE-BODIED

AIRCRAFT

1992 2005 1993 1999 1998 2005

American 21% American 19% Kellogg 35% Kellogg 30% Boeing 70% Boeing 51%

United 20 United 17 General Mills 25 General Mills 30 Airbus 30 Airbus 49

Delta 15 Delta 15 Post/Nabisco 18 Post/Nabisco 13

Northwest 14 Northwest 11 Quaker 8 Private Label 11

Continental 11 Continental 9 Private Label 6 Ralston 7

US Airways 9 Southwest 7 Ralston 5 Quaker 6

Sources: Wall Street Journal (December 21, 2001), p. A8; (December 27, 1996), p. A3; (October 16, 1998), p. B4; and (July 14, 2006), p. B1.

Example Auto Rental and Retail Gasoline Firms Consolidate: Enterprise Rent-A-Car, and Exxon/Mobil In auto rentals and retail gasoline, consolidation has occurred. The market shares of individual rental car companies have been very stable for the past decade, but Avis bought Budget, thereby reaching a 30 percent market share compared to Hertz’s 29 percent. Then Enterprise combined with National and Alamo to achieve near parity (27 percent) with Avis and Hertz. The top three auto rental companies had 69 percent of the relevant market in 2000, whereas by 2008, the top three firms had 86 percent (see Figure 12.2). Massive consolidation also occurred in the retail gasoline industry, where company after company sought a large partner with whom to merge. Scale economies in exploration and development, as well as the closing of redundant gas stations, drove the trend.

414 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

across five or six companies. And in three industries heavily influenced by the disruptive technology of Internet computing (namely, music recording, laptop computers, and basic cell phones), the forces of competition have dispersed the shares across five firms. Similarly in wireless operators and pharmaceuticals, market shares are more dispersed as the huge investments required to dominate the market are potentially unrecoverable.

INTERDEPENDENCIES IN OLIGOPOLISTIC INDUSTRIES The nature of interdependencies in these oligopolistic industries can be illustrated using an airline pricing example.

The Cournot Model One standard approach to the interdependency problem among oligopolists is merely to ignore it—that is, for a firm to assume that its competitors will act as if it does not exist.

FIGURE 12.2 The Relative Sizes of Competitors in the Auto Rental and Gasoline Industries

Rental Cars (market share)

Hertz 30% 29% 27% 12%

2000 2008

Retail Gasoline (market share)

Shell Exxon/Mobil 24% Shell 20% BP/Amoco/Arco 18% Chevron/Texaco 16% Total/Fina/Elf 10% Conoco/Phillips 7%

1992 2001

29%

Dollar Thifty

Dollar/Thifty

7% 4%

Enterprise

Enterprise/National/Alamo

8% Budget 10% National/Alamo 20% Avis

Avis/Budget 20%

Phillips 4% Sun 4% Marathon 5% Citgo 5% BP 6% Mobil 7% Amoco 7% Exxon 8% Texaco 8% Chevron 8%

9%

Hertz

Source: Wall Street Journal (December 21, 2001), p. A8, B6; (November 1, 2005), p. A2.

Chapter 12: Price and Output Determination: Oligopoly 415

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Because of the wide scope of oligopoly industry configurations in Table 12.1, several simplifying models have been used to describe oligopolists’ competitive behavior regard- ing price, output, and other market conditions. The Cournot oligopoly model, proposed by the French economist Augustin Cournot, asserts that each firm, in determining its profit-maximizing output level, assumes that the other firm’s output will not change.

For example, suppose that two duopolists (Firms A and B) produce identical prod- ucts. If Firm A observes Firm B producing QB units of output in the current period, then Firm A will seek to maximize its own profits assuming that Firm B will continue producing the same QB units in the next period. Firm B acts in a similar manner. It at- tempts to maximize its own profits under the assumption that Firm A will continue pro- ducing the same amount of output in the next period as Firm A did in the current period. In the Cournot model, this pattern continues until reaching the long-run equilib- rium point where output and price are stable and neither firm can increase its profits by raising or lowering output. The following example illustrates the determination of the long-run Cournot equilibrium.

Example Airline Pricing: The Pittsburgh Market Consider the case of the airline route between Pittsburgh and Dallas. One can fly this route on a number of different airlines, but only American and US Airways offer nonstop service between these cities. Initially, both airlines were charging $1,054 for a round-trip coach-class ticket. American then introduced a discounted fare for only $640, a reduction of $414. US Airways was then faced with the deci- sion of whether to maintain its current $1,054 fare, match American’s new $640 fare, or undercut American’s $640 fare. American’s demand function (and reve- nues) in the Pittsburgh-Dallas market depended on the reaction of US Airways to the fare reduction. A decision by US Airways to charge a higher fare (e.g., the cur- rent $1,054 fare) would result in additional market share for American, because many travelers would choose American’s lower-priced service. A decision by US Airways to match American’s new fare would result in American retaining its ex- isting market share on the Pittsburgh-Dallas route. However, depending on the price elasticity of demand and the mix of full-fare and discounted tickets sold, the price reduction could actually increase American’s revenues and profits. Finally, a decision by US Airways to undercut American’s new $640 fare would lead to a lower market share and a likely further price reduction by American.

Example The Cournot Oligopoly Solution: Siemens and Lucent- Alcatel Suppose that two European electronics companies, Siemens (Firm S) and Lucent- Alcatel (Firm T), jointly hold a patent on a component used in airport radar sys- tems. Demand for the component is given by the following function:

p = 1,000 − QS − QT [12.1]

(Continued)

416 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

CARTELS AND OTHER FORMS OF COLLUSION Oligopolists sometimes reduce the inherent risk of being so interdependent by either for- mally or informally agreeing to cooperate or collude in decision making. Collusive agree- ments between oligopolists are called cartels. In general, collusive agreements are illegal

where QS and QT are the quantities sold by the respective firms and P is the (mar- ket) selling price. The total cost functions of manufacturing and selling the compo- nent for the respective firms are

TCS = 70,000 + 5QS + 0:25 2 S [12.2]

TCT = 110,000 + 5QT + 0:15 2 T [12.3]

Suppose that the two firms act independently, with each firm seeking to maximize its own total profit from the sale of the component.

Siemens’s total profit is equal to

πS = PQS − TCS = ð1,000 − QS − QTÞQS − ð70,000 + 5QS + 0:25Q2SÞ = − 70,000 + 995QS − QTQS − 1:25Q2S [12.4]

Note that Siemens’s total profit depends on the amount of output produced and sold by Lucent-Alcatel (QT). Taking the partial derivative of Equation 12.4 with respect to QS yields

∂πS ∂QS

= 995 − QT − 2:50QS [12.5]

Similarly, Lucent Alcatel’s total profit is equal to

πT = PQT − TCT = ð1,000 − QS − QTÞQT − ð110,000 + 5QT + 0:15Q2TÞ = −110,000 + 995QT − QSQT − 1:15Q2T

[12.6]

Note also that Lucent-Alcatel’s total profit is a function of Siemens’s output level (QS). Taking the partial derivative of Equation 12.6 with respect to QT yields

∂πT ∂QT

= 995 − QS − 2:30QT [12.7]

Setting Equations 12.5 and 12.7 equal to zero yields

2:50QS + QT = 995 [12.8] QS + 2:30QT = 995 [12.9]

Solving Equations 12.8 and 12.9 simultaneously gives the optimal levels of out- put for the two firms: Q*S = 272.32 units and Q*T = 314.21 units. By substituting these values into Equation 12.1, we calculate an optimal (equilibrium) selling price of P* = $413.47 per unit. The respective profits for the two firms are obtained by substituting Q*S and Q*S into Equations 12.4 and 12.6 to obtain π*S = $22,695.00 and π*T = $3,536.17.

cartels A formal or informal agreement among firms in an oligopolistic industry that influences such issues as prices, total industry output, market shares, and the division of profits.

Chapter 12: Price and Output Determination: Oligopoly 417

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

in the United States and Europe; however, some important exceptions exist. For example, prices and quotas of various agricultural products (e.g., milk, oranges) are set by grower cooperatives in many parts of the country with the approval of the federal government. The International Air Transport Association (IATA) airlines flying transoceanic routes jointly set uniform prices for these flights. And ocean shipping rates are set by hundreds of collusive “conferences” on each major transoceanic route.

In addition, illegal collusive arrangements also arise from time to time. For example, cement and paving companies, as well as cardboard box manufacturers, often are in- dicted for price fixing. In a 2008 case, South Korean electronics giant LG Electronics paid a $400 million fine (the second-largest antitrust fine ever) for conspiring with Japa- nese Sharp ($129 million) and Taiwanese Chunghwa Display to fix the wholesale price of LCD monitors in laptops, cell phones, and televisions.3 The grain-processing giant Archer-Daniels-Midland (ADM) pled guilty in 1996 to organizing an explicit quota and pricing system among five firms in the lysine market (see Figure 12.3); lysine is an amino acid food supplement that speeds the growth of livestock. ADM paid $100 million in antitrust penalties, and ADM executives went to jail.4 Roche and BASF, large Swiss and German industrial conglomerates in pharmaceuticals, chemicals, fragrances, and vita- mins, agreed to pay $500 million and $225 million fines, respectively, to the U.S. Justice Department for their leadership of a price-fixing conspiracy in vitamin supplements. This 1999 antitrust settlement reduced Roche’s profitability by 30 percent.5 Worldwide the fines arising from the vitamin price-fixing conspiracy totaled $1.6 billion. These se- vere penalties indicate how serious the inefficiencies arising from cartelization of an in- dustry can be. Businesses are wise not to ignore the prohibition against price fixing.

FIGURE 12.3 Lysine Manufacturers Who Pled Guilty to Price Fixing

1995 World Market Share

Ajinomoto (Japan) 34.0% Archer-Daniels-

Midland (U.S.) 26.4%

Kyowa Hakko Kogyo (Japan)

18.1%

Sewon (South Korea)

14.2%

Cheil Jedang (South Korea)

6.3%

Source: Wall Street Journal (July 9, 1998).

3“3 LCD-Makers Plead Guilty to Price Fixing,” Wall Street Journal (July 7, 2008), p. B1. 4“In ADM Saga, Executives Now on Trial,” Wall Street Journal (July 9, 1998), p. B10. 5“Scandal Costs Roche,” Wall Street Journal (May 25, 1999), p. A20.

418 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Factors Affecting the Likelihood of Successful Collusion The ability of oligopolistic firms to engage successfully in collusion depends on a number of factors:

Number and Size Distribution of Sellers Effective collusion generally is less difficult as the number of oligopolistic firms involved decreases. In the 1990s, the De Beers diamond cartel in Switzerland and South Africa was effective in part because Russia agreed in 1995 to sell 95 percent of its total wholesale supply through De Beers. De Beers’s central selling organization and Russia together accounted for more than 75 percent of world supply at that time.

Product Heterogeneity Products that are alike in their characteristics are said to be homogeneous, and price is the only distinction that matters. When products are hetero- geneous (or differentiated), cooperation is more difficult because competition is occurring over a broader array of product characteristics, such as durability, fashion timing, war- ranty, and after-sale policies.

Cost Structures The more cost functions differ among competing firms, the more difficult it will be for firms to collude on pricing and output decisions. Also, successful collusion is more difficult in industries where fixed costs are a large part of total costs. A higher percentage of fixed costs implies higher contribution margins to recover those fixed costs. And, as we saw in Equation 10.2 in Chapter 10, higher margins mean a lower break-even sales change that makes discounting more attractive and restraining discoun- ters more difficult. Therefore, breakdowns in collusively high prices often occur in indus- tries that require highly capital-intensive production processes, such as petroleum refining, steel making, and airlines.

Size and Frequency of Orders Successful oligopolistic cooperation also depends on the size distribution of customer orders over time. Effective collusion is more likely

Example How Ocean Shipping Conferences Have Affected Shipping Rates6

Since the Shipping Act of 1916, ocean freight companies have been exempt from the antitrust laws of the United States. Shipping rates on a transoceanic route are set jointly by 10 to 50 competitors acting as a “shipping conference.” Two studies in 1993 and 1995 by the U.S. Agriculture Department and the Federal Trade Com- mission (FTC) found that rates were 18 or 19 percent lower when ocean-shipping companies broke out of these conference arrangements and negotiated as indepen- dents. Nevertheless, the conferences maintain their market power by signing exclusive-dealing contracts with large-volume customers. The enormous capacity of the shipping conferences allows more schedule frequency and greater reliability than the independents can offer. In exchange for exclusive contracts, shipping con- ferences have offered attractive settlement of disputed cargo claims, using speedy liquidated damages processes. In 2000, the U.S. Congress held hearings about this antitrust immunity and proposed its removal, but no legislation was passed.

6Based on “Making Waves,” Wall Street Journal (October 7, 1997), p. A1; J. Yong, “Excluding Capacity-Constrained Entrants through Exclusive Dealing: Theory and Applications to Ocean Shipping,” Journal of Industrial Economics 46, no. 2 (June 1996); and “Shipmates,” Wall Street Journal (February 20, 2003), p. A1.

Chapter 12: Price and Output Determination: Oligopoly 419

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

when orders are small, frequent, and received regularly, as in the purchase of autos. When large orders are received infrequently at irregular intervals, as in the purchase of aircraft engines, it is more difficult for firms to collude on pricing and output decisions. Hence, Pratt & Whitney, Rolls-Royce, and General Electric have never colluded on jet engine prices.

Threat of Retaliation An oligopolistic firm will be less tempted to grant secret price concessions to selected customers if it feels that other cartel members would detect these price reductions and then retaliate. The toilet tissue manufacturers’ collusive agree- ment allegedly operated through public bids for institutional customers such as schools and hospitals. Sealed bids might have prevented the collusion, surprisingly.

Percentage of External Output Most cartels contain the seeds of their own de- struction. Rising prices and profits attract the entry of new competitors. Any increase in supply from outside the cartel means that larger restrictions on output must be imposed on cartel members in order to sustain any given market price. At one point in 1999, De Beers had to purchase for its own inventory $3.96 billion in diamonds (in only an $8 bil- lion market) in order to stabilize prices because so many Canadian, Australian, and Rus- sian diamonds (external at that point to the De Beers cartel) had flooded the market.8

Finally, in 2000, with 37 percent of total diamond supply outside the cartel, De Beers declared the end of its 65-year cartel. Similar events ended the OPEC I cartel when Mex- ican, Venezuelan, and Norwegian oil flooded onto the market. Ocean shipping prices are breaking down today because the rate-setting “conferences” now control less than 70 percent of the $85 billion North Atlantic market and less than 50 percent of the $262 billion trans-Pacific market. External suppliers reduce the likelihood of successful coordi- nation among cartel members to maintain prices above their competitive level.

Example DRAM Chipmakers Pay Enormous Fines for Forming a Global Cartel7

The world’s top four manufacturers of inexpensive random access memory chips, a key component of all consumer electronic devices, agreed to fines and jail terms for several executives because of 1999–2002 price fixing. The criminal conspiracy raised prices 400 percent in a six-month period from $1 to $4 per 100 megabits and then orchestrated maintaining the price at $3. DRAM chips are generic and easily substitutable between suppliers. As a result, a cartel agreement to limit pro- duction is necessary to maintain price above competitive levels. Samsung and Hynix, two Korean firms that produce the majority of the chips, paid $300 million and $185 million fines, respectively. Infineon Technologies of Germany paid a $160 million fine, and four executives went to jail for several months and paid in- dividual fines of $250,000. Micron Technology of Boise, Idaho, received immunity for cooperating with the prosecutors and complainants Dell and HP in making the case.

7Based on “Samsung to Pay,” Wall Street Journal (October 14, 2005), p. A3; and “Hynix Pleads Guilty,” Wall Street Journal (April 22, 2004), p. B6.

8Based on “De Beers to Abandon Monopoly,” Wall Street Journal (July 13, 2000), p. A20; and “Atlantic Ocean Shipping Cartel Makes Concessions,” Wall Street Journal (February 7, 1997), p. A2.

420 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Cartel Profit Maximization and the Allocation of Restricted Output Under both legal cartels and secret collusive agreements, firms attempt to increase prices and profits above the level that would prevail in the absence of collusion. The profit- maximization solution for a two-firm cartel, E and F, is shown graphically in Figure 12.4. The industry demand D, marginal revenue MR, and marginal cost ∑MC curves are shown in the third panel. The industry marginal cost curve is obtained by summing hor- izontally across outputs the marginal cost curves of the individual firms in the first two panels: that is, ∑MC = MCE + MCF. Total industry profits are maximized by setting total industry output (and consequently price) at the point where industry marginal revenue equals industry marginal cost (i.e., Q*Total units of output at a price of P* per unit).

If the cartel maximizes its total profits, the market share (or quota) for each firm should be set at a level where marginal cost of all firms is identical and the industry (summed) MC = MR. The optimal joint output is for Firm E to produce a quota of Q*E units and for Firm F to produce a quota of Q*F units. If Firm E were producing at a level where its marginal costs exceeded Firm F’s, cartel profits could be increased by shifting output from E to F until marginal costs were equal.9

Cartel pricing agreements are hard to reach, but the central problem for a cartel lies in monitoring these output shares or quotas. Detecting quota violations and effectively enforcing punishment schemes are nearly impossible. Consequently, most cartels are un- stable, like the price-fixing agreements among cardboard box manufacturers—these col- lusive agreements form approximately once a quarter and break up within a few weeks. The longevity of the Organization of Petroleum Exporting Countries (OPEC) and the De Beers diamond cartels is exceptional. Let’s return to Figure 12.4 and see why.

Suppose you are Firm F facing a cartel-determined price for crude oil P* of $20 per barrel. Your marginal costs are presently running $12 per barrel at your assigned quota of QF/QTotal. The Aramco pipelines, which once consolidated all your throughput from the production wells to shipping terminals, have now been superseded by numerous in- dependent shipping terminals, where the crude is relatively undifferentiated. Should you abide by your quota commitment? Is it in your best interest to do so? The answer

FIGURE 12.4 Price-Output Determination for a Two-Firm Cartel

Firm E output (units)

MCE

Q*

P* $20

$12

$20

$12 ATCE

Firm F output (units)

MCF

Q*

ATCF

Cartel output (units) = ΣQi

ΣMC

Q*

MR

PS

Pc Dmkt

0 0 0E F Total

C os

t an

d pr

ic e

9Note that the average total costs of the two firms are not necessarily equal at the optimal (profit-maximizing) output level. Note also that Firm E is given a sizable share of the total output even though its average total costs are higher than Firm F’s.

Chapter 12: Price and Output Determination: Oligopoly 421

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

depends on whether your additional sales beyond quota are detectable and whether your additional output will increase total supply enough to place downward pressure on the car- tel price. If the answer to both questions is no, then because a 40 percent profit margin ($8) awaits your selling another barrel, a profit maximizer will be tempted to increase out- put and capture the hatched area of incremental profit in the middle panel of Figure 12.4.

Of course, the problem is that other cartel members may think exactly the same way. If everyone takes the cartel price as given and independently profit maximizes, then car- tel supply increases to ∑MC, and the black market price must fall to the competitive level Pc of perhaps $17 just to clear the market. Enforcement of the ideal quotas QF and QE is

INTERNATIONAL PERSPECTIVES

The OPEC Cartel10

The Organization of Petroleum Exporting Countries (OPEC) was founded in 1960 by five Persian Gulf nations who hosted ARAMCO, a joint venture set up in 1947 by the international oil companies for the exploration and development of the Mideast oil fields. ARAMCO set the price of crude oil and paid oil concession royalties used by the host nations to purchase back the ARAMCO assets. Controlling 80 percent of world petroleum output in 1973–1974, the OPEC members decided to restrain production in order to sustain a 400 percent increase in the price of crude oil from $3 to $12 per barrel. The OPEC I cartel was born. Saudi Arabia is the most influential member of this price-fixing cartel because of the tre- mendous size of its production capacity—almost one-half of OPEC’s total output at the inception of OPEC and still 32 percent of OPEC output today.

By the early 1980s, with the price of oil at $32 to $41 ($80 in 2006 dollars), covert price cuttingwas rampant. Nigeria, for example, engaged in secret price cutting by reducing income taxes for the oil companies working there. Other OPEC members bartered and extended paymenttermsforoilpurchases, therebyreducinginter- est expenses on the funds required to finance the pur- chase.During this period (often referred to asOPECII), Saudi Arabia regularly stabilized declining oil prices by actingasa“swingproducer,”cutting itsproduction toas low as 2 million barrels per day (bpd) in 1980 when its authorized quota was 4.35 million and its capacity was 10 million bpd. OPEC II ended effectively in October 1985, when Saudi Arabia reversed its policy and began increasing its output to as much as 6 million bpd. The equilibriummarket price of crude fell as low as $12.

Today OPEC controls less than 40 percent of world oil output, half of what they once did. Throughout the 1990s, production expanded in non- traditional oil-producing regions like Prudhoe Bay, Alaska; in Russia; and in the North Sea despite ex- traction costs three to five times higher than the $3- per-barrel exploration, development, and extraction cost in the Middle East. Venezuela has publicly chal- lenged the role of Saudi Arabia as swing producer and price leader, especially in the western hemi- sphere. And Russia at 9.27 million bpd has posed the same challenge in other parts of the world.

With production breaking out all over, in 1998 and early 1999, crude oil prices collapsed to $9.96 (see Figure 12.5). To stabilize the market, OPEC members agreed in March 1999 (and again in Sep- tember 2000) to a production quota system. Saudi Arabia accepted a 585,000-barrel-per-day cutback, which equaled 7 percent of its February 1999 average daily production of 8.8 million barrels. Iran, with a 12 percent share, agreed to a 264,000-barrel cutback, which also equaled a 7 percent reduction of its 3.6-million-barrel output. Venezuela accepted a 125,000-barrel-per-day cutback, which equaled a 4 percent reduction of its 3.4-million-barrel output. Crude oil prices responded almost immediately, ris- ing more than threefold from a $10 trough to $33 per barrel in 15 months (again see Figure 12.5). The OPEC III cartel was in place, effectively restricting output to raise prices.

10“Why the Saudis Won’t Back Down Soon,” Wall Street Journal (April 8, 1986); and J. Griffin and W. Xiong, “The Incentive to Cheat: An Empirical Analysis of OPEC,” Journal of Law and Economics 60, no. 2 (1997).

422 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the weak point of every cartel. In OPEC, Saudi Arabia plays a pivotal role in absorbing quota violations by other OPEC members and thereby stabilizes the cartel.

FIGURE 12.5 How OPEC III Production Quotas Affected Crude Oil Prices

1992

Gulf War

OPEC III

1994 1996 1998 2000 2002

50

60

70

40

30

20

10

0

O il

pr ic

e, in

c on

st an

t 20

00 U

.S .$

1992–2002 Real price per barrel

Mean +/– 2 Standard deviations

Source: Federal Reserve Bank, St. Louis, National Economic Trends.

Example Coffee Pricing Agreement Dissolves amidst Dilemma11

When the coffee bean harvest is larger than projected, the top Colombian and Bra- zilian coffee producers along with several African and Central American smaller producers often agree in principle to withhold millions of tons of coffee beans from the market in an effort to raise collapsing wholesale prices. Brazilian produ- cers may propose to hold back 2 million bags of a projected 18-million-bag crop. Colombian producers may agree to hold back 1.3 million bags. However, both countries oppose a formal quota system that assigns production ceilings, imposes monitoring mechanisms, and penalizes violators. In 1989 and again in 1993, Inter- national Coffee Agreements collapsed over the refusal to accept assigned quotas.

If all major coffee bean producers could rely upon one another to withhold ex- traordinarily large production in exceptionally good weather years, all would have higher profitability. Instead, some cartel members maximize self-interest by releas- ing excess supplies to the world market at just below the officially agreed-upon price. Because other cartel members think that same way, equilibrium market price then plummets. Only dupes then continue to restrain output when world market prices erode, signaling that other coffee producers are violating the agreement.

11“Brazil, Colombia Form Cartel for Coffee Exports,” Wall Street Journal (September 8, 1991), p. B12.

Chapter 12: Price and Output Determination: Oligopoly 423

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Exhaustible Natural Resources: Saudi Arabia Plays a Waiting Game12

Some resources like crude oil, coal, natural gas, and diamonds are formed over tens of thousands of years. Although limited and fixed in this geological sense, more intense exploration and development can often locate additional resources. Renew- able resources like fisheries and timberland replenish themselves if harvesting is restrained to prevent exhaustion, but the only reason for an exhaustible resource owner to not extract crude oil, or coal, or natural gas is if she believes that the resource price is going to rise in the near future. Price changes and price change expectations are therefore the key to exhaustible resource decisions.

Define consensus expectations for future prices in time period T (PT) as

PT = P0ð1 + rÞT [12.10] where r is the real rate of interest (more precisely, the inflation-adjusted risk- adjusted rate of interest). Dividing each unit of time into n subperiods and taking the limit as n goes to infinity, the compound growth version of these consensus price expectations may be written as

PT = P0½ lim n → ∞

ð1 + r=nÞnT� = P0erT [12.11]

where e is 2.7183 …, the base of the natural logarithms. We are now in a position to express the harvest now-or-wait decision in terms of the opportunity cost of waiting (the real rate of interest r) relative to the percentage rate of growth of re- source prices:13

ΔPT=ΔT T

= rP0erT

PT

which reduces, using equation 12.11, to

ΔPT=ΔT PT

= r [12.12]

This result states that as long as the expected rate of price increase (say, 8 per- cent) exceeds the interest rate (say, 4 percent), one should leave the crude oil, coal, or natural gas in the ground and harvest later. If interest rates rise above this per- centage growth rate of the exhaustible resource prices, the resource should be ex- tracted and sold now.

In 2008–2009, crude oil prices rose to $147 per barrel and then fell like a rock to $39 even though demand continued to grow. The worldwide financial crisis crushed the speculative demand, and greener hybrid-electric and all-electric vehi- cles were beginning to replace gasoline-engine autos. In the face of projections of flat to declining crude oil prices, Saudi Arabia decided to increase production and harvest now! From 8 million barrels per day (mbd) in 2002, the Saudis increased production to almost 11 mbd in 2009 (see Figure 12.6). As a result gasoline re- mains substantially less expensive than electricity for powering cars and only about

(Continued)

424 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

80 percent as expensive as the full cost of hydrogen fuel for powering cars. Again, these capacity utilization policies were designed to discourage the development of substitute fuels by slowing the rate of price increase for crude oil. Any other ex- haustible resource owner with a 67-year supply of proven reserves would have done the same thing.

12Based on OPEC Annual Statistical Bulletin; and “Why the U.S. Is Still Hooked on Oil Imports,” Wall Street Journal (March 18, 2003), p. A1. 13This step is based on the calculus result that

derT

dT = rert

FIGURE 12.6 Saudi Arabian Crude Oil Production

1970

(Millions of barrels per day)

2

4

6

8

10

10.9

1975 1980 1985 1990 1995 2000 2005 2010

Source: U.S. Energy Information Agency.

Chapter 12: Price and Output Determination: Oligopoly 425

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Saudi Arabia has an enormous 264 billion barrels of known but untapped crude oil reserves that are expected to last 66.5 years at current rates of production (see Figure 12.7). By comparison, the United States has only 21 billion barrels and 12 years of reserves left. Therefore, the Saudis prefer to discourage the emergence of alternative fuels such as sugar cane-based ethanol in Brazil, corn-based ethanol in the United States, and the extraction facilities for the extensive tar sands in Canada. In this sense, U.S. oil interests in Texas and Oklahoma are often at odds with Saudi Arabia. The former want policies designed to raise the oil price quickly before their reserves run out. The latter want to hold prices below levels that would trigger replacing internal combustion gasoline engines with all-electric vehicles or hydrogen fuel cells since Saudi and other OPEC oil reserves will last for nearly a century.

Cartel Analysis: Algebraic Approach The profit-maximizing price and output levels for a two-firm cartel can be determined algebraically when the demand and cost functions are given. Consider again the Siemens (Firm S) and Lucent-Alcatel (Firm T) example discussed in the previous section. The de- mand function was given by Equation 12.1 and the cost functions for the two firms were given by Equations 12.2 and 12.3. Suppose that Siemens and Lucent decide to form a cartel and act as a monopolist to maximize total profits from the production and sale of the components.

Example What Drives the Cost of $3 per Gallon Gasoline?14

In recent years, the OPEC III cartel has been overshadowed by the Iraq War and by unanticipated growth of gasoline consumption in China and India. In 2006 and again in 2009, crude oil sold for $70–$80 a barrel and gasoline prices rose above $3 per gallon. What are the component costs of this $3+ retail price of gasoline?

One explanation might be that state and federal excise taxes or state and local sales taxes have risen, but they have remained about $0.50 to $0.60 per gallon for a decade (see Figure 12.8). Another explanation might be that retail station owners are gouging customers, but perfect competition characterizes the retail gas market, and consequently, retail margins are low and have remained unchanged at $0.15 to $0.20 for many years. Distribution bottlenecks occasionally are responsible for price spikes in a local region when pipelines break, but in general, only $0.07 of the price of gas is attributable to distribution costs. Scarcity of refining capacity is another explanation, and $0.80 of the $3.00 is attributable to refining costs that have risen somewhat in recent years.

Nevertheless, by far the largest component cost reflected in retail gas prices is crude oil (again, see Figure 12.8). Fully half of the cost of $3-per-gallon gasoline ($1.54) is attributable to the cost of crude oil itself. And it is this component that has increased enormously from $35 per barrel in 2004 to $70–$80 per barrel in 2005–2006 and again in 2009.

14Based on “Oil Nations Move Closer to a New Round of Cuts,” Wall Street Journal (March 12, 1999), p. A3; “Crude Cuts: Will Oil Nations Stick or Stray?” Wall Street Journal (March 26, 1999), p. A19; “The Next Oil Shock,” The Economist (March 6, 1999); “Standstill Britain,” The Economist (September 16, 2000), p. 64; and “At OPEC Some Say There’s Enough Oil,” Wall Street Journal (September 12, 2000), p. A2.

426 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Total industry profits (πTotal) are equal to the sum of Siemens’s and Thomson’s profits and are given by the following expression:

πTotal = πS + πT = PQS −TCS + PQT −TCT [12.13]

FIGURE 12.7 Proven Oil Reserves

Canada

United States

Algeria

Tazakhstan

Libya

Russia

AE

Venezuela

Kuwait

Iraq

Iran

0 50 100 Billion barrels Years remaining at

2008 production rate150

Saudi Arabia 264.1 66.5

86.9

100+

99.6

100+

89.7

21.8

64.6

70.0

45.6

12.4

24.1

Source: The Economist (June 13, 2009), p. 101.

FIGURE 12.8 Components of the Price of Gasoline per Gallon (1990–2009)

1990 1995

Crude oil

Refining

Distribution Tax

2000 2005 and 2009

100

150

200

300

250

50

0

Pr ic

e (c

en ts

p er

g al

lo n)

Retail price $3.00/gallon* (includes $0.07 marketing and distribution cost)

Crude oil $1.40

Refining $0.80

Retail markup $0.19

Federal, state, and local taxes

$0.54

Source: Changing Gasoline Prices, Federal Trade Commission, June, 2005.

Chapter 12: Price and Output Determination: Oligopoly 427

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Substituting Equations 12.1, 12.2, and 12.3 into this expression yields

πTotal = ð1,000−QS −QTÞQS − ð70,000+ 5QS + 0:25Q2SÞ + ð1,000−QS −QTÞQT − ð110,000+ 5QT + 0:15Q2TÞ

= 1,000QS −Q2S −QSQT − 70,000− 5QS − 0:25Q 2 S

+ 1,000QT −QSQT −Q2T − 110,000− 5QT − 0:15Q 2 T

= − 180,000+ 995QS − 1:25Q2S + 995QT − 1:15Q2T − 2QSQT [12.14]

To maximize πTotal, take the partial derivatives of Equation 12.14 with respect to QS and QT:

∂πTotal ∂QS

= 995 − 2:50QS − 2QT

∂πTotal ∂QT

= 995 − 2:30QT − 2QS

Setting these expressions equal to zero yields

2:5QS + 2QT − 995 = 0 [12.15]

2QS + 2:3QT − 995 = 0 [12.16]

Solving Equations 12.15 and 12.16 simultaneously gives the following optimal output le- vels: Q*S = 170.57 units and Q*T = 284.39 units.

Substituting these values into Equations 12.13 and 12.14 gives an optimal selling price and total profit for the cartel of P* = $545.14 per unit and π*Total = $46,291.43, respec- tively. The marginal costs of the two firms at the optimal output level are equal to

MC*S = dðTCSÞ dQS

= 5 + 0:50QS

= 5 + 0:50ð170:57Þ = $90:29 MC*T =

dðTCTÞ dQT

= 5 + 0:30QT

= 5 + 0:30ð284:29Þ = $90:29 As in the graphical solution illustrated earlier in Figure 12.2, the optimal output (or

market share) for each firm in the cartel occurs where the marginal costs of the two firms are equal.

Table 12.3 summarizes the results of the Siemens and Thomson example: (a) where the two companies acted independently to maximize their own company profits (Cour- not equilibrium), and (b) where they formed a cartel to maximize total industry profits. Several conclusions can be drawn from this comparison. First, total industry output (Q*Total) is lower and selling price (P*) is higher when the firms collude. Also, total indus- try profits (π*Total) are higher when the firms set prices and output jointly than when they act independently. Finally, although it may not be true in all collusive agreements, one firm’s profits (i.e., Siemens’s) are actually lower under the cartel solution than when it acts independently. Therefore, to get Siemens to participate in the cartel, Thomson prob- ably would have to agree to share a significant part of the cartel’s additional profits with Siemens.

428 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PRICE LEADERSHIP Besides cartels, another model of price-output determination in some oligopolistic indus- tries is price leadership. Many industries exhibit a pattern where one or a few firms

TABLE 12.3 COMPARISON OF PRICING, OUTPUT, AND PROFITS FOR SIEMENS AND THOMSON

OPTIMAL VALUE

(A) NO COLLUSION: SIEMENS AND THOMSON ACT INDE- PENDENTLY TO MAXIMIZE THEIR OWN COMPANY ’S

PROFITS

(B) COLLUSION: SIEMENS AND THOMSON FORM A CARTEL TO MAXIMIZE

TOTAL INDUSTRY PROFITS

Q*S (Siemens’s output) 272.32 units 170.57 units

Q*T (Thomson’s output) 314.21 units 284.29 units

Q*Total = Q*S + Q*T (Total industry output) 586.53 units 454.86 units

P* (Selling price) $413.47/unit $545.14/unit

π*S (Siemens’s profit) $22,695.00 $14,858.15

π*T (Thomson’s profit) $3,536.17 $31,433.28

π*Total = π*S + π*T (Total industry profit) $26,231.17 $46,291.43

Example Revenue-Sharing in Major League Baseball15

Major League Baseball (MLB), a cartel of professional team owners, has been ex- empted from the antitrust laws since a U.S. Supreme Court decision in 1922. MLB restricts entry, approves transfers of ownership, and regulates the selection and em- ployment of apprentice players for their first six years in professional baseball. In 1975, Curt Flood of the St. Louis Cardinals successfully challenged baseball’s re- strictive labor practices (the “reserve clause”) for major leaguers beyond their sixth year, and a collective bargaining agreement then granted such players free agency status. Experienced players could then offer their services to the highest bidder whenever their contracts were due for renewal.

As a result, salaries skyrocketed and star players concentrated in the biggest mar- kets (New York and Los Angeles) where owners with larger ticket revenue, higher concession sales, and bigger television contracts offered to pay more. Even the average player profited from the end of baseball’s reserve clause. Average salary for major leaguers rose in inflation-adjusted dollars from $160,000 in 1972 to $1,015,000 in 1992. Because owners spend 58 percent of team revenue on salaries and another 13 per- cent on the scouting system and the minor league apprentice teams, the MLB cartel intervened to restore competitive balance. MLB began a revenue-sharing system whereby the wealthiest teams are taxed 34 percent of total revenue to subsidize salaries for teams with a smaller fan base, either because of smaller markets (Minneapolis) or less success on the field (Baltimore). By 2003, subsidies totaled more than $260 million.

15Based on “Let the Market Rule,” Wall Street Journal (November 10, 1998), p. A22; “Just Not Cricket,” The Econo- mist (May 31, 2003), p. 34; and Gerald Scully, The Market Structure of Sports (Chicago: Chicago University Press, 1995).

price leadership A pricing strategy followed in many oligopolistic industries. One firm normally announces all new price changes. Either by an explicit or an implicit agreement, other firms in the industry regularly follow the pricing moves of the industry leader.

Chapter 12: Price and Output Determination: Oligopoly 429

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

normally set a price and others tend to follow, frequently with a time lag of a few days. In the case of basic steel products, for example, the price that prevails within a week is generally uniform from one producer to another.

Effective price leadership can only happen if price movements initiated by the leader have a high probability of being adopted and no maverick or nonconforming firms exist. The fewer the number of firms in the industry (i.e., the greater the interdependencies of decision outcomes among firms), the more effective price leadership is likely to be. Two major price leadership patterns have been observed in various industries from time to time: barometric and dominant price leadership.

Barometric Price Leadership In barometric price leadership, one firm announces a change in price that it hopes will be accepted by others. The leader need not be the largest firm in the industry. In fact, this leader may actually change from time to time. The leader must, however, be reasonably cor- rect in its interpretation of changing demand and cost conditions so that suggested price changes will be adopted industry-wide. In essence, the barometric price leader merely initi- ates a reaction to changing market conditions that other firms find in their best interest to follow. These conditions might include such things as cost increases (or decreases) and sluggish (or brisk) sales accompanied by inventory buildups (or shortages) in the industry.

Dominant Firm Price Leadership In dominant firm price leadership, one firm establishes itself as the leader because of its larger size, customer loyalty, or lower cost structure in relation to other competing firms.

Example Barometric Price Leadership: American Airlines and Continental Airlines16

In the second week of March 2002, American Airlines announced a de facto increase in business airfares. Three-day advance purchase fares were no longer available on many nonstop routes from American hubs. Instead, American returned to the old 7-day advance purchase requirement to obtain a 20 percent off full coach class fare ($1,629 from Dallas to New York or $1,684 from Dallas to Miami, for example). Other much cheaper Saturday overnight fares bought 7 days or even 14 days in ad- vance were not affected because those prices primarily target leisure travelers.

American Airlines was hoping that its major competitors, Continental, Delta, United, US Airways, and Northwest, would take this opportunity to follow its lead and increase margins. Only Continental did so. Indeed, Northwest took ad- vantage of the situation and promoted a deeply discounted $198 round-trip fare on American’s principal nonstop routes. Within a few days, American rescinded its pricing changes on routes where it competed with Northwest but retained them where it had a dominant hub, as at Dallas-Ft. Worth. In addition, American simultaneously announced a week of $198 fares on 10 nonstop routes from United’s Chicago hub, 10 nonstop routes from Delta’s Atlanta hub, 10 nonstop routes from US Airways’ Pittsburgh hub, and 10 nonstop routes from Northwest’s Minneapolis hub. Only Continental’s Houston hub was spared.

16Based on “Airfare Skirmish Shows Why Deals Come and Go,” Wall Street Journal (March 19, 2002), p. B1.

430 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The leader may then act as a monopolist in its segment of the market. What is the in- centive for followers to accept the established price? In some cases it may be a fear of cutthroat retaliation from a low-cost dominant firm that keeps smaller firms from under- cutting the prevailing price. In other cases, following a price leader may be viewed as simply a convenience.

The price-output solution for the dominant-firm model is shown in Figure 12.9. DT shows total market demand for the product, MCL, represents the marginal cost curve for the dominant (leader) firm, and ∑MCF constitutes the horizontal summation of the mar- ginal cost curves for the follower firms, each of which may well have costs higher than MCL. In the following analysis, assume that the dominant firm sets the price knowing that follower firms will sell as much output as they wish at this price. The dominant firm then supplies the remainder of the market demand.

Given that the follower firms can sell as much output as they wish at the price PL established by the dominant firm, they are faced with a horizontal demand curve and a perfectly competitive market situation. The follower firms view the dominant firm’s price PL as their marginal revenue and maximize profits, producing that level of output where their marginal cost equals the established price. The ∑MCF curve therefore shows the to- tal output that will be supplied at various prices by the follower firms. The dominant firm’s residual demand curve DL is obtained by subtracting the amount supplied by the follower firms’ ∑MCF from the total market demand DT at each price. For example, at a price of PL, Point G on the DL curve is obtained by subtracting EC from ED. Other points on the DL curve are obtained in a similar manner. At a price of P1 the quantity supplied by the follower firms’ Q1 is equal to total market demand (Point A), and the dominant firm’s residual demand is therefore zero (Point F). The dominant firm’s mar- ginal revenue curve MRL is then obtained from its residual demand curve DL.

The dominant firm maximizes its profits by setting price and output where marginal cost equals marginal revenue. As shown in Figure 12.9, MRL = MCL at Point B. There- fore, the dominant firm should sell QL units of output at a price of PL per unit. At a

FIGURE 12.9 Price-Output Determination for the Dominant Firm

QL

Output Q (units)

PL

P1

QF

MCL

Q1 QT

F

E G

B C D

A

ΣMCF

DL

MRL

DT

0

Pr ic

e an

d co

st (

$/ un

it )

Chapter 12: Price and Output Determination: Oligopoly 431

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

price of PL, total demand is QT units, and the follower firms supply QT – QL = QF units of output.

The following example illustrates the application of these concepts.

Example Price Leadership: Aerotek Aerotek and six other smaller companies produce an electronic component used in small planes. Aerotek (L) is the price leader. The other (follower [F]) firms sell the component at the same price as Aerotek. Aerotek permits the other firms to sell as many units of the component as they wish at the established price. The company supplies the remainder of the demand itself. Total demand for the component is given by the following function:

P = 10,000 − 10QT [12.17]

where

QT = QL + QF [12.18]

that is, total output (QT) is the sum of the leader’s (QL) and followers’ (QF) output. Aerotek’s marginal cost function is

MCL = 100 + 3QL [12.19]

The aggregate marginal cost function for the other six producers of the compo- nent is

∑MCF = 50 + 2QF [12.20]

We are interested in determining the output for Aerotek and the follower firms and the selling price for the component given that the firms are interested in max- imizing profits.

Aerotek’s profit-maximizing output is found at the point where

MRL = MCL

Its marginal revenue function (MRL) is obtained by differentiating the firm’s total revenue function (TRL) with respect to QL. Total revenue (TRL) is given by the following expression:

TRL = P ·QL

QL is obtained from Equation 12.18:

QL = QT − QF

Using Equation 12.17, one can solve for QT:

QT = 1,000 − 0:10P [12.21]

To find QF, we note that Aerotek lets the follower firms sell as much output (i.e., components) as they wish at the given price (P). Therefore, the follower firms are faced with a horizontal demand function. Hence

MRF = P [12.22]

(Continued)

432 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

To maximize profits, the follower firms will operate where

MRF = ∑MCF [12.23]

Substituting Equations 12.22 and 12.20 into Equation 12.23 gives

P = 50+ 2QF [12.24]

Solving this equation for QF yields

QF = 0:50P − 25 [12.25]

Substituting Equation 12.21 for QT and Equation 12.25 for QF in Equation 12.20 gives

QL = ð1,000− 0:10PÞ − ð0:50P − 25Þ = 1,025− 0:60P [12.26]

Solving Equation 12.26 for P, one obtains

P = 1,708:3333− 1:6667QL [12.27]

Substituting this expression for P in defining total revenue yields

TRL = ð1,708:3333− 1:6667QLÞQL = 1,708:3333QL − 1:6667Q2L [12.28]

Differentiating this expression with respect to QL, one obtains Aerotek’s marginal revenue function:

MRL = dðTRLÞ dQL

= 1,708:3333− 3:3334QL [12.29]

Substituting Equation 12.29 for MRL and Equation 12.16 for MCL and equating the two gives the following optimality condition:

1,708:3333− 3:3334Q*L = 100+ 3Q*L [12.30]

Solving this equation for Q*L yields

Q*L = 253:945 units

or an optimal output for Aerotek of 253.9 units of the component. Substituting this value of QL into Equation 12.27 gives

P* = 1,708:3333− 1:6667ð253:945Þ = $1,285:083

or an optimal selling price of $1,285.08. The optimal output for the follower firms is found by substituting this value of P into Equation 12.25,

Q*F = 0:50ð1,285:083Þ − 25 = 617:542 units

or an optimal output of 617.5 units.

Chapter 12: Price and Output Determination: Oligopoly 433

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

THE KINKED DEMAND CURVE MODEL Sometimes when an oligopolist cuts its prices, competitors quickly feel the decline in their sales and are forced to match the price reduction. Alternatively, if one firm raises its prices, competitors rapidly gain customers by maintaining their original prices and hence have little or no motivation to match a price increase. In this situation, the demand curve facing an individual oligopolist would be far more elastic for price increases than for price decreases. If an oligopolist raises its price and others do not follow, the increase in price will lead to a declining share of the market as illustrated in Figure 12.10. Demand segment KD0 is the share-of-the-market demand curve where all rivals match price and this firm’s market share remains unchanged, for example, at 21 percent. For price increases above P, however, if rival firms do not match price, the demand segment facing this firm is more elastic. For price increases, its market share declines, perhaps to 15 percent.

The oligopolist’s demand curve is thereforeDKD0 with the prevailing price as P and out- put as Q. The marginal revenue curve is discontinuous because of the kink in the demand curve at K. Hence, marginal revenue is represented by the two line segments MRX and YMR0. If the marginal cost curve MC passes through the gap XY in the marginal revenue curve, the most profitable alternative is to maintain the current price-output policy. The profit-maximizing level of price and output remains constant for the firm, which perceives itself to be faced with a fixed unit price, even though costs may change over a rather wide range (e.g., MC2 and MC1). This model explains why stable prices have been observed to exist in some oligopolistic industries. But the kinked demand model is incomplete in that it offers no reason why the prevailing price level rather than some other one is chosen.

AVOIDING PRICE WARS Knowing how to avoid a price war has become a critical success factor for many high- margin businesses in tight oligopolistic groups. Recall from our discussion of break-even

FIGURE 12.10 The Kinked Demand Curve Model

Output Q (units)

P

Q

MC2

Y

X

D�

K MC

MC1

D

MR�

MR

MC2

MC

MC1

0

Pr ic

e an

d co

st (

$/ un

it )

434 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

sales change analysis in Chapter 10 that the higher the margin, the more tempted com- panies are to use price discounting to increase incremental sales. Because each additional sale incurs few additional costs, high margins encourage price discounting to gain market share. So building a business plan or adopting a strategy that reduces the power of sub- stitutes, entrants, buyers, and suppliers, and thereby generates high profit margins, is no guarantee of success. To sustain profitability, oligopoly firms also must avoid discounting tactics in their high-margin business.

The ready-to-eat (RTE) cereal, beer, camera film, cigarette, book/DVD, and video game industries have all experienced classic price wars. In some cases, the catalyst for the price war was the fast-rising market share of private labels in what had previously been a heavily branded category. In the 1990s, generic cigarettes such as “Basic” took a substantial amount of market share from premium brands Marlboro, Benson and Hedges, Winston, Merit, and Salem. The R.J. Reynolds company introduced a mid- priced “fighting brand” Doral, promoted it heavily, and grew market share quickly. Ulti- mately, Philip Morris (now Altria) was losing so much market share that they took 20 percent ($0.40) off the average $1.92 price of Marlboro. Similarly, a tiny cereal man- ufacturer, Ralston, began supplying many grocery store chains and Target with private label cereals (e.g., Kroger Raisin Bran) that sold at price points 30 percent less than the premium brands such as Kellogg’s Raisin Bran. The fourth largest manufacturer, Quaker Oats, with a 7 percent market share, began selling branded cereals such as Cap’n Crunch and Life in large “value-priced” bags for $3.50 in the Target and Walmart distribution channels. The market share of these private label store brands grew rapidly, sometimes as much as 30 percent per year.

Growing the Market One key to avoiding price wars in tight oligopolies is to rec- ognize the ongoing nature of the pricing rivalry and attempt to mitigate the intensity of the price competition by growing the market. United Airlines cannot hope to get rid of American Airlines. Pepsi foresees a perpetual rivalry with Coke. Consequently, each rival must anticipate retaliation for aggressive discounting designed to attract away the other company’s regular customers. It is better to maintain high prices and expect your rivals to do the same. Then each company can focus on opening new markets and selling more volume to established customers. Coke Classic now sells an average of six servings per day to heavy Coke drinkers. In the past five years, Coke introduced dozens of new soft drinks to countries throughout the world. As a result, the Coca-Cola concentrate syrup has never been discounted in 80 years.

Customer Segmentation with Revenue Management Customer segmenta- tion with differential pricing is another way to avoid price wars. If low-cost new en- trants attack a major airline, one effective response that avoids initiating a price war with other major carriers involves matching prices to a targeted customer segment and then carefully controlling how much capacity is released for sale to that segment. “Fencing” restrictions such as 7-day advance-purchase requirements and Saturday night stay-overs prove crucial in segmenting the price-sensitive discretionary traveler from the regular business expense account customer. The incumbent carriers can “meet the competition” in these restricted fare classes while reserving sufficient capac- ity for those who desire to pay for the reliability, convenience, and change order re- sponsiveness of business-class and full-coach seats. Most importantly, established competitors can maintain high prices on unaffected departures, segments, and routes. In Chapter 4, we discuss how revenue management techniques can help accomplish these goals.

Chapter 12: Price and Output Determination: Oligopoly 435

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Reference Prices and Framing Effects In addition to segmenting the target customers into more and less price-sensitive submarkets, product line extensions can re- duce gainshare price discounting by providing reference prices and framing effects that help sell the mid-range product at undiscounted prices. Consumers of unbranded prod- ucts typically remember the last price they encountered on the shelf in deciding whether to purchase at the quoted price today. Branded products, however, trigger much longer reference pricing. Discounting with a major branded item such as Tide detergent tends to etch in the customer’s mind a new lowball price that can be expected thereafter for many months or even years. Therefore, what one really might like to do in the face of

Example Price Wars at General Mills and Post17

The price cut that triggered a price war in ready-to-eat (RTE) cereals was also a 20 percent discount ($1.00 off the $4.80 average price for a full-size box of RTE cereal). The price war in cereals was started by Post Cereals, the distant third player in the industry with a 13 percent market share. Post had carefully analyzed the tactical situation and decided it could better maintain regular customers and compete for price-sensitive new customers if Kellogg and General Mills re- duced advertising. Post believed they would do so only in response to a massive industry-wide price cut.

General Mills was experiencing a slowly eroding 25 percent market share, while Kellogg faced a rapidly declining 35 percent market share. Every share point in the U.S. ready-to-eat cereal industry is worth $80 million in sales. In part because of a panic-stricken determination to arrest the erosion of their market shares, both Kel- logg and General Mills quickly decided to match the Post price cut. Full-size boxes of branded products such as General Mills’ Wheaties and Kellogg’s Frosted Flakes were cut in price from $4.80 to $3.88. Just as Post had predicted, each of the lead- ing firms then scaled back their advertising campaigns. Cereals such as Post Raisin Bran and Post Grape Nuts then gained share rapidly, at least for a short time until Kellogg matched the Post price cut on two-thirds of its premium brands. Two years later, cereal prices in the all-important grocery store distribution channel be- gan to return to their pre-price-war levels.

Kellogg has the strongest brands in the cereals industry with 12 of the 15 top- selling cereals. Rather than match Post’s price cuts, Kellogg might have poured not two but three scoops of raisins into every box of Kellogg’s Raisin Bran. In the first two months after the price cuts by Post and General Mills, Kellogg lost three share points (from 35 percent to 32 percent) and Post gained four (from 16 percent to 20 percent). At $80 million per share point and 55 percent gross margins (on av- erage across the affected brands), Kellogg’s contributions on the lost sales totaled $132 million (−3 × $80 million × 0.55). To retrieve that $132-million-per-year op- erating profit, Kellogg slashed prices 19 percent on two-thirds of its brands, sacrificing more than $305 million (−0.19 × $2.4 billion sales × 0.66). Market share continued to decline to 29 percent in 1999, and the capitalized value of Kellogg fell by $7 billion. Many observers wondered whether expending $305 million (or half that much) on advertising or product innovation would have accomplished more.

17Based on “Denial in Battle Creek,” Forbes (October 7, 1996); “Cereal Thriller,” The Economist (June 15, 1996); and P. Cummins, “Cereal Firms in Cost-Price Squeeze,” Reuters News Service (May 15, 1996).

436 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

private-label discount competition is to introduce a super-premium product offered at price points well above your traditional product. Loyal customers of the regular brand will remember these high references prices.

Because the opportunity losses in going from a mid-range Chrysler Town and Coun- try to bargain-basement products like a Dodge minivan (while saving $5,000, for exam- ple) tend to weigh more heavily upon consumers than the perceived satisfaction of moving upscale to super premiums like a Mercedes M class (at an equal $5,000 higher cost), the mid-range products like Chrysler Town and Country are expected to sell even better in the presence of the framing effect provided by super-premium products. In the 2000s, Town and Country provided 28 percent of Chrysler’s sales with relatively little discounting.

The Role of Innovation Another way to avoid or at least reduce the effects of price wars is to differentiate through innovation. Rather than matching price cuts, a higher-priced brand can highlight conspicuous product innovations the discounters have missed. Sony’s Mavica was an easy-to-use point-and-shoot digital camera that re- corded images onto thumb drives or disks. The disk popped out of the camera and into any PC for easy editing, storing, and printing. While digital camera competitors Kodak and Casio were improving picture resolution to justify expensive and complicated home printer hardware peripherals using Kodak chemicals and paper, Sony simplified the pro- cess and increased customer value. As a result, the Mavica earned a premium price rela- tive to its competitors.

Figure 12.11 analyzes an oligopolistic market with extreme brand loyalty based on in- novation, customer risk avoidance, or effective brand name advertising. Kellogg’s Raisin

WHAT WENT RIGHT • WHAT WENT WRONG

Good-Better-Best Product Strategy at Kodak and Marriott18

Marriott Corporation and Kodak have responded to fierce price competition in their respective industries by intro- ducing upscale, high-quality mid-range, and down- market product lines to their respective target customers. Ritz-Carlton, Courtyards by Marriott, and Fairfield Inns all operate as subsidiary hotel chains under the parent Mar- riott Corporation but as distinct offerings.

Similarly, in the early 1990s, in response to declining perceived quality differentials, collapsing market share, and price pressure from private label film, Kodak intro- duced a new lineup that included Royal Gold, Kodak Gold Plus, and Funtime Film. Successful segmentation is the key to such a product strategy for avoiding ruinous price discounting. Funtime Film and the Kodak disposable cameras that followed are positioned for everyday use to capture the hundreds of events, posed people, and scenery that highly accessible cheap film sold through convenience store distribution channels makes possible. These photo shots customers will later find “lost” in great stacks in file

cabinets, desk drawers, and old shoeboxes. Although not generally used to memorialize anything of significance, the snapshot accentuates the experiential event, as it happens.

“Kodak moments,” however, pursue a different set of value drivers. Kodak Royal Gold provides exceptional pic- ture resolution in many different lighting conditions. Al- though slower, Gold Plus is also able to memorialize subtleties of expressions of surprise, exaltation, pride in fulfilling challenging tasks, and so on. Kodak’s marketing research found that many of its customers would pay a price premium to memorialize a personal emotion (such as when a woman demonstrably triumphs amidst the worst whitewater rapids in a raft filled with her brothers). Heavy advertising and event marketing further established this product image.

18Based on “Film-War Spoils: A Buck a Roll?” Wall Street Journal (Novem- ber 11, 1998), p. B1; “Eastman Kodak Company: Funtime Film,” Harvard Business School Publishing (1998); and “Kodak Is Rolling Out Digital Photo Processing,” Wall Street Journal (February 9, 1999), p. A4.

Chapter 12: Price and Output Determination: Oligopoly 437

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example WhatWent Right at Interlink Surgical Steel andGillette?19

In the 1980s and early 1990s, Interlink sold replacement hypodermic syringes by the thousands to hospitals for 10 cents per syringe. Each time a catheter was changed, a new hypodermic syringe would be inserted into the patient’s vein. A Japanese surgical steel company entered the market with an identical product for 3 cents each. Interlink promptly introduced a replacement device that only needs insertion one time; that is, any new saline or pharmaceutical drip lines can be hooked directly to an Interlink syringe device that need not be removed and re- placed. This new process reduces the risk of patient infection and the inherent haz- ard to the nursing staff of exposure to patient blood. Interlink again dominates the market, and prices have stabilized at higher levels than before.

Similarly, Gillette Co. responded to a four-bladed new product introduction of Quattro by Schick-Wilkinson Sword. Quattro had stolen 3 percent of Gil- lette’s 83 percent market share in the men’s razor market. Instead of an ongo- ing parade of discounts, coupons, and promotions, Gillette rolled out its own innovation: a battery-powered vibrating razor called Mach3Power. M3Power improves the closeness and longevity of the shave by exciting chin hairs to push up out of their follicles. The new product costs two-thirds more than the Gillette Mach3Turbo it replaced, and replacement cartridges are 20 percent higher-priced.

19Based on “How to Fight a Price War,” Harvard Business Review (March/April 2000), pp. 107–116; “How to Escape a Price War,” Fortune (June 13, 1994), pp. 82–90; and “Gillette to Launch,” Wall Street Journal (January 16, 2004), p. A8.

FIGURE 12.11 Segmented Oligopoly with Extreme Brand Loyalty

Quantity (000s), Qd

Pr ic

e an

d co

st (

$/ bo

x)

$11

$1

$2

$4

$6

1175

$6

$4

MR Kroger MR PostMR Kellogg

DRaisin Bran

438 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Bran faces an inverse demand segment, ($11 – Qd) = Price, that includes the highest willingness-to-pay customers. Setting MR in this segment, $11 – 2Qd, equal to a marginal cost of $1, Kellogg’s Raisin Bran maximizes operating profit at Q* = 5(000) and a price per box of ($11 – 5) = $6. Without as established a brand image, Post Raisin Bran must sell under $6 and accordingly faces a different segment whose inverse demand may be written as ($6 – Qd) = Price (i.e., the line segment from $6 downward to the right along D in Figure 12.11). Setting MR at Post, $6 – 2Qd, equal to a higher $2 marginal cost per box yields a profit-maximizing output for Post of 2(000) at a profit-maximizing price of ($6 – $2) = $4 per box. These (5/11 = 45 percent) and (2/11 = 19 percent) market shares for Kellogg and Post, respectively, approximate their actual market shares in the ready- to-eat cereal market for raisin bran products. Additional firms with still less brand loy- alty, such as Kroger Raisin Bran, would supply the remaining segments illustrated still farther downward to the right on DD0 of the raisin bran demand curve.

Matching Price Cuts with Increased Advertising Perhaps the best way to avoid a price war in a small oligopolistic rivalry group is to not start one in the first place. If someone else does start a price war, often the best response is simply to match the competition and then accentuate nonprice elements of the marketing mix by increasing services or advertising. When Phillip Morris cut 20 percent from the price of premium cigarettes such as Marlboro, rather than furthering the downward price spiral, Reynolds instead matched the price cut only in its premium brands, Winston and Salem, and expanded advertising. At the $2.00 per pack all-time high price before the price war, the heavy smoker had a $35 per week incentive to quit. For Marlboro, with an 82 percent contribution margin, the 20 percent price cut ne- cessitated a 32 percent (0.82/[0.82 – 0.20] = 1.32) increase in incremental sales to achieve increased short-term profit. Instead, Marlboro market share rose only about 17 percent.

A final key in avoiding price wars comes through the tactical insights often available from game theory analysis. Being able to identify a rival’s payoffs using competitor sur- veillance helps predict the competitor’s response to one’s own price cuts. In other cir- cumstances, cooperative high-price outcomes may arise out of mutual interest. Simply recognizing the detailed structure of the pricing “game” can be a first step in altering the competitive environment in ways that increase profitability. In the following chapter, we present game theory techniques that provide useful managerial insights for effective tactical decision making.

Example Nonprice Tactics in a Price War: Coors20

Coors implemented exactly the same nonprice response in the midst of a costly price war between Anheuser-Busch and Miller Brewing. As Miller and Bud prod- ucts reduced the category to more and more of a commodity with ever deeper discounts, Coors and Stroh’s decided to realign their product positioning with Co- rona and Heineken. Amidst heavy advertising, Coors gained two share points de- spite prices $2-per-case higher than Miller and Bud.

20Based on “Big Brewers Find Price War Seems to Have No End,” Wall Street Journal (July 2, 1998), p. B6.

Chapter 12: Price and Output Determination: Oligopoly 439

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

SUMMARY

� An oligopoly is an industry structure characterized by a relatively small number of firms in which rec- ognizable interdependencies exist among the ac- tions of the firms. Each firm is aware that its actions are likely to evoke countermoves from its rivals.

� In the Cournot model of oligopoly behavior, each of the firms, in determining its profit-maximizing output level, assumes that the other firm’s output will remain constant.

� A cartel is a formal or informal agreement among oligopolists to cooperate or collude in determining outputs, prices, and profits. If the cartel members can enforce agreements and prevent cheating, they can act as a monopolist and maximize industry profits.

� A number of factors affect the ability of oligopolis- tic firms to engage successfully in some form of formal (or informal) cooperation, including the number and size distribution of sellers, product heterogeneity, cost structures, size and frequency of orders, secrecy and retaliation, and the percent- age of industry output from outside the cartel.

� Price leadership is a pricing strategy in an oligopo- listic industry in which one firm sets the price and, either by explicit or implicit agreement, the other firms tend to follow the decision.

� In the kinked demand curve model, it is assumed that if an oligopoly firm reduces its prices, its com- petitors will quickly feel the decline in their sales and will be forced to match the reduction. Alterna- tively, if the oligopolist raises its prices, competi- tors will rapidly gain customers by maintaining their original prices and will have little or no moti- vation to match a price increase. Hence, the de- mand curve for individual oligopolists is much more elastic for price increases than for price de- creases and may lead oligopolists to maintain sta- ble prices.

� To avoid price wars, oligopoly firms can grow the market, engage in product line extensions, expand into new geographic areas, segment customers and employ differential pricing, or innovate to retain profitable customers.

Exercises 1. Assume that two companies (C and D) are duopolists that produce identical prod- ucts. Demand for the products is given by the following linear demand function:

P = 600−QC −QD

where QC and QD are the quantities sold by the respective firms and P is the sell- ing price. Total cost functions for the two companies are

TCC = 25,000+ 100QC TCD = 20,000+ 125QD

Assume that the firms act independently as in the Cournot model (i.e., each firm assumes that the other firm’s output will not change).

a. Determine the long-run equilibrium output and selling price for each firm. b. Determine the total profits for each firm at the equilibrium output found in

Part (a).

Answers to the exercises in blue can be found in Appendix D at the back

of the book.

440 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

2. Assume that two companies (A and B) are duopolists who produce identical prod- ucts. Demand for the products is given by the following linear demand function:

p = 200−QA −QB

where QA and QB are the quantities sold by the respective firms and P is the sell- ing price. Total cost functions for the two companies are

TCA = 1,500+ 55QA +Q2A

TCB = 1,200+ 20QB + 2Q2B

Assume that the firms act independently as in the Cournot model (i.e., each firm assumes that the other firm’s output will not change). a. Determine the long-run equilibrium output and selling price for each firm. b. Determine Firm A, Firm B, and total industry profits at the equilibrium solu-

tion found in Part (a).

3. Consider Exercise 2 again. Assume that the firms form a cartel to act as a monop- olist and maximize total industry profits (sum of Firm A and Firm B profits). a. Determine the optimum output and selling price for each firm. b. Determine Firm A, Firm B, and total industry profits at the optimal solution

found in Part (a). c. Show that the marginal costs of the two firms are equal at the optimal solution

found in Part (a).

4. Compare the optimal solutions obtained in Exercises 2 and 3. Specifically: a. How much higher (lower) is the optimal selling price when the two firms form

a cartel to maximize industry profits, compared to when they act independently? b. How much higher (lower) is total industry output? c. How much higher (lower) are total industry profits?

5. Alchem (L) is the price leader in the polyglue market. All 10 other manufacturers (follower [F] firms) sell polyglue at the same price as Alchem. Alchem allows the other firms to sell as much as they wish at the established price and supplies the remainder of the demand itself. Total demand for polyglue is given by the follow- ing function (QT = QL + QF):

P = 20,000− 4QT

Alchem’s marginal cost function for manufacturing and selling polyglue is

MCL = 5,000+ 5QL

The aggregate marginal cost function for the other manufacturers of polyglue is

∑MCF = 2,000+ 4QF

a. To maximize profits, how much polyglue should Alchem produce and what price should it charge?

b. What is the total market demand for polyglue at the price established by Al- chem in Part (a)? How much of total demand do the follower firms supply?

Chapter 12: Price and Output Determination: Oligopoly 441

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

6. Chillman Motors, Inc., believes it faces the following segmented demand function:

P =

� 150− 0:5Q when 0≤Q≤ 50 200− 1:5Q forQ> 50

a. Indicate both verbally and graphically why such a segmented demand function is likely to exist. What type of industry structure is indicated by this relationship?

b. Calculate the marginal revenue functions facing Chillman. Add these to your graph from Part (a).

c. Chillman’s total cost function is

TC1 = 500+ 15Q+ 0:5Q 2

Calculate the marginal cost function. What is Chillman’s profit-maximizing price and output combination?

d. What is Chillman’s profit-maximizing price-output combination if total costs increase to

TC2 = 500+ 45Q+ 0:5Q 2

e. If Chillman’s total cost function changes to either

TC3 = 500+ 15Q+ 1:0Q 2

or

TC4 = 500+ 5Q+ 0:25Q 2

what price-output solution do you expect to prevail? Would your answer change if you knew that all firms in the industry witnessed similar changes in their cost functions?

7. Library Research Project. It was observed in the chapter that collusion among oli- gopolists can be facilitated in part by information sharing. As a consequence, the sharing of price information among rival oligopolists can violate U.S. antitrust laws. You can see how the U.S. Supreme Court has interpreted antitrust law as it pertains to sharing price information by reading a summary of the case of U.S. v. U.S. Gypsum Co. et al. (438 U.S. 422), which is available at www.stolaf.edu/ people/becker/antitrust/summaries/438us422.htm. In what manner was price in- formation shared, and why did the court find these actions to be an antitrust violation?

Case Exercise CELL PHONES DISPLACE MOBILE PHONE

SATELLITE NETWORKS21 Motorola’s Iridium, a go-anywhere mobile phone system that beamed signals down from 66 satellites, was called “the eighth wonder of the world” by Motorola CEO

21Based in part on “Apple, RIM Outsmart the Phone Market,” Wall Street Journal (July 20, 2009), p. C6.

442 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Chris Galvin. However, at $1,500 for a handset the size of a brick, consumers balked, and few business customers needed the security and reliability offered in remote cor- ners of the globe like Katmandu or Tibet. As a result, Motorola’s 25 percent market share in cell phones declined steadily to 13 percent in 2001, and Motorola stock fell 16 percent from 1997–2001, during a period when the S&P 500 was up 76 percent.

Questions 1. Characterize the product space for mobile phones when Iridium began. 2. What trends did Nokia pursue as it designed mobile phone products in the late

1990s? (Refer to the Managerial Challenge at the beginning of this chapter.) 3. What might a more proactive Motorola have done differently had it correctly per-

ceived the steps its rival Nokia would take? 4. Smart phones from Apple and RIM today have imposed upon Nokia competitive

pressure once associated with Motorola. What would you advise Nokia to do in light of the success of the iPhone with all its thousands of applications from inde- pendent software providers?

Chapter 12: Price and Output Determination: Oligopoly 443

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

13 CHAP T E R

Best-Practice Tactics: Game Theory CHAPTER PREVIEW Businesses and potential entrants into product markets who compete against a few rivals need effective tactics for best practices decision making. Effective tactics in turn require anticipating rival response and counter- response. Noncooperative simultaneous and sequential games were designed for just this purpose, including entry deterrence and accommodation games, bidding games, manufacturer-distributor games, product development or research and development games, and pricing and promotion games.

All such noncooperative games prohibit side payments and binding contracts between rivals. Instead, they depend on self-enforcing relationships to maintain strategic equilibrium. For example, each airline in a posted pricing game must decide whether it is in its own best interest to resist discounting to gain market share, based on the best reply responses it anticipates from rivals. In some circumstances, mutual discounting proves to be a dominant strategy that provides protection from renegade discounters, while in other situations, mutual forbearance leads to higher margins.

The order of play can matter in such games if credible threats and commitments influence the endgame outcomes. In this chapter we explore the role that first-mover and last-mover advantages, non-redeployable assets, credible punishment schemes, hostage mechanisms, matching price guarantees, and imperfect information can play in business strategy and tactics.

MANAGERIAL CHALLENGE Large-Scale Entry Deterrence of Low-Cost Discounters: Southwest, People Express, Value Jet, Kiwi, and JetBlue1

Since the deregulation of the U.S. airlines industry, leg- acy carriers such as United, American, US Airways, and Delta have faced a progression of low-cost competitors. Beginning with Southwest Airlines, continuing though People Express, Value Jet (now Air Tran), Kiwi, Inde- pendence Air, and now JetBlue, these firms established point-to-point operations in unserved or underserved cities and thereby created profitable business models at much lower prices. For example, from San Antonio to

Los Angeles, Southwest is profitable at $300 for last- minute walk-up one-way service whereas American Air- lines charges $520. From San Antonio to Philadelphia, Southwest’s cost-covering fare is $280, whereas Ameri- can’s is $495.

Not surprisingly, one focus of the low-cost discoun- ters has been the relentless pursuit of cost savings. Point-to-point service simplifies a carrier’s operations, and Southwest with 15-minute turnarounds gets 10.3

444

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

OLIGOPOLISTIC RIVALRY AND GAME THEORY Most oligopolistic competition takes place today in product submarkets between a few rival incumbent firms, each with some market power over price. Consider Bayer Aspirin, Bufferin, Excedrin, and St. Joseph’s in pain relievers; Pepsi and Coke in colas; Six Flags

flight hours per plane per day, 46 percent more than the industry average. By increasing capacity utilization 46 percent, Southwest lowers its indirect fixed cost per seat for interest, overhead, and time depreciation on the airframes by 31 percent (1 – 1/1.46 = 1 – 0.69). By stripping out the first-class seats and the galley, South- west increases its seating capacity by 33 percent from 90 to 119 seats on Boeing 737s, thereby also lowering direct fixed costs per seat for flight crews, fuel, and mainte- nance by another 25 percent (1 – 1/1.33 = 1 – 0.75). In addition, the low-cost start-ups have remained largely nonunionized with labor costs well below indus- try averages; as a result, legacy carriers have gone back time and time again to their machinist and flight crew unions seeking wage and salary concessions. In 2005, the total cost per available seat mile (ASM) varied from a high of 11.62 cents at Delta and 10.89 cents at US Airways to 7.70 cents at Southwest and 6.74 cents at JetBlue. American and United were in the middle at 9.80 cents and 10.12 cents, respectively.

Two prominent game theory questions in such set- tings involve the discounter’s optimal capacity choice and the legacy airline’s optimal pricing to deter or ac- commodate the entry. Southwest’s entry strategy in any new city is to offer a uniform low-price, no-frills seat with high-frequency scheduling. By unbundling all ser- vices, adopting quick turnaround times, working longer crew shifts, and converting all first-class and galley space into additional coach-class seats, Southwest typi- cally achieves 20 to 25 percent lower operating cost than American. Southwest’s prototypical target customer is a manufacturer’s trade representative who often needs to travel on short notice, but is seldom fully reimbursed through a company expense account.

In essence, a Southwest entry creates a new segment of the market not previously served by much more ex- pensive and infrequently scheduled legacy carriers (e.g., passengers who often drive rather than take the plane). These new low-willingness-to-pay customers often quickly reserve most if not all of Southwest’s capacity, leaving almost none available to other air travelers. As a result, a Southwest entry may not erode the market for

higher-priced incumbent airlines but rather creates a so- called “Southwest effect” of increased volume and load factors throughout the market, but at much lower prices. Eventually, this lower price environment drags down the profitability of all carriers.

The established airlines have to decide whether to match Southwest’s deeply discounted fares right away or accommodate Southwest by maintaining high fares. Southwest has to decide whether to enter with a large or small capacity. In this chapter, we will see how businesses use game theory reasoning to make such decisions.

Discussion Questions

� What service characteristics not offered by Southwest might you be willing to pay for at American? What about a vacationer? How would the answer be different for a trial lawyer traveling whenever the judge’s schedule per- mits or for a mid-level executive traveling on an expense account?

1Based on “UAL Hopes Latest Cost Cuts Will Yield,” Wall Street Journal (May 12, 2005), p. A10; “Southwest’s Dallas Duel,” Wall Street Journal (May 10, 2005), p. B1; and F. Harris, “Large Scale Entry Deterrence of Low-Cost Discounters: An Early Success of Revenue Management,” Inter- national Journal of Revenue Management 1, no. 1 (2007).

MANAGERIAL CHALLENGE Continued

© Ki m

St ee le /P ho to gr ap he r’s

Ch oi ce /G et ty Im ag es

Chapter 13: Best-Practice Tactics: Game Theory 445

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

and Disney in theme parks; and Delta, US Airways, and American in air travel to Florida. Smaller competitors selling more generic products are often present in periph- eral markets, but these oligopolists are notable because of some brand name or other barrier to effective entry and because of their extraordinary interdependence.

Recall that in a purely competitive industry, such as real estate development of tract home subdivisions, each competitor can act quite independently. Each takes price as “given,” that is, determined externally in the open market, because any decision to expand or embargo his or her own supply has no appreciable effect on the enormously larger industry supply. Even if one real estate developer were to purchase many subdivisions in a community, the barriers to entry are so low that any price above cost would surely attract enough new competitors to restore the competitive price-taking equilibrium.

In contrast, each firm in an oligopolistic market must pay close attention to the moves and countermoves of its rivals. Correctly anticipating entry and exit, product de- velopment, pricing, and promotions several steps ahead of actual events and at least one step ahead of the competition are often the keys to a successful business. Despite one’s best efforts, sometimes a competitor takes the lead, and then quickly adaptive behavior is preferable to reactive behavior. The best option of all is proactive behavior, and proactive behavior requires accurate and reliable predictions of rival initiatives and rival response.

The managerial purpose of game theory is to provide these predictions of rival behav- ior. To execute defensive strategy as well as plan strategic initiatives, each oligopolist must try to predict well in advance the actions, responses, and counter-responses of their rivals and then choose optimal strategies accordingly. Modern game theory was invented for precisely this purpose.

A Conceptual Framework for Game Theory Analysis A general definition of a strategy game is any consciously interdependent choice behav- ior engaged in by purposeful individuals or hierarchical groups who share a common goal (e.g., tribes, sports teams, or value-maximizing companies). As such, strategy games have always been a part of human interactions. Some of the earliest formal analyses of strategy games involved strategic voting in the Roman Senate, bargaining among Phoeni- cian traders, and the ancient Chinese military tactics of Sun Tzu.

Consider, for example, how the private property rights to a person’s belongings might evolve in a setting like the TV show Survivor. Table 13.1 displays the normal form of the strategy game in which communities of hunter-gatherers had to decide between agricul- tural pursuits combined with guarding consolidated property versus continuously hunt- ing and marauding against targets of opportunity. History records that the private property consolidators (the Aggies) won out; let’s see why.

TABLE 13.1 PRIVATIZATION OF PERSONAL BELONGINGS IN SURVIVOR :

THE MARAUDER-GUARDER GAME

Randle

Guard Maraud

Guard 1st 4th Kahn Better Worst

Maraud 2nd 3rd Worse Best

Note: Column-player payoffs are above the diagonal. Row-player payoffs are below the diagonal. Randle ranks outcomes from 1st to 4th. Kahn ranks outcomes from best to worst.

game theory A theory of interdependent decision making by the participants in a conflict-of-interest or opportunity- for-collaboration situation.

strategy game A decision-making situation with consciously interdependent behavior between two or more of the participants.

446 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Two competing players (Randle and Kahn) struggle for resources by selecting between two actions: Maraud, which occasionally yields unguarded windfall treasures but leaves one’s own possessions vulnerable to counterattack; or Guard, which frees time between defensive strug- gles for consolidating and multiplying the fruits of one’s labors with agriculture. Kahn has a tactical advantage in marauding against anything but strongly guarded positions. However, no matter what action Kahn decides to take, an examination of the payoff matrix in Table 13.1 reveals that the Aggie named Randle is always better off selecting Guard. In particular, the outcome in the NW cell (above the diagonal) is ranked first by Randle, whereas the out- come in the NE cell (again above the diagonal) is ranked fourth. Similarly, the outcome in the SW cell (above the diagonal) is ranked second by Randle, versus the outcome in the SE cell, which is ranked third. Therefore, no matter what action Kahn decides to take, Randle is always better off guarding his own consolidated property rather than marauding himself.

Guard is Randle’s dominant strategy because his outcomes from Guard exceed the out- comes from any alternative strategy, regardless of the opponent’s behavior. Knowing this fact or discovering it through trial and error, Kahn predicts his rival Randle will continue to Guard. On that presumption, Kahn then iterates back to his own choice decision and finds he prefers Guard himself. {Guard, Guard} therefore emerges as a strategic equilibrium, and the game provides a sense of how and why private property arrangements evolved.

Components of a Game The essential elements of all strategy games are present in the preceding example and include the following: players, actions, information sets, payoffs, an order of play, focal outcomes of interest, strategies, and equilibrium strategies. Let’s illustrate each compo- nent in a game of service quality competition. Suppose two copier repair players, Lanier Now and Sharp ER, must choose whether to offer fast-response copier repair service six or seven territories removed from their respective regional headquarters located in two different cities 100 miles apart (see Figure 13.1). Six versus seven territories of

FIGURE 13.1 Fast-Response Copier Service and Repair

Lanier Now

2

4

6 7

2

4

6 7

Sharp ER

100 miles

dominant strategy An action rule that maximizes the decision maker’s welfare independent of the actions of other players.

Chapter 13: Best-Practice Tactics: Game Theory 447

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

fast-response service repair are the actions, which must be announced simultaneously at next week’s industrial trade show. The payoffs from the decisions are shown in Table 13.2. This payoff matrix is the normal form of the game, which is an appropriate way of representing any simultaneous-play (versus sequential-play) game.

Sharp finds that fast-response service repair in the more distant seventh territory is ex- pensive. Cutting back to six territories reduces cost by $15 per week per customer and raises Sharp’s profit from $55 to $70 per week when Lanier also cuts back, and from $45 to $60 per week when Lanier does not. The improved effectiveness of Sharp’s service in the remain- ing six territories lowers the prices that rival Lanier can charge and reduces its profit from an initial $45 down to only $30 should Lanier continue servicing all seven territories. By cutting back to six territories itself, Lanier can restrict its losses to just $5 ($45 now to $40 > $30). The common information set known to both players includes knowledge of all these payoffs.

What strategy should Lanier adopt? First, using the concept of dominant strategy, it is clear that Sharp ER will discontinue service in the seventh territory. Sharp is better off cutting back to six territories independent of what Lanier does. For Sharp, seven territo- ries is dominated (unambiguously less preferred than six territories). Lanier wishes it were not so, because its most successful operation entails head-to-head, seven-territory competition against Sharp. Nevertheless, predictable reality lies elsewhere, and Lanier must predict six-territory behavior on the part of its rival and proceed to reexamine its remaining options. Having eliminated Sharp’s dominated strategy in the second column, Lanier now has an unambiguously preferred strategy of providing fast-response repairs in only six territories itself. {Six, Six} is therefore the equilibrium strategy pair. That is, by applying the concept of a dominant strategy equilibrium to the prediction of its rival’s behavior, Lanier can iterate back to analyze its own best action. {Six, Six} is therefore referred to as an iterated dominant strategy equilibrium.

This concept of eliminating dominated strategies in simultaneous games and then iter- ating back to one’s remaining choices first appeared in The Theory of Games and Eco- nomic Behavior by John von Neumann and Oskar Morgenstern.2 Von Neumann and Morgenstern confined their analysis primarily to cooperative games, in which players can form coalitions, arrange side payments, and enter into binding agreements. John Nash, Reinhard Selten, and John Harsanyi won the 1994 Nobel Prize in Economic Sciences for their extension of strategic equilibrium concepts to noncooperative games, sequential games, and games of imperfect information. John Nash’s life was celebrated in the book A Beautiful Mind by Sylvia Nasar and the subsequent movie starring Russell Crowe.

TABLE 13.2 SIX OR SEVEN TERRITORIES?

Sharp

Six Territories Seven Territories

Six $70 $55 Lanier TTTerritories $40 $35

Seven $60 $45 Territories $30 $45

Note: Payoffs are profits. Sharp payoffs are above the diagonal, and Lanier payoffs are below the diagonal.

normal form of the game A representation of payoffs in a simultaneous-play game.

iterated dominant strategy An action rule that maximizes self- interest in light of the predictable dominant- strategy behavior of other players.

2Two other useful volumes on modern game theory are A. Dixit and S. Skeath, Games of Strategy (New York: Norton, 1999); and Eric Rasmussen, Games and Information, 2nd ed. (Cambridge, MA: Basil Blackwell, 1993).

448 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Cooperative and Noncooperative Games The fact that in a cooperative game players can form coalitions, make side payments, and communicate to one another their private information about their own prices, profit margins, or variable costs has limited the usefulness of cooperative game theory in busi- ness settings. An illustration of a side payment in cooperative games is the mandatory compensation scheme a manufacturer might impose when one sales representative vio- lates another’s exclusive territory. Or, suppose in the previous Lanier Now and Sharp ER example that the two firms got together to arrange a side payment to ensure a strategic equilibrium of {Six, Six}. As you may already suspect, most such cooperative game agree- ments between arm’s-length competitors to exchange price information or arrange side payments are per se violations of the antitrust laws in the United States and Western Europe.3 For these reasons, business strategists paid relatively little attention to game theory until noncooperative strategic equilibrium concepts were developed.

Noncooperative games prohibit collusive communication, side payments, and third- party enforceable binding agreements. Instead, such games focus on self-enforcing reli- ance relationships to characterize strategic equilibrium and predict rival response. One example we already encountered in Chapter 10 is the mutual reliance between buyers of high-priced experience goods such as used cars and sellers with non-redeployable assets (e.g., CarMax advertising). Other examples include computer companies who build oper- ating systems to a common standard that can communicate across PC platforms, or competing airlines who announce high fares day after day despite the quick but short- lived attraction of breaking out as a renegade discounter. Clearly, these noncooperative games differ from cooperative games in important ways that make them more applicable to business strategy.

Other Types of Games Strategy games are also classified according to the number of players involved, the com- patibility of their interests, and the number of replays of the game. We analyzed both prior games as single-period (“one-shot”) games, but ongoing rivalry between the players in “Guarder-Marauder” and in “Six or Seven Territories” is highly pertinent to the stra- tegic situation. We will turn our attention next to the distinct and somewhat paradoxical implications of so-called repeated games. In a two-person game, each player attempts to

Example The Nobel Prize Goes to Three Game Theorists Nash, Selten, and Harsanyi won a 1994 Nobel Prize for their work on equilibrium strategies in sequential games ranging from chess and poker to central bank inter- ventions, limit pricing to deter entry, research-and-development competitions, and the auctioning of the radio magnetic spectrum. Not infrequently, multiple equilib- ria arise in such games. Another implication is that the order of play can affect strategic decisions; moving first in a preemptive product development can often bar a competitor’s threatened entry. In other circumstances, making the last re- sponse in the endgame, as dynamic technology takes a new direction, can secure a strategic advantage. Distinguishing between these and other complex paths to the most profitable strategy is the role of equilibrium strategies in game theory.

cooperative game Game structures that allow coalition formation, side payments, and binding third-party enforceable agreements.

3For example, the antitrust opinions in U.S. v. National Gypsum, 428 U.S. 422 (1978) and U.S. v. Airline Tariff Publishing Co., et al., 92-52854 (1992) expressly prohibited the exchange of preannouncement price lists between competitors.

noncooperative games Game structures that prohibit collusion, side payments, and binding agreements enforced by third parties.

Chapter 13: Best-Practice Tactics: Game Theory 449

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

obtain as much as possible from the other player through whatever methods of coopera- tion, bargaining, or threatening are available. N-person games are more difficult to ana- lyze because subsets of players can form coalitions to impose solutions on the rest of the players. Coalitions can be of any size and can break up and re-form as the game pro- ceeds. Parliamentary government is the classic example of n-person games. Although the possibility of coalitions adds greatly to the richness of the types of situations that can be considered by game theory, coalitions add substantial complexity to the theory required to analyze such games.

In a two-person zero-sum game, the players have exactly opposite interests; one player’s gain is the other player’s loss, and vice versa. “Guarder-Marauder” serves as an intuitive example. Although a number of parlor games and some military applications can be analyzed with zero-sum games, most real-life conflict-of-interest situations do not fit within this category. In contrast, in a two-person non-zero-sum game, both players may gain or lose depending on the actions each chooses to take. “Six or Seven Territories” is a non-zero-sum game; limiting competition to six territories raises the total profit from the interaction to $110 rather than $90. In all such games at least one outcome is jointly pre- ferred, and consequently, the players may be able to increase their payoffs through some form of coordination. Perhaps the most famous generic structure for non-zero-sum games is the Prisoner’s Dilemma. Many real-world situations, such as duopoly pricing between Pepsi and Coke, a used car sales transaction, and bargaining with channel partners in manufacturer-distributor games, can be represented as a Prisoner’s Dilemma.

ANALYZING SIMULTANEOUS GAMES The Prisoner’s Dilemma In the Prisoner’s Dilemma, two suspects are accused of jointly committing a crime. To convict the suspects, however, a confession is needed from one or both of them. They are separated and no information can pass between them, making it a noncooperative game. If neither suspect confesses, the prosecutor will be unable to convict them of the crime and each suspect will receive only a short-term (1-year) prison sentence. If one suspect con- fesses (i.e., turns state’s evidence) and the other does not, then the one confessing will re- ceive a suspended sentence, and the other will receive a long 15-year prison sentence. If both suspects confess, then each will receive an intermediate 6-year prison sentence. Each suspect must decide, under these conditions, whether to confess. This conflict-of-interest situation can be represented in a game matrix such as the one shown in Table 13.3.

TABLE 13.3 PRISONER ’S DILEMMA PAYOFF MATRIX

Suspect 2

Not Confess Confess

1-year prison term for each suspect

Suspended sentence for Suspect 2 15-year prison term for Suspect 1

Not Confess

Suspect 1 15-year prison term for Suspect 2 Suspended sentence for Suspect 1

6-year prison term for each suspect

Confess

two-person zero-sum game Game in which net gains for one player necessarily imply equal net losses for the other player.

450 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

This game can be examined by using the concept of a minimum security level that arises when players “worst case” the situation. A maximin strategy then selects the max- imum payoff when worst-case scenarios arise. For Suspect 2 (the column player), the minimum payoff from his choosing “Not Confess” is a 15-year prison sentence arising when Suspect 1 confesses (in the bottom row) and the minimum payoff from his choos- ing “Confess” is a 6-year prison sentence arising again when Suspect 1 confesses. So, maximizing the security level would therefore motivate Suspect 2 to choose the second alternative of confessing in order to avoid the possibility of a still worse outcome from not confessing. Similar reasoning holds true for Suspect 1, and she also would be moti- vated to choose the alternative of confessing her guilt. Thus, the “Confess” alternative dominates the other strategy “Not Confess” and {Confess, Confess} constitutes a domi- nant strategy/equilibrium strategy pair and isolates the predictable solution of the players.4

In all such Prisoner’s Dilemma games, both suspects would clearly receive a better payoff (i.e., a shorter sentence) if they both decided to choose their first alternatives (“Not Confess”). However, in seeking to maximize their predictable payoffs (or, more accurately, to maximize their security levels), the first alternative is not a rational choice for either suspect. The players could of course agree in advance to maintain their inno- cence. But without strong sanctions to force each other to adhere to the agreement, each would be tempted to double-cross the other by confessing his or her guilt. Remember that whichever suspect breaks the agreement first has the possibility of reducing his or her sentence from a six-year prison term to a suspended sentence.

The analogy to pricing and output decisions among firms in oligopolistic industries is striking. Suppose two cruise lines—Carnival Cruise and Royal Caribbean (RC)—operate the only three-day Caribbean cruises from Miami. If each firm acts independently to maximize its own profits, the long-run (Cournot equilibrium) profit-maximizing price is $300 per person. If two firms act jointly to maximize total industry profits, the profit-maximizing price is $450. Assume that these two prices are the only prices under consideration.

Both firms must decide their action without knowing their rival’s decision, which is the essence of a simultaneous game. Although sequential game reasoning is critical to the successful conduct of many business strategies, some decisions must be made simul- taneously with one’s rivals. Consider offers in a silent auction, release dates for fashion clothing collections, promotional ads to meet a newspaper deadline, and posted price an- nouncements at an electronic clearinghouse sponsored by the airline or cruise ship industry.

The payoffs to each cruise ship firm are shown in Table 13.4. The below-diagonal number in each cell is the payoff to Royal Caribbean, and the above-diagonal number is the payoff to Carnival. Each firm is reluctant to choose the (jointly) more profitable $450 price. If either firm reneges and discounts to $300, then the firm that charges $450 will earn significantly lower profits than the rival. This game has a typical Prison- er’s Dilemma ordering of outcomes. As we have seen, unilaterally cooperating by an- nouncing high prices under such circumstances is foolish. For example, the payoff for Carnival from unilateral defection ($375,000) exceeds the payoff from mutual coopera- tion at high prices ($275,000), which itself exceeds the payoff from mutual defection at low prices ($185,000), which finally exceeds the payoff from unilateral cooperation ($60,000). Therefore, Carnival’s dominant strategy is to defect.

maximin strategy A criterion for selecting actions that minimize absolute losses.

4Maximin strategy will often yield actions that are not aligned with dominant strategy equilibrium if the deci- sion maker is focused on maximizing gains or expected value of net gains rather than just minimizing absolute losses. A related strategy focuses on the minimization of opportunity losses, sometimes called minimax regret.

simultaneous game A strategy game in which players must choose their actions simultaneously.

sequential game A game with an explicit order of play.

Chapter 13: Best-Practice Tactics: Game Theory 451

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Royal Caribbean has no such dominant strategy. However, because RC can predict Carnival’s behavior, by eliminating the prospect of Carnival’s dominated $450 strategy, RC can iterate to a preferable strategy itself. Therefore, Royal Caribbean’s behavior is also quite predictable, and the iterated dominant strategy equilibrium proves to be {$300, $300} or {Defect, Defect} just as in Prisoner’s Dilemma itself. The Prisoner’s Dilemma facing Royal Caribbean and Carnival is a noncooperative positive-sum game of coordination. In the next section we will study how to escape the dilemma by changing the structure of such games.

Dominant Strategy and Nash Equilibrium Strategy Defined Note that a dominated strategy is not necessary for both cruise ship companies to reach an iterated dominant strategy equilibrium. The reason is that a dominant strategy re- quires no particular optimal or suboptimal response behavior on the part of anyone else. It is defined as an action for player i that is an optimal action fa*i g in the strong sense that no matter what other players do, the payoff for player i, Πifa*i ; a−ig exceeds the payoff for player i from any other action, Πifai; a−ig5

Πifa*i , a−ig > Πifai, a−ig [13.1]

Consequently, one dominant strategy is quite enough to predict rival behavior and there- fore identify the strategic equilibrium in any two-person one-shot simultaneous game. Once Carnival’s dominant strategy (i.e., to defect and cut prices to $300) has been iden- tified, Royal Caribbean’s behavior (i.e., to also defect) is easily predictable. We have seen this outcome in “Six or Seven Territories?” and in “Marauder-Guarder.”

What about simultaneous games without any dominant strategy? To examine this question, we now turn to a one-shot simultaneous price announcement game between PepsiCo and Coca-Cola. Each week both firms must choose whether to maintain or dis- count in their grocery store distribution channels. The payoffs per week per store are displayed in Table 13.5. As shown in the northeast cell, if Coca-Cola unilaterally defects to discounting, then Coca-Cola’s payoff increases from $13,000 to $16,000, while Pepsi- Co’s payoff declines by 25 percent from $12,000 to $9,000. Similarly, PepsiCo can turn the tables on Coca-Cola by unilaterally discounting to increase operating profits by 16 percent from $12,000 to $14,000, while Coca-Cola profits would decline from $13,000

TABLE 13.4 CRUISE SHIP PRICING WITH DOMINANT STRATEGY

Carnival

$450 $300

$450 $275 $375

Royal Caribbean

$350 $50

$350 $60 $185

$320 $175

Note: Column-player payoffs (in thousands) are above the diagonal. Row-player payoffs are below the diagonal.

5A starred action refers to a maximizing choice; here, it is an action that results from maximizing profit.

452 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

to $10,500. Table 13.5 contains no dominant strategy. PepsiCo wants to discount when Coca-Cola maintains higher prices ($14,000 > $12,000), but just as clearly, PepsiCo wants to maintain higher prices when Coca-Cola discounts ($9,000 > $6,300). And the same con- tingent ambiguity is present for Coca-Cola. What criteria allow the prediction of rival be- havior in this game of “Renegade Discounting”?

The answer lies in a reflexive application of the concept of best-reply response. If an action were the best reply to a rival’s action, which in turn was the best reply to the orig- inal action, the parties would have identified an equilibrium strategy. More formally, a Nash equilibrium strategy is defined as an action for player i that is conditionally optimal fa*i g in that the payoff for player i, given best-reply responses by rivals Πifa*i ,a*−ig, exceeds the payoff for player i from any other action Πifai; a*−ig given best- reply responses of rivals:

Πifa*i , a*−ig > Πifai, a*−ig [13.2] In Renegade Discounting, the two pure Nash equilibria are {Maintain*p, Discount*c}

and {Discount*p, Maintain*c}, where the subscripts refer to PepsiCo and Coca-Cola. Recall that the order of play is not important in this game; we could have just as easily reversed these strategy pairs and listed Coca-Cola rather than PepsiCo first. The actual rivals ap- pear to have perceived precisely this point because, for 42 weeks in 1992, they took turns discounting on the endcaps in grocery stores across America.

What is notable about these Nash equilibrium strategies is that they are non-unique. The multiple equilibria occur because the Nash equilibrium concept is less demanding (i.e., easier to satisfy) than dominant strategy equilibrium. The latter requires that an ac- tion be optimal for every possible rival response, whereas Nash equilibrium requires only that an action be optimal for a best-reply rival response. However, this knowledge does not help solve PepsiCo’s problem as to what price to announce next. Remember that each bottler is announcing its price without knowing until afterwards what its rival announced.

If PepsiCo believed Coca-Cola would discount half the time and maintain half the time, the expected value of PepsiCo’s maintaining is $10,500 (namely, 0.5 × $12,000 + 0.5 × $9,000), whereas the expected value of PepsiCo’s discounting is smaller (i.e., only $10,150). These results would seem to suggest a preference for maintaining high prices, but again, if PepsiCo kept its prices predictably high, Coca-Cola could unilaterally defect and earn $16,000, whereas PepsiCo would then realize only $9,000. So how can PepsiCo avoid tipping its hand and ending up with the $9,000 outcome rather than its own $14,000 defection outcome too often?

TABLE 13.5 RENEGADE DISCOUNTING IN SOFT DRINKS WITH NO

DOMINANT STRATEGY

Coca-Cola

Maintain High Prices Discount Low Prices

$13,000 $16,000 Maintain High Prices $12,000 $9,000

PepsiCo $10,500 $8,000

Discount Low Prices $14,000 $6,300

Note: Column-player payoffs (in thousands) are above the diagonal. Row-player payoffs are below the diagonal.

best-reply response An action that maximizes self-interest from among feasible choices.

Nash equilibrium strategy An equilibrium concept for nondominant strategy games.

Chapter 13: Best-Practice Tactics: Game Theory 453

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The answer lies in PepsiCo’s randomizing the pricing process. PepsiCo must figure out what automated pricing response would make Coca-Cola indifferent between main- taining and discounting and thereby willing to randomize its own price announcement. That is, what probability of discounting by PepsiCo would make Coca-Cola indifferent by equating Coca-Cola’s expected payoff from maintaining to its expected payoff from discounting? Interestingly, because the payoffs are asymmetrical, the desired probability is not 0.5. Let’s see what the solution is. Using p to represent the probability that PepsiCo maintains and (1 – p) if it discounts, we calculate

ðpÞ$13;000 + ð1 − pÞ$10;500 = ðpÞ$16;000 + ð1 − pÞ$8;000 [13.3] where the Coca-Cola payoffs are arranged to correspond to the columns of Table 13.5. The solution probabilities p = 0.454 and (1 – p) = 0.546 make Coca-Cola indifferent and therefore PepsiCo less vulnerable to unilateral defection.

Note the mirror-image reflexivity associated with this Nash equilibrium solution: Coca-Cola faces a comparable payoff structure and strategy dilemma to that of PepsiCo, and would presumably want to know the probabilities of maintaining and discounting that would make PepsiCo indifferent between the two choices. Calculating as before

ðp0Þ$12;000 + ð1 − p0Þ$9;000 = ðp0Þ$14;000 + ð1 − p0Þ$6;300 [13.4] where the PepsiCo payoffs are arranged to correspond to the rows of Table 13.5, we obtain p0 = 0.574 and (1 – p0) = 0.426. If randomized choice by PepsiCo is a best-reply response to Coca-Cola, and if Coca-Cola can then do no better, this renegade discounting game must have a third Nash equilibrium strategy—namely, {Maintain by PepsiCo with p = 0.454, Maintain by Coca-Cola with p0 = 0.574}. This strategy pair is called a mixed Nash equilib- rium strategy. A 0.454 probability weight on maintaining and a 0.546 probability weight on discounting by PepsiCo yields $11,634 expected value for each of Coca-Cola’s price an- nouncement strategies. Similarly, a 0.574 probability weight on maintaining and a 0.426 probability weight on discounting by Coca-Cola yields $10,720 expected value for each of PepsiCo’s price announcement strategies. The strategic equilibrium solution for this game therefore contains two pure and one mixed Nash strategy: {Maintain*p, Discount*c}, {Discount*p, Maintain*c}, and {Maintain*p = 0.454, Maintain*c = 0.574}.

Using a computer program that randomizes an unfair coin toss is one way to imple- ment this mixed Nash strategy. In principle, however, none of these three Nash equilib- rium strategies is preferable to any other. In a one-shot play of Renegade Discounting, all four cells in Table 13.5 still arise. The {$6,300, $8,000} outcome in the southeast cell and the {$12,000, $13,000} outcome in the northwest cell, as well as the two asymmetric out- comes that correspond to our two pure Nash strategies, will all sometimes arise. In a non- cooperative simultaneous one-shot game that allows no communication in advance, no side payments, and no binding agreements, players simply cannot avoid this multiplicity of possible strategic equilibria. In practice, therefore, a one-shot play of any of the three Nash strategies in the Renegade Discounting game can work out well or badly.

Of course, the {$12,000, $13,000} outcome is best of all. In the next section we will see how to secure this win-win outcome by introducing repeated plays, imperfect informa- tion, and credibility mechanisms to convert this simultaneous game into a sequential game. Barry Nalebuff, Yale professor and author of the widely read Thinking Strategi- cally, calls it “changing the nature of competition” and distinguishes it from “collusion,” which would violate the antitrust laws.6

mixed Nash equilibrium strategy A strategic equilibrium concept involving randomized behavior.

6See A. Dixit and B. Nalebuff, Thinking Strategically (New York: Norton, 1993); and “Businessman’s Di- lemma,” Forbes (October 11, 1993), p. 107.

454 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

THE ESCAPE FROM PRISONER’S DILEMMA Multiperiod Punishment and Reward Schemes in Repeated Play Games In this section, we relax the assumptions of single-play, complete, and perfect informa- tion games. Let’s again look at the PepsiCo and Coca-Cola example with the Prisoner’s Dilemma payoff structure as shown in Table 13.6. Both PepsiCo and Coca-Cola are worse off if either unilaterally defects from maintaining high prices. Each soft drink bot- tler would like to pursue the $12,000 payoff, but the only way to avoid the vulnerability of a unilateral defection is by defecting oneself! Dominant strategy drives both players to discount their 12-packs in the one-shot game. However, surely PepsiCo and Coca-Cola recognize they are engaged in an ongoing competitive process, not a one-shot (i.e., single-play) game. Week after week, they will encounter each other in many future re- plays of this pricing game at grocery and convenience stores nationwide. Consequently, tacit cooperation rather than dogmatic price cutting has a chance to evolve.

Suppose Coca-Cola begins the process by announcing a high price in Period 1. Coke’s intention is to play that price continuously until PepsiCo defects and thereafter to never announce High again, which is a so-called grim trigger strategy. Any move by PepsiCo away from cooperative High pricing, and Coca-Cola’s punishment is immediate and never-ending. Multiperiod punishment schemes are a key to inducing cooperation in Pris- oner’s Dilemma games, whether it is cruise ship, airline, or soft drink companies. In this case, PepsiCo compares the perpetuity opportunity loss of ($12,000 – $8,000) discounted at the interest rate r per period to the one-time gain from defection of ($17,000 – $12,000):

$4,000=r > $5,000 if r < 0:8 [13.5]

The interpretation is straightforward. At any discount rate less than 80 percent, PepsiCo’s future gains from cooperatively maintaining high prices outweigh the one-time gains from defection. Thus, the dominant strategy to defect in one-shot games is no longer attractive. This calculation and conclusion reflect a generalizable Folk theorem, which states that for any payoff structure, a discount rate always exists that is low enough to induce cooperation in an infinitely repeated Prisoner’s Dilemma. So, a grim trigger strategy can induce cooperation in an infinitely repeated Prisoner’s Dilemma.7

TABLE 13.6 REPEATED PRISONER ’S DILEMMA IN SOFT DRINKS

Coca-Cola

Maintain High Prices Discount

$12,000 $17,000 Maintain High Prices

$12,000 $6,000 PepsiCo

$6,000 $8,000 Discount $17,000 $8,000

trembling hand trigger strategy A punishment mechanism that forgives random mistakes and miscommunications.

grim trigger strategy A strategy involving infinitely long punishment schemes.

Folk theorem A conclusion about cooperation in repeated Prisoner’s Dilemma.

7Of course, one transparent disadvantage of grim trigger strategies is that cooperative outcomes cannot survive a single small mistake in reasoning or miscommunication by either player. Selten’s concept of a trembling hand trigger strategy allows one grace period of misplay by the other party before imposing the grim punish- ment for defection. Of course, a wily rival who understands this strategy will take advantage of his opponent by claiming just as many one-period “mistakes” of defection as he can get away with.

Chapter 13: Best-Practice Tactics: Game Theory 455

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

However, because companies do not last forever, the Folk theorem raises an obvious question, “What about for shorter periods, say 20 weeks?” The 20-period calculation is easily done; r now must be less than 79 percent. But if 20 weeks, what about 10, and if 10, what about for 2 weeks? Suppose it is now the beginning of Week 2. We know we are out of this “cooperative” structure next week (i.e., Week 3), so our remaining incentive to maintain high prices is only $4,000/(1 + r), and our incentive to defect is $5,000. Now all of a sudden, for any discount rate, each player is better off defecting. This result is also generalizable. The last play of a finitely repeated Prisoner’s Dilemma has the same incen- tives as a one-shot Prisoner’s Dilemma; everybody defects. Therefore, one period away from the endgame of a finitely repeated Prisoner’s Dilemma, neither party has an incen- tive to maintain its reputation for cooperating.

Unraveling and the Chain Store Paradox The prospects for cooperation in any finitely repeated Prisoner’s Dilemma are poor, because what is true for a 2-period game must be true by backwards induction for a 3-period game. If you know in the 2-period game that it pays to defect, then in the 3-period game you must know that a certain defection is only one period away; there- fore, you should defect now. And if that is so for a 3-period game, then so too for a 4-period game, and so on, even for a 20-period game. Reinhard Selten investigated this unraveling problem for finitely repeated Prisoner’s Dilemmas in the context of chain store incumbents facing repeated entry threats from rivals.8 In a Prisoner’s Dilemma set- ting like those we’ve been examining, the established firm has a dominant strategy to accommodate the new entrant. But one’s intuition says that in the face of enough repeti- tions of the chain store competition, the established firm’s reputation for fighting entry can pay off. And in the extreme, this intuition is absolutely correct. In infinitely repeated games, the Folk theorem does apply. However, with any fewer repetitions, in even the enormous number of chain store competitions that might face a McDonald’s or a Walmart, the cooperative equilibrium unravels.

Reinhard Selten invented the concept of endgame reasoning to show this paradoxi- cal result and to emphasize the sequential nature of reputation effects. Endgame rea- soning always entails looking ahead to the last play in an ordered sequence of plays, identifying the player whose decisions will control the outcome of the endgame, and then predicting that player’s best-reply response. In Figure 13.2 we have a chain store Incumbent (I) who accommodates or fights in response to a Potential Entrant (PE) who stays out or enters. Accommodation forgoes $20,000 of the incumbent chain store’s profit ($100,000 – $80,000) and induces future entry, but fighting the present entry to acquire a reputation for toughness in future possible entry situations entails actual losses now (–$10,000). Conceive of the displayed game tree as the last three en- counters of a 20-chain store competition perceived by both players from the start. Looking ahead to the endgame, it is clear that the incumbent will accommodate in the last submarket. At decision C, the $100,000 to accommodate exceeds the $60,000 to fight, and the $80,000 to accommodate at decision node B exceeds the –$10,000 to fight. More importantly, a tough reputation gains no future payoff thereafter because it is truly the endgame. Since the Potential Entrant also knows it is the endgame, entry will surely take place in that last submarket.

unraveling problem A failure of cooperation in games of finite length.

8See J. Harsanyi and R. Selten, A General Theory of Equilibrium Selection in Games (Cambridge, MA: MIT Press, 1988); or for a less technical treatment, E. Rasmussen, Games and Information, 2nd ed. (Cambridge, MA: Blackwell, 1994), Chapter 5.

infinitely repeated games A game that lasts forever.

endgame reasoning An analysis of the final decision in a sequential game.

456 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Now, in looking back to the previous submarket (i.e., the 19th store), the incumbent realizes that its rival’s subsequent entry in the 20th submarket is certain and therefore that, again, any attempt to acquire an enhanced reputation for fighting is useless in that 19th submarket. Accommodate is therefore the best-reply response in the proper sub- game from node B onward to the endgame. Because the entrant can predict this decision as well, entry occurs in the 19th submarket at node A. But what is true of the 19th must therefore be true of the 18th, and the 17th, and so forth, right back to the start of the game.

This backwards induction reasoning leads to the chain store paradox. We can calcu- late in Submarket 1 that at reasonable rates of discount the incumbent may well have a sufficient net present value of profits from deterring future entry to justify fighting now rather than accommodating. Yet, the predictability of its future accommodation jeopar- dizes the credibility of the incumbent’s present fighting. This predictability of selecting accommodation as a best-reply response all the way out to the endgame means the rep- utation effects of any present fighting unravel. Accommodation therefore occurs in every submarket or every period in the 20-submarket/20-period game just as we argued earlier it would in the 2-submarket/2-period game.

Our intuition tells us otherwise, especially when a long line of potential entrants is waiting in the wings—hence, the term chain store paradox. And, as we shall now see, changing some other features of the chain store decision problem can overturn this counterintuitive result.

FIGURE 13.2 The Chain Store Paradox

PE

I

I

I

I

PE

I

I

PE

($80, $10)*

Fight (–$10, –$10)

($100, 0)

($60, 0)

N.A.

N.A.

N.A.

N.A.

G

F

Stay Out

H

E

Enter

I Stay Out

D ($80, $10)*

Fight

(–$10, –$10)

B

C

Enter

($100, 0)

Fight

($60, 0)

Stay Out

A Accommodate

($80, $10)*

18th Store 19th Store 20th Store

Note: Payoffs are listed (Incumbent, Potential Entrant) in thousands. N.A. (not applicable).

Accommodate

Accommodate

Accommodate

Accommodate

Accommodate

Fight

Fight

Fight

Enter

Accommodate

chain store paradox A prediction of always- accommodative behavior by incumbents facing entry threats.

Chapter 13: Best-Practice Tactics: Game Theory 457

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Mutual Forbearance and Cooperation in Repeated Prisoner’s Dilemma Games One way to short-circuit the reasoning of the chain store paradox is to introduce an un- certain ending of the game. If the incumbent can never be sure whether future encoun- ters beyond Submarket 20 will arise, then the reputation effect of fighting in the 19th period returns. Any positive probability that the game will continue is sufficient (again, at low enough discount rates) to restore the deterrent effect of fighting in Period 20. If fighting is rational in Period 20, then the incumbent is willing to fight in 19, 18, and so forth, back to Period 1. And if the incumbent is willing in Period 1, then it may not have to because the other firm will not enter. The analogous implication in a finitely repeated pricing game such as Repeated Prisoner’s Dilemma in Soft Drinks (in Table 13.6) is that the rivals will cooperate by maintaining high prices as long as the endgame is uncertain. With one period remaining, we can then write Equation 13.5 as

$4,000 + $4,000 × 1

ð1 + rÞ × p > $5,000 [13.6]

where p is the probability of the game continuing beyond the next period. For r = 0.1, a probability as low as 0.28 is sufficient to elicit cooperation in maintaining high prices and a {$12,000, $12,000} northwest cell outcome in Table 13.6. Therefore, infinite repetition is not required to induce cooperation in Prisoner’s Dilemmas; an uncertain ending will suffice.

Example Solving the Chain Store Paradox: Semiconductor Pricing at Intel, NEC, and AMD9

These insights seem especially important in industries with fast-changing technol- ogy, such as computer chips and consumer electronics, where cost disadvantages that might end an incumbent’s business are seldom permanent because the tech- nology changes so often. One illustration is the semiconductor industry, where In- tel, AMD, and Motorola have recently returned to dominance after almost being displaced by Japanese firms such as Hitachi, NEC, and Toshiba in the 1990s.

With each successive generation of chips (see Table 13.7), the market leaders practice life cycle pricing techniques. After a period of high target pricing and

TABLE 13.7 MAJOR INTEL MICROPROCESSORS AND CLOCK SPEED

YEAR MICROPROCESSOR MHZ

1979 8088 5

1982 286 6

1985 386 16

1989 486 25

1993 Pentium 60

1995 Pentium Pro 150

1997 Pentium II 233

1999 Pentium III 333

2002 Pentium IV 550

2004 Celeron M 1200

2005 Pentium M 1600

2006 Core2 Duo 2130

458 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Bayesian Reputation Effects A second ingenious escape from Prisoner’s Dilemma incorporates Bayesian reputation effects about opponent type, based on the work of Nobel laureate John Harsanyi. It in- volves estimating the likelihood of various opponent moves based on past events. If some irrational “crazies” who do not always maximize their payoffs are known to exist in the market, a perfectly sane incumbent might end up taking actions that seem crazy. The intent of the incumbent is to secure a disreputation in which the incumbent is indistin- guishable from the crazies.10 An example would be an automobile manufacturing incum- bent that “predates”—that is, prices its product below its variable cost, even though the operating losses from such a strategy may not be recoverable in excess profits later. Jap- anese automobile manufacturers are often accused of such “dumping” in the offshore auto markets, especially in Europe.

Winning Strategies in Evolutionary Computer Tournaments: Tit for Tat Robert Axelrod was intrigued by the reasons why people who are ardently pursuing their own goals often end up cooperating with competitors in long-term interactions.11 He in- vestigated the question of optimal strategy in repeated Prisoner’s Dilemma by conduct- ing a computer simulation in which 151 strategies competed against one another 1,000 times. He discovered that those strategies that finished highest in the computer tourna- ment had several characteristics in common. First, winning strategies are clear and simple in order to avoid triggering mistakes by one’s potential cooperators. Second, win- ning strategies make unilateral attempts to cooperate; they never initiate defection—just the reverse—they initiate niceness. Third, as we would expect, all winning strategies are provokable—they have credible commitments to some punishment rule. Interestingly, limited-duration punishment schemes that displayed forgiveness won out over maximal-punishment grim trigger strategies that do not. The reason seemed to be that winning strategies recover from misperceptions, miscommunications, and strategic mis- takes; reprisals need not become self-perpetuating.

value-based pricing, Intel, with more than 70 percent of the worldwide market, limits price rather than accommodate AMD with 21 percent, Motorola with 5 per- cent, and numerous smaller competitors. That is, chip prices are slashed in an at- tempt to deter entry by the imitators. Then, with uncertain timing, the whole process repeats itself. New chips are introduced at high prices, imitators reverse engineer the design, and limit pricing begins again. The uncertain endpoint of the successive chip generation games leads to a solution to the chain store paradox and an increased likelihood of higher prices, just as in soft drink pricing between Coca- Cola and PepsiCo.

9Based on Investor’s Business Daily (January 13, 1998), p. A8.

10On the asymmetric information pooling equilibrium behind this strategy, see R. Gibbons, “An Introduction to Applicable Game Theory,” Journal of Economic Perspectives 11, no. 1 (Winter 1997), pp. 140–147; and Rasmussen, op. cit., pp. 352–356. 11Robert Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984). See also “Evolutionary Eco- nomics,” Forbes (October 11, 1993), p. 110; and Jill Neimark, “Tit for Tat: A Game of Survival,” Success (May 1987), p. 62.

Chapter 13: Best-Practice Tactics: Game Theory 459

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

What types of actual strategies would you guess best fit these winning criteria? Surpris- ingly, “Tit for Tat” won the tournament! Repeating what your opponent did on the last round is simple and clearly provokable, but consistent with unilaterally initiating coopera- tion yourself. And perhaps most importantly, Tit for Tat is forgiving. After a single-period punishment, it reverts to cooperating as soon as the opponent/cooperator does so.

For example, one possible approach to managing competition for cruise ship compa- nies Carnival and Royal Caribbean (RC) in Table 13.8 is to follow a Tit-for-Tat (TFT)

Example Brown and Williamson’s Reputation for Predating The U.S. Supreme Court has addressed these issues and thereby set a standard for judging predatory pricing behavior by U.S. firms. In Brooke Group Ltd. v. Brown and Williamson Tobacco Company, 113 U.S. 2578 (1993), the Court held that pric- ing generic cigarettes below cost was not evidence of an undesirable predatory in- tent to monopolize a market because Brown and Williamson had no opportunity thereafter to earn excess profits and recoup its losses from the alleged predatory conduct. Similarly, when Kodak priced its Instamatic film camera at $11.95 despite a $28 direct manufacturing cost, it faced little prospect of later recovering the $16.05 operating loss per camera. Instead, this pricing tactic was reasonable as an attempt to clear inventory quickly before exiting the Instamatic submarket.

Whether the Court in these cases looked deeply enough into the long-term ef- fect of deterring effective entry through disreputation effects is a hotly debated an- titrust issue. Under incomplete information about opponent types, behaving like a “crazy,” who predatorily prices below cost when recovering the losses is unlikely, may deter an opponent’s entry. This disreputation effect of becoming known as a firm who might well price below cost is more valuable when the cost of new en- trants is high, brand loyalty to incumbents is weak, and the number of potential entrants is large.

American Airlines discouraged the discounters Vanguard, Sunjet, and Western Pacific from remaining in the Dallas–Ft. Worth airline market using such tactics. In May 2001, American’s indictment for predation was dismissed on the grounds that at no point did American lower price below its average variable cost. Hard- nosed competitive tactics that remain within the boundaries of recovering average variable cost are legal, because antitrust law exists to protect competition (encour- aging lower prices for consumers), not to protect individual competitors.

TABLE 13.8 CRUISE SHIP PRICING WITH PRICE MATCHING

Carnival Pricing Policy

$450 $300 “Match”

$450 $275 $375 $275

Royal $350 $150 $350

Caribbean $300 $160 $185 $185 Pricing $320 $175 $175 Policy “Match” $275 $185 $275

$350 $175 $350

Note: Column-player payoffs above the diagonal in $ thousands.

460 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

decision rule. Royal Caribbean, who has a dominant $300 strategy, could signal a conspicuous focal point by promoting “staterooms” (rather than smaller, less well- appointed “cabins”) as an industry standard and then choosing the $450 pricing strategy in the first period. Thereafter, Royal Caribbean would select the same pricing strategy in the next period as Carnival chose in the previous period. For example, if Carnival charges $450 in the current period, then Royal Caribbean would do likewise in the next period. On the other hand, if Carnival defects and charges $300 in the current period, then Royal Caribbean would retaliate by charging the same $300 price next period. Through repeated plays, the participants may learn the Tit-for-Tat decision rule being applied by their competitor.

Price-Matching Guarantees How should Carnival respond to a Tit-for-Tat decision rule by Royal Caribbean? Let’s look at the analogies between this limited duration punishment scheme and a match- ing price guarantee. In Table 13.8, a matching price guarantee by Royal Caribbean substantially reduces Carnival’s incentive to discount down to $300 when RC has an- nounced $450 prices. Under the heading “$300” in the second column, one sees that Carnival’s $300 discounted price can no longer generate the $375,000 payoff of the first row but instead simply realizes the $185,000 payoff from a matching price policy by RC. This $185,000 payoff that arises is the same when both firms discount to $300. Because RC’s customers will monitor and enforce RC’s matching price guarantee by requesting rebates of ($450 – $300 =) $150 from RC whenever Carnival discounts to $300, Carnival cannot hope to gain a significant share of RC’s customers by discounting.

To place Royal Caribbean in the same position, Carnival will likely announce a matching price guarantee as protection in those times when RC might try a sneak at- tack on Carnival’s market share by discounting unexpectedly. Assuming, as we did ear- lier, that Royal Caribbean initiates play with a $450 price announcement, both cruise companies will maintain $450 prices, effectively playing “Match, Match” and escaping the Prisoner’s Dilemma by realizing the {$350,000, $275,000} payoff in the far north- west and far southeast cells. Like double-the-difference price guarantees, matching price guarantees increase the expected price level and hence the profitability in a tight oligopoly market.

Now, how does this outcome compare to Tit for Tat? Assume that the “Match” alter- native is not available in the game. Nevertheless, Carnival should see Royal Caribbean’s TFT decision rule as a delayed matching price guarantee. That is, with a one-period lag, Royal Caribbean is going to match any discount that Carnival tries and subsequently match (again with a one-period lag) any return to high prices as soon as Carnival re- turns. These payoff paths are certain; no amount of apologizing by Carnival about mis- takes and miscommunications can prevent RC’s one-period punishment. Therefore, Carnival simply compares the profits from discounting unilaterally this period ($375,000 – $275,000) to a discounted opportunity loss from punishment next period ($275,000 – $160,000):

$100,000 < $115,000=ð1 + rÞ if r < 0:15 [13.7] As long as the discount rate is less than 15 percent and the continuation of this particu- lar cruise route is certain for both firms, Carnival should not discount and thereby defect on the industry leader’s pricing policy of $450.

Of course, if the probability of continuance (p) falls below 1.0, a limited duration pun- ishment scheme such as Tit for Tat becomes much less effective immediately. For

conspicuous focal point An outcome that attracts mutual cooperation.

Chapter 13: Best-Practice Tactics: Game Theory 461

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

example, multiplying the future opportunity loss from punishment next period by just 10 percent less than certainty of continuance,

$100,000 < $115,000ð1 − 0:1Þ=ð1 + rÞ = $103,500=ð1 + rÞ $100,000 < $103,500=ð1 + rÞ if r > 0:035 [13.8]

implies that Carnival will defect and discount to attempt to gain market share any time the interest rate is greater than 3.5 percent.12 Tit for Tat therefore is a more effective coordination device for oligopolists that expect to encounter one another again and again, such as PepsiCo and Coca-Cola, United and American Airlines, Anheuser-Busch and Miller, and Carnival and Royal Caribbean.

Because the {$450, $450} actions yield $90,000 more for Royal Caribbean than the iterated dominant strategy equilibrium {$300, $300}, Royal Caribbean may well initiate cooperation and thereafter play Tit for Tat. With rational, unconfused, and well- informed competitors, communication of conspicuous focal points and multiperiod pun- ishment schemes can induce conditional cooperation in repeated Prisoner’s Dilemma. Perhaps for this reason, U.S. courts have prohibited airlines from signaling such coordi- nation information to one another through their centralized reservation systems.

Example Signaling a Punishment Scheme: Northwest13

America West once announced a $50 fare reduction for 21-day advance purchase tickets on the busy Minneapolis–Los Angeles route dominated by Northwest Air- lines. Rather than cutting its own $308 fare from its Minneapolis hub to match the America West $258 fare, Northwest responded by signaling a multiperiod punish- ment scheme. In particular, Northwest announced a $40 reduction (from $208 to $168) for 21-day advance purchase tickets on the busy route from New York to America West’s hub in Phoenix. This retaliatory fare was labeled on the Airline Tariff Publishing computer system as available for only the next two days, with possible renewal thereafter. Five days later, America West canceled its $50 discount promotion on Minneapolis to Los Angeles travel.

Antitrust law makes it illegal for companies to conspire to fix prices. Price sig- naling the particulars of a multiperiod punishment scheme in order to elicit coop- eration in maintaining high prices is seen as a violation of this provision of the Robinson-Patman Act. Northwest defended its actions as “competitive initiatives and responses consistent with independent self-interest.” However, the Third Cir- cuit Appeals Court was quite clear; signaling a limited-duration punishment scheme involving prices is not legal. U.S. v. Airline Tariff Publishing Co. et al., 92- 52854 (1992) expressly prohibited such preannouncements of price changes that might facilitate price coordination.

13Based on “Fare Game,” Wall Street Journal (June 28, 1990), p. A1; “Fare Warning,” Wall Street Journal (October 9, 1990), p. B1; and “Why Northwest Gives Competition a Bad Name,” BusinessWeek (March 16, 1998), p. 34.

12Equation 13.8 can also be written to highlight the interplay between p and r as

$100,000 < $115,000ð0:9Þ=ð1 + rÞ = $115,000=ð1 + RÞ

where R is the effective rate of interest 1/(1 + R) = p/(1 + r) = 0.9/(1 + r). In the Tit for Tat example, because the probability of continuance is near 1.0, the effective rate of interest and the actual rate of interest are quite similar. For r = 10 percent and p = 0.9, p/(1 + r) = 0.9/1.1 = 0.82, and therefore the effective rate of interest is 22 percent: 1/(1 + 0.22) = 0.82. As p gets smaller, the effective and actual rates of interest diverge exponen- tially. For example, for an actual interest rate of 10 percent and p = 0.55, p/(1 + r) = 0.55/1.1 = 0.5, and there- fore the effective rate of interest is 100 percent: 1/(1 + 1.0) = 0.5.

price signaling A communication of price change plans, prohibited by antitrust law.

462 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Industry Standards as Coordination Devices Mandatory industry standards or regulatory constraints are often a way of changing the structure of a simultaneous-play Prisoner’s Dilemma into a sequential-play game. Java programming language for the Internet, the digital signal specifications CDMA for cell phones, the Blu-ray specifications for high-definition television, or for wireless recharg- ing of cell phones, blenders, and power tools are examples of industry standards used in this way.14 By restricting the flexibility of one another’s responses, rivals can often secure an escape from the {Defect, Defect} dominant strategy payoffs of a simultaneous-play Prisoner’s Dilemma and achieve more profitable outcomes in a sequential game.

Consider the business-to-business sale of electrical equipment illustrated in Figure 13.3. General Electric would like to manufacture and distribute a high specifications (“Gold-plated”) halogen recessed lighting fixture supported with full installation and after-sale service. Unfortunately, however, the GE distributor has higher payoffs from not providing full installation. Under those circumstances, GE is better off manufactur- ing a fixture that meets only minimal specifications. Because of the distributor’s domi- nant strategy, the two companies earn payoffs {Worse, Better} and find themselves in a Prisoner’s Dilemma. They would prefer the northwest cell {Better, Best}, but each would then be vulnerable to a defection by the other company, resulting in their Worst outcome.

FIGURE 13.3 Electrical Industry Standard Allows GE Distributor to Escape Prisoner’s Dilemma

Note: Payoffs are listed (GE, GE Distributor).

GED

GED

GETP Minimal

Gold-plated

Full installation

No installation

Full installation

No installation

Best, Better*

Below code (illegal)

Worse, Worst

Below code (illegal)

Installation standard

Better

Best

Worst

Worse

Best

Worst

Worse

Better

High resale price (full installation)

Gold-plated product

Minimal specifications

product

Low resale price (no installation)G E

D is

tr ib

ut or

General Electric

14See “Adaptor Die,” The Economist (March 7, 2009), pp. 20–21.

Chapter 13: Best-Practice Tactics: Game Theory 463

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

By enlisting third parties (TP) such as Underwriters Laboratory in specifying an installation standard or encouraging the adoption of local building codes that require full installation, General Electric and its distributors can escape the Prisoner’s Dilemma. A General Electric distributor would then be engaged in an illegal (“below code”) sale if it provided anything less than full installation, so General Electric can anticipate full installation and will therefore proceed to manufacture the high specifications product. The payoffs will then improve to {Better, Best}.

ANALYZING SEQUENTIAL GAMES To illustrate the importance of the sequential order of play in many tactical situations, consider another manufacturer-distributor coordination game that arises between heavy truck manufacturers and independent retail distributors. The payoffs for the promotion and sale of a heavy truck like those sold by Volvo-GM Truck are displayed in normal form in Table 13.9. Let’s first examine the actions and payoffs in the left-hand column. The manufacturer wants the retail distributors to continue providing personal selling ef- forts and all after-sales service, rather than discontinue these activities, and thereby in- crease their retail margins. In return, the manufacturer agrees to advertise the product. If full services continue and advertising occurs, the customers will tolerate higher manu- facturer’s suggested retail prices (MSRP). In that case, the retail distributor and the man- ufacturer can earn additional profits of $180,000 and $300,000, respectively. However, if retail selling efforts and some after-sales services are discontinued and if MSRP increases (as in the northwest cell of Table 13.9), unit sales volume declines so sharply that the retail distributor receives only $120,000 profit while the manufacturer makes only $280,000 profit per day.

Independent retail distributors may feel tempted to deliver less service in many small, inconspicuous ways, especially if they suspect a lack of manufacturer-based ad- vertising of this product will make their time and effort spent on other products more valuable. Sales volume will eventually decline, but at a substantially higher margin that may well be in the retail distributor’s best interests. This outcome is represented in the northeast cell of Table 13.9; with retail selling effort discontinued and no manufacturer- based advertising, both parties incur fewer expenses and realize $130,000 profit for the retailer but only $150,000 profit for the manufacturer). On the other hand, if the man- ufacturer does not advertise, but retailer services continue (i.e., the southeast cell), man- ufacturer profit skyrockets to $380,000 but the retailer only clears $60,000 because of much higher expenses.

TABLE 13.9 SIMULTANEOUS MANUFACTURER/DISTRIBUTOR 1

Truck Manufacturer

Price increase No price increase Advertise Do not advertise

Increase margins (Discontinue services)

$280,000 $150,000

$120,000 $130,000 Retail Distributor $300,000 $380,000

Continue services $180,000 $60,000

464 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

What would you do as the retail dealer/distributor in this situation? Would you try for the increased margin by economizing on selling expenses and after-sale services? Re- member that your best payoff occurs when the manufacturer anticipates your continua- tion of full selling effort and after-sales services, and chooses therefore to advertise and raise the customers’ expectations by announcing a higher price point. And the manufac- turer’s best payoff occurs when you, the retailer, provide extensive dealer services, and the manufacturer economizes on advertising expenses. Note that Table 13.9 contains no pure Nash equilibrium strategies! So, how would you coordinate this on again–off again relationship?15 Note how much additional predictability of rival behavior emerges in this coordination game if we introduce a small but pivotal change in the structure of the game: a sequential order of play.

A Sequential Coordination Game Suppose as in Figure 13.4 that the manufacturer (M) must commit first to the release of a product update that warrants higher pricing, and that this decision is easily observable and irreversible. Then the retail distributor (R) must decide whether to continue personal selling effort and after-sale services or discontinue, and finally the manufacturer will thereafter decide on whether to contribute to cooperative advertising with the retailer. It turns out that introducing this sequential order to the decision making will make it possible to predict unambiguously the optimal strategic behavior by both parties and resolve the coordination problem of the simultaneous game.

FIGURE 13.4 A Sequential Coordination Game: Manufacturer/Distributor II

M

$280K$120K

$120K$40K

$100K $350K

Product update

No product update

M Manufacturer

R Retail distributor

M1

R2

Continue

Advertise

Not advertise

Discontinue R1

M2

Not advertise

Not advertise

Equilibrium prediction of rival response: {Update, Discontinue, Advertise}

Advertise

Not advertise

M3

M4

Continue

Discontinue

$180K $300K

Retailer Manufacturer

$150K$130K

$380K$60K

15We are assuming that merging the two entities into one vertically integrated firm is infeasible. In Chapter 15, we will see how these coordination problems can be resolved by, and indeed motivate, private voluntary contracting over vertical requirements between manufacturers and distributors.

Chapter 13: Best-Practice Tactics: Game Theory 465

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The structure of the sequential game can be represented as a game tree or decision tree as in Figure 13.4. The order of the decisions is read from left to right, and each circle represents a decision node. Update or Not update identifies possible actions that Player M can take at the first decision node M. Continue or Discontinue identifies possible ac- tions of Player R at the second decision nodes R1 to R2, and Advertise or Not advertise identifies possible actions of the manufacturer at nodes M1 to M4. The payoffs for the retailer and then for the manufacturer, associated with each sequence of possible actions, are listed in the last two columns. Note that some of these payoffs mirror those in Table 13.9 while others are totally new.

The manufacturer can look ahead and foresee that an Update of the product will make it advantageous for the retail distributor to Discontinue full retail selling effort. Out of self-interest, the manufacturer commits to an update, increases MSRP prices, and follows through with advertising. That is, the manufacturer can look ahead and ana- lyze what subsequent choices are in the retail distributor’s best interest (i.e., best-reply responses) and then reason back to detect what actions are in its own (the manufac- turer’s) self-interest. Each party in Figure 13.4 is able to look ahead and reason back using the concept of best-reply response to predict the rival’s behavior. None of this se- quential reasoning was available in the simultaneous-play version of the game.

Endgame reasoning always entails looking ahead to the last play in an ordered se- quence of plays, identifying the player whose decisions will control the outcome of the endgame, and then predicting that player’s best-reply response. In this instance, knowing that the manufacturer controls the outcome from node M1 and that the manufacturer is better off with $350,000 from Not advertise forces the retail distributor to dismiss the prospect of the $180,000 outcome in the first row. That possible outcome is not consis- tent with best-reply response by the manufacturer who does control the endgame. There- fore, that branch should be removed (“pruned”) from the game tree; the retail distributor should assume that if the product is updated and the retailer continues extensive selling effort, the manufacturer will not engage in cooperative advertising. Therefore {$100,000, $350,000} is the predictable outcome of deciding to Continue at node M1. However, your analysis is far from finished.

The conclusion about the endgame reasoning allows you, the retailer, to employ back- wards induction and rethink whether you would prefer to Continue or Discontinue at node R1. If the manufacturer’s best-reply response from M2 onwards is to Advertise (which yields $280,000 rather than $120,000 for M and $120,000 for you), it appears that your self-interest at the prior node R1 is to Discontinue. The strategy pair that pro- vides a best-reply equilibrium for the subgame beyond R1 is then {Discontinue, Adver- tise}, implying distributor and manufacturer payoffs of {$120,000, $280,000}, respectively. In sum, the strategy triplet that provides a Nash equilibrium for the sequen- tial coordination game Manufacturer-Distributor II is then {Update, Discontinue, Advertise}.

Subgame Perfect Equilibrium in Sequential Games Looking ahead to the rival’s best-reply responses in the endgame and then reasoning back to preferred strategy at earlier decision points is Reinhard Selten’s concept of a sub- game perfect equilibrium strategy for sequential games, a concept for which he and John Nash won the 1994 Nobel Prize in economics. Like many other path-breaking ideas, this intuitive strategic equilibrium concept is quite deceptive in its simplicity. Re- call that a Nash equilibrium strategy is a decision maker’s optimal action such that the payoff, when all other players make best-reply responses, exceeds that decision maker’s payoff from any other action, again assuming best-reply responses. Selten applied this

game tree A schematic diagram of a sequential game.

backwards induction Reasoning in reverse time sequence from later consequences back to earlier decisions.

subgame perfect equilibrium strategy An equilibrium concept for noncooperative sequential games.

466 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Nash equilibrium concept to sequential play and invented the concept of Nash equilib- rium in a proper subgame of sequential play.

Some nodes of the game tree, such as R2 in the bottom half of Figure 13.4 and the subgames thereafter, can be eliminated from consideration because they cannot be reached by best-reply responses. Such decision points are “off the equilibrium path.” Sel- ten’s idea was that only in the proper subgame nodes would the Nash equilibrium con- cept hold. Specifically, the $380,000 outcome in the next-to-bottom row of Figure 13.4 is the highest payoff in the entire exercise. Yet, the manufacturer should not consider this logical possibility precisely because M3 and beyond is not a proper subgame; the payoffs {$60,000 and $380,000} cannot be reached by best-reply responses. Knowing that the manufacturer never cooperatively advertises a product that has not been updated, the re- tailer at R2 rejects Continue in favor of a best-reply response Discontinue, to capture $130,000 rather than the alternative $60,000. Subgame perfect equilibrium strategy re- quires analyzing the outcomes associated with actions and best-reply responses at R1 and M2, the only proper subgame nodes of Figure 13.4.16 Again, {Update, Discontinue, Advertise} proves then to be the subgame perfect equilibrium strategy for Manufacturer- Distributor II.

Sometimes this identification of proper and improper subgames can get quite compli- cated when many endgames are possible. To illustrate, consider the three-way compara- tive advertising duel in Exercise 6 at the end of this chapter. With varying degrees of success, three firms attack one another with combative advertising in pairwise, sequential competitions until just one firm remains. It can take two complete rounds of advertising attacks and almost 20 endgames to analyze the subgame perfect equilibrium strategy for that problem.

BUSINESS RIVALRY AS A SELF-ENFORCING SEQUENTIAL GAME It is important to emphasize that the subgame perfect equilibrium concept is self- enforcing. It predicts stable rival response, not because of effective monitoring and third-party enforcement, but because each party would be worse off departing from the equilibrium strategy pair than it would be implementing it. Ultimately, it is this best- reply response idea that identifies whether a commitment is credible. And credibility can work both ways; credible commitments can also become credible threats. Let’s see how.

Consider a well-established pharmaceutical manufacturer of ulcer relief medicine, who presently markets the only effective ulcer therapy, possessing no known side effects, and earns $100,000. This incumbent (let’s call the firm “Pastense”) faces a small potential entrant (“Potent” for short). Potent has discovered a new therapeutic process that also has the potential to cure stomach ulcers. Potent must decide whether to enter the mo- nopoly market or stay out and license its trade secrets to any one of several interested buyers. Pastense must decide whether to maintain its present high prices, moderate its prices, or radically discount its prices. The payoffs are displayed in Figure 13.5. If the potential entrant Potent enters, and the incumbent Pastense does not moderate or

16The reader may wonder about the relevance of M1 if a miscommunication or strategic mistake is made by the retail distributor at node R1. This concern is valid because mistakes and miscommunication do happen in the reality of business rivalry. Indeed, a refinement of subgame perfect equilibrium strategy allows for just such mistakes and describes equilibrium strategy for the manufacturer in this game less uniquely as {Update and Advertise if Retailer Discontinues} but {Update and Do Not Advertise if Retailer Makes a Strategic Mis- take and Continues}.

Chapter 13: Best-Practice Tactics: Game Theory 467

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

discount to lower price points, suppose all the ulcer relief business goes to the new en- trant and the incumbent realizes nothing. In contrast, with entry and discount prices, suppose the incumbent’s product enjoys a slight cost advantage and earns a $10,000

Example Business Gaming at Verizon17

Former Verizon chair Ray Smith employed the techniques, exercises, and lessons of game theory throughout his organization. In “war games,” teams of Verizon man- agers assumed the role of major competitors and explored tactics that could defeat Verizon’s business plans. Other teams detailed future contingencies in a large game tree that allowed Verizon to map its future moves and countermoves as well as uncover the competitive effects of new technological developments (e.g., digital voice and video transmission) before they happened. Traditional planning models lock managers into assumptions the importance of which they can only gauge through sensitivity analysis. But sequential game analysis constantly reminds man- agers to shape the game, not just play it. It can mean reversing the order of play by recommending preemptive strikes in some circumstances (e.g., merging with Ny- nex) but highlighting the value of “fast-second” best-reply responses in other cir- cumstances (e.g., in following rather than leading basic research and product development at Lucent Technologies).

In addition, Verizon has learned to recognize endgames that are unfavorable to the company and reshape the structure of the competitive rivalry in those busi- nesses. Verizon recently redefined the scope of the telephone industry’s local net- work strategy game by winning approval in the courts for telephone companies to own the content transmitted over their phone lines. Verizon managers are now hard at work analyzing the new larger game that includes business directories, dig- itized movies, and video production.

17Based on “Business as a War Game: Report from the Battlefront,” Fortune (September 30, 1996), pp. 190–193.

FIGURE 13.5 Entry Deterrence I: Incumbent Pricing in Response to Entry Threat

PE

Moderate price

High price

Low price

High price

Moderate price

Low price

Enter

Stay out/license

Note: Outcomes are listed (Incumbent, Potential Entrant) in $000s.

($0, $80)

($35, $50)

($50, $40)*

($80, $30)

($70, $60)

($40, $20)

I2

I1

468 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

greater payoff (i.e., $50,000 and $40,000 in the third row of Figure 13.5). Moderate in- cumbent prices result in a $35,000 payoff for Pastense and a $50,000 payoff for Potent.

To prevent the reduction of its profit from $100,000 as a monopolist to $50,000 post- entry, Pastense itself might be a prime candidate for the purchase of Potent’s trade se- cret. Realistically, however, it is liable to run up against antitrust constraints that restrict mergers between dominant incumbents and new entrants. Note also from node I2 in Figure 13.5 that what another established pharmaceutical manufacturer will pay to li- cense the trade secret, with all the attendant technology transfer and marketing chal- lenges, bears little correlation to what Potent itself could hope to earn upon entry. Potent receives its second highest payoff ($60,000) when it licenses its trade secret in a moderate price environment. Potent earns the least (namely, $20,000) when it stays out and licenses, and Pastense discounts anyway.

First-Mover and Fast-Second Advantages As is now obvious, “Who can do what, when?” is the essence of any sequential strategy game. The order of play determines who initiates and who replies, which determines the best-reply response in the endgame, and thus the strategic equilibrium. If Potent enters, Pastense strongly prefers a Low pricing response, because $50,000 far exceeds the zero or $35,000 outcomes from either the High or Moderate alternatives. This anal- ysis of the incumbent’s best-reply response allows Potent to predict that its own $80,000 and $50,000 outcomes are infeasible. Even though each is theoretically associ- ated with its entry, neither can be obtained if Pastense makes a best-reply response in this proper subgame.

Similarly, in the bottom endgame node I2, if Potent stays out, its royalty payoffs of $60,000 cannot be obtained, because Pastense will price High to secure $80,000 for it- self rather than accept its lower $70,000 and $40,000 alternatives. Only two focal out- comes of interest remain for Potent in making its entry decision: the shaded payoffs of $40,000 from entering and $30,000 from staying out. Being a value-maximizing firm, Potent decides to enter, predictably, and the events of the subgame perfect strategic equilibrium {Enter, Discount} then unfold. Notice that both players could be better off with the {$70,000, $60,000} outcome in the lower node, but Potent cannot expect Pastense to respond with Moderate rather than High prices should Potent stay out and license.

However, to illustrate the pivotal importance of the order of play, let’s mix things up a bit. From the incumbent’s point of view, too, the outcomes {$50,000, $40,000} are not entirely satisfactory. Given its second-mover timing, Pastense did as well as could be ex- pected. But the incumbent may wonder whether seizing the first-mover initiative would have worked to its advantage. However, no general rule on this point exists—sometimes it will, and sometimes it will not. Each sequential game situation is in this way unique.

To analyze the question, in Figure 13.6 we reverse the order of play in Entry Deter- rence II. Now, the potential entrant controls the endgame, and the incumbent must an- nounce irreversible pricing policies in advance. Saying they are irreversible does not make it so, but more on that in the next section. Analyzing the three endgame nodes, Pastense realizes that Potent will choose to enter when high prices are precommitted, stay out when moderate prices are precommitted, or enter when discount prices are pre- committed. Knowing these outcomes, Pastense announces a moderate pricing policy, and the starred {Moderate, Stay Out} strategic equilibrium is the result. Not only has the po- tential entrant’s behavior changed, but in addition, the payoff to Pastense has risen from $50,000 to $70,000. In this instance, a first-mover advantage proved to be just what the name implies.

focal outcomes of interest Payoffs involved in an analysis of equilibrium strategy.

Chapter 13: Best-Practice Tactics: Game Theory 469

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIGURE 13.6 Entry Deterrence II: Response to Incumbent Price Commitment

I I

Enter

Stay out

Enter

Stay out

Enter

Stay out

High price

Moderate price

Low price

Note: Outcomes are listed (Incumbent, Potential Entrant) in $000s.

($0, $80)

($80, $30)

($35, $50)

($70, $60)*

($50, $40)

($40, $20)

Price commitment signal PE2

PE3

PE1

Example Technology Leader or Fast-Second: IBM18

Whether to secure first-mover advantages in the development of new computing technologies or instead engage in a pattern of quick imitation (i.e., a fast-second strategy) poses a more difficult choice than one might think. In the absence of sunk-cost investments as a barrier to entry, hit-and-run entry often proves effec- tive. Apple Computer commercialized the graphical user interface (GUI) that Xerox invented, and Apple’s breathtaking but unsuccessful Newton led the way for Palm. Microsoft made a fast-second and very successful challenge to Netscape’s early dominance of Internet browsers. And Sun Microsystems developed the re- duced instruction-set computing (RISC) that IBM pioneered. Even in consumer perishables, Dunkin’ Donuts’ Coolatta is fast imitating one of Starbucks’s most profitable offerings, Frappuccino.

By restraining up-front investments in basic research and focusing instead on the development of products, IBM has recently switched from a technology leader to a “first of a kind” systems problem solver for high-margin hospital customers. One example has been a blending of computer imaging and voice-recognition de- vices that allows hospital radiologists and surgeons to superimpose X-ray images and text on any PC throughout the local area network in a medical center. Doctors speak to one another while viewing PC-based images, and the IBM hardware and software creates a digital transcript of their diagnostic findings and expert opinions.

On the other hand, IBM Microelectronics division leveraged the company’s long-standing basic R&D effort in materials science into a breakthrough in silicon chips. IBM’s engineers discovered how to form copper rather than aluminum cir- cuits and yet prevent the copper atoms from bleeding into the surface of the sili- con. Copper is a more conductive material and therefore can be laid down in narrower circuits than aluminum. The more circuits etched on a square centimeter of silicon, the more powerful and cost effective the computer chip. IBM’s copper- on-silicon circuitry patent promises to increase computing power 40 percent for any given size chip.

18Based on “Einstein and Eraser-Heads,” Wall Street Journal (October 6, 1997), p. 1.

470 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

CREDIBLE THREATS AND COMMITMENTS In multiperiod games, the credibility of all threats and commitments ultimately de- rives from whether the threat-maker or commitment maker successfully identifies and adopts subgame perfect strategies. In Entry Deterrence I (see Figure 13.5), Pas- tense’s threat to discount the ulcer relief medicine if Potent entered was credible pre- cisely because discounting was, in fact, a best-reply response. Any other response would have made Pastense worse off (i.e., lowered its payoff). A credible threat is therefore defined as a conditional strategy that the threat-maker is worse off ignoring than implementing. By the same token, a commitment by Pastense to maintain high prices (i.e., not to discount and thereby spoil the royalty value of Potent’s trade secret) if Potent would stay out of the market is a credible commitment. Again, the reason is that this action is the incumbent’s best-reply response to Potent’s staying out and just earning royalties from the ongoing value of its trade secret. Therefore, without any monitoring or third-party enforcement whatsoever, one can fully rely upon Pastense to honor its commitment, because it would not be in its own best interest to do otherwise.

In Figure 13.5, if Potent wanted to secure a commitment from Pastense to price at the moderate level in exchange for some portion of the much larger $60,000 royalties, Potent would need to employ a binding, third-party-enforceable contractual agreement. It is simply not in Pastense’s best-reply-response interest to fulfill such a commitment otherwise.

You can now begin to see why purposeful individual behavior and a shared objective in groups is so critical to game theory reasoning. To predict choices of highly interde- pendent players, one must know what makes them tick, what true goals they seek, and what the consequence of various actions is on those goals, which is sometimes harder than it sounds. For example, performance-based incentives and takeover threats often align management objectives quite closely with stockholder value, but what motivates a closely held family-run business is sometimes difficult to fathom. Moreover, consistently transmitted signals of business strategy are often jammed or misinterpreted by the re- ceiver. Therefore, to ensure the effective communication of credible threats and credible commitments requires some guidelines. This situation can be illustrated by returning to the Entry Deterrence game.

As we have seen, Pastense found the switch to first-mover status highly advantageous. By committing to maintain moderate prices rather than discount, its profits increased from $50,000 to $70,000 when Potent sold out rather than entered. The question we must now reexamine, however, is “Why did Potent believe Pastense would maintain Moderate prices?” After all, it is clear from the original game tree in Figure 13.5 that once Potent licensed its trade secret to another less capable potential entrant (let’s call the new firm “Impotent”), Pastense was really better off raising its price back to the high level it had once enjoyed. Note that thereby Pastense would then receive the $80,000 payoff from high prices rather than a $70,000 payoff from moderate prices. Thus, Pastense’s commitment to maintain a moderate price was not a credible commit- ment because Pastense was worse off making good on the commitment than ignoring it.

One might be inclined to respond that likewise Potent can renege on its commitment to stay out of the ulcer relief business. Licensing a trade secret for royalty revenue today need not preclude Potent’s potential entry tomorrow. Indeed, such royalty agreements seldom include a no-competition clause. However, the difference here is that Potent’s payoff is maximized by staying out! Its commitment to staying out if the incumbent maintains moderate prices is in Potent’s own best interest. Staying out is a best-reply re- sponse; therefore it is a credible commitment.

credible threat A conditional strategy the threat-maker is worse off ignoring than implementing.

credible commitment A promise that the promise-giver is worse off violating than fulfilling.

Chapter 13: Best-Practice Tactics: Game Theory 471

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

MECHANISMS FOR ESTABLISHING CREDIBILITY19 As second mover, Potent controls the endgame and therefore finds itself in a position to insist on the necessary assurances from Pastense. Among the alternative mechanisms for establishing credibility, Pastense might establish a bond or contractual side payment, which would be forfeited if Pastense raised prices. Some such contracts, referred to as maximum resale price maintenance agreements, do exist between retailers and their sup- pliers. Another possible credibility mechanism would be for Pastense to invest heavily in its moderate price strategy to establish a reputation for moderate prices. Loss of this non-redeployable reputational asset would discourage reneging on its commitment to maintain moderate prices. Third, Pastense could short-circuit or interrupt the repricing process by preselling its ulcer relief medicine with forward contracts. Forward sale con- tracts establish credible commitments because the courts generally refuse to excuse for- wards or futures contract breaches for any reason. Fourth, Pastense could enter into teamwork or an alliance relationship with Potent that would sufficiently dilute the re- wards from reneging on its commitment, perhaps by taking an equity stake in Potent. Fifth, Pastense could change the structure of the game to require that both it and Potent only “take small steps.” In the next section, we analyze leasing as a way to pursue this small-steps credibility mechanism.

And finally, the most practical response to this situation would be for Pastense to ar- range an irreversible and irrevocable hostage mechanism, whereby likely future custo- mers were granted a moderate price guarantee. Sometimes referred to as “most favored nation” clauses, these price guarantees promise double refunds if the customer discovers any lower-price Pastense transaction during the next or the previous year. As long as Potent observed at least one moderate price transaction before licensing its trade secret, it could rest assured that Pastense had now offered a credible commitment not to raise prices. The resulting double refunds, should Pastense raise prices, and the sacrifice of fu- ture transactions with its own repeat-purchase customers, should it renege on the re- funds, ensure that Pastense will finally be better off honoring its commitment to moderate prices than ignoring it. And again, notice that these agreements are entirely self-enforcing; no third party needs to be relied upon to make the rival behaviors predictable.

Example Double-the-Difference Price Guarantees: Best Buy At times, Best Buy offers to rebate twice the differential purchase price of a DVD player to preferred customers should those customers find the same DVD player selling for less anywhere in the local area over the next three months. This rebate guarantee will be enforced by the courts. As in the simultaneous-play pricing game between PepsiCo and Coca-Cola, Best Buy normally would be better off discount- ing (maybe even steeply discounting) when competitors such as HHgregg and Sound Warehouse maintain high prices. But in the face of this double- the-difference low-price guarantee, Best Buy would lose more money on rebates than it could possibly gain from any amount of incremental business it could

(Continued)

19This section relies heavily on A. Dixit and B. Nalebuff, Thinking Strategically: The Competitive Edge in Business, Politics, and Everyday Life (New York: Norton, 1993), especially Chapters 5 and 6.

non-redeployable reputational asset A reputation whose value is lost if sold or licensed.

472 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

REPLACEMENT GUARANTEES As we discussed in Chapter 10, all buyers rationally discount experience goods such as used cars and computer components if they cannot verify independently at the point of purchase the seller’s quality claims. A replacement guarantee or a product performance repair warranty are other good examples of hostage mechanisms—in this case, hostage mechanisms that establish the credibility of a seller’s commitment to deliver high- quality components in the goods it offers for sale. Should the seller violate his or her commitment, a third party (usually the courts) will impose on the seller monetary judg- ments that are larger than the incremental cost of upgrading from lower to higher qual- ity inputs in the first place. Therefore, the buyer is assured of a higher quality machine when the seller offers to include a replacement guarantee or repair warranty for the same (or a slightly higher) price. These guarantees and warranties illustrate a credible commit- ment mechanism (i.e., third-party-enforceable promises that the promise-giver would be worse off violating than keeping).

What exactly constitutes a credible replacement guarantee? Claims by Dooney & Bourke handbags, Revo sunglasses, and Sewell Cadillac for lifetime repair or replacement provide credible commitments. Why? The key is repeat customer business. Because in- cremental sales to established or referral customers are much less expensive than attract- ing new customers, the customer-for-life relationships at these companies provide a

reasonably expect to take away from the competitors. In effect, Best Buy has given its competitors a hostage that supports a commitment to maintain high prices.

In Figure 13.7, Best Buy provides a bond of its intentions to maintain high prices by preannouncing the double-the-difference price guarantee. Sound Ware- house must then decide whether to discount or maintain high prices in light of the Best Buy rebate program. Like all good hostage mechanisms, the hostage is worth more to the hostage giver than its value in use to the recipient. That is, Sound Warehouse could trigger double-rebate payments at Best Buy by discounting its own price. And harming a competitor is a reasonable secondary goal, but it’s only secondary. Securing your own highest payoff perhaps through legal coopera- tion with a competitor is the primary goal. Because Best Buy would respond to a Sound Warehouse discount by matching the lower price point, Sound Warehouse would gain no additional market share by using the hostage in this way. Indeed, for the hostage recipient such a decision would lead to the payoff labeled “Worse” in the top right-hand corner of Figure 13.7.

Knowing that Best Buy controls the endgame and that it would be in Best Buy’s best interest after announcing the rebate program to match a discount price, Sound Warehouse finds itself preferring to maintain high prices. Because Best Buy is also best off by maintaining high prices, the payoff {Best*, Better} results. Thus, by in- troducing a price guarantee that limited its own ability to take advantage of its op- ponent’s vulnerability at high prices, Best Buy secured first-best outcomes when the alternative was Worse (i.e., compare the shaded and unshaded boxed payoffs in Figure 13.7). A hostage mechanism that establishes one’s credible commitment to maintain high prices if the rival maintains high prices will often elicit high prices from that rival. Thus, from the point of view of both companies, double- the-difference low-price guarantees are unambiguously preferred. Of course, con- sumer advocates will not rejoice, but complaining about double-the-difference price guarantees attracts few sympathizers to the consumer cause.

Chapter 13: Best-Practice Tactics: Game Theory 473

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

hostage mechanism. Dooney & Bourke’s, Revo’s, or Sewell Cadillac’s normally preposter- ous replacement or service guarantees backed by a brand name, unique distribution channel, or other non-redeployable asset become credible because of the seller’s depen- dence on repeat or referral business. In effect, Sewell Cadillac says, “My sunk cost invest- ments cannot be recovered (and, by definition, cannot be liquidated at anything near their historical cost) unless I earn your repeat business.”

FIGURE 13.7 Double-the-Difference Price Guarantees

Note: Payoffs are listed (Best Buy , Sound Warehouse ).

BB

BB

BB SW

SW

Discount

Maintain high prices

Match

Maintain

Discount

Maintain

(Worse*, Worse)

(Worst, Best)

(Best, Worst)

(Better*, Better)

BB

BB

SW

BB

Discount

Maintain high prices

Match

Maintain

Discount

Maintain

(Worse*, Worse)

(Worst, Best)

(Better, Worst)

(Best*, Better) Price guarantee

No price guarantees

Example Noncredible Commitments: Burlington Industries Classic examples of business strategies that flounder because of the absence of credible commitments include the quota “commitments” in a cartel and the “com- mitment” not to compete after purchasing surplus equipment in a declining indus- try. Burlington Industries has experienced many problems with its overseas sales of old textile looms, often acquired in mergers and then liquidated at scrap value. The foreign buyers restore the old equipment and then back ship their production into the United States despite no-competition clauses in the equipment purchase con- tracts. Burlington has now begun to destroy old equipment, not just dismantle it, especially in declining product lines where it wishes to pursue a niche strategy as “the last iceman.” The idea is to preserve high margins by becoming the last com- pany in one’s locality to sell block ice to cruising boat owners or to wedding plan- ners for ice sculptures. To this purpose, IBM bought up Amdahl Millennium and Hitachi Skyline mainframe computers and physically crushed them in a scrap yard.

474 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Hostages Support the Credibility of Commitments In Chapter 10 we encountered a noncooperative sequential games mechanism for secur- ing cooperation in a repeated Prisoner’s Dilemma game through the use of credible com- mitments. Potentially notorious firms selling low-quality experience goods (e.g., PC components) for high prices were identifiable in Table 10.2 as firms with entirely rede- ployable assets. That is, firms selling out of temporary locations, with unbranded products and no company reputation, could be reasonably expected to follow the domi- nant strategy of producing low quality. Consider eNow Components illustrated in Figure 13.8. This company is not likely to plan on more than one transaction with any cus- tomer. It may not even plan on doing business through its present post office box or e-business site for long. Consequently, these are firms to whom no customer should offer a high price.

On the other hand, we argued that firms who asked high prices but also exhibited verifiable sunk cost investments that dissipate the rent from such prices were much bet- ter bets. Reputational advertising of nontransferable company logos (say, Apple Compu- ters or CarMax) or investment in non-redeployable assets, such as product-specific showrooms (Ethan Allen) and unique retail displays (L’eggs), present a hostage to buyers. Because sellers offering hostages are worse off if they fail to deliver on the prom- ise of high quality, a buyer can rely on these credible commitments even if unable to verify quality at the point of purchase. Although the credible commitments are noncon- tractual in nature, they establish reliance relationships that are as predictable as enforce- able contracts.

Finally, credibility arises from cooperative game mechanisms that involve binding (third-party-enforceable) contracts such as franchise agreements, escrow bonds, and re- fund guarantees. These contractual mechanisms also provide hostages that support win- win exchange despite a dominant strategy that would otherwise lead players to defect. The key to the credibility of such mechanisms is the same as in noncooperative games.

FIGURE 13.8 An Illustration of a Noncredible Lifetime Guarantee

eNow Components, plc.

Reconditioned replacement hard drives for any laptop

Check here High quality Durable* $159.00

or check here Standard 90-day warranty $89.00

Model laptop (name, no.)

Mail to: Box K The Docks Bayonne, N.J.

All orders by mail must be accompanied by cash or money order.

*A lifetime replacement guarantee accompanies this hard drive.

Chapter 13: Best-Practice Tactics: Game Theory 475

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

First, in the light of its warranty obligations, is the promise-giver better off fulfilling its promise than ignoring it? And second, can the warranty or bond be revoked for any reason other than just causes the promise-giver cannot control? If the answer to both questions is yes, then these contractual commitments are credible and consistent with best-reply response. If not, then they are noncredible and should not be relied upon.

Credible Commitments of Durable Goods Monopolists What buyers will pay for a capital equipment purchase, such as a corporate jet, a main- frame computer, or a business license depends in part on how well the seller resolves some credible commitment issues. If a piece of durable equipment has a working life that will extend over several market periods, an early adopter of a new model worries about (1) obsolescence risk, (2) uncertain reliability of the new technology, and (3) the risk of falling prices subsequent to his or her purchase. How well a manufacturer ad- dresses these three perceived risks of buying early will determine the rate of adoption and the prices paid to acquire new capital equipment.

Planned Obsolescence The competitive advantages Cisco’s newest data servers might offer an information tech- nology user such as a direct marketer are seriously compromised whenever Cisco intro- duces a still newer model and makes the direct marketer’s machine obsolete. In addition, other potential buyers who see somewhat less advantage in the newest server equipment will likely benefit from a Cisco price reduction at some later date. Knowing these likeli- hoods, the first buyers hesitate, adopt later, and offer to pay less than they otherwise would for the new technology. To overcome this persistent problem that comes with ev- ery new generation of equipment, Cisco must somehow credibly commit to maintaining high prices and to a controlled rate of planned obsolescence that allows early buyers time to recover their investment costs.

In an industry with slow-moving technology, a dominant firm could make contractual commitments to phase in new updated equipment. At times, tractor-trailer trucks have been sold this way, and to a certain extent, the limited body style change from year to year in some automobile models (Mini Coopers, Camrys, and Accords) reflects the same idea. However, in the data server industry, Cisco cannot afford such restrictions; technol- ogy simply moves too fast.

So what alternatives remain? Buyers of durable equipment can’t be expected to risk a lot of capital soon after the rollout of a new model; yet, companies such as Cisco cannot

Example Resale Value of a Mini Cooper20

BMW’s Mini Cooper has the highest resale value after five years as a percentage of the purchase price of all autos sold in America ($11,800 on a $20,000 car, or 59 percent). On average, American cars are worth only 35 percent of their pur- chase price after five years. The giant land yacht Ford Expedition is worth the least—only 19 percent. Toyota has the highest average resale value across manufac- turers after three years—52 percent—to General Motors’ 43 percent.

20Based on “Value-Packed Vehicles,” Forbes (November 2, 2006), p. 51–53; and “U.S. Auto-Makers Fail to Improve Resale Value,” Wall Street Journal (November 19, 2008), p. D3.

476 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

lock themselves in to delays of the future upgrades. One approach is to continuously up- grade the product at higher and higher prices, what Carl Shapiro and Hal Varian call versioning.21 Microsoft adopted this model upgrade strategy with their Windows operat- ing system. The buyer was told, in effect, don’t hesitate and wait for the price to fall; the next model will be even more expensive. But of course competitive pressure may keep that from happening as technologies morph from patent-holding monopolies to fast- second imitators.

Moreover, even if no competing product appears in the marketplace, Cisco’s operat- ing systems are not consumed on the spot; they don’t wear out. Like any durable goods monopolist, Cisco is perceived by the potential early adopters to be competing against itself. Similarly, Microsoft’s biggest competitor for Windows 7 is Windows XP, just as the best substitute for Windows XP was Windows 2000. Another approach altogether is needed. One is to ask buyers to take small steps by leasing the equipment one market period at a time. Recall that this was one of the mechanisms we identified earlier in the chapter for establishing the credibility of commitments. Although this approach fails to slow (and may actually quicken) the pace of new product introductions, buyers risk less up-front capital and therefore can be induced more easily to take on the new model and update their capital equipment more frequently at higher prices.

IBM employed exactly this approach for many years by offering only to lease their mainframe computers. Similarly, Dell Computer advertises, “How many companies will let you return your computer when it becomes obsolete?” and leases its PCs for $99 per month with the opportunity to renew for a new updated computer two years later. And BMW auto leases provide bumper-to-bumper scheduled maintenance and unexpected repairs for the lifetime of the lease. So leasing mitigates obsolescence and maintenance risk. But what about the early adopters’ risk of subsequent price reductions? How does leasing address that risk?

Post-Purchase Discounting Risk Understanding the tactical advantage of leasing requires careful analysis of the manufac- turer’s asymmetric information in making planned obsolescence and price discount deci- sions. Because the manufacturer knows the marketing plans and can estimate the pace of technology and the risk of obsolescence much better than the end-user customer, lease terms can be more favorable when the seller undertakes to absorb the risk of price pro- motions and planned obsolescence. That is, in a competitive marketplace for capital equipment leases (i.e., the corporate jet lease market), one would expect sellers to offer closed-end leases with residual values that reflect their accurate estimates of what a two-year-old corporate jet will be worth. This fixed residual value is what really estab- lishes the credibility of the manufacturer’s commitment over the lease period to refrain from discounting or introducing a new model that would render the current model ob- solete. If the lease writer (the lessor) violated this promise, the assets returned at the end of the lease would be worth less than the residual value at which the manufacturer-lessor has agreed to take them back. In effect, the manufacturer has given a hostage to the leaseholder (the lessee). By agreeing to take back the capital equipment for a preset amount and dispose of it in the resale market, the manufacturer-lessor credibly commits to a limited set of price promotions and to a limited rate of planned obsolescence.

versioning A new product rollout strategy to encourage early adoption at higher prices.

21See C. Shapiro and H. Varian, “Versioning: The Smart Way to Sell Information,” Harvard Business Review (November–December 1998), pp. 106–118.

Chapter 13: Best-Practice Tactics: Game Theory 477

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Leasing Digital Moviehouse Projectors: Hughes-JVC22

U.S. households spent more on movie theater admissions in 2009 ($9.87 billion) than on DVDs for the first time since 2002. Digital cameras and projectors are clearly involved. George Lucas’ Star Wars movies are now filmed entirely on digital movie cameras. Cinema companies such as General Cinema and Carmike much pre- fer digital film over the 60-pound celluloid film prints that often reach diameters of five feet. Downloading compressed-signal digital films using high-speed secure data networks will allow the movie houses much more flexibility in their scheduling. In addition, the sound and projection quality will no longer deteriorate after a few dozen showings. The movie production companies also like the new technology because a full national rollout of a celluloid movie costs approximately $10,000,000 to produce the 5,000 prints required, and it necessitates a large fleet of trucks to move the film canisters about the country.

The biggest hurdle to the fast adoption of this new technology is the $150,000 re- placement cost for each of the projectors in a small (five-screen) cinema. With the U.S. movie theatre industry badly overbuilt (some estimates suggest by as many as 10,000 more cinemas than needed), General Cinema, Carmike, and others are under- standably hesitant to commit three-quarters of a million dollars of additional capital per movie house. Each would like to wait for the discounted digital projector prices that they believe will come later. Hughes-JVC, who manufactures one of the projec- tors, plans to lease the projectors to cinemas. With digital rather than physical distri- bution, leasing fees for the projectors can be tied to actual showings. This will allow cinemas in overbuilt markets to participate in earlier adoption of the new technology.

22Based on “Curtains for Celluloid,” The Economist (March 27, 1999), p. 81; and “Moving Images into the Future,” The Economist (December 6, 2008), pp. 8–10; and “Cinema Surpasses DVD Sales,” Wall Street Journal (January 4, 2010), p. B10.

Example NetJets Fractional Ownership Plans for Learjet and Gulfstream Aircraft and Lexus23

FlexJets offers guaranteed access on four hours’ notice to a fleet of Learjet and Challenger business aircraft for as little as $175,000 per year cost. NetJets, a divi- sion of Warren Buffett’s Berkshire Hathaway Company, offers fractional shares in “the world’s largest and finest fleet of 450 Gulfstream aircraft with guaranteed availability, guaranteed costs, and guaranteed liquidity of your asset.” NetJets ex- pects to schedule more than 500,000 flights next year. A one-sixteenth share of a Citation seven-passenger jet leases for a $620,000 upfront commitment, a $7,909 monthly fee, plus $1,675 per flight hour to cover operating cost. These contractual arrangements are essentially operating leases.

One illustration of how crucial these fractional ownership-leasing arrangements are in protecting the early adopter against resale price instability is that a $44- million Gulfstream VI, the top-of-the-line business jet, sold as a two-year-old used aircraft in late 2002 for only $18 million. More normal resale prices would be $25–$28 million. Similarly, deal making on luxury new car sales in 2003 drove the resale value of two-year-old Lexus LS 430s and Saab 9-5s down by 23.4 percent from $53,500 to $41,000 in just one year, compared to a 14.7 percent reduction for two- year-old models one year earlier. Again, someone (either lessor or lessee) must bear

(Continued)

478 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

these repricing risks, but closed-end leasing with fixed residual values offers a cred- ible commitment from manufacturers to early adopters that the seller will not flood the market at discounted prices before the early adopter’s equipment needs to be replaced. In the luxury car segment, closed-end leases with a fixed residual value have therefore grown to 29 percent of all U.S. auto industry transactions.

23“Prices on Private Planes Dive,” Wall Street Journal (September 5, 2002); and “The Bargain Jaguar,” Wall Street Journal (March 20, 2003), p. D1.

Example The Licensing of Taxi Medallions and Cell Phones24

Similarly, by selling a taxi medallion or a cellular phone authorization as a renewable license, a municipality can credibly limit the supply of the city’s transportation and communication infrastructure. If the city were to insist on an outright purchase, taxi and cellular entrepreneurs would be concerned that soon thereafter the city would flood the market with additional taxis and cell phone companies. Consequently, the amounts bid for the right to do business would decline substantially. For exam- ple, in Washington, D.C., the city council authorized essentially open entry, attracted 12 cabs for every 1,000 residents, and found that the equilibrium license fee was $25. In contrast, New York City has restricted entry to 11,797 taxi medallions, two taxis per 1,000 residents, and as a result the New York taxi medallion asset transfers for $140,000.

The point is not that potential license holders wish to avoid being duped, al- though of course all of us are motivated to avoid embarrassments. Instead, it’s that licenses authorizing a business are a property right that the license holders may need to resell. Random events happen to every company, and license holders cannot assume that they will be able to operate forever. Licenses are durable capital assets, and their resale value is as much a concern as would be the value of a main- frame computer or corporate jet. Municipalities can raise more money, therefore, with renewable leases for all business licenses. What occurs in business licensing by municipal and state governments also occurs in the licensing of trade secrets and patents. Again, credible commitments by seller-lessors to actions that will maintain forward asset values are the key to eliciting higher buyer-lessee willingness to pay among early adopters.25

Even though renewable licensing and leasing offer tactical advantages in estab- lishing credible commitments not obtainable with outright sales, always remember that in competitive markets leasing will not be cheaper than buying. Any costs im- posed on the seller by the credibility mechanisms (e.g., a higher residual value) will be priced into the lease. The point is simply that some credible commitments im- pose additional costs on the asymmetrically informed manufacturer as a lessor that are lower than the reduction in price that would be required to accomplish a com- parable sale. Consequently, manufacturer profitability increases with renewable li- censing and leasing, relative to the alternative profitability available from the outright sale of durable equipment, business licenses, or patents.

24“New York Taxi Policy,” Wall Street Journal (March 17, 1992), p. A14; and “Put the Brakes on Taxicab Monopo- lies,” Wall Street Journal (November 6, 1984), p. A20. 25See M. Waldman, “Durable Goods Theory for Real-World Markets,” Journal of Economic Perspectives (Winter 2003), pp. 131–154.

Chapter 13: Best-Practice Tactics: Game Theory 479

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Lease Prices Reflect Anticipated Risks Of course, the risk of technological developments and competitor discounts that the manufacturer cannot control still remain. The lessor and lessee have credibly committed some things and left others to chance. All such remaining risks will be priced into the terms of the residual value lease. As a result, over the lifetime of the equipment it will not be cheaper to lease rather than to buy. In other words, the buyer who chooses an extensive repair and replacement warranty contract is imposing on the seller-lessor (say, BMW) the risk of product failure; this risk is then fully priced into a higher lease payment for a 3 Series or a Mini Cooper.

Nevertheless, manufacturers need some way of credibly committing themselves to maintaining high asking prices and a limited rate of planned obsolescence over the buyer’s holding period. Only then will early adopters pay the higher prices manufac- turers wish to charge in the early mature phase of an upgraded product’s life cycle. Closed-end leases with fixed residual values offer such a credible commitment because they demonstrate and certify just what the manufacturer’s best estimates of forward value truly are. Such leases therefore raise the acquisition prices early adopters of durable equipment are willing to pay.

SUMMARY

� Proactive oligopolists require accurate predictions of rival initiatives and rival response. The manage- rial purpose of game theory is to predict just such rival behavior. In a game theory analysis, each analyzes its competitors’ optimal decision- making strategy and then chooses its own best counterstrategy.

� Business strategy games may be classified as simultaneous-play or sequential-play, one-shot or repeated, zero-sum or non-zero-sum, two-player or n-player, and cooperative or noncooperative.

� Cooperative games allow coalition formation, side payment agreements, and third-party enforceable contracts, whereas noncooperative games prohibit these characteristics.

� Simultaneous-play games occasionally arise in pricing and promotion rivalry, but the essence of business strategy is sequential reasoning.

� Dominant strategy equilibrium entails actions that maximize at least one decision maker’s payoff, no matter what any other player chooses to do.

� Nash equilibrium strategy involves actions that maximize each decision maker’s payoff, given best-reply responses of the other players.

� Nash equilibrium for simultaneous games identi- fies both pure and mixed strategies. Stability of

tactical prediction arises from the fact that the players’ choices reflect best-reply reactions to one another, even though no sequential timing of the actions is involved.

� Most Nash equilibrium strategies are non-unique; multiple pure Nash strategies exist.

� Mixed strategy provides an optimal rule for ran- domizing one’s actions among multiple Nash equi- librium strategies.

� Mutual cooperation in a repeated Prisoner’s Di- lemma game can be secured with uncertain endgame timing, adoption of an industry standard, multiper- iod punishment schemes such as Tit-for-Tat or grim trigger strategy, and strategic hostage or bonding me- chanisms for establishing credible commitments and threats.

� Cooperation in noncooperative games is more likely if strategies are clear, provokable, take coop- erative initiatives unilaterally, and are forgiving so as not to perpetuate mistakes. The Tit-for-Tat strategy has these characteristics.

� Playing against Tit-for-Tat strategies necessitates comparing the additional profit from unilateral de- fection against the discounted opportunity loss from limited-duration certain punishment next period. As the probability of continued replay

closed-end leases with fixed residual values A credible commitment mechanism for limiting the depth of price promotions and the rate of planned obsolescence.

480 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

declines, Tit for Tat becomes a less-effective coor- dination device for escaping the Prisoner’s Dilemma than a matching price policy.

� The order of play matters in sequential games of coordination between manufacturers and distribu- tors, entry deterrence and accommodation, service competition, R&D races, product development, and so on, because rivals must predict best-reply responses and counter-responses all the way out to an endgame.

� Endgame reasoning looks ahead to the last play in an ordered sequence of plays, identifies the player whose decisions control the available outcomes in the endgame, and then predicts that player’s pre- ferred action.

� Subgame perfect equilibrium strategy looks ahead to analyze endgame outcomes and then reasons back to prior best-reply responses.

� Credible threats and credible commitments are the key to endgame reasoning, and therefore credibility mechanisms are the key to subgame perfect equi- librium strategy.

� Advantages may accrue to either first-movers or fast- seconds in a business rivalry. The former can credibly threaten or credibly precommit and therefore pre- empt some outcomes, whereas the latter replies and can determine the best-reply response in the end- game. Which is more advantageous depends on the particulars of the tactical and strategic situation.

� A credible threat is a conditional strategy the threat- maker is worse off ignoring than implementing. A credible commitment is an obligation the commit- ment maker is worse off ignoring than fulfilling.

� Mechanisms for establishing credibility include es- tablishing a bond or contractual side payment, in- vesting in a non-redeployable reputation asset, short-circuiting or interrupting the response pro- cess, entering into a profit-sharing alliance, taking small steps, or arranging an irreversible and irrev- ocable hostage mechanism.

� Closed-end leases with fixed residual values are a mechanism for establishing a durable goods man- ufacturer’s credible commitment to early adopters of new models not to discount deeply after the sale.

Exercises 1. Suppose that two Japanese companies, Hitachi and Toshiba, are the sole produ- cers (i.e., duopolists) of a microprocessor chip used in a number of different brands of personal computers. Assume that total demand for the chips is fixed and that each firm charges the same price for the chips. Each firm’s market share and profits are a function of the magnitude of the promotional campaign used to promote its version of the chip. Also assume that only two strategies are available to each firm: a limited promotional campaign (budget) and an extensive promo- tional campaign (budget). If the two firms engage in a limited promotional cam- paign, each firm will earn a quarterly profit of $7.5 million. If the two firms undertake an extensive promotional campaign, each firm will earn a quarterly profit of $5.0 million. With this strategy combination, market share and total sales will be the same as for a limited promotional campaign, but promotional costs will be higher and hence profits will be lower. If either firm engages in a limited promotional campaign and the other firm undertakes an extensive promotional campaign, then the firm that adopts the extensive campaign will increase its mar- ket share and earn a profit of $9.0 million, whereas the firm that chooses the lim- ited campaign will earn a profit of only $4.0 million. a. Develop a payoff matrix for this decision-making problem. b. In the absence of a binding and enforceable agreement, determine the domi-

nant advertising strategy and minimum payoff for Hitachi. c. Determine the dominant advertising strategy and minimum payoff for Toshiba. d. Explain why the firms may choose not to play their dominant strategies

whenever this game is repeated over multiple decision-making periods.

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

Chapter 13: Best-Practice Tactics: Game Theory 481

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

2. Consider the following payoff matrix:

Player B Strategy

21

$1,000 $2,000 1 $2,000 $1,000

Player A Strategy $1,000 $2,000

2 $2,000 $1,000

a. Does Player A have a dominant strategy? Explain why or why not. b. Does Player B have a dominant strategy? Explain why or why not.

3. Suppose that two mining companies, Australian Minerals Company (AMC) and South African Mines, Inc. (SAMI), control the only sources of a rare mineral used in making certain electronic components. The companies have agreed to form a cartel to set the (profit-maximizing) price of the mineral. Each company must de- cide whether to abide by the agreement (i.e., not offer secret price cuts to custo- mers) or not abide (i.e., offer secret price cuts to customers). If both companies abide by the agreement, AMC will earn an annual profit of $30 million and SAMI will earn an annual profit of $20 million from sales of the mineral. If AMC does not abide and SAMI abides by the agreement, then AMC earns $40 million and SAMI earns $5 million. If SAMI does not abide and AMC abides by the agreement, then AMC earns $10 million and SAMI earns $30 million. If both companies do not abide by the agreement, then AMC earns $15 million and SAMI earns $10 million. a. Develop a payoff matrix for this decision-making problem. b. In the absence of a binding and enforceable agreement, determine the domi-

nant strategy for AMC. c. Determine the dominant strategy for SAMI. d. If the two firms can enter into a binding and enforceable agreement, deter-

mine the strategy that each firm should choose.

4. Two insurance companies that manage employee benefit programs are bidding for additional business in their area of expertise at a market rate of $200 per hour. The potential customers refuse to leave their current suppliers and award benefit management contracts to the new firms unless billing rates are cut by $50. Abbott, Abbott & Daughters (AA&D) decides to do just that. Your firm, Zekiel, Zekiel & Sons (ZZ&S), must decide whether to match the price cut and then allow custo- mers to choose randomly between the two firms, or whether to lower rates still further to $100 per hour. Past experience suggests, however, that the price cutting may well not stop there. The clients will surely take their best current offer back and forth between the two firms, forcing a downward price spiral. The question therefore is, “How low will you go?” Crucially, this game has a stopping rule: At a price below your $40 cost, the additional business becomes unprofitable and must be refused. AA&D has higher costs at $66 per hour.

Again, your decision depends on an analysis of the sequence of predictable fu- ture events represented with a game tree or decision tree. Provide one. To sim- plify, assume that all rate cuts must be in $50 increments, that customers choose

482 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

quickly between equal rate quotes using fair coin tosses (represented by capital letter N for Nature), that once a rate quote has been matched it cannot be low- ered, and that many potential customers are present in the market. It is now your turn at node Z1 with rates at the $150-per-hour level. What should you do? Match rates or cut rates further?

5. How does the analysis and the strategic equilibrium outcome differ in Exercise 4 if the other firm enjoys a cost advantage (e.g., $35 at AA&D)? Then does the or- der of play (i.e., who goes first in making price cuts) matter in this bidding game with asymmetric costs?

6. Consider an ongoing sequence of pairwise marketing competitions between three companies with promotional campaigns of varying degrees of success. Each cam- paign involves comparative advertising belittling the target company. The com- pany with the most loyal customers (call this firm “Most”) enjoys 100 percent success when it attacks either of the others. The company with the least loyal cus- tomers (“Least”) has a 30 percent success rate when it belittles either Most or “More.” More experiences an 80 percent success rate. The firms each launch their advertising attacks one at a time in an arbitrary sequence. Least goes first and can attack either Most or More. More attacks second, and Most attacks third. If more than one of the opponents survives the first round of competition, the order of play repeats itself: Least, then More, then Most. Any player can skip his or her turn; that is, the three actions available to Least to initiate the game are as follows: attack More, attack Most, or do nothing and pass the turn.

Diagram the game tree and employ subgame perfect equilibrium analysis to identify the strategic equilibrium. What should the most vulnerable firm with the least loyal customers do to initiate play? What would be More’s best-reply re- sponse if attacked and More survives? What if Least did nothing? What would Most do when and if its turn arose?

7. Why should the early adopters of an information technology system provided by IBM Systems Solutions be willing to pay more for a closed-end lease of the servers and other hardware required than for an outright purchase?

8. People who are regularly late often don’t bother to carry watches. In response, other people tend to adjust to their tardiness by starting meetings 10 minutes after they’re scheduled, coming to lunch appointments 10 minutes late, and so on. An- alyze the following coordination game and explain why.

Harry

Be Punctual Always Late

Be Punctual

Tom

Always Late

07001

100 50

50 95

70 95

9. Nike and Adidas face the following coordination problem in trying to decide whether to conduct heavy or light combative advertising against the other firm. What should each firm do?

Chapter 13: Best-Practice Tactics: Game Theory 483

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Nike

Light Ads Heavy Ads

Light Ads

Adidas

Heavy Ads

$12 M $6 M

$10 M $4 M

$4 M $9 M

$5 M $8 M

10. The outcomes in the bottom half of the game tree describing the last (the 20th) submarket of the chain store paradox in Figure 13.2 are labeled N.A. (not appli- cable). Why? What specific equilibrium concept in sequential games rules out the applicability of these outcomes? Hint: How would you describe the game tree from node E onward as opposed to the game tree from node D onward?

11. What problems arise in PepsiCo’s couponing customers every other week to try to attract additional business? Would mail-order segmentation of PepsiCo versus Coca-Cola customers help this process? How?

12. Suppose you have announced you will “meet the competition” in response to en- try threats by a potential rival who has done marketing research in your target market and is offering a lower price point. What difference does it make, if any, if technology is moving very fast in the market so that this game proves to be one-time-only simultaneous play?

13. Analyze the following sequential game and advise Kodak about whether they should introduce the new product, Picture CD.

Kodak

Sony

Sony

Kodak

Kodak

Introduce picture CD

Do not introduce

Increase ads

Maintain ads

Increase ads

Maintain ads

High

Moderate

Low

High

Moderate

Low

New product Introduction

Rival advertising Pricing policy

$380m

$610m

$560m

$710m

$620m

$570m

$400m

$580m

$620m

$590m

$540m

$550m

$610m

$540m

$720m

$600m

Kodak Sony

14. Calculate the eight-hour-shift costs of operating a taxi with a medallion license that cost $125,000 borrowed at 10 percent interest assuming two shifts for 365

484 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

days per year, plus a $25,000 car that depreciates 50 percent in one year, plus $22 for gas and maintenance per shift per day. Would you pay $60/shift for a taxi operator’s license? Why or why not?

15. A math graduate student explains to her friend how to approach a group of smart attractive guys who have brought along famous actor Russell Crowe. What should her friend do? Ignore Russell Crowe or fixate on Russell Crowe? Explain the equi- librium reasoning underlying your answer.

Student 1

Ignore R.C. Fixate on R.C.

Ignore R.C.

Student 2

Fixate on R.C.

No date tonight Date with R.C. (Best)(Worse) Date

with other guys (Better)

No date tonight (Worse)

Date with other guys (Better)

No date ever (Worst)

Date with R.C. (Best)

No date ever (Worst)

Note: Best payoff—date with R.C., Better—date with other guys, Worse—no date tonight, Worst—no

date ever with any of these guys.

Case Exercise

INTERNATIONAL PERSPECTIVES

The Superjumbo Dilemma26

Boeing and Airbus complete wide-bodied commer- cial aircraft in several sizes at the rate of about one per day. Customers first pay a deposit of one-third of $84 to $127 million for a 767, one-third of $134 to $185 million for a 777, and one-third of $165 to $200 million for a 747, depending on how the planes are equipped. The second third is due after final assem- bly when the aircraft is painted, and the final third is due at delivery. Final assembly requires 15–25 days, the entire production schedule is 11 months long, and of course, design modifications add months to the front end of each project. The largest of the Boe- ing planes (the 747-400) carries 432 passengers; by comparison, the largest Airbus plane (the A380) car- ries 550 passengers.

As early as 1993, Boeing and Airbus entered into discussions to jointly develop a very large commercial transport (VLCT) with perhaps 1,000 seats. If each firm proceeded independently, the market for VLCTs is so small relative to the massive R&D costs that sizeable losses were assured. Either firm had superior profit available if it proceeded alone. Analyze this simulta- neous play noncooperative product development game and predict what Boeing and Airbus would do and why.

In fact, both the two competitors decided to enter into a strategic alliance with the option to develop a superjumbo or withdraw and maintain a wide-bodied aircraft focus. Analyze Boeing’s decision in light of its $45 million contribution margin on each 747 produced and sold. Net operating profit is about $15 million.

Chapter 13: Best-Practice Tactics: Game Theory 485

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Questions 1. In light of the foregoing payoffs, why did Airbus go ahead with the A380 Super-

jumbo even though its $10.7 billion development cost required as many as 250 planes to break even by 2010?

2. In 2004, Boeing produced fewer planes than Airbus (see Figure 13.9). If Boeing finds itself less profitable at 60 percent market share than at 45 percent, what is the likely impact on the Airbus-Boeing tactical competition?

Enter Strategic Alliance but Don’t Agree

to Develop VLCT Enter Strategic Alliance

and Jointly Develop VLCTBoeing\Airbus

Enter Strategic Alliance and Jointly

Develop VLCT

Enter Strategic Alliance but Don’t Agree to Develop

VLCT

Reduced development risk

Loss from alliance costs

Max default risk, possibly net profit

Cannibalize wide-bodied business

Max default risk, possibly a net profit

Loss from alliance costs

Loss from alliance costs but preserve wide-bodied business

Ongoing net profit of $15mm per wide-bodied plane

26Based on M. Kretschmer, “Game Theory: The Developer’s Dilemma, Boeing v. Airbus,” in Booz, Allen, and Hamilton, Strategy & Business (Second Quarter 1998); “Towards the Wild Blue Yonder,” The Economist (April 27, 2002), p. 67; “Giving ’em Away,” Busi- nessWeek (March 5, 2001), pp. 52–55; and “Global Dogfight,” Wall Street Journal (June 1, 2005), p. A1.

FIGURE 13.9 Wide-Bodied Aircraft Deliveries by Year

1974 1980 1990 2000

600

400

200

0 2010

Boeing

Airbus

Source: Wall Street Journal (October 14, 2003), p. A2; and The Economist (August 15, 2009), p. 11.

486 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

3. The EU claims that Boeing was given $3.2 billion in tax exemptions by the State of Washington to support the Boeing Dreamliner 787 project. The United States claims that Airbus received $6 billion in loans that do not need to be repaid to support the research, development, and launch aid for the Airbus 380 Super- jumbo. These charges and countercharges at the World Trade Organization per- tain to whether either firm is “dumping” when it brings the first 787s and 380s to market. What category of cost must be covered by the early penetration prices in order to avoid such charges of predatory pricing?

Chapter 13: Best-Practice Tactics: Game Theory 487

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

13A APPENDIX

Entry Deterrence and Accommodation Games

In this appendix, we examine the tactical issues that arise when an incumbent firm faces an imminent threat of entry. We analyze whether to accommodate or attempt to deter the potential entrant and what capacity planning or limit pricing or sunk cost investment tactics to employ in that effort. At the end, we characterize contestable markets as de- pendent on both entry and exit barriers.

EXCESS CAPACITY AS A CREDIBLE THREAT One type of credible threat or commitment that can markedly influence the subse- quent competition is an investment in non-redeployable excess capacity. Irreversible investment in excess capacity credibly commits a high-priced incumbent to serve the price-sensitive new customers who might be attracted into the market by a potential entrant’s discounting. If these and other regular customers can be expected to favor doing business with the incumbent, then excess capacity investment can substantially enhance the deterrent effect of an incumbent’s threat to cut prices in response to entry.

Why exactly does excess capacity enhance an incumbent’s threat to reduce prices should low-price entrants appear in the market? Is it that the incumbent can thereby prevent the new entrant from acquiring a large market share? Is it that the incumbent can deny the new entrant a unique reputation for low prices? Is it that the incumbent can become more profitable than before the entry threat? The answer to all these ques- tions is no. The sole reason any action or communication is credible is if it makes the threat-maker worse off ignoring the threat than carrying out the threat. In Figure 13A.1, the competitive firm that invests in excess capacity by expanding from Plant 1 to Plant 2 is worse off with unchanged output Q1 and unit costs of $180 at A than selling the larger output Q2 with unit costs of $120 at B. A noncompetitive firm that must lower price to carry out a threat thereby increases sales and also moves from Q1 to Q2. Ignor- ing the threat would leave the incumbent worse off with higher unit costs at A now that Plant 1 has been replaced by Plant 2.

PRE-COMMITMENTS USING NON- REDEPLOYABLE ASSETS To address the tactical ramifications of installing excess capacity, consider the capacity decision of a well-established hospital that faces an entry threat from an outpatient clinic specializing in obstetrics and elective plastic surgery. The hospital is constructing a new surgical wing. The hospital’s business manager can build a new facility to meet the future

488

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIGURE 13A.1 Excess Capacity Enhances Credibility in Entry Deterrence

Q1 Q2

$120

$180

$

A

B

SRACt SRAC1

LRAC

Output

Example Excess Capacity in the World Car Market: Samsung and Hyundai1

Auto sales in the United States, Europe, and Japan plunged during the Financial Crisis of 2008–2009. Why add excess capacity in such a business environment?

Economies of scale do not seem to be involved. Even Hyundai has already reached minimum efficient scale (see Table 13A.1). Moreover, the GM Opel- Fiat-Saab-Daewoo and Ford-Jaguar-Volvo-Land Rover-Mazda global alliances focus on designing common platforms for vehicle families so that the multimillion-dollar body-stamping machines and assembly plants can produce Opel Astra sedans one week and Fiat Zafira seven-seat minivans the next. Increasingly, up to a dozen differ- ent vehicles share the same platform and assembly line. Even with less popular vehi- cles, minimum efficient scale has therefore become much less difficult to achieve.

A second explanation for capacity expansions highlights the location of the new capacity, much of which is appearing in Asia, especially South Korea and Thailand. Two-thirds of the growth in new car sales between 2000 and 2010 has arisen in the de- veloping countries of China and India. Korean conglomerate Samsung recently opened a new robot-equipped 500,000-vehicle plant at an investment of $5 billion even though Korean production (at 6 million vehicles) is already substantially ahead of domestic consumption (1.5 million vehicles) plus export sales (3.5 million vehicles). Predictably, local retail auto prices inKorea collapsed as the looming overcapacity forced profit margins down to levels that no longer attract new auto industry investment.

But that may have been exactly the idea. Incumbent manufacturers like Hyun- dai and Kia want to deter further entry into an economy that can easily ship to the Asian growth market. Pre-committing to enough capacity that no potential entrant will doubt their threat to aggressively cut prices is a way to defend market share. If this tactical initiative works and potential entrants stay out, the incumbent firms will never have to make good on their threat.

1Based on Ward’s Automotive Yearbook, “Car Making in Asia: Politics of Scale,” The Economist (June 24, 2000), pp. 68–69; and “In Asia, GM Pins Hopes on a Delicate Web,” Wall Street Journal (October 23, 2001), p. A23.

Appendix 13A: Entry Deterrence and Accommodation Games 489

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

demand projected at their currently high prices, or she can include some considerable excess capacity in her expansion plans. Suppose that the birthing rooms and type of operating theater used in obstetrics and plastic surgery are not redeployable to general surgical or other specialized uses. Instead, the excess capacity, if built, will serve as a non- redeployable excess capacity pre-commitment by the hospital to compete for all the new price-sensitive business that a lower-priced outpatient surgical clinic might attract into the market.

The structure of this game is presented in the decision tree in Figure 13A.2. The hos- pital chooses excess capacity or not; the outpatient surgical clinic chooses thereafter to enter or stay out, and the hospital then controls the pricing endgame. If the hospital builds excess capacity, it is more likely to cut prices in the face of entry, and the clinic is then better off staying out. If the hospital does not build excess capacity, it is more likely to accommodate the entrant by maintaining high prices, and the clinic is then bet- ter off entering. Therefore, looking ahead to predict the hospital and the clinic’s best- reply responses in the various proper subgames and endgames, the hospital’s likely choices narrow to two strategies shaded in Figure 13A.2: {Excess capacity, Stay out, Limit pricing} and {No excess capacity, Enter, Accommodate with moderate price}.

Clearly, hospital business as usual is no longer an option. In particular, the very profitable prior elective surgical business with high prices, no excess capacity, and no

FIGURE 13A.2 Excess Capacity Pre-Commitment Game

Maintain high price

Limit pricing

Match low price

Maintain high price

Cut price below entrant

Stay out

Enter

Stay out

Enter

No excess capacity

Excess capacity

Note: I refers to the incumbent hospital, and PE refers to the potential entrant clinic.

Accommodate with moderate price

Limit pricing

Accommodate with moderate price

PE

I

PE

I

I

I

I

TABLE 13A.1 2005 WORLDWIDE AUTO AND LIGHT TRUCK SALES (TOP 10)

General Motors 17.3% Peugot Group 5.9%

Toyota 15.7% Nissan 5.6%

Ford 11.4% Chrysler 5.4%

VW Group 8.3% Hyundai Group 4.6%

Honda 7.3% Renault Group 4.2%

Source: International Organization of Motor Vehicle Manufacturers

490 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

competition in the top row of the game tree is no longer a focal outcome of interest. The entry threat may require that the incumbent hospital now maximize its remaining profit by pre-committing itself to constructing some excess capacity. In that scenario, then, the clinic will consider hit-and-run entry, but probably decide instead to stay out and enter the same market in another community with less capacity present or projected.

In general, whether incumbents will choose to deter potential entrants (e.g., in the bottom half of the game tree in Figure 13A.2) by the use of excess capacity pre- commitments or will actually prefer to accommodate (in the top half of the game tree) by retaining their smaller capacities and lowering prices is a complex question that depends on several factors. As we saw earlier, the answer depends in part on whether the incumbent can secure a first-mover advantage. Without it, in Entry Deterrence I, the incumbent Pastense discounted, but with it in Entry Deterrence II, Pastense moder- ated prices and accommodated entry.

This deterrence/accommodation decision also depends on whether the post-entry competition will be in prices among differentiated product sellers, each with some mar- ket power over price, or in quantities among homogeneous product sellers with no mar- ket power over price. Finally, the decision to deter or accommodate depends on how old and new customers in various segments of the market sort between an incumbent with excess capacity and a capacity-constrained lower-priced new entrant.

CUSTOMER SORTING RULES If the entrant attracts only new price-sensitive customers, that’s one thing. If, on the other hand, the new entrant takes away high-willingness-to-pay regular customers of the incumbent, that’s something else. Not surprisingly, the former situation more typi- cally leads to accommodation; the latter often leads to deterrence.

Probably the simplest customer sorting pattern of all is extreme brand loyalty to in- cumbents. In this case, even in the face of higher prices, customers reject the new en- trant’s offered capacity and instead back-order and reschedule when denied service at the incumbent. Competitive pressure from imitators normally erodes this degree of mar- ket power, but Microsoft Windows and popular local restaurants provide examples of products and services whose brand loyalty has sustained such a customer sorting pattern. At the other extreme, under efficient rationing of capacity, customers allocate them- selves across the fixed-priced capacity of new entrant discounters in a manner that achieves maximum consumer surplus. This customer sorting rule implies that those with the highest willingness to pay will exert the effort, time, and inconvenience to seek out, queue up, and order early to secure the lowest priced capacity. Of course, one obvi- ous qualification is that these customers may also have the highest opportunity cost of their time.

A third alternative is inverse intensity rationing, a much less threatening customer sorting pattern posed by new low-priced capacity in a segmented market. In this in- stance, the low-willingness-to-pay customers quickly absorb all the capacity of the low- priced entrant. Starting with a customer just willing to pay the entrant’s low price, one proceeds conceptually up the demand curve only as far as required to stock out the new entrant. In this instance, the demand of the incumbent may be largely unaffected if the discounter’s capacity remains relatively small. Finally, there is random rationing of the low-priced capacity. Under random rationing, all customers willing to pay the low prices —that is, both regular customers of the incumbent and the new customers attracted into the market by the entrant’s discounting—have an equal chance of securing the low- priced capacity. For example, if 70 customers were present in the market at the incum- bent’s original high price, and 30 additional customers appear in response to the

brand loyalty A customer sorting rule favorable to incumbents.

efficient rationing A customer sorting rule in which high- willingness-to-pay customers absorb the capacity of low-price entrants.

inverse intensity rationing A customer sorting rule that assures that low- willingness-to-pay customers absorb the capacity of low-price entrants.

random rationing A customer sorting rule reflecting randomized buyer behavior.

Appendix 13A: Entry Deterrence and Accommodation Games 491

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

discounts, the probability of any 1 of the 100 potential customers securing service from 40 units of low-priced capacity is 40/100 = 0.4. Conversely, the probability of not being served is (1 − 0.4) = 0.6, such that any customer may need to seek out higher-priced capacity with this probability. Under random rationing, therefore, the incumbent’s ex- pected demand falls as a result of the entry from 70 to (70 × 0.6) = 42.

Because of the pivotal nature of the decision timing and of these customer sorting rules in predicting deterrence versus accommodation behavior, game-theoretic analysis must often be intertwined with an industry study in order to discriminate among the many possible implications. Otherwise, the rational business decisions of incumbents in these models may vary over a wide range of alternatives from the relatively passive ac- quisition of excess capacity all the way to the aggressive incumbent who occasionally predates by pricing below cost with no prospect of later recovering the loss. For the pur- pose of predicting rival behavior, this state of game-theoretic knowledge presents some- thing of an embarrassment of riches. Hence, we reiterate the importance of doing sufficient field research to discover the particulars of the industry or firm-specific situation.

In Web Appendix D, we explore the entry deterrence and accommodation game be- tween US Airways and People Express and between United and JetBlue. Detailed cost, price, and realized revenue data allow us to distinguish among several pricing and capac- ity choice implications of sequential game theory. The analysis lends support to the im- portance of customer sorting “rules” in explaining why People Express met with little resistance and indeed was accommodated by incumbents in mid-Atlantic city-pair mar- kets, but encountered effective deterrence from US Airways in the southeastern city-pair markets. Eventually, People Express was forced to withdraw and exited from all of its southeastern routes.

A Role for Sunk Costs in Decision Making In both theory and practice, sequential games of entry deterrence and accommodation have uncovered a very rich variety of strategic incumbent behavior in response to entry or potential entry. These include the excess capacity pre-commitments just discussed as well as credible price discount threats. However, they also include price discrimination and capacity allocation schemes. Such yield management or revenue management sys- tems can provide incumbents with an effective way to deter new entrant discounters. We discuss revenue management in Chapter 14.

Finally, entry deterrence and accommodation strategy may also be expressed through advertising campaigns or other promotional investments in non-redeployable assets. As we mentioned in Chapter 13, some examples would be reputational investments in com- pany logos (such as CarMax) or unique stand-alone retail stores like McDonald’s. Such investments pre-commit the incumbent to aggressively defend market share and cash flow in order to recover the cost of these non-redeployable investments.

Non-redeployable investments in specific assets are a reality in many businesses; these fixed asset expenses are said to be “sunk.” Industrial machinery is often specialized to the purpose at hand and sometimes even to a particular supplier. For decades, Sara Lee Hosiery bought twisted nylon fiber for their highest quality hosiery from a sole source supplier, MacField Industries; the upstream nylon production equipment and the down- stream hosiery spinning equipment were only usable as complements to this one suppli- er’s twisted fiber input. Similarly, much of the trade secret knowledge discovered by Microsoft programmers is not easily packaged and separated out for redeployment and sale to another firm. Markets in which non-redeployable assets are common will be markets whose sunk cost conditions deter entry.

492 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Perfectly Contestable Markets Contestable markets are strategic industry groups in which new firms can enter and exit on short notice without anticipating losses due to sunk costs. Jet taxis provide an illus- tration. Even if only a few firms like Berkshire Hathaway’s NetJets dominate such a mar- ket, prices seldom rise above break-even levels because of the constant “hit-and-run” tactics of the frequent entrants. Rival firms jump in and scallop off the profits whenever prices rise and then escape quickly once the profits are dissipated. This hit-and-run entry/exit pattern ensures little divergence from cost-covering competitive equilibrium.

In the perfectly contestable markets scenario, incumbents react more slowly to entry threats than their regular customers, who chase after the most inexpensive supplier of the moment. In contrast, as we have seen in Figure 13A.2, proactive incumbents invest in excess capacity and non-redeployable assets in order to deter entry. That may sound like sunk-cost reasoning, and indeed it is. Recognizing the sequential nature of business strategy and the role of credible threats and credible commitments therein has led to a rehabilitation of the role of sunk costs in managerial decision making.

Example Contestable Market in Bicycle Helmets: Bell Sports2

Bell Sports began as a motorcycle helmet manufacturer with a small side-bet busi- ness in bicycle helmets and accessories. Today, Bell has $100 million sales of bicy- cle helmets, 85 percent of which occur in the United States. Twenty-seven states have initiated regulations making bicycle helmets mandatory for young riders. The potential growth in Europe, where bicycle helmets have become a fashion statement, is even greater. Prices range from $30 for colorful hard-shell designs to $140 for ultra-lightweight models.

The trouble with running a fast-growing niche business is that without sunk- cost investments, Bell inevitably attracted many new entrant competitors. Bicycle helmets are easy to fabricate and quickly sell themselves. All one needs is plastic molding machines and a foam extrusion process. These technologies are easily con- verted from many other industries and, more importantly, can be redeployed back to those other uses upon exit. The product sells well in bike shops and such dis- count stores as Kmart and Walmart without any significant sales force, point-of-sale actions, or after-sale service to differentiate one seller from another. Consequently, the bicycle helmet market is a classic case of contestable markets. Bell Sports is con- stantly subject to hit-and-run entry from other niche manufacturers—for example, Giro, Aurora, and Troxel Cycling.

The theory of contestable markets suggests that with no barriers to entry or exit and low customer costs of switching manufacturers, Bell Sports can never make more than a competitive profit in this business. As soon as prices rise above cost, temporary competitors enter the business, customers switch their allegiance, and Bell must lower prices. As a consequence, gross margins are low (averaging 8 per- cent) and fluctuate by as much as 50 percent from year to year. Bell’s only alter- natives are to outdo the hit-and-run entrants on new designs or to commit enough marketing investment dollars to establish a non-redeployable brand asset “Bell Hel- mets,” like they have established in the motorcycle business. Until then, entry de- terrence will prove infeasible, and entry accommodation must continue.

2Based on “Bell Sports,” Forbes (February 13, 1995), pp. 67–68.

contestable markets An industry with exceptionally open entry and easy exit where incumbents are slow to react.

Appendix 13A: Entry Deterrence and Accommodation Games 493

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

In fact, it is precisely because firms can do nothing about their sunk-cost investments, precisely because they are irreversible, irrevocable, and otherwise unrecoverable, that a particular threatened plan of action involving the use of these non-redeployable assets is credible. The player with sunk-cost investments has burned bridges. No better alterna- tives exist than to deliver on credible threats and to make credible commitments to re- main in the business of serving repeat customers until the equipment becomes obsolete or wears out. Again, best-reply reasoning is the key to credibility, and credibility is the key to subgame perfect equilibrium strategy.

Brinkmanship and Wars of Attrition Sometimes the question of tactical interest is not one of deterring or accommodating the entry of other firms but rather how long should your own firm stay in an obviously declin- ing business. In competing to win an exclusive license (e.g., to host the Olympics), to define an industry standard (e.g., for digital HDTV), to earn FDA approval for a new class of drugs, or to capture the product loyalty of fickle customers with advertising, successive rounds of bloodletting by multiple competitors may prevent profitability until someone con- cedes and drops out. Hence, these entry/exit games are sometimes called “wars of attrition.”

The first period of a multi-period sequential game representing a “war of attrition” is displayed on the left-hand side of Figure 13A.3. Each period requires a $10 million “ante” at the start of the period just to remain in the game. No competitor knows her rivals’ decision about staying when deciding whether to “hold ’em or fold ’em.” It’s as though auction bids were sealed and then opened simultaneously. If either firm concedes and leaves, it pays nothing, and the other firm’s $10 million ante is immediately recover- able by the firm who stays.

FIGURE 13A.3 A War of Attrition for HDTV Industry Standard

We

We

We

They

They

They

They

Stay

Concede and leave

They

They

Stay

Concede and leave

Stay

Concede and leave

Period one Period two

They We/Us

–10

0

92.6

46.3

–10

92.6

0

46.3

Stay

Concede

Stay

Concede

Stay

Concede and leave

Stay

Concede and leave

Stay

Concede and leave

Stay

Concede and leave

Note: The shaded box above refers to the veil of ignorance surrounding the decision “We” and “They” make each period, not knowing the rival’s choice before announcing our own. All payoffs are in $ millions.

494 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The market payoffs come at the end of the period and are equal to $100 million if one firm concedes and $50 million if both concede. These market prizes repeat every year until the end of the game. If at the start of any period the rivals (“They”) leave and concede the $100 million market to “Us,” a payoff of [$100/(1 + r)] awaits—for example, at 8 percent interest rates, ($100/1.08) – $10 + $10 = $92.6 million. If “We” concede when “They” stay, $92.6 million is their payoff. If both firms concede, they immediately merge, and the mar- ket is split 50–50 with no further cost. If “They” hang tough and stay in competition, and “We” do the same, both firms lose $10 million—that is, no one wins the prize that period. Then each proceeds to make the decision as to whether to spend another $10 million ante to compete in the next period. The question is, “How long should you stay?”

Consider the three-period game. If “We” leave now, the payoff is zero when “They” stay, and $50 million when “They” concede. Let p be the probability that our opponent is a type who will concede immediately. Then the expected payoff from conceding imme- diately ourselves is just3

$50p + 0ð1 − pÞ ≥ 0 [13A.1] If we leave at the start of the second period, our expected payoff would be equal to

$100p + $50q − $10 ≥ 0 [13A.2]

where q is the probability that our rival will concede at the start of the second period, and (1 – p – q) is the probability “They” will stay until the third period—that is, Hang Toughers never concede. If we stay to the end, our expected payoff is as follows:

$200p + $100q − $20 ≥ 0 [13A.3]

Setting Equation 13A.1 equal to 13A.2 and 13A.2 equal to 13A.3, then solving simulta- neously, yields the values of p and q that would leave “Us” just indifferent between con- ceding and hanging tough. Collecting terms and simplifying, we have

50p + 50q = 10 [13A.4] −100p − 50q = −10 [13A.5]

which together imply p = 0, q = 0.2, and (1 – p – q) = 0.8. In other words, “We” are indifferent as to conceding or hanging tough and paying

$10 million each of the first two periods to stay until the endgame if and only if there is no less than a 20 percent chance “They” will leave in the second period and no more than an 80 percent chance “They” will stay until the endgame. “We” decide whether to stay or leave by assessing the actual situation and actual rival and then comparing these 20 percent and 80 percent derived probability break points against our subjective proba- bility estimates of the actual situation.

TACTICAL INSIGHTS ABOUT SLIPPERY SLOPES Note that p = 0 implies that neither party concedes immediately. Instead, q = 0.2 indi- cates a middle ground strategy of opponent types who test the competitive waters be- fore conceding at the start of Period 2. This positive probability of “middle grounders” is a reflection of a slippery slope. Once you enter into a war of attrition and make your

3In the discussion that follows, we ignore discounting to simplify the analysis. The payoff of $92.6 million therefore becomes $100 million.

slippery slope A tendency for wars of attrition to generate mutual losses that worsen over time.

Appendix 13A: Entry Deterrence and Accommodation Games 495

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

own first “ante” of −10 (because the Bayesian probability of Hang Toughers was less than 0.8), the probability of middle grounders who will step on that slippery slope with you is NOT zero. Instead, we have seen in the simultaneous solution to Equations 13A.4 and 13A.5 that the equilibrium probability of middle grounders in this game is 0.2. That is, with the parameters assumed in this war of attrition, there is a 0.8 probabil- ity of a sequence of mutual losses, a death spiral until the less deep pocket is empty. Hence, these so-called brinkmanship games have serious and largely uncontrollable consequences even for the player with an apparently advantaged position. If you misesti- mate the depth of an opponent’s pocket, 80 percent of the time you will be up against Hang Toughers and stepping onto a slippery slope to financial ruin—your own financial ruin.

What is a best-reply response in the n period game? Continuing to ignore the dis- counting of future cash flows for the moment to simplify the analysis, if we hang tough until the rivals leave and if the rivals leave in period t, we realize the $100 million market prize for n – t periods and pay $10 million for t periods. Since if we leave now, zero (i.e., $50p = 0) is the payoff, and the expected value of all other alternatives can be no worse than this; otherwise we’ll just leave now. Combining these facts

ðn − tÞ$100 − t$10 > 0 [13A.6] 0:91n > t [13A.7]

where 0.91 is the ratio of the $100 million prize to the sum of (the periodic cost $10 million plus the $100 million prize).

How should we interpret Equation 13A.7? If “We” believe the rivals will stay 91 per- cent of the total time or less, we should stay ourselves. If we believe “They” will stay more than 91 percent (or, with discounting, 90.25 percent) of the total time, we should concede immediately and save our $10 million ante to invest in another competitive

Example Circuit City Driven over the Brink4

Consumer electronics has often been a brutally competitive business with retailers offering one price promotion after another (10 percent below the best price else- where, 20 percent below, 30 percent below) to try to attract business away from rivals. In 2004, Circuit City with $4.4 billion in sales got 35 percent of its revenue from consumer electronics, 24 percent from audio products and entertainment software, and 41 percent from videos, video games, and video game equipment. Best Buy with $11.6 billion in sales in 2004 got 37 percent from consumer elec- tronics, 19 percent from entertainment software, 6 percent from appliances, and 38 percent from home office equipment.

Best Buy and Circuit City did enter a war of attrition. DVD players launched in 1997 at an initial price of $840 were discounted to $571 by 1998, $467 by 1999, $345 by 2000, and collapsed to cost in 2001. Similarly, Blu-ray players launched in 2006 at a price of $800 were discounted to $497 by 2007, $388 by 2008, and $322 by 2009, and then collapsed to a cost of $221. Circuit City was forced into bankruptcy and exited the market.

4“Prices No Longer Red Hot,” Wall Street Journal (December 23, 2009), p. D9.

496 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

encounter elsewhere. For the three-period case, if “We” believe the rivals will stay less than 0.9025 × 3 = 2.71 years, we should hang tough and stay to the end ourselves.

Obviously, these calculations apply equally to the symmetrically positioned “They” as well, so wars of attrition quickly become a matter of bluffing and signaling. The most useful insight the previous analysis offers in such entry/exit games is that each player should assess the probability of her rivals’ leaving in light of all the available evidence and should hinge her own decision on the ratio of the prize to the sum of the prize plus periodic cost. Very much like playing straight draw poker, you hold rather than fold the bigger the prize and the smaller the periodic cost of hanging tough and calling your rivals’ bluff.

SUMMARY

� Incumbents may seek to deter potential entrants through the use of excess capacity pre-commit- ments or credible threats of advertising campaigns and price discounts.

� Whether incumbents deter or accommodate po- tential entrants depends in general on the presence or absence of first-mover advantages, on the struc- ture of competition in prices versus quantities, and on how customers sort across alternative firms when the low-priced capacity stocks out.

� Customer sorting patterns include the following: random rationing in which all customers are equally likely to obtain the low-priced capacity; efficient rationing in which the highest (then next highest) willingness-to-pay customers obtain

the low-priced capacity until it is exhausted; extreme brand loyalty in which none of the regular customers seek the low-priced capacity; and inverse intensity rationing in which the lowest (then next lowest) willingness-to-pay customers obtain the low-priced capacity until it is ex- hausted. With inverse intensity rationing, the customer sorting implies a segmented market and is most likely to lead to accommodation of entry.

� In wars of attrition, whether to hang tough and stay in competition for a market price or concede defeat and leave depends on the ratio of the prize to the sum of the prize plus the periodic cost of competing.

Exercises 1. Dunkin’ Donuts and McDonald’s McCafé have entered the specialty coffee busi- ness pioneered at Starbucks’s 5,439 locations. Prices are 20% lower (99 cent espresso shots versus $1.45 at Starbucks), orders are simpler (Large Mocha Swirl Latte at $2.69 versus Venti Caffè Mocha at $3.35), and the wait time is under a minute versus three to five.5 Have Starbucks Frappuccinos become an affordable luxury? Or are espresso coffee and flavored coffee going mainstream where Dun- kin’ Donuts and McDonald’s have 17 percent and 15 percent of the fast-food out- let brewed coffee business, respectively, to Starbucks’s 6 percent? Which is happening in your city? Moreover, as McCafé approaches Starbucks’ market, what customer sorting rule will likely apply? Is there any reason to believe Star- bucks will have a different strategy in responding to Dunkin’ Donuts’ 4,100 stores versus the several hundred McCafés under development? Why?

5Based on “Latte versus Latte,” Wall Street Journal (February 10, 2004), p. B1.

Appendix 13A: Entry Deterrence and Accommodation Games 497

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

2. Sony’s PlayStation 2 (PS2) dominated the game console market 1997–2003 with 123 million units sold. Today, Nintendo’s Wii has 62 percent of the market. Prices to achieve this spectacular growth have fallen continuously from $400 at launch to $250. Sony’s PS3 with 20 percent market share at $1,000, then $600, and now $399 competes in a more up scale segment of the game console market against Microsoft’s Xbox at 18 percent market share with prices that started at $700. Initially Xbox and PS3 tried to compete head-to-head on add-ons such as extra hand controller sets. Now, brinkmanship pricing has broken out with Xbox 360 discounted from $760 to a range of $400 to $343. Which firm is likely to blink first?

498 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

14 CHAP T E R

Pricing Techniques andAnalysis CHAPTER PREVIEW This chapter builds on the price and output determination models developed in Chapters 10–13 as it considers more complex pricing issues. The first two sections examine a value-based pricing conceptual framework. Then we characterize differential pricing in segmented markets where different target customers are charged non-uniform prices. Differential pricing is often accomplished with bundled pricing, couponing, and two-part tariffs (an access or entry fee combined with a user fee). Finally, we discuss the concept of pricing throughout the product life cycle including target pricing, penetration pricing, pricing for organic growth, limit pricing, and niche pricing. We conclude with a section on pricing of goods and services sold over the Internet. Finally, applications of revenue management in airlines, fashion clothing, consulting firms, and baseball are explained. Together, the pricing practices presented in this chapter provide an extensive overview of the way actual managers apply pricing techniques to maximize shareholder wealth.

Two additional pricing topics closely related to accounting (pricing of joint products and transfer pricing) are presented in Web Appendix E.

MANAGERIAL CHALLENGE Pricing of Apple Computers: Market Share versus Current Profitability1

Apple Computer manufactures and sells iPad tablet computers, iMac personal computers, Mac Mini home entertainment systems, iPhones, as well as the famous iPod. All these products are innovative in design and stylistically striking. Some people compare the lifestyle choice of a Mac to driving a sports car rather than a minivan.

Historically, Apple priced its products higher than similar models of other computer makers. For example, despite price cuts throughout the 2000s, iMacs are still priced $300 higher than somewhat comparable Dell lap- tops and iPads cost $500 relative to Amazon Kindle’s $300. Nevertheless, after opening Apple stores in major cities in 2007–2009, Apple’s margins increased (from 24 percent to 29 percent) and market share exploded upward in several product lines.

Apple emphasizes the high value-in-use of its dis- tinctive products even though the resulting lower mar- ket share discourages independent software providers (ISPs) from writing software applications for the Mac. Some observers believe that the company would have to increase its market share from 12 percent to as much as 20 percent to accomplish the ISP objective. Apple’s ex- ecutives defend the company’s high prices, saying that Apple has to improve its balance sheet so it can con- tinue research and marketing efforts and idea creation. Is Apple’s decision to charge premium prices for its products the right approach?

Steve Jones, former chief marketing officer of Coca- Cola, agrees that a company with leading-edge products must remain positioned at a high-end price point. “Create sparks, engage in reverse process thinking, clarify

499

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

A CONCEPTUAL FRAMEWORK FOR PROACTIVE, SYSTEMATIC-ANALYTICAL, VALUE-BASED PRICING In the past, pricing decisions were often treated as an afterthought. Companies either routinely marked up prices or reacted in an ad hoc fashion to a competitor’s discounting. Today, systematically analyzing the customer value basis for an asking price and there- after carefully selecting which orders to accept and which to refuse has become a critical success factor for many businesses.

Pricing decisions must always be systematic and analytical, based on hard facts in- stead of ad hoc hunches. In the men’s aftershave business, an established incumbent re- cently encountered a new entrant with a penetration price 40 percent below the leading brands Skin Bracer, Old Spice, and Aqua Velva. The incumbent increased advertising but maintained its original price point and was astounded to observe a 50 percent decline in market share through its grocery store distribution channel. Only after the fact was a

the innovation, define the destination for the brand, in- spire, and then get out of the way. Trend-setter customers

along with company staff will then co-create exciting new sources of customer value.”2 The events taking place daily in Apple stores all across North America, Europe, Japan, East Asia, and Australia personify this way of thinking.

Discussion Questions

� Have you visited an Apple store? Did the ex- perience introduce you to value-added features of the products you were considering?

� Is the price premium for Apple products like iPhones and iPads sufficiently small to warrant selecting its offerings over the competitors’?

� Have you owned several versions of upgrades of computer products? In a product life cycle sense, why would Apple not include all its value-added features in its original rollout of a breakthrough product like the iPad?

1Jim Carlton, “Apple’s Choice: Preserve Profits or Cut Prices,” Wall Street Journal (February 22, 1995), p. B1; and “Apple Gets Vote of Confidence,” Wall Street Journal (September 14, 2006), p. B1. 2Remarks at the Wake Forest MBA Marketing Summit, “Using Advertising and Creativity to Drive Your Brand,” Winston-Salem Journal, North Carolina (February 13, 2006).

MANAGERIAL CHALLENGE Continued ©

Ru bb er ba ll/ Ge tty

Im ag es

500 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

systematic analysis undertaken. Careful demand estimations showed that customers in the grocery store channel were price elastic and advertising inelastic.

Proactive pricing must also be tactically astute and internally consistent with a company’s operations strategy. A high-cost, full-service, hub-and-spoke airline cannot slash prices dramatically even if it means 10 or 20 percent increases in market share in a high-margin segment. It must instead anticipate a matching price reaction by its lower-cost rivals, perhaps followed by still further price cuts below its own cost. Knowing these probable reaction paths in advance makes the attempt to gain market share through discounting much less attractive despite additional incremental sales at a high margin.

Most importantly, pricing should be value based. Prestone and Zerex sell leading anticorrosive radiator fluids whose product characteristics warrant a price premium. Under apparent price pressure, Zerex often simply matches any competitor’s discount as long as competing prices on generic radiator fluid cover Zerex’s cost. A thorough value analysis reveals, however, that this reactive cost-based pricing fails to realize about one-third of Zerex’s sustainable profit margin. Cost-based pricing has been called one of the “five deadly business sins” by Peter Drucker; what firms should do instead is “price- based costing.” That is, firms should segment customers, perform an extensive customer value analysis, and then develop products whose costs allow substantial profitability in each product line the firm chooses to enter. Each firm’s marketing and operations capa- bilities are then key to sustaining that profitability.

Costs are not irrelevant. Indeed, a key to effective pricing management is to know precisely what activity-based costs are associated with each type of order from each cus- tomer segment. Knowing these costs allows firms with optimal differential prices to iden- tify which orders to refuse. This insight—that every company has orders that it should refuse—is the key to a new set of pricing techniques known as “yield management,” or more generally, “revenue management (RM).” In an RM approach, costs become the consequence of a value-based pricing and product development strategy.

The appropriate conceptual framework for setting prices is the target customer’s value-in-use. Value-in-use is the cost savings that arise from the use of your product or service relative to a next best competitor. A faster commute on a toll road or a nonstop jet to a distant city saves the $220-an-hour attorney or accountant’s time value per hour. A Google ad with a documented click-through rate saves the advertising expenditure on magazine ads or TV commercials. An integrated and easy-to-use Canon digital photo- graphy system saves the casual photographer time, money, and inconvenience in image capture, photo editing, developing, distribution, and storage.

Table 14.1 lists various tangible and psychological sources of value-in-use, including product specifications and ease of use, delivery reliability, service frequency, change order responsiveness, loyalty programs, and empathy in order processing. Many of these sources of cost savings are functional points of differentiation, but others are relationship-based. In addition, marketing communications seeks to position and brand the product through advertising, personal selling, and event marketing. Viral marketing identifies trend setters among the target customer group and seeks to place the product with those individuals, hoping that others will follow their lead. Because consumers strive to avoid psychological dissonance, products that affirm a particular lifestyle or group identity can often generate perceived value well beyond tangible cost savings. Coca-Cola and Starbucks each offers a lifestyle association and identity that would other- wise necessitate much larger clothing, travel, auto, and entertainment expenditures to achieve a similar result.

Importantly, lowest price is seldom what triggers a purchase. Instead, a target custo- mer’s purchase is triggered by either (1) value through functions, cost savings, and

value-in-use The difference between the value customers place on functions, cost savings, and relationships attributable to a product or service and the life cycle costs of acquiring, maintaining, and disposing of the product or service.

Chapter 14: Pricing Techniques and Analysis 501

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

relationships that exceed the product’s asking price or (2) a ratio of value to asking price that is greater than a competitor’s. In Table 14.1, $0.50 is not the lowest price for a digi- tal photography print, but Kodak’s offering will nevertheless trigger a purchase if the ex- cess customer value with the Kodak digital photography system ($2.00 − $0.50) exceeds the excess value from less valuable $0.29 products—perhaps ($1.00 − $0.29).

Consequently, firms should begin their pricing decisions by identifying the value dri- vers in each and every customer segment. Business air travelers, for example, value con- formance to the schedule, delivery reliability, and the ability to change itineraries on

Example Coated Coronary Stents Reduce the Cost of Later Surgeries3

Each year in the United States, 1.3 million angioplasty heart surgeries are per- formed at an average cost of $48,399 and 448,000 coronary bypass (open heart) surgeries are performed at an average cost of $99,743. This $104 billion of cardiac surgery is very big business. In the relatively simple noninvasive angioplasty proce- dure, a balloon is threaded into the femoral artery along the thigh and then di- rected upward into a clogged coronary artery, where it is inflated to clear the blockage. Eighty percent of the time, an $800 medical device (a stent) made of coiled wire resembling the spring in a ballpoint pen is then inserted into the coron- ary artery to hold it open. The coronary stent market grew from introduction in the mid-1990s to $6 billion worldwide in 2006 with over a million stents sold an- nually in the United States and an almost equal number sold abroad. In 15 to 30 percent of these cases, however, scar tissue grows around the wire stent, reclosing the artery. That complication then necessitates $99,743 for follow-up open heart surgical procedures and additional treatments.

Recent advances by Cordis, a subsidiary of Johnson & Johnson, and by Boston Scientific and Medtronic have coated the coronary stent with an antibiotic or a cell-killing cancer drug to stop the scar tissue from forming. In its most advanced form, hollow stents can be made to eleut just the right amount of time-release phar- maceutical over a 45-day post-op period to prevent the vessel wall scarring. J&J’s sirolimus-eleuting stent called CYPHER has practically eliminated the problem in routine angioplasties, reducing the incidence of re-blockage to 3 percent in various human trials. Achieving a 12 to 27 percent reduction in the likelihood of $99,743 open heart surgery saves the patient at least 0.12 × $99,743 = $11,969 in expected future medical expenses as well as avoiding the risks of the additional surgery.

Hospitals and their attending surgeons charge on average $9,700 for the over- night angioplasty procedure plus insertion of an uncoated bare metal stent (billed to patients separately at an additional $1,165) for a total cost of $10,865. The sim- plest coasted stent is billed to patients at just under $2,000 for a total cost of $11,700 relative to the $11,969 minimum value-in-use savings it creates. The so- phisticated drug-eleuting CYPHER coated stent, priced at $4,150 for a total bill of $13,850, has experienced less acceptance in the marketplace. For patients with the lowest (15 to 17 percent) chances of scar tissue closure, value-in-use is less than the asking price.

3Based on “Medical Device Maker Sees Vast Market for Cardiac Stent,” Miami-Herald (March 16, 2003); “How Doc- tors Are Rethinking Drug-Coated Stents,” Wall Street Journal (December 9, 2006), p. A1; “Use of Coated Stents on the Rise,” Wall Street Journal (July 16, 2008), p. D2; and “Alternative Medicine Is Mainstream,” Wall Street Journal (January 9, 2009), p. A13.

502 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

short notice more than they value frequent flights, good meals, and wide seats. Because that first set of process-based value drivers is harder to imitate, sustainable price pre- miums are often associated with those operations processes rather than the product or service characteristics of flights, meals, and seats. Because business travelers account for only 27 percent of the traffic but 80 percent of the profitability, a critical success factor for legacy airlines is to have hub processes and route planning that sustain these hard- to-imitate processes on nonstop flights.

In sum, pricing decisions should be proactive and systematically analytical, not reac- tive and ad hoc. Most importantly, pricing should be value-based, not cost-based. The value-in-use conceptual framework leads naturally to a differential pricing environment in which mass-produced products or services are customized to the requirements of tar- get customer segments.

OPTIMAL DIFFERENTIAL PRICE LEVELS The first step in setting optimal differential prices is then to estimate demand by market segment—say, for each of two customer classes (business and nonbusiness air travelers) on the Thursday 11:00 A.M. flight from DFW to LAX. The expense account business trav- eler tends to make less flexible travel plans and reserve space later and thus faces fewer close substitute alternatives than the nonbusiness traveler. Average revenue and marginal revenue schedules for business travelers therefore prove to be less price elastic than for nonbusiness travelers, as indicated in Figure 14.1.

TABLE 14.1 CONCEPTUAL FRAMEWORK FOR VALUE-BASED PRICING

Consumer analysis of segmented wants (trends, motivation,

unmet needs)

Products/services positioning

Competitor analysis of differentiation

(strengths, vulnerabilities, rival positioning)

Marketing strategy

Product functions/uses Conformance to specs

Ease of use

Distribution/service Delivery reliability Delivery frequency Change order responsiveness

Customer relationships Loyalty programs Empathy in order processing

Communications Advertising Personal selling Event marketing Viral marketing Other activities

Perceived value to target customers Rival marketing, Rival products

Pown [Perceived excess value] Prival

Excess value then triggers a purchase CASE A: [Value-in-use P*own] 0 PURCHASE [(Functions Cost-saving processes Relationships) (Acquisition Pown Life cycle costs)] ($2.00 worth of an emotional moment ($0.50) memorialized/edited/printed/distributed/stored)

CASE B: [Excess own value-in-use] [Excess rival value-in-use] PURCHASE [$2.00 – $0.50] [$1.00 $0.29]

Chapter 14: Pricing Techniques and Analysis 503

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Graphical Approach Previously, the airline’s capacity planning department will have summed all the expected marginal revenues E(MR) from the various segments and determined an optimal total capacity by setting summed marginal revenue [∑E(MR)] equal to the marginal cost of the last seat sold (MClss).

4 The result in Figure 14.1 is that a plane with 170 seats should be scheduled for the Thursday 11:00 A.M. flight departure.

One may think of the optimal differential pricing decision as determining how this total capacity of 170 seats should be allocated across the customer segments. Because at the margin a firm forgoes revenue unless the last customer in each segment contributes a marginal revenue equal to the marginal cost of the last seat sold (MClss), the optimal al- location of seating capacity results from equating the segment-level MRs to one another:

MRbus = (MClss) = MRnonbus [14.1]

which in Figure 14.1 is at MR = $130. Consider a case in which this condition does not hold. Suppose the 62nd seat sold in the business class contributed $160 of marginal rev- enue and the 108th seat sold in the nonbusiness class contributed $120. Clearly, one could raise $40 additional revenue for unchanged costs by selling one less seat in non- business and one more in business, leaving both classes with, say, an MR = $130.5

FIGURE 14.1 Optimal Differential Pricing and Capacity Allocation (45 Days in Advance) for Thursday 11:00 A.M. Flight from Dallas to Los Angeles

$261

$188

$130

Pr ic

e ($

)

Seats17010763

∑E(MR)E (MRnonbus)E (MRbus)

E (Dnonbus) MClss

E (Dbus)

0

4To find aggregate demand, remember that individual demands (and MRs) are horizontally summed for rival goods that cannot be shared (such as airplane seats and bite-sized candy bars), whereas demands for nonrival goods (e.g., outdoor statues, tennis courts, and national defense) are vertically summed. 5Note that the MR of each segment is not set equal to MC. Rather, the summed MR of all segments has been set equal to MC. The individual MRs are set equal to the MC of the last unit sold (i.e., $130), and therefore equal to one another.

504 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

What prices can achieve the capacity allocation of 63 seats to the business class and 107 seats to the nonbusiness class? The answer is deceptively simple. Optimal differential prices are whatever asking prices will clear the market if the firm supplies 63 and 107 seats in these two fare classes. In Figure 14.1 the answer appears to be $261 and $188 with some effective barrier or “fencing” that prevents resale from the lower to the higher fares. The difficulty, of course, is predicting demand sufficiently well to know what prices will have this effect for the 11:00 A.M. flight next Thursday.

Algebraic Approach Table 14.2 shows the spreadsheet data on which such a decision would be based in prac- tice. The first three columns show the number of seats demanded, fares, and marginal

TABLE 14.2 ALLOCATING AIRLINE CAPACITY WITH DIFFERENTIAL FARES FOR LEISURE

AND BUSINESS

BUSINESS CLASS LEISURE CLASS

EXPECTED SEAT

DEMAND FARE

EXPECTED MARGINAL REVENUE

EXPECTED SEAT

DEMAND FARE

EXPECTED MARGINAL REVENUE

TOTAL SEATS

MARGINAL COST

1 $1,084 $1,084 1 $87

2 1,032 980 2 87

3 974 858 3 87

4 907 705 4 87

5 835 550 5 87

10 613 390 10 87

1 $342 $342 87

2 331 320 95

3 319 294 95

4 311 288 95

20 456 280 5 305 280 25 95

10 280 256 95

20 260 240 95

30 381 230 30 250 230 60 100

40 240 210 100

50 231 194 100

40 331 180 60 222 180 100 112

70 214 162 112

50 295 150 80 206 150 130 112

90 198 140 120

60 268 133 100 192 133 160 125

63 261 130 107 188 130 170 130

110 186 128 140

70 252 122 120 181 122 190 155

130 176 115 170

80 235 110 140 173 110 220 190

Chapter 14: Pricing Techniques and Analysis 505

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

revenue for business-class travelers. For example, at a fare of $1,084, only one seat on the entire plane would be sold, and it would go to a business-class passenger. If the fare falls to $1,032, two seats are taken by business-class passengers. At a fare of $974, three seats are taken, and so on. Expected marginal revenue is the increase in total revenue realized from selling one more seat in the business class. For example, when a single seat is sold at $1,084, total revenue is also $1,084. When two seats are sold at a fare of $1,032, how- ever, total revenue jumps to $2,064, and marginal revenue, which is the difference in to- tal revenue realized from selling one more seat, is $2,064 minus $1,084, or $980. Similarly, the marginal revenue associated with the third seat sold is $2,922 minus $2,064, or $858.

Table 14.2 also shows corresponding information for leisure-class passengers. Note that the first leisure-class seat is sold at $342, the second at $331, and so on. The last two columns depict total seats sold and marginal cost, which is the variable cost associ- ated with serving one additional passenger in either class.

Using this simple two-booking-class example, marginal revenue equals rising mar- ginal cost at $130 per seat. (Marginal cost increases by steps with the addition of flight attendants needed to serve additional passengers and the additional fuel consumed be- cause of worsening aerodynamics at high load factors.) At MC = $130, optimal fares are obtained by equating individual marginal revenues of both segments and the mar- ginal cost of the last seat expected to be sold (the 170th seat in this example). Business and leisure traveler marginal revenues equal $130 at 63 and 107 seats, respectively, and fares of $261 and $188 are optimal at these seat allocation levels.

Multiple-Product Pricing Decision Figure 14.2 illustrates an analogous decision with five products; D1 represents the de- mand for Product 1, D2 for Product 2, and so on. Again, profits are maximized when the firm produces and sells quantities of the five products such that marginal revenue is

FIGURE 14.2 Multiple-Product Pricing

Output Q (units)

MC

Q2

I

Q1 Q3 Q4 Q5

EMR

MR5

MR4 MR3MR2

MR1

P1 P2 P3

P4 P5

D5

D4

D3 D2

D1

0

Pr ic

e an

d co

st (

$/ un

it )

506 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

equal in all markets and equal to marginal cost. The line EMR represents equal marginal revenue, the firm’s marginal revenue opportunity in other product lines. Because it is as- sumed that new product markets were entered in order of their profitability, the prices charged for the five products are arranged in declining order, from P1 to P5, and the elasticity of demand increases from D1 to D5. The height of the EMR line is determined by the intersection of the firm’s marginal cost curve MC and the marginal revenue curve for the last product market that may be profitably served, MR5 at Q5.

The equilibrium condition in the marginal market D5 where P, MR, and MC are vir- tually equivalent illustrates the well-known fact that nearly all firms produce some prod- ucts that generate little or no incremental operating profit and are on the verge of being dropped or replaced because the contribution margin approaches zero.

Differential Pricing and the Price Elasticity of Demand In all of the preceding examples, an inverse relationship exists between optimal price and the price elasticity of demand in separate submarkets. Recall that for profits to be maxi- mized, marginal revenue must be equal in each of the separate submarkets. In Chapter 3 the relationship between marginal revenue (MR) and price (P) was shown to be the fol- lowing (Equation 3.7):

MR = P 1 + 1 ED

� � [14.2]

where ED is the price elasticity of demand. If P1, P2, E1, and E2 represent the prices and price elasticities in the two submarkets, we may equate marginal revenue by setting equal

MR1 = P1 1 + 1 E1

� � and MR2 = P2 1 +

1 E2

� � [14.3]

Hence,

P1 1 + 1 E1

� � = P2 1 +

1 E2

� �

P1 P2

= 1 +

1 E2

� �

1 + 1 E1

� � [14.4]

Example Supermarket Pricing Supermarkets provide an illustration of this multiple-product pricing model. The primary resource constraint of a supermarket is shelf space, which can be allocated among a wide variety of product categories (meat, dairy products, canned goods, frozen foods, and produce). Canned goods have only a 1–2 percent profit margin. Generally the markups and profit margins on staple items, such as bread, milk, and soap, are lower than on nonstaple items, such as imported foods and deli items. In an effort to increase their overall profitability, many supermarkets have added higher-profit-margin categories, such as delicatessens, in-store bakeries, fresh fish, and floral departments by reallocating existing shelf space to these new categories.6

6Allocation of shelf space within each product category also involves a consideration of profit margins when making de- cisions about stocking private label versus national brand canned goods, prepackaged versus fresh-cut meat, and so on.

Chapter 14: Pricing Techniques and Analysis 507

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Perhaps JetBlue Airways has determined that the price elasticity of demand for two customer segments (New York to Los Angeles unrestricted coach and for Super Saver Saturday night stay overs) is −1.25 and −2.50, respectively. To determine the relative prices (P1/P2) that JetBlue should charge if it is interested in maximizing profits on this route, substitute E1 = −1.25 and E2 = −2.50 into Equation 14.4 to yield

P1 P2

= 1 +

1 −2:50

� �

1 + 1

−1:25

� � = 3:0

or

P1 = 3.0 P2

Thus the price of an unrestricted coach seat (P1) should be 3.0 times the price of a Super Saver coach seat (P2).

Price elasticity is the key; the larger the number of close substitutes, the higher the price elasticity of demand, and therefore the lower the optimal markup and price-cost margin. In electricity pricing, industrial customers such as factories and hospitals can now buy their power from competing public utilities. The industrial customer has so many more close substitute alternatives that the price per kilowatt hour is less than half the price of residential or small commercial users. Again, the higher the price elasticity, the lower the optimal price, ceteris paribus.

The increase in profitability from engaging in differential pricing as opposed to uni- form pricing across all customer segments can be illustrated with the following example.

Example Differential Pricing at Taiwan Instrument Co. Taiwan Instrument Company (TIC) makes computer memory chips in Formosa, which it ships to computer manufacturers in Japan (Market 1) and the United States (Market 2). Demand for the chips in the two markets is given by the follow- ing functions:

Japan: P1 = 12 − Q1 [14.5]

United States: P2 = 8 − Q2 [14.6]

where Q1 and Q2 are the respective quantities sold (in millions of units), and P1 and P2 are the respective prices (in dollars per unit) in the two markets. TIC’s total cost function (in millions of dollars) for these memory chips is

C = 5 + 2(Q1 + Q2) [14.7]

Case I: Differential Prices TIC’s total combined profit in the two markets equals

π = P1Q1 + P2Q2 − C [14.8]

= ð12 − Q1ÞQ1 + ð8 − Q2ÞQ2 − ½5 + 2ðQ1 + Q2Þ� = 12Q1 − Q21 + 8Q2 − Q

2 2 − 5 − 2Q1 − 2Q2

= 10Q1 − Q21 + 6Q2 − Q 2 2 − 5 [14.9]

(Continued)

508 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

To maximize profit with respect to Q1 and Q2, find the partial derivatives of Equa- tion 14.9 with respect to Q1 and Q2, set them equal to zero, and solve for Q*1 and Q*2:

∂π

∂Q1 = 10 − 2Q1 = 0

Q*1 = 5 ðmillionÞ units ∂π

∂Q2 = 6 − 2Q2 = 0

Q*2 = 3 ðmillionÞ units Substituting Q*1 and Q*2 into the appropriate demand and profit equations yields

P*1 = $7 per unit

P*2 = $5 per unit

π* = $29 ðmillionÞ This optimal solution is illustrated graphically in Figure 14.3, Panel (a).

Maximizing π with respect to Q1 and Q2 is equivalent to setting MR1 = MR2. The equivalence of MR1 and MR2 may be proved by taking the partial derivatives of the TR function with respect to Q1 and Q2:

TR = P1 · Q1 + P2 · Q2

= ð12 − Q1ÞQ1 + ð8 − Q2ÞQ2 = 12Q1 − Q21 + 8Q2 − Q

2 2 [14.10]

and substituting the solution values, Q*1 = 5 and Q*2 = 3:

MR1 = ∂TR ∂Q1

= 12 − 2Q1

MR*1 = 12 − 2ð5Þ = $2 per unit

MR2 = ∂TR ∂Q2

= 8 − 2Q2

MR*2 = 8 − 2ð3Þ = $2 per unit which equals the total marginal cost, that is, the derivative of Equation 14.7 with respect to (Q1 + Q2).

The respective elasticities in the Japanese and U.S. markets at the optimal solu- tion are

E1 = dQ1 dP1

· P1 Q1

= −1 7 5

� � = −1:40

(Continued)

Chapter 14: Pricing Techniques and Analysis 509

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIGURE 14.3 Demand and Cost Functions for Memory Chips: Taiwan Instrument Company

P*

Q* Q1

0

MR1

D1 MC

DT

QT = Q1 + Q2Q* Q2

0 0

MR2

D2

P*

Q* Q1 QT = Q1 + Q2Q* Q2

p*

MR1

D1 MC

DT

MR2

D2

MRT

Q*

MRT

Q*

Pr ic

e an

d co

st (

$/ un

it )

Pr ic

e an

d co

st (

$/ un

it )

1

1

2

2 T

1 2 T

15 14 13 12 11 10

9 8 7 6 5 4 3 2 1

15 14 13 12 11 10

9 8 7 6 5 4 3 2 1

(a) Case I: Differential pricing

(b) Case II: Uniform pricing

Japan United States Total

Japan United States Total

1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 101112131415

0 0 01 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 101112131415

and

E2 = dQ2 dP2

· P2 Q2

= −1 5 3

� � = −1:67

Hence we see that, as in the JetBlue Airways example, when the elasticity of demand is less (in absolute value) in Japan (Market 1) than in the United States (Market 2), the price in Japan is greater than in the United States.

(Continued)

510 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Case II: Uniform Prices Now suppose that TIC is not permitted to engage in differential pricing.

To determine the profits TIC will earn if it does not discriminate between the two markets, solve the two demand equations for Q1 and Q2 and add them to get a total demand function:

Q1 = 12 − P1

Q2 = 8 − P2

QT = Q1 + Q2

= 12 − P1 + 8 − P2

Because price discrimination is no longer possible, P1 must equal P2, and

QT = 20 − 2P

or

P = 10 − QT 2

Total profit is now

π = PQT − C

= 10QT − Q2T 2

− 5 − 2QT

= 8QT − Q2T 2

− 5 [14.11]

To find the profit-maximizing level of QT, differentiate Equation 14.11 with respect to QT, set it equal to zero, and solve for Q*T :

dπ dQT

= 8 − QT = 0

Q*T = 8 ðmillionÞ units Substituting Q*T into the appropriate equations yields

P* = 10 − QT 2

= $6 per unit

π* = 8QT − Q2T 2

− 5 = $27 ðmillionÞ Q*1 = 12 − 6 = 6 ðmillionÞ units Q*2 = 8 − 6 = 2 ðmillionÞ units

MR*1 = 12 − 2ð6Þ = $0 per unit MR*2 = 8 − 2ð2Þ = $4 per unit

This uniform price solution is illustrated graphically in Figure 14.3, Panel (b). As summarized in Table 14.3, note that TIC’s profits are higher when it engages in differential pricing ($29 million) than when it does not ($27 million).

Chapter 14: Pricing Techniques and Analysis 511

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

DIFFERENTIAL PRICING IN TARGET MARKET SEGMENTS After identifying the different value drivers for various segments of the target market and setting an optimal differential price, firms must prevent resale between the segments using a variety of “fences.” Two of the most frequent methods of direct segmentation that prevent resale involve intertemporal pricing (1) by time-of-day or day-of-week and (2) differential pricing by delivery location.

Congestion-based pricing at peak-demand periods on roadways, bridges, and subway systems is an example of intertemporal pricing, illustrated in Figure 14.4. Peak-period

TABLE 14.3 TAIWAN INSTRUMENT COMPANY: EFFECTS OF PRICE

DISCRIMINATION

CASE I DIFFERENTIAL

PRICING CASE II

UNIFORM PRICING

Market 1 (Japan) 2 (U.S.) 1 (Japan) 2 (U.S.)

Price P* ($/unit) 7 5 6 6

Quantity Q* (million units) 5 3 6 2

Marginal Revenue MR* ($/unit) 2 2 0 4

Profit π* ($ million) 29 27

FIGURE 14.4 Congestion Tolls with Peak and Off-Peak Demand: Dulles Toll Road

0 QOP

POP = MCOP

PP

P�

DOFF-PEAK(OP)

QPEAK(POP)QP

MCP

MCOP

MCOP + KCAPACITY COST

QC Traffic volume (cars/mile)

Pr ic

e ($

/c ar

)

$3

($1)

$2

$7 P

DPEAK(P)

Congestion toll (PP – MCOP)

Notes: P refers to peak period; OP refers to off-peak period.

512 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

drivers place demands on the Dulles toll road between 6:00 A.M. and 9:00 A.M. far in ex- cess of its carrying capacity (QC). Charging commuters a toll equal to just the wear and tear imposed by their vehicle passing over the toll road pavement (i.e., an off-peak marginal cost, MCOP) induces many more cars to enter the highway (QPEAK) than can be accommodated (i.e., QP > QC). The result is slowdowns, stoppages, and a markedly increased travel time for each commuter. Beyond QC, the traffic volume at which this congestion begins, MCP rises steeply, representing the incremental fuel and time costs imposed (by one additional car) on all the other drivers along a 10-mile stretch of congested toll road.

The advantage of a congestion toll such as (PP – MCOP) = $2 is that it induces discre- tionary peak-period travelers to switch to other travel times and alternative modes of transportation. If a toll road authority set peak-period prices of $3 just sufficient to cover this congestion cost plus the off-peak MC, traffic volume would decline from QPEAK (POP) to QP, and the equilibrium differential prices PP and POP would emerge. Such congestion pricing reflects the true resource cost of the scarce transportation system capacity at peak travel times.

Like peak–off-peak roadway pricing, many other examples of differential pricing en- tail charging differential prices for the same capacity at different times. Hence, such cus- tomers are not in rivalry for the same capacity. Parking meters in San Francisco can now raise price between 10:00 A.M. and 2:00 P.M. Coca-Cola has new cold drink machines that vary the price by time of day, as well as by the predicted high temperature for the day. The demanders of matinee ($5) and evening movie theaters ($9) are not in rivalry for the same theater seats. First-run movies and subsequent movie videos, hardback and later paperback editions of books, seasonal discounts in the resort and cruise ship busi- nesses, and weekend discounts in hotels all represent effective segmentation of different target customer classes by time of purchase.

Direct Segmentation with “Fences” Direct segmentation of target customer classes not in rivalry with one another for the same capacity can also be accomplished by selling various versions of a product custom- ized for target segments, or by varying the price by delivery location. Customers who arrive at the suburban neighborhood rental counters of Hertz and Avis have flexibly timed, convenience-based uses for rental cars. Consequently, demand is much more price sensitive than the demand at the airport by business travelers. A recent study found that weekday rates for a midsize sedan were $43 in neighborhood rental locations versus $69 on average in airport rental locations.7 Because round-trip taxi fares from airports to the neighborhood locations would typically far exceed the $26 price difference, Avis and Hertz customers are effectively segmented by rental location.

Another example of location-based segmentation would be fashion clothing from France’s Arche or Ralph Lauren sold less expensively in discount outlets along interstate highways than in suburban storefronts or vacation resorts. Outlet mall shoppers almost never overlap with the customers these companies find in their trendy boutiques. Hence, geographic segmentation works. Outlet shoppers will also buy a less costly, less durable version of the product (e.g., a lighter-weight chemise cloth in Polo golf shirts), so in dif- ferential pricing, Ralph Lauren accomplishes more than just inventory clearance without any danger of cannibalizing full-price sales. So, total sales expands to address this new segment created by the new location.

congestion pricing A fee that reflects the true marginal cost imposed by demand in excess of capacity.

7“Playing the Car-Rental Game,” Wall Street Journal (July 31, 2002), p. D3; and “Highway Nirvana at a Price,” Wall Street Journal (July 6, 2004), p. A15.

Chapter 14: Pricing Techniques and Analysis 513

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Hal Varian and Carl Shapiro argued that such “versioning” is an especially good way to sell information economy items such as software.9 A voice recognition package sells for $79 as general-purpose Voice ProPad, for $795 as Office Talk, and for $7,995 as Voice Ortho, a special-purpose medical transcriber for surgical theatres. All three

Example Congestion Charges on Los Angeles, San Diego, Houston, Minneapolis, Denver, and Dulles Toll Roads8

The typical American spent over 40 hours per year in traffic jams in 2007 due to peak-period road congestion. Los Angeles drivers spent 93 hours, equivalent to 2.5 weeks of unpaid work. Urban traffic grew by 20 percent during 1995–2005 but ur- ban roadways increased by only 2 percent. The federal excise tax on gasoline of 18.4 cents per gallon (which is earmarked to repair and build federal highways) has not increased since the early 1990s. But is the answer simply more highways? Communities in Southern California, New Jersey, Houston, and Washington, D.C., think not.

Congestion tolls that charge commuters for the incremental time loss and fuel increase their cars impose on other drivers during peak periods have been adopted on several public and private toll roads around the United States. Every day, 24,000 drivers pay a peak-period congestion toll of $3.30 per trip for a 10-mile stretch of true expressway in Orange County, California. Toll booths, which themselves cause delays, have been replaced by credit-card-size transponders mounted in dashboards from which overhead wireless receivers assess tolls as the cars speed by. Such on- board-units (OBUs) will be commonplace on all cars and trucks by 2010, and they may become the communications hub of the vehicle, offering information on road conditions ahead, directions to inexpensive gas stations, and emergency dispatch services based on the built-in global positioning satellite (GPS) device.

In Figure 14.4 the cost of additional vehicles on increased peak-period road con- gestion raises the peak period toll to $3 from an off-peak toll of $1. High occu- pancy vehicle (HOV) lanes, normally reserved for buses and carpools, can be accessed for as little as $0.25 off peak but cost $8 per trip during peak traffic vol- ume. Adopting real-time dynamic pricing with traffic volume measured every six minutes, and opening HOV lanes when they are underutilized, San Diego has in- creased its total road capacity by 64 percent. Electronic debits for driving on inner- city streets in London costs £8/day (about $12) between 7:00 A.M. and 6:00 P.M. The congestion toll expenses of a typical commuter mount up quickly. But the effect of the tolls has been noticeable; traffic in London’s designated congestion zone is down 30 percent. And time is money, so faced with the slow 25–40 m.p.h. com- mute on adjacent “freeways,” many peak-period travelers in Los Angeles, San Diego, Houston, Minneapolis, Denver, and Washington, D.C., are opting for the differential pricing of a 65 m.p.h. toll road.

8Based on “How to Make Traffic Jams a Thing of the Past,” Fortune (March 31, 1997), p. 34; “A Survey of Commut- ing,” The Economist (September 5, 1998), p. 62; Economist Technology Quarterly (June 12, 2004), pp. 30–32; “Steep Increases Set for Toll Roads,” Wall Street Journal (June 21, 2007), p. D1; and “Transportation Infrastructure,” Chapter 6, Economic Report of the President (Washington, DC: U.S. Government Printing Office, 2007).

9C. Shapiro and H. Varian, “Versioning,” Harvard Business Review (November/December 1998), pp. 106–114.

514 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

versions derive from the same source code, but the more comprehensive version gener- ates 100 times as much value to particular target customers. In contrast, when Amazon sells the same book or DVD at different prices to customers with different clickstreams, that degree of differential pricing for identical versions of the product often leads to ad- verse customer reactions. Coca-Cola is finding the same resistance to its differential time- of-day pricing in soft drink machines. As a result, many sellers adopt the techniques of indirect segmentation using two-part tariffs, couponing, and bundling. With indirect seg- mentation, the customer herself selects what differential price to pay from a variety of available alternatives.

Optimal Two-Part Tariffs Two-part tariffs entail charging both a lump-sum entry fee for access to the facility or service and a per-unit user fee for each unit sale consumed. Amusement parks, night- clubs, golf and tennis clubs, copier leasing companies, cellular phone providers, Internet access providers, and rental car companies often employ such pricing. Their revenue per unit sale is a nonlinear function of two parts: a lump-sum monthly or daily fee that pro- vides access to the facility, phone, computer, or rental car independent of use, and a per- hour or per-minute or per-mile fee that varies with usage. The magnitude per unit of user fees should at least cover marginal costs so that heavy demanders “pay the freight” through higher total user fees. Tying the price for a leased copier to a metering counter that effectively measures intensity of use results in a differential monthly leasing fee across customer segments plus a cheap incremental cost per copy.

Companies differ on whether to set high or low entry fees and whether to charge high or low user fees. AT&T Wireless and Gillette practically give away their cell phones and razors but then charge steep prices for the calls and blades. In contrast, iPods are pricey at the front end but iTunes thereafter are quite cheap. Similarly, most golf and tennis clubs charge substantial membership fees and annual dues, but thereafter adopt trivial user fees (e.g., $5 per court hour or $25 for greens fees).11 As we will see, just how

Example Dynamic Pricing for Electricity10

In principle, as experience has shown in Britain, Australia, and New Zealand, de- regulation of electricity can work well if peak-load customers are asked to pay a price that reflects marginal cost. As much as 55 percent of the variation in intraday cost is attributable to extraordinary transmission line fees and old inefficient plants fired up to meet the last 5 percent of peak demand. For example, running a clothes dryer at 6:00 P.M. in August imposes wholesale costs on Pacific Gas and Electric of $220 per megawatt hour (MWh), relative to approximately $100 at 10:00 P.M. in August, and $20/MWh in April. Smart appliances can now turn off the dryer when California electricity on time-of-day meters rises in price during the 4:00 to 7:00 P.M. peak period.

10Based on “Making Meters Smarter,” BusinessWeek On-Line (October 5, 2009); “How to Do Deregulation Right,” BusinessWeek (March 26, 2001), p. 112; “PG&E Gropes for a Way Out,” Wall Street Journal (January 4, 2001), p. A1; and A. Faruqui and K. Eakin, eds., Pricing in Competitive Electricity Markets (New York: Kluwer, 2000).

11When the Pebble Beach golf course and the Wimbledon tennis club have $350 greens fees for a round of golf or two sets of tennis, those user fees reflect congestion-based pricing rather than optimal two-part tariffs.

Chapter 14: Pricing Techniques and Analysis 515

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

much above marginal cost to set the optimal user fee depends on how dissimilar are the segments of customer demand.

Let’s investigate how to analyze an optimal two-part tariff. Consider the situation de- picted in Figure 14.5 for separate customer segments with relatively elastic (D1) and rel- atively inelastic demand (D2) for rental autos. These might be young couples who are renting cars for vacationing (D1) and manufacturers’ trade representatives renting cars for making sales calls (D2). The challenge is to find a uniform daily rate (the lump-sum access fee) and a mileage charge (the user charge) that maximize profit and keep both customer segments in the market. One alternative would be to price the mileage at its marginal cost (MC) = height OA and elicit Q1 and Q2 usage from each segment while realizing from both the maximum daily rate that the price-sensitive vacationer along D1 will pay (namely, hatched area AEF).

Perhaps, however, a better alternative is available. Suppose the car rental agency raises the mileage charge to P* and reduces the daily access fee to the hatched and shaded area P*DF. Mileage will decline in both segments (to Q01 and Q

0 2, respectively), and area

P*DEA will be net revenue lost by virtue of the reduced daily access fee in both seg- ments. However, the additional net revenue from mileage charges (P*DGA in one seg- ment and P*HIA in the other segment) will more than offset the lost access fees. In particular, profits of the car rental agency will increase by the difference of area DHIE—area DEG. This result is generalizable to other optimal two-part tariff decisions.

Consequently, in addition to charging a positive lump-sum access fee, a price- discriminating monopolist will adopt two-part tariffs that price usage above its marginal cost. The more similar the price elasticity of demands of the target customer segments,

FIGURE 14.5 Optimal Two-Part Tariffs for Auto Rentals

F

C

D

G

E B

H

I MC

P*

A

0 Q� Q� Q2

D2D1

Q1 Consumption (miles)

Pr ic

e ($

)

1 2

Vacationer Trade rep

516 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the closer the user charge should be to marginal cost. The more dissimilar the segment demands, the higher the user charge should be raised above MC.

Couponing Another pricing mechanism for indirectly segmenting the market and allowing the cus- tomer to select what level of consumption and total price to pay is couponing. The $49 billion spent on direct-mail marketing with accompanying rebate and coupon offers sur- passed the amount spent on newspaper ($45 billion) and television ($43 billion) adver- tising for the first time in the United States in 2003. This direct marketing approach is made possible by successful forecasting based on consumer spending patterns. Compa- nies have access to cash register and credit card data as well as property tax and public utility usage records. These sources allow Lowe’s Home Improvement, for example, to project with more than 80 percent accuracy the month in which a particular household will buy a gas grill.

Such laser-like precision in targeting encourages differential pricing. If coupons worth 25 cents off the price of a box of cereal, 40 percent off the price of fashion clothing, or $50 rebates off the price of an expensive gas grill are redeemed religiously by some

WHAT WENT RIGHT • WHAT WENT WRONG

Two-Part Pricing at Disney World The original studies of optimal two-part tariffs were commissioned for the Disney theme parks in California and Florida. Disney World in Orlando opened with an optimal entrance fee plus a user charge per ride. Specifi- cally, customers purchased booklets of tickets for the rides once they were already inside the theme park, having ear- lier paid uniform lump-sum fees at the entrance gates. For several years this system worked quite well as first-time visitors encountered all the Disney-theme displays and ve- nues and dispersed across the grounds taking the occa- sional ride. User charges per ride were sufficient to discourage persistent riding and re-riding of favorites like Space Mountain.

Before long, however, repeat-visitor demand became the key to profitable operation of the theme parks. The

target customer was a couple with two children who might be on their third or fourth Disney vacation. As certain attractions and rides became the magnet that drew families back, long lines developed at the popular rides and shows. Surveys showed that customers began to feel that the tick- ets they purchased were not valid for rides and that entry fees to enter the park were too steep in light of the disap- pointing congestion inside. These negative perceptions of two-part tariffs led Disney to replace them with a menu of differential entry fees based on projected usage of the Magic Kingdom, Epcot, and Disney Studios.

More recently, Universal Studios theme parks have ex- perimented with reducing congestion through time- stamped passes to popular rides and with preferential treatment for customers who have paid for shorter wait times.

WHAT WENT RIGHT • WHAT WENT WRONG

Price-Sensitive Customers Redeem Pillsbury measured the price elasticity of demand for its cake mix customers who redeem coupons and found it to be −0.43, whereas that of the nonredeemers was −0.21. Similarly, Purina measured the price elasticity of coupon redeemers among its cat food customers as −1.13 and

among nonredeemers as −0.49. Coupons for frozen potato dishes at Ore-Ida are redeemed by households with price elasticity of −1.33, and nonredeeming households have price elasticity of −1.97. Clearly, in all these cases, coupon- ing is a way to segment the market and offer a discount to the more price-sensitive segment.

Chapter 14: Pricing Techniques and Analysis 517

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

segments but ignored by others, Kellogg, Neiman Marcus, and Lowe’s will segment the market with these direct mail promotions. The price-sensitive customers who con- stantly redeem coupons and file for rebates receive a lower net price, consistent with Equation 14.4.

Bundling Bundling is another highly effective pricing mechanism that sellers use to capture profit from differential pricing across target customer segments. Have you ever wondered why Time Warner Cable offers Showtime, the channel for popular first-run movies, only in a bundled package that includes the History channel? One insight is that this paired bun- dle of product offerings occurs because some other Time Warner customer is a history buff who is wondering why the History channel comes with access to largely unwatched movies. That is, the operating profit to a seller from bundling negatively correlated de- mands is always larger than the operating profit from selling equally costly products sep- arately. Let’s see why.

Suppose that two sets of customers have the following reservation prices for two ca- ble channels, each of which incurs variable licensing fees of $1 for a single showing to a single household. Movie buffs would pay $9 for access to first-run movies and $2 for access to historical documentaries. History buffs would pay $8 for access to the History channel and $3 for access to Showtime. If the channels are priced uniformly to both cus- tomer segments as separate products, Time Warner can realize at most $8 (or $9 – $1) on Showtime and $7 (or $8 – $1) on the History channel for a total of $15 operating profit.12 However, note that both types of customers would pay up to $11 for the com- bined pair of channels rather than do without. If Time Warner made them available only as a bundled package, sales revenue would be $22 minus $4 licensing fees for a total of $18 operating profit, which is greater than the $15 we previously calculated.

As long as one customer is willing to pay more for product A that another customer wants less than product B, the seller who is restricted to charging the two customers the same uniform price will always be better off bundling the two items, assuming all reser- vation prices exceed variable cost. Such inversely correlated demands occur in many set- tings. All-inclusive Caribbean resorts such as $595 per day Bitter End on Virgin Gorda or Hotel Isle de France in St. Bart’s have guests who value $75 gourmet lunches and $350 per night fabulous cabanas but would not pay much for all the water sports activi- ties and equipment, while other guests value $170 per day water sports activities but would not pay such high prices for better food or cabanas. Similarly, Elizabeth Arden’s $225 all-inclusive half-day spas have clients who would not pay high-margin $50 prices for at least one of the following five services: facials, mud wraps, massage, manicures, and pedicures. Bundling all these services together always raises profitability when target customer demand is inversely correlated across the offerings as long as variable costs of the separate components are below the customers’ willingness to pay for each.

Now suppose the variable costs are higher at, say, $3. The History channel valued at $2 by the movie buff is no longer a profitable sale. Pure bundling includes this unprofitable sale and generates the same $22 revenue but now incurs $12 of total

reservation prices The maximum price a customer will pay to reserve a product or service unto their own use.

12In this example, selling both products separately to both customer segments does not pay because of the much lower prices required. Specifically, Showtime priced separately into both segments would have to sell for as little as $3, thereby earning operating profits of $6 − $2 = $4, and the History channel would have to sell for as little as $2, earning $4 − $2 = $2. Thus, the total profit of $6 from selling all products to all custo- mers at a uniform price would substantially diminish the potential profit of $15 from selling each product to its target market alone. If the asymmetric demands in the two segments were not so different, this result could reverse, as long as all reservation prices were greater than variable cost.

518 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

variable cost, yielding a profit of only $10. Forgoing the sale of the History channel to the movie buff by selling each product separately at a $9 price for Showtime and an $8 price for the History channel generates $6 (or $9 – $3) on Showtime and $5 (or $8 – $3) on the History channel, yielding a total of $11 operating profit. Quite intuitively, pure bundling will be less attractive than pricing separately when some of the bundled sales are unprofitable.

It is also easy to see why positively correlated demand across customers works against bundling. Figure 14.6 displays reservation prices along a “budget” line that our customers in the earlier example are willing to spend on the two products.13 The y-intercept is the total willingness-to-pay constraint for the two products—namely, Ph + Pm = $11. With Showtime reservation prices on the vertical axis and History channel reservation prices on the horizontal axis, each customer’s mix of reservation prices lies along the line

Pm = $11 − 1Ph [14.12]

The −1 in Equation 14.12 signifies the perfect negative correlation between the reserva- tion prices (demand) of our movie buff and those of our history buff. But suppose Time Warner has a third type of customer whose reservation prices are positively correlated

FIGURE 14.6 Reservation Prices for Three Customer Segments

1 ($8, $9)

2

3

$2

$2 $4 $6 $8 Ph

Pm

$10 $11

History ($/showing)

$4

$6

$8

$10

$11 IV

III

I

II

Sh ow

ti m

e ($

/s ho

w in

g)

13This budget line is analogous to the budget line of a household making consumption decisions, except in this case it is the firm that is constrained by the maximum expenditure the customer is willing to make on the two goods.

Chapter 14: Pricing Techniques and Analysis 519

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

with those of the movie buffs—that is, a third type of customer who values Showtime at $8 and the History channel at $5. These reservation prices are high when the movie buff’s reservation price is high and low when the movie buff’s reservation price is low. Such positively correlated demand lies above the budget constraint in Figure 14.6 be- cause the total willingness to pay on the left-hand side of Equation 14.12 is no longer $11 but rather is now $13, as shown for Point 3.

With positively correlated demands across two of the three customer types, Time Warner could sell the Showtime-History bundle to all three for $11 and earn $15 [3 × ($11 − $6)].14 However, a better alternative is available. Mixed bundling sells the products both separately and as a bundle with the bundled price discounted below the sum of the two separate prices. In our three-customer-type example, Time Warner could sell Showtime for $9 and the History channel for $8, while making the Showtime-History bundle available for the package price of $13. The third type of customer would opt for the bundle, whereas each of the other types of customers would buy one product only. Revenue for this mixed bundling approach totals $30, but only four license fees are required, therefore earning $18 in profit. In general, pure bundling generates less profit than mixed bundling when positively correlated demands are involved. That’s why Arden’s salons sell their beauty treatments bundled for $225 or $50 each.

Figure 14.6 can be used to characterize the attractiveness of pure bundling for the seller. If all customers have perfectly negatively correlated demands, their reservation prices lie, as we have seen, along the $11 budget constraint. If customers have positively correlated demands, their reservation prices will lie consistently either above or below this reservation budget constraint. With separate product prices of Pm = $9 and Ph = $8, customers with reservation prices in Quadrant I will always buy both products rather than one of the separate products alone (Quadrants II and IV) while those in Quadrant III will never buy either product sold separately. In addition, however, we know that customers with reservation prices above the reservation budget constraint will buy the bundled package and those below will not. Optimally, Customer 3 will therefore purchase the bundle, Customer 1 will purchase Showtime alone, and Customer 2 will purchase the History channel alone. Only mixed bundling can achieve this result.

Example McDonald’s Introduces Mixed Bundles as “Extra Value Meals” In the United States, fast food consumption skyrocketed in the last two decades as 70 percent of households became two-worker households, and families began eat- ing out several times a week. Beer and pizza or a soft drink, burger, and fries be- came the standard supper for many high time-cost households. With greater health consciousness, however, not everyone who wanted a burger wanted the fries. In other cases, some consumers wanted the fries but not the burger, and preferred lower fat chicken sandwiches instead. McDonald’s Corporation has done its best to respond to these customer dissimilarities by introducing “Extra Value Meals” that bundled chicken sandwiches with fries and a medium soft drink.

(Continued)

14Here we are again assuming that variable costs are at the higher level of $3 per showing.

mixed bundling Selling multiple products both separately and together for less than the sum of the separate prices.

520 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

To summarize, two-part tariffs, couponing, and bundling are pricing mechanisms that induce customers to segment themselves indirectly. Two-part tariffs are particularly ef- fective in capturing higher profits than uniform prices when customer segments are more nearly identical in their price elasticity of demand. Couponing works best when target segments are extraordinarily different in their price elasticity of demand. Bundling captures additional profit when segmented target customer demand is inversely corre- lated across multiple products.

Price Discrimination Price discrimination is defined as selling the same product or service out of the same distribution channel at different prices to different buyers during the same period of time. Examples of price discrimination include the following:

• Doctors, dentists, hospitals, lawyers, and tax preparers who charge clients who reside in wealthy zip codes more for the same service than those who reside in poorer zip codes

• Dell Inc.’s ultralight laptop, which it sells for $2,307 to small business customers, for $2,228 to health care companies, and for $2,072 to state and local governments

• Firms that sell the exact same product under two different labels at widely varying prices (Hotpoint and Kenmore appliances, Michelin and Sears Roadhandler radial tires)

• Athletic teams that sponsor family nights and ladies’ nights at discount prices, while other customers pay the full price

• Hotels, restaurants, and other businesses that offer discounts to senior citizens • Differentially priced seats in coach class on a particular flight based on how far

ahead you book the reservation or Saturday night stay overs • Korean TV manufacturers who sell products direct to the customer at a lower price

in the United States than in Japan

The 2006 prices for some of McDonald’s most popular menu items are listed here:

Examining only the last two columns, we see that some bundled “Extra Value Meal” customers (McChicken Sandwich Meal and Double Cheeseburger Meal) are getting very little discount.

MENU ITEM SEPARATE PRICE BUNDLED

PRICE

TOTAL IF PURCHASED SEPARATELY

Large French Fries $1.39

Medium Soft Drink $1.09

McGrill Chicken Sandwich $2.69 $4.29 $5.17

Chicken McNuggets $2.79 $4.29 $5.27

McChicken Sandwich $1.00 $3.39 $3.48

Big and Tasty Burger $1.59 $3.49 $4.07

Double Cheeseburger $1.00 $3.39 $3.48

Big Mac $2.19 $3.79 $4.67

Quarter Pounder $2.19 $3.79 $4.67

price discrimination The act of selling the same good or service, distributed in a single channel, at different prices to different buyers during the same period of time.

Chapter 14: Pricing Techniques and Analysis 521

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Most differential pricing to a firm’s retail customers is perfectly legal.15 It raises profits because it transfers some of the excess customer value (the satisfaction gained from the purchase of the product) from buyer to seller relative to the excess value generated for customers who pay a lower uniform price.

If a coffee aficionado were willing to pay $4.00 for a large cup of fresh brewed java for which Dunkin’ Donuts charges $1.95, and $5.00 for the same cup of fresh brewed java bundled (on request) with a shot of espresso for which Starbucks charges $3.50, the con- sumer’s excess value would decline from $2.05 at Dunkin’ Donuts to $1.50 at Starbucks. Nevertheless, in the case where this customer declines the shot of espresso, Starbucks must have offered something else of value because the customer’s willingness to pay rose from $4.00 to $5.00. If this customer kept returning to Starbucks, we should assume that the lifestyle and group identity available at Starbucks attracted his or her business.

When Nationwide and GMAC auto insurance lower rates based on the reduced theft and collision risk exposure of the places you drive, that is not price discrimination. A GPS tracking device in the car confirms that the loss protection service is different.

Example eBusiness Clickstreams Allow Price Discrimination: Personify and Virtual Vineyards16

Personify, an Internet service company, created software that allows Web busi- nesses to categorize buyers based on their clickstream patterns. Using this software, Virtual Vineyards, CDNow, and Amazon.com have experimented with charging different prices for the same wine, CD, or book based on the clickstream path. Let’s see how this capability might work for Virtual Vineyards.

Incremental variable plus direct fixed costs in making table wines run at least $8 for the following inputs: $0.50 for the bottle itself, $0.30 for the cork, $0.20 for the label, $2 per bottle for the quickly decaying barrel used for aging, plus anywhere from $5 to $50 for the grapes. If it costs $220 to produce and stock a case of 20 bottles, and if five customers will pay $24 a bottle while another 15 customers will pay $10 a bottle, uniform pricing across all 20 customers will result in a $20 loss [($10 × 20 = $200) − $220 = −$20]. At a $10 uniform price, all 20 customers will buy but the Virtual Vineyards will stop producing this wine. Similarly, at a $24 price, only five customers will buy, and the Virtual Vineyards again discontinues this wine because of now even more substantial losses [($24 × 5 = $120) − $200 = −$100].

But suppose five customers are asked to pay $20 per bottle, while 15 more are asked to pay $9. Each group pays less than their willingness to pay and less than their proposed uniform prices of $24 and $10, yet the winery now stands to make a profit: (5 × $20) + (15 × $9) = $235 − $220 = $15. Virtual Vineyards could perhaps segment this market and prevent resale by charging $20 in the retail distribution channel and $9 for limited quanti- ties at the winery. Alternatively, they could price discriminate by the clickstream of the customers visiting their Web site. New customers who first clicked through to the history of the awards won by the winery would be asked to pay $20 for the new release, whereas returning customers who renew their membership online in the frequent buyer program and whose last purchase order was a case would be asked to pay $9 for the new release.

16Based on “I Got It Cheaper Than You,” Forbes (November 2, 1998), pp. 83–84; and “The Art and Science of Pricing Wine,” CNet (July 3, 2003).

15The Robinson-Patman Act prohibits price discrimination in wholesale business-to-business transactions where the product is going to be resold but allows whatever the market will bear in retail transactions not ac- companied by duress, misrepresentation, or outright fraud.

522 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

In the limiting case of perfect price discrimination (PPD), sometimes called first degree price discrimination, the seller discovers the maximum price each individual is willing to pay for each unit purchased. A PPD monopolist then charges each pur- chaser this maximum reservation price in order to capture the consumer’s entire perceived excess value above the cost-covering price. However, because the information required for such pricing is so extensive, perfect price discrimination almost never occurs. Instead, as we have seen in studying two-part tariffs, couponing, and bundling, firms often price discriminate by allowing customers within indirectly segmented groups to determine their own price through intensity of use, redemption behavior, or selection of packages of products such as Disney World entrance fees (so-called second degree price discrimination). Finally, firms may attempt to price discriminate through directly segmenting classes of customers by time or location of purchase and then charging one uniform price within each customer class (so-called third degree price discrimination).

PRICING IN PRACTICE To this point, the chapter discussed firms that seek to maximize short-run profits. How- ever, pricing is an area where a longer-run life cycle view of the firm’s decision making is helpful.

Product Life Cycle Framework17

In the early stages of life cycle pricing, the marketing, operations, and financial man- agers decide what the customer will value, how the firm can manage the supply chain to consistently deliver those characteristics, and how much it will cost, including the financing costs. If the value-based prices can cover this long-run full cost, the product becomes a prototype. Each proposed product or service then proceeds to marketing re- search, where the demand at various price points in several distribution channels usually is explored. Marketing research will identify a target price that the cross-functional prod- uct manager or the general managers will know is required on average over the product life cycle in order for the new product to provide sufficient revenue to cover fully allocated cost.

Once a product or service rollout takes place (usually at target price levels), the mar- keting plan often authorizes promotional discounts. In this stage of the life cycle, the firm is interested in penetrating the market. To do so requires coupons, free samples, name recognition advertising, and slot-in allowances on retail shelves. Penetration pricing therefore characterizes this early stage of the product life cycle at which net prices to the manufacturer fall below the firm’s target price, as shown in Figure 14.7.

When a new product is introduced by a firm, pricing for that product is a difficult and critical decision, especially if the product is a durable good—one that has a relatively long useful life. The difficulty of pricing the new product comes from not knowing the level of demand with confidence. If the price is initially set too low, some potential cus- tomers will be able to buy the product at a price below what they are willing to pay. These lost profits will be gone forever. This problem is accentuated when the firm ini- tially has limited production capacity for the new product.

17For a discussion of the conceptual framework of value-based pricing over a product’s life cycle, see T. Nagle, J. Hogan, and J. Zale, The Strategy and Tactics of Pricing, 5th ed. (Upper Saddle River, NJ: Prentice Hall, 2011), Chapter 7.

life cycle pricing Pricing that varies throughout the product life cycle.

Chapter 14: Pricing Techniques and Analysis 523

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Under these circumstances, many firms adopt a strategy of price skimming, or pric- ing down along the demand curve. The initial price is set at a high level, even though the firm fully intends to make later price reductions. When the product is first intro- duced, a group of fashion-conscious or technology-conscious early adopters will pay the high price established by the firm. Once this source of demand is exhausted, the price is reduced to attract a new group of customers. Flat-screen TVs and handheld computers such as the Blackberry and Palm’s Treo are excellent examples of this phe- nomenon. As we discussed in Chapter 13, manufacturers who engage in price skim- ming on industrial equipment (e.g., mainframe computers and corporate jets) need credibility mechanisms to assure early full-price customers that later discounting will be limited.

In the mature stage of the product or service life cycle, organic growth comes from focusing on product differentiation and commitment to building the brand. Marketing team initiatives will add value in both product refinements and order management pro- cesses through brand-name advertising, product updates, or increased flexibility in ac- cepting change orders from regular customers. Each decision at this mature stage is motivated by a desire to realize the highest value-based pricing allowed by the competi- tive conditions and potential entry threats. Although at times this approach to pricing can be overwhelmed by the necessity of short-term tactics to defend market share, the product life cycle remains a planning framework to which the pricing manager often returns.

At a late mature stage of the product or service life cycle, product managers may de- cide to limit price, reducing it well below the value-based pricing level in order to deter entry. Limit pricing appears to be inconsistent with profit maximization but in fact is motivated by a long-term profitability objective.

Because competitors are constantly devising lower-cost ways of imitating leading products, limit pricing sometimes has only temporary success. If the entry threat materi- alizes into a real live new entrant, many incumbent firms then decide to accommodate by raising prices in a particular high-price, high-margin market niche. This pricing

FIGURE 14.7 The Price Life Cycle

Product life (months)

Pr ic

e, C

os t

Variable cost

Target price

Niche price

Limit price

Penetration price

Value-based price

price skimming A new-product pricing strategy that results in a high initial product price being reduced over time as demand at the higher price is satisfied.

524 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

practice is often referred to as niche pricing. Concluding that declining market share from entry into the mass market is inevitable, the incumbent moves upmarket and sells its experience and expertise at high prices in the top-end segments of the market, much as it did at the start of the product life cycle.

Full-Cost Pricing versus Incremental Contribution Analysis Some inadvisable pricing practices are widely adopted: two examples are full-cost pricing and target return-on-investment pricing. Full-cost pricing requires that not only direct

Example Loss of Patent Protection Limits the Price of Prozac: Eli Lilly When brand-name pharmaceuticals reach the end of their 20-year patent protec- tion, sales may plummet unless prices are radically reduced. Some formerly pat- ented drugs lose as much as 80 percent of their sales in the first year after generic substitutes are introduced. The ulcer relief medicine Zantac, which was at the time Glaxo’s biggest seller, plummeted 51 percent in the first half-year after loss of pat- ent protection. By year-end, 10 rival products were on the shelves. Zovinax, an anti-herpes medication, lost 39 percent in the first six months after generics costing only 20 percent of Zovinax’ price appeared in the marketplace. And sales of Bristol-Myers Squibb’s Capoten, at $0.57 per pill, declined 83 percent the year that a substitute generic pill at $0.03 was introduced.

In light of these disastrous experiences throughout the pharmaceutical industry, Eli Lilly limited the price of the depression treatment Prozac to variable plus direct fixed costs in order to arrest or at least slow the onslaught of imitators into its antidepressants market. We saw in Chapters 11 and 13 that such limit pricing strat- egies can deter entry and thereby raise the discounted present value of long-term cash flow to stockholders relative to maximizing short-run profitability.

Example Niche Pricing at Pfizer18

After Pfizer’s No. 1 LDL cholesterol-reducing drug Lipitor, which has almost $10 billion in sales and one more year of patent life, Novasc was Pfizer’s second flag- ship product, with $4.34 billion in sales. Unfortunately for Pfizer, the patent on this leading hypertension drug expired in 2007. Pfizer estimates that 3 million of the 30 million Americans who suffer from both hypertension and high cholesterol are be- ing treated for both diseases. These patients are candidates for a combined pill, Caduet (Novasc + Lipitor), that should preserve the pricing power on Novasc as it experiences an onslaught of generic competition. The profit potential from such combo pharmaceuticals is substantial relative to the 85 percent discounts that would otherwise be necessary to limit entry into Novasc’s stand-alone market. Niche pricing is always of limited applicability, and Pfizer must expect that com- petitors will work hard to chip away at the appeal of one high-priced pill relative to a generic pill plus high-priced Lipitor.

18“Drug Makers’ Combo,” Wall Street Journal (January 29, 2004), p. B1.

full-cost pricing A method of determining prices that cover overhead and other indirect fixed costs, as well as the variable and direct fixed costs.

Chapter 14: Pricing Techniques and Analysis 525

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

fixed costs of a particular product line such as licensing and maintenance and advertising be considered in pricing, but even indirect fixed costs of overhead and capital financing be added to variable costs to arrive at a final price. Indirect costs may be allocated among a firm’s several products in a number of ways. One typical method is to estimate total indirect fixed costs assuming the firm operates at a standard level of output, such as 70–80 percent of capacity, and then allocate the indirect costs by volume.

Target return-on-investment pricing begins by selecting an acceptable profit rate on investment, usually defined as earnings before interest and taxes (EBIT) divided by total gross operating assets. This return is then prorated over the number of units expected to be produced over the planning horizon. Advocates of full-cost and target return pricing argue that it is important to allocate all fixed costs among the various products produced by the firm and that each product should be forced to bear its fair share of the fixed-cost burden.

However, each product should instead be viewed in the light of its incremental con- tributions to covering the business plan’s fixed costs. Incremental contribution analysis provides a better basis for considering whether the manufacture and sale of a product should be expanded, maintained, or discontinued in favor of some higher-profit alterna- tive. Every firm should have an effective control system in which a general manager con- tinually monitors the overall contribution of the firm’s complete product line. This person can then ensure that value-based prices contribute to both the variable cost of each product and the total fixed costs of the firm. Such target pricing is especially rele- vant at the launch of a product line and later at the decision to exit (see Figure 14.7).

Example Full-Cost Pricing Results in the Loss of a Big Contract at J.P. Morgan: British Telephone Telecommunications is a fiercely competitive business. British Telephone (BT) once found that its $13 million bid to provide secure long-distance microwave business communications for the investment bank J.P. Morgan ended up $4 million higher per year than Sprint’s rival bid of $9 million. When BT executives did a follow-up study to see why they had been so undercut by Sprint, they discovered that the vice president of the BT subsidiary in the United States had attempted to recover the en- tire annual overhead for the subsidiary headquarters from this one account. Needless to say, BT lost J.P. Morgan’s business with a full-cost bid of $13 million when Sprint had offered to do essentially the same thing for $9 million. Full-cost pricing always runs the risk of such undercutting by rivals.

Example Incremental Contribution Analysis at Continental Airlines At one point Continental was filling only about 50 percent of its available seats, or about 15 percent less than the industry average. Eliminating 5 percent of its flights would have resulted in a substantial increase in this load factor but would have reduced profits as well. The airline industry is characterized by extremely high in

(Continued)

target return- on-investment pricing A method of pricing in which a target profit, defined as the (desired profit rate on investment × total gross operating assets) is allocated to each unit of output to arrive at a selling price.

incremental contribution analysis An incremental managerial decision that analyzes the change in operating profits (revenue – variable costs – direct fixed costs) available to cover indirect fixed costs.

526 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Pricing on the Internet19

E-business encounters several problems unique to Web-based transactions. First is the anonymity of buyers and sellers who often are identified by only a Web address. Offers to buy (and sell) may be reneged, receivables may never be collected, and items delivered may not be what buyers thought they had bought. The incidence of all these events is much greater in the virtual sales environment. As a result, offers are higher, and bids are lower. From another perspective, the bid-ask spread in an Internet transaction rises to cover the cost of fraud insurance.

A second problem that the Internet accentuates is the inability to confirm variable product quality with hands-on examination. Internet pricing of commodity products such as crude oil, sheet metal, and newsprint paper, shown in Table 14.4, often pursues a

direct fixed costs, which are incurred whether a plane flies or not: time deprecia- tion costs on the aircraft, interest charges, and the cost of continuous pilot training, the expense of ground crews, as well as headquarters staff overhead. Consequently, Continental has found it profitable to operate a flight as long as it covers variable plus direct fixed costs of the flight.

The analysis of whether to operate a flight proceeds as follows. First, manage- ment examines the majority of scheduled flights to be certain that depreciation, overhead, and insurance expenses are met for this basic schedule. Then the possi- bility of scheduling additional flights is considered, based on their impact on oper- ating profit. If revenues on a flight exceed actual variable costs plus direct fixed costs, the flight should be added. These relevant costs are determined by soliciting inputs from every operating department that specify exactly what extra expenses are incurred as a result of the additional flight’s operation. For instance, if a ground crew that can service the additional flight is already on duty, none of the costs of this service are included in actual operating costs. If, on the other hand, overtime must be paid to service this flight, then that direct fixed cost varies with the deci- sion to operate this flight and should be included among its costs.

Another example of such incremental contribution analysis is the case of a late- night Continental flight from Colorado Springs to Denver and an early morning return flight. Even though the flights often go without a passenger and little or no freight, the cost of operating them is less than an overnight hanger rental in Color- ado Springs. Hence, the flights are operated, not shutdown.

In performing this type of incremental contribution analysis, two important points must be stressed. First, someone in management must have coordinating authority to ensure that overall objectives are met before facing decisions based solely on incremental analysis. In the case of Continental, the vice president of flight planning assumed this task. Second, every reasonable attempt must be made to identify actual incremental costs and revenues associated with a particular decision. Once this information is determined, incremental analysis becomes a use- ful and powerful tool in considering a wide range of decision problems facing the firm.

19An excellent survey of pricing strategy for Internet products is provided in John de Figueiredo, “Finding Sus- tainable Profitability in Electronic Commerce,” Sloan Management Review (Summer 2000), pp. 41–52. See also “The Click Here Economy,” BusinessWeek (June 22, 1998), pp. 122–126.

Chapter 14: Pricing Techniques and Analysis 527

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

low-cost strategy. The availability of quick resale at predictable commodity prices reas- sures buyers and sellers, and here Internet pricing at tight bid-ask spreads proves quite efficient. However, as one moves to the right in Table 14.4, product quality becomes harder and harder to detect at the point of sale. Firms such as Amazon and CDNow seek to substitute brand equity for the inability of customers to examine the product. America Online (now AOL Inc.), Amazon, and Priceline spent tens of millions of dollars establishing their brand equity.

When it comes to toys, suits, homes, and new autos, consumers search for that look-and-feel for which they’re willing to pay. Brands again play an important role in certifying quality, but in this case it is product branding (e.g., Game Boy, Hart Schaff- ner Marx, Harris Tweed) that matters, not Web site brands. Customers rely on the hostage associated with the sunk cost investment in the product brand names to estab- lish credibility. Finally, with highly variable quality in tires, PCs, produce, and lumber, only strong warranties, escrow accounts, and replacement guarantees or deep discounts can replace the reputation effects that help sell these experience goods in nonvirtual settings.

Internet sellers can add value and reduce some transaction costs in these markets by customizing and selling direct to the customer like Dell, who provides order fulfillment and manufactures almost nothing. For this reason, services have grown quickly on the Net; the travel industry itself accounted for 35 percent of all online sales in 2002. Table 14.5 shows that the growth rate of services far surpassed growth in consumer products online.

In business-to-business (B2B) transactions, pricing is more complex than in business- to-consumer transactions. In B2B, multiple attributes come into play in the price negoti- ation. B2B customers haggle over date of shipment, delivery costs, warranty service times and locations, delivery reliability, and replacement guarantees. These additional consid- erations typically mean pricing is a part of a two- or three-step process. First, customers match their nonnegotiable requirements to the suppliers with those attributes, and those firms become the order-qualified suppliers. Then, the remaining attributes may be nego- tiated away against demands for a lower price point. In the heyday of the Internet bub- ble, B2B Internet sales grew twenty-fold from $8 billion in 1997 to $183 billion in 2002; see Table 14.5.

Internet pricing in these B2B settings requires a matching process to qualify for an order and then a dynamic pricing algorithm to trade off the remaining attributes. Infor- mation technology complexity in these B2B transactions arises because customers are heterogeneous, and the attributes that qualify a firm to supply one group of customers may not match the requirements of other customers. In addition, as we shall see in the

TABLE 14.4 PRICING STRATEGY FOR VARIOUS INTERNET PRODUCTS

COMMODITY PRODUCTS

QUASI-COMMODITY PRODUCTS

LOOK-AND-FEEL SEARCH GOODS

EXPERIENCE GOODS VARIABLE QUALITY

Crude oil

Newsprint

Sheet metal

Paper clips

Books

CDs

Videos

Suits

Homes

New autos

Toys

PCs

Produce

Tires

Lumber

Low-cost, low-price strategy

Differentiate with reliable delivery and extra services

Employ differential pricing based on brands and time of adoption in fashion cycle

Customize and build to order with low- and high-price tiers

dynamic pricing A price that varies over time based on the balance of demand and supply, often associated with Internet auctions.

528 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

next section, delivery reliability (i.e., the probability of stockout and back order) is a con- tinuous variable that should be optimized with a revenue management solution, not a simple on-again/off-again attribute to promise or refuse a potential customer in exchange for a somewhat larger or smaller markup.

THE PRACTICE OF REVENUE MANAGEMENT,20 ADVANCED MATERIAL Differential pricing is sometimes complicated by capacity choices that must be made before demand is known. Consider an airline, printing press operator, or elective surgery clinic, each of which must schedule capacity before the respective demands for the 11:00 A.M. flight, the press run next Thursday, or elective surgeries tomorrow are known. If no reve- nue can be realized after scheduled delivery from empty airline seats, from underutilized printing presses, or empty surgical theatres, random customer arrivals force a firm with fixed capacity to choose between underutilizing excess capacity or imposing service denials and stockouts on regular customers.

The spoilage from unsold capacity and the spill of high-margin repeat customers for whom no capacity remains are serious problems that may affect the firm’s financial suc- cess and indeed its survival. In Figure 14.8, a reduction in capacity from Qd1 to a level just sufficient to meet mean demand at P0 reduces the spoilage for low-demand events (Qd2 ) from AB to CD but introduces spill (i.e., Spill2) for high-demand events (Qd1 ). Rev- enue or yield management (YM) is an integrated set of managerial economics techniques designed to deal with these pricing and capacity allocation problems under fixed capacity and random demand.

TABLE 14.5 GROWTH IN ONLINE SALES

INTERNET BUBBLE YEARS COMPOUND ANNUAL GROWTH RATE1997 2001

Consumer Services

Travel $654 million $ 7.4 billion 83%

Event Tickets 79 million 2 billion 124%

Financial Services 1.2 billion 5 billion 43%

Consumer Products

Apparel $ 92 million $514 million 53%

Books/CDs 156 million 1.1 billion 63%

PCs 863 million 3.8 billion 45%

B2B $ 8 billion $ 183 billion 119%

Source: BusinessWeek (June 22, 1998), pp. 122–126, Forrester Research.

20F. Harris and P. Peacock provide a thorough overview of YM techniques and potential industry applications in “Hold My Place Please: Yield Management Improves Capacity Allocation Guesswork,” Marketing Manage- ment 4, no. 2 (Fall 1995), pp. 34–46.

stockouts Demand in excess of available capacity.

spoilage Perishable output that goes unsold.

spill Confirmed orders that cannot be filled.

Chapter 14: Pricing Techniques and Analysis 529

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Spill and Spoilage at Sport Obermeyer21

The selling season in fashion retailing is short (lasting no more than several months), and customer demand at the product-line level is fickle and hard to fore- cast. Consequently, buyers for retail merchants like Neiman Marcus, Blooming- dale’s, Saks Fifth Avenue, Rich’s, and Macy’s must place orders far in advance of actual sales without really knowing which fashion trends will sell well and which will sell poorly. Sport Obermeyer faces this problem with ski clothes. In a particu- lar winter ski season, Pandora ski parkas may become a fashion statement and quickly sell out. If Pandora parkas go on back order, creating frustrated buyers, the store will lose that customer’s goodwill and future sales. In addition, the lost retail contribution margin every time Sport Obermeyer “spills” one of these custo- mers is $15.

On the other hand, Pandora’s line of ski parkas may not “catch on” this season. Instead, they may end up as spoilage (i.e., a large inventory of unsold winter clothes). The merchant would then incur losses on the unsold merchandise and forgo the opportunity to sell another Champion sweatshirt that could have occu- pied the ski parka’s shelf space. Sport Obermeyer can use the tools of yield man- agement to balance these costs of spill and spoilage and thereby determine how many parkas to order and what shelf space to devote to parkas versus sweatshirts.

21Based on M. Fisher et al., “Making Supply Meet Demand in an Uncertain World,” Harvard Business Review (May/ June 1994), pp. 83–93.

FIGURE 14.8 Spill and Spoilage with Random Demand and Fixed Prices

P0Pr ic

e

Qd2 (P0) Qd1 (P0)Mean demand

Excess capacity

Spill 2

Spoilage 1

A Spoilage 2

C

B

ED

Output/unit sales

Mean Qd (P)

Qd1 high

Qd2 low

Customer demand distribution

0

Idle capacity

530 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

A Cross-Functional Systems Management Process Firms might respond to unanticipated demand fluctuations in the presence of capacity constraints by simply auctioning off their scarce product to the highest bidder or by holding massive clearance sales when faced with inventory overhang. Much to their det- riment, department store retailers take a myopic marketing view of what markdown prices can accomplish. Fashion-conscious customers shop numerous stores to “win” the trendy items and do so with little repeat purchase loyalty. Everyone else has become ha- bituated to wait for the inevitable and deep discount sales. The proportion of department store revenue earned through transactions at clearance sale prices rose from 8 percent in 1970 to 55 percent in 2001.22 Even regular customers of leading department stores report buying at discount prices almost as often (46 percent) as at regular prices (54 percent). Not surprisingly, profitability in the department store retailing sector has collapsed, and consolidation mergers have taken some of the best-known retailers out of business. What might these stores have done differently?

One alternative would be for the retail merchant’s suppliers to develop flexible manufacturing systems (FMSs) so they could respond to demand fluctuations more quickly. If reorder cycles could occur several additional times within the fashion season, merchants could stockpile less inventory and yet experience fewer stockouts.

Another alternative to resolve the problems of high-margin spill is simply to acquire more capacity. Of course, no company can afford to build unlimited additional capac- ity. Aggregate capacity planning incorporates a careful financial analysis of the capital budgeting problem that identifies the optimal fixed capacity for any line of business. A better alternative is to reserve some of this optimal fixed capacity for late-arriving, high-margin customers. Reserving capacity as the moment of delivery approaches should not be interpreted as “excess capacity” but rather as idle capacity that presents a sustainable revenue opportunity. This insight is one that comes out of revenue man- agement analysis.

Every company has some orders that it should refuse. Revenue management (RM), often called yield management, is fundamentally an order acceptance and refusal process that links marketing decisions about demand creation and pricing with operations deci- sions about scheduling and financial decisions about capacity planning. The objective is to decide which orders to accept at particular prices and which to refuse. These relation- ships are depicted in Figure 14.9 as a cross-functional triangle of account management, forecasting, and scheduling decisions. Practitioners of yield management believe that the sources of sustainable price premiums lie in these cross-functional systems management processes. In this view, innovative products and successful advertising campaigns are quickly reverse engineered and readily imitated. Advertising and product design cannot therefore provide sustainable competitive advantage. Process advantages, on the other hand, prove much more difficult for competitors to imitate.

Sources of Sustainable Price Premiums Yield management processes add conspicuous tangible value for which customers gladly pay higher prices. In most cases, the added value arises through customizing and optimiz- ing the account and order management. In the airline industry, for example, some custo- mers want extensive flexibility of reservations that allows frequent changes in departure and arrival times. If an airline has the operations capability and information technology to provide this service, business travelers with unconfirmed meeting schedules will offer

22B.P. Pashigian, “Demand Uncertainty and Sales,” American Economic Review 78, no. 5 (December 1993), pp. 936–953; and “Priced to Move,” Wall Street Journal (August 7, 2001), p. A6.

revenue management A cross-functional order acceptance and refusal process.

Chapter 14: Pricing Techniques and Analysis 531

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

large price premiums to secure this change order responsiveness. To take another example, Disney offers substantial price premiums to gift product suppliers who can deliver high quality on time as promised. When Disney order clerks submit an error-laden request that needs to be changed within the normal 30-day reorder cycle, Disney volunteers to pay even more. Such supplements to revenue go exclusively to firms that have the systems management processes that can handle extraordinary change order requests.

Firms compete on other aspects of order processing as well. Some customers want short scheduling delays (e.g., just-in-time retailers without warehouses). Others want high delivery reliability and a small probability of being denied service in the event of a stockout (e.g., business executives traveling to a stockholders’ meeting). Still others value conformance to product or service specifications. For time-sensitive deliveries of organ transplants, for example, excellent on-time service records warrant paying high airfares. The alternative would be a much more expensive jet charter service. Manufacturers as well as service firms can establish sustainable price premiums based on these same order-processing characteristics of change order responsiveness, minimal scheduling de- lay, delivery reliability, and conformance to specifications.

All these sources of sustainable price premiums are prominent in airline services at “fortress hubs” where one carrier controls more than 65 percent of the seat departures. Figure 14.10 displays the major hub airports in the continental United States and pro- vides pricing and market share data for the top two carriers. In each of these cities, the dominant firm(s) has sufficient operating capacity and systems control to provide high- quality service. At Dallas–Ft. Worth, for example, American Airlines has high schedule convenience, with departures quite close to the time preferences of a DFW-origin trav- eler. Similarly, at this airport American can offer high delivery reliability, schedule con- formance to expectations, and change order responsiveness. Passengers, especially business travelers, will pay substantial premiums for these high-service-quality character- istics because of the additional value such flights create in their own business activities. YM systems reserve idle capacity to meet these high-value demands when the late arrival of such requests is forecasted. YM systems also “protect” fewer seats and release more

FIGURE 14.9 Cross-Functional Revenue Management

Account managementScheduling

Pricing CustomersDemand estimation and forecasting

Order acceptance and refusal process

Capacity planning

532 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

capacity to deep-discount leisure segments of the market. Overall, fares per revenue passenger mile at a fortress hub such as Atlanta (where Delta has a 79 percent share) are 84 percent higher than at Orlando, where Delta again is the leading carrier but has only 32 percent of the market.

Revenue Management Decisions, Advanced Material Revenue management (RM) can be divided into three decisions: (1) a proactive pricing and aggregate capacity planning decision, (2) an inventory or capacity reallocation deci- sion, and (3) an overbooking decision. Figure 14.11 places these decisions in an RM con- ceptual framework, showing a flowchart of the components of a revenue management process. What all three decisions have in common is a tactical focus that depends on anticipated rival responses; a systems management philosophy that integrates marketing,

FIGURE 14.10 Ratio of Business to Leisure Fares and Airline Market Shares at Hub Airports

Market shares of top two carriers

Charlotte Minneapolis Pittsburgh Detroit Dallas/Ft. Worth Atlanta St. Louis Chicago Denver Phoenix Newark Seattle Orlando

US Airways Northwest US Airways Northwest American Delta TWA United United America West Continental Alaska Delta

93% 84% 81% 80% 65% 79% 69% 47% 69% 39% 52% 31% 32%

Delta United Southwest Delta Delta ASE Southwest American Delta Southwest United United Continental

3% 4% 4% 3%

19% 4%

14% 34% 5%

27% 10% 16% 10%

TotalSecond largestLargest carrierHub airport

96% 88% 84% 84% 84% 83% 83% 81% 74% 68% 62% 47% 42%

2.14

2.29

2.18

2.42

2.27

2.70

2.18

2.13 2.09

2.07

2.55

1.39

1.93

Source: USA Today (February 23, 1998), p. 3B; and BusinessWeek (July 20, 1998), p. 121.

Chapter 14: Pricing Techniques and Analysis 533

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

operations, and finance; and finally a multiproduct orientation that continuously recon- figures the firm’s product offerings. We now address the managerial economics of each RM decision.

Proactive Price Discrimination Proactive price discrimination involves maximiz- ing profits in the light of anticipated late-arriving demand and rival firm responses. In principle, computerized decision support systems (DSS) make it possible to reauction the remaining seats on a flight or the remaining runs on the printing press each time a new customer arrives on a reservation system. Conceivably, each customer would then experience first-degree price discrimination and pay a unique price that reflects time of delivery, service costs, and price elasticity. Few RM practitioners have adopted bid price systems. Instead, most set prices and initially allocate capacity with the familiar tech- niques of marginal analysis that we discussed earlier in the chapter for differential fares that allocate the capacity between business and nonbusiness air travelers on the 11:00 A.M. Thursday flight DFW to LAX.

Capacity Reallocation The second step in revenue management is to reallocate capacity or inventory as delivery times approach in the light of advance sales and con- firmed orders. Suppose you forecasted advance sales for business class on Thursday de- partures from Dallas to Los Angeles in accordance with the exponential function (tickets purchased = aBt) estimated in semilog form as

ln(tickets) = ln a + lnB(t) = α + βt [14.13]

FIGURE 14.11 Revenue Management Flowchart

4. Setting prices 5. Capacity scheduling and allocation

6. Tactical price adjustment

7. Capacity reallocation

2. Market segmentation (fencing)

3. Aggregate capacity choice

1. Demand estimation

8. Results measurement and overbooking

534 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

where t is a simple time trend variable. Similarly, suppose you forecasted all nonbusiness travel with a sales penetration function such as

Tickets purchased = ek1−k2ð1=tÞ [14.14]

estimated as

ln(tickets) = k1 − k2(1/t) [14.15]

where k1 and k2 are constants defining the rate of sales growth throughout the advance sales period. These forecasted business and nonbusiness “booking curves” are plotted in Figure 14.12. Note that the curves exhibit the early- and late-arriving characteristics of demand in the nonbusiness and business markets, respectively.

The forecasted booking curves reflect new demand arrivals and cancellations and are, in that sense, net bookings. If such bookings require substantial nonrefundable deposits on advance sales, then they reflect realized revenue rather than just potential sales.23 For the 11:00 A.M. flight to Los Angeles, the final demand target in Figure 14.12 is 63 busi- ness and 107 nonbusiness passengers. This initial allocation of total capacity applies when customer reservations open 180, 120, 60, or, in our case, 45 days prior to depar- ture. It is subject to change (and in fact often is changed) by the revenue managers.

Confidence intervals based on the different demand arrival distributions are then used to determine when actual bookings deviate a given amount, which triggers an exception report. For example, business travel appears to be greater than forecasted ticket sales at day t – 10 by a statistically significant amount. This violation of the threshold sales curve raises the question of whether to stop sales in the nonbusiness class where contri- bution margins (Pnonbus − MC = $188 − $130 = $58) are clearly lower than in the busi- ness class segment, where Pbus − MC = $261 − $130 = $131.

FIGURE 14.12 Advance Sales Forecasts, Bookings, and Threshold Curves

–38–45 –33 –28 –23

Days prior to departure

–18 –8–10–13 –3

Act ual

busin ess bo

okin g

Bus ine

ss f ore

ca st

107

63

42 40

Nonbusiness forecast

T ic

ke ts

p ur

ch as

ed

Busines s thres

hold c urve

23In the last RM decision, we consider the effects of no-shows on authorized overbookings.

threshold sales curve A level of advance sales that triggers reallocation of capacity.

Chapter 14: Pricing Techniques and Analysis 535

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The answer to this capacity reallocation question lies in applied statistics and in the initial profit-maximizing capacity choice. Marginal capacity expansions are justified as long as the expected incremental revenue minus variable cost (i.e., the additional ex- pected contribution to fixed cost) is greater than the incremental cost of additional ca- pacity. In capacity reallocation, the cost of additional capacity in the business class is an opportunity cost, or the forgone contribution from selling one less seat in the nonbusi- ness class. At the margin, we would reallocate capacity as long as the expected contribu- tion margin from allocating another seat to business travelers would exceed the lost contribution margin from a boarding denial in nonbusiness. That is,

ðPbus − VCÞðProb ShortagebusÞ = ðPnonbus − VCÞ $131ðProb ShortagebusÞ = $58 [14.16]

where the Prob Shortagebus is the probability that the business class will in fact be full and, therefore, the probability that the extra business-class seat will realize its $131 mar- ginal contribution.24

The preannounced prices and resulting contribution margins of $58 and $131 indi- cated (using Equation 14.16) that high yield spill should occur 44.1 percent of the time:

(Prob shortagebus) = $58/$131 = 0.441

For any business-class demand distribution (say, normally distributed with a mean of 60 seats and a standard deviation of 20 seats), we can calculate the optimal capacity choice as

μseats + zασseats = 60 + 0.148 × 20 = 63 seats [14.17]

where zα is the absolute value of the standard normal critical value (z value) for one- tailed alpha from Table 1 in Appendix B. These calculations correspond to the initial situation in which business-class capacity was set at 63 seats (see Figure 14.1). On this flight, 63 seats is often referred to as the protection level for business-class seats. Simi- larly, 107 seats is the authorization level for nonbusiness-class seats.

Now recall that at day t – 10 in Figure 14.12, we received an exception report: It ap- pears that the arrival distribution for next Thursday’s flight is not normally distributed with mean 60 and standard deviation 20, that is, N(60,20). Instead, the exception report may indicate mean demand has increased to N(62,20). Again, using the fact that with prices of $261 and $188 the optimal probability of stockout in business class is 0.441, we can use Equation 14.17 to calculate the new optimal capacity allocation at 62 seats + 0.148(20) seats = 65 seats. This result implies that a stop-sales policy of 105 seats should apply to the bookings accepted in nonbusiness travel and two seats (107 − 105) should be reallocated to business class. Continued monitoring of bookings relative to the fore- cast thresholds may result in a return of these seats to nonbusiness class or a still further reallocation of capacity toward business travelers.

The same questions and analyses of capacity allocation apply in assemble-to-order manufacturing when a sport apparel manufacturer or a customized paper products man- ufacturer must decide which orders to accept and which to refuse (i.e., how to allocate fixed total capacity). As yield management moves out of the service sector (e.g., airlines, hotels, rental cars, advertising agencies, hospitals, professional services) and into manufacturing, these managerial economics techniques become increasingly important.25

24It is quite possible to have a positive probability of stockout on both sides of Equation 14.16. Here, however, we assume an unlimited demand in the nonbusiness segment at the low fare of $188 for Dallas–Ft. Worth– LAX. That is, the probability of stockout is assumed to be 1.0 on the right-hand side.

protection level Capacity reserved for sale in higher margin segments.

authorization level Capacity authorized for sale in lower margin segments.

25See F. Harris and J. Pinder, “A Revenue Management Approach to Order Booking and Demand Management in Assemble-to-Order Manufacturing,” Journal of Operations Management (December 1995), pp. 299–309.

536 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Optimal Overbooking The third and final yield management decision is an optimal overbooking decision. Here the airline authorizes the reservation clerks to sell more seats than are available on each departure to combat the lost revenue from “no-shows.” Of course, many tickets entail discount fares that require advance purchase, but some business-class tickets are not purchased until check-in time. This means that a confirmed sale is not realized revenue until delivery time. In some industries orders can be canceled or shipments refused. At times the air carriers experienced up to 35 percent no-shows in certain city-pair markets.

The optimal overbooking decision illustrates marginal analysis in practice. Each airline seeks to minimize the summed costs of spoilage and spill. In Figure 14.13, as expected de- mand of business travelers approaches planned capacity and the expected load factor ap- proaches 100 percent, the total cost of spoilage (i.e., unsold seats × contributionbus) declines toward zero. In contrast, as the expected load factor approaches 100 percent, the costs of high-yield spill rise for three reasons. First, oversales represent lost contributions, which might have been captured by other service offerings (e.g., later flights); that is, some customers will “balk” and go to a competitor. Second, oversales necessitate out-of-pocket expenses to compensate passengers who board the airplane and then volunteer to give up their seats. And third, oversales and the resulting stockouts sacrifice customer goodwill and brand loyalty, thereby causing lost future sales. These rising total costs of high-yield spill also are depicted in Figure 14.13.

Total summed costs are reduced by a rising load factor as long as the rising costs of oversales are more than offset by the falling cost of spoilage. From load factors below 92–97 percent, the declining cost of spoilage more than offsets the rising cost of oversales for nonbusiness travel. Beyond 97 percent, the rate of spoilage cost reduction is less than the rate of oversales cost increase. This relationship is shown in the lower diagram by com- paring theMCnonbusiness oversales to the marginal benefit of reduced spoilage,MBspoilage reduction, which is the MC of unsold seats saved by planning a higher load factor. For the nonbusiness class, the optimal planned load factor appears to be 97 percent. In contrast, in the business class theMCbusiness oversales is so much higher as load factor increases that the optimal planned load factor is associated with a higher quality 94 percent.

Both decisions are referred to as overbooking decisions because a 97 percent expected load factor for the 105 seats now allocated to nonbusiness travelers may necessitate actually booking not 0.97 × 105 = 102 seats, but rather 127 seats [(127 × (1 − 0.2) = 102] in per- iods when no-shows are averaging 20 percent. Similarly, optimal overbooking in business

Example The Optimal Probability of a Stockout at Sport Obermeyer Recall that Sport Obermeyer must allocate its fixed retail shelf and display rack space between Pandora parkas and Champion sweatshirts. The lost retail contribu- tion margin every time Sport Obermeyer “spills” a Pandora parka customer is $15. The lost retail contribution margin on a Champion sweatshirt is $4. Knowing these margins and the relative sales effectiveness of particular shelf space, Sport Ober- meyer can use the tools of yield management to balance the costs of spoilage and spill in order to decide the optimal incidence of stockouts in Pandora parkas. Using Equation 14.16, Sport Obermeyer calculates that Pandora parkas should stock out 27 percent of the time: Prob (Shortageparka) = 0.27. With demand distri- bution data, Sport Obermeyer can calculate that a probability of shortage of $4/$15 = 0.27 requires stocking 85 size 8 parkas in each of its stores.

optimal overbooking A marginal analysis technique for balancing the cost of idle capacity (spoilage) against the opportunity cost of unserved demand (spill).

Chapter 14: Pricing Techniques and Analysis 537

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

may imply confirming reservations not for 61 seats in heavy (35 percent) no-show periods, but rather for 94 seats (94 × 0.65 = 61). On average, 102 nonbusiness and 61 business travelers enplane for the Thursday 11:00 A.M. flight to Los Angeles. Of course, these 163 total passengers expected are only an average of actual passenger counts, which may vary on any particular departure from large spoilage to severe oversales.

FIGURE 14.13 How the Overbooking Decision Minimizes the Summed Cost of Spoilage and Spill

Expected load factor

Expected load factor

$

1009992 93 94 95 96 97 98

Total costsbusiness

Total costsnonbusiness

Oversales costsbusiness

Oversales costsnonbusiness

Spoilage

94 97

MBspoilage reduction MCbusiness oversales

MCnonbusiness oversales

Example Pinpoint Booking Accuracy at American Airlines26

Price differentials on American Airlines’ popular 5:30 P.M. flight (Flight 2015) from Chicago to Phoenix are huge, ranging from $238 to $1,404 round trip. American constantly adjusts the capacity allocation to each of seven fare classes as advance sales data deviate from forecast. Four weeks prior to a recent departure, American had already sold 69 of the 125 coach seats at Super Saver fares. With three weeks to departure, all three fare classes below $300 had reached their maximum autho- rization levels and were closed to further reservations. One day before departure,

(Continued)

538 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

130 passengers were booked on the 125-seat flight, but American was still autho- rizing up to five additional full coach reservations. The yield management compu- ters predicted that cancellations and no-shows might go as high as 10. The next day, Flight 2015 departed full with no one denied boarding. The systems manage- ment objective of yield management is “to sell the right seat to the right customer at the right price at the right time” (Sabre Solutions). Such outcomes of the RM system are a critical success factor for American in a period when the index of U.S. airfares has risen only 8 percent during 1995–2010.

26Based on “High-Tech Pricing Boosts Business Fares,” Charlotte Observer (November 9, 1997), p. 1D; and “For U.S. Airlines, a Shakeout,” Wall Street Journal (September 19, 2005), p. A1.

Example Revenue Management in Baseball: The Baltimore Orioles27

Recent applications of revenue management have taken the techniques out of travel services and into private-pay elective surgeries, radio and television advertising, op- era and symphony concerts, law firms, consulting firms, golf courses, and now baseball. Like airlines, all these businesses have fixed capacity with perishable in- ventory; once the last out of the fifth inning has been called, empty seats offer no realizable value. Although season ticket holders are prominent in the planning of any professional sports franchise, single-game and three-game ticket packages re- main a substantial source of revenue. And unstable demand makes prediction of sales in these more immediate segments a challenging and potentially highly prof- itable process to do well, especially in baseball.

Most professional teams celebrate their sellouts, but some fail to realize that spare capacity (however slight) as game time approaches is a substantial revenue opportu- nity. Allowing discount ticket packages and promotions to displace last-minute walk- in customers often sacrifices high-margin repeat purchase business. At the same time, overall attendance in professional baseball remains below prestrike levels of the 1980s, and many games are played in ballparks only half full. The Baltimore Orioles revenue manager attempts to balance both of these errors of understocking and of overstock- ing. Single-game seats purchased well in advance are available at a discount.

However, a substantial capacity of well-placed seats is protected in anticipation of late-arriving, high-willingness-to-pay, game-day-only customers. Advance sales are tracked, and variances are noted relative to previous sales histories for that home stand against similar opponents. As game day approaches, authorization levels for release of discount tickets gradually adjust to reflect the probability of stockout in higher-margin segments. Ideally, on game day, perhaps 93 percent of the seats are filled with fans in a variety of different segments paying a variety of different prices, each reflecting the location, customer responsiveness, reliability, timing, and other ticketing services that particular customers prefer, thereby adding maximum value.

27Based on “Managing Baseball’s Yield,” Barron’s (September 11, 1995), p. 50; and “Tickets with Flex,” Sports Illus- trated (February 23, 2009), p. 62.

Chapter 14: Pricing Techniques and Analysis 539

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Yield management continuously reallocates capacity and adjusts these overbooking authorizations as advance sales data roll in. The incremental revenues from effective yield management can be significant. For example, American Airlines recently calculated its additional revenue from attending to these problems at $467 million per year. Marriott International estimates that revenue management contributes as much as $200 million each year to its revenue stream. And the Canadian Broadcasting Corporation realized a $2 million revenue gain the first two weeks after it adopted revenue management techniques.28

SUMMARY

� All pricing decisions should be proactive, systematic-analytical, and value-based, not reac- tive, ad hoc, and cost-based.

� Two conditions are required for effective differen- tial pricing:

1. One must be able to segment the market and prevent the transfer of the product (or service) from one segment to another.

2. Differences in the elasticity of demand at a given price between the market segments must be discernible.

� To maximize profits using differential pricing, the firm must allocate output in such a way that mar- ginal revenue is equal in the different market segments.

� Differential pricing is often implemented through the direct segmentation of intertemporal pricing or pricing by delivery location.

� Indirect segmentation to support differential pric- ing is often accomplished through two-part pric- ing. Optimal two-part prices entail a lump-sum access fee and a user charge that equals or exceeds marginal cost and varies with units consumed.

� Couponing is another way to price discriminate while charging the same list prices to different cus- tomers, some of whom are highly price sensitive and will redeem coupons and others who will not.

� Bundling is a third pricing mechanism that indi- rectly segments customers with inversely correlated demand across multiple products.

� Price discrimination is the act of selling at the same time the same good or service produced by a given distribution channel at different prices to different customers.

� A good’s pricing strategy varies throughout the product or service’s life cycle. A frequent pattern is target pricing, followed by penetration pricing, price skimming, value-based pricing, limit pricing, and finally niche pricing.

� Full-cost pricing and target pricing are inconsistent with the marginal pricing rules of economic theory. Incremental contribution analysis is a widely appli- cable method of economic analysis that can help pricing managers achieve a more efficient and profitable level of operation.

� Pricing on the Internet suffers from anonymity and lack of reputation effects, along with search across various product qualities being especially difficult to verify prior to purchase. These complications imply distinctly different pricing approaches for commodity-like products, search goods, and expe- rience goods.

� B2B pricing on the Internet requires a two-step process of multi-attribute matching to qualify for consideration as a supplier and then a dynamic pricing scheme to trade off additional features and functions as sources of value-in-use against lower price point alternatives.

� Yield management (YM) or revenue management (RM) consists of pricing and capacity allocation techniques for fixed-capacity manufacturers or

28As cited in R. Cross, Revenue Management: Hardcore Tactics for Market Domination (New York: Broadway Books, 1997).

540 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

service firms with perishable inventory and ran- dom demand.

� Flexible manufacturing systems and production- to-order with JIT delivery can seldom fully resolve the spill and spoilage problems addressed by RM.

� RM provides an optimal order acceptance and re- fusal process with cross-functional resolution of ac- count management, demand forecasting, and scheduling decisions.

� Proactive price discrimination equates the mar- ginal revenue from different segments of the target market. It does so with differential value-based prices that reflect delivery reliability, change order responsiveness, scheduling convenience, confor-

mance to expectations, and the value of these ser- vice quality characteristics to the particular class of customers (i.e., third-degree price discrimination).

� RM reallocates inventory or service capacity in ac- cordance with the condition (P − MC)a(Prob Shortage)a = (P − MC)b. This procedure identifies optimal protection levels for high-margin seg- ments, accounts, and customers and an optimal authorization level for release of capacity to lower-margin segments, accounts, and customers.

� The optimal overbooking decision equates the de- clining marginal cost of spoilage as load factor or capacity utilization increases with the rising mar- ginal cost of spill (i.e., oversales).

Exercises 1. The price elasticity of demand for a textbook sold in the United States is esti- mated to be −2.0, whereas the price elasticity of demand for books sold overseas is −3.0. The U.S. market requires hardcover books with a marginal cost of $40; the overseas market is normally served with softcover texts on newsprint, having a marginal cost of only $15. Calculate the profit-maximizing price in each market.

Hint: Remember that MR = P 1 + 1 ED

� �� �

2. The price elasticity of demand for air travel differs radically from first-class (−1.3) to unrestricted coach (−1.4) to restricted discount coach (−1.9). What do these elasticities mean for optimal prices (fares) on a cross-country trip with incremen- tal variable costs (marginal costs) equal to $120?

3. American Export-Import Shipping Company operates a general cargo carrier ser- vice between New York and several Western European ports. It hauls two major categories of freight: manufactured items and semimanufactured raw materials. The demand functions for these two classes of goods are

P1 = 100 − 2Q1 P2 = 80 − Q2

where Qi = tons of freight moved. The total cost function for American is

TC = 20 + 4(Q1 + Q2)

a. Determine the firm’s total profit function. b. What are the profit-maximizing levels of price and output for the two

freight categories? c. At these levels of output, calculate the marginal revenue in each market. d. What are American’s total profits if it is effectively able to charge different

prices in the two markets? e. If American is required by law to charge the same per-ton rate to all users,

calculate the new profit-maximizing level of price and output. What are the profits in this situation?

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

Chapter 14: Pricing Techniques and Analysis 541

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

f. Explain the difference in profit levels between the differential pricing and uniform pricing cases. Hint: First calculate the point price elasticity of demand under the uniform price-output solution.

4. Sort the following products into those priced with two-part tariffs, user charges only, or lump sum access fees only: pay-per-view movies on cable TV, pay phones, Netflix, iTunes, country club membership, soda from vending machines, laundromats, cell phones, season ticket holders with seat rights.

5. Phillips Industries manufactures a certain product that can be sold directly to re- tail outlets or to the Superior Company for further processing and eventual sale as a completely different product. The demand function for each of these markets is

Retail Outlets: P1 = 60 − 2Q1

Superior Company: P2 = 40 − Q2

where P1 and P2 are the prices charged and Q1 and Q2 are the quantities sold in the respective markets. Phillips’ total cost function for the manufacture of this product is

TC = 10 + 8(Q1 + Q2)

a. Determine Phillips’ total profit function. b. What are the profit-maximizing price and output levels for the product in

the two markets? c. At these levels of output, calculate the marginal revenue in each market. d. What are Phillips’ total profits if the firm is effectively able to charge dif-

ferent prices in the two markets? e. Calculate the profit-maximizing level of price and output if Phillips is re-

quired to charge the same price per unit in each market. What are Phillips’ profits under this condition?

6. In the face of stable (or declining) enrollments and increasing costs, many colleges and universities, both public and private, find themselves in progressively tighter financial dilemmas that require basic reexamination of the pricing schemes used by institutions of higher learning. One proposal advocated by the Committee for Economic Development (CED) and others has been the use of more nearly full- cost pricing of higher education, combined with the government provision of suf- ficient loan funds to students who would not otherwise have access to reasonable loan terms in private markets. Advocates of such proposals argue that the private rate of return to student investors is sufficiently high to stimulate socially optimal levels of demand for education, even with the higher tuition rates. Others argue against the existence of significant external benefits to undergraduate education to warrant the current high levels of public support.

As with current university pricing schemes, proponents of full-cost pricing generally argue for a standard fee (albeit higher than at present) for all students. Standard-fee proposals ignore relative cost and demand differences among activi- ties in the university.

a. Discuss several possible rationales for charging different prices for different courses of study.

b. What are the income-distribution effects of a pricing scheme that charges the same fee to all students?

542 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

c. If universities adopted a system of full-cost (or marginal cost) pricing for various courses, what would you expect the impact on the efficiency of resource allocations within the university to be?

d. Would you complain less about large lecture sections taught by graduate students if these were priced significantly lower than small seminars taught by outstanding scholars?

e. What problems could you see arising from a university that adopted such a pricing scheme?

7. General Medical makes disposable syringes for hospitals and doctor supply com- panies. The company uses cost-plus pricing and currently charges 150 percent of average variable costs. General Medical learned of an opportunity to sell 300,000 syringes to the Department of Defense if they can be delivered within three months at a price not in excess of $1 each. General Medical normally sells its syringes for $1.20 each.

If General Medical accepts the Defense Department order, it will have to forgo sales of 100,000 syringes to its regular customers over this time period, although this loss of sales is not expected to affect future sales.

a. Should General Medical accept the Defense Department order? b. If sales for the balance of the year are expected to be 50,000 units less be-

cause of some lost customers who do not return, should the order be accepted? (Ignore any effects beyond one year.)

8. The Pear Computer Company just developed a totally revolutionary new personal computer. It estimates that it will take competitors at least two years to produce equivalent products. The demand function for the computer is estimated to be

P = 2,500 − 0.0005Q

The marginal (and average variable) cost of producing the computer is $900. a. Compute the profit-maximizing price and output levels assuming Pear acts

as a monopolist for its product. b. Determine the total contribution to profits and fixed costs from the solution

generated in Part (a). Pear Computer is considering an alternative pricing strategy of price

skimming. It plans to set the following schedule of prices over the coming two years:

TIME PERIOD PRICE QUANTITY SOLD

1 $2,400 200,000

2 2,200 200,000

3 2,000 200,000

4 1,800 200,000

5 1,700 200,000

6 1,600 200,000

7 1,500 200,000

8 1,400 200,000

9 1,300 200,000

10 1,200 200,000

Chapter 14: Pricing Techniques and Analysis 543

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

c. Calculate the contribution to profit and overhead for each of the 10 time periods and prices.

d. Compare your results in Part (c) with your answers in Part (b). e. Explain the major advantages and disadvantages of price skimming as a

pricing strategy.

9. Explain the effect on capacity reallocations of advance sales data indicating mean demand of 55 rather than 60 during a slow travel week for business class, using the information in Figure 14.1, Table 14.2, and Equation 14.17.

10. Suppose the frequent-flyer program raised the cost of high-yield spill twofold be- cause business customers who are denied boarding now take their business to other carriers for several future trips, not just the current one. Reanalyze the over- booking decision in Figure 14.13 under these circumstances. Will overbooking of business-class service increase or decrease?

11. An aircraft with 100 seats serves passengers through two types of fares: full ($550) and discount ($250). Extra passengers have $50 marginal cost. Demand for dis- count tickets is unlimited, while demand for full-fare tickets is evenly distributed between 11 and 30 seats. How many seats should be protected for full-fare pas- sengers and not authorized for release to the discounted $250 segment?

544 Part 4: Pricing and Output Decisions: Strategy and Tactics

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PART5 ORGANIZATIONAL ARCHITECTURE

AND REGULATION

Part 5 addresses the new institutional economics of organizational architecture aswell as the regulation of business. Major themes include incentive contracting,the choice of organizational form (e.g., vertically integrating to redraw the bound- aries of the firm), and the regulation/deregulation debate. Chapter 15 discusses the the- ory of business contracting, managerial incentive contracts, the principal-agent problem, corporate governance, licensing of trade secrets, the dissolution of partnerships, and ver- tical integration. Appendix 15A explores optimal mechanism design in auctions and in- centive compatibility for joint ventures. Chapter 16 then addresses the economic regulation of business, including antitrust, patenting, and licensing, as well as regulatory and private market approaches for controlling externalities. Chapter 17 discusses capital budgeting techniques used in acquisition, merger, and spin-off activities to change the organizational boundaries of the firm.

ECONOMIC ANALYSIS AND DECISIONS

1. Demand Analysis and Forecasting

2. Production and Cost Analysis 3. Pricing Analysis 4. Capital Expenditure Analysis

ECONOMIC, POLITICAL, AND SOCIAL ENVIRONMENT

1. Business Conditions (Trends, Cycles, and Seasonal Effects)

2. Factor Market Conditions (Capital, Labor, Land, and Raw Materials)

3. Competitors’ Responses 4. External, Legal, and Regulatory

Constraints 5. Organizational (Internal)

Constraints

Cash Flows Risk

Firm Value (Shareholders’ Wealth)

545

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

15 CHAP T E R

Contracting, Governance, and Organizational Form CHAPTER PREVIEW This chapter explores the coordination and control problems faced by every business organization and the institutional mechanisms designed to solve these problems in a least-cost manner. The most important organizational architecture decision is determining the boundary of the firm (i.e., the span of hierarchical control). In dealing with external suppliers, outsource partners, internal divisions, authorized distributors, franchisees, and licensees, every firm must decide where the internal organization stops and where market transactions take over.

Contracts between business organizations provide an ex ante framework that defines these relationships, but all contracts are purposefully incomplete. Consequently, every firm must address the potential for post-contractual opportunistic behavior by business partners and then design governance mechanisms to reduce these contractual hazards. Should Dell make or buy subassembly components for their PCs? Should Canon license its digital camera technology for Internet distribution by Verizon, or should it invest in a strategic partnership with Verizon? Should Microsoft vertically integrate into media delivery devices by buying WebTV? Should Red Hat continue to adopt open source architecture and allow its licensees to duplicate, modify, and redistribute its Linux-based software without charge?

We address these questions initially from the perspective of the coordination game between manufacturers and distributors using the game theory techniques of Chapter 13.

MANAGERIAL CHALLENGE Controlling the Vertical: Ultimate TV1

Enormous business opportunities loom on the horizon for companies that operate at the intersection of Web-based Internet services and digital TV. Over the next five years, 220 million analog television sets will be replaced by $150 billion worth of PC-enabled digital televisions.

Microsoft has invested heavily in digital entertain- ment programming for these smart televisions and television-enabled PCs. Their know-how and trade

secret investments are largely non-redeployable and include the operating system and user interface back- bone for everything from interactive museum tours to distance learning virtual courses. WebTV’s (now MSN TV’s) patented signal compression chip has freed con- tent providers of the bandwidth limitations that prevent the Net from transmitting high-speed Web images and video. Digital TV manufacturers quickly established

546

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

INTRODUCTION Organizational form and institutional arrangements play an extensive role in eliciting ef- ficient behavior. Incentive contracts can motivate manager-agents to pursue the interests of owner-principals. Incentive-compatible revelation mechanisms can increase the mar- ket value of joint ventures between partners such as Nokia and Siemens. On another front, allowing the freewheeling of electricity from one public utility to another or priva- tizing Conrail, British Telecom, Japan Air Lines, Teléfonos de México, and Société Gén- érale can improve the incentives to maximize capitalized value in these formerly bloated public monopolies.

Institutional choices also involve the form of organization that companies adopt. For example, some manufacturers, such as Goodyear Tires, develop franchise dealerships rather than engage in the vertical requirements contracting with independent retailers preferred by other manufacturers such as Michelin. Perhaps the most important applica- tion of these concepts occurs in deciding the boundary of the firm—whether to vertically integrate throughout the supply chain like Exxon or outsource like Dell.

THE ROLE OF CONTRACTING IN COOPERATIVE GAMES In Chapter 13, we saw that once a manufacturer commits to updating a product, distrib- utors sometimes find that their best-reply response is to continue extensive selling efforts and post-sale services. If so, the required coordination of manufacturer and distributor actions can be achieved by a self-enforcing reliance relationship. At other times, however, the payoffs are such that coordination requires something more than the best-reply re- sponse concept we examined. Consider again the decisions in Figure 15.1. These are the

partnerships with WebTV to assess the danger and take the first steps toward acquiring an equity stake in this emergent technology.

Microsoft then decided to vertically integrate and preemptively bought WebTV for $425 million. Micro- soft intended to combine its one-way dependent and reliant digital entertainment assets with WebTV’s tech- nology to produce digital consumer products for smart phones and handheld PCs. Microsoft also sought to

become a cable TV industry standard by investing more than $10 billion in equity ownership in AT&T, Telewest, Comcast, and three European cable firms.

Discussion Questions

� Brainstorm about why Microsoft, a software giant and content provider, found it necessary to buy a controlling interest in the technology by which its content would be delivered to end users?

� What is another example of a manufacturing company that controls the retailing of its products?

� What analogies can you draw between the two companies and the two products?

1Based on “Why Microsoft Is Glued to the Tube,” BusinessWeek (September 22, 1997), p. 96; “Microsoft to Buy WebTV for $425 Million,” Wall Street Journal (May 7, 1997), p. A8; “Microsoft’s Blank Screen,” The Economist (September 16, 2000), p. 74; and “Smart TV Gets Even Smarter,” BusinessWeek (April 16, 2001), pp. 132–133.

MANAGERIAL CHALLENGE Continued ©

An ge lo Ca va lli /B ue na

Vi st a Im ag es /P ho to di sc /G et ty Im ag es

Chapter 15: Contracting, Governance, and Organizational Form 547

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

same actions and payoffs we examined earlier in the Manufacturer-Distributor II game (Figure 13.4).

Recall that in the subgame perfect equilibrium {Update, Discontinue, Advertise}, the retail distributor is better off discontinuing some selling effort, knowing full well that the manufacturer’s best-reply response is to advertise anyway. See the boxed and starred out- come in Figure 15.1. This odd combination of actions dominates all other sequential pat- terns that meet the conditions of best-reply response for each player at each proper subgame node in the decision tree. However, sales volume would decline, so the manu- facturer is clearly worse off than would have been the case had the distributor continued all selling efforts. In that event, the manufacturer would realize either $300,000 or $350,000, whereas updating and advertising to prevent a sales collapse from the distribu- tors’ reduced selling effort results in a manufacturer’s payoff of only $280,000.

Moreover, the institutional arrangements under which defection by the retail distribu- tor occurs regularly are not value maximizing. The subgame perfect equilibrium strategy {Update, Discontinue, Advertise} creates total value after expenses of $120,000 + $280,000 = $400,000. {Update, Continue, Not Advertise} generates $450,000 total payoffs, and {Update, Continue, Advertise} generates $480,000 total payoffs. One might then ex- pect some organizational form to emerge to realize this additional value. One alternative is vertical integration. By buying the distributor firm (for something slightly more than the discounted present value of the retail distributor’s present value $120,000), the man- ufacturer could impose the value-maximizing actions and resolve coordination and con- trol with internal monitoring and incentive systems within the consolidated firm.

Alternatively, manufacturers might suspend shipments to retail distributors and change distributors frequently as one after another violated the expectations of the rela- tionship and pursued the “Discontinue” strategy. Because instability in the distribution channel imposes substantial start-up costs, the manufacturer may instead enter into a cooperative game of credible promises and side payments (i.e., a relational contract).

FIGURE 15.1 Vertical Requirements Contracting Is Required to Maximize Value

M

$60,000 $380,000

$100,000 $350,000

$40,000 $120,000

$180,000 $300,000

Update

Not update

M2

R2

Discontinue selling effort Advertise

Not advertise

Continue R1

M1

Not advertise

Not advertise

Advertise

Not advertise

M4

M3

Discontinue selling effort

Continue

$120,000* $280,000**

$130,000 $150,000

Retail distributor Manufacturer

548 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Vertical Requirements Contracts Contracts are binding, third-party enforceable agreements designed to facilitate deferred exchange. A promisee undertakes some costly action (perhaps making a sidepayment that the law terms a consideration) in exchange for and relying upon the promisor’s pledge of a subsequent performance. Here, the manufacturer updates the product, relying upon the re- tail distributor to subsequently perform presale selling efforts. The manufacturer then deci- des about manufacturer-sponsored advertising by the retail distributor. Of course, distributors can appear to promise one thing and then deliver another, but vertical require- ments contracts are one method of establishing the credibility of such promises.

For the contracting problem in Figure 15.1, a vertical requirements contract might offer the distributor more than the $120,000 of the equilibrium noncooperative actions. A cooperative surplus of $480,000 (in the first row) − $400,000 (in the third row) = $80,000 is generated by retail distributors who can be induced to provide full-selling services when the manufacturer updates and advertises the product. Therefore, the manufacturer might offer the distributor’s threat point (the $120,000 payoff from defecting) plus one-half of this in- crease in value from performing as promised (namely, another $40,000). Such an agree- ment would yield actions that increase the manufacturer’s payoff from $280,000 to ($480,000 − [$120,000 + $40,000]) = $320,000. Assuming alternative distributors were available, this contract would be accepted by the present distributor, and both players would be $40,000 better off than in the subgame perfect equilibrium {Update, Discontinue, Advertise}, which made no use of third-party-enforceable contracting.2

Again, credible commitments are the key. Given the decision timing, the manufacturer could promise to advertise but then not do so. Because of this second possibility of post- contractual opportunism, the retail distributor may require escrow accounts for coopera- tive advertising or the parties could stipulate a $51,000 damage penalty should the manu- facturer breach its duty to advertise after the distributor expended full effort in attempting to sell an upgraded product. In that event, the manufacturer in Figure 15.1 would have $300,000 from advertising and $350,000 − $51,000 = $299,000 from not advertising. The retail distributor would be better off continuing its selling effort ($100,000 + $51,000) rather than discontinuing its selling effort ($120,000). We could then anticipate a value- maximizing {Update, Continue Effort, Advertise} outcome with a maximum value of $180,000 + $300,000 = $480,000 profit payoff.

Negotiating position may in the end reallocate some of the cooperative surplus back to the manufacturer in the form of franchise fees. For example, a profit-sharing franchise con- tract with the $51,000 stipulated penalty clause for manufacturer breach on advertising might start by asking the distributor to pay a $50,000 per period franchise fee. On net, then, the distributor would receive updated products, $180,000 in operating profits minus $50,000 in franchise fees, plus a stipulated damages agreement regarding co-op advertising funded by the manufacturer. The manufacturer would receive continued full selling effort by the distributor, $300,000 in operating profits, plus the $50,000 fee as franchisor.

The Function of Commercial Contracts Forming such a contract provides a hostage beyond the mere reputational asset that prospective distributors might offer. In exchange for an agreed consideration, the

contracts Third- party enforceable agreements designed to facilitate deferred exchange.

vertical requirements contract A third-party enforceable agreement between stages of production in a product’s value chain.

2The actual vertical requirements contract here is likely to be structured around an offer of a percentage of the profits. By acknowledging the uncertainty of the final product value, the manufacturer and distributor could agree to share this risk. In particular, a vertical requirements contract that offered to grant 33 percent of the summed profits ($160,000 of $480,000) to an authorized distributor in exchange for full-selling effort and after-sale service of an updated manufacturer-advertised product would maximize the value of this business opportunity.

Chapter 15: Contracting, Governance, and Organizational Form 549

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

promisee receives a credible promise. The promisor’s commitment to perform is credi- ble because the legal rules of contract interpretation and enforcement provide assurance that any expectations clearly spelled out by the parties will be met. Although courts seldom order recalcitrant contractors to perform specifically as was promised, they are quick to award expectation damages that leave the parties no worse off than was an- ticipated under the contract. Standard contract remedies therefore provide incentives for efficient precaution by the promisor and for no more than efficient reliance on the promise by the promisee.

A stipulation procedure in the law of contract works exceptionally well for fully anticipated events. However, the rules of contract case law that have evolved (several of which are summarized in the bottom half of Table 15.1) also reduce the transaction costs of renegotiation and settlement when unanticipated events occur. For example, the market price that can be realized on a Volvo-GM truck can change dramatically between making an investment at amanufacturing facility to design and produce an updated truck that only runs on ethanol and the subsequent promotion and sale of such trucks six months later. Suppose in the meantime that the market price collapses because a competitor’s new and improved hybrid-fuel truck is introduced.

If the initial truck manufacturer and distributor agreed to a fixed-price contract six months earlier, the manufacturer will get the agreed-upon revenue because the distribu- tor took that risk. On the other hand, if changes in the regulatory environment over the six months make it illegal to sell that ethanol-based model of truck, then the frustration of purpose doctrine of the Uniform Commercial Code (UCC) would excuse the distrib- utor from the contractual obligation to pay.

Example Crankshaft Delivery Delay Causes Plant Closing The role of contract remedies as incentives is well illustrated by the historical case of Hadley v. Baxendale, Court of Exchequer 1854, 9 Exch 341. A mill owner or- dered a replacement for a broken crankshaft from a machine shop that agreed to a standard repair and return of the mill owner’s equipment. When return delivery was delayed because of poor road conditions, the mill owner sued for lost profits resulting from his extended plant closing. The court rejected this claim for extraor- dinary damages because the machine shop had taken the customary shipping pre- cautions and would have been expected to do more (perhaps by arranging for an expedited delivery by express coach) only if the mill owner had stipulated the extraordinary damages that would arise from further delay.

In other words, the machine shop was entitled to expect that the mill owner would not rely excessively on the promise of a three-day repair unless informed to the contrary. If the mill owner had time-sensitive business scheduled immedi- ately thereafter and no temporary substitute crankshaft available, it was his respon- sibility to disclose those potentially destructive private facts and thereby elicit a different level of precaution. Therefore, the mill owner’s reliance was excessive and inefficient, not deserving of the reinforcement that would have resulted from the court awarding lost profit.3

3An excellent extended discussion of the role of contract remedies as incentives for efficient reliance and efficient pre- caution against nonperformance appears in R. Cooter and T. Ulen, Law and Economics, 5th ed. (Reading, MA: Addison-Wesley, 2008).

expectation damages A remedy for breach of contract designed to elicit efficient precaution and efficient reliance on promises.

frustration of purpose doctrine An illustration of the default rules of contract law that can result in excusal of contract promises.

550 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Contracts facilitate deferred exchange. The rules of contract law, as embodied primar- ily in the common law but also codified in statutes including the UCC, provide predict- able outcomes at low transaction costs. For example, as we saw earlier, the courts almost always impose a liability for expectation damages on parties that breach their contract promises. Because circumstances change, expectation damages are often more efficient than forcing a promisor to fulfill the contract. In these cases, the costs of the expectations damages are often less than the costs of taking the actions specified in the contract.

In some cases, contract promises are excused altogether. These excusals fall into two categories: exceedingly rare formation excusals and more frequent performance excusals. If I sell you a damaged Learjet without disclosing the damage, you (the buyer) can ask to be excused. On the other hand, an astute buyer of a damaged jet who recognizes the potential for enhancing the value through inexpensive repair can profit from his or her asymmetric information without concern about whether the court might later set aside the sales contract and restore the plane to its original owner. Contract law supports this delicate balance of requiring the disclosure of destructive facts without reducing the incen- tive to develop asymmetric information that enhances value (i.e., constructive facts).

Spot market transactions pose the least informational and incentive problems. For example, buying electricity off the grid at quarter until the hour for delivery on the hour avoids pricing risk and the possibility of opportunistic behavior. Complete and cer- tain information plus competitive entry and exit in efficient markets implies that equilib- rium market prices will reflect all relevant information. This makes the current price the best forecast of future prices. However, the availability of simple commodity transactions for immediate delivery fails to solve several issues that arise in most business contracting.

TABLE 15.1 A SPECTRUM OF ALTERNATIVE CONTRACT ENVIRONMENTS

FOR MANUFACTURERS AND DISTRIBUTORS

SPOT MARKET TRANSACTIONS

VERTICAL REQUIRE- MENTS CONTRACT

RELATIONAL CONTRACT

Timing Instantaneous, one-time-only Deferred exchange; promise of future performance for immediate consideration

Repeat business

Players Anonymous buyers/sellers Contract partners Well-known dealers/ agents

Enforcement Barter or consideration for a consideration

Enforced by impartial third parties

Self-enforcing; best-reply responses

Information Perfect (complete and certain) information + competition leads to efficient markets

Purposefully incomplete con- tracts embrace ambiguity; governance mechanisms

Reputation; signaling/ bluffing games

SOME RULES OF CONTRACT LAW Contracts facilitate deferred exchange by addressing uncertain performance outcomes (the incomplete contract problem), unobservable effort in assuring performance (the moral hazard problem), and recontracting hazards (the holdup problem)

BASIC FUNCTIONS OF CONTRACT ILLUSTRATIVE CONTRACT RULE

1. Providing incentives for efficient precaution and efficient reliance

2. Encouraging the discovery of asymmetric information

3. Providing risk allocation mechanisms 4. Reducing transaction costs

1. Award of expectation damages 2. Required disclosure of destructive but not

constructive facts 3. Frustration of purpose doctrine 4. Nonexcusal in forward sales contracts

spot market transactions An instantaneous one- time-only exchange of typically standardized goods between anonymous buyers and sellers.

Chapter 15: Contracting, Governance, and Organizational Form 551

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Consider the deferred exchange of a present consideration for a promise of future for- ward sales. Suppose, that is, the truck manufacturer and distributor entered into a forward sales contract for diesel fuel to be used as a promotion to enhance the distributor’s selling effort, and subsequently the price of diesel fuel tripled, what would happen?

The default rule of the UCC for forward sales contracts is very explicit and subject to few, if any, excusals. If the truck manufacturer sold forward to the distributor 100,000 gallons of diesel at $2.33 per gallon in June 2010 for truck sales promotions at delivery in December 2010, and if the December spot market price rises to $4, the manufacturer took that risk of rising cost in fulfilling the contract. A manufacturer’s plea that it would be ruinous financially to deliver as promised will fall on deaf ears; the manufacturer must either deliver the 100,000 gallons in December or face an immediate court judgment of ($4 − $2.33) × 100,000 = $167,000 awarded to the distributor. Every commercial contract must either stipulate the allocation of such risks or operate under the UCC default rules that are intended to increase predictability and thereby reduce the transaction costs of forward business contracting.

Table 15.1 summarizes several other characteristic differences between spot market transactions, reputation-based relationships, and vertical requirements contracts between manufacturers and distributors (that might include forward sales agreements for the die- sel gasoline promotion). Whether manufacturers and distributors will decide to employ spot market transactions, relational contracting, fixed profit-share franchise contracts, or vertical integration depends on the relative transaction costs of coordination and control in the various contractual settings.

Example The Enforcement and Excusal of Contract Promises: The Extraordinary Case of 9/114

In a typical performance excuse, contingent events like a change in regulatory con- straints may frustrate the purpose of a contract. If the U.S. FDA withdraws approval for a pharmaceutical’s claim to be safe and effective, Merck may obtain an excusal of its contractual agreement with the inventor to license and market the drug.

Contracts are also excused because of unforeseen natural disasters or acts of war. On the morning of September 11, 2001, Bank of New York was obligated to clear and provide cash settlements for about 84,000 government security transac- tions. The client firms, such as J.P. Morgan Chase & Co., had invested large sums in real-time hard-wired data feeds and sophisticated telecommunications connec- tions to Bank of New York. Yet, three of the bank’s buildings in lower Manhattan were either damaged or forced to close because of the terrorist attack. At one point, Bank of New York owed Citigroup and Morgan $30 million each on settlements the bank could not authorize for final clearance because of the chaos at their busi- ness facilities.

Under other circumstances, each day said settlement was delayed would have resulted in a claim against the Bank of New York for expectation damages of ap- proximately (1/365) × 5% × $30 million = $4,109 per contract per day. With razor- slim margins on these clearing and settlement operations, such damages would have quickly exhausted all profit. However, under the unforeseen circumstances of 9/11, an act of war had prevented performance of the settlement and clearing con- tract, and the Bank of New York was entitled to a performance excuse.

4Based in part on “Little Changes at Bank of New York,” Wall Street Journal (March 8, 2002), p. C11.

forward sales contracts A consensual agreement to exchange goods delivered in the future for cash today, with no possibility of performance excuse.

552 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Incomplete Information, Incomplete Contracting, and Post-Contractual Opportunism Practically all exchanges, whether for products, financial claims, or labor services, are con- ducted under conditions of incomplete information. On the one hand, decision makers often face random disturbances on the outcome of their actions. This type of incomplete information is typically handled in routine ways by risk spreading in insurance markets, which pool such casualty risks and thereby reduce the loss exposure to any individual business or household. Randomly occurring injuries at a consumer electronics assembly plant seldom coincide with injuries in a firm’s delivery trucks or severe weather disrup- tions at a firm’s shipping facility. As a result, modest insurance premiums can cover the cost of the anticipated claims involving such diversifiable risk events.

However, incomplete information as to the existence and probability of remote risks (i.e., what possible outcomes might occur) often prevent the parties at risk from writing insurance contracts. Consider the full contingent claims contract you and your surgeon would need to write before an organ transplant operation or two pharmaceutical compa- nies would need before one licensed the rights to produce a pregnancy-related drug to the other. To develop all the accurate information required for an agreement about po- tential losses and full compensation in all possible future contingencies is simply prohib- itively expensive. Consequently, few transplant patients and few business partners attempt to negotiate full contingent claims contracts. Such prohibitively expensive infor- mation costs explain why contracts are often incomplete by design.

One immediate consequence of incomplete contracts is the possibility of post- contractual opportunistic behavior that is not specifically prohibited by the contract. Employees who receive on-the-job training (OJT) may moonlight with their new skills. Managers may reconfigure assets following a labor contract concession in ways their on- going employees did not anticipate. Baseball players may attempt a hold out at the time of contract renewals just before a World Series. Knowing this tendency, companies provide less OJT, workers agree to fewer wage concessions, and owners develop more farm team replacement players than they otherwise would. So, the incompleteness of contracts results in inefficient behavior as an inescapable consequence of costly and therefore incomplete information. To reduce these inefficiencies, companies adopt governance mechanisms such as mandatory arbitration agreements to help resolve post-contractual disputes.

CORPORATE GOVERNANCE AND THE PROBLEM OF MORAL HAZARD Oliver Williamson, winner of the 2009 Nobel Prize in Economic Sciences, emphasizes that contracts pose the ex ante framework but that governance mechanisms provide the ex post implementation required to maximize value:

The parties to commercial contracts are held to be perceptive about the nature of the con- tractual relations of which they are a part, including an awareness of potential contractual hazards. However, because complex contracts are unavoidably incomplete—it being impos- sible or prohibitively costly to make provision for all possible contingencies ex ante—much of the relevant contractual action is borne by the ex postmechanisms of governance.5

Hence, because of post-contractual opportunistic behavior, any vertical requirements contract between the manufacturer and distributor in Figure 15.1 will be only the

incomplete information Uncertain knowledge of payoffs, choices, and other factors.

full contingent claims contract An agreement about all possible future events.

post-contractual opportunistic behavior Actions that take advantage of a contract partner’s vulnerabilities and are not specifically prohibited by the terms.

governance mechanisms Processes to detect, resolve, and reduce post-contractual opportunism.

5Oliver Williamson, “Economics and Organization: A Primer,” California Management Review (Winter 1996), p. 136. See also Williamson’s The Mechanisms of Governance (New York: Oxford University Press, 1996).

Chapter 15: Contracting, Governance, and Organizational Form 553

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

beginning of their agreement. In addition, the firms will need to resolve a pivotal coordi- nation and control issue: the inherent unobservability of selling effort on the part of the retail distributor. Unobservable effort in fulfilling contract promises illustrates a difficult but standard business contracting problem of “moral hazard.” After securing terms to their liking, all contract partners must be wary of the potential for shirking on the agree- ment in inconspicuous and hard-to-detect but potentially ruinous ways.

In the next section, we will apply the moral hazard idea to managerial contracting, but first let’s consider the commercial lender’s moral hazard problem with eliciting un- observable borrower effort in selecting safe working capital projects. A known reliable borrower may approach a lender with a randomly occurring liquidity crisis that necessi- tates an extension of its bank line of credit.6 The lender then offers terms for the loan renewal: an interest rate, principal amount, collateral requirements, and loan term. The situation is depicted in Figure 15.2. If the borrower decides to accept, a line of credit extension is granted. Then a random process intervenes presenting one of several uncer- tain business opportunities to the borrower. As in earlier chapters, we signify the role of uncertainty at a node in the decision tree with a large N for a choice by Nature. The possible business opportunities are a spectrum from the relatively safe investments in in- ventory during periods of back order to an extension of the company’s receivables policy allowing customers to pay within 90 days rather than pay cash at the point of purchase. Because product sales may be sensitive to credit terms such as “90 days same as cash,” the latter use of working capital has a higher expected return but is much more risky in that uncollected customer accounts may skyrocket. Finally, borrowers may go so far as to

FIGURE 15.2 The Problem of Moral Hazard in Line-of-Credit Lending: A Game-Theoretic Model of Workouts

L–Lender; B–Known, Reliable Borrower; N–Nature

Other default risk mitigation mechanisms?

B

Interest rate

L

A well-known customer

applies for line-of- credit extension Principal

amount

Term

Collateral

High

Moderate Low risk and return (inventory)

Moderate risk and return

High risk and return (international expansion)

B

N

Refuses

Accepts

B

N

Terms of loan Uncertain business

opportunities (possible asset substitution)

Potential moral hazard problem in screening projects

B

High effort in screening projects

Low effort

Large

Small

Long

Short Extensive

Minimal

Liquidity crisis

No crisis

moral hazard problem A problem of post- contractual opportunism that arises from unverifiable or unobservable contract performance.

6In the situation examined here, we abstract from the hidden information problem of adverse selection (by as- suming that the borrowers are known customers of the lending institution) to focus attention on the hidden action problem of moral hazard in commercial lending.

554 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

use the new line of credit for an overseas expansion with all the attendant massive risk of failure and default accompanying such initiatives.

The moral hazard problem for the lender is then motivated. The commercial bank wants the borrower to exercise great care, high effort, and good judgment in selecting projects on which to expend its newly granted working capital from the line-of-credit extension. However, banks must move carefully in setting the loan terms to elicit this largely hidden action. Remember, the bank does not know in advance what business op- portunity the borrower will be facing. This is not a project financing situation where the commercial banker can take part in assessing the company’s capital budgeting proposals and directly monitor its ROI and the attendant risks. Instead, the bank makes the funds available and must then elicit subsequent borrower effort in appropriately screening pro- jects that present themselves for possible investment.

What loan terms should the lender offer? Large loans with long repayment terms are most desired by the borrowers in order to gain the most financial flexibility in the face of their current liquidity crisis. Having more funds and more time to straighten out a busi- ness plan that has gone awry is preferable, but do these terms elicit more or less effort in screening projects so as to prevent or contain a loss? Surprisingly, higher interest rates to reflect the high default risk are not the answer. A high interest rate forces reliable bor- rowers to seek out riskier working capital projects in order to secure higher expected re- turns and therefore be in a position to repay the loan. More moderate rates supported by extensive collateral pledges of security would seem to move the reliable borrower in the direction of more effort to find safer projects.7

The Need for Governance Mechanisms Perhaps the most effective way to manage the risks of loan delinquency and nonpayment may be for the lender to establish a governance mechanism through which it regularly reassesses a borrower’s financial ratios and makes decisions about ongoing access to the extended line of credit. The more frequent, more convenient, and more audited these fi- nancial reports, the better the governance mechanism will work. In essence, the bank be- comes almost a project financing partner for every major use of its funds in order to motivate the desired care, effort, and judgment from the borrower. In the end, a project-by-project financing approval process is precluded by the definition of the prob- lem and by the fact that bankers are not experts in all the businesses to whom they lend, but the closer a governance mechanism can come to this result, the less likely a default will be.

Hiring managerial talent involves a similar moral hazard problem because managerial effort is inherently unobservable. In particular, managers are paid for their creative ingenu- ity in proactively solving problems; they are paid to think hard about problems that haven’t happened yet. As such, managers can easily shirk their duties and instead devote their cre- ative ingenuity to activities unrelated to their job (so-called mental moonlighting). Because of this inability to directly observe managerial effort, we mistakenly blame managers at times for poor performance attributable really to nothing more than bad luck, and we fail to acknowledge managerial merit at times because we incorrectly assume that good perfor- mance is attributable to good luck. In the next section, we will see that “incentive contract- ing” with a minimum salary guarantee and a performance-based bonus (e.g., a stock option

7In other circumstances, lenders would face an adverse selection problem of detecting whether unknown bor- rowers are from a fraudulent or reliable subpopulation of loan applicants, and the offered terms of the loan will then affect the acceptance and refusals that determine the proportion of the loan portfolio originating from each group. Moderate interest rates and high collateral are intended in that situation to allow borrowers to signal their reliable intent to repay.

Chapter 15: Contracting, Governance, and Organizational Form 555

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

or restricted stock grant) can be an efficient solution to the moral hazard problem in man- agerial contracting.

There is another issue with incomplete managerial contracting, however, and that one relates to contract renewals. When executives have acquired unique company-specific knowledge and skills as well as pensions and severance packages not redeployable to others, they are in a favorable position in negotiating with Compensation Committees of the Board of Directors. The executive’s institutional memory is at least somewhat irreplaceable. Consequently, when it comes time to renew their contracts, senior execu- tives often engage in “holdups.”

Evidence of such holdups is plentiful: massive executive “loans” are often forgiven, incentive-based compensation lost in the face of poor performance is often reinstated or replaced through golden parachutes triggered by a hostile takeover, and option strike prices are often reset to lower levels in down markets. Considerable effort to retain expe- rienced senior managers is clearly value maximizing, but the Compensation Committee of the Board must be an independent body prepared to monitor, benchmark, and whistle blow on these contract renewals if necessary. Beyond holdup, outright deception and fraud remain a frequent problem. For example, backdating option grants to dates when the strike price was in the money are clearly criminal events, and the SEC is treating them as such. Table 15.2 provides a list of the implementation mechanisms of corporate governance available to address this holdup issue.

TABLE 15.2 IMPLEMENTATION MECHANISMS OF CORPORATE

GOVERNANCE

• Internal monitoring by independent board of director committees • Internal/external monitoring by large creditors • Internal/external monitoring by owners of large blocks of stock • Auditing and variance analysis • Internal benchmarking • Corporate culture of ethical duties • High employee morale supportive of whistle-blowers

WHAT WENT RIGHT • WHAT WENT WRONG

Moral Hazard and Holdup at Enron and WorldCom8

Misaccounting for short-term business expenses as long- term capital investments required a $3.8 billion restatement of lower operating profits at WorldCom in fiscal year 2000. Enron executives depleted pension reserve accounts while heralding the attractiveness of employee stock option plans for retirement planning. Business news of one scandal after another during the financial crisis of 2007–2009 made it abundantly clear that governance mechanisms are needed, but still the question remains, “Why, exactly?” Why don’t debt contracts of bond holders, personal loan contracts that senior executives use to relocate their households, and performance-based incentive contracts aligning owner and managerial interests prevent these abuses? Is it just that too

many gratuitous payments have been extracted from com- pensation committees, too many executive loans have been forgiven, and too many deferred stock options have been reset to lower strike prices when stock prices fell? That is, are the incentives in all this incentive contracting just misaligned? The answer is decisively “No.” The more funda- mental problem is that incomplete contracts invite post- contractual opportunistic behavior, requiring vigorous corporate governance mechanisms even in the presence of strong incentives.

8Based on “Taken for a Ride,” The Economist (July 13, 2002), p. 64; and “WorldCom Aide Conceded Flaws,” Wall Street Journal (June 16, 2002), p. A3.

556 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

THE PRINCIPAL-AGENT MODEL Many types of owner-principals hire manager-agents to stand in and conduct their busi- ness affairs. Parent companies set up principal-agent relationships with subsidiaries. Manufacturer principals employ retail distributor and advertising agents. And most im- portantly, equity owners hire executives with managerial incentive contracts. The owners’ objective in such principal-agent relationships is to offer incentives to managers to forego alternative employment opportunities and to act in a value-maximizing manner on be- half of the owner-principals.

The Efficiency of Alternative Hiring Arrangements Managerial hiring contracts may take on several pure or hybrid forms, including straight salary, wage rate, or profit sharing. In straight salary contracts, the manager and firm agree on a total compensation package and specific conditions of employment. In other contexts, such as consulting, the managerial consultant may receive an hourly wage rate equal to the best alternative employment opportunity in the competitive labor market for his or her type of consulting services. In Figure 15.3, the managerial consultant is hired for, say, 50 hours per week at wage rate Wa. Dl is the firm’s input demand, which is the marginal revenue product of these labor services—namely, the marginal output of addi- tional hours times the marginal revenue from selling that additional output. Because each firm is atomistic in the labor market for these management consulting services, Sl is the perfectly elastic supply facing any given employer at the going market wage. Be- yond 50 hours, the declining Dl no longer exceeds the incremental input cost along Sl.

Managers also may secure employment under a pure profit-sharing contract. Like pure commission-based salespeople or manufacturer’s trade representatives, the manager may accept a percentage (say 40 percent) of the receipts directly attributable to his or her efforts in lieu of wage or salary income. Think of the percentage finder’s fee sometimes offered for cost-saving suggestions in big corporations or the federal government. Again in Figure 15.3, we can represent this third alternative hiring arrangement as the ray AB, wherein the manager receives 40 percent of the owner’s willingness to pay for each hour

FIGURE 15.3 Alternative Managerial Labor Contracts

F

O D

10

E

Dl G

A

C

B

U

$

I J

22 50

Sl � Wa

Hours

H

0

Chapter 15: Contracting, Governance, and Organizational Form 557

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

of management services. Initially, this profit share will exceed the wage rate alternative. For example, during the first 22 hours of work, the profit-sharing contract will overcom- pensate by area ADJ (shaded area O). Thereafter the profit share falls below the man- ager’s market wage rate per hour.

If 40 percent proves to be an equilibrium profit share, the overcompensation (shaded area ADJ labeled O) will just equal the undercompensation for the last 28 hours of work (shaded area DCF labeled U). This level leaves both the owners and the manager indif- ferent between this hiring arrangement and the alternative 50-hour-per-week wage rate contract at Wa. If the profit share were reduced to, say, 35 percent (represented by the ray IB), the dark-shaded amount of overcompensation for the first 10 hours would fail to offset the massive undercompensation for hours 10 to 50. The manager would then reject the profit-sharing contract in favor of the wage rate offer. By raising the profit share back to 40 percent, the firm appears able to restore the attractiveness of each contract, at least for certain types of workers. In reality, as we shall now see, the actual situation in hiring managerial talent is often rather different.

Work Effort, Creative Ingenuity, and the Moral Hazard Problem in Managerial Contracting Pure profit-sharing contracts contain the seeds of their own destruction. Suppose several individuals are involved in generating pharmaceutical sales, and the input that the profit sharer contributes to team production is largely unobservable. No time card can success- fully monitor the input, perhaps because a measure of work effort rather than work hours is really what is required. The rational employee then considers his or her alterna- tives. As long as the profit-sharing compensation exceeds the alternative wage rate, he or she dedicates unobservable work effort to this job. Beyond 22 hours of work effort, how- ever, the employee can earn more by working for someone else at the alternative wage rate Wa. Therefore, the disloyal (but rational) trade representative underworks the terri- tory; he or she moonlights. This predictable response is another aspect of the moral haz- ard problem. Only a moral sense of duty to one’s employer prevents this problem from becoming a real hazard to the business.

Predicting such behavior, the employer may decide to withdraw the offer of a pure profit-sharing contract. Let’s see why. If the territory is under worked by 28 hours, the em- ployer saves profit-sharing payments equal to area DFGH in Figure 15.3 but loses output valued at the much larger ECGH and therefore is out the net value [ECGH – DFGH – ADJ (the overpayment area for the first 22 hours of work), all of which is equal to area EDC] relative to a wage contract that just paid piece rates for 50 hours of work at an implicit wage rate of Wa per hour. The fact that work effort is largely unobservable makes the pure profit-sharing contract unattractive to the employer relative to a piece-rate contract.

This generalization is not always true. For example, in hiring attendants for parking garages, the time clock and customer complaints (e.g., horn blowing and broken parking gate barriers) monitor the required input quite well. A dismissal policy in the employ- ment contract making time-on-task a condition of employment motivates the required input. Similar time-on-task constraints and output quotas are employed in hiring share- croppers and retail sales clerks. In these instances, the firms and their employees have evolved ways to resolve the moral hazard problem. Nobel laureate Ronald Coase empha- sized that private voluntary bargaining between principals and agents will often find ways to contract around such problems.9

9See J. Farrell, “Information and the Coase Theorem,” Journal of Economic Perspectives (Fall 1987), pp. 113–129.

558 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The problem of moral hazard arises then only when an employee’s action such as work effort is directly observable only at a prohibitive cost. Consider again the pharma- ceutical sales representative for whom appointment logbooks and random follow-up monitoring simply cannot detect the persuasive effort necessary to secure orders from physician customers. One could trail around after the sales representative and interview each physician after the sales calls were completed to try to detect the ingenuity and per- severance the sales rep displayed, but quite obviously, this monitoring practice would be prohibitively expensive. Instead, in the face of truly “hidden actions,” the pharmaceutical company is more likely to favor some performance-based incentive contract involving benchmarking rather than the pure profit-sharing contract.

During a period of benchmarking, the employer reassigns previously low productivity sales territories to above-average trade representatives to see whether their effort can al- ter the success rate per sales call. If so, the employer concludes that lack of effort by prior sales representatives was responsible for the low sales. After several such benchmarkings, the employer is able to identify those sales representatives to be kept and those to be dismissed. Importantly, the “keepers” are then allowed to retain all the productive ac- counts they have developed.

For managerial jobs, however, the moral hazard problem is significantly harder to re- solve. The input senior management contributes to team production is not time on task at the desk, but rather what we have called creative ingenuity: creatively formulating and solving problems that may not even have arisen as yet. Managers are paid to think, not to shuffle papers. The difficulty is in detecting when creative ingenuity is being applied to the employer’s business, rather than another business for which the manager may be mentally moonlighting. Of course, eventually the difference will show up in performance, but over how long a period and how big a difference? These questions are tough to an- swer satisfactorily after a senior manager has shirked his or her duties to the detriment of shareholder value and been let go.

More problematically, the shirking manager may never be let go, and the hard- working manager may never be rewarded. If random disturbances affect the company’s performance, it is difficult even after the fact to separate unobservable shirking from neg- ative random disturbances. How, then, are owners to know when to blame senior man- agers for downturns in company performance and when to give them credit for upturns? One governance mechanism often used to analyze these variances is the company audit. Managers report on the sources and uses of funds in accordance with generally accepted accounting principles (GAAP). Any period-to-period variances are then assessed and verified by independent auditors.10 Despite dedicated efforts, auditors can seldom sepa- rate the effects of management decisions from random disturbances in company perfor- mance. That is, the moral hazard problem is much harder to solve when combined with the performance uncertainty most firms face.

Some companies address the problem of managerial moonlighting by benchmarking one manager against another (say, in comparable plants or geographic divisions). They hope that the effects of business cycle factors and random time-series disturbances will be highly correlated across plants and divisions, and that the manager’s effort and crea- tive ingenuity will therefore correspond with the plant or division’s differential perfor- mance. Unfortunately, they are usually wrong. As a result, Japanese companies rely on intense loyalty-building exercises, peer pressure, and lifetime employment contracts (fol- lowing some probationary period) to reduce mental moonlighting and other forms of shirking.

benchmarking A comparison of performance in similar jobs, firms, plants, divisions, and so forth.

company audit A governance mechanism for separating random disturbances from variation in unobservable effort.

10This audit mechanism is explored at greater length at the end of the chapter in the first Case Exercise.

Chapter 15: Contracting, Governance, and Organizational Form 559

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

In isolation, neither hidden effort (unobservability) nor performance uncertainty poses any special difficulty for owner-principals hiring manager-agents. The moral hazard result- ing from the unobservability of a manager’s input is, by itself, resolvable by assigning the manager lagged residual income claims (e.g., deferred stock options or restricted stock). Set- tling up ex post with a manager, after all the effects of his or her effort and ingenuity have had time to influence performance, creates more effective performance-based incentives.

Example Indexed Stock Options at Adobe Systems, Dell, and Cisco11

To align managerial incentives with equity owner interests, most companies regu- larly award deferred stock options to their managers. These performance-based bo- nuses entitle the holder to purchase company stock at a slight discount to its current value. If the firm’s performance subsequently improves, capitalized value rises and both shareholders and the managers stand to gain. To exercise their op- tions, managers often must wait three to five years, but they sometimes realize gains of 50 to 80 percent or more. In 2009, CEO compensation in Fortune 500 companies averaged $7 million, 71 percent of which involved deferred-stock or options-based compensation for superior performance.

To acquire the stock for these deferred compensation programs, some compa- nies dilute equity by issuing new shares, while others repurchase shares on the open market. To reduce the cost of these “buybacks,” especially in a rising market, some companies, such as Adobe Systems, Dell, and Cisco, index the exercise price for their deferred options to the average stock price of their industry. When all the related companies do well, the value of the option rises, but so does the exercise price. As a result, the managers do not exercise their options at that juncture and instead are motivated to outperform their peers in related companies in both good times and bad.

11Based on “Stock Options That Don’t Reward Laggards,” Wall Street Journal (March 30, 1999), p. A26; “Corporate America Faces Declining Value of Options,” Wall Street Journal (October 16, 2000), p. A1; “The Gravy Train Just Got Derailed,” BusinessWeek (November 19, 2001), p. 118; and “CEO Compensation Survey,” Wall Street Journal (April 11, 2005), p. R1.

Example P&G Pays Ad Executives Based on Their Performance Procter & Gamble places more than $3 billion per year in advertising through Saatchi & Saatchi, Leo Burnett, and other ad agencies. Traditionally, agencies earned flat-rate fees assessed as 15 percent of the ad dollars expended for the client. In the 1990s, Ford, Colgate-Palmolive, and P&G broke out of this flat-rate system and began paying a baseline fixed fee plus a performance bonus. Now, account ex- ecutives at the agencies earn a fixed salary if their creative communications are less than compelling and P&G sales stay flat. On the other hand, a hugely successful ad campaign can earn multimillion-dollar bonuses if P&G’s sales growth can be at- tributed to the advertising. Both the clients, the ad agency owners, and the account execs now share in the risks of consumer whimsy, but a base salary provides a safety net should random misfortune occur.

560 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Similarly, performance uncertainty taken alone creates a risk-allocation problem that may be easily resolved with insurance. Managers are somewhat less able to diversify than owners because of the specific human capital the manager often invests in a long-term relationship with his or her corporate employer. This situation results in risk-averse owners and risk-averse managers frequently structuring some sort of risk-sharing agree- ment, such as a guaranteed baseline salary combined with a performance-based bonus.

Formalizing the Principal-Agent Problem The real difficulty for managerial contracting arises then when the unobservability of management’s creative ingenuity and performance uncertainty occur simultaneously. The coexistence of these problems constitutes the so-called principal-agent problem most firms face. Settling up ex post facto with management teams then no longer creates the desired incentives. Some managers get unlucky and receive blame they did not de- serve, and others get lucky and receive credit they did not earn.

The principal-agent problem can be formalized as an optimization problem subject to dual constraints. The principal chooses a profit-sharing rate and a manager’s salary guar- antee to maximize the expected utility of the risk-averse owner-principals’ profit where profit depends on the manager-agent’s effort, on the cost of the managerial incentives contract, and on random disturbances. An incentive compatibility constraint then aligns the effort chosen by the manager in response to the share and salary offer with the effort that maximizes the expected utility of the owner-principals. That is, an incentive-compatible profit share and salary elicit the managerial effort and creative in- genuity required to maximize the owner’s value. Third and finally, the participation constraint ensures that the manager will reject his or her next best offer of alternative employment.

In the next section, we illustrate the meaning of each of these three elements with a linear optimal incentives contract, which can solve the principal-agent problem. How- ever, understand that an optimal managerial incentives contract is easier to describe than to attain.12 See the Case Exercise at the end of this chapter to try your hand at bal- ancing all the competing objectives involved.

Screening and Sorting Managerial Talent with Optimal Incentives Contracts Asymmetric information arises in all hiring decisions, but it often plays an especially prominent role in managerial hiring decisions. Job applicants know all the information, but potential employers have access only to the information that applicants select for their résumés. Thus, among perhaps 19 résumé facts that the personnel department might like to know, the applicant discloses only 14. Let’s see how linear share contracts can be used to sort managerial talent based on one of these undisclosed characteristics— namely, a manager’s risk aversion.

Suppose a large bank has two openings for which it desires managers of different risk aversion. One position is the assistant vice president for commercial construction loans in a city with overbuilt office developments and, consequently, very high vacancy rates.

principal-agent problem An incentives conflict in delegating decision-making authority.

incentive compatibility constraint An assurance of incentive alignment.

participation constraint An assurance of ongoing involvement.

12Even the solution we describe is limited to separable functions in effort and money income. The general principal-agent problem with risk-averse owners and managers has multiple solutions and requires nonlinear incentive contracts relating salary and profit share to company performance. See J. Tirole, The Theory of In- dustrial Organization (Cambridge, MA: MIT Press, 1988), pp. 35–54; and D. Kreps, A Course in Microeco- nomic Theory (Princeton, NJ: Princeton University Press, 1990), Chapter 16.

Chapter 15: Contracting, Governance, and Organizational Form 561

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The other position is an assistant vice president to manage the venture capital loan port- folio, to interact with owners of new start-up businesses, and to represent the bank at the entrepreneurship club of the city. As you might suspect, the bank has two rather differ- ent people in mind as ideal candidates for these openings. In the commercial construc- tion area, the bank seeks an instinctively cautious and safety-conscious manager who will take every opportunity to reduce the large default risk already present in this portion of the bank’s business. Both jobs are simply listed as assistant vice president positions; no further details are given.

Two managers with the requisite training and experience apply for the bank positions. Their résumés are similar. However, unbeknownst to the bank, one drives an old Porsche, not insured against collision damage, and has skydiving as an undisclosed for- mer hobby. Instead of skydiving, this individual (let’s call him Dashing) now prefers bungee jumping, which he understandably decides would be inappropriate to list as a hobby on an application for a bank job. The other individual (you guessed it, Smooth) drives a dealer-serviced Land Rover on which she carries the maximum auto insurance coverage. Despite never leaving town, Smooth keeps the Rover in four-wheel drive at all

WHAT WENT RIGHT • WHAT WENT WRONG

Why Have Restricted Stock Grants Replaced Executive Stock Options at Microsoft?13

In 2004, Microsoft announced that it would use approxi- mately $30 billion of its $60 billion in cash and short-term investments to buy back stock in 2004–2008. Between 1995 and 2005, the use of restricted stock grew from 18 percent of CEO compensation to 22 percent. But Microsoft went further. Microsoft joined numerous other companies in granting restricted stock rather than stock options as a performance-based bonus to more than 10,000 of its em- ployees. Why change the Microsoft performance bonuses?

Several reasons are behind this change. First, restricted stock cannot be sold if an executive leaves the company. In contrast, once stock options vest, they are typically sold. During the heyday of the booming information economy in the late 1990s, Microsoft’s options created literally thou- sands of multimillionaires among Microsoft’s senior man- agers. Too many of these valuable human resources simply chose to retire early and move on to other pursuits. One former executive took up professional bowling. Another senior executive, Paul Allen, bought a basketball team and built a guitar museum. The change of compensation contracts to feature restricted stock is expected to enhance the retention of pivotal employees relative to granting options.

A second reason a Board of Directors prefers restricted stock is that senior managers are often able to negotiate option features that work against optimal incentive con- tracting. For example, few option exercise prices are in- dexed relative to the industry group or strategic competitors. So stock options reward mediocrity when all

share prices in an industry rise together. Third, many se- nior executive options are “reloaded.” As soon as the op- tion contracts vest (in 2–5 years) and are exercised, senior managers negotiate the issue of new options with new exercise prices but the old expiration date (of typically 10 years). This clause allows executives to profit from in- duced volatility in their share prices. They have a powerful incentive to “swing for the fences” with high-risk projects whenever options in the money can be converted to cash at an intermediate date and then replaced with new options expiring at the original dates.

Finally, few stock option contracts restrict in any way the executive’s ability to “unwind” his or her risk exposure by hedging the risks the options create. Selling short a long option position that ties the manager’s wealth to that of the shareholders is hardly justifiable given that the objective of performance-based pay is the aligning of managerial in- centives with shareholder interests. But the hedge fund scandal at Goldman Sachs in April 2010 suggests that such short selling of customer positions occurs on Wall Street.

Corporations should be determined to prevent the practice on Main Street. Restricted stock has none of these drawbacks.

13“Microsoft Ushers out Era of Options,” Wall Street Journal (July 9, 2003), p. A1; L. Bebchuk et al., “Managerial Power and Rent Extraction in the Design of Executive Compensation,” University of Chicago Law Review 69 (2002), pp. 751–846; B. Hall and K. Murphy, “The Trouble with Stock Options,” Journal of Economic Perspectives (Summer 2003); “Finance 2.0: An Interview with Microsoft’s CFO,” The McKinsey Quarterly, no. 1 (2005), “CEOs Get Paid,” Knowledge Wharton (May 3, 2006), p. 3.

562 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

times to secure the extra traction. Once while attending a formal cocktail party with a few close friends, Smooth revealed that she spent her Christmas bonus on “more insur- ance, of course.” Thinking none of these details to be of any real significance to the bank, she too omits that information from the job application.

The bank’s problem is to sort these two types of individuals, both of whom are well qualified, into the jobs appropriate for their different risk aversions. In Figure 15.4, we display the guaranteed base salary and profit-sharing rate, the two components of a linear incentives contract. On the horizontal axis are various percentages that represent what additions to or subtractions from one’s pay occur as a result of the profit-sharing agreement. A greater share rate initially elicits more effort and creative ingenuity and re- sults in greater expected profit contribution from the manager’s activities. Eventually, at still higher share rates, the profit contribution actually declines. The two hill-shaped curves in Figure 15.4 represent expected profit-sharing payouts that would allow the firm to just break even on its incentive payments to the two managers. The lower ex- pected profit curve corresponds to the commercial construction loans job, and the higher curve corresponds to the venture capital loans job.

Let’s suppose that the company has expected sales of $100 million and a net cash flow from sales of 12 percent. Hence, expected profits available for distribution to owners and managers run $12 million. The bank first elicits responses to two tentative contract offers for their assistant vice president jobs. Contract A offers $48,000 salary plus or minus 0.1 percent of the net cash flow or $12,000, which implies a $36,000 to $60,000 possible range of income outcomes. Contract B offers $96,000 salary plus or minus 0.2 percent, or a $72,000 to $120,000 range. Contracts A and B are not equally attractive. One dom- inates the other in that minimum outcomes under Contract B exceed the maximum out- comes under Contract A. Because risk increases only modestly from 0.1 percent to 0.2 percent, both prospective employees are likely to select B, and therefore, contract B is said to result in a pooling equilibrium. Such an outcome is illustrated in Figure 15.4, where the indifference curves for both Smooth (IS) and Dashing (ID) indicate that Con- tract B would be preferred by both applicants.

FIGURE 15.4 Sorting Managers with Linear Share Contracts

0

Profit-sharing rate

96

C

B

A48 55

70 E

D

IS

ID

B as

e sa

la ry

( th

ou sa

nd s

of d

ol la

rs )

�0.3% �0.4% �$12(000) �$24(000)

�0.2%�0.1%

More risk averse

Less risk averse

Venture capital personal

banker

Commercial construction loan officer

linear incentives contract A linear combination of salary and profit sharing intended to align incentives.

pooling equilibrium A decision setting that elicits indistinguishable behavior.

Chapter 15: Contracting, Governance, and Organizational Form 563

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

To separate the two applicants according to their risk aversion, the bank withdraws Contract B and instead introduces a more demanding risk-return trade-off. To see how this would work, we begin by indicating on Figure 15.4 the indifference curve for the more risk-averse applicant, Smooth (IS), that establishes a line of demarcation between contracts to the northwest, which are more preferred than A, and those to the southeast, which are less preferred. To induce a revelation of the risk-aversion differences between Smooth and Dashing, the bank then offers a contract pair such as A and C. Contract C imposes a much larger 0.4 percent profit share and offers only $55,000 expected base salary, just $7,000 above Contract A. Smooth finds C much less preferred than the original ContractA and immediately says so. In contrast, less risk-averse Dashing (ID) is so close to being risk neutral (i.e., with almost flat indifference curves between expected salary and profit share), that Dashing may well prefer Contract C. This separating equilibrium in which Smooth reveals her stronger risk aversion by rejecting C in favor of A, whereas Dashing does just the reverse, attains several of the employer’s objectives. First, these profit-sharing contracts are incentive-compatible contracts in that they elicit appropriate effort and creative ingenuity from both managers while maximizing the shareholders’ value. Second, these profit-sharing contracts sort Dashing as the manager to head up the venture capital personal banking group and Smooth as the bank manager for the commercial construction group. The reductions in expected profit to compensate each manager would have been larger without this matching of risk aversion and job types.

However, one aspect of an optimal incentives contract remains unaddressed. The par- ticipation constraint has not yet been satisfied. An alternative employer can offer Con- tract D, which attracts Smooth with both more expected salary and lower profit risk while retaining the separating properties of the (A,C) contract pair. As long as such im- provements in both risk and return are possible, Smooth will continue to resign and move. Only with contract pair (A,E) will both the incentive compatibility and participa- tion constraints be satisfied; Dashing selects Contract E while Smooth selects Contract A, and both remain in the bank’s employ. In addition to resolving the sorting of managerial talent, if it induces the appropriate effort from both managers and prevents their being bid away to alternative employment opportunities, this linear incentives contract consti- tutes a solution to the principal-agent problem in managerial contracting.

CHOOSING THE EFFICIENT ORGANIZATIONAL FORM Ultimately, the choice of organizational form (e.g., spot market transactions, relational contracting, franchise profit sharing, or vertical integration) depends upon what best suits the governance needs of the asset owners involved. Assets can be non-redeployable, specific assets with little or no value in second-best use, such as re- mote plants and work-in-process inventories, or they may be nonspecific investments that are fully redeployable, such as corporate jet aircraft, popular finished goods in in- ventory, and Treasury cash. In addition, some assets are dependent on unique comple- mentary investments (e.g., specially designed computer hardware and closed-architecture software), while others are not (e.g., a flexible power plant designed to run on natural gas, coal, or biodiesel). One classic example of these asset-characteristic dichotomies is a hot-rolled steel-making plant belonging to companies such as Republic and U.S. Steel, which requires a blast furnace, converter, reduction furnace, and rolling mill. Because melting the ore and then the metal requires substantial energy, such plants are located beside one another to avoid the reheating expense. However, the organization form

separating equilibrium A decision setting that elicits distinguishable behavior.

non-redeployable, specific assets Assets whose replacement cost basis for value is substantially greater than their liquidation value.

564 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

question is not whether the operations will be physically integrated and in adjacent loca- tions, but rather whether they will be jointly or separately owned and managed.

At one end of the spectrum, spot market recontracting is efficient for fully redeploy- able durable assets, not dependent upon other complementary assets. Rental cars provide a good illustration of such assets that may be allocated through spot markets with no loss of efficiency, as shown at the top left of Table 15.3. The limiting feature of this organiza- tional form, however, is the potential for “holdup” inherent in the frequent renewal of spot market contracts. If one party has non-redeployable assets (e.g., a steel mill’s blast furnace or a stadium of a sports franchise), spot market recontracting provides too many opportunities for unique engineers or star players to appropriate the surplus value in any business relationship. Owners of non-redeployable assets wish to avoid this holdup haz- ard by securing longer-term supply contracts.

Reliant assets are non-redeployable durable assets that when resold must be sold in thin markets for less than their value in first-best use. These assets are highly specific to their current use because of substantial unrecoverable sunk cost investments either in acquisition, distribution, or promotion. Specialized equipment in remote locations such as a bauxite mine is the most common reliant asset. Where reliant assets are dependent on unique complements (like an aluminum plant) in order to achieve any substantial value added, one has the maximum potential for holdup in spot market recontracting. In addition to all the bargaining costs such a situation engenders, these “all- eggs-in-one-basket” ventures tend to receive underinvestment relative to what would be optimal.

Dependency relations between assets may be either one way or bilateral. Manufac- turers with independent distributors are a good example of a bilateral dependent rela- tionship involving reliant assets. Each party in a Volvo-GM Heavy Truck Corporation manufacturer/distributor relationship is equally dependent on the other. In such cases, independent dealers often gravitate toward vertical requirements contracts with a fixed profit share, as shown in the bottom right corner of Table 15.3.

When assets are dependent on unique complements but not reliant because of their substantial redeployability, the parties often adopt long-term performance-based relational contracts. Redeployable corporate jets and pilots provide a good illustration. Pilots need not own the planes or secure fixed profit-share contracts to operate the planes. Instead, as indicated in Table 15.3, the efficient organizational form of a jet char- ter company is normally one of long-term relational contracts with standby pilots who report on short notice for piecemeal assignments. This alliance works well, and both the pilots and plane owners understand that the longevity and reliability of the relation- ship enhances value relative to spot market recontracting.

Pepsi and Starbucks entered into an alliance to sell cold Frappuccino through soft drink machines. Starbucks’ assets in this alliance are redeployable but one-way

TABLE 15.3 EFFICIENT ORGANIZATIONAL FORM DEPENDS ON ASSET

CHARACTERISTICS

FULLY REDEPLOYABLE DURABLE ASSETS

NON-REDEPLOYABLE RELIANT ASSETS

Not dependent on unique complements

Spot market recontracting Long-term supply contracts + risk management

One-way dependent assets Relational contracts (alliances) Vertical integration

Bilateral dependent assets Relational contracts (joint ventures) Fixed profit-sharing contracts

reliant assets At least partially non-redeployable durable assets.

relational contracts Promissory agreements of coordinated performance among owners of highly interdependent assets.

Chapter 15: Contracting, Governance, and Organizational Form 565

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

dependent on any of several companies with tens of thousands of drink machines. So, again referring to Table 15.3, a Starbucks–Pepsi alliance is the efficient organizational form. To take a related example, Genentech’s biotechnology is fully redeployable but one-way dependent on marketing behemoth partners such as Pfizer or GlaxoSmithKline. As a result, Genentech has consistently entered into 10 marketing partnerships, 20 licensing agreements, and numerous product development alliances with large drug makers.

Example Schwinn and Sylvania Dealers Offered Exclusive Territories14

The Schwinn and Sylvania companies sold their bicycles and TVs through authorized dealers who were prohibited from reselling to unauthorized bike shops and electron- ics stores. Although the dealers could carry other product lines, the resale restriction provided dealers with an exclusive territory. Prosecutors in the U.S. Justice Depart- ment saw this vertical territorial restriction as anticompetitive, not as a transaction cost-reducing contractual alternative to the governance mechanisms that would oth- erwise be necessary to protect Schwinn and Sylvania’s brand-name capital.

Fortunately for Sylvania, in Continental T.V., Inc., et al. v. GTE Sylvania Inc., 433 U.S. 36 (1977), the Supreme Court disagreed. In this case, the Court explicitly recog- nized that brand-name capital is a specialized, non-redeployable asset that would be compromised if third-party dealers were allowed to sell the Sylvania product with unauthorized marketing plans. Exclusive territories therefore became subject to a rule of reason analysis that weighs a legitimate manufacturer interest in presale sell- ing effort and post-sale services against an illegal attempt to prevent reductions in the manufacturer’s suggested retail price. Sylvania’s fixed profit-sharing contracts with its dealers when combined with exclusive territories were seen as an efficient and legal organizational form.

14See S. Dutta, J. Heide, and M. Bergen, “Vertical Territorial Restrictions and Public Policy,” Journal of Marketing 63 (October 1999), pp. 121–134.

Example Kodak and Time Warner Form a Digital Photography Alliance When Kodak built a fully integrated, five-step digital photography solution of im- age capture, editing, processing, distribution, and storage Web site, it clearly re- quired a phone or cable company to provide broadband services. However, just as clearly, Kodak’s Web site was redeployable to more than one broadband provider. Consequently, although Kodak was dependent on broadband complementors, it foresaw little contractual hazard of ex post holdup by any of them.

In addition to online delivery of digital prints from film submitted to one of its 30,000 retail developing locations, Kodak believes that it will soon be one-way de- pendent on an online partner as customers increasingly edit, store, share photos, and order reprints over the Internet. Of course, Time Warner is fully redeployable to many other uses. Because Kodak is one-way dependent but not reliant upon Time Warner, and because Time Warner is neither dependent nor reliant on Kodak, a relational contract to establish a Kodak–Time Warner alliance to dis- tribute digital photography online is the efficient organizational form.

566 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

In contrast, consider the bilateral dependency of a relational contract between a PC assembler and a chip supplier. Without specially designed Motorola computer chips, Apple iMacs have little value, and without Apple iMacs, these Motorola chips have little value. Yet, each manufacturer makes reliance investments that are specific to the other partner’s design decisions. Hence, as indicated in the bottom left cell of Table 15.3, a joint venture is the efficient organizational form. The term joint ventures is often reserved for bilateral relationships that establish a separate corporate legal identity. Appendix 15A addresses incentive-compatible revelationmechanisms for eliciting asymmetric cost information from joint ventures partners.

Finally, when reliant assets are one-way dependent on unique complementary re- sources, the most efficient organizational form is vertical integration. Remote aluminum plants are one-way dependent on nearby bauxite mines. In contrast, because the bauxite can be shipped anywhere, the mine owners are not dependent on the local aluminum plant. Both assets entail substantial sunk cost investment, but only the remote aluminum plant is a non-redeployable durable asset (i.e., one with little value to other companies should the nearby bauxite source no longer be available). In this situation, upstream

WHAT WENT RIGHT • WHAT WENT WRONG

Cable Allies Refuse to Adopt Microsoft’s WebTV as an Industry Standard15

Demand for interactive television with Internet surfing, Web shopping, interactive sports, and e-mail has grown quickly in hotel and airport lounges but slowly elsewhere. Cable companies such as AT&T and Time Warner ap- pear most likely to trigger the adoption of these smart TVs in households through their leasing of set-top con- trol boxes to residential customers. After acquiring WebTV (now renamed MSN TV) for $425 million in 1997, Microsoft shifted to an alliance strategy to secure the adoption of its complex software by the cable TV operators. Microsoft’s interactive WebTV product known as UltimateTV was fully redeployable across competing cable service companies, and the cable com- panies sought to remain fully redeployable across inter- active TV software providers. Because Microsoft/ WebTV was one-way dependent on cable providers, but cable had numerous other ways to generate value without Microsoft, an alliance was the efficient organizational form for these asset characteristics.

Microsoft’s product offering required the cumbersome software architecture of Windows CE. Standard set-top control boxes don’t have enough memory or fast enough microprocessors to support Microsoft’s operating system. Consequently, Microsoft invested more than $10 billion in co-designing digital entertainment networks and new set-top control boxes with seven cable companies world- wide (AT&T: $5 billion; Telewest Comm in the United Kingdom: $2.6 billion; Comcast: $1 billion; and another $1.2 billion in Rogers, NTL, and UPC in Canada and

Europe). In return, AT&T Broadband and its subsidiary TCI promised forward sales contracts for a total of 10 mil- lion set-top control boxes employing Microsoft CE soft- ware in a Motorola-built unit, the DCT5000. Today, the first 250,000 DCT5000s sit idly stacked in a Seattle ware- house. Microsoft’s software was simply too complex, too costly, and too late.

The full installation costs for the Microsoft DCT5000 cable networks skyrocketed to $500 per household control box. Yet marketing research showed that cable subscribers would willingly add only $5 per month to their cable bills in order to secure these enhanced services. Ongoing delays induced Europe’s largest cable company, UPC, to order set-top digital entertainment software from Liberate, a Mi- crosoft rival. AT&T then announced that it had no plans to deploy interactive WebTV software and that Microsoft would build only the replacement for the scrolling online TV Guide.

Had the cable companies allowed Microsoft/WebTV to become an industry standard, full-scale vertical integration would have been warranted. Microsoft’s digital entertain- ment assets would then have been one-way dependent on cable service providers, whose assets would have been no longer redeployable. Once when Bill Gates was presenting UltimateTV as a possible industry standard, a cable com- pany president, Brian Roberts, half-jokingly suggested that Microsoft buy the entire cable industry.

15Based on “Microsoft’s Blank Screen,” The Economist (September 16, 2000), p. 74; and “Set-Top Setback: Microsoft Miscues,” Wall Street Journal (June 14, 2002), p. A1.

Chapter 15: Contracting, Governance, and Organizational Form 567

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

vertical integration by the aluminum manufacturer is required in order to prevent opportunistic holdup by the bauxite mine owners.

Sometimes capitalized value is one-way dependent on a unique downstream comple- mentary firm. The huge success eBay has had in attracting sellers of one-of-a-kind items to their network of auction buyers necessitated an electronic payments platform. By 2002, PayPal had captured that market, signing up 12 million households and 3 million businesses (28,000 new users a day) who wished to authorize credit card charges with nothing more exchanged than an e-mail address and a charge amount. At one point, 61 percent of 30 million settlements on PayPal were transactions generated off eBay’s 150 million postings. Consequently, in July 2002, eBay paid $1.4 billion to acquire Pay- Pal and thereby vertically integrate downstream. PayPal had become a unique comple- ment to non-redeployable eBay assets, and, as indicated in Table 15.3, vertical integration was the efficient organizational form.

Prospect Theory Motivates Full-Line Forcing At times the span of hierarchical control (i.e., the boundary of the firm) is determined by the most effective marketing. Social psychologists have long noticed that people buy both lottery tickets and disability insurance. Daniel Kahneman and Amos Tversky hy- pothesized that people are risk preferring at wealth levels below their current socioeco- nomic position and risk averse above (as in Figure 15.5). That is, their absolute value of

INTERNATIONAL PERSPECTIVES

Economies of Scale and International Joint Ventures in Chip Making16

Approximately one dozen large electronics companies in the United States, Europe, and Japan were once involved in producing memory chips. Up-front costs for each item were staggering. For example, the cost of developing the 64-megabit memory chip design and production technology was estimated to range from $600 million to $1 billion. Once developed, memory chips then required investment of an additional $600 million to $750 million in a “fab,” a fabrication plant that produced up to 10 million chips a month.

Because of the asset characteristics and the massive scale economies available in such a cost structure, many of the semiconductor companies involved in these re- search and development efforts formed international joint ventures (JVs) to share the huge fixed costs and risks involved. Some of these partnerships include:

U.S. COMPANY FOREIGN PARTNER

AT&T NEC (Japan)

Texas Instruments Hitachi (Japan)

Motorola Toshiba (Japan)

IBM Siemens (Germany)

Initially these joint ventures took various forms. For example, AT&T and NEC agreed to swap basic chip-making technologies. Their redeployable trade secrets were each enhanced by unique complemen- tary knowledge of the joint venture partner, consis- tent with Table 15.3. Similarly, Texas Instruments and Hitachi agreed to develop a common design and manufacturing process and then do low- volume production together, but mass production and marketing remained separate. Joint marketing was not redeployable to other companies but did require the coordination of a fixed profit-sharing agreement. Again, see Table 15.3.

Finally, a Motorola and Toshiba JV consolidated the production of millions of memory chips per month to realize the potential for massive scale econ- omies in their codependent and fully redeployable outsourcing.

16Based on “The Costly Race Chipmakers Cannot Afford to Lose,” Business- Week (December 10, 1990), pp. 185–187; and “Two Makers of Microchips Broaden Ties,” Wall Street Journal (November 21, 1991), p. 84.

568 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

utility losses is perceived as greater, often much greater, than an equal-dollar-valued utility gain.17

In a product category such as minivans, their observation is that a household contem- plating a Caravan SE model with full options for $22,000 will perceive the decline in sat- isfaction from dropping down to a $17,000 base model Caravan as much greater than the increase in satisfaction associated with giving up another $5,000 to purchase a $27,000 Caravan LX or SXT model. The fact that perceived losses outweigh equal-value prospective gains reflects a theorem of prospect theory that has many implications for the optimal boundaries of the firm.

For one thing, marketers are well advised to distribute trial products (“try now, pay later”) because the utility loss if the consumer contemplates returning the product and owes nothing will be greater than the utility gained by prospective additional consump- tion with the money saved. Second, receivables managers are well advised to ask consu- mers to forgo something prospective to pay for their product (i.e., a projected year-end bonus, tax refund, or frequent flyer award). Again, the payment in prospective additional consumption forgone will be perceived as less undesirable than an equivalent cash ex- pense that means giving up other current consumption.

Finally, firms with good-quality house brands should encourage premium brands in the channel, and those with premium brands should encourage the introduction of super-premium brands. If no such distribution channel arrangements can be secured, the implication is that firms selling moderately priced brands should buy premium brands. If none exists, firms are advised to develop their own and force them into the channel. Hanes Hosiery pursues a “good, better, best” product strategy with private la- bels, L’eggs, and Hanes. Marriott developed its Fairfield Inns, Courtyard by Marriott, and Marriott Resorts and International Hotels, and now Ritz Carlton. The Gap launched

FIGURE 15.5 Full-Line Forcing

Sa ti

sf ac

ti on

o r

pe rc

ei ve

d va

lu e

(u ti

le s)

Monetary value $17,000 $27,000 $32,000 $38,000

Luxury import

Town and

Country minivan

SXT

SE with full options

Base SE

�160

�80

100

140 160

$22,000

17D. Kahneman and A. Tversky, “Prospect Theory: An Analysis of Decision under Risk,” Econometrica 47, no. 2 (March 1979), pp. 263–291; and C. Camerer, “An Experimental Test of Several Generalized Utility The- ories,” Journal of Risk and Uncertainty 2, no. 1 (April 1989), pp. 61–104.

prospect theory A basis for hypothesizing that the satisfaction from avoiding losses exceeds the anticipation of equal-value prospective gains.

Chapter 15: Contracting, Governance, and Organizational Form 569

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

not only the down-scale fighting brand Old Navy but also went out and acquired the upscale brand Banana Republic.

Let’s return to Figure 15.5 to see exactly why full-line forcing works. Auto dealers face an especially difficult competitive environment with mobile customers, innumerable sub- stitute makes and models, and Internet search engines comparing price discounts by competing dealers on a real-time basis. Moving their repeat-purchase “customers for life” up along a base-car-to-luxury-car product spectrum is one key to profitability in the auto dealership business. Another key is to avoid selling only base automobiles.

In Figure 15.5, stepping down from an SE model minivan with full options to a base model saves a customer $5,000 but sacrifices 50 percent of the perceived value of the minivan product line (−160 satisfaction utiles out of 320 total). Spending another $5,000 to step up to an SXT or LX luxury model garners only an additional 30 percent (+100/320) of the product line’s perceived value. Although the unit sales of the dealer- ship do not change, the mere presence of the LX and SXT model drives such a customer to justify spending $5,000 to avoid the greater disutility of the base model while congrat- ulating himself or herself for avoiding spending another $5,000 for a smaller relative in- crease in satisfaction.

After providing high-quality (and high-margin) service and maintenance for the SE with full options over several years, the dealership then plans to see that same customer back in the showroom eyeing the SXT models. Again, full-line forcing can encourage the “upsell” to the $27,000 SXT model. Saving $5,000 by continuing to purchase SEs with options will result in a 30 percent disutility of (−100/320), whereas purchasing the $5,000-more-expensive Town and Country model in Figure 15.5 would increase satisfac- tion by only 40 utiles, 12.5 percent. Again, the careful shopper, after comparing marginal benefits and marginal costs, decides to go with the SXT, and the dealership rejoices.

Example Full-Line Forcing in Pens, Aspirin, and Multivitamins at Revco and Eckerd Perhaps the most amazing aspect of prospect theory is the extent to which retail market shares can be altered by the practice of full-line forcing. Eckerd and Revco fully control the distribution channel policy in their own drugstores. In-house store brands can be sold beside national brands. Suppose Revco Aspirin at $2.89 for 100 80-mg tablets secures a 30 percent market share against a generic aspirin product selling 100 80-mg tablets at $1.50. Revco can allocate additional shelf space to non- store-brand painkillers and chooses between Bayer Aspirin at $2.89 and Tylenol at $5.29. Not surprisingly, Revco Aspirin’s market share falls if the cheaper Bayer product is introduced. But what is astounding is that, in case after case, introduc- ing the Tylenol product raises Revco Aspirin sales to a 40 percent market share, reduces generic aspirin from 70 to 40 percent, and raises Tylenol to 20 percent.

Because some people experience side effects from aspirin but not from Tylenol, a fairer comparison perhaps is to investigate the same experiment with absolutely identical multivitamins. Revco multivitamins at $3.29 per 100 tablets might again secure 30 percent of the market relative to generic multivitamins at $1.99 with 70 percent market share. Introducing One A Day brand multivitamins at $5.19 into the channel will actually raise the market share of Revco’s product. That is, with the good-better-best product strategy, market shares might distribute as follows: 40 percent generic, 40 percent Revco, and 20 percent One A Day. The disutility of lost perceived consumption outweighs equal-value utility gains.

570 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

VERTICAL INTEGRATION Search, bargaining, and holdup costs are all reduced when internal transfers and the monitoring and incentive systems within the firm replace the spot market contracting and recontracting necessitated by operating at arm’s length with outside suppliers and independent distributors. As we have seen, Nobel laureates Ronald Coase and Oliver Williamson argue that these factors explain why the firm emerged as an organizational form despite the diseconomies of ever-wider spans of managerial control.18 Another mo- tive for a manufacturer to vertically integrate upstream to suppliers or downstream to retail distributors involves the inefficiency of successive monopolization (i.e., the pres- ence of market power over price at more than one stage of production). For example, the transfer of Disney Studio’s downstream equity to Pixar (the upstream digital enter- tainment content provider) is a method of pre-commitment by Pixar to exercise up- stream price restraint and not spoil the downstream market. We now illustrate these ideas further with a detailed study of vertical integration in the hosiery industry.

Consider, first, an upstream yarn supplier who operates in a competitive intermediate product market and a downstream hosiery manufacturer who enjoys the market power to mark up the wholesale price for pantyhose above its marginal cost. Figure 15.6

FIGURE 15.6 Hosiery Integration Analysis with Upstream Competitor

Q*

Py

Ph

CD

A B

MCy

MRh

MCh

MRh � MCh

MCy � MCh

Demandh

Quantity (lbs.)

Pr ic

e ($

p er

lb .)

18For more extensive discussion of this topic, see S. Hamilton and K. Stiegert, “Vertical Coordination,” Journal of Law and Economics (April 2000).

Chapter 15: Contracting, Governance, and Organizational Form 571

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

illustrates the situation each firm faces when the yarn inputs are combined in fixed pro- portions with manufacturing labor and machinery to yield hosiery output. The outside demand curve and its marginal revenue capture the hosiery manufacturer’s revenue op- portunities in the wholesale pantyhose product market. Given the marginal cost of ho- siery production (MCh) and the competitive price of yarn (Py = MCy), the manufacturer sets summed marginal cost of hosiery production and yarn inputs (MCh + MCy) equal to hosiery marginal revenue (MRh) at output Q*. This joint profit- maximizing output decision maximizes hosiery profits by setting the marginal cost of yarn equal to the net marginal revenue product of the yarn supplier. So, subtracting the downstream marginal cost (MCh) from the downstream marginal revenues (MRh) leaves the net revenue opportunity available to the upstream yarn supplier—that is, MRh – MCh. Setting this derived demand for yarn inputs equal to upstream marginal costs (MCy) identifies Q* as the yarn supplier’s preferred throughput rate as well as the hosiery manufacturer’s preferred output rate. Thus, the upstream supplier who prices yarn so as to just recover marginal cost imposes no throughput constraint on downstream hosiery operations.

The hosiery manufacturer in Figure 15.6 would not change the yarn input prices, nor the wholesale output prices, nor the throughput quantity if it were to vertically integrate upstream and operate the yarn supplier. Therefore, vertical integration can only result in disadvantages associated with a wider span of managerial control. For profits ABCD to remain unchanged, these disadvantages would need to be offset by another factor such as reduced transaction costs. In general, in the absence of other factors, we would conclude that in Figure 15.6, the hosiery manufacturer has no profit motive for backward integra- tion into the competitive yarn supplier’s business.

In contrast, however, consider the case in which the yarn supplier has a proprietary process that is unique and therefore adds substantial value itself to the hosiery manufactur- ing process. In Figure 15.7, the derived demand for the yarn input is again MRh – MCh, and everything else about the hosiery operations remains the same as in Figure 15.6, ex- cept that now the upstream firm has the market power to mark up its own marginal cost (MCy). By taking a second marginalization of the revenue, subtracting off the hosiery pro- duction cost, and setting MMRh − MCh = MCy, the yarn supplier maximizes upstream profits EFGH by choosing a price P´y at throughput Q´. If P´y exceeds the upstream mar- ginal cost MCy, the summed marginal cost facing the hosiery manufacturer is now higher, and, consequently, the desired output declines from Q* to Q´. Although hosiery prices rise to P´h, the higher costs and smaller output of hosiery operations cause the profits of the downstream firm (the manufacturer) to decline (i.e., IJKL in Figure 15.7 < ABCD in Figure 15.6). That is, the presence of profit margins upstream results in a throughput constraint that unambiguously reduces downstream profitability.19

Backward vertical integration by the hosiery manufacturer can squeeze out the mar- gins upstream by simply setting an internal transfer price for yarn Py = MCy. This change will return the optimal throughput to Q*, the profit-maximizing level for the consolidated yarn and hosiery operations. That is, even after paying the upstream profits EFGH to secure the control rights from the yarn company, the downstream hosiery manufacturer has higher net profits (ABCD − EFGH) than its profit from independent operations IJKL. Consequently, we would expect these two firms to coordinate their op- erations either as a joint venture or as a vertically integrated hosiery manufacturer with its own retail distribution.

19This implication holds without qualification here because of fixed proportions production, i.e., the efficient input mix remains unchanged despite the reduction in output. Under variable proportions, vertical integration may be motivated or not, depending on the input substitutability and possible cost savings.

572 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIGURE 15.7 Hosiery Integration Analysis with Upstream Market Power

Q*

P�

P�

L

JI

E F

G

H

MCy MRh

MCh

MRh � MCh

Demandh

Quantity (lbs.)

Pr ic

e ($

p er

lb .)

K

MMRh � MCh

Q�

y

h

yP� � MCh

WHAT WENT RIGHT • WHAT WENT WRONG

Dell Replaces Vertical Integration with Virtual Integration20

New developments in information technology, such as the enterprise resource planning system SAP, have widened the efficient span of hierarchical control. Rather than en- abling larger vertically integrated companies, SAP enables virtual integration. Dell Inc. owns almost no PC compo- nent manufacturing operations. Instead, it outsources its requirements to several hundred supplier-partners who are tied together in a real-time monitoring, adaptation, and control system using the Internet. Its patented build- to-order business model must handle effectively an extra- ordinarily complex set of just-in-time component flows to support a final product assembly that ships 10,000 possible product configurations direct to a customer.

Information technology plays a key role in the gover- nance mechanisms for this type of virtually integrated sup- ply chain management. When this business model is successful, plant and equipment requirements decline, in- ventories shrink, and operating leverage rises substantially. With less fixed capital investment than a vertically inte- grated competitor, the return on invested capital climbs accordingly. Dell stock appreciated 300-fold in its first fifteen years of operation.

20Based on “Identity Crisis,” Wall Street Journal (October 10, 2000), p. C1; “Direct Hit,” The Economist (January 9, 1999), pp. 55–58; and J. Margretta, “The Power of Virtual Integration,” Harvard Business Review (March/April 1998), pp. 72–85.

Chapter 15: Contracting, Governance, and Organizational Form 573

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Dissolution of Assets in a Partnership Vertical integration is the most extreme organizational form on a continuum from out- sourcing with spot market transactions, through relational contracting, vertical require- ments contracting, and extending to alliances and joint ventures. In Appendix 15A, we will see how subtle institutional arrangements can align the interests of partners in a joint venture. But what happens when a joint venture or partnership must be dissolved?

The optimal mechanism design for fair division necessitates sequential game analysis when the prize is decaying with the passage of time as the players negotiate. Suppose a partnership must divide $3 million in assets. The only catch is that the assets, which the two partners must agree upon dividing, decline in value by $1 million each time either party refuses the proposed division. Perhaps the assets are perishable pharmaceuticals or intellectual property that decays in value quickly. Or, alternatively, once the capital mar- kets get wind that the partnership is fracturing, any synergistic value of the two subsets of assets quickly evaporates.

The two principals toss a coin to see who should make the first offer of how to split the $3 million. Suppose Joe wins the toss. How much should he offer to Kim to secure her ac- ceptance of the offer and forgo her rights of first refusal? Kim gets to structure a second- round offer should this first proposed division be refused, and she will also respond to Joe’s third and final offer should the second proposal be refused. Notice that the endgame to this problem happens after three rounds when the assets are gone. Think about how little Kim would be willing to accept in the third and final round. Can Kim hold out for more than this amount in the second round? What, therefore, is the maximum Joe needs to offer to trigger an acceptance in the first round? Does this game have a first-mover advantage, or is it better to play second and be the respondent in the third and final round?

These and related questions can be addressed with the sequential game techniques of Chapter 13. In Figure 15.8, Joe makes the final offer at the far right but, as you may suspect, in one sense Kim’s right of refusal controls this endgame. Because $100,000 is

FIGURE 15.8 Dissolution of Assets in a Partnership

Note: Payoffs are listed in the following order: (Joe’s payoff, Kim’s payoff)

Assumptions: No communication other than the offers. Each party pursues the objective of maximizing absolute gain (blue payoffs). Alternatively, if Kim pursues maximizing relative gain, the shaded payoffs emerge. Minimum increment is $100,000.

{$900,000, $2.1 million}

{$1.9 million, $1.1 million}

Joe I

Kim I

Kim II

Offer to split

Offer to split

Accepts

Refuses

($500,000, $1.5 million)

($1 million, $1 million)

Joe II

Joe III

Offer to split

Accepts

Refuses

($400,000, $600,000)

($900,000, $100,000)

($0, $0)

Kim III

Accepts

Refuses

574 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

preferable to zero, Kim’s best-reply response is to accept an offer to split the final million as $900,000 for Joe and $100,000 for Kim. This analysis implies that in the preceding (second) round, Kim can offer Joe $1 million (i.e., $100,000 more than Joe’s endgame outcome), and Joe’s best-reply response is to accept. Stepping the backwards induction one more round back, this second-round analysis implies that if Joe offers Kim $1.1 mil- lion in the first round (i.e., $100,000 more than Kim’s second round outcome), Kim will accept. Note that we are short-circuiting threats and modes of communication other than simply stating the offers. Also note that we assumed each party believes the other party is maximizing absolute gain, without regard to the relative distribution. In such circumstances, making the first and the last offer appears advantageous to Joe.

Mechanism design features of this problem include the rotating offer and right of re- fusal, the lack of communication about objectives and credible threats, and the coin toss to decide who offers first. Any of these could be changed, and it might make quite a dif- ference. For example, suppose Kim’s reputation was that she pursued relative wealth, not absolute wealth. In that event, the outcome ($0; $0) would quite literally be preferable to the alternative of her partner Joe going $800,000 up at ($900,000; $100,000). Even split- ting the last million dollars ($500,000; $500,000) would not be preferable to ($0; $0). So only an asymmetric distribution favoring Kim (say, $400,000; $600,000 in the shaded box) would elicit her acceptance.

In earlier rounds, Joe now knows that he must increase the unequal distribution in Kim’s favor in order to secure her acceptance, and Kim knows that the same is not true of Joe. Given the $100,000 minimum increments, Kim reasons by backward induc- tion from the ($400,000; $600,000) outcome of Round 3 that she can offer as little as ($500,000; $1.5 million) to secure Joe’s acceptance in Round 2. Finally, this intermediate outcome implies by the relative gain reasoning motivating Kim that Joe must elicit her acceptance at the start of the negotiations by offering a ($900,000; $2.1 million) split in Round 1. Comparing the blue and shaded offers in Round 1, it makes $1 million of dif- ference whether the mechanism design allows communication of Kim’s motivation to maximize relative (not absolute) gain as opposed to a mechanism design with unknown silent partners responding through intermediaries. Small changes in the institutional pro- cedures or the organization architecture often make this large a difference in tactical encounters.

SUMMARY

� Businesses make choices about organizational form that define the span of hierarchical control from the vertically integrated oil company at one ex- treme to the virtual manufacturer Dell, which out- sources all manufacturing and assembly to supplier-partners.

� All external and internal business relationships re- quire a solution to the twin problems of coordina- tion and control. External business relationships can be organized through spot market transactions, long-term contracts, or reputation effects in rela- tional contracting. These forms of organization

differ in their timing, players, enforcement, and information structure.

� Long-term vertical requirements contracts provide an ex ante framework for resolving coordination and control problems between manufacturers, sup- pliers, and distributors. Because all contracts are purposefully incomplete, ex post opportunistic be- havior requires governance mechanisms to reduce several types of contractual hazards. The moral hazard problem arises because of the unobservabil- ity of effort in contract performance. The post- contractual opportunistic behavior, called “holdup,”

Chapter 15: Contracting, Governance, and Organizational Form 575

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

presents another commonly occurring contractual hazard.

� Exchange under incomplete information and un- der asymmetric information differs. Incomplete in- formation refers to the uncertainty that is pervasive in practically all transactions and motivates insur- ance markets. Asymmetric information, on the other hand, refers to private information one party possesses that the other party cannot indepen- dently verify.

� Asymmetric information leads to the problem of adverse selection.

� Contracts are seldom complete because full contin- gent claims contracting is often prohibitively ex- pensive. Intentionally incomplete contracting allows post-contractual opportunistic behavior and leads to holdup. Resolving the holdup problem in managerial contracting necessitates the use of gov- ernance mechanisms.

� Governance mechanisms include internal monitor- ing by director subcommittees and large creditors, internal/external monitoring by large block share- holders, auditing and variance analysis, bench- marking, an ethically dutiful corporate culture, and whistle-blowing.

� Managerial labor can be hired in several ways, for example, straight salary, wage rate, or profit shar- ing. Pure profit sharing results in moonlighting, however, because the manager’s inputs (effort and creative ingenuity) are largely unobservable. Unob- servable effort leads to the moral hazard problem, which can be resolved by setting up ex post facto (e.g., with deferred stock options).

� In combination, random disturbances in firm per- formance and unobservable managerial effort pre- sent a more difficult principal-agent problem to resolve. Owner-principals do not know when to blame manager-agents for weak performance or give credit for strong performance. Optimal incen- tives contracts involving some guaranteed salary

and a profit-sharing bonus can, in theory, resolve the principal-agent problem.

� Linear combinations of salary and profit sharing can also be used to elicit asymmetric information about managerial preferences, sort managers by their own personal risk aversion, and prevent ad- verse selection in managerial hiring. Boards of di- rectors face a holdup problem in renewing senior executive contracts, necessitating governance mechanisms.

� What form of organization to adopt (e.g., spot market transactions, long-term vertical require- ments contracts, relational contracts, or vertical in- tegration) depends on the contractual hazards that need to be avoided. What contractual hazards arise in business relationships depends on the asset characteristics, redeployability or specificity of the fixed assets, and the relative dependence of those fixed assets on unique complementary assets.

� Perceived utility losses are often larger (in absolute value) than equivalent-value utility gains, implying a hybrid utility function suggested by prospect theory.

� Prospect theory implies sellers should distribute trial products, take prospective income in payment, and full-line force.

� Vertical integration is an optimal organizational form when the assets are one-way dependent on complementary assets and are largely non- redeployable.

� Optimal mechanism design seeks to motivate value-maximizing behavior while reducing transac- tion costs.

� Mechanism design features such as first offer, right of first refusal, lack of communication, and credi- ble threats in a dissolution of assets in a partner- ship can be analyzed as a sequential game.

� Small changes in the institutional procedures often make a large difference in the outcome and distri- bution of payoffs.

Exercises 1. Suppose an enhanced effectiveness of cooperative advertising occurs if the distrib- utor shares its superior on-the-spot information about current trends in the mar- ketplace with the manufacturer. Explain how each of the following would affect the information-sharing objective: a. All details of co-op advertising are agreed to up front in the franchise contract.

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

576 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

b. Advertising is pursued independently by the manufacturer and the retail distributor.

c. Cooperative advertising allowances are rebated to the distributor out of the franchise fees.

2. If contract promises were not excused because of acts of war, would the clearing and settlements clients of Bank of New York change their behavior? If so, how? What reliance behavior would be considered efficient? What reliance behavior would be considered excessive?

3. For each of the possible loan terms the commercial lender might offer, identify what choices by the lender would be the worst for combating the moral hazard problem in Figure 15.2.

4. If coal mine tonnage can be shipped elsewhere cheaply, but an adjacent coal-fired power plant is not redeployable to other uses, what organizational form would be adopted by the power plant owners? Why?

5. Would warehouse operators insist on owning their own trucking companies? Why or why not? What coordination and control problems and contractual haz- ards would these companies encounter?

6. What organizational form would warehouse operators and truck hauling compa- nies adopt?

7. In benchmarking sales representatives against one another, what problems arise from continuing to reassign the above-average trade representatives to previously unproductive sales territories?

8. Explain how the optimal incentives contract would differ if the less risk-averse bank officer (Dashing in Figure 15.4) had generated the smaller expected profit (i.e., the lower hill-shaped curve).

9. In the Division of a Decaying Business game, what should you offer if the assets at the start of the game are $4 million rather than $3 million? Now is there a first- mover or second-mover advantage? Why?

10. Analyze the pure Nash equilibrium and mixed Nash equilibrium strategies in the following manufacturer-distributor coordination game. How would you recom- mend restructuring the game to secure higher expected profit for the manufacturer?

Manufacturer

Product Update/ No Update/ Higher MSRP Same MSRP

Discontinue Special Selling Services

Distributor

Continue Special Selling Services

$0 2 M

$1 M $4 M

$6 M 0

$2 M 0

Chapter 15: Contracting, Governance, and Organizational Form 577

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Case Exercises BORDERS BOOKS AND AMAZON.COM

DECIDE TO DO BUSINESS TOGETHER Borders Books, the leading chain of bookshop retailers, entered into an agreement with Amazon.com, the online retailer, to fulfill Internet orders from the Borders.com Web site. Using Table 15.3 and the following questions, what organizational form would you predict for this business relationship?

Questions 1. Are Amazon.com’s warehouses, Web pages, and one-click sales methods fully re-

deployable to other products? If so, name a few such products. If not, why not? 2. Are Borders’ fixed assets fully redeployable? If so, suggest how. If not, why not? 3. Is Borders dependent on Amazon as a unique complement? That is, is Amazon

.com the only potential company that could process Internet-based orders for Borders?

4. Is Amazon dependent on Borders for referrals, or does it have its own Internet- based order flow?

5. Is your answer consistent with the multiyear fee-for-service contract between Bor- ders and Amazon.com whereby Amazon processes the order, ships the book, re- cords the sales, and pays Borders a referral fee? One Borders executive described this as a “low-risk, low-return” approach to online sales while retaining Borders’ focus on its core mission of running bookshops.

DESIGNING A MANAGERIAL INCENTIVE CONTRACT Recall that in Chapter 1, Specific Electric Co. asked you to implement a pay-for-performance incentive contract for its new CEO. Using your deeper knowledge of the principal-agent problem, try it again.

THE DIVISION OF INVESTMENT BANKING FEES IN A SYNDICATE You are the lead underwriter among a syndicate of five investment banks composed of yourself, the syndicate co-manager, and syndicate members 3, 4, and 5. Your syn- dicate finds a deal worth $100 million in fees. You must submit a proposal as to how the fees should be divided, and the syndicate then votes by majority rule.

Your syndicate is rational and democratic in the sense that the division of fees will be decided based upon maximization of absolute gain in this single deal, and the members also have reputational reasons in future deals (where the lower-ranked members hope to achieve more influence and get a higher-ranked position) to abide by a majority decision.

If your proposal is rejected by vote of the syndicate, you are displaced as lead un- derwriter, removed altogether from the deal making, and replaced by the co-manager who then makes a proposal to the remaining four. If his deal is rejected, he too is

578 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

removed, and syndicate member 3 makes a proposal to the three firms remaining, and so on.

Questions 1. What should you as the lead underwriter offer and to whom? [Hint: Employ the

methods of sequential game reasoning. Start at the endgame and work backward.] 2. Is there anything unusual about the divison of these fees across the syndicate? 3. Why would any investment bank volunteer to be a syndicate co-manager?

Chapter 15: Contracting, Governance, and Organizational Form 579

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

15A APPENDIX

AuctionDesign and Information Economics In this appendix, we discuss optimal mechanism designs for auctions and several other institutional procedures. Priceline sells airline and concert tickets by soliciting sealed bids with purchase guaranteed for given dates of travel, while eBay posts ascending price of- fers with amendment and cancellation. These are instances of alternative mechanism de- signs for auctions. Often the purpose of such mechanism design choices is to induce the revelation of asymmetric information that is needed to maximize shareholder value for the auction buyers or sellers.

Mechanism design can also address coordination and control problems in joint ven- tures and partnerships. Specifically, an incentive-compatible contract (IC contract) can induce partners to reveal their private proprietary information about preliminary cost projections to one another. In an IC contract, each partner will incur the net cost effects of his or her own information revelation on the other partner. Hence, the IC contract imposes a mechanism design for sharing profits that serves the self-interest of each part- ner through revealing the true and complete cost information to the other partner. This incentive-compatible revelation mechanism provides a powerful tool with unique capabilities.

OPTIMAL MECHANISM DESIGN Queue Service Rules A simple business application of the concept of optimal mechanism design is the queue service rule for filling customer orders from those waiting to purchase. The traditional first-come, first-served procedure induces an inefficient pattern of customer arrivals at, for example, a concert site. If the box office opens at 9:00 A.M., a few potential customers arrive three hours earlier or even the night before. Others stand on queue for two hours, and many more show up to wait in line as the ticket booth opens.

What customers are willing to pay for tickets is surely affected by the inconvenience of this wait. And the subpopulation of customers who have low opportunity cost of time (and are therefore willing to arrive the earliest, wait the longest, and obtain tickets with the greatest probability) may not be the subpopulation that will pay the most for tickets. For this reason, many ticket agencies have no objection to a wealthy patron paying someone with lower opportunity cost of time to stand on queue, purchase the ticket, and transfer it at face value to the higher-willingness-to-pay customer.1 In any case, all this waiting time is time wasted, and time is money. As a result, if other more efficient queue service rules were adopted, more of the customer’s willingness to pay could be captured by the concert promoters and sports teams.

optimal mechanism design An efficient procedure that creates incentives to motivate the desired behavioral outcome.

1Scalpers, of course, charge higher prices still, but note that such gray markets reveal to the ticket agency what those customers who are not willing to show up at the ticket window and wait are in fact willing to pay. This information helps the ticket agency set an optimal price.

580

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIRST-COME, FIRST-SERVED VERSUS LAST-COME, FIRST-SERVED As a provisional alternative, consider last-come, first-served. Under this queue service rule, a customer has no incentive to stand on queue and wait. Indeed, any time a queue forms, all those in front of the last person to arrive have an incentive to leave and go about their other business, returning later when fewer people are likely to show up. In essence, the last-come, first-served mechanism design has removed the incentives for in- efficient behavior artificially created by first-come, first-served. Customers do not inher- ently prefer to arrive early and peak-load their demands. Instead, it was the non-optimal queue service rule that artificially created incentives to arrive early, stand around, and wait.

With last-come, first-served, in contrast, customers have an incentive to spread their arrivals throughout the ticket window’s normal hours of operation. Once a more or less uniform distribution of customer arrivals throughout the day can be established, the ticket agency can adjust its capacity and set its service rate to deal with the steady stream of customers who arrive and purchase with little or no waiting.

Few ticketing operations have adopted a last-come, first-served queue service rule. Why not? Under last-come, first-served, recall that any customer should leave the queue whenever a later arrival preempts his or her priority ranking as last in line. But

Example Containerized Shipping at Sea-Land/Maersk Historically, ocean shipping rates have been heavily regulated based on categories of cargo (e.g., paper, film, frozen fish) and shipping lanes (e.g., Rotterdam to New York, Liverpool to Jacksonville, Seoul to San Francisco). Conferences of ocean shipping companies announced common carrier shipping rates for first-come, first-served customers. More than half the world’s cargo still moves under publicly mandated ocean shipping contracts at these uniform shipping rates. With no abil- ity to adjust prices, salespeople maximized the volume of cargo. The only good ship was a full ship. Because empty slots on a container ship perish as a revenue opportunity the moment the ship sails and because a container ship such as the Regina Maersk has space above decks for more than 700 containers, shippers queued up large volumes of cargo in anticipation of each ship’s departure. In this mechanism design, the customers’ waiting time became a substantial implicit cost totaling millions of dollars a day that otherwise might have gone to the shipper as additional revenue.

Today, deregulation is fast approaching the ocean shipping industry, and a spot market auction for space in containerized ships has emerged. The immedi- ate consequence has been the delaying of low-priority shipments such as rolls of newsprint from one voyage to the next in favor of higher-rate cargo such as perishable pharmaceuticals. In response to this new business environment, Sea- Land/Maersk optimized the placement of their containers around the world. Each empty container at each freight terminal is assigned a forecasted net reve- nue opportunity at that location and at other potential locations along the shipping route. Shippers who offer less waiting time can charge higher rates; last-come, first-served policies are being explored for certain high-margin cargo.

Appendix 15A: Auction Design and Information Economics 581

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

customers will not wish to return many times to the ticket window. Again, time is money. So, customers ahead in line are likely to offer side payments to later arrivals to induce them to leave. Predictably, those with the highest opportunity cost of time will end up bribing those with lower opportunity cost to leave and return.

The problem is that this side payment system again reduces the ticket agency’s re- ceipts because in many ways it has just replaced the inefficiency of arriving early and waiting with the inefficiency of arriving often, departing, and returning. The recipients of the side payments are no worse off because they voluntarily decide to leave and return later, but those who make the side payments now turn to the window and surely offer less than they otherwise would have paid. How can the ticket agency’s mechanism design deal with this subtle complication?

Stratified Lotteries for Concerts Lotteries and online auctions may hold the key to an effective mechanism design for ticket sellers. Online sales are cheaper than phone sales because they require no op- erators and no toll-free line charges, if customers will accept an Internet or faxed ticket, thereby saving the ticketing agent’s mailing expense. Only 3 in 10 live enter- tainment tickets is distributed online today, but this channel is growing much faster than industry sales as a whole. Suppose rather than announcing in advance what po- sition in the customer queue would be served first, the ticketing agency simply picked a position at random. In effect, that’s exactly what a lottery for the right to purchase a ticket does. Anytime prior to the day of sale, a customer stops by an Internet Web site to pick up a lottery number.

Example Online Auctions and Stratified Lotteries at Ticketmaster2

Because those customers with low willingness to pay are equally likely to get the winning lottery numbers as those with high willingness to pay, Ticketmaster and other companies adopt a stratified lottery scheme. Rights to purchase high-price seats are distributed in one lottery, medium-price seats in another, and low-price seats in a third. At a designated date, the winning numbers are chosen at random and posted on the Web site and public-access TV channels. Only those customers holding the winning lottery numbers arrive to buy tickets, and because seat avail- ability is assured, customers have no reason to arrive early, queue up, and wait. This lottery mechanism design does reduce waiting time while allowing the ticket agency to charge higher uniform prices for each class of seats.

Holding auctions online can present a conflict of interest with Ticketmaster’s traditional business of adding a $3 to $5 convenience fee for computerized ticket distribution by mail and in music stores. As the exclusive ticketing agent for 70 million tickets worth $3 billion, Ticketmaster is wary of reselling tickets above face value. The sports teams, concert halls, and promoters would then suspect Tick- etmaster of underpricing the original tickets to charge convenience fees on not one but two sales. Ticketmaster could, of course, simply defer to the arenas and promo- ters in setting the face value prices, but a better alternative may be available.

(Continued)

stratified lottery A randomized mechanism for allocating scarce capacity across demand segments.

582 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

AUCTIONS Types of Auctions Everything from Monday Night Football to mineral rights, forest land, airline tickets, used equipment, and the electromagnetic spectrum have been allocated to their highest valued use through auctions. The choices in auction design are numerous, as illustrated in Table 15A.1. Bidding can be sequential with rebid opportunities like eBay and most estate auctions or simultaneous as on Priceline and in the sealed-bid auctions for newly issued government debt securities. Bid prices can be continuous or constrained by mini- mum discrete bid improvements. The New York Stock Exchange recently switched to continuous decimalization of their auction prices from the one-eighth-tick size restric- tions on minimum bid improvements. Bids may be sealed, revealed by open outcry, or posted anonymously as on eBay. Bidding can be one-time only or dynamic, repeated in multiple rounds with cancellation and amendment of prior bids allowed (so-called open bidding). Finally, owners can place a minimum reservation price (the “reserve”), below which the item will not sell, or allow an auction to proceed with no minimum.

Perhaps the most important design differences between major types of auctions are who pays, what amount, and how the winner is determined. In some auctions all bidders

Ticketmaster and StubHub.com are both conducting “official” Internet auctions for various sports teams. Last year, the New York Jets earned close to $100 million in auctioned ticket sales on behalf of season ticket holders, with most transactions occurring at about 30 percent over face value. The team takes approximately 10 per- cent of this “gate” (incremental revenue), and the rest goes to the season ticket holder. But what type of auction should Ticketmaster design?

2Based on “Ticket Scalpers Find a Home on the Web,” Wall Street Journal (February 4, 1999), p. B1; “A Winning Ticket,” The Economist (August 22, 1998), p. 52; S. Rosen and A. Rosenfeld, “Ticket Pricing,” Journal of Law and Eco- nomics 40, no. 2 (1997), pp. 351–377; and “Don’t Scalp Us. We’ll Scalp You,” BusinessWeek (April 19, 2004), p. 44.

TABLE 15A.1 A COMPARISON OF AUCTION MECHANISM DESIGN

CHARACTERISTICS

EBAY PRICELINE

Sequential Simultaneous

Minimum bid improvement Continuous

Posted prices Posted offers to buy (reverse auction)

Multiple rounds One-time-only if seller “hits”

Open bidding Credit card immediately authorized

Reserve No reserve

Highest wins and pays All accepted bids pay

First (highest) price Whatever price was bid

English ascending price Dutch discriminatory descending price

Consolidated feedback on seller Seller anonymous

Appendix 15A: Auction Design and Information Economics 583

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

pay (e.g., on Priceline.com, credit card information must accompany all offers, and con- ceivably the seller could collect from all bidders if the prices were acceptable). Of course, most auctions adopt the highest-wins-and-pays allocation rule. What the winner pays (whether it is his or her own highest bid or, at times, the second-highest bid) and how the auctioneer arrives at the winning bid can differ. English auctions ascend to higher and higher prices with open outcry or posted bids until the last bidder to make an offer exceeding all other offers is declared the winner.

Dutch auctions work in the opposite direction by identifying the first bidder to regis- ter an acceptance as the auctioneer announces a succession of descending asking prices. The wholesale flower market in the Netherlands operates in this way, hence the term Dutch auction. In ascending-price auctions the winner takes all, but in descending-price auctions, the winner is often given the opportunity to purchase less than the total capac- ity available for sale, and the auction then continues downward.

Which of these and other auction design characteristics maximize the revenue to the seller and which allocate resources to their highest-valued use are important business questions and public policy issues. One well-understood insight from mechanism design theory is that asymmetric information will lead to timid bidding in ascending price auc- tions because of the winner’s curse. To illustrate the winner’s curse, consider the follow- ing auction situation:3 You are developing a bidding strategy for an asset whose value to the seller is a random variable distributed uniformly between $0 and $100. The seller observes this value and desires some profit on the transaction to cover the auction ex- penses, but places no minimum reservation price on the auction. You anticipate that a rational seller will refuse all offers below his or her personal value. The asset might be a baseball player’s labor contract or a set of maps of the subterranean geological forma- tions in a petroleum-rich area. Because of different complementary assets and skill, your value is certain to be 50 percent higher than the seller’s personal asset value. What offer should you make?

Winner’s Curse in Asymmetric Information Bidding Games If neither party knows the true value, an expected value of $50 plus a small premium (i.e., well below $75) appears initially to be a reasonable offer that will be accepted. However, if the seller knows the true value, consider what reasonable offers will be ac- cepted and what reasonable offers will be refused. To simplify the analysis, suppose just three offers and three realizations of the seller’s value are possible: $0, $55, and $100. In Figure 15A.1, we see that should the true value be zero, only offers that overpay for the asset will be accepted. These payoffs are shown in the shaded boxes to the right of the decision tree. Should the true value be $55, $55 offers will be refused and again only $100 offers that overpay (even relative to the $77.50 value to the bidder) will be accepted (see the lowest boxed payoff). If the assets have the maximum possible value to the seller of $100, offers of $0, $55, and $100 will be refused. In short, all reasonable offers will be refused. Therefore, surprisingly, you should offer nothing at all! If you win such an auction, you are cursed with having overpaid for the asset. Welcome to the winner’s curse!

English auctions An ascending-price auction.

Dutch auctions A descending-price auction.

3Adapted from M. Bazeman and W. Samuelson, “I Won the Auction But Don’t Want the Prize,” Journal of Conflict Resolution (December 1983), pp. 618–634. See also R. McAfee and J. McMillan, “Auctions and Bid- ding,” Journal of Economic Literature (September 1987), pp. 699–738.

winner’s curse Concern about overpaying as the highest bidder in an auction.

584 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Table 15A.2 lists several auction price sequences for offshore oil tracts and FCC spec- trum rights. The huge gaps between the winning bid and the second highest bid suggest that a winner’s curse was present. The winning bidder for offshore oil in the Santa Bar- bara channel paid $11.4 million more than the second highest bid. Similarly, Wireless Co. paid $12.2 million more than GTE bid for the cell phone spectrum license in the Dallas Metro area. Bidding sequences for star players in professional sports look similar.

Mechanism design theory reveals several insights about this asymmetric information bidding game. First, most bidders will figure out the winner’s curse in an auction design such as Table 15A.2 and therefore bid timidly, if at all.4 To induce more aggressive bid- ding in repeated, multiple-round versions of such asymmetric information games, sellers such as De Beers find they must carefully sort their rough-cut diamonds, grading them into “sights.” Reputation for reliability in grading the sights more economically than the bidders could pick, cull, and resell the rejects is what brings De Beers’ buyers back,

FIGURE 15A.1 Winner’s Curse in an Asymmetric Information Bidding Game

Notes: Payoffs are listed in this order: (Bidder, Seller). Starred payoffs refer to best-reply responses.

S

Refuse

Accept

�$17.50 $45*

B

All offers refused by seller

Bidder behavior Timeline: Information of

random asset value Seller behavior

�$55 $55*

0 0

0 0

0 0

$50 Loss

$27.50 Loss

0 0*

$95 �$45 5

0 0*

0 0*

$0 Offer

$0

$55

$100

$0

$55

$100

$55 Offer

$100 Offer

N

S

S

S

S

S

Refuse

Accept

Refuse

Accept

Refuse

Accept

Refuse

Accept

Refuse

Accept

N

�$100 $100*

4Notice that the same conclusion applies to a continuous-bid version of the auction, though experimental evi- dence suggests that most first-time players misperceive the asymmetric information nature of the seller’s right of refusal and incorrectly bid $50 to $75. See C. Camerer, “Progress in Behavioral Game Theory,” Journal of Economic Perspectives 11, no. 4 (Fall 1997), pp. 167–188.

Appendix 15A: Auction Design and Information Economics 585

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Winner’s Curse at ESPN5

ABC debuted Monday Night Football and sold its ad slots for 30 years, but it over- paid and ended up losing approximately $150 million per year. When bids for the right to televise this show for the eight years of 1998–2005 for $4 billion, NBC decided that the winner would be cursed with losses and dropped out of the bid- ding. ABC (a Disney company) continued bidding and eventually won the “prize” for $4.4 billion ($550 million per year). For the 2006–2011 seasons, another media division of Disney, ESPN, agreed to pay almost twice as much ($1.1 billion per year) for the next eight years. ESPN hopes to garner sufficient revenue to break even from cable subscriber fees it charges Comcast and Time Warner as well as from the traditional advertising slots.

Viewership has tumbled 33 percent to 26 million households since the peak in- terest in professional football in the early 1980s. Although the new television rights do allow for more TV time-outs in which to sell commercials, $50 million per three-hour game works out to eighty-four 15-second spots (28 per hour), each costing $600,000 that will be required just to break even. By comparison, 15- second spots on the Oscars award show sell for $840,000 to reach 40 to 50 million viewers, and the Super Bowl with 90 million viewers sells for $1.25 million.

Professional football can increase other prime-time ratings, as CBS demon- strated by successfully shifting the Sunday NFL Game of the Week audience right into their highest-rated show 60 Minutes. And when the Fox Broadcasting network won the rights to the Sunday NFL games, the 60 Minutes audience share shrank from 30 percent to 22 percent. Nevertheless, we agree with ABC’s recent assess- ment of the winner’s curse. ESPN may be hard pressed to come close to recovering their investment.

5Based on “Thrown for a Loss by the NFL,” Time (January 26, 1998), p. 52; “A Ball ESPN Couldn’t Afford to Drop,” BusinessWeek (May 2, 2005), p. 42; and “Marketers Rely on Oscar,” Wall Street Journal (February 2), 2006, p. B3.

TABLE 15A.2 BIDS FOR OFFSHORE OIL CONTRACTS AND

FCC SPECTRUM RIGHTS

OFFSHORE OILa FCC SPECTRUMb

SANTA BARBARA CHANNEL

ALASKA NORTH SLOPE

MIAMI METRO AREA

DALLAS METRO AREA BIDDER

$43.5 $10.5 $131.7 $84.2 Wireless Co.

32.1 5.2 126.0 72.0 GTE Inc.

18.1 2.1 125.5 68.7 Wireless Co.

10.2 1.4 119.4

6.3 0.5 119.3

0.4 113.8

113.7

108.4

aIn millions, 1969 dollars. bIn millions, 1995 dollars. Source: Adapted from Tables II and IV in R. Weber, “Making More for Less,” Journal of Economics and Management Strategy, 6, no. 3 (Fall 1997), pp. 529–548.

586 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

auction after auction. De Beers then has a minimum participation rule that insists on a certain number of bids if one wishes to be asked back.

Secondly, if the asymmetric information is discoverable by appraisals, marketing re- search, or other similar services, another insight from auction design theory is that the seller should conduct a multiple-round auction with open bidding. Open bidding allows the bidders to react to asymmetric information revealed in prior rounds and therefore reduces the winner’s curse. This idea was used by the Federal Communications Commis- sion in the spectrum auctions for personal communication systems (PCS) such as cell phones, mobile fax and data service, and voice-mail pagers.

Information Revelation in Common-Value Auctions To illustrate this role of open bidding, consider two PCS bidders: Wireless Co., an alli- ance of Sprint and several large cable TV companies that spent $2.1 billion and won the rights to serve 145 million customers in 29 metropolitan service areas, and PCS PrimeCo, an alliance of three regional Bell companies that spent $1.1 billion and won the rights to serve 57 million customers in 11 metropolitan service areas. Several winning bids are listed in Table 15A.3. For example, Wireless paid $46.6 million for the Louis- ville, Kentucky, service area. How did Wireless decide what to bid?

Suppose that both bidders know that the net present value of the rights to transmit PCS services in Louisville is a random variable uniformly distributed from $10 million to $60 million with six discrete values possible—in particular, $10 million, $20 million, $30 million, $40 million, $50 million, and $60 million. Also assume (provisionally) that both parties value the asset identically, a so-called common-value auction. The problem then from the bidders’ point of view is to elicit sufficient information from the market environment and from the offers of other bidders to correctly identify the value and en- sure a profit (i.e., not overpay for the asset). In advance, each company conducts market- ing research experiments to narrow the possible outcomes and thereby better inform its own bid. Suppose Wireless Co.’s marketing research results are unable to exclude the two tails of the uniform distribution of possible values (i.e., $10 million and $60 million) but can exclude with certainty $20 million and $30 million as well as $50 million. Taken by itself, this information allows Wireless to narrow its probability assessments to $10 mil- lion, $40 million, and $60 million.

TABLE 15A.3 WINNING BIDS IN BROADBAND PCS AUCTION

MARKET POPULATIONa WINNER SECOND HIGHEST BIDb

PRICE/ POP.

New York 26.4 Wireless Alaacr $442.7 $16.76

San Francisco 11.9 PacTel AmerPort $202.2 $17.00

Charlotte 9.8 BellSouth CCI $ 70.9 $ 7.27

Dallas 9.7 Wireless GTE $ 84.2 $ 8.68

Houston 5.2 PrimeCo. Wireless $ 82.7 $15.93

New Orleans 4.9 PrimeCo. Powertel $ 89.5 $18.17

Louisville 3.6 Wireless PrimeCo. $ 46.6 $13.10

Salt Lake City 2.6 Wireless GTE $ 46.2 $17.95

Jacksonville 2.3 PrimeCo. GTE $ 44.5 $19.56

aIn millions from the 1990 census. bPrice paid for 30 MHz Block B spectrum rights, March 1995. Source: P. Cramton, “The FCC Spectrum Auctions,” Journal of Economics and Management Strategy, 6, no. 3 (Fall 1997), pp. 431–496.

common-value auction Auction where bidders have identical valuations when information is complete.

Appendix 15A: Auction Design and Information Economics 587

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Weighting each outcome equally yields an expected value bid of $36.7 million as follows:

1 3 ð$10 millionÞ + 1

3 ð$40 millionÞ + 1

3 ð$60 millionÞ = $36:7 million

Similarly, PCS PrimeCo conducts its own marketing research that, let’s assume, excludes $10 million, $30 million, and $50 million as possible outcomes for the Louisville service area. That is, PCS PrimeCo has access to separate information that causes it to calculate a different expected value bid:6

1 3 ð$20 millionÞ + 1

3 ð$40 millionÞ + 1

3 ð$60 millionÞ = $40 million

These best estimates of the common value are based on the asymmetric information available to the two firms. Consequently, in a simultaneous sealed-bid auction, the most a seller could hope to realize is $40 million. With sealed bids, no information is conveyed to the competitor, and an optimal bidding strategy is therefore simply to shade your bid slightly below the Bayesian expected value based on your own information set. PCS PrimeCo would therefore bid something just under $40 (perhaps $39.6) million and win the spectrum rights for the Louisville service area.

Bayesian Strategy with Open Bidding Design Notice, however, from the seller’s point of view ex post facto (after receiving the sealed bids) that the joint information set of the two parties suggests PCS PrimeCo has under- paid. To review, the union of the two sets of marketing research outcomes excludes $10 million, $20 million, $30 million, and $50 million. Said another way, the combined marketing research results have narrowed the possible outcomes for the value of the Louisville service area to $40 million and $60 million. Neither firm has access to this much information. Each simply knows a subset of all the marketing research available. But as a seller in such a setting, the FCC wishes to elicit full revelation of all asymmetric information because it affects the winning bid. If $40 million and $60 million are equally likely, and the bidders can somehow discern this information, the Louisville service area is worth just under $50 million, not PCS PrimeCo’s bid of just under $40 million.

One way to bring all the asymmetric information into play is to adopt a sequential open- bidding auction design in which each company is chosen at random to post its bid (in one round, then the random order of posting procedure is redone for Round 2, Round 3, etc.).7

6The equally weighted probabilities of 1/3 are actually Bayesian probabilities of each possible remaining value based on a perfectly accurate forecast that $10 million, $30 million, and $50 million (the prime numbers in the set of possible asset values) have been ruled out. It is helpful if we think of the marketing research as identify- ing Prime and Not Prime numbers between 1 and 6 multiplied by 1 million. Then, the Bayesian probability ($20 million/Perfect Forecast of Not Prime) = (0.167 × 1.0)/[0.167 + (0.833 × 0.4)] = 0.33 where 0.167 is the prior probability before the marketing research is conducted that $20 million will be the realized asset value. The number 1.0 is the accuracy of the forecasting instrument; for example, the conditional probability that when $20 million is the true value, the conclusion from the marketing research will be that the value is Not Prime, meaning “not a prime number between 1 and 6.” The number 0.833 is the prior probability that the asset value will be something other than $20 million. And finally, the number 0.4 is the probability that when something other than $20 million is the true asset value, the perfectly accurate forecasting instrument will still say Not Prime. That happens with $40 million and $60 million (i.e., twice in five possibilities).

The analysis here is easily modified to incorporate less-than-perfect forecasts from the marketing research, which is helpful because imperfect forecasts are the reality of business. See E. Rasmussen, Games and Informa- tion, 3rd ed. (Cambridge: Basil Blackwell, 2001), Chapter 13, Section 5. 7Open bidding with a nonrandom, structured sequence of role reversals on multiple auctions allows bidders to signal and punish one another (tit for tat) and therefore increases the likelihood of tacit collusion.

588 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Then, whichever company bids first, the other company will deduce the first bidder’s additional marketing research results and proceed to increase its bid in light of the more complete information available. For example, if PCS PrimeCo bids first, and bids $40 million based on its own asymmetric information, Wireless Co. will then be in a position to deduce that PCS PrimeCo’s marketing research excluded $10 million, $30 million, and $50 million as possible values. That’s the only information that would be consistent with a bid of $40 million in a simultaneous sealed-bid auction over an asset with a uniform distribution from $10 million to $60 million with only these six possible outcomes. Knowing from its own marketing research that $20 million, $30 million, and $50 million have also been ruled out, Wireless will immediately place a winning bid of just under $50 million:

1 2 ð$40 millionÞ + 1

2 ð$60 millionÞ = $50 million

With many more than these two bidders and other service areas in which Wireless Co. would be required to bid first and in which PCS PrimeCo had a turn playing the fast second, this sequential open-bidding auction design would work well. Winning bids would rise to the Bayesian expected asset values reflecting all available information, and highest-value users would receive the assets.

Strategic Underbidding in Private-Value Auctions9

One serious drawback of English open outcry auctions is the strategic reticence (tendency to underbid) that bidders exhibit. If the bidders have common information but different valuations (i.e., a so-called private-value auction), those with high

Example Open Bidding Simultaneous Auction of PCS Spectrum Rights8

Thirty firms ultimately participated in the broadband spectrum auctions. The FCC specified two 30-MHz blocks for each of 51 metropolitan service areas. A special feature of these metropolitan service areas was strong interdependencies in provid- ing service in contiguous service areas. Bidders were encouraged therefore to as- semble and reassemble efficient bundles of licenses as the auction progressed. Consequently, the FCC adopted multiple-round simultaneous auctions with open bidding to allocate spectrum rights. Each bidder was told there would be several rounds of bidding, all bids in each round were announced, and each bidder was allowed to cancel or amend bids from round to round. All bids in every metro area remained open as long as any bidding activity continued in any service area. The auction lasted 112 rounds over a four-month period. Using this auction de- sign, the FCC raised $7.7 billion. AT&T paid $49.3 million and Wireless Co. paid $46.6 million for the A block and B block spectrum rights in Louisville.

8Based on “Market Design and the Spectrum Auctions,” Journal of Economics and Management Strategy 6, no. 3 (Fall 1997); and “Sale of Wireless Frequencies,” Wall Street Journal (March 25, 1998), p. A3.

9Two excellent elaborations of this and the next topic are J. McMillan, “Bidding in Competition,” Games, Strategies, and Managers (New York: Oxford University Press, 1992), Chapter 11; and E. Rasmussen, “Auc- tions,” Games and Information, 2nd ed. (Cambridge, MA: Basil Blackwell, 1993), Chapter 12.

private-value auction Auction where the bidders have different valuations when information is complete.

Appendix 15A: Auction Design and Information Economics 589

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

willingness to pay have an incentive to refrain from aggressive bidding in an attempt to just exceed the bid of the player with the second highest valuation. For example, in the FCC’s spectrum auctions, the cellular phone incumbents already established in a metropolitan area had higher valuation than other bidders. In the early rounds of any such open bidding auction over private values, eventual high bidders hold back. Analysis of the FCC data suggests only 53 percent of the eventual winners were the high bidders after the early rounds. Sellers worry that this strategic reticence dam- pens the overall level of bidding throughout the auction and may well reduce final revenue.

Suppose two bidders each values a service or asset between $0 million and $10 mil- lion. No information about the actual net present values is known. That is, no common- value information, asymmetric or otherwise, is available. In this purely private-value auction that lasts only one round, the bids are sealed, and the highest bid wins. Your valuation is $6 million. What should you bid?

With two bidders present, each must assume that the other will offer something less than his or her private value, say k × v where k is a proportion and v is the private value.10 Any bids by Alice (Pa) that are greater than k times Bob’s value (vb) will win. That is, any time the value to Alice va = Pa/k > vb, Alice wins the auction and realizes a profit of va − Pa. With uniform density, the probability that Bob’s value is any given number between $0 and $10 million is 1/10 million. Again, Alice wins when this value is between $0 and Pa/k dollars. Therefore, Alice’s cumulative probability of winning is Pa/k events, each of which has a marginal probability of 1/10 million—that is, Alice’s cumula- tive probability of winning is Pa/(k × 10 million). Alice’s expected profit from the auc- tion may therefore be written as follows:

EðProfitaÞ = ðva − PaÞ Pak · 10,000,000 [15A.1]

Differentiating Equation 15A.1 with respect to Pa and setting the derivative equal to zero, Alice’s expected profit from the auction is maximized when

ðva − 2PaÞ 1k · 10,000,000 = 0 [15A.2]

that is, when Pa = va/2. Alice maximizes her expected profit from participating in the auction, conditional on Bob’s choosing a kvb bidding rule, by choosing to reduce her own private value by one-half. Because Alice and Bob are symmetrically situated in this bidding game, Bob too should reduce his private value by one-half. With k = 1/2, the players are in a Nash equilibrium. Each maximizes self-interest, conditional on the other player’s bidding v/2, by bidding half of his or her own private value. In a two-player, simultaneous, highest-price-wins-and-pays sealed-bid auction, the magnitude of rational underbidding is fully 50 percent!

In the case of five bidders, it is easy to show that a rational bidder should reduce true private value by one-fifth, and if n bidders, by 1/n.11 Quite intuitively, therefore, the more bidders, the smaller is the rational underbidding. See Figure 15A.2. Sellers under- stand this reasoning and therefore provide sorting services (in De Beers’ case) and hand- some catalogs and live exhibitions (in Christie’s case) to draw bidders into the auction

10This example relies on McMillan, op. cit., pp. 138, 208–209. 11See Rasmussen, op. cit., p. 296; and P. Klemperer, “Spectrum Auctions,” European Economic Review 46 (2002), pp. 829–845.

590 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

process. Paul Klemperer of Oxford University describes how this fundamental insight de- termined many of the auction design decisions for the British spectrum auctions. Thir- teen different companies entered the bidding, resulting in the highest per-capita revenue raised by all the European and Asian 3G auctions.

If sellers expend enough resources to expand the pool of bidders, in the limit the seller can realize (n – 1)/n of themaximumprivate value (vmax). Five bidders implies 80 percent of vmax. Ten bidders implies 90 percent of vmax. Twenty bidders implies 95 percent of vmax. If increasing the number of bidders from two to five results in a 30 percent increase in value, then doubling the number of bidders from five to ten results in a 10 percent

Example Exponential Valley Inc. Auctions a Chip Patent12

Exponential Valley Inc., a Silicon Valley microprocessor start-up, decided to auc- tion its portfolio of 45 issued and pending patents rather than move into pro- duction. The computer chip patents include features that would allow a competitor to match forthcoming chip products from industry leader Intel Corp. Intel regularly sues companies who make clones of their chips and has been effective in deterring entry with this strategy. The Exponential Valley pa- tents appear to offer an opportunity to provide protection from Intel’s patent infringement suits. At considerable expense, Exponential developed a large pro- spectus of technical and bidding information that it targeted to possible bidders: AMD, Chromatic Research Inc., and Texas Instruments. Exponential Valley did so because the larger the number of bidders, the less the strategic underbidding will be.

12Based on “An Auction of Chip Patents May Ignite Bidding War,” Wall Street Journal (August 1, 1997), p. B5.

FIGURE 15A.2 Strategic Underbidding Declines and Seller Revenue Rises with Bidder Entry in Private-Value Auctions

2 5 10 Number of BiddersMaximum Private Value

�10%

�20%

�50%

Strategic Underbid

Strategic Underbid

Strategic Underbid

Appendix 15A: Auction Design and Information Economics 591

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

increase in value, and doubling them again results in only a 5 percent increase in value. Diminishing returns to these efforts by sellers to expand the pool of bidders implies that the fundamental problem of strategic underbidding will be reduced but never eliminated.

Of course, underbidding in a private-value auction is rational only if bidders can be assured of winning in the final round. One way to reduce strategic underbidding in private-value auctions is to reduce the available information about other valuations by sealing the bids. A less extreme approach is to end the auction without warning, after several preliminary rounds, thereby in effect sealing the bids unpredictably. Of course, as we saw in the previous section, it can be in the seller’s interest to induce the revelation of all asymmetric information. In fact, as noted earlier, sellers often have an incentive to preannounce expert appraisals of value (as Christie’s and Sotheby’s auction houses do) in order to reduce the winner’s curse.

This concern is important in auction design, but not always controlling. The reason is that some asymmetrically held information can, if revealed, lower the rational bid (see the first Case Exercise at the end of this appendix). So sealing bids is a design alternative that becomes more attractive the greater the variation in private values and the more symmetric the information pertaining to valuation among the bidders. Obviously, open bidding or pre-announcing appraisals is more attractive whenever favorable information is known to the seller. However, even when the seller is in the dark, open bidding has a positive expected value for the seller because the exchange of common-value information always reduces the winner’s curse.

Second-Highest Sealed-Bid Auctions: A Revelation Mechanism Strategic underbidding is especially troubling, of course, if the seller is collecting bid rev- enue from all participants in the auction. A “seller” may be attempting to assess whether buyer willingness to pay justifies investment in a new facility (e.g., a ballpark, a pool, a set of tennis courts, or a clubhouse). Each potential user is asked what he or she would pay for access if the facility were to be built. If sufficient demand exists, the facility manager then builds the facility and collects the highly divergent, discriminatory prices from each “bidder.” The key to such an assessment is designing an incentive-compatible revelation mechanism. The same thing is true in designing a private-value auction. If, as an auction designer, one could remove the incentive to underbid and at the same time prevent a winner’s curse, one could align incentives with true revelation of value. Think through the following illustration of an ingenious incentive-compatible auction mecha- nism design that won William Vickrey the Nobel prize!13

Vickrey suggested that rather than requiring the winning bidder to pay the high bid, what if the auction mechanism design specified in advance that the highest bid wins but that the winner would pay a reduced amount only, equal to the second- highest bid? Think through the consequences of this radical idea. By definition, under the rules of a highest-pays-second-highest sealed-bid auction, the payment triggered by bidding one’s true private value cannot exceed the next best alternative selling price. This use of a verifiable second opinion and exit option to shore up the bidder’s pro- tection from suffering a winner’s curse was the key insight of William Vickrey’s

incentive-compatible revelation mechanism A procedure for eliciting true revelation of privately held information from agents with competing interests.

13Every game theory and mechanism design theory book describes the Vickrey auction, also known as the second-highest sealed-bid auction, or the uniform price auction. Vickrey’s original article is also revealing and insightful; see W. Vickrey, “Counterspeculation, Auctions, and Competitive Sealed Tenders,” Journal of Finance 16, no. 8 (1961), p. 37.

592 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

incentive-compatible mechanism design. True revelation of maximum willingness to pay proves compatible, through an ingenious mechanism design, with the bidder’s de- sire to avoid the winner’s curse.

To sum up, no bidder in a Vickrey auction has any incentive to underbid. Reducing your bid below your private value has no effect on the payment due should you win. Instead, underbidding in a second-highest sealed-bid auction when you are the highest- willingness-to-pay participant simply increases the probability of losing an auction asset that it would have been possible to acquire for less than it is worth to you. And if some- one else values the asset more than you, no payment is triggered by bidding up to your own private value. Therefore, for all possible cases, true revelation of private value dom- inates underbidding as a bidding strategy. And because bids are sealed and the auction

Example Second-Highest Sealed-Bid Auction: U.S. Treasury Bills14

Auction design decisions in security markets select procedures that will raise the most revenue for the issuers. On this question, worldwide debate currently rages about the optimal design of Treasury bill new-issue security auctions. Denmark and Sweden adopt diametrically opposed designs. The Swedes sell government bills and bonds at discriminatory Dutch auction prices; the Danes sell in Vickrey auc- tions at a uniform second-highest sealed-bid price. Given this diversity of expert opinion and practice, the U.S. Federal Reserve authorized the New York Federal Reserve Bank to experiment with both designs for two-year and five-year notes. Most Treasury auctions in the United States (and indeed around the world) are discriminatory descending-price (Dutch) auctions; buyers pay whatever they bid for quantities of T-bonds along a demand schedule submitted by each bidder. If the market-clearing price implies a yield of 5.03 percent, a typical bidder may get $5 million worth of T-bonds at her highest submitted price yielding 5.00 percent, $7 million worth of T-bonds at a lower price yielding 5.02 percent, and perhaps another $10 million at the market-clearing (lowest) price yielding 5.03 percent.

Uniform price second-highest sealed-bid auctions are different. Every bidder in this Vickrey auction might pay the uniform-slightly-higher price associated with a yield of 5.02 percent, and the Treasury’s marginal cost of raising debt capital will decline from 5.03 percent to 5.02 percent. However, some still higher-priced trans- actions no longer take place, and other lower-price transactions also no longer take place. Therefore, it is unclear analytically what will happen to Treasury revenue; it depends on the interest rate elasticity of T-bill demand.

This result highlights the insight that the principal advantages of second-highest sealed-bid Vickrey auctions are not to raise seller revenue, but rather to minimize strategic underbidding, reveal true private valuations, and reduce bidder collusion among small numbers of bidders. In contrast, security markets are highly efficient, with numerous potential buyers willing to pay a common value for these T-bond and T-bill assets. Therefore, second-highest sealed-bid mechanism design is more appropriate for the private placement IPO market employed in the Google IPO. Combining managerial insight about auction design with a careful analysis of the particulars of the business setting becomes crucially important.

14Based on “Bidding up Debt Auctions,” BusinessWeek (September 8, 1997), p. 26; and S. Nandi, “Treasury Auctions: What Do the Recent Models and Results Tell Us?” Federal Reserve Bank of Atlanta Economic Review (Fourth Quarter 1997).

Vickrey auction An incentive- compatible revelation mechanism for drawing out sealed bids equal to private value.

Appendix 15A: Auction Design and Information Economics 593

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

lasts only one round, no strategic bidding, false carding, or signaling can have any effect on other participants in the auction. Therefore, No Underbidding is a dominant strategy equilibrium for all bidders.15 Of course, realized revenue is still reduced by the full differ- ence between the highest and second-highest valuations, but at least the seller learns the true value of the auctioned asset.

Revenue Equivalence of Alternative Auction Types16

Under particular circumstances, the four simplest types of auctions (English ascending price, Dutch descending uniform price, first-price sealed-bid, and highest pays second- highest sealed bid) yield equivalent revenue to the seller-auctioneer. For example, both first-price sealed-bid auction participants and Dutch descending uniform price auction participants must pay whatever they bid if their bids prove to be the winning market- clearing bids. In addition, participants in both types have no access to information about bidders with lesser valuations than themselves. Consequently, the optimal bidding strat- egy in a Dutch uniform price or a first-price sealed-bid auction is identical, and thus the winning bids will be identical (see Table 15A.4).

Similarly, for private-value auctions, as English auction participants learn more and more in the course of the bidding about the independent valuations of other bidders, the person with the highest valuation will ultimately offer an amount just in excess of the second highest bidder. Second-highest sealed bid auctions induce, as we have seen, true revelation of value from all bidders, but the winner also pays an amount just equal to the second high- est bidder. Hence, an English ascending price auction and a second-highest sealed bid auc- tion yield essentially the same expected revenue to the seller in private-value auctions. Indeed, it turns out that with risk-neutral bidders and private values, all four simple auction types result in the same expected revenue (i.e., an amount just equal to the highest value minus the full difference to the next highest bid). From a seller’s perspective, this revenue equivalence theorem (RET) is a depressing result; one would hope to do better.

In several instances, one auction type does raise more expected revenue than another. These optimal mechanism design differences depend upon the risk preference of the bid- ders and upon the common-value or private-value nature of the item being auctioned. Continuing first with the private-value auctions, if bidders are risk averse and they are operating under the lack of information of a Dutch or first-price sealed-bid design, they will seek to reduce the probability of losing the item when their valuation is highest. Consequently, relative to the situation in an English or second-highest sealed-bid auction in which the winning bidder pays essentially the second-highest bid, risk-averse Dutch or first-price sealed-bid auction participants raise their bids in order to reduce the probabil- ity of being outbid by a close second valuation. Strategic underbidding to avoid the win- ner’s curse is still present but it is mitigated by risk aversion. So, a seller-auctioneer can raise more revenue on average when the bidders are risk averse and the values are inde- pendent and private by conducting a Dutch or first-price sealed-bid auction. These re- sults for private-value auctions for items such as patent licenses, sales territories, antiques, and fine art are summarized in the top section of Table 15A.4.

As to common-value auctions for items that have thick resale markets such as crude oil, mineral leases, forest logging rights, equity and debt securities, and easily redeployable

15The bidder’s degree of risk aversion has no bearing on this result. However, more risk-averse bidders do en- sure against losing in winner-take-all first-price sealed-bid auctions by increasing their bid relative to the seller revenue from a second-highest sealed-bid auction. We discuss the role of risk aversion further in the next section. 16For a more extensive discussion of this topic, see A. Dixit, D. Reiley, and S. Skeath, Games of Strategy, 3rd ed. (New York: Norton, 2009), Chapter 15; and E. Rasmussen, op. cit., Chapter 12.

594 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

surplus equipment such as commercial delivery trucks and corporate jet aircraft, the source of uncertainty in the valuation is an estimation risk. Every bidder knows that the true value at resale is identical across all auction participants; it’s just that this true value is an unknown while the oil is still in the ground, the logs still in the forest, the IPOs yet unissued, and so on. Each bidder must therefore assess this true value from his or her own forecast information in a Dutch uniform price or first-price sealed-bid auction and from any additional information that can be gleaned from the sequence of bids in the En- glish auction. The Bayesian updating of the initial estimates (the “priors”) in the sequen- tial process of an English auction will tend to result in a pooling of the bidders’ information; therefore, winning bids in the English auction will tend toward the mean estimate of the population of forecasts. Mechanisms that facilitate the learning of this best unbiased estimate of what the harvestable common value of the resource truly is re- duce strategic underbidding to avoid the winner’s curse.

Hence, as shown in the bottom section of Table 15A.4, in selling common-value items, a seller-auctioneer can raise the most revenue with an English ascending price auction. And as we saw in the previous section, second-highest sealed-bid designs can also substantially reduce the winner’s curse in these common-value auctions, raising the seller’s projected revenue relative to Dutch uniform price auctions and first-price sealed bid auctions that make it hard for the bidders to reduce their estimation risk.17

A still larger auction design issue is whether to continue the traditional ascending-price (English) auction or adopt descending-price (Dutch) auction procedures. Basement.com and OutletZoo.com start prices high and drop them in increments until all the units for sale have been demanded. Clearly, sellers can realize more revenue in discriminatory Dutch auctions by charging differentially higher prices to the early bidders than can be realized in uniform-price Dutch or English auctions that identify a market-clearing price. Of course, sophisticated institutional and industrial buyers understand this concept as well, and this

TABLE 15A.4 SELLER EXPECTED REVENUE FROM AUCTION TYPES

PRIVATE-VALUE AUCTIONS

(PATENT LICENSE, SALES TERRITORY, ESTATE ANTIQUES)

Risk-Neutral Bidders Dutch = First-Price = English = Second-Highest

Uniform Price Sealed-Bid Sealed-Bid

Risk-Averse Bidders Dutch = First-Price > English = Second-Highest

Uniform Price Sealed-Bid Sealed-Bid

COMMON-VALUE AUCTIONS (MINERAL LEASE, LOGGING RIGHTS, AIRCRAFT)

Risk-Neutral Bidders English > Second-Highest > Dutch = First-Price

Sealed-Bid Uniform Price Sealed-Bid

Risk-Averse Bidders English > Second-Highest 0 Dutch = First-Price

Sealed-Bid Uniform Price Sealed-Bid

17Risk aversion introduces the same complication into the analysis as earlier—that is, Dutch uniform price and first-price sealed-bid participants raise their bids to reduce the probability of losing an item to a close second competitor. Still, the benefits of information pooling tend to favor English or second-highest sealed-bid designs for common-value auctions.

Appendix 15A: Auction Design and Information Economics 595

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

segment may well prefer a traditional non-auction Web site such as Grainger.com where sellers post initial best offers and then focus on availability, delivery, installation, technical support, and other after-sale services. Establishing the value-in-use of these extras in the “to- tal offering” may prove as critical as seller warranties and replacement guarantees to B2B customers.19

Example Internet Auction Design Becomes a Big E-Business Debate: eBay versus Priceline18

Online auctions experienced a truly explosive rate of Internet site development and attracted tremendous investor interest. Eight sites existed in July 1998, 400 existed in July 1999, and more than 3,000 sites existed by July 2000. Some of the sites is- sued public shares and realized massive equity market value. In 2005, eBay Inc. reached 150 million listings. Shortly after its initial public offering in 2000, Price- line exceeded the combined market value of the major airlines (American, Delta, United) whose unsold last-minute tickets it offered to liquidate.

In the information economy, online auctions are a key business model that may displace the two traditional price-setting processes: (1) one-on-one negotiation (haggling) and (2) a menu of fixed price quotes provided by the seller. Business- to-business (B2B) transactions may continue to require one-on-one negotiation over the time, quality, availability, and delivery dimensions of the “deal,” but in business-to-consumer transactions, auction prices themselves establish most of what is needed for a “deal.”

Between 2000 and 2002, however, Priceline’s share price fell 90 percent. For one thing, Priceline failed to secure a broad patent for its reverse auction design, sug- gesting Priceline will not be protected from subsequent imitators. Priceline employs an ascending-posted-price highest-wins-and-pays auction mechanism design. Bids are listed anonymously to reduce collusion and posted continuously to mitigate the strategic underbidding that accompanies any private-value auction (i.e., the win- ner’s curse). Bid payments must be guaranteed with a credit card at the time the buyers’ offers are placed; subsequently, bids cannot be cancelled and are executed automatically if the offer is accepted.

eBay has a more transparent ascending-price auction design with posted-prices and highest-wins-and-pays rules for declaring a winner. Rather than allowing sell- ers to remain anonymous until they decide to “hit” a Priceline buyer’s posted offer, eBay consolidates feedback on the seller’s past performance and hot-links it to the bidding site. Also, unlike Priceline, eBay allows open bidding (i.e., offers are re- peated in multiple rounds, and prior offers may be cancelled or amended). If open bidding proves to be optimal for auctioning airline tickets, automobiles, and machinery, eBay rather than Priceline will thrive.

18Based on “The Heyday of the Auction,” The Economist (July 24, 1999), pp. 67–68; “Redesigning Business: Priceline,” Harvard Business Review (November/December 1999), pp. 19–21; “Dotty about dot.commerce?” The Economist (Feb- ruary 26, 2000), p. 24; “Going, Going, Gone, Sucker!” BusinessWeek (March 20, 2000), pp. 124–125; and “Inside: Is Priceline Vulnerable?” Harvard Business Review (December 3, 1999), pp. 19–21.

19See A. Kambil and E. van Heck, Making Markets (Boston: HBS Press, 2002); J. Anderson and J. Narus, Busi- ness Market Management (New York: Prentice Hall, 1999); and R. Oliva, “Sold on Reverse Auctions,” Market- ing Management (March/April 2003), p. 45.

596 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Contractual Approaches to Asymmetric Information in Online Auctions In some ways, Priceline.com is like couponing without the brand-name exposure. This provides a way to price discriminate to the price-sensitive customer segments without degrading one’s brand identity. Perhaps this explains why Delta Airlines took a substan- tial equity stake in Priceline; Delta could liquidate its inventory without cannibalizing higher fare sales. But anonymous sellers offering unverifiable product or service claims on Priceline must credibly commit to higher quality products if they hope to attract any- thing more than rock-bottom prices. As we have seen, hostage or bonding mechanisms are the key to credible commitments.

First, sellers can invest in and disseminate appraisals to try to distinguish their auc- tion offerings from fraudulently advertised products. Independent certified appraisals, such as buy-back guarantees, place a floor under the value of the product offered for sale. Perhaps a tract of development land or the intellectual property in a patent lease is up for bids; the seller can pay for an appraisal about the range of values typically paid in resale markets for land or intellectual property assets with similar characteristics. Two drawbacks prevent the widespread adoption of this approach, however. Certified appraisals are expensive, and appraisals seldom establish maximum or unique asset value.

The second contractual approach sellers can adopt to establish credibility for their auction claims is to signal their respective quality by offering warranties and replacement guarantees. These signaling features of the product offering are observable to the buyer at the point of sale and are highly correlated with the unobservable product or service qual- ity at issue. Automobile tire manufacturers who take shortcuts vulcanizing the tire cas- ings that embed steel wire belts in their tire treads will make shorter treadwear warranties and provide guarantees against fewer blowout hazards. The buyer can there- fore pay more for tires with extensive warranties and feel confident that the tire quality is above that offered by non-warrantied suppliers. Encouraging buyers to screen alternate suppliers in this way therefore achieves a separating equilibrium of fraudulent versus re- liable suppliers without the more costly extensive independent appraisal of each tire.

Another way for sellers to credibly commit to the delivery of a high-quality durable product is to offer to lease rather than sell the product and then to accept lease terms with a high residual value. This feature of leasing offers a net advantage over buying in that sellers with informational advantages will credibly commit to forward value through the take-back provisions of the lease. If one seller says you can lease with a 60 percent residual at the end of four years and another quotes a 40 percent residual at the end of four years, all other things the same, the lease payments will recover depreciation only two-thirds as great in the 60 percent residual lease. In that case, buyers may well be will- ing to lease a higher-quality/higher-priced product. Of course, such an approach to es- tablishing credibility will not be adopted if sellers foresee substantial obsolescence risk. Therefore, high residuals seldom occur alone; usually other lease terms (e.g., financing charges, initial asset prices, or lease closing fees) adjust upward to “price in” the extra seller risk when residuals increase.

Finally, sellers can agree to accept contingent payments—that is, seller revenue depen- dent on the performance that the buyer experiences. Suppose the due diligence required to establish clear value in an asset sale is prohibitively expensive. If a tract of development land has buried fuel tanks that necessitate substantial environmental remediation but the pres- ence of which is unknown, the land seller can agree ex ante to pay for the restoration of the land. Thereby, the contingent risk is insured away by a seller’s credible commitment to re- store the land should it prove damaged. Similarly, if timberland sales or oil field leases prove

appraisals An estimate of value by an independent expert.

contingent payments A fee schedule conditional on the outcome of uncertain future events.

Appendix 15A: Auction Design and Information Economics 597

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

particularly productive, the buyer and seller can agree to larger money payments than if the actual harvested timber and oil pumped out prove disappointing.

These contingent payment contracts can be arranged as progress payments, so that the buyer and seller take small and continuous steps while money remains owed. Check- lists of building problems that remain to be corrected often trigger the final small contin- gent payments between new homeowners and their builders. Sometimes corporations employ contingent payments while exchanging hostages in an asset sale by requiring the seller to take a financial stake (perhaps 15 percent of the cash flow to equity owners) in the buyer’s subsidiary spun off to manage the new assets. Given the low priority of an equity owner should the spin-off declare bankruptcy and need to liquidate, contingency payments increase the seller’s incentives to reveal hidden features that determine the true cash flows realizable from the asset.

INCENTIVE-COMPATIBLE REVELATION MECHANISMS Perhaps the most powerful mechanism design tool for drawing out privately held asym- metric information is William Vickrey’s self-enforcing revelation mechanism. Later, Ed- ward Clarke and Ted Groves added the idea of multiple agents in group decision making such that the mechanism design objective became demand or cost revelation in a part- nership that had to be compatible with the incentives of all the parties: an incentive- compatible (IC) revelation mechanism.

Example Intel and Analog Inc. Form Partnership to Develop DSP Chip20

Intel Corp. has dominated the manufacture of semiconductor chips for computer microprocessors for more than a decade. With AMD and Siemens beginning to pose some threat in this traditional market, Intel formed a joint venture in 1999 with Analog Devices Inc. to move into the chip market for communications de- vices such as cell phones, pagers, and wireless videophones. Intel and Analog jointly developed a new line of digital signal processor (DSP) chips. DSP chips take analog signals like voice, photo images, and video and convert them to digital signals to be transmitted over wireless systems. Given the enormous growth in wireless communications, new signal compression and encryption capabilities of the Intel-Analog chip are expected to compete well against rival DSP suppliers Texas Instruments (TI) and Lucent Technologies. DSP chips also appear to have application in modems and other networking devices that provide high-speed ac- cess to the Internet. Speech-recognition systems for machine-controlled applica- tions are a complementary value-adding technology. By 2010, DSP chips will provide the functionality of a laptop computer on a thumbtack that fits into wristwatch-size devices. DSP chip sales growth has recently approached 30 percent per year, reaching total sales of $5.7 billion.

Groups of Intel and Analog engineers will collaborate on designing the core ar- chitecture of the chip, and the two companies will then separately develop and sell products based on the design. The contingent payoffs to each firm are therefore based in part on their cooperative design success and in part on their separate product development and marketing efforts. This joint venture contract provides

(Continued)

598 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Cost Revelation in Joint Ventures and Partnerships When pivotal information in such partnerships is privately held and verification by third parties is infeasible or undesirable, the partners seek to adopt procedures to assure true revelation of this asymmetric information. Consider a joint venture to develop several new personal computer products between a PC designer-manufacturer, such as Apple Computer and Motorola, a leading supplier of computer chips.21 The Apple operating system depends on the capabilities of the Motorola chips, and the chips are produced in anticipation of the requirements of the operating system. The partners believe they can better sustain a competitive advantage in this fast-moving technology by jointly de- veloping new products. After the joint venture covers development and production costs, they agree to split the profits equally.

Each partner in the joint venture has private information about cost to which the other partner does not have access. For example, as it develops the iPad, Apple discovers its operating system development costs, and Motorola discovers its computer chip design and production costs. Although neither can independently verify the other partner’s asymmetric information, the success of a joint venture often depends on each partner’s ability to generate enough operating profits to recover these development costs. As the partnership reaches project milestones, determining profit potential requires an accurate revelation of true costs. Let’s see why and what can be done to achieve this goal.

The study of incentive-compatible revelation mechanisms can provide some answers. Each partner faces random disturbances in its cost factors.22 Sometimes software devel- opment is delayed by inconspicuous but debilitating bugs in the programming, which increase the cost from, say, $80 to $120 million. Similarly, sometimes chip development and production necessitates redesign (e.g., Intel’s problems with the Pentium chip), in- creasing that cost from, say, $50 to $70 million. Neither partner can hope to discover and rectify all such problems in advance. However, each can detect early warning sig- nals of cost overruns and, if need be, cancel that aspect of their joint venture.

Cost Overruns with Simple Profit-Sharing Partnerships When both cost overruns happen simultaneously, the joint venture should shut down, because the development costs of proceeding to full-scale production ($120 million + $70 million) will exceed the projected revenue available, say, $180 million. These pro- jected operating profits and losses appear in Table 15A.5. If Apple experiences $120 mil- lion cost (the column labeled High Costs), the partnership will cancel the project

incentives for continuing cooperation far beyond those generated by a simple profit-sharing agreement, yet it preserves each company’s option to pursue some business plans privately.

20Based on Intel Corporation and Analog Design Inc. joint press releases (February 3, 1999); and “TI Lays Out DSP Plans until 2010,” Hardware Reviews and News on the The-View.com (December 6, 1999).

21The general structure of this section relies on A. Dixit and B. Nalebuff, Thinking Strategically (New York: Norton, 1991), pp. 306–319. The illustration here is based on “Apple Wants Other PC Makers to Build Com- puters to Use Macintosh Software,” Wall Street Journal (January 28, 1994), p. B5; and “IBM, Apple in PC De- sign Accord,” Wall Street Journal (November 8, 1994), p. B5. 22Similar arguments can be made about asymmetric information regarding random disturbances in demand.

Appendix 15A: Auction Design and Information Economics 599

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

whenever Motorola also experiences high cost of $70 million (the row labeled High Costs) because proceeding would result in a $10-million operating loss. By the same token, when only one partner or neither partner experiences high costs, the joint venture project should go forward and realize profits of $30 million, $10 million, and $50 mil- lion, respectively. Only with correct operate and shutdown decisions can the joint ven- ture generate its maximum value.

The incentive problem is that initially each partner has an incentive to overstate true costs in order to be overcompensated from the joint venture revenues. For exam- ple, in Table 15A.5, if Apple reveals true costs of $80 million and Motorola claims costs of $70 million when in fact its true costs are $50 million, Motorola’s joint profit share declines by $10 million from one-half of $50 million (top left cell) to one-half of $30 million (lower left cell). But with $20 million extra reimbursement from overstat- ing its cost, Motorola ends up with (1/2) $30 million + $20 million, which exceeds one-half of $50 million by $10 million. Similarly, if Apple overstates its costs, the Apple profit share falls from $25 million to $5 million, but this decline is more than offset by the $40 million extra reimbursement for overstating its $80 million actual cost to $120 million.

If low cost and cost overruns are equally likely at Motorola and if the probability of a cost overrun at Apple is 0.3, then expected costs are $60 million at Motorola and $92 million at Apple. With true revelation of costs, expected net profit from the joint venture is then (0.5 × 0.7) $50 million + (0.5 × 0.3) $10 million + (0.5 × 0.7) $30 million + (0.5 × 0.3) $0 = $29.5 million, or $14.75 million for each partner.23 However, if one or both partners overstate costs, the projects with mixed costs in the southwest and northeast cells of Table 15A.5 will also be canceled, and the expected net profit from the joint ven- ture then declines. For example, if Apple falsely reveals $120 million when low costs of $80 million are present, the joint development project is canceled whenever Motorola experiences $70 million cost. This cancellation results in the partners forgoing the $30 million profit on the mixed-cost project in the southwest cell and reduces the expected value of the joint venture to $19 million (i.e., $9.5 million per partner).24 Value- maximizing managers facing asymmetric information seek some revelation mechanism that will provide appropriate incentives to induce the revelation of true costs, thereby preserving and capturing the full $14.75-million-per-partner expected value of both the low-cost and the mixed-cost projects.

TABLE 15A.5 JOINT PROFITS (IN MILLIONS) FROM A SIMPLE PROFIT-

SHARING PARTNERSHIP WITH $180 MILLION IN REVENUE

APPLE

LOW COSTS ($80) HIGH COSTS ($120)

Motorola Low Costs ($50) $50 $10

High Costs ($70) $30 −$10

23Note that the project in the southeast cell of Table 15A.5 is canceled because of mutual early warnings of high cost and therefore a projected operating loss. 24This expected value is calculated as (0.5 × 0.7) $50 million + (0.5 × 0.3) $10 million = $19 million.

600 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Clarke-Groves Incentive-Compatible Revelation Mechanism One such revelation mechanism is known as the Clarke tax mechanism.25 Edward Clarke’s path-breaking idea was that to create appropriate incentives for asymmetric cost (or demand) revelation in a partnership, each party’s revelation should trigger an imposition of the expected costs on (and the forgone profit opportunity losses suffered by) the other partners. In this way, the maximizing incentives of each of the asymmetri- cally informed partners could be made compatible. For our PC product development ex- ample, Table 15A.6 indicates the revenue shares each partner would receive under a Clarke tax mechanism. The row player Motorola gets the below-diagonal payoffs in each cell, and the column player Apple gets the above-diagonal payoffs in each cell.

After the other party’s expected costs are covered, each partner’s payoff is then recal- culated as the residual or net revenue share from all non-canceled projects triggered by its own cost revelations. To illustrate, if Motorola reveals Lowm cost, the current project will proceed independent of Apple’s cost, and Motorola will realize $88 million, which is $180 million total partnership revenue minus the $92 million expected cost of Apple:

Expected Net Revenue Share ðLow for MotorolaÞ = $180 million − ½ð0:7 × $80 millionÞ + ð0:3 × $120 millionÞ�

= $88 million

This figure appears in the third column of Table 15A.7. However, if Motorola announces Highm cost, the project is canceled whenever Apple detects early warning signs that its own cost is Higha. Consequently, should Motorola decide to reveal high cost when low cost is present, its net revenue would fall from $88 million to 0.7 ($180 million – $80 million) + 0.0 ($180 million – $120 million) = $70 million, also listed in the third (Net Revenue Shares) column of Table 15A.7. Motorola’s net revenue share declines because of a zero probability of realizing the $60 million revenue share in the northeast cell of Table 15A.6. The false overstatement of cost by Motorola results in that project being canceled, and everyone loses. Under a Clarke tax mechanism, not just actions but infor- mation revelations themselves have consequences. And, as we shall see, building these consequences into a reimbursement system can induce the true revelation of partnership costs.

TABLE 15A.6 INDIVIDUAL REVENUE SHARE NET

OF PARTNER COST (MILLIONS)

Apple

Lowa ($80) Higha ($120)

Lowm ($50)

Motorola

Highm ($70)

P (LOWa) 0.7 P (HIGHa) 0.3 P (LOWm) 0.5 P (HIGHm) 0.5

$130 $130 $100 $60

$110 $0 $100 $0

Note: Column-player payoffs are above diagonal. Row-player payoffs are below diagonal.

25This revelation mechanism is also referred to as the Clarke-Groves revelation mechanism after Ted Groves, who formalized the concept, thereby showing a different connection to William Vickrey’s earlier work on incentive-compatible auction design.

Appendix 15A: Auction Design and Information Economics 601

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The importance of the discovery of such incentive-compatible revelation mechanisms can hardly be overemphasized; they have led to many path-breaking private sector and public policy applications. Clarke first developed the concept in the context of the true demand revelations needed in consumption partnerships to finance a jointly consumed park, pool, or playground.26 The demand revelation problem in assessing an optimal user tax share in a consumption partnership is analogous to the cost revelation problem in assessing an optimal profit share in a business partnership.

An Optimal Incentives Contract To organize a joint venture or partnership around a Clarke-Groves revelation mecha- nism usually involves the implementation of a so-called optimal incentives contract. Each party agrees in advance to a set of partnership net revenue shares associated with the expected payoffs from a revelation mechanism (see the third column of Table 15A.7). The important thing to appreciate is that the problem of independently verifying asym- metric information has not gone away. A third party attempting to enforce the contract (e.g., a U.S. district court) would still have just as much trouble verifying the claims for cost reimbursement arising under this contract as the parties had in trying to verify their own partner’s cost. Entering into a partnership incentives contract does not escape the asymmetric cost information problem. Instead, the revelation mechanism creates incen- tives for a self-enforcing reliance relationship between the partners, not unlike the incentive-compatibility constraint we described in Chapter 15 as characterizing the opti- mal incentives contract between owner-principals and manager-agents.

TABLE 15A.7 EXPECTED NET PROFIT SHARES (MILLIONS) WITH

TRUE COST REVELATION UNDER AN OPTIMAL

INCENTIVES CONTRACT

APPLE

PROBABILITY

NET REVENUE SHARES

PROJECTED COSTS

NET PROFIT SHARES

Lowa 0.7 $120 $80 $28

Higha 0.3 $ 65 $60 $ 1.5

Expected Value $103.5 $74 $29.5

MOTOROLA

PROBABILITY

NET REVENUE SHARES

PROJECTED COSTS

NET PROFIT SHARES

Lowm 0.5 $88 $50 $19

Highm 0.5 $70 $49 $10.5

Expected Value $79 $49.5 $29.5

26To build the appropriate size urban park or swimming pool requires private information about use value and willingness to pay. However, if one asks potential demanders who assume their answer will determine their tax share, the respondents will understate their willingness to pay. See Edward Clarke, Demand Revela- tion and the Provision of Public Goods (Boston: Ballinger, 1980). For more on the applications of revelation mechanisms, see R. Cornes and T. Sandler, “Clarke’s Demand-Revealing Mechanism,” Theory of Externalities, Public Goods, and Club Goods (New York: Cambridge University Press, 1986), pp. 105–108; and Hal Varian, Intermediate Microeconomics, 8th ed. (New York: Norton, 2009).

optimal incentives contract An agreement about payoffs and penalties that creates appropriate incentives.

self-enforcing reliance relationship A non-contractual, mutually beneficial agreement.

602 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The structure of incentives underlying Table 15A.7 is fully capable of inducing the partners to reveal their true costs; each would be worse off not doing so. We have al- ready seen how Motorola would be worse off overstating its cost. Similarly, if Apple were to overstate its cost, profitable projects in the southwest cell of Table 15A.6 would be canceled. Rather than realizing 0.5 ($130 million) + 0.5 ($110 million) = $120 million from the good fortune of incurring Lowa cost, Apple would instead realize only 0.5 ($130 million) = $65 million, which fails to cover its own low-cost realization of $80 million. In addition, this false overstatement of cost and the cancellation of the profitable project in the southwest cell reduces Apple’s expected receipts from the partnership to just (0.5 × 0.7) $130 million + (0.5 × 0.3) $130 million = $65 million, whereas with true revelation it realizes $103.5 million: (0.5 × 0.7) $130 million + (0.5 × 0.7) $110 million + (0.5 × 0.3) $130 million = 0.7 ($120 million) + 0.3 ($65 million) = $103.5 million. Truth-telling dominates false revelation for both partners.

We can now also explain why both Apple and Motorola would adopt an optimal in- centives contract that credibly commits each partner to a true revelation of asymmetric cost information. Apple realizes an expected net profit with true revelation of $103.5 mil- lion expected receipts minus expected costs of $74 million (i.e., $29.5 million), shown in the last column of Table 15A.7. And similarly, Motorola realizes an expected net profit of $79 million expected receipts minus $49.5 million expected costs (i.e., $29.5 million). Each of these amounts equals the $29.5 million joint profits potentially available in the original simple profit-sharing contract of Table 15A.5. However, recall that each party knows in advance that the other party will have private information about cost overruns. Each could therefore predict that the simple profit-sharing partnership would lead to cost overstatement, cancellation of the mixed cost projects, and loss of value. This proactive reasoning implies that only the mutual low-cost outcome in Table 15A.5 will escape cancellation and actually generate profit. Therefore, only a much smaller expected profit—that is, just 0.5 × 0.7 ($50 million) = $17.5 million—is assured by the simple profit-sharing contract. This smaller amount from simple profit sharing is what rational parties choosing among partnership contracts would compare to the $29.5 million expected net profit from an optimal incentives contract.

INTERNATIONAL PERSPECTIVES

Joint Venture in Memory Chips: IBM, Siemens, and Toshiba27

IBM once entered into an agreement with Siemens and Toshiba to co-produce computer memory chips. At the same time, AMD and Intel announced similar joint ventures to develop flash memory chips with Fujitsu and Sharp, respectively. Flash chips retain the information needed to restart computer operating systems when the power is interrupted. In all three cases, the Japanese firm will contribute its superior manufacturing capability, and the American and German firms will contribute their product design and innovative research capabilities.

The key question in such joint ventures is whether the Western companies will simply give away their technological knowledge while their Japanese part- ners deliver little asymmetric information in ex- change. To ensure an evenly balanced partnership, the manufacturing know-how of the Japanese will be dissected as production cost information under various market conditions to be revealed and ana- lyzed by the joint venture partners.

27Based on “Pragmatism Wins as Rivals Start to Cooperate on Memory Chips,” Wall Street Journal (July 14, 1993), p. B1.

Appendix 15A: Auction Design and Information Economics 603

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The application of incentive-compatible revelation mechanisms and optimal incen- tives contracts has led to many exciting new types of asymmetric information partner- ships. The same principles also underlie the concept of an efficient breach of contract in the economics of contract law. When one partner breaches a contractual relationship, the legal remedies take into account the opportunities forgone and expectation damage costs imposed on the partner who does not breach.28 These concepts have become a key for achieving partnership or joint venture success in both small firms and large corpora- tions under asymmetric information.

Implementation of IC Contracts Incentive-compatible contracts are implemented with contingent claims contracting, a standard form instrument for sophisticated business relationships. The parties agree on the projected probabilities for the various levels of cost, the likely auditable joint operating profit for the partnership, and the unobservable reimbursable cost to each party in each contingency. These agreements form the contract expectations and de- fine the contract damages should unforeseen events induce either party to breach the contract.

The cost information revelation leads to efficient cancellation or go-ahead decisions. More generally, of course, consequences other than project cancellations can result from the information revelation of one partner. The information revealed can cause an expan- sion or contraction of the R&D efforts, prototype development efforts, marketing re- search efforts, etc. of the other partner. And each such revelation of misinformation fails to maximize what both can agree in advance would be the optimal course of action in each contingent state. It is those expected opportunity losses that the Clarke tax mech- anism then deducts from joint profits to find IC contract receipts. This process, of course, can prove quite a bit more complicated than the preceding example suggests; it presents a big challenge for the negotiating team of corporate attorneys.

In addition, you may have already noticed one further issue. The sum of the individ- ual net revenue shares in every cell of Table 15A.7 when the project goes ahead is greater than $180 million, the projected operating profit from the partnership. Thus, IC revela- tion mechanisms generally fail to break even; instead, they “break the budget.” Specifi- cally, if both parties declare high cost, each is entitled to an IC payout ($100 million in the case of Motorola and $130 million in the case of Apple) that together break the bud- get. To emphasize the generality of this result, the particular example has been con- structed to exhibit “budget breaking” in each contingent event, except cancellation. More typically, some cells would exhibit surplus, and others would exhibit deficit. Still, the deficit cells may arise first. What is the partnership to do? What implementation procedure can handle this deficit budgeting problem?

Recall that both partners have substantially greater net profit shares under the IC con- tract ($29.5 million) than the expected profit from a simple profit-sharing agreement ($17.5 million). Eliciting true information revelation really does have value, and both parties therefore would be willing to make ex ante commitments to cover such a deficit in the partnership. Examining the third column of Table 15A.7 shows that Apple’s ex- pected net revenue share is $103.5 million, while Motorola’s is $79 million. Conse- quently, $182.5 million − $180 million = $2.5 million per period (perhaps $2.5 million/0.05 = $50 million as a capital sum to cover the perpetual expected deficit)

28An excellent supplemental reading on efficient breach of contracts is R. Cooter and T. Ulen, Law and Eco- nomics, 5th ed. (Glenview, IL: Pearson Addison-Wesley, 2007).

604 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

must be posted as a bond to implement the IC contract procedure. Each partner would be asked to establish credible commitment to the IC contract partnership by investing $25 million ex ante to achieve an increase in expected profit of ($29.5 million − $17.5 million) per period, worth perhaps $12 million/0.05 = $240 million.30 Although compli- cated, these IC contracts clearly make sense for value-maximizing managers.

INTERNATIONAL PERSPECTIVES

Whirlpool’s Joint Venture in Appliances Improves upon Maytag’s Outright Purchase of Hoover29

Sometimes joint ventures are designed to increase the value of assets sold in a phase-out partnership rather than an immediate sale. As a potential buyer of Phi- lips’s European appliance division, Whirlpool sought access to more private information than due diligence by its merger and acquisition attorneys could uncover. Philips had a consumer franchise of nine appliance brands and a pan-European network of retail dealers who were second only to Electrolux in market share. But like other intangible assets (e.g., pivotal human resources and technical know-how), brands and distri- bution relationships are notoriously hard to value. In a new corporate organization and culture, could the Phi- lips brands be redeployed without Philips’ extremely strong reputation in European electronics? Would the fragmented network of independent dealers remain loyal once Whirlpool’s name was substituted for Phi- lips? And most importantly, what cost savings could be realized by sourcing all of the design, procurement, and production of Whirlpool and Philips components glob- ally to achieve economies of scale?

These questions were best answered by a joint venture in which Philips retained a 47 percent own-

ership stake, and Whirlpool immediately assumed management control in exchange for $381 million. After both parties shared cost and demand informa- tion for three years and fully assessed potential value, the remainder of the business was sold to Whirlpool for $610 million.

In contrast, Maytag satisfied its strategic plan to enter the European market by purchasing outright Chicago Pacific Corporation, whose Hoover Appli- ance division had a substantial retail dealer network in Britain. However, Maytag knew little about the growing retail power of superstore chains near British shopping malls and still less about the marketing re- search on British households. Consequently, Maytag stumbled from one promotional blunder to another and eventually sold the Hoover European subsidiary at a $130 million loss. Again, with carefully designed incentives, a joint venture could have elicited the rev- elation of valuable asymmetric information for the greater success of Maytag’s European initiative.

29Based on A. Nanda and P. Williamson, “Use Joint Ventures to Ease the Pain of Restructuring,” Harvard Business Review (November/December 1995), pp. 119–128.

SUMMARY

� First-come, first-served is a mechanism design for servicing a queue that reduces seller revenue be- cause of predictable congestions and expected

waiting time. Last-come, first-served introduces leave-and-return transaction costs and therefore also reduces seller revenue.

30Technically, the investment to cover projected deficits will change an individual household’s behavior unless we impose the restriction of quasi-linear preferences. In company settings, this assumption is plausible; see Varian, op. cit., pp. 274–277.

Appendix 15A: Auction Design and Information Economics 605

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

� Stratified lotteries and auctions can relieve conges- tion and reduce transaction costs in the queue, raising seller revenue.

� Auction design choices are multifaceted but at the simplest level always include who pays, what amount, and how the winner is determined. Simple auction types are English ascending-price auctions, Dutch descending-price auctions, first-price sealed- bidauctions, andsecond-highest sealed-bidauctions.

� Auctions also differ in the resale opportunities available to the participants. Common-value auc- tions have thick resale markets where the items can be easily resold at a consensual fair market value. Private-value auction items have no common re- sale value and instead involve assets with differing valuations to the auction participants.

� The winner’s curse implies that strategic underbid- ding is rational when the seller or other buyers have asymmetrically advantaged information about a common-value auction.

� Open bidding is a procedure for posting the offers in multiple rounds with cancellation and modifica- tion privileges to induce auction participants to pool their information about estimates of value. Open bidding reduces the winner’s curse and raises expected auction revenue in common-value auctions.

� What simple auction types raise the greatest ex- pected revenue for the seller-auctioneer depends upon the common-value or private-value nature of the item being auctioned and on the auction participants’ risk aversion.

� Dutch auctions and first-price sealed-bid auctions have identical information structures and identical bidding strategies, and therefore they generate identical expected revenue to the seller.

� Relative to Dutch or first-price sealed-bid auctions, English ascending-price and second-highest sealed- bid auctions raise the seller’s expected revenue in common-value auctions for items such as crude oil, forest logging rights, and aircraft because they en- courage the most pooling of bidder information. In private-value auctions, bidders who are risk averse offer higher bids and therefore generate more auctioneer-seller revenue in Dutch and first-price sealed-bid auctions.

� To escape adverse selection and elicit high-quality auction goods necessitates some sort of bonding mechanism to induce self-enforcing reliance relation- ships between buyers and sellers. Warranties, inde- pendent appraisals, leases with a high residual, collateral, irrevocable money-back guarantees, con- tingent payments, and brand names all provide assur- ance to buyers that the seller will notmisrepresent the product quality. Such hostage mechanisms support asymmetric information exchange.

� Joint ventures and partnerships face an asymmetric information problem in reimbursing each member for privately known costs that are unverifiable. In both demand revelation problems for funding pub- lic goods and cost revelation problems for partner- ships, each member has an initial incentive to falsely reveal (overstate) his or her private (cost) information.

� Both understatement of demand and overstate- ment of cost result in the cancellation of profitable partnership projects. Yet, each individual member may be better off with exaggerated cost reimburse- ment than with a simple profit share. Preserving the maximum value of the partnership requires an incentive-compatible revelation mechanism.

� Under an incentive-compatible (IC) mechanism, partners making cost revelations incur the expected costs imposed on and opportunities forgone by the other partners. Each partner agrees that not just ac- tions, but information revelations themselves, have consequences for profit-share payout. Such a gover- nance mechanism must be self-enforcing, however, because the asymmetric information problem has not disappeared. A court would have just as much troubleverifying theclaims for reimbursementunder this incentives contract as it would under the initial simple profit-sharing contract.

� Incentive-compatible revelation mechanisms do motivate partners to reveal their true projections of costs.

� IC revelation mechanisms are implemented through contingent claims contracts and often re- quire ex ante posting of a bond to solve the breaking-of-the-budget problem. Quasi-linear preferences are then required to assure a unique and efficient Clarke-Groves revelation mechanism.

606 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Exercises 1. What auction design features reduce the winner’s curse and therefore reduce stra- tegic underbidding?

2. Why don’t airlines and hotel chains worry about self-destructive cannibalization of their own higher-priced live sales when they list seats and rooms for sale in the virtual marketplace on Priceline.com?

3. What advantage does eBay’s open bidding provide to sellers? Why? 4. Which two of the following are most clearly common-value auction items: Viper

sports cars, electricity, patent licenses, T-bills, antiques, or fine art? 5. If some auction participants for crude oil field leases have estimates that the oil in

the ground is worth $1.2 million, $1.3 million, or $1.5 million with certainty; and other auction participants have estimates that the same oil field lease is worth $1.1 million, $1.3 million, or $1.5 million with certainty; and a third group of auction participants have estimates that the same oil field lease is worth $1.1 million, $1.2 million, or $1.3 million, and all three forecasts contain the true common value, what is that value? How would you as auctioneer-seller design an auction to re- duce strategic underbidding and realize this true value?

6. Distinguish common-value and private-value auctions; provide examples of each. Distinguish descending-price (Dutch) auctions and ascending-price (English) auc- tions; provide examples of each.

7. You are developing a bidding strategy for an ascending-price sealed-bid auction of a crude oil field worth between $1 million and $51 million to the seller. Because your extraction costs are lower, your value is 20 percent greater than the seller’s value. The seller faces transaction costs of conducting the sale and therefore will not ac- cept an offer unless it exceeds her personal value. How much should you bid?

8. How can a second-highest sealed-bid ascending-price auction design diminish the “winner’s curse” and reduce the strategic underbidding that arises in highest- wins-and-pays typical ascending-price auctions with sealed bids?

9. Some newly issued T-bills are auctioned by discriminatory pricing in Dutch auc- tions, whereas other newly issued T-bills are auctioned by uniform prices second- highest sealed-bid ascending-price auctions. Which auction design is more like private placement of corporate newly issued bonds and IPO stock? Which auction design is more likely to increase seller revenue?

10. Fast Second and Speedo are trying to decide what to bid for a license in a cellular phone auction where the possible values of the new license are $10 million, $20 million, $30 million, $40 million, $50 million, and $60 million, each equally likely. The auction is single-round sequential, both parties have exactly the same value for the asset but neither knows its true value from the possible distribution (i.e., a so-called common-value auction), and Fast Second gets to bid after Speedo.

Each company has invested in marketing research about the value of the li- cense, which can come out one of two ways: possible values of $20, $30, or $50 million, or possible values of $20, $40, or $60 million. Whichever result arises is known to be 100 percent accurate (i.e., the license is worth one of the three iden- tified amounts with certainty). Speedo proceeds with a bid of $33 million. Fast Second has marketing research saying that the value of the license is $20, $40, or $60. How much should Fast Second bid?

Set up the Bayesian probability rule for Fast Second of BAYES PROB ($40 mil- lion value/Forecast of $33 million bid by Speedo).

Answers to the exercises in blue can be found in Appendix D at the back

of the book.

Appendix 15A: Auction Design and Information Economics 607

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

11. Show that not just overstatement but also understatement of cost is dominated by truth-telling in the joint venture of Motorola and Apple.

12. What payoffs would be required under an optimal incentives contract, similar to Table 15A.7, if the cost overruns at Apple became as likely as those at Motorola?

13. In Appendix 15A, we have assumed bidder’s valuations are independent, but sup- pose they are affiliated. That is, suppose on eBay you wish to affiliate with those who think Beatles albums are valuable, and that affects your own personal valua- tion. How will this change from independent to affiliated valuations affect bidding strategy on eBay? Do you observe such behavior on the site?

Case Exercises SPECTRUM AUCTION

Continuing the analysis of broadband spectrum auctions from the appendix, suppose that two bidders know that the net present value of the rights to transmit PCS services in Louisville is a random variable uniformly distributed from $10 million to $60 mil- lion with six discrete values possible: $10 million, $20 million, $30 million, $40 mil- lion, $50 million, and $60 million. Also assume that both parties value the asset identically, making it a common-value auction. In advance, each company conducts marketing research experiments to narrow the possible outcomes and thereby better inform its own bid. Suppose Wireless Co.’s marketing research results exclude the two tails of the uniform distribution of possible values (i.e., $10 million and $60 million) as well as $40 million. Similarly, PCS PrimeCo conducts its own marketing research that excludes $10 million, $30 million, and $50 million as possible outcomes for the Louisville service area.

Questions 1. What should Wireless Co. bid in a single-round sealed-bid common-value auc-

tion? What should PCS PrimeCo bid in this same auction? 2. If Wireless goes first in a sequential posted-price auction with multiple rounds to

follow, what should PCS PrimeCo respond in Round 2? In Round 3, will Wireless then wish to amend its earlier bid? Why or why not?

3. What auction design would be in the seller’s best interest: single-round sealed-bid or multiple-round open bidding?

4. Identify other factors that could affect the optimal auction design.

DEBUGGING COMPUTER SOFTWARE: INTEL31 Debugging has been a way of life in the computer industry from its inception. Indeed, the origin of the term debugging derives from the daily process of removing dead moths from the thousands of electronic tubes in the ENIAC, the first electronic com- puter. Every piece of computer hardware or software ever shipped has likely had logic

31Based on “It’s Not a Bug, It’s a Feature,” Forbes (February 13, 1995), p. 192.

608 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

faults. Indeed, most popular software programs contained thousands of known “bugs” in their first-generation products. In 1994, incomplete debugging of the floating point division calculator in the Pentium I computer chip caused a massive product recall that cost Intel $475 million dollars.

Why do computer component manufacturers release products with known bugs? One obvious answer is that delayed release may allow competitors to preempt the market with new technologies that render your product obsolete. Another important answer is a central insight of managerial economics that everything worth doing is not necessarily worth doing well. Computer design and manufacturing firms face a rising marginal cost of correcting thousands of bugs detected by their beta testing pro- cess. At some point, each firm must balance the lost sales and replacement costs from product recalls against the ever-increasing cost of design perfection from continuously more debugging.

A somewhat surprising third answer may, however, hold the key: fixing bugs in subsequent generations of software sells upgrades. Early Microsoft Windows products had a nasty bug that caused the program to crash with the finality of a hopeless error message—“unrecoverable application error.” Microsoft fixed the bug and proceeded to sell millions of copies of the upgrade. Bugs in programs limit their durability, and in technology businesses the selling of upgrades is a part of the business plan. Hal Var- ian, chief economist of Google, calls this practice “versioning.”

Questions 1. Discuss the practice of selling upgrades as a mechanism design. 2. What dual objectives are being served when a breakthrough new product is not

equipped with all its known value added features? 3. How does the versioning of products address the problems of a durable goods

monopolist who must compete with his or her own discounted used products that are in good working order?

Appendix 15A: Auction Design and Information Economics 609

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

16 CHAP T E R

Government Regulation CHAPTER PREVIEW Managerial decisions designed to maximize shareholder wealth face many constraints. Some of these constraints arise from a business’s moral social responsibilities. Other constraints take the form of laws or other legal obligations. A wide array of government regulations are constraining (e.g., prohibiting price fixing collusion) but in other cases enabling (e.g., protecting trade secrets). To make value-maximizing decisions, managers must fully understand the regulatory aspects of their environment. This chapter explores several types of regulatory issues: antitrust, business permits, licensing and patents, and the cap-and-trade market-based approach to environmental regulation.

MANAGERIAL CHALLENGE Cap and Trade, Deregulation, and the Coase Theorem

Professor Ronald Coase from the University of Chicago Law School received the Nobel Prize in Economics for his work on the relationship among property rights, transac- tion costs, and the role of government. Coase challenged the prevailing view that economic externalities, such as water, air, and noise pollution, could only be resolved through governmental action. Coase argued that exter- nalities should often not be viewed as one party inflicting harm on another but rather as reciprocal imposition of side effects. For example, a steel fabrication plant might use the surrounding factory buildings to absorb noise from its production process. The owners of nearby old factories might tear them down to clear land for an am- phitheatre where less noise could attract more patrons.

The Coase theorem claimed that such reciprocal ex- ternalities could be resolved without government inter- vention if the transaction costs of arriving at a private voluntary bargaining solution were kept low. Coase ar- gued the issue was one of the appropriate specification and assignment of property rights. For example, planes landing at Logan Airport in Boston might need to avoid violating a sensible decibel level entitlement assigned to property owners in Winthrop and Revere surrounding the airport. Otherwise, damage claims would arise, and

monetary compensation or other noise-proofing settle- ments would be awarded.

In air pollution control, this Coasian approach to allocating “rights to pollute” has now been adopted on a worldwide basis. Under the conditions of the U.S. Clean Air Act, the Environmental Protection Agency (EPA) grants allowances for about 50 percent of the sulfur dioxide emissions from electric utility plants. Congress then gave polluters the right to trade these pollution allowances among themselves. For example, if one firm already has emission levels at its plants that are within EPA bounds, it can sell its excess allow- ances. Other firms that do not meet the emissions stan- dards can choose either to buy the pollution rights at a market price or to install the needed pollution control equipment—whichever is cheaper. The Chicago Board of Trade quickly created a market on which these pollu- tion allowances are actively traded. Take a look at the auction prices for sulfur dioxide allowances (a compo- nent of acid rain) at http://www.epa.gov/airmarkets.

Interest in this so-called cap-and-trade approach, al- lowing market forces to operate in a constrained envi- ronment rather than relying on command-and-control regulation of pollution, continues to grow. After the

610

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

THE REGULATION OF MARKET STRUCTURE AND CONDUCT Antitrust regulation is designed to increase competition by eliminating attempts to monop- olize an industry, as well as by attacking certain patterns of per se illegal conduct, such as price fixing and exclusionary contracts, that are always believed to harm competition.

Market Performance Ultimately, what society would like from the producers of goods and services is a multi- dimensional performance concept that includes these elements:

1. Resources should be allocated in an efficient manner, sometimes labeled static efficiency.

successful U.S. development of emissions trading for sulfur dioxide in the 1990s, the European Union in 2005 introduced a carbon dioxide emissions trading sys- tem. Short selling and illiquid, thin trading initially caused a ton of emissions from burning coal to rise in cost almost as high as the cost of the coal itself. Since most baseline electricity comes from coal-fired plants, electricity prices rose steeply and then seesawed until regulatory institutions stabilized. One desired

consequence was that natural gas, which costs only slightly more than coal per 1,000 BTU but pollutes the environment with only half as much CO2 emis- sions, became the fuel preferred by electricity genera- tors. If adopted in the United States, the additional costs for CO2 emissions were projected by the U.S. Energy Information Administration to lower industrial shipments from 1 to 3 percent and GDP from 0.5 to 1 percent over the period 2012–2030.1

Deregulation is another trend that has characterized the 1990s and 2000s. Deregulation of most aspects of the transportation industries is complete; natural gas pipelines and telephone companies have been greatly deregulated; and the electric utility industry is moving toward deregulation. Greater deregulation will open new opportunities for future managers and confront them with new challenges. For example, the deregula- tion of mortgage-backed securities led to an extraordi- nary financial crisis in the United States 2007–2009, followed by a massive loss of real estate and stock mar- ket wealth, and a resulting severe recession.

Discussion Questions

� Contrast the United States’ and the European Union’s experience with cap and trade.

� Why do you think the cap-and-trade bill for carbon dioxide has stalled in the U.S. House of Representatives?

� In order to reduce greenhouse gases like CO2, would you support paying 10 percent more for electricity? Why or why not?

1“Cumulative Impact of House Cap-and-Trade Bill,” The Economist, August 15, 2009, p. 24.

MANAGERIAL CHALLENGE Continued ©

Ph ot oL in k/ Ph ot od is c Gr ee n/ Ge tty

Im ag es

Chapter 16: Government Regulation 611

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

2. Producers should be technologically progressive; that is, they should attempt to develop and quickly adopt new technologies that will result in lower costs, improved quality, or a greater diversity of better products.

3. Producers should operate in a manner that encourages full employment of produc- tive resources, including human capital.

Market Conduct A structure-conduct-performance model of the factors that influence market perfor- mance is illustrated in Figure 16.1. Market performance is dependent on the conduct of firms in their

1. Pricing behavior 2. Product policy 3. Sales promotion and advertising policy 4. Research, development, and innovation strategies

Market Structure Bothmarket performance andmarket conduct dependon the structure of the particularmar- ket. The concept of market structure refers to threemain characteristics:

1. The degree of seller and buyer concentration in the market, as well as the size distri- bution of these sellers or buyers: On the seller side, concentration determines whether an industry is classified as monopoly, oligopoly, pure competition, or some

FIGURE 16.1 A Conceptual Market Structure-Conduct-Performance Model

Fundamental market and environmental conditions

1. Location and ownership of raw material 2. Product durability 3. Technology 4. Labor organization 5. Regulation

1. Price elasticity 2. Cross elasticity 3. Growth prospects 4. Type of product 5. Method of purchase

Supply Demand

1. Seller and buyer concentration 2. Actual or perceived product differentiation 3. Conditions of entry 4. Vertical integration 5. Diversification or conglomeration 6. Contestability

Market structure

1. Pricing behavior 2. Product policy 3. Sales promotion and advertising strategies 4. Research, development, and innovation strategies

Market conduct

1. Efficient allocation of resources 2. Technologically progressive 3. Full employment 4. Equitable income distribution 5. Resource conservation 6. Satisfactory product performance and safety characteristics

Market performance

612 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

variant thereof. Buyer concentration is also important because the bargaining power of buyers affects the gross margin sellers can earn.

2. The degree of objective or perceived differentiation between the products or services of competing producers:When buyers perceive the product of one firm to be different from that of another, these buyer preferences will impart a degree ofmarket power to the seller that ultimately affects that seller’s market conduct and performance.

3. The conditions surrounding entry into and exit from the market: When significant barriers to entry exist, competition may cease to become a disciplining force on existing firms, making performance less than the competitive ideal. Exit barriers diminish the competitive discipline imposed by potential (as opposed to actual) competitors. Entry barriers allow established sellers to raise prices above the mini- mum average cost of production and distribution without motivating new sellers to enter the industry. Barriers to entry may be classified into four types: product differentiation, absolute cost advantage, economies of scale, and limited access to distribution. These general types of entry barriers, how they arise, and their consequences are summarized in Table 16.1.

TABLE 16.1 TYPES AND CONSEQUENCES OF BARRIERS TO ENTRY

TYPES CONSEQUENCES FOR NEW ENTRANTS

A. Product differentiation barriers arise from 1. Buyer preferences, conditioned by advertis-

ing, for established brand names 2. Patent control of superior product designs

by existing firms 3. Ownership or control of favored distribution

systems (e.g., exclusive auto dealerships)

B. Absolute cost advantages of established firm’s production and distribution arise from 1. Control of superior production techniques

by patent or secrecy 2. Exclusive ownership of superior natural

resource deposits 3. Inability of new firms to acquire necessary

factors of production (management, labor, equipment)

4. Superior access to financial resources at lower costs

C. Economies of large-scale production and distribution (or sales promotion) arise from 1. Capital-intensive nature of industry

production processes 2. High initial start-up costs

D. Limited access to distribution channel

A. 1. New entrants cannot sell their products for as high a price as existing firms can

2. Sales promotion costs for new entrants may be prohibitive

3. New entrants may be unable to raise suffi- cient capital to establish a competitive dis- tribution system

B. 1. Costs of new entrants are higher than for existing firms, so even though existing firms may charge a price that results in above- normal profits, new entrants may be unable to make even a normal profit at that price

C. 1. The entry of a new firm at a sufficient scale will result in an industry price reduction and a disappearance of the profits anticipated by the new entrants

2. New firms may be unable to acquire a sufficient market share to sustain efficient operations

D. Closed shelf-space or Internet portals will necessitate massive slot-in investments and may prohibit certain business models

Source: Joseph Bain, Industrial Organization, 2nd ed. New York: John Wiley (1968), pp. 237–265.

Chapter 16: Government Regulation 613

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Contestable Markets2

A perfectly contestable market is easily accessible to potential entrants and easy to exit because capital investments are redeployable (trucks, patented inventions, information). The potential competitors use the incumbent firms’ pre-entry price to evaluate whether entry would be profitable. With freedom of entry and exit and fully redeployable assets, potential competitors are not worried about incumbents’ pricing reactions. If profit po- tential disappears after initial entry, new entrants can simply leave the industry. The pos- sibility of hit-and-run profits by potential entrants will cause even a dominant incumbent firm to set prices equal to average cost, because any higher price leaves an opportunity for profitable entry. Contestable markets yield perfectly competitive market performance despite a market structure comprising just two of three firms.

ANTITRUST REGULATION STATUTES AND THEIR ENFORCEMENT Since 1890, a number of federal antitrust laws have been passed with the intent of pre- venting monopoly and of maintaining competition in U.S. industry. These laws were ini- tially directed at the large stockholder trusts such as Standard Oil, American Tobacco, and several coal and railroad trusts. Under a trust agreement, the voting rights to the stock of a number of directly competitive firms were conveyed to a legal trust. The trust then managed the firms collectively, thereby maximizing profits, but high prices and re- stricted outputs resulted. In this section, we summarize the most important of these an- titrust laws and their effects on business decisions.

The Sherman Act (1890) The Sherman Act’s important provisions are brief, but they are wide ranging. It declares illegal “every contract, combination in the form of a trust or otherwise, or conspiracy in restraint of commerce among the several States, or with foreign nations . . . [and] every person who shall monopolize, or attempt to monopolize, or combine or conspire with any other person or persons, to monopolize any part of the trade or commerce among the several States, or with foreign nations, shall be deemed guilty. . . .”

Example Why City-Pair Airlines Are Not Contestable Markets Aircraft, of course, seem to be the classic redeployable asset. Put an aircraft on the resale market, and within several weeks, one should be able to realize close to the replacement value of the asset. However, several features of the airline business do not meet the conditions of contestable markets. First, hub airport investments are sunk costs often not redeployable into other airline route structures. Second, the costs of switching from one airline to another are often raised by frequent flyer programs, flight schedules, and ticket promotions that restrict interline transfers. Finally, airline incumbents change prices two or three times a day, adjusting to competitive threats much more quickly than hit-and-run entrants can move in and out of city-pair markets. So, think of trucking (not airlines) as an example of a contestable market.

2Based on William J. Baumol, J.C. Panzar, and R.D. Willig, Contestable Markets and the Theory of Industry Structure (New York: Harcourt Brace Jovanovich, 1982).

antitrust laws A series of laws passed since 1890 to limit monopoly power and to maintain competition in most American industries.

614 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Clayton Act (1914) The Clayton Act prohibited four anticompetitive business practices:

1. Price discrimination at wholesale was deemed illegal, except to the extent that it was based on differences in grade, quality, and quantity of the product sold.

2. Section 3 prohibited sellers from leasing or making “a sale or contract for the sale . . . on the condition that the lessee or purchaser thereof shall not use or deal in the . . . commodity . . . of a competitor.” This prohibition against “exclusive dealing and tying contracts”was not absolute. It applies to the extent that the practice foreclosed customers from purchase agreements they desired tomake.

3. Section 7, the antimerger section, barred any corporation engaged in commerce from acquiring the shares [or assets] of a competing firm or from purchasing the stocks of two or more competing firms, where substantial damage to competition could be proven.

4. Interlocking directorates, defined as a case where the same person is on the board of directors of two or more firms, were declared illegal if they served to eliminate competition among parties who would otherwise compete.

The Federal Trade Commission Act (1914) The Federal Trade Commission (FTC) Act was passed as a supplement to the Clayton Act. Its major antitrust provision, found in Section 5, merely states “that unfair methods of competition in commerce are hereby declared illegal.” The Clayton Act established the Federal Trade Commission as an independent government antitrust agency to identify and prosecute anticompetitive trade practices.

The Robinson-Patman Act (1936) The Robinson-Patman Act is summarized here:

1. Section 2(a) makes it illegal to discriminate in prices at wholesale when selling goods of “like grade and quality” where the effect may be to “substantially lessen competition.” A seller who is charged with price discrimination has two affirmative defenses

Example Why Miller Beer Is So Hard to Find in Mexico3

Mexico is the world’s eighth largest beer market. Since the 1994 NAFTA agreement, Corona and Modelo beer exports grew fivefold to account for 11 percent of the U.S. market. Anheuser Busch owns a 50 percent non-controlling stake of the Modelo brewer. The Miller Brewing Co. attempted the same sort of penetration into the Mex- ican domestic market without any real success. Modelo and its rival FEMSA have 99 percent of the market. For one thing, FEMSA owns large convenience store chan- nels in Mexico, and Miller is not stocked on their shelves. In other cases, bars are paid to deal exclusively with Modelo. As a consequence, a six-pack of Modelo Especial costs $4.60 in Mexico versus $1.80 for the best seller in Brazil and $2.20 for the best seller in Chile. Such exclusive dealing contracts that foreclose the Modelo customers from doing business with a competitor would be prohibited in the United States.

3Based on “Why Corona Is Big Here,” Wall Street Journal (January 17, 2003), p. B1.

Chapter 16: Government Regulation 615

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

enumerated in Section 2(b): First, the “cost defense” permits differentials in price that “make only due allowance for differences in the cost of manufacture, sale, or delivery.” Second, the Act permits a lower price to be charged in one segment of themarket tomeet “an equally low price of a competitor.”

2. Sections 2(d) and 2(e) prohibit the seller from allowing discounts to a buyer for merchandising services rendered the seller by the buyer, unless similar discounts are offered to all buyers. Secret rebates are prohibited. For example, Liz Claiborne can- not rebate 15 percent to the Gap without offering the same deal to department stores like Macy’s. Also, advertising or promotional allowances must be made avail- able to all buyers, not just a few selected large firms.

The Hart-Scott-Rodino Antitrust Improvement Act (1976) The Hart-Scott-Rodino Act requires companies with assets over $100 million to provide notification and information concerning any proposed merger to the Antitrust Division of the Department of Justice and to the Federal Trade Commission. Over a 30-day wait- ing period, the DOJ and FTC examine the competitive effects of the merger proposal. They then either challenge the proposed merger in federal court or allow the merger to be completed, possibly with some modifications. Companies can appeal rulings by the FTC, and private complainants can bring antitrust suits to the federal courts. State attorneys general can also initiate federal antitrust suits.

Government agencies can use various methods to enforce the antitrust laws. Most an- titrust cases are settled with consent decrees negotiated between the company and en- forcement officials. Under a consent decree, a company agrees to take certain actions (or cease and desist from other actions) in return for the government agreeing not to seek additional penalties in the courts. In cases filed by antitrust agencies against a com- pany, the courts may issue an injunction requiring (or prohibiting) certain actions by the company. The courts may also impose fines and, in certain instances, prison sentences. In cases involving charges of monopolization, the courts may require divestiture of certain assets by the company. For example, the antitrust commission of the European Union insisted that British Airways (BA) divest itself of 353 landing slots at London’s Heathrow Airport if BA and American Airlines wished to merge. Rather than lose that many of its prized assets, BA decided to continue competing with American.

Example California’s Class Action Suit against Microsoft Settled for $1.1 Billion In January 2003, a U.S. district court in Washington found Microsoft guilty of maintaining a 92 percent monopoly of desktop operating system software by using anticompetitive practices. Two and a half years later in 2005, the defendant settled a class action lawsuit filed by the California Attorney General on behalf of 13 mil- lion California individuals and businesses. Microsoft agreed to pay $5 to $29 vou- chers for each California buyer who licensed either Windows 95 or Windows 98 between 1995 and 2001 as compensation for the alleged overcharges. The vouchers could be used for laptop, desktop, or tablet computers and for software from any computer company. If all 51 million vouchers are submitted, Microsoft stands to lose $1.1 billion. Although this figure seems enormous, Microsoft in 2005 had $61 billion in cash and short-term investments.

616 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

ANTITRUST PROHIBITION OF SELECTED BUSINESS DECISIONS Collusion: Price Fixing Explicit agreements among competitors to fix prices along with other overt forms of col- lusion, such as market-sharing agreements, are per se illegal under the Sherman Act. That is, the courts generally declared such agreements illegal, regardless of whether or not they cause obvious injury to competition. During 2010, EMI Group, Sony BMG Mu- sic Corp., Bertelsmann Inc., Vivendi SA’s Universal Music Group, and Warner Music Group Corp. were indicted for creating joint ventures (with 85 percent of the market) to conspire to inflate and maintain the price of digital music. The defendants argued that coordination was needed to deal with the double trouble presented by pirated music downloading (Napster and its successors) and Apple’s 72 percent share of the online song market with the wildly popular iTunes. In a few cases, such as the sugar and ocean shipping industries, firms have been legislatively exempted from the antitrust laws and are legally permitted to jointly set prices and allocate output.

Mergers That Substantially Lessen Competition When industry sales, assets, or contributions to value added are concentrated in a few hands, market conduct and performance are less likely to be competitive in nature. One widely used index of market concentration is the market concentration ratio. It may be defined as the percentage of total industry output (measured in terms of sales, employment, value added, or value of shipments) attributable to the 4, 8, 20, or 50 largest companies.

Data on market concentration ratios are regularly made available from the Census Bureau, based on the Census of Manufacturers. The Census Bureau defines industries in terms of SIC product categories. The SIC system consists of up to a seven-digit category code, indicating increasing specificity of industry and product as the number of digits increases. All manufacturing, for example, is specified by the first digit, food and kindred products by a two-digit category, candy and other confectionary products by a four-digit category, and sugar- or chocolate-coated nuts by a five-digit category.

Table 16.2 provides concentration ratios for selected industries. Some industries have become highly concentrated, such as breakfast cereals, turbine generators, aluminum, and lightbulbs. Some industries, such as hosiery, concrete block and brick, as well as sporting goods, are highly fragmented at the national level. One hypothesis of the structure-conduct-performance paradigm is that performance will be more competitive in the lower concentration markets. This hypothesis does not always hold; occasionally very concentrated industries such as airlines at fortress hub airports are highly competi- tive because of the contestability of those markets.

Consolidation mergers to reduce excess capacity in such an industry occur regularly. However, efficiency gains from further consolidation in already highly concentrated in- dustries such as microprocessors are typically tiny. Because of sizable premiums paid to target-firm shareholders, the returns to acquiring-firm shareholders are consistently neg- ative. For these reasons, mergers between firms such as AMD and Intel are seldom ap- proved by the antitrust authorities.

Another important measure of market concentration is the Herfindahl-Hirschman Index, or HHI:

HHI = ∑ N

i=1 S2i

where Si is the market share of the ith firm and N is the number of firms in the industry. For example, in a relevant market consisting of just three firms, such as baby food

market concentration ratio The percentage of total industry output produced by the 4, 8, 20, or 50 largest firms.

Herfindahl-Hirschman Index A measure of market concentration equal to the sum of the squares of the market shares of the firms in a given industry.

Chapter 16: Government Regulation 617

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

(where Gerber has 70 percent, Beech-Nut has 16 percent, and Heinz has 14 percent mar- ket share), the HHI is 702 + 162 + 142, which is equal to 5,352. HHI has a maximum value of 10,000 and decreases as the number of firms (N) increases. The HHI measures market concentration but accentuates the potential influence of leading firms with asymmetrically large market shares. HHI values for selected industries are also shown in Table 16.2. Not shown in the table is the most extreme case of all. Microsoft’s 92 percent market share makes the HHI in software operating systems 8,526.

TABLE 16.2 CONCENTRATION RATIOS AND HERFINDAHL-HIRSCHMAN

INDEX FOR SELECTED INDUSTRIES

SHARE OF VALUE ADDED ACCOUNTED FOR BY THE 4, 8, AND 20 LARGEST

COMPANIES IN EACH MANUFACTURING INDUSTRY

SIC INDUSTRY NAME 4-FIRM RATIO

8-FIRM RATIO

20-FIRM RATIO

HERFINDAHL- HIRSCHMAN

INDEX

31123 Breakfast cereals 82 93 100 3,000

311511 Fluid milk 46 57 71 1012

31511 Hosiery and socks 35 45 64 318

32561 Soap and detergents 49 62 72 949

32411 Petroleum refining 47 67 92 809

32721 Flat glass 76 98 100 1,677

327331 Concrete block and brick 24 32 44 206

331315 Aluminum sheet, plate, and foil 75 89 98 2,286

333611 Turbine and turbine generator sets 88 91 96 2,403

33511 Electric lamp bulbs 90 94 98 2,848

32992 Sporting and athletic goods 24 32 46 199

33991 Jewelry and silverware 18 26 39 142

Source: Census of Manufacturers, U.S. Department of Commerce.

Example INTEL and AMD Fight It Out in Microprocessors4

Intel has acquired most of its rivals, but two remaining firms compete vigorously. Intel and AMD, who together control all of the chip production for microproces- sors in desktop PCs, compete vigorously and their resulting margins are very slim. In general, Intel technology has an advantage in laptop computers and servers, but starting in 2004 Intel lost its dominant role in lower-end desktop machines (see Figure 16.2). More recent dual core chips have reestablished Intel’s product advan- tage in Internet-enabled handheld computers.

4Based on “Sued for Stifling Competition,” Wall Street Journal (December 18, 2009), p. 25; and Robert Bruner, “Does M & A Pay?” Journal of Applied Finance (Spring/Summer 2002), pp. 45–68.

618 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Merger Guidelines (1992 and 1997) The FTC and the Antitrust Division of the Department of Justice (DOJ) in 1992 issued merger guidelines based on the Herfindahl-Hirschman Index (HHI) that they use in de- ciding whether to challenge a proposed merger:

1. For markets with an HHI greater than 1,800, the government is likely to challenge a merger that increases the index by 50 to 100 points, or more.

2. For markets with an HHI between 1,000 and 1,800, a merger challenge by the gov- ernment is unlikely unless the index increases by 100 or more points.

3. Formarkets with anHHI less than 1,000, the government is unlikely to challenge amerger.

A merger increases the HHI by two times the product of the market shares of the candi- date firms. So, when Beech-Nut and Heinz’s baby food division wished to merge, the merger was challenged because the 5,352 point HHI changed as follows:

HHI before = S2Gerber + S 2 Beech-Nut + S

2 Heinz = 5,352

HHI after = S2Gerber + ðSBeech-Nut + SHeinzÞ2 = S2Gerber + ðS2Beech-Nut + S2Heinz + 2SBeech-NutSHeinzÞ = 702 + ð162 + 142 + 2 · 16 · 14Þ = 5,352 + ΔHHI

= 5,352 + 448 = 5,800

The merger guidelines also list other factors that are considered in the analysis, including the ease with which competitors can enter the industry, likely failure of the to-be-acquired firm without the merger, and possible gains in efficiency for the (combined) firm.

Generally, a narrower definition of the market will heighten the measure of potential monopoly power and raise the probability of a merger, substantially lessening competi- tion. If the FTC or Antitrust Division comes to this conclusion, they will seek an injunc- tion in federal court to stop the proposed merger.

Monopolization As we saw earlier, firms engaged in overt forms of collusion with other companies can be successfully prosecuted under the Sherman Act. Companies acting alone also can be

FIGURE 16.2 Market Share Dynamics in Microprocessors for Desktop PCs

60%

70%

50%

40%

30%

20%

10% 1998 1999 2000 2001 2002 2003 2004 2005 2006

Intel

AMD

Source: ZD InfoBeads.

Chapter 16: Government Regulation 619

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

charged under the act with illegally attempting to monopolize a market or engaging in monopolistic practices. However, proving such alleged violations of the laws often is quite difficult. Before the Microsoft case in 1998, the last large monopolization case brought by the U.S. antitrust officials resulted in the breakup of AT&T in 1984.

Wholesale Price Discrimination A large company that operates as a manufacturer or distributor in two (or more) differ- ent geographic (or product) markets and cuts wholesale prices in one market and not in

Example Trustbusters Reappear: DISH Network-DIRECTV Merger Disapproved5

In 2007, the FTC prohibited a merger between Whole Foods Inc. and Wild Oats Markets in the premium natural and organic supermarkets, where these are the leading competitors. Similarly, European antitrust authorities blocked the $41 bil- lion merger of General Electric and Honeywell in 2005 on market concentration grounds. The two companies dominate the regional and corporate jet engine mar- kets. Higher concentration in a two- or three-firm industry appears prohibited.

In 2000, the FTC blocked a proposed merger between Heinz and Beech-Nut with 16 percent and 14 percent, respectively, of the baby food market. The ratio- nale was that the Herfindahl-Hirschman index (HHI) for baby foods measured 5,352, relative to the 1,800-point presumptive benchmark that causes concern. Be- hemoth Gerber has 70 percent of the market and is itself responsible for 4,900 of the 5,352 HHI points. The FTC did not accept the Heinz and Beech-Nut argument that together their two firms could realize scale economies, lower distribution and R&D costs, and thereby compete more effectively against the dominant firm.

Finally, Charlie Ergen’s privately held EchoStar (DISH Network) launched a $26 billion bid for Hughes Electronics’ subsidiary DIRECTV. DISH Network and DIRECTV were the top two satellite TV providers in the United States. Ergen ar- gued that the relevant market included cable TV systems, with a combined indus- try share distribution as follows:

Comcast 33%

DIRECTV 17%

Time Warner 17%

DISH Network 13%

Charter Comm. 10%

Cox Comm. 10%

HHI 2,036

The FTC and the Antitrust Division disagreed, defined the relevant market nar- rowly as satellite TV, and prohibited the merger. Even under the broader definition, HHI was already in excess of 1,800, the cutoff under the 1997 Merger Guidelines. Ergen’s case, for an exception, rested on potential efficiencies from the removal of duplicate satellite transmissions, freeing bandwidth for additional channels.

5Based on “Competition Policy,” The Economist (May 3, 2007), p. 79; “Is the FTC Defending Goliath?” BusinessWeek (December 18, 2000), pp. 160–162; and “Murdoch Wins DIRECTV,” Wall Street Journal (April 10, 2003), p. B1.

620 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Example Potentially Anticompetitive Practices: Microsoft’s Tying Arrangements6

Netscape alleged that dominant software maker Microsoft used its leading position in the market for computer operating systems to gain an anticompetitive advan- tage in the market for applications software such as Microsoft’s Internet Explorer and Media Player. Netscape complained that Microsoft illegally tied its Internet access software (Microsoft Explorer) to sales of Windows 95, which provided the operating system for 92 percent of the personal computers in the United States. Microsoft distributed Explorer free with every sale of Windows 95 to Compaq and Dell computers, priced Windows 95 without Explorer much higher, and threatened to remove the Windows 95 license if any Web browser other than Microsoft Explorer was preinstalled on the PCs Compaq shipped. Over four quar- ters in late 1996 and 1997, Microsoft’s share of the Web browser market grew from 20 percent to 39 percent. By 1999, Netscape’s share had fallen from a high of 84 percent to 47 percent, and Microsoft’s product accompanied by the allegedly anticompetitive practices had resulted in a 53 percent share. Today Netscape has about 8 percent of a market that Internet Explorer dominates. Was this evidence of “substantial harm to competition,” or just substantial harm to a particular competitor?

Tying arrangements that extend the monopoly power of a dominant firm in one market to another distinct product and relevant market are illegal per se. Because Microsoft’s sales practices precluded Netscape from selling its Web browser, Microsoft was required under a consent decree to unbundle the two products and change its pricing practices. Nevertheless, in May 1998, the Justice Department and 20 state attorneys general filed suit, alleging illegal tying arrangements and other anticompetitive practices. Microsoft was found guilty of the alleged violations.

In 2004, the European Court of Justice ruled against Microsoft in a complaint first filed by Sun Microsystems that Microsoft illegally bundled their Media Player product with Windows. Windows without Media Player was priced much higher. Because Microsoft has a dominant firm monopoly in PC operating systems, with market share that grew between 1997 and 2005 from 86 percent to 93 percent, the European Court of Justice ruled that Microsoft must unbundle its Media Player and offer the diminished capability version of Windows more cheaply than the ex- panded capability version. The Court levied a €497 million fine. In 2006, another €281 million fine was imposed for noncompliance. Finally, in 2009 Microsoft lost their final appeal and paid (with interest) €1.4 billion. In another antitrust case in 2007, the European authorities fined elevator and escalator manufacturers €992 for price fixing. Clearly, losses of this magnitude suggest that assuring compliance with antitrust policy is an appropriate priority for shareholder wealth-maximizing managers.

6Based on “Browse This,” U.S. News & World Report (December 5, 1997), p. 59; “U.S. Sues Microsoft over PC Browser,” Wall Street Journal (October 21, 1997), p. A3; “Knowing the ABCs of the Antitrust Case against Microsoft,” Wall Street Journal (October 30, 1997), p. B1; “Microsoft’s Browser: A Bundle of Trouble,” The Economist (October 25, 1997), p. 74; “Microsoft on Trial,” Wall Street Journal (April 4, 2000), p. A16; “Microsoft Is Dealt Blow by EU Judge,” Wall Street Journal (December 23, 2004), p. A3; and “Europe’s Antitrust Chief Defies Critics,” Wall Street Journal (February 25, 2008), p. A1.

Chapter 16: Government Regulation 621

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the other market can be accused under the Robinson-Patman Act of engaging in illegal price discrimination. Differential pricing directly to final product customers is allowed (and often based on “what the market will bear”) but not so in pricing to intermediate product resellers (wholesalers, distributors, etc.). For example, the publisher Penguin Books paid a large judgment to independent booksellers after it was proved that Penguin offered volume discounts and other trade promotions to Barnes & Noble and Borders that were unrelated to the cost of serving those accounts. Similarly, six vending machine companies sued Philip Morris, alleging that other distributors and retail merchants received rebates, buybacks, and promotional allowances intended to lower costs and allow the favored distributors to drive the complainants out of business. The Robinson-Patman Act was designed to prohibit pre- cisely this kind of favoritism in wholesale trade.

Refusals to Deal In general, a manufacturer can refuse to deal with any retail distributor who fails to follow company policies that are based on legitimate business justifications. However, this authority is subject to three limitations. First, the orders of a renegade discounter can be refused if and only if the manufacturer acts independently of compliant dealers whose sales at higher price points are suffering because of the increased competition (United States v. GM, 1966). Second, an explicit well-justified policy must be in place in advance; the manufacturer cannot pressure individual dealers, threaten suspension of shipments of new “hot” products, or offer to reinstate if the offending discounters agree to raise their prices (FTC v. Stride Rite, 1996). Finally, manufacturers cannot “lock up” buyers of durable products by refusing to supply parts to independent ser- vice organizations (ISOs), especially if the ISO prices are far below the manufacturer’s service prices. In Eastman Kodak v. Image Technical Services (1992), the Supreme Court argued that customers should be able to select independent service providers and nonwarranty repair from whomever they choose. Kodak’s defense that ISO main- tenance and repair failed to meet Kodak’s quality standards was disproved by the evidence.

Resale Price Maintenance Agreements Manufacturers often wish to limit their distributors’ flexibility to initiate price discounts. Resale price maintenance (RPM) agreements prohibit retailers from cutting the price at which they resell the product below a manufacturer’s suggested retail price (MSRP). Most such restrictions are illegal, especially when they appear motivated by the desire of competing retail dealers to reduce price competition. For example, Chevrolet dealers in Los Angeles once approached GM about using an RPM agreement to sanction the steep discounting of a “renegade” dealer in Orange County, California, who had markedly cut into everyone’s possible profit margins. In United States v. GM (1966), the Supreme Court declared this anticompetitive practice illegal per se since it was moti- vated by an express desire to lessen competition.

COMMAND AND CONTROL REGULATORY CONSTRAINTS: AN ECONOMIC ANALYSIS Federal, state, and local governments are involved in the regulation of business enter- prises. Table 16.3 contains a partial listing of the federal regulatory agencies and de- partments. In addition to the Federal Trade Commission and Antitrust Division of

622 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the Department of Justice discussed earlier, many other agencies regulate business deci- sions. State regulations encompass a wide range of activities, including regulation of public utility companies and licensing of various businesses, such as health care facili- ties, and numerous professions, such as law and accounting. Local governments fre- quently set and enforce zoning laws and building codes. Regulatory constraints can be imposed in non-discriminatory ways on any set of similar business. For example, the European Union prohibits direct-to-consumer advertising of prescription drugs. These constraints can affect a firm’s operating costs (both fixed and variable), capital costs, and revenues.

Example RPM at Stride Rite and Leegin Creative Leather Products7

In the late 1990s, Nintendo, New Balance Athletic, and Stride Rite all paid multimillion-dollar fines to settle charges that the manufacturers cut shipments to retail outlets that refused to charge the full MSRP. In Stride Rite’s case, leading retailers were cut off if they refused to sell six styles of women’s Keds at the full, undiscounted MSRP price of $45. Although Stride Rite insisted that it could suspend contracts with retailers who violated other company marketing policies and procedures, Stride Rite had refused to deal with only those particu- lar dealers who discounted Keds. The courts ruled Stride Rite guilty of an anticompetitive business practice because the suspended dealers had been pres- sured to raise prices. Vertical requirements contracting between dealers and manufacturers about matters other than resale price is widespread and perfectly legal.

Occasionally, however, a manufacturer or distributor can demonstrate a “legiti- mate manufacturer’s interest” in regulatory or industry standards that places a floor under the resale prices of its products (e.g., a rare book distributor). When one of these special exceptions is made and an RPM agreement is allowed, the new contours of antitrust policy rules are carefully examined.

Such an event occurred in June 2007. Leegin Creative Leather Products of Dallas cut off shipments of their high-end handbags, clutch purses, and key fobs to Kay’s Kloset, an independent retailer who refused to stop discounting Leegin’s products. Leegin argued that its business model to compete against Coach and Gucci relied upon its brand equity as an elite line of leather accessories and that this brand equity was tarnished beyond repair by discounting. On this reasoning, restrictive pricing agreements that enhance inter-brand competition (between Lee- gin and Coach) could be allowed even though they diminish intra-brand compe- tition (among Leegin’s distributors). The U.S. Supreme Court agreed in part, ruling that RPMs should be subject to a “rule of reason” rather than being illegal per se. Leegin’s policy of refusing to supply dealers who violate MSRP is under review by the original court. A final ruling is imminent and will be much discussed.

7Based on “StrideRite Agrees to Settle,” Wall Street Journal (September 28, 1993), p. A5; “Retail Price Maintenance Policies,” Knowledge @ Wharton (August 9, 2007); “Price Fixing Makes a Comeback,” Wall Street Journal (August 18, 2008), p. B1; and Leegin Creative Products Inc. v. PSKS Inc., 571 U.S. 877 (2007).

Chapter 16: Government Regulation 623

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Deregulation Movement Beginning in the late 1970s and continuing through the 2000s, the business environ- ment moved toward relying less on government regulation and more on the market- place to achieve desired economic objectives. For example, the Ocean Shipping Reform Act of 1998 deregulated freight rates for the 40-foot containers that ship apparel, con- sumer electronics, autos, and ores from Asia to the United States and ship computer software, forest products, and grain back to Asia. The one-way cost of $3,500 for a 6,000-garment container quickly declined to $1,500. Other consensus successes in de- regulation have come in airlines, railroads, telephones, and natural gas pipelines. How- ever, reduced regulatory supervision has at times proven problematic. Several large

TABLE 16.3 PARTIAL LISTING OF FEDERAL GOVERNMENT REGULATORY AGENCIES

DEPARTMENT/AGENCY PURPOSE

Environmental Protection Agency (EPA) Regulates pollution of air, water, and land

Consumer Product Safety Commission (CPSC) Protects against unreasonable risks of injury associated with consumer products

Equal Employment Opportunity Commission (EEOC) Enforces laws on employment discrimination based on race, religion, and sex

Labor: Employment Standards Administration Enforces minimum wage and overtime laws

Labor: Occupational Safety and Health Administration (OSHA)

Regulates safety and health conditions in the workplace

Labor: National Labor Relations Board (NLRB) Regulates labor relations between employers and employees (and their unions)

Interstate Commerce Commission (ICC) Regulates interstate surface transportation

Nuclear Regulatory Commission (NRC) Regulates civilian use of nuclear energy

Securities and Exchange Commission (SEC) Regulates issuance of new securities and trading of existing securities

Federal Communications Commission (FCC) Regulates radio and television broadcasting and interstate telephone service

Federal Reserve System Regulates commercial banks and bank holding companies

Agriculture: Food Safety and Inspection Service Regulates meat and poultry industry for safety and accurate labeling

Health and Human Services: Food and Drug Administration (FDA)

Regulates safety of food, drugs, and cosmetics

Energy: Federal Energy Regulatory Commission (FERC) Regulates interstate rates for transportation and sale of natural gas and transmission and sale of electricity

Transportation: Federal Aviation Administration (FAA) Regulates safety of airplanes, airports, and airline operations

Transportation: National Highway Traffic Safety Administration (NHTSA)

Regulates safety of motor vehicles and tires

Labor: Mine Safety and Health Administration Regulates safety and health in mines

Treasury: Office of Comptroller of the Currency Regulates national banks

Treasury: Bureau of Alcohol, Tobacco, and Firearms (BATF)

Regulates manufacture and sale of alcoholic beverages, tobacco, explosives, and firearms

624 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

investment banks, commercial banks, and insurance companies (e.g., Bear Stearns, Leh- man Brothers, Wachovia, and AIG) failed or were bailed out at taxpayer expense dur- ing the financial crisis of 2007–2009 when excessive leverage exposed these firms to massive default risk.

WHAT WENT RIGHT • WHAT WENT WRONG

The Need for a Regulated Clearinghouse to Control Counterparty Risk at AIG8

American International Group (AIG) sold $2 trillion ($2,000 billion) worth of loss protection against mortgage default risk to all the nation’s commercial banks like Bank of America, Wells Fargo, and Wachovia. By referring to these derivative contracts as credit default swaps (CDS), AIG escaped insurance industry regulation. When mort- gage delinquency rates rose in 2007 from their historical average of one-half of 1 percent to 2 percent in the prime mortgage market and from 2 percent to 13 percent in the subprime mortgage market, AIG was immediately insolvent, meaning that their liabilities exceeded the present value of their future cash flows and other assets. In fact, with less than $100 billion in capital and $2,000 billion owed on their derivative positions to the banks, AIG was an insolvent counterparty of grand proportions. Only approximately one in twenty of the CDS contingent claim contracts under which AIG had promised to cover the mortgage losses of the nation’s banks could be honored. Because of the systemic risk to the entire banking system, the Federal Reserve bailed out AIG in September 2008 and provided a guarantee that all AIG’s losses would be paid.

How did such a violation of capital adequacy require- ments in the regulation of banks and insurance companies ever happen? The full answer is complex but one key in- gredient was the absence of a central clearinghouse coun- terparty and its associated margin calls as AIG’s loss exposure worsened. In a typical derivative security transac- tion, investors are required to maintain escrow capital in their brokerage accounts to offset possible losses as the underlying security price on which they have written or purchased a future or an options contract moves against the position they have taken. Specifically, the clearinghouse who handles settlement of Chicago Board of Trade or other exchange-traded derivatives requires these so-called “margin calls” in the escrow accounts of the brokers. And the brokers then in turn require additional escrow deposits from their clients. But not so in the case of AIG! Why not? One reason is that the fragile CDS market totalled $42 trillion, three times the size of the entire U.S. GDP.

Mortgage-backed derivative securities such as CDSs were not traded through a regulated exchange or settle- ment clearinghouse. Instead, in 2003 the U.S. Congress took the advice of then Federal Reserve chairman Alan Greenspan to exempt mortgage-backed derivatives from these procedures. Greenspan testified that such instru- ments could reliably transact over the counter (OTC) with- out the scrutiny of a regulated exchange or clearinghouse. But what was sacrificed was precisely the crucial margin requirements that the regulated clearing and settlement process would have imposed on AIG. Instead, AIG simply entered into OTC private contracts with the commercial banks that were seeking to lay off their mortgage default risk. With 20 to 1 leveraging of their capital to loss expo- sure, AIG became too big and too interconnected to the U.S. banking system to be allowed to fail. So, banks con- tinued to buy AIG’s mortgage default risk products even though it was perfectly obvious that if the housing market turned sour, mortgage holders would default in large num- bers, thereby creating liabilities much bigger than AIG could possibly pay.

As the commercial bank’s counterparty on thousands of CDS mortgage loss protection contracts, AIG would normally have been required to escrow additional capital as properties in California and Florida collapsed in value by 20, 30, and ultimately 40 percent. At these asset values, mortgage borrowers found their real estate worth signifi- cantly less than their mortgage obligations, triggering mas- sive numbers of loan defaults. In a regulated clearinghouse, AIG would have been required to raise more capital to cover these impending losses, but instead they were al- lowed to simply write still more mortgage loss protection contracts. The Financial Reform Act of 2010 proposes to reverse the sadly mistaken 2003 decision that allowed mortgage-backed derivatives to “trade over the counter” with little regulatory supervision.

8Based on G. Gensler, “The Derivatives Debate,” Wall Street Journal (April 21, 2010), p. A21; and L. Ausbel and P. Cramton, “Auction Design Critical for Rescue Plan,” The Economist’s Voice, Berkeley Electronic Press (Septem- ber 2008).

Chapter 16: Government Regulation 625

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

REGULATION OF EXTERNALITIES In the normal course of business, every firm faces decisions that impose spillover costs upon third parties. The sharpest debate today is over air pollution, carbon footprints, and the need to reduce greenhouse gases to prevent catastrophic global warming. Both managers and the public have a keen interest in least-cost implementation of the kinds of remedies society mandates for controlling these externalities.

Externalities exist when a third party receives benefits or bears costs from consump- tion or production activities for which the market system does not enable them to re- ceive full payment. Pollution by-products of trucking deliveries, for example, combine with certain atmospheric conditions and cause smog. In places such as Los Angeles, this smog may impose significant costs on asthmatic residents and some businesses such as the Pasadena Sightseeing Company. In short, externalities arise with any interde- pendency of household utility or firm production functions that is not reflected in mar- ket prices.

Only externalities that are not conveyed through the price system result in inefficiencies. Thus when mad cow disease causes preference for meat to shift from beef to chicken, the price of beef will fall and that of chicken will rise, making beef producers and chicken consumers worse off, and chicken producers and beef consu- mers better off, because of the price change. But all of these so-called pecuniary exter- nalities have operated through the market price system, and they therefore pose no inefficiency.

The legal doctrine of “coming to the nuisance” in Spur Industries v. Del Webb Devel- opment, S.C. Arizona, 1972 (108 Ariz. 178, 494 P.2d 700) illustrates why pecuniary exter- nalities result in no inefficiency. If the land a developer purchases for a subdivision development is located next to a cattle feedlot, the price paid per acre will reflect the stench. The reduced price of the land internalizes the spillover effects. Later, if residents of the subdivision complain about the stench and the feedlot is declared a public nui- sance, the developer may have to pay to relocate the cattle feeding business. Again, when external effects are reflected in prices, all affected parties directly participate in the transaction, and there is no inefficiency.

When non-pecuniary externalities are present, however, resources are likely to be mis- allocated. Producers or consumers are less likely to engage in an action that contributes to society’s well-being if they are not fully compensated for all benefits generated. Simi- larly, in the case of negative externalities, a producer or consumer will likely over allocate resources to some production or consumption activity if part of the cost is shifted to others.

In general, an external cost should be reduced to the point where the marginal spill- over costs saved by any further reduction just equal the marginal lost profits from the externality-generating activity. Similarly, an action that generates external benefits should be expanded to the point where the marginal benefits to all of society from such an ex- pansion just equal the marginal costs.

Coasian Bargaining for Reciprocal Externalities In many cases, externalities arise because of incompatible uses of air, land, or water re- sources. For example, late-night takeoffs and landings by FedEx jets may disturb sleep in houses around the airport. Feeding of thousands of animals in a small enclosed feedlot create offensive odors in adjacent subdivisions. Agricultural land runoff of nutrient-rich water may adversely affect downstream intake by a bottled water plant. No adverse con- sequences would occur if either party were absent.

externality A spillover of benefits or costs from one production or utility function to another.

pecuniary externality A spillover that is reflected in prices and therefore results in no inefficiency.

626 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Ronald Coase’s famous paper, “The Problem of Social Cost,” illustrates a reciprocal externality between a spark-throwing railroad and a farmer with adjacent flammable fields. Coase’s ingenious and intriguing claim was that under certain conditions involving full in- formation and low transaction costs, the answer to the question, “Who’s liable and therefore who should pay damages?” had no effect on the resource allocation decisions of these parties. In particular, if the railroad had the property right to throw sparks along its right-of-way, the trains scheduled down this track and the acreage planted along it would be exactly the same as if the railroad had the liability for all spark-induced fire damage along its tracks.

Example Coase’s Railroad To see how this remarkable result arises through Coasian bargaining, consider the payoffs in Table 16.4. If the railroad has the property right, i.e., Panel (a), the farmer incurs $600 worth of crop destruction per train per 10 acres planted along the tracks. Initially, the railroad ignores these external spillover costs and chooses an activity level of trains that maximizes its own profits, i.e., two trains in the bot- tom row of Panel (a). The farmer would plant 10 rather than 20 acres along the tracks in order to earn $300 and avoid losing $800 (in the extreme southeast cell). If substantial impediments to bargaining were present, no further action would take place in an unregulated market environment. And yet, a mutually ben- eficial private voluntary bargaining opportunity would exist.

In particular, if the railroad were to cut back to one train, the farmer’s profit would rise from $300 to $900, while the railroad’s profit would decline by $500 (from $1,500 to $1,000). Accordingly, $501 is a minimally sufficient bribe to elicit the lower train-activity level, and $600 is the savings in fewer crops burned. Thus, Coase predicted that if the parties have few impediments to bargaining, the farmer would offer a side payment sufficient to abate the incremental (second) train and its spark hazard, because the second train is worth less (to the railroad) than the in- cremental agricultural losses cost the farmer. Just how much the farmer will pay and how little the railroad will accept is not addressed, but one thing is clear: Potential gains from trade do motivate a bargain to reduce railroad activity from two trains to one and crops acreage planted adjacent to the railroad from 20 to 10 acres.

Now, consider the case in which the railroad has the liability for spark-induced crop damages. Initially, the farmer prepares to plant 20 acres along the tracks as this activity level maximizes his or her independent profit (at $1,600). However, no trains are profitable with this much acreage in production because $600 in damages per train per 10 acres (i.e., $1,200 altogether) is owed when the railroad has $1,000 gross profit with one train, and $2,400 in damages is owed when the railroad offers to compensate the farmer for not only crop damages but also lost profit if the farmer would plant fewer acres. In particular, the railroad can offer the farmer $101 to plant 10 acres rather than 20 acres since the farmer’s gross profit differs by only $100 (i.e., $1,500 versus $1,600). If the railroad then also compensates the farmer $600 for one train’s crop damage on 10 acres, the railroad owes $701 and earns a gross profit of $1,000.

Table 16.4, Panel (b), displays the gross profits before crop damages have been com- pensated. Shown in the middle row of Panel (b), the railroad offers the farmer compen- sation in excess of $100 (perhaps, $101) to scale back the acreage planted from 20 acres, where farmer gross profit is $1,600, to 10 acres, where farmer gross profit is $1,500. This reallocation of activities is worth $600 in damage savings to the railroad. Again, Coasian bargaining leads the parties to agree upon one train and 10 acres.

reciprocal externality A spillover that results from competing incompatible uses.

Chapter 16: Government Regulation 627

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The Coase theorem states that reciprocal externality generators and recipients will choose efficient activity levels whatever the initial liability assignment. It makes no claim about the distributional consequences of reversing the direction of a liability assignment. Quite obviously, making the railroad liable in one instance, instead of asking the farmer to cover his or her own crop losses from burned fields in the other, results in quite different net profit outcomes. However, what the Coase theorem does assert is that in reciprocal externality settings with small numbers of affected parties, resource allocation as to the externality- generating and externality-receiving activity levelswill be unchanged independent of the initial liability assignment; one train will be scheduled, and 10 acres will be planted.

Qualifications of the Coase Theorem Some qualifications are in order, many of which Coase himself recognized. First, techni- cal transaction costs of searching for and identifying the responsible owners and affected parties, of detecting violations of one’s property rights, and of internally negotiating the side payments within a group of claimants must all remain low and be unaffected if lia- bility assignment is reversed. Second, neither party can operate in a purely competitive market, for then the net profits required for side payments would be nonexistent. And, third, and perhaps most important, one party quickly makes an offer the other is just willing to accept only when the information regarding the payoffs in Table 16.4, Panels (a) and (b), is complete, certain, and known to both parties.

When information is incomplete, private voluntary bargaining doesn’t necessarily lead to resource allocation that is variant to the direction of liability assignment. This qualifi- cation of the Coase theorem holds even if property rights are fully specified, completely assigned, and enforced at little or no cost. In all incompatible use lawsuits involving asymmetrically known cost information, the precautionary actions of the plaintiff and the defendant have some bearing on the assignment of liability. For example, the parties in Coase’s railroad example would avoid liability in part by employing spark arresters or land setbacks as long as the benefit in crop-loss savings exceeded the cost. However, the problem posed by asymmetric information is that some aspects of precaution are inher- ently unobservable or unverifiable (e.g., attentiveness to subtle signals of impending haz- ard) while others are observable but affect accident avoidance in a nondeterministic way

TABLE 16.4 COASIAN BARGAINING

(a) Gross profits

(RR has the property right)

Farmer (Acres planted)

0 10 20

0

Railroad 1

(Trains per day)

2

0 $1,500 $1,600

0 0 0

0 $900 $400

$1,000 $1,000 $1,000

0 $300 $800

$1,500 $1,500 $1,500

(b) Gross profits

(RR has the liability)

Farmer (Acres planted)

0 10 20

0

Railroad 1

(Trains per day)

2

0 $1,500 $1,600

0 0 0

0 $1,500 $1,600

$1,000 $400 $200

0 $1,500 $1,600

$1,500 $300 $900

Source: Adapted from R. Coase, “The Problem of Social Cost,” Journal of Law and Economics 2 (October 1960), pp. 1–44.

Coase theorem A prediction about the emergence of private voluntary bargaining in reciprocal externalities with low transaction costs.

628 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

(e.g., good brakes may lock up on rain-slickened roads when less effective brakes would not). Recall that unobservability and random disturbances together result in the problem of moral hazard, which we discussed in Chapter 15.

No incentive-compatible mechanism can both preserve the voluntary nature of the Coasian bargaining and also elicit true revelation of the unobservable damages. There- fore, contrary to the traditional understanding of the Coase theorem, disputants in recip- rocal externality conflicts might be expected not to engage in private voluntary bargaining alone, but rather to delegate the question of damage assessment and recovery to third-party court systems.9 Civil procedural rules in an impartial court system can be seen as credible commitment mechanisms. Through these mechanisms potential dispu- tants submit to liability assignments and wealth transfer remedies that motivate efficient accident avoidance despite frequently asymmetric information. So, the implication of the Coase theorem holds: Parties that are disputing over incompatible uses will contract their way to an efficient allocation of resources unless high transaction costs impede the re- quired bargaining.

Impediments to Bargaining Regulation will continue to have a role, however, because of impediments to bargaining. Several impediments are recognized in the legal system. Prohibitive notification and search costs (to identify absentee owners and notify all the affected parties) are the justi- fication for certifying class action suits in the case of oil spills and other large-scale ex- ternalities affecting many claimants. Voluntary private bargaining about incompatible uses may also be impeded by the need for continuous monitoring of an unverifiable bar- gain, such as an agreement to restrain one’s catch to the maximum sustainable rate of harvest of a deep sea fishery. Therefore, public agencies must regulate the catch.

Finally, the most significant impediment to bargaining in large-numbers externality cases is the strategic holdout or free rider. When a court grants an injunction against a polluter’s operation, relief from the injunction may require that the polluter obtain a unanimous waiver from all the affected parties. If many claimants are certified as posses- sing such a right of waiver, each claimant has an incentive to hold out for more compen- sation than would be required to cover his or her damages. The predictable presence of strategic holdouts short-circuits Coase’s private voluntary bargaining hypothesis. In such cases, the courts therefore adopt other mechanisms such as assigning liability for payment of permanent damages.

Example Mandatory Auto Inspections Auto inspections are a good example of a solution by regulatory directive. Recog- nizing that significant external benefits are to be gained from reducing serious au- tomotive accidents caused by poorly maintained equipment, no state government gives the consumer the choice of refusing regular inspection and maintenance of brakes, lights, and other safety equipment; it simply mandates inspection and re- pair for all vehicles. Cancer-causing lead additives in gasoline and ozone shield- depleting chlorofluorocarbon (CFC) refrigeration gases have also been massively reduced by such mandatory regulatory directives (see Figure 16.3).

9See F. Harris, “Economic Negligence, Moral Hazard, and the Coase Theorem,” Southern Economic Journal 56, no. 3 (January 1990), pp. 698–704.

class action suits A legal procedure for reducing the search and notification costs of filing a complaint.

strategic holdout A negotiator who makes unreasonable demands at the end of a unanimous consent process.

Chapter 16: Government Regulation 629

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Resolution of Externalities by Regulatory Directive Another approach to resolving externality problems is to prohibit the action that gener- ates the external effects. In most cases, however, this simplistic approach is suboptimal and frequently impractical. Auto emissions could be cut to zero if autos were banned, but the economic effects of such a move, at least in the short run, would be disastrous for all developed economies. Moreover, seldom does an optimal solution require that ex- ternalities be completely eliminated; a strict zero-pollution policy often entails excessive pollution-abatement costs.

However, it is seldom transparently obvious exactly what regulatory directive to is- sue. Consider the problem of setting an overall emissions standard when multiple sources of pollution are present, as in the acidification of rain along the Eastern forests from coal-fired power plants in the Midwest. Under command and control regulation, each of the polluting entities (each point source) must be directed as to how it should act. A simple proportionate distribution of “pollution rights” to each plant would not take into account the dramatic difference in the cost of abatement from one plant to another. Instead, optimality would equalize the marginal effectiveness of the last dollar spent on pollution abatement by each polluter. So, a low-cost point source’s regulatory directive should require more abatement than a permit for a high-cost point source. Yet this sort of detailed point-source regulation is seldom achieved.

Resolution of Externalities by Taxes and Subsidies Another potentially efficient solution to externality problems is to provide subsidies (either in the form of cash or tax relief) to those whose activities generate significant ex- ternal benefits and to levy a tax on those whose activities create external costs. Such a tax

Example Boomer v. Atlantic Cement Co., Inc., 26 N.Y. 2d 219 A large cement plant valued at almost $200 million spewed cement dust regularly across a neighborhood of Albany, New York. Some of the affected households were unable to continue their laundry operations; small airborne particulates re- quired frequent washing and repainting of cars and homes. Asthmatics suffered more health problems. The Atlantic Cement plant was declared a public nui- sance, and the court chose among three types of injunctions: (1) an order to cease operations until the air pollution could be abated, (2) an order to cease operations until a waiver could be obtained from each affected household, or (3) an order declaring the cement plant liable for $740,000 in permanent da- mages and requiring a cessation of operations until these court-specified damages were paid.

Because the first injunction hinged on undeveloped technology and the second created strategic holdouts, the New York Court of Appeals in effect licensed the ongoing nuisance to Atlantic Cement Co. for a one-time fee of $740,000 in 2006 dollars. No private voluntary bargain to reduce the cement dust could have over- come the strategic free-rider/strategic holdout problem. And, as result of having to pay court-mandated damages, the plant’s owners did begin to internalize the social cost of cement production when establishing new plants.

630 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

and subsidy scheme, however, requires a tremendous amount of information to admin- ister the program effectively.

Consider the analysis required for a per-unit pollution tax T* in Figure 16.4. Demand per truck is the private willingness to pay (WTP) for trucking deliveries throughout the Los Angeles area. Setting marginal private cost (MPC) equal to private WTP, the truck- ing company will put 50,000 miles a year on its typical delivery truck. Still more delivery miles are avoided by the managers because the price additional customers are willing to pay for deliveries is less than the MPC of operating the truck. The problem is that trucking mileage generates the by-product nitrous oxide (NO2), which causes smog. Through careful environmental science, businesses such as the Pasadena Sightseeing Co. and asthmatic citizens of Los Angeles estimate that the air pollution causes damages from lost tourist business as well as eyes, nose, and throat irritants of area BCD. Consequently, although marginal benefit (Po) equals MPC at the private market equilibrium Point A, additional costs attributable to the NO2 externality—namely, CB at 50,000 miles—suggest that full costs are substantially higher: at Point E, MPC (AB) + MExC (CB) > Po.

When private plus external costs exceed marginal benefits to the delivery truck’s cus- tomers at 50,000 miles, the joint product trucking mileage/NO2 is produced in excess of its optimal level. Fifty thousand miles per year is too much trucking. However, the ques- tion, as always, comes down to how much less trucking and associated pollution abate- ment is optimal?

In Figure 16.4, the reduction inmileage at which marginal social cost (MSC), which is the sum of MPC +MExC, just equals marginal willingness to pay for trucking is 40,000 miles at Point F. Clearly, the smog victims have damages (area GHBC) great enough to compensate

FIGURE 16.3 CFC Production

0

50

1988 1990 1992 1994 1996

100

150

200

250

300

350

T ho

us an

d to

ns

India

China

France

Germany

Japan

United States

Source: Tomorrow’s Markets, Global Trends and Their Implications for Business, World Resources Institute (2002), p. 27.

Chapter 16: Government Regulation 631

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

the trucking company for its lost profits (area FIA) associated with a 10,000-mile reduction in mileage. And the maximum side payment smog victims would offer for the next 10,000- mile reduction from 40K to 30K (area JKHG) is smaller than the minimally sufficient bribe (area LMIF) that the trucking company would accept. But zeroing in on 40,000 miles as the optimal mileage (not 35K, 38K, 42K, or 45K) is difficult because accurate marginal ex- ternal cost information is so hard to come by. An optimal per-unit tax of T* levied on delivery truck mileage reduces the mileage chosen from 50,000 to 40,000 through user charges in the amount Po + T* that reflect the marginal social cost of the (trucking/NO2) joint product. But again, T* assumes heroically that the regulators know that 40,000 con- stitutes the optimal mileage.

In practice, a tax or emission charge would be placed on a firm’s pollutants, such as pounds of particulate matter emitted from a delivery truck or power plant smokestack. A firm could continue to pollute if it pays the per-unit tax, or it could find that it is cheaper to buy pollution control equipment. If, after a reasonable period of time, a com- munity still believes the level of particulate matter in the air is too high, the tax per pound of pollutant would be increased in a stepwise fashion until the community was satisfied with the result. The per-unit tax solution avoids the rigidity of all-or-nothing regulatory directives but the exact amount of an optimal effluent tax for water pollution or emission tax for air pollution is extremely difficult to estimate.

Resolution of Externalities by Sale of Pollution Rights: Cap and Trade Another increasingly popular approach to the problem of pollution is the issuance of transferable pollution rights. In effect, licenses are sold that give the license holder a right to pollute up to some specific limit during a particular period of time. This approach was adopted under the 1990 Clean Air Act. The U.S. Environmental Protection Agency (EPA)

FIGURE 16.4 Optimal per Unit Pollution Tax

B

C

M

L

H

G

I A

E F

K J

D

T*

10k 20k 25k 40k 50k

Po+T*

Po

Trucking Miles/NO2

Pr ic

es a

nd C

os ts

( $)

Minimally Sufficient Bribe

Maximum Offer Side Payment

Marginal Social Cost (MSC) MPC + MExC

Demandtrucking

Marginal Private Cost (MPC)

Marginal Damage (MExC)

632 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

sets a maximum level of some pollutants that may be safely emitted in an area. The fed- eral government then sells, at auction, licenses to individual firms giving them the right to pollute up to that specified amount. The licenses can be freely traded, which permits their price to fluctuate with market demands and innovations in abatement technology. Be- cause it is essentially market oriented, this approach causes pollution costs to be internally recognized in all the price and production decisions of individual firms.

GOVERNMENTAL PROTECTION OF BUSINESS In addition to regulating business enterprises, numerous government programs and poli- cies protect businesses by restricting the entry of competitors.

Licensing and Permitting When the government requires and issues a license permitting someone to practice a particular business, profession, or trade, it is by definition restricting the entry of some other competitors into that business. Licensing is generally used to protect the public from fraud or incompetence in those cases where the potential for harm is quite large. Nevertheless, by restricting output, government licensure generates market power for license or permit holders.

In 2003, Mecklenburg County, North Carolina, which surrounds Charlotte, learned that its role as a ground transportation hub was threatened because the smog permits authorized by the North Carolina Environmental Commission were unavailable. North Carolina regulatory agencies foreclosed the construction of additional freight trucking terminals until some other dust-generating facility was closed. If no additional permits will be authorized, those businesses with preexisting permits become more valuable.

Patents Patents are by definition a legal governmental grant of monopoly power. A patent holder may prevent others from manufacturing or selling a patented product or process and may grant a license permitting others to make limited use of the patent in exchange for royalty payments. The monopoly granted by a patent is not, however, an absolute

Example Plant Expansion Requires Purchase of Open-Market Pollution Rights: Times Mirror Co. The United States and the European Union have active markets for sulfur dioxide and nitrous oxide pollution rights. These markets allow electric utilities, trucking companies, and manufacturing firms to buy and sell pollution credits in continu- ous auctions. Also, a private placement market exists in which brokers, as well as some states, arrange customized contracts for pollution credits between companies that have excess rights to sell and companies that need rights to comply with envi- ronmental regulations. For example, the Times Mirror Company was able to com- plete a $120-million expansion of a paper-making plant near Portland, Oregon, after buying the right to emit 150 tons of additional hydrocarbons into the air an- nually. The pollution rights were acquired for $50,000 from the owners of two businesses that held surplus emission rights—a chemical plant that had gone out of business and a dry-cleaning firm that adopted a pollution-free cleaning fluid.

patent A legal government grant of monopoly power that prevents others from manufacturing or selling a patented article.

Chapter 16: Government Regulation 633

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

one. First, it is limited to a 17-year period, and few renewals are permitted. It is possible that a shorter patent monopoly period would provide sufficient incentives to encourage a high level of inventive activity. Serious proposals suggest shortening the patent period to just four years for computer software, for example. Second, competing firms are not pro- hibited from engineering around an existing patent and bringing out a closely compet- ing, alternative design. Third, many patents are successfully challenged by competitors, especially in the European Union, where patent applications are not kept secret.

THE OPTIMAL DEPLOYMENT DECISION: TO LICENSE OR NOT Finally, we discuss the decision as to whether to license patents and trade secrets to com- petitors. Think of Apple’s source code for the graphical user interface, Pfizer’s automated drug discovery technology, and Disney’s films and characters. Fifty years ago, 78 percent of the assets of U.S. nonfinancial corporations were tangible assets (real estate, inventories, and plant and equipment). Today, that figure is just 53 percent; intangible assets such as patents, copyrights, and goodwill have grown to nearly dominate the balance sheet.

In 2006, 10 firms passed Thomas Edison’s record of 1,093 patents acquired (see Table 16.5). IBM acquired 2,972 patents, Canon acquired 1,837, and Hewlett-Packard acquired 1,801. IBM Corporation was seeking patent protection at the astounding rate of 10 pat- ent applications per working day.10 Some of this activity is strategic patenting of technol- ogy portfolios where companies do not wish to make a new device immediately, but they can plausibly describe how they would make it, what the device is used for, and the nov- elty of the idea. These evidentiary requirements are part of obtaining a patent.

Not just electronic devices, genetic engineering, and computer software, but also busi- ness process methods are “hot” current areas of patenting activity. Dell received a patent on the direct-to-the-consumer business model. Patent filings for business processes ex- ceeded 10,000 per year for the first time in 2007. Patent attorneys believe the ATM ma- chine, frequent flyer programs, and even credit cards could be patented as business processes if they were invented today.

TABLE 16.5 TOP 10 FIRMS WITH PATENTS GRANTED IN 2006

IBM 2,972

Canon 1,837

Hewlett-Packard 1,801

Matsushita 1,701

Samsung 1,645

Micron Technology 1,561

Intel 1,551

Hitachi 1,293

Toshiba 1,149

General Electric 906

Source: The Economist (May 10, 2008), p. 75.

10“The Knowledge Monopolies,” The Economist (April 8, 2000), pp. 75–78; “Business Methods Patents,” Wall Street Journal (October 3, 2000), p. B14; “Mind Over Matter,” Wall Street Journal (April 4, 2002), p. A1; and “Innovation.” Forbes (July 5, 2004), pp. 142–146.

634 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Table 16.6 shows that the financial markets are definitely capitalizing this “knowledge capital” into the equity market value of companies with patents, trade secrets, and proprietary know-how. Almost half of the $22 billion market value of Dow Chemical and more than a third of the $140 billion market value of Merck are discounted cash flow from intangible assets, mostly intellectual property. Much of Amazon.com’s $11 billion market value arises from licensing fees on their business methods patents. Between 1994 and 1999, IBM boosted its annual revenues from licensing its intellectual property by more than

WHAT WENT RIGHT • WHAT WENT WRONG

Delayed Release at Aventis11

The following list shows the top 10 patented drugs in 2000 and 2006. The instability on this list is striking. Once a drug goes off patent (e.g., Prozac in 2001 and Zocor in 2004), the sales evaporate quickly. Generic drugs have be- come 46 percent of the U.S. prescription drug market. Generics usually enter the market at a 60–75 percent dis- count and eventually sell for as little as 20 percent of the original patented drug’s price. Blue Cross and Blue Shield estimated this price difference to be $84 per prescription. In 2011, more than twice as many patented pharmaceuti- cals ($42 billion worth) are coming off patent than in any previous year.

When Eli Lilly’s patent on Prozac ended, the firm lost 70 percent of its sales to generic rivals within one month. More typically, the loss of sales to generics for U.S. drugs coming off patent is 85 percent the first year. To prevent such disruptive shocks to the recovery of enormous fixed

R&D costs, some pharmaceutical companies routinely file frivolous patent extensions that change simply the coating or the delivery system (tablet to liquid, for example). When Prilosec, the anti-heartburn pill, went off patent in 2005, AstraZeneca introduced the copycat prescription drug Nexium. Of equal concern, Aventis (a Franco-German drug group headquartered in Strasbourg, France) is ac- cused of bribing the American generic drug producer An- drix with $90 million to delay the introduction of its cheaper substitute heart attack drug. Such sums suggest just how extensive the monopoly markup on patented pharmaceuticals must be.

11Based on “Don’t Look Down,” The Economist (January 6, 2001), p. 62; “Bloom and Blight,” The Economist (October 26, 2002), p. 60; “Protection Racket,” The Economist (May 19, 2001), p. 58; www.imshealth.com (ac- cessed September 2006); “Friends for Life,” The Economist (August 25, 2009), pp. 55–56; and “Something Rotten,” The Economist (August 8, 2009), p. 12.

BEST-SELLING PATENTED DRUGS, 2000

DRUG PATENT OWNER TREATMENT DRUG

PATENT OWNER TREATMENT

1. Losec AstraZeneca ulcers 6. Prozac Eli Lilly depression

2. Lipitor Pfizer cholesterol 7. Celebrex Pfizer arthritis

3. Zocor Merck cholesterol 8. Seroxat GlaxoSmithKline depression

4. Norvasc Pfizer high blood pressure

9. Claritin Schering-Plough allergies

5. Orgastro Abbott Labs ulcers 10. Zyprexa Eli Lilly schizophrenia

BEST-SELLING PATENTED DRUGS, 2006

DRUG PATENT OWNER TREATMENT DRUG

PATENT OWNER TREATMENT

1. Lipitor Pfizer cholesterol 6. Enbel Angen Wyeth arthritis

2. Nexium AstraZeneca acid reflux 7. Effexor Wyeth depression

3. Plavix Sanofi platelet aggregation

8. Orgastro Abbott Labs ulcers

4. Serentide GlaxoSmithKline allergies 9. Zyprexa Eli Lilly schizophrenia

5. Norvasc Pfizer high blood pressure

10. Singulair Schein allergies

Chapter 16: Government Regulation 635

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

200 percent from $500 million to $1.6 billion. In consumer products, too, Reebok recently paid $250 million in royalties to obtain a 10-year exclusive license to market National Football League–branded uniforms, hats, and equipment and to have its trademark on the players’ apparel of all NFL teams. So, substantial revenue is available from licensing, but of course licensing better enables one’s competitors to compete for a firm’s own regular customers.

Pros and Cons of Patent Protection and Licensure of Trade Secrets Much debate continues about whether patent and trade secret protection motivates first- mover companies to innovate or stifles technological research by fast-second companies. Imitators often substantially advance some aspect of a new technology, but they must license the original patents or run the risk of defending themselves against patent in- fringement lawsuits. Jeff Bezos of Amazon.com proposed that 20-year patent protection for computer software and business methods be reduced to only 3–5 years. Patent pro- tection outside the United States is already diminished. In Europe, patent applications invite legal challenge because they are publicized, and a majority of initial EU patents have been overturned. The EU has also decided not to issue patents for either computer software or business methods. In that environment, trade secrets, proprietary know-how, and internal business practices take on added importance.

Whether to “bury” the trade secrets or to acknowledge openly their existence and license them to competitors is a significant strategic decision about the firm’s con- tracting and governance mechanisms. No less important is the decision as to whether to develop know-how in-house or to license proprietary know-how from competitors. Many executives believe that manufacturing managers and R&D experts should inter- act on an ongoing basis in order to develop a reservoir of non-codifiable tacit knowl- edge. If licensing is pursued, two-way licensing of trade secrets represents the

TABLE 16.6 KNOWLEDGE CAPITAL, 1998 ($ BILLIONS)

SALES BOOK VALUE

TOTAL MARKET VALUE

MARKET VALUE OF INTANGIBLES*

Merck 23.6 12.6 139.9 48.0

Bristol-Myers Squibb 16.7 7.2 107.0 30.5

Johnson & Johnson 22.6 12.4 92.9 29.7

DuPont 39.9 11.3 87.0 26.4

Dow Chemical 20.1 7.7 21.8 10.2

Monsanto 7.5 4.1 33.2 6.0

BUSINESS METHOD PATENTS

PATENT NUMBER DATE ISSUED DEVICE/ PROCESS INVENTORS AFFILIATE

5,797,127 8/18/98 Reverse auctions Jay Walker et al. Priceline.com

5,960,411 9/28/99 One-click buying Jeff Bezos et al. Amazon.com

*Estimated market value attributable to intangible, nonfinancial assets. Source: Baruch Lev, CFO (February 1999); and The Economist (June 12, 1999), p. 62.

636 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

exchange of hostages recommended by game theorists for credible long-term reliance relationships.

In Table 16.7, Motorola and Lucent Technologies (the spin-off of Bell Labs from AT&T) are trying to decide whether to develop in-house or license proprietary trade se- crets in telecommunications engineering and software. Because of Lucent’s long-term ex- perience in this arena, if Motorola develops and patents the process and devices, Lucent expects to be successful as an imitator earning $9 billion. Should licensing of some pro- prietary know-how prove necessary, Lucent believes an inexpensive limited license will be sufficient. Consequently, Motorola will be unable (in that circumstance) to recover

WHAT WENT RIGHT • WHAT WENT WRONG

Technology Licenses Cost Palm Its Lead in PDAs12

In 1996, Palm single-handedly created the personal digital assistant (PDA) craze. Like Apple with its later iPhone and RIM’s BlackBerry, Palm builds its own software and hard- ware. Initially, Palm operating systems ran three-quarters of all handheld devices that are capable of surfing the In- ternet. Palm’s Pre and Pixi devices are $150 and $80, re- spectively, much cheaper than the Apple iPhone, but their applications number about 1,000 while the Apple iPhone has 100,000. Unlike Apple, Palm decided to license its OS technology to competing manufacturers Handspring and Sony. Within two short years, Handspring surpassed Palm in manufacturing sales of PDAs by offering expansion slots and peripheral equipment such as phones and music players.

Licensing always entails such risks, but Palm really had little choice in the matter. Cell phone giant Nokia had

licensed its Series 60 mobile-phone software to Siemens and Matsushita. The three firms together control 47 percent of the global cell phone market. Series 60 technology enables a cell phone to send and receive digital camera pictures and e-mail and, most importantly, to browse the Net. Palm knew that is Nokia succeeded in getting its OS adopted as an industry standard for handheld Web surfing, Nokia would set in motion a virtuous circle of increasing returns. Recently, Palm and Handspring merged into pal- mOne to achieve a larger installed base because worldwide Palm has shrunk to less than 10 percent of Nokia’s market share.

12“Matsushita to Use Nokia’s Cellphone Software,” Wall Street Journal (December 20, 2000), p. B10; “One Palm Flapping,” The Economist (June 2, 2001), p. 65; and “As Its Phones Flop, Palm Shares the Blame,” Wall Street Journal (February 26, 2010), p. B1.

Example Competing Business Plans at Celera Genomics and Human Genome Sciences Genomics has revolutionized drug discovery and development, with some re- spected industry analysts predicting that “all drug discovery efforts will soon be genomics-based.”13 Celera Genomics, the company that finished reading the hu- man genome sequence in 2000, expects to sell information, in effect to license its genome database, for as much as $90 million a year. It is hoped that comparing which genes are expressed and which remain recessive in various diseases will lead drug scientists to new blockbuster therapies and early detection of harmful side effects. However, an in-depth understanding of the biology of therapeutic mechanisms at the molecular level will also be key. Human Genome Sciences (HGS) has decided therefore to position itself as a drug maker, attempting to pat- ent drug processes, not simply license genetic information to traditional drug com- panies. HGS’s first product is a gene therapy that speeds the healing of wounds.

13BusinessWeek (January 8, 2001), p. 113.

Chapter 16: Government Regulation 637

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

its fully allocated research and development costs and will therefore lose money (i.e., the −$1 billion payoff in the southeast cell). In contrast, if Lucent develops and patents the needed process, its first-mover advantage would bring in substantial license fees from Motorola, who would need the proprietary knowledge gained through the trade secret license. Hence, the payoffs describing this situation are $4 billion/$3 billion in the northeast cell.

To complete the description of the payoff matrix, if neither firm develops the process, no profits accrue to either party. And if they compete head to head in a patent race, we assume the development costs will rise such that total profits fall from $7 billion to $6 bil- lion, divided $5 to $1 between the technology leader Lucent and follower Motorola. What should Lucent do?

If Lucent could be sure Motorola was proceeding, Lucent would most prefer to wait and play “fast second.” Relative to $5 billion in the northwest cell, $9 billion in the southwest cell is certainly attractive. However, Motorola can be expected to avoid the

TABLE 16.7 TO LICENSE OR DEVELOP EXPERTISE IN-HOUSE?

Motorola

Develop/Patent Imitate/License

Develop/Patent

Lucent

Imitate/License

$1 billion $3 billion

$5 billion $4 billion

$1 billion 0

$9 billion 0

WHAT WENT RIGHT • WHAT WENT WRONG

Motorola: What They Didn’t Know Hurt Them14

Motorola, Inc., was a pioneer in communications engineer- ing with many of the early analog devices in radio, televi- sion, and military signal processing to its credit. More recently, Motorola developed and successfully launched the first handheld cell phones and also took the lead in satellite-based wireless communications with Iridium, a global cellular network project. Ambitious future projects include a satellite-based, high-speed, high-security video- conferencing network for corporate customers and a satellite-based transcontinental and transoceanic connec- tion for land-based cell phone companies.

Network reliability problems began to arise, however, when Motorola insisted on slowly developing its own digi- tal wireless proprietary know-how rather than licensing the needed trade secrets and patents from Lucent or QUAL- COMM Incorporated. Motorola had little expertise in dig- ital switches, computing equipment, and communications software. Yet proprietary knowledge in these areas proved

critical in attempting to integrate Motorola’s satellite sys- tem with land-based cell phone networks. At one point, Motorola launched a cell phone system whose software essentially blocked any other user from simultaneously connecting through the same cell tower and receiving sta- tion. In effect, this device crashed the local cell network any time it was in use.

Perhaps it is not surprising that QUALCOMM and Lu- cent, a former division of AT&T, experienced less trouble adding know-how in wireless technology to their long- standing expertise in wire-based telecommunications net- works than Motorola experienced trying to add know-how in digital switches and communications software to their long-standing expertise in analog wireless hardware. Mo- torola should have licensed the proprietary know-how rather than attempt to develop it in-house.

14Based on “Unsolid State: Motorola Struggles to Regain Its Footing,” Wall Street Journal (April 22, 1998), p. A1; and “How Motorola Roamed As- tray,” Wall Street Journal (October 26, 2000), p. B12.

638 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

development expense and attempt itself to wait, imitate, and license as required to fill in the gaps in its own trade secrets and proprietary know-how. Indeed, Motorola has a dominant strategy to wait, imitate, and license. Consequently, Lucent anticipates that the payoff ($4 billion/$3 billion) in the northeast cell will emerge as an iterated dominant equilibrium. Recall that an iterated dominant equilibrium strategy is a self-interest maxi- mizing action by Lucent that is consistent with dominant-strategy responses of Motorola. At the northeast cell ($4 billion/$3 billion), neither party wishes to deviate to another action; therefore {DevelopLucent, LicenseMotorola} is a Nash equilibrium strategy and the only Nash equilibrium in Table 16.7.

Although the numbers in Table 16.7 are only illustrative, thinking through the game theory analysis can often provide insight in predicting rival reaction to company moves and countermoves. In this case, an analysis such as Table 16.7 clearly indicates the ad- vantages of the licensing alternative rather than the in-house development Motorola actually pursued.

Conclusion on Licensing The decision to license depends in part on the availability of increasing returns and the sustainability of a company’s competitive advantage from cost reductions. In Europe, where few industry standards have emerged for information technology products and where patents are often successfully challenged, first-mover firms have licensed to com- petitors rather than simply watch their trade secrets and proprietary know-how be steadily eroded by imitators. The result is increased competition, lower prices for consu- mers, and a faster rate of technological adoption. For example, in Europe prices for some digital TV components (e.g., digital video broadcasting chips) keep dropping, and the digital technology is quickly being incorporated into related products such as cell phones, pagers, and secure video business networks for corporate meetings.

In the United States, Red Hat used a general public license to penetrate as quickly as possible into the operating system market with its Linux-based software that is intended to compete with Windows NT and eventually with Windows itself. Red Hat allows its suppliers and customers to copy, modify, and redistribute Red Hat software at no charge as long as they do so without charge. This open-source software strategy is an attempt to achieve the inflection point for increasing returns that Microsoft’s competitors, including Apple, never reached. Apple Computers pursued the opposite nonlicensing strategy, thereby effectively slowing the rate of adoption of Macs, and lost to IBM and Microsoft. Today, Google is pursuing Red Hat’s open source strategy with the operating system for its new smartphone offering, the Android phone.

A final tactical advantage of licensing comes from reducing re-contracting hazards. In purchasing high-end Alpha chips from Digital Equipment Corporation (now a division of Hewlett-Packard), many workstation manufacturers worry about later being subject to a “hold-up.” At contract renewal, once their designs are optimized for the Alpha technology, the manufacturers are vulnerable to major price increases for these sole-source-supplied chips. Digital can credibly commit to more stable prices and thereby increase the rate of adoption for its product by licensing to both AMD and Intel. By licensing and allowing customers to dual source the Alpha chip technology, Digital credibly commits to renew its supply contracts without price gouging.

Chapter 16: Government Regulation 639

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

SUMMARY

� Market performance refers to the efficiency of re- source allocation within and among firms, the technological progressiveness of firms, the ten- dency of firms to fully employ resources, and the impact on the equitable distribution of resources.

� Market conduct refers to the pricing behavior; the product policy; the sales promotion and advertis- ing policy; the research, development, and innova- tion strategies; and the legal tactics employed by a firm or group of firms.

� Market structure refers to the degree of seller and buyer concentration in a market, the degree of ac- tual or imagined product differentiation between products or services of competing producers, and the conditions surrounding entry into the market.

� Contestable markets are assumed to have freedom of entry and exit for potential competitors, slow- reaching incumbents, and low switching costs for consumers. In a perfectly contestable market, the resulting set of prices and outputs approaches those expected under perfect competition.

� Measures of market concentration include: � the market concentration ratio, defined as the

percentage of total industry output attributable to the 4, 8, 20, or 50 largest companies.

� the Herfindahl-Hirschman Index (HHI), which is equal to the sum of the squares of the market shares of all firms in an industry.

� A group of antitrust laws has been passed to pre- vent monopoly and to encourage competition in U.S. industry. The most important of these acts are the Sherman Act of 1890, the FTC and Clayton Acts of 1914, the Robinson-Patman Act of 1936, the Celler-Kefauver Antimerger Act of 1950, the Hart-Scott-Rodino Antitrust Improvement Act of 1976, and Merger Guidelines of 1992 and 1997.

� Federal, state, and local governments all impose reg- ulations on business enterprises. Regulatory con- straints can affect a firm’s operating costs (both fixed and variable), capital costs, and revenues.

� The current political and economic environment favors a significant reduction in the amount of

government regulation and interference in the op- eration of the private sector of the economy. Re- cent deregulation includes banking, transportation, natural gas pipeline, electric utility, and telecom- munications industries.

� A number of regulatory policies, such as licensing and issuing patents, are designed to restrict competition.

� Externalities exist when a third party receives benefits or bears costs arising from an economic transaction in which he or she is not a direct participant. The impact of externalities is felt outside of (external to) the normal market pricing and resource-allocation mechanism.

� Pecuniary externalities, in which spillover effects are reflected in the market pricing mechanism, re- sult in no inefficiencies.

� Ronald Coase has shown that an efficient alloca- tion of resources can generally be achieved in the case of small-numbers externalities by contractual bargaining between the creator and recipient of the externality.

� Impediments to private voluntary bargaining in- clude prohibitive search and notification costs, in- ternal negotiation costs among large numbers of affected parties, prohibitive monitoring costs, and an absence of the surpluses required for making side payments.

� Many possible solutions to problems of externali- ties exist and include solution by voluntary side payment, governmental prohibition, regulatory directive, imposition of pollution taxes or subsi- dies, and a sale of rights to create the external- ity, and merger followed by a cap and trade market.

� Whether to develop and license or wait and imitate is an organizational form decision about the protection afforded by patents, the relative impor- tance of proprietary know-how, the availability of industry standards, technological lock-in, value- enhancing complements, and other sources of increasing returns.

640 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Exercises 1. If Apple iPod only played iTunes, and iTunes only could be heard on the Apple iPod, could Apple price the technologically integrated bundle any way they wanted? If other electronic music can play on an iPod, what determines whether there are any limitations on the bundled pricing of iPods and iTunes? What are those limitations?

2. An industry is composed of Firm 1, which controls 70 percent of the market, Firm 2 with 15 percent of the market, and Firm 3 with 5 percent of the market. About 20 firms of approximately equal size divide the remaining 10 percent of the market. Cal- culate the Herfindahl-Hirschman Index before and after the merger of Firm 2 and Firm 3 (assume that the combined market share after the merger is 20 percent). Would you view a merger of Firm 2 with Firm 3 as procompetitive or anticompeti- tive? Explain.

3. Suppose an industry is composed of eight firms with the following market shares:

A 30% E 8%

B 25 F 5

C 15 G 4

D 10 H 3

Based on the (revised 1997) merger guidelines, would the Antitrust Division likely challenge a proposed merger between a. Firms C and D (assume the combined market share is 25 percent)? b. Firms F and G (assume the combined market share is 9 percent)?

Explain your answer.

4. What are the incentives to innovate for a monopoly firm as compared with a firm in a competitive market if patent protection is not available?

5. Would you consider the fractional ownership jet taxi industry (NetJets, FlexJets, etc.) to be a contestable market? Why or why not?

6. The industry demand function for bulk plastics is represented by the following equation:

P = 800 − 20Q

where Q represents millions of pounds of plastic. The total cost function for the industry, exclusive of a required return on

invested capital, is

TC = 300 + 500Q + 10Q2

where Q represents millions of pounds of plastic. a. If this industry acts like a monopolist in the determination of price and out-

put, compute the profit-maximizing level of price and output. b. What are total profits at this price and output level? c. Assume that this industry is composed of many (500) small firms, such that

the demand function facing any individual firm is

P = $620

Compute the profit-maximizing level of price and output under these condi- tions (the industry’s total cost function remains unchanged).

d. What are total profits, given your answer to Part (c)?

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

Chapter 16: Government Regulation 641

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

e. Because of the risk of this industry, investors require a 15 percent rate of return on investment. Total industry investment amounts to $2 billion. If the monopoly solution prevails, as calculated in Parts (a) and (b), how would you describe the profits of the industry?

f. If the competitive solution most accurately describes the industry, is the in- dustry operating under equilibrium conditions? Why or why not? What would you expect to happen?

g. The Clean Water Coalition proposed pollution control standards for the in- dustry that would change the industry cost curve to the following:

TC = 400 + 560Q + 10Q2

What is the impact of this change on price, output, and total profits under the monopoly solution?

h. Assume these standards are being proposed only in the state of Texas, which has 50 of the 500 producers. What impact would you expect the new stan- dards to have on Texas firms? The rest of the industry?

7. Discuss the problems of aircraft noise around an airport from an externality per- spective and propose a possible solution if (a) housing existed in the airport area before the airport was built and (b) housing was built adjacent to the airport after the airport was built.

8. A sheep rancher leased the mineral rights beneath her grazing land to an oil com- pany. She fears that discharges from the oil wells will pollute her underground water resources. Consequently, the contract for the sale of mineral rights requires that the rancher and the oil company reach a mutually agreeable solution to the water contamination problem should it occur. If this bargaining fails to reach a conclusion acceptable to both sides, the mineral rights lease will be terminated automatically, and the rancher will be required to return a portion of the lease proceeds to the oil company. The portion that must be returned to the oil com- pany is to be determined through a process of binding arbitration. Discuss likely outcomes should this problem arise.

9. Branding Iron Products, a specialty steel fabricator, operates a plant in the town of West Star, Texas. The town has grown rapidly because of recent discoveries of oil and gas in the area. Many of the new residents have expressed concern at the amount of pollution (primarily particulate matter in the air and waste water in the town’s river) emitted by Branding Iron. Three proposals have been made to rem- edy the problem: a. Impose a tax on the amount of particulate matter and the amount of waste

water emitted by the firm. b. Prohibit pollution by the firm. c. Offer tax incentives to the firm to clean up its production processes.

Evaluate each of these alternatives from the perspectives of economic efficiency, equity, and the likely long-term impact on the firm.

10. An industry produces its product, Scruffs, at a constant marginal cost of $50. The market demand for Scruffs is equal to

Q = 75,000 − 600P

642 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

a. What is the value to a monopolist who is able to develop a patented process for producing Scruffs at a cost of only $45?

b. If the industry producing Scruffs is purely competitive, what is the maximum benefit that an inventor of a process that will reduce the cost of producing Scruffs by $5 per unit can expect to receive by licensing her invention to the firms in the industry?

11. If the decision to develop and license or wait and imitate in Table 16.7 is a simultaneous-play repeated game between Lucent and Motorola for each new generation of technology, what happens if the Motorola payoff in the southwest cell is positive $2 billion? How should Motorola “play” in this modified licensing game? How should Lucent play?

Case Exercises MICROSOFT TYING ARRANGEMENTS

1. Which of the following is a violation of the antitrust laws in the United States and why? (a) Microsoft monopolizes the market in PC operating systems with a 92 percent market share; (b) Microsoft attempts to monopolize the market in In- ternet portals with a pattern of anticompetitive tactics (tying arrangements, refu- sals to deal, etc.); (c) Microsoft sells Windows plus Microsoft Internet Explorer for less than Windows without Internet Explorer installed as the default browser; (d) Microsoft gives Internet Explorer away free to individual adopters with vari- able cost estimated at $0.0067; (e) Microsoft threatens to delicense Compaq and Dell, who would then be unable to preinstall Windows on PCs they ship unless Compaq and Dell exclude Netscape’s Internet browser from the user interface.

2. What difference does it make to the tying arrangement issues if Internet Explorer is a functionally integrated component of Windows? What if it’s more like a radio in an automobile than a steering post interlock device?

3. How should Microsoft market long-distance telephone services in the new wire- less telecommunications devices that also include Internet portals?

MUSIC RECORDING INDUSTRY BLOCKED FROM CONSOLIDATING Given the following market share distributions, the U.S. Antitrust Division blocked a merger of BMG and EMI in 2001, and the European Commission blocked a merger of Time Warner’s music division and EMI in 2000. Analyze these decisions and present arguments both pro and con.

U.S. MARKET SHARES WORLD MARKET SHARE

Vivendi Universal 20% Vivendi Universal 21%

Sony 20% Sony 19%

BMG 15% EMI 13%

Time Warner 13% Time Warner 12%

EMI 11% BMG 12%

What else should be involved in such merger policies?

Chapter 16: Government Regulation 643

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

17 CHAP T E R

Long-Term Investment Analysis CHAPTER PREVIEW Investment analysis (capital budgeting) is the process of planning for the purchases of assets whose returns (cash flows) are expected to continue beyond one year. When making capital budgeting decisions, the firm’s managers commit resources to the expansion of its productive capacity, improvement of its cost efficiency, or diversification of its asset base. Each of these decisions affects future cash flows the firm expects to generate and the risk of those cash flows. Capital expenditures are a bridge between short-term price and output decisions and longer-term strategic decisions that wealth- maximizing managers must make to remain competitive. Public sector and not- for-profit managers use cost-benefit analysis and cost-effectiveness analysis when considering many such long-term resource allocation decisions. These techniques also are presented in this chapter.

MANAGERIAL CHALLENGE Multigenerational Effects of Ozone Depletion and Greenhouse Gases1

The long-term effects of ozone depletion from hydro- chlorofluorocarbon (HCFC) emissions and of CO2 and other greenhouse gases from the burning of fossil fuels are controversial. Environmental scientists insist that the release of HCFCs opened a gaping hole in the ozone shield that provides protection from the sun’s ultravio- let (UV) rays. More recently, some scientists have argued that the increasing concentration of greenhouse gases has raised global temperatures. What is less controversial is that these environmental events have massive consequences for human health and wealth. Increasing incidence of skin cancers, melting polar icecaps, and rising sea levels imply tangible losses of catastrophic proportions—perhaps many billions of dollars annually. Some of these losses are immediate, but others are perhaps 100 years off. Benefit-cost analy- sis normally considers projects no more than 20 to 30 years long and employs discount rates of 2 to 7 percent. How should one discount such an uncertain and distant

future as is involved in ozone depletion and greenhouse gases?

Assuming a constant discount rate equal to the rate of return on long-term government bonds (5.3 percent), the discount factor that should be applied to find the present value of projected benefits or losses avoided in year 100 would be (1/1.053100) = .00517 or $5,716,930 present value per billion dollars of future losses avoided in year 100. Notice what happens, how- ever, if uncertainty about the appropriate discount rate varies from 2 to 8 percent. The discount factor at 100 years would then vary from (1/1.02100) = .13803 or $138,032,967 per billion dollars for 2 percent, to as little as (1/1.08100) = .000454, or $454,595 per billion dollars for 8 percent.

The possibility of lower discount rates implies that more than $138 million dollars should be spent today to avoid the projected $1 billion in delayed damage 100 years from now! Of course, the higher 8 percent rate

Cont.

644

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

implies spending less than one-half a million to avoid the $100 billion future loss. This range of present value estimates from $138 million to $454,595 million is be- yond what any analyst can work with in doing sensitiv- ity analysis. What should a benefit-cost analyst conclude? And what should businesses whose cash flows depend on “fun in the sun” recreation like golf courses and theme parks or physical assets built near sea level like the New York and Shanghai business districts and beachside hotels conclude about the justifiable capital expenditure to slow down or reverse global warming?

One insight is that it’s not the 5 percent average dis- count rate between 2 and 8 percent that matters in such circumstances, just as the average depth of a pool does not determine the hazard to someone who cannot swim. Instead, the lowest applicable discount rate largely de- termines the present value of very distant cash flows because higher discount rate scenarios like 7 and 8 per- cent inevitably sum to essentially zero at 100 years.

Martin Weitzman has calculated what discount rate is implied by assuming the discount rate begins at 4 percent and then follows a random path with equi- probable higher and lower rates and a standard devia- tion of 3 percent. The results led him to recommend a sliding scale discount rate of 2 percent for 25- to 75-year and 1 percent for 76- to 300-year cash flows. The fact that lower discount rates have such dramatically non- symmetrical effects on present value when very distant benefits are involved means that the present value of reducing CO2 emissions and other greenhouse gases may be much higher than previously thought.

As decisionmakersworldwide consider the options for reducing greenhouse gases (CO2, methane, nitrous oxide, etc.), one glaring fact surfaces quickly. The U.S. Energy Information Administration estimates that 41 percent of all the CO2 emissions we generate in the United States come from electricity generation, and 83 percent of that comes from burning coal (see Figure 17.1). So, 34 percent (0.41 × 0.83) of all U.S. CO2 emissions comes from coal- fired power plants. Not surprisingly, much interest in pre- venting global warming has therefore centered on alterna- tive sources of electrical energy generation. The popular inequality RE < C refers to sources of renewable energy (RE) for powering our households and workplaces that might be less expensive than coal (C).

What exactly are the alternative sources of electrical energy? Wind power, solar power, biomass, hydro, biofuel, geothermal, and ocean tidal power all contain

some attractive common features for the United States. They are renewable resources in abundant local supply with low carbon footprints. At present, only 5 percent of electrical generation comes from all these alternative energy sources rather than coal (83 percent), natural gas (15 percent), or negligible fuel oil (again see Figure 17.1). Natural gas is as cheap today as coal per British thermal unit (BTU), but its carbon footprint (though 80 percent cleaner than coal) is still quite substantial. Nuclear power is nonrenewable because of the inordinate half-life of nuclear fission waste. So, coal remains the electric utilities’ preferred fuel. In part, this is because we Americans are at present spending almost $1 billion a day in net wealth transfer for foreign crude oil, and coal is our most plentiful energy resource. Indeed, at present rates of consump- tion, the United States has more proven reserves of coal (90 years’ worth) than Saudi Arabia has proven reserves of oil (87 years).

But one ton of coal generates a megawatt of electric- ity plus one ton of CO2 byproduct. The marginal cost for delivered coal has varied from $45 to $85 per ton in 2007–2010 and the electricity is worth from $0.06/kWh in Washington State to $0.12/kWh in New York State.

MANAGERIAL CHALLENGE Continued

© M al co lm

Fi fe /G et ty Im ag es

Chapter 17: Long-Term Investment Analysis 645

Cont.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Taking the median cost of coal ($65), which also hap- pens to be the 2010 cost, and the weighted average value of electricity across the United States ($0.10), the elec- trical generation industry pays $65 per ton for coal on the spot market when additional energy is demanded and produces a $100 megawatt of electricity. The prob- lem is that when one includes the market cost of a ton of CO2 byproduct, using a long-term average price of the CO2 emission trading contract in the European Un- ion (EU) of $30, a typical public utility in the United States must recover all its capital cost from the remain- ing tiny $5 operating profit ($100 − $65 − $30 = $5) per megawatt hour. Even a giant 500-megawatt U.S. power plant would therefore project under a cap-and-trade carbon emissions trading regime cash flow of only $2,500 per hour for perhaps 20 hours a day or $50,000 per day to recover capital equipment costs and earn a profit. That sums to $17.6 million per year, insufficient to recover the $700 million cost of a 500-megawatt coal- fired power plant plus smokestack scrubbers.

Discussion Questions

� At a 5 percent discount rate, what annual cash flow is needed to make the 500-megawatt plant profitable?

� What is the effect on your analysis of CO2 trading credits varying in price from $16 to $39 in recent years?

� Do you think the collapse of the Copenhagen conference on carbon emissions reduction hastens or delays the time when coal-fired power plants will become unprofitable? What would happen before that occurred?

1Based on Richard Newell and William Pizer, “Discounting the Benefits of Climate Change Mitigation,” Resources for the Future (December 2001); Martin Weitzman, “Gamma Discounting,” American Economic Review (March 2001); and Frederick Harris, “Alternative Energy: A Symposium,” Wake Forest University, September 19, 2008.

FIGURE 17.1 Greenhouse Gas Emissions in the United States

Energy-related carbon dioxide 5,735.5 81.3%

Other carbon dioxide 103.8 1.5%

High-GWP gases 175.6 2.5%

Nitrous oxide 300.3 4.3%

Methane 737.4 10.5%

Coal 83%

Natural gas 15%

Fuel oil 2%

By gas, 2008 (million metric tons)

Total = 7,052.6

CO2 emissions from electricity generation by fuel type, 2008

Source: U.S. Energy Information Administration, Greenhouse Gases in the United States (December 2009).

MANAGERIAL CHALLENGE Continued

646 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

THE NATURE OF CAPITAL EXPENDITURE DECISIONS Previous chapters were mostly concerned with analytical tools and decision models that help managers make the most efficient use of existing resources. This chapter considers decisions to replace or expand the firm’s capital investment outlays. Capital outlays, by definition, have a long-range impact by determining products that will be produced, markets to be entered, the location of plants and facilities, and the type of technology (with its associated costs) to be used. Capital expenditures require careful analysis be- cause they are costly to make and often more costly to reverse.

A capital expenditure is a cash outlay that is expected to generate a flow of future cash benefits lasting longer than one year. It is distinguished from a normal operating expenditure, which is expected to result in cash benefits during the coming one-year pe- riod. Capital budgeting is the process of planning for and evaluating capital expendi- tures. In addition to asset replacement and expansion decisions, other types of decisions that can be analyzed using capital budgeting techniques include research and develop- ment expenditures, investments in employee education and training, lease-versus-buy de- cisions, and mergers and acquisitions.

A BASIC FRAMEWORK FOR CAPITAL BUDGETING This basic capital budgeting decision-making framework is illustrated in Figure 17.2. Suppose a company has nine investment projects under consideration, labeled A, B, C, . . . , I. The model assumes that all projects have the same risk. This schedule of pro- jects is often called the investment opportunity curve. For example, Project A requires an investment of $2 million and is expected to generate a 24 percent rate of return. Project B will cost $1 million ($3 million minus $2 million on the horizontal axis) and generate a 22 percent rate of return, and so on. Graphically, the projects are arranged in descend- ing order by their rates of return, indicating that no firm has a limitless number of pos- sible new products that all generate high rates of return. As new products are produced, new markets entered, and cost-saving technologies adopted, the number of highly profit- able investment opportunities tends to decline.

The marginal cost of capital curve represents the marginal cost of capital to the firm; that is, the cost of each additional dollar of investment capital raised in the capital mar- kets. Using this basic model, the firm should undertake Projects A, B, C, D, and E because their returns exceed the firm’s marginal cost of capital.

THE CAPITAL BUDGETING PROCESS The process of selecting capital investment projects consists of the following important steps:

1. Generate alternative capital investment project proposals. 2. Estimate cash flows for the project proposals. 3. Evaluate and choose investment projects to implement. 4. Review the investment projects after they have been implemented to assure

assumptions were accurate. Otherwise, one should modify assumptions as required for similar future projects.

capital expenditure A cash outlay designed to generate a flow of future cash benefits over a period of time extending beyond one year.

capital budgeting The process of planning for and evaluating capital expenditures.

Chapter 17: Long-Term Investment Analysis 647

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Generating Capital Investment Projects Ideas for new capital investments can come from many sources both inside and outside the firm. Proposals can originate at all levels in the organization, from factory workers all the way up to the board of directors. Most medium- and large-sized firms have depart- ments whose responsibilities include searching for and analyzing capital expenditure pro- jects. These departments include cost accounting, industrial engineering, marketing research, research and development, and corporate planning.

FIGURE 17.2 A Simplified Capital Budgeting Model

1

Investment (millions of dollars)

A 24

20

16

12

8

4

0 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

B

C

D

E F

G

H I

Marginal cost of capital curve

Investment opportunity curve

R at

e of

r et

ur n

(p er

ce nt

)

Example Capital Expenditures at Chrysler: The Grand Cherokee2

Chrysler Corporation developed its Grand Cherokee sport utility vehicle using an unusual (for Chrysler) “platform team” approach. Seven- to eight-hundred engineer- ing, marketing research, and design personnel brought the vehicle to market more quickly and at a lower cost than had been typical for other American auto compa- nies. The Grand Cherokee was developed and a new plant built in Detroit to produce it, all for about $1.1 billion in capital outlays. Chrysler planned to sell up to 175,000 of these vehicles each year at a price point of $32,000, realizing a profit of $5,500 per unit. For almost two decades, these cash flow projections have been realized.

The Grand Cherokee capital expenditures contain elements of both demand management and cost reduction. In addition, the decision to build the Grand Cherokee in an efficient new plant in Detroit rather than in its older underutilized plant in Toledo reflected a commitment to hold costs of production at a minimum. Chrysler saw sales of its older, smaller Jeep Cherokee decline from a peak of nearly 200,000 units per year to about 125,000. Chrysler’s intention was to cut the price of

(Continued)

648 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Estimating Cash Flows Certain basic guidelines are helpful when estimating cash flows. First, cash flows should be measured on an incremental basis. In other words, the cash-flow stream for the proj- ect should represent the difference between the cash-flow streams to the firm with and without acceptance of the investment project. Second, cash flows should be measured on an after-tax basis, using the firm’s marginal tax rate. Third, all the indirect effects of the project throughout the firm should be included in the cash-flow calculations. If a depart- ment or division of the firm is contemplating a capital investment that will alter the rev- enues or costs of other departments or divisions, then these external effects should be incorporated into the cash-flow estimates. Fourth, sunk costs should not be considered when evaluating the project. A sunk cost is an outlay that has been made (or committed to be made). Because sunk costs cannot be avoided, they should not be considered in the decision to accept or reject a project. Fifth, the value of resources used in the project should be measured in terms of their opportunity costs. Recall from Chapter 8 that op- portunity cost is the value of a resource in its next best alternative use.

For a typical investment project, an initial investment is made in year 0, which gen- erates a series of yearly net cash flows over the life of the project (n). The net investment (NINV) of a project is defined as the initial net cash outlay in year 0. It includes the ac- quisition cost of any new assets plus installation and shipping costs and tax effects.3

The incremental, after-tax net cash flows (NCFs) of a particular investment project are equal to cash inflows minus cash outflows. For any year during the life of the project, these may be defined as the difference in net income after tax (ΔNIAT) with and without the project plus the difference in depreciation (ΔD):

NCF = ΔNIAT + ΔD [17.1]

ΔNIAT is equal to the difference in net income before tax (ΔNIBT) times (1 – T), where T is the corporate (marginal) income tax rate:

ΔNIAT = ΔNIBTð1 − TÞ [17.2] ΔNIBT is defined as the difference in revenues (ΔR) minus the differences in operating costs (ΔC) and depreciation (ΔD):

ΔNIBT = ΔR − ΔC − ΔD [17.3]

its old Jeep Cherokee and market it aggressively as a low-cost, sporty utility vehicle alternative. Thus, Chrysler created for itself the option either to retain the older model if it sold well at a lower price, or to phase out the older model and close the Toledo plant. The evaluation process of this and many other major capital ex- penditure projects contains elements of cost reduction, demand management, and the creation of embedded real options.

2Based on “Iacocca’s Last Stand at Chrysler,” Fortune (April 20, 1992), p. 63ff.

3When the new asset is replacing an existing asset, one must also include in the net investment calculation the net proceeds from the sale of the existing asset and the taxes associated with its sale. See R. Charles Moyer, James R. McGuigan, and William J. Kretlow, Contemporary Financial Management, 11th ed. (Cincinnati: South-Western, 2010), pp. 306–308, for a discussion of the cash flow calculations for replacement decisions.

Chapter 17: Long-Term Investment Analysis 649

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Substituting Equation 17.3 into Equation 17.2 yields

ΔNIAT = ðΔR − ΔC − ΔDÞð1 − TÞ [17.4] Substituting this equation into Equation 17.1 yields the following definition of net cash flow:

NCF = ðΔR − ΔC − ΔDÞð1 − TÞ + ΔD [17.5]

Evaluating and Choosing the Investment Projects to Implement Once a capital expenditure project has been identified and the cash flows have been esti- mated, a decision to accept or reject the project is required. Acceptance of the project results in a cash-flow stream to the firm—that is, a series of either cash inflows or

Example Cash-Flow Estimation: Hamilton Beach/ Proctor-Silex, Inc. To illustrate the cash-flow calculations, consider the following example. Suppose that Hamilton Beach/Proctor-Silex, a manufacturer of small electric appliances, has been offered a contract to supply a regional merchandising company with a line of food blenders to be sold under the retail company’s private brand name. Hamilton Beach/Proctor-Silex’s treasurer estimates that the initial investment in new equipment required to produce the blenders would be $1 million. The equip- ment would be depreciated (using the straight-line method)4 over five years with a zero (0) estimated salvage value at the end of the five-year contract period. Based on the contract specifications, the treasurer estimates that incremental revenues (additional sales) would be $800,000 per year. The incremental costs if the contract is accepted will be $450,000 per year. These include cash outlays for direct labor and materials, transportation, utilities, building rent, and additional overhead. The firm’s marginal income tax rate is 40 percent.

Based on the information, NINV and NCF can be calculated for the project. The net investment (NINV) is equal to the $1 million initial outlay for the new equip- ment. The difference in revenues (ΔR) with and without the project is equal to $800,000 per year and the difference in operating costs (ΔC) is equal to $450,000 per year. The difference in depreciation (ΔD) is equal to the initial outlay ($1 mil- lion) divided by 5, or $200,000 per year. Substituting these values, along with t = 0.40, into Equation 17.5 yields the following:

NCF = ð$800,000 − $450,000 − $200,000Þð1 − 0:40Þ + $200,000 = $290,000

Hamilton Beach/Proctor-Silex must decide whether it wants to invest $1 million now to receive $290,000 per year in net cash flows over the next five years. The next section illustrates two of the criteria used in evaluating investment proposals.

4This depreciation method is just one of several possible methods that can be used. See Moyer, McGuigan, and Kre- tlow, op. cit., pp. 319–322, for a discussion of the various depreciation methods.

650 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

outflows for a number of years into the future. Typically, a project will result in an initial (first-year) outflow (investment) followed by a series of cash inflows (returns) over a number of succeeding years.

Various criteria are used to assess the desirability of investment projects. This section focuses on two widely used discounted cash-flow methods:5

• Internal rate of return (r) • Net present value (NPV)

Internal Rate of Return The internal rate of return (IRR) is defined as the discount rate that equates the present value of the net cash flows from the project with the net investment. The following equation is used to find the internal rate of return:

∑ n

t=1

NCFt ð1 + rÞt = NINV [17.6]

where n is the life of the investment and r is the internal rate of return. An investment project should be accepted if the internal rate of return is greater than

or equal to the firm’s required rate of return (cost of capital)—if not, the project should be rejected.

Example Calculation of Internal Rate of Return: Hamilton Beach/Proctor-Silex The internal rate of return for the Hamilton Beach/Proctor-Silex investment proj- ect is calculated as follows:

∑ 5

t=1

290,000

ð1 + rÞt = 1,000,000

∑ 5

t=1

1

ð1 + rÞt = 1,000,000 290,000

= 3:4483

The term ½∑5t=11=ð1+rÞt � represents the present value of a $1 annuity for five years discounted at r percent and is equal to 3.4483. The value 3.4483 in the Period 05 row of Table 5 in Appendix B falls between 3.5172 and 3.4331, which corre- sponds to discount rates of 13 and 14 percent, respectively. Interpolating between these values yields an internal rate of return of

r = 0:13 + 3:5172 − 3:4483 3:5172 − 3:4331

ð0:14 − 0:13Þ = 0:1382

or 13.8 percent. If Hamilton Beach/Proctor-Silex requires a rate of return of 12 percent on pro-

jects of this type, then the project should be accepted because the expected return (13.8 percent) exceeds the required return (12 percent). Later in this chapter we consider how to determine the required return (i.e., the firm’s cost of capital).

5For those not familiar with discounting (present value) techniques, Appendix A at the end of this book pro- vides a review of these concepts.

internal rate of return (IRR) The discount rate that equates the present value of the stream of net cash flows from a project with the project’s net investment.

Chapter 17: Long-Term Investment Analysis 651

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Net Present Value The net present value (NPV) of an investment is defined as the present value, discounted at the firm’s required rate of return (cost of capital), of the stream of net cash flows from the project minus the project’s net investment. Algebrai- cally, the net present value is equal to

NPV = ∑ n

t=1

NCFt ð1 + kÞt − NINV [17.7]

where n is the expected life of the project and k is the firm’s required rate of return (cost of capital).

An investment project should be accepted if the net present value is greater than or equal to zero and rejected if its net present value is less than zero. This is so because a positive net present value translates directly into increases in stock prices and increases in shareholder wealth.

Example Net Present Value Calculation: Hershey Foods Hershey Foods is considering an investment in a new wrapping machine for its Kiss candies. The machine has an initial cost (net investment) of $2.5 million. It is expected to produce cost savings from reduced labor and to generate additional revenues because of its increased reliability and productivity. Over its anticipated economic life of five years, the new Kiss wrapping machine is expected to generate the following stream of net cash flows (NCFt).

YEAR ( t) NET CASH FLOW (NCFt)

1 $600,000

2 800,000

3 800,000

4 600,000

5 250,000

If Hershey requires a return (k) of 15 percent on a project of this type, should it make the investment?

Hershey can solve this problem by computing the net present value of the cash flows from this project (using Equation 17.7) as follows:

YEAR CASH FLOW PRESENT VALUE INTEREST

FACTOR AT 15%* PRESENT VALUE (1) (2) (3) (4) = (2) × (3)

0 ($2,500,000) 1.00000 ($2,500,000)

1 600,000 0.86957 521,742

2 800,000 0.75614 604,912

3 800,000 0.65752 526,016

4 600,000 0.57175 343,050

5 250,000 0.49718 124,295

($379,985)

*Appendix B, Table 4.

Because this project has a negative net present value, it does not contribute to the goal of maximizing shareholder wealth. Therefore, it should be rejected.

net present value (NPV) The present value of the stream of net cash flows resulting from a project, discounted at the required rate of return (cost of capital), minus the project’s net investment.

652 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Net Present Value versus Internal Rate of Return Both the net present value and the internal rate of return methods result in identical decisions to either accept or reject individual projects. Net present value is greater than (less than) zero if and only if the internal rate of return is greater than (less than) the required rate of return k. In the case of mutually exclusive projects—that is, projects where the acceptance of one al- ternative precludes the acceptance of one or more other alternatives—the two methods may yield contradictory results; one project may have a higher internal rate of return than another and, at the same time, a lower net present value.

Consider, for example, mutually exclusive projects X and Y shown in Table 17.1. Both require a net investment of $1,000. Based on the internal rate of return, Project X is pre- ferred, with a rate of 21.5 percent compared with Project Y’s rate of 18.3 percent. Based on the net present value with a discount rate of 5 percent, Project Y ($270) is preferred to Project X ($240). Thus, it is necessary to determine which of the two criteria is the correct one to use in this situation. The outcome depends on what assumptions the deci- sion maker chooses to make about the implied reinvestment rate for the net cash flows generated from each project. The net present value method assumes that cash flows are reinvested at the firm’s cost of capital, whereas the internal rate of return method assumes that these cash flows are reinvested at the computed internal rate of return. Generally, the cost of capital is considered to be a more realistic reinvestment rate than the computed internal rate of return because it is the rate the next (marginal) investment project can be assumed to earn. In Figure 17.1, the last project invested in, Project E, offers a rate of return nearly equal to the firm’s marginal cost of capital. Consequently, the net present value approach is normally superior to the internal rate of return when choosing among mutually exclusive investments. Table 17.2 summarizes the two techniques.

ESTIMATING THE FIRM’S COST OF CAPITAL The cost of capital is concerned with what a firm has to pay for the capital (i.e., the debt, preferred stock, retained earnings, and common stock) it uses to finance new investments. It also can be thought of as the rate of return required by investors in the firm’s securities. As such, the firm’s cost of capital is determined in the capital markets and is closely related to the degree of risk associated with new investments, existing assets, and the firm’s capital structure. In general, the greater the riskiness of a firm as perceived by investors, the greater the return investors will require and the greater will be the cost of capital.

The cost of capital also can be thought of as the minimum rate of return required on new investments. If a new investment earns an internal rate of return that is greater than the cost of capital, the value of the firm increases. Correspondingly, if a new investment earns a return less than the firm’s cost of capital, the firm’s value decreases.

TABLE 17.1 NET PRESENT VALUE VERSUS INTERNAL RATE OF RETURN FOR MUTUALLY

EXCLUSIVE INVESTMENT PROJECTS

PROJECT X PROJECT Y

Net investment $1,000 $1,000

Net cash flows

Year 1 667 0

Year 2 667 1,400

Net present value at 5% $240 $270

Internal rate of return 21.5% 18.3%

cost of capital The cost of funds that are supplied to a firm. The cost of capital is the minimum rate of return that must be earned on new investments undertaken by a firm.

Chapter 17: Long-Term Investment Analysis 653

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The following discussion focuses on the two major sources of funds for most firms: debt and common equity. Each of these sources of funds has a cost.

Cost of Debt Capital The pre-tax cost of debt capital to the firm is the rate of return required by investors. For a debt issue, this rate of return kd equates the present value of all expected future receipts— interest I and principal repayment M—with the offering price V0 of the debt security:

V0 = ∑ n

t=1

I

ð1 + kdÞt +

M ð1 + kdÞn [17.8]

The cost of debt kd can be found by using the methods for finding the discount rate (that is, yield to maturity) discussed in Appendix A.

Most new long-term debt (in the form of bonds) issued by companies is sold at or close to par value (normally $1,000 per bond), and the coupon interest rate is set at the rate required by investors. When debt is issued at par value, the pre-tax cost of debt, kd, is equal to the coupon interest rate. Interest payments made to investors, however, are deductible from the firm’s taxable income. Therefore, the after-tax cost of debt is com- puted by multiplying the pre-tax cost by 1 minus the firm’s marginal tax rate T:

ki = kdð1 − TÞ [17.9]

TABLE 17.2 SUMMARY OF THE CAPITAL BUDGETING DECISION CRITERIA

CRITERION PROJECT ACCEPTANCE

DECISION RULE BENEFITS WEAKNESSES

Net present value (NPV) method

Accept project if project has a positive or zero NPV; that is, if the present value of net cash flows, evaluated at the firm’s cost of capital, equals or exceeds the net investment required

Considers the timing of cash flows

Provides an objective, return- based criterion for acceptance or rejection; most conceptually accurate approach

Difficulty in interpreting the meaning of the NPV computation

Internal rate of return (IRR) method

Accept project if IRR equals or exceeds the firm’s cost of capital

Easy to interpret the meaning of IRR

Considers the timing of cash flows

Provides an objective, return- based criterion for acceptance or rejection

Sometimes gives decision that conflicts with NPV; multiple rates of return problem*

*See Moyer, McGuigan, and Kretlow, op. cit., p. 334, for a discussion of the multiple internal rates of return problem.

Example Cost of Debt Capital: AT&T To illustrate the cost of debt computation, suppose that AT&T sells $100 million of 8.5 percent first-mortgage bonds at par. Assuming a corporate marginal tax rate of 40 percent, the after-tax cost of debt is computed as

ki = kdð1 − TÞ = 8:5ð1 − 0:40Þ = 5:1%

654 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Cost of Internal Equity Capital Like the cost of debt capital, the cost of equity capital to the firm is the equilibrium rate of return required by the firm’s common stock investors.

Firms raise equity capital in two ways: (1) internally, through retained earnings, and (2) externally, through the sale of new common stock. The cost of internal equity to the firm is less than the cost of new common stock because the sale of new stock requires the payment of flotation costs.

The concept of the cost of internal equity (or simply equity, as it is commonly called) can be developed using several different approaches, including the dividend valuation model.

Dividend Valuation Model Recall from Chapter 1 that shareholder wealth was defined as the present value, discounted at the shareholder’s required rate of return ke, of the expected future returns generated by a firm (see Equation 1.1). For the typical firm, these future returns can take two forms: the payment of dividends to the share- holder or an increase in the market value of the firm’s stock (capital gain). For the share- holder who plans to hold the stock indefinitely, the value of the firm (shareholder wealth, according to the dividend valuation model) is

V0 = ∑ ∞

t=1

Dt ð1 + keÞt

[17.10]

where Dt is the dividend paid by the firm in period t. 6 If the shareholder chooses to sell

the stock after n years, his or her wealth (V0) is

V0 = ∑ n

t=1

Dt ð1 + keÞt

+ Vn

ð1 + keÞn [17.11]

where Vn is the market value of the shareholder’s holdings in period n. Equation 17.11 is identical to 17.10, because the value of the firm in period n is based on the future returns (dividends) of the firm in periods n + 1, n + 2, . . .7

If the dividends of the firm are expected to grow perpetually at a constant compound rate of g per year, then the value of the firm (Equation 17.10) can be expressed as

V0 = D1

ke − g [17.12]

where D1 is the dividend expected to be paid in period 1 and V0 is the market value of the firm. If D1 is the dividend per share (rather than total dividends) paid in period 1, then V0 represents the market price per share of common stock. Solving Equation 17.12 for ke yields

ke = D1 V0

+ g [17.13]

The following example illustrates how Equation 17.13 can be applied in estimating the cost of equity capital.

dividend valuation model A model (or formula) stating that the value of a firm (i.e., shareholder wealth) is equal to the present value of the firm’s future dividend payments, discounted at the shareholder’s required rate of return. It provides one method of estimating a firm’s cost of equity capital.

6A profitable firm that reinvests all its earnings and never distributes any dividends would still have a positive value to stockholders, because its market value would be increasing, and shareholders could sell their stock and obtain a capital gain on their investment in the firm. 7The value of the firm in period n is

Vn = ∑ ∞

t=n+1

Dt ð1 + keÞt−n

When this expression is substituted in Equation 17.11, Equation 17.10 is obtained.

Chapter 17: Long-Term Investment Analysis 655

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Cost of External Equity Capital The cost of external equity is greater than the cost of internal equity for the following reasons:

• Flotation costs associated with new shares are usually high enough that they cannot realistically be ignored.

• The selling price of the new shares to the public must be less than the market price of the stock before announcement of the new issue, or the shares may not sell. Before any announcement, the current market price of a stock usually represents an equilibrium between supply and demand. If supply is increased (all other things being equal), the new equilibrium price will be lower.

When a firm’s future dividend payments are expected to grow forever at a constant per-period rate of g, the cost of external equity k0e is defined as

k0e = D1 Vnet

+ g [17.14]

where Vnet is the net proceeds to the firm on a per-share basis.

Example Cost of Internal Equity Capital: Fresno Company Suppose the current price of the common stock of Fresno Company (V0) is $32. The dividend per share of the firm next year, D1, is expected to be $2.14. Dividends have been growing at an average compound annual rate of 7 percent over the past 10 years, and this growth rate is expected to be maintained for the foreseeable fu- ture. Based on this information, the cost of equity capital is estimated as

ke = 2:14 32

+ 0:07 = 0:137

or 13.7 percent.8

8Another technique that can be used to estimate the cost of equity capital is the capital asset pricing model. See Moyer, McGuigan, and Kretlow, op. cit., Chapters 6 and 12, for a more detailed discussion of the CAPM theory and its use in calculating the cost of equity capital.

Example Cost of External Equity Capital: Fresno Company To illustrate, consider the Fresno Company example used in the cost of internal equity discussion, where V0 = $32, D1 = $2.14, g = 0.07, and ke = 13.7 percent. Assuming that new common stock can be sold at $31 to net the company $30 a share after flotation costs, k0e is calculated using Equation 17.14 as follows:

k0e = 2:14 30

+ 0:07

= 0:141 or 14:1%

Because of the relatively high cost of newly issued equity, many companies try to avoid this means of raising capital. The question of whether a firm should raise capital with newly issued common stock depends on the firm’s investment opportunities.

656 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Weighted Cost of Capital Firms calculate their cost of capital to determine a discount rate to use when evaluating proposed capital expenditure projects. Recall that the purpose of capital expenditure analysis is to determine which proposed projects the firm should actually undertake. Therefore, it is logical that the capital whose cost is measured and compared with the ex- pected benefits from these proposed projects should be the next or marginal capital the firm raises. Typically, companies estimate the cost of each capital component as the cost they expect to have to pay on these funds during the coming year.9

In addition, as a firm evaluates proposed capital expenditure projects, it normally does not specify the proportions of debt and equity financing for each individual project. Instead, each project is presumed to be financed with the same proportion of debt and equity contained in the company’s target capital structure.

Thus, the appropriate cost of capital figure to be used in capital budgeting is not only based on the next capital to be raised but also weighted by the proportions of the capital components in the firm’s long-range target capital structure. This figure is called the weighted, or overall, cost of capital.

The general expression for calculating the weighted cost of capital ka is as follows:

ka =

" equity fraction of capital structure

#" cost of

equity

# +

" debt

fraction of capital structure

#" cost of debt

#

= E

D + E

� � ðkeÞ + DD + E

� � ðkiÞ [17.15]

where D is the amount of debt and E the amount of equity in the target capital structure.10

Example Weighted Cost of Capital: Columbia Gas Company To illustrate, suppose that Columbia Gas has a current (and target) capital structure of 75 percent equity and 25 percent debt. (The proportions of debt and equity should be the proportions in which the firm intends to raise funds in the future.) For a firm that is not planning a change in its target capital structure, these proportions should be based on the current market value weights of the individual components (debt and common equity). The company plans to finance next year’s budget with $75 million of retained earnings (ke = 12 percent) and $25 million of long-term debt (kd = 8 per- cent). Assume a 40 percent marginal tax rate. Using these figures, the weighted cost of capital being raised to finance next year’s capital budget is calculated using Equation 17.15 as

ka = 0:75 × 12:0 + 0:25 × 8:0 × ð1 − 0:40Þ = 10:2%

This discount rate should be used to evaluate projects of average risk.

9Stated another way, the cost of the capital acquired by the firm in earlier periods (the historical cost of capi- tal) is not used as the discount rate in determining next year’s capital expenditures. 10If the target capital structure contains preferred stock, a preferred stock term is added to Equation 17.15. In this case, Equation 17.15 becomes

ka = E

E + D + P

� � ðkeÞ + DE + D + P

� � ðkiÞ + PE + D + P

� � ðkpÞ

where P is the amount of preferred stock in the target capital structure and kp is the component cost of preferred stock.

Chapter 17: Long-Term Investment Analysis 657

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

COST-BENEFIT ANALYSIS The remainder of this chapter is devoted to some techniques of analysis that may be used to assist in public and not-for-profit sector resource allocation decisions. The pri- mary analytical model examined is cost-benefit analysis, although cost-effectiveness stud- ies are also discussed.

Cost-benefit analysis is used to evaluate programs and investments based on a com- parison of their benefits and costs. Cost-benefit analysis is the logical public sector coun- terpart to the capital budgeting techniques discussed earlier.

Accept-Reject Decisions Cost-benefit analysis may be used to determine whether a specific expenditure is eco- nomically justifiable. For instance, one might examine a program designed to eradicate tuberculosis considering the current costs of the disease that could be averted by a spe- cific expenditure of funds. Benefits (averted costs) may be divided into four categories:

1. Expenditures on medical care, including physician and nurse fees, drug costs, and hospital and equipment charges

2. Loss of gross earnings during the disease 3. Reduction in gross earnings after the disease because of decreased employment

opportunities resulting from the social stigma attached to the illness 4. The pain and discomfort associated with having the disease

Suppose a particular program to eradicate tuberculosis requires a one-time outlay of $250 million (Table 17.3). Assume that the total benefits (averted disease costs) of this one-year program are expected to accrue for a period of five years. If one accepts, for the moment, that an appropriate social discount rate is 15 percent, the program may be evaluated in the net present-value analysis framework developed in the capital budget- ing discussion. The decision rule is to accept the project if the (discounted) benefits are greater than or equal to the (discounted) costs. Because the program has a positive cal- culated net discounted benefit, in this case $81.83 million, it is an acceptable project.

Alternative decision-making criteria include the internal rate of return and the benefit-cost ratio. According to the internal rate of return criterion, a project is accept- able if the IRR is greater than or equal to the appropriate social discount rate. In the case

TABLE 17.3 NET COST-BENEFIT ANALYSIS

END OF YEAR

ACTUAL DOLLAR

BENEFIT (COST)

PRESENT VALUE INTEREST FACTOR

AT 15%*

DISCOUNTED BENEFITS AND

COSTS ($ MILLION)

(1) (2) (3) (4) = (2) × (3)

0 ($250) 1.000 ($250.00)

1 150 0.870 130.50

2 125 0.756 94.50

3 100 0.658 65.80

4 50 0.572 28.60

5 25 0.497 12.43

Net benefits = $81.83

*Appendix B, Table 4.

cost-benefit analysis A resource allocation model that can be used by public and not- for-profit sector organizations to evaluate programs or investments on the basis of the magnitude of the discounted benefits and costs.

social discount rate The discount rate to be used when evaluating benefits and costs from public sector investments.

658 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

of the tuberculosis eradication program, the IRR for the benefits and costs shown in Table 17.3 is 32.4 percent. Because this exceeds the social discount rate of 15 percent, the project is acceptable. According to the benefit-cost ratio criterion, a project is accept- able if the benefit-cost ratio is greater than or equal to 1.0, where the benefit-cost ratio is equal to the present value of the benefits (discounted at the social discount rate) divided by the present value of the costs (similarly discounted). For the tuberculosis eradication program, the benefit-cost ratio is equal to

Benefit-cost ratio = 130:50 + 94:50 + 65:80 + 28:60 + 12:43

250 = 1:33

Because this ratio exceeds 1.0, the project is acceptable according to this criterion. All three decision criteria give identical decisions to accept or reject individ- ual projects.

Program-Level Analysis In addition to being used to evaluate whether an entire program is economically jus- tifiable, cost-benefit analysis can also determine whether the size of an existing program should be increased (or reduced) and, if so, by what amount. This determi- nation may be made using traditional marginal analysis as developed earlier in the text.

Returning again to the tuberculosis control program, assume that, because of strong lobbying from the American Medical Association, a number of expenditure levels be- yond the originally proposed $250 million are being considered. Table 17.4 summarizes these proposed programs and their expected benefits. An analysis that looked at only one of these proposed program expenditure levels would have concluded that any of the pro- gram levels was worthwhile because each proposal generates positive expected net pro- gram benefits.

When analyzed as a group, however, these program levels clearly show a limit to the economically justifiable expenditure of funds for tuberculosis control. The required anal- ysis is summarized in Table 17.5. A level of expenditure of $300 million is best because it generates an additional (marginal) $164.17 million in benefits, but the marginal program cost (in comparison to the $250 million program level) is only $50 million. To increase the program to $350 million would be counterproductive, because only $44 million in benefits are generated for the additional $50 million outlay (marginal costs exceed mar- ginal benefits).

TABLE 17.4 SCHEDULE OF PROGRAM BENEFITS FOR VARIOUS COST

LEVELS

PROGRAM COST ($ MILLIONS)

DISCOUNTED PROGRAM BENEFITS

($ MILLIONS)

NET PROGRAM BENEFITS

($ MILLIONS)

$250 $331.83 $ 81.83

300 496.00 196.00

350 540.00 190.00

400 565.00 165.00

benefit-cost ratio The ratio of the present value of the benefits from a project or program (discounted at the social discount rate) to the present value of the costs (similarly discounted).

Chapter 17: Long-Term Investment Analysis 659

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

STEPS IN COST-BENEFIT ANALYSIS The general principles of cost-benefit analysis may be summarized by answering the fol- lowing set of questions:

1. What is the objective function to be maximized? 2. What are the constraints placed on the analysis? 3. What costs and benefits are to be included, and how may they be valued? 4. What investment evaluation criterion should be used? 5. What is the appropriate discount rate?

The decision-making process in cost-benefit analysis may be traced in the flowchart presentation of Figure 17.3. Program objectives are set by the public through their politi- cal representatives. Alternatives are enumerated, explored, and revised in the light of constraints that may be operative in the system. These alternatives are then compared by enumerating and evaluating program benefits and costs in a present-value framework. Discounted benefits are compared with discounted costs, and intangibles are considered so a recommendation may be made about the merits of one or more alternative programs.

OBJECTIVES AND CONSTRAINTS IN COST-BENEFIT ANALYSIS Cost-benefit analysis is merely a way of evaluating alternative choices in decision mak- ing. As such, we need to examine it in light of several criteria proposed by welfare econ- omists as important for evaluating alternatives. One such criterion is Pareto optimality. A change is said to be desirable or consistent with Pareto optimality if at least one per- son is made better off (in his or her own judgment) and no one is made worse off (in their own judgment).

Cost-benefit analysis is often tied to a weaker notion of social improvement, some- times called the Kaldor-Hicks criterion, or merely the notion of a “potential” Pareto im- provement. Under this criterion, a change (or an economic program) is desirable if it generates sufficient gains that when distributed, the gains are sufficient to make all peo- ple in the community at least as well off as they were before the change. The fact that gainers may not in fact compensate losers is not a matter of direct consideration in cost-benefit analysis, but the income distributional impacts of a program are an ex- tremely important side issue.

TABLE 17.5 MARGINAL ANALYSIS OF BENEFITS AND COSTS

PROGRAM COST ($ MILLIONS)

MARGINAL COST ($ MILLIONS)

DISCOUNTED MARGINAL BENEFITS ($ MILLIONS)

$ 0 — —

250 $250 $331.83

300 50 164.17

350 50 44.00

400 50 25.00

660 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

It is important to recognize that not all projects with benefits that exceed costs will necessarily be adopted because of the following:

1. Physical constraints. The type of program alternatives considered is ultimately limited by the currently available state of technology and by the production possibilities derived from the relationship between physical inputs and outputs. For example, it is not yet possible to prevent cancer; therefore, a major emphasis must be directed toward early detection and treatment.

FIGURE 17.3 Schematic of the Cost-Benefit Analysis Process

Enumerate alternatives

Enumerate benefits: 1. Primary 2. Secondary a. Technological b. Pecuniary 3. Intangibles

Evaluate alternatives

Determine program objectives

Recognize constraints

Reevaluate

Evaluate benefits (state in terms

of dollars)

Choose an appropriate

discount rate

Compare discounted benefits and costs for

all alternatives

Evaluate relative significance of

intangibles

Choose or recommend the best alternative

Congress

State legislatures City, other local governments

Enumerate costs: 1. Primary 2. Secondary a. Technological b. Pecuniary 3. Intangibles

Evaluate costs (state in terms

of dollars)

Chapter 17: Long-Term Investment Analysis 661

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

2. Legal constraints. Domestic as well as international laws relating to property rights, the right of eminent domain, due process, and constitutional limits on a particular agency’s activities, among others, must be considered.

3. Administrative constraints. Effective programs require that individuals are available or can be hired and trained to carry out the program objectives. Even the best-conceived program is worthless unless individuals with the proper mix of technical and administrative skills are available.

4. Distributional constraints. Programs affect different groups in different ways because gainers are rarely the same as losers. When distributional impacts are of concern, the objective of cost-benefit analysis might be presented in terms of maximizing total benefits less total costs, subject to the constraint that benefits-less-costs for a particular group reach a pre-specified level.

5. Political constraints. What may be optimal may not be feasible because of the slow- ness and inefficiency of the political process. Many times what is best is tempered by what is possible, given the existence of strong competing interest groups as well as an often cumbersome political mechanism.

6. Financial or budget constraints. More often than not, agencies work within the bounds of a predetermined budget. In other words, the objective function must be altered to the suboptimizing form of maximizing benefits given a fixed budget. Virtually all programs have some absolute financial ceiling above which the program may not be expanded, in spite of the magnitude of social benefits.

7. Social and religious constraints. It is futile to tell Indian Hindus to eat sacred cattle to solve their nutritional problems. This example is just one of the social and religious constraints that may limit the range of feasible program alternatives.

ANALYSIS AND VALUATION OF BENEFITS AND COSTS Cost-benefit analysis is quite similar to traditional private sector profit-and-loss account- ing. In the private sector, the firm is guided by the criterion that private revenues must equal or exceed private costs over the long run for the firm to survive. In contrast, in cost-benefit analysis, the economist asks whether society as a whole will be better off by the adoption or nonadoption of a specific project or by the acceptance of one project to the exclusion of alternatives.

Benefits may be measured by the market price of the outputs from a public program or by the price consumers would be willing to pay based on certain assumptions in the analysis. Such assumptions obviously matter, and the analyst should consider a range of plausible assumption values. Moreover, any assumptions to which the recommendations are sensitive must be reported to the decision maker. In general, sensitivity analysis should always accompany benefit-cost recommendations.

Direct Benefits Benefits and costs may be categorized in a number of ways. Primary or direct benefits of a project consist of the value of goods or services produced if the project is undertaken compared to conditions without the project. The primary benefit of an irrigation project is the value of the additional crops produced on the irrigated land less the cost of seeds, labor, and equipment required to produce the crops. The primary benefits attributable to a college education might be considered as the increase in gross earnings of the graduate over what would have been earned without a college degree.

662 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Direct Costs Direct or primary costs are generally easier to measure than direct benefits. They include the capital costs necessary to undertake the project, operating and maintenance costs in- curred over the life of the project, and personnel expenses. Remember that the costs be- ing measured are opportunity costs, or the social value forgone elsewhere when factors of production in an alternate area of activity are used. If a proposed project will draw 20 percent of the required labor from the ranks of the unemployed, the market cost (wage payments) of these workers’ services will overstate the true social cost. A similar conclusion applies for the use of idle land. With no alternative use, the opportunity cost of the use of this land is zero (for as long as no productive alternative uses exist), no matter what the government happens to actually pay its owner in compensation. Such compensa- tion to the owner only affects the distribution of the benefits derived from land usage.

Indirect Costs or Benefits and Intangibles In addition to the primary impacts of a project, government investment invariably cre- ates secondary or indirect effects. Secondary costs and benefits may be of two types: real effects and pecuniary effects. Real secondary benefits may include reductions in necessary outlays for other government projects, for example, when an early glaucoma-detection campaign reduces the number of people who go blind, thereby reducing future govern- ment disability transfer payments. Similarly, an irrigation dam may reduce flooding and create a recreational area. These secondary benefits should be counted in a cost-benefit study. The same argument applies in accounting for secondary costs. For example, the Wallisville Dam Project in Texas was alleged to cause in excess of $500,000 in damages annually to saltwater fishing because of its impacts on the tidal marshlands. This real secondary cost should have been counted in the cost-benefit analysis of the Wallisville project.

Pecuniary benefits should generally not be included in the enumeration of “countable” benefits in a study. They generally arise in the form of lower input costs or changes in land values resulting from a project. For example, an improved highway may lead to greater business volume and profitability of gas stations, souvenir shops, and restaurants along that road, as well as higher land values and consequently higher rents to the landlords. Many of these benefits are purely distributional in nature because some busi- ness will be drawn from firms along other roads once the new road is completed.

A final group of program benefits and costs is intangibles. For these recognizable im- pacts of a project, it is either extremely difficult or impossible to calculate a dollar value. Intangibles may include such notions as quality of life and aesthetic contributions (or detriments). Intangibles may be analyzed by making trade-offs against tangibles in such a manner that the cost of additional increments of intangible improvement, for instance, may be compared with the forgone tangible benefits of a project.

THE APPROPRIATE RATE OF DISCOUNT When the benefits or costs of a program extend beyond a one-year time limit, they must be discounted back to some common point in time for purposes of comparison. Most people prefer current consumption to future consumption, so the social discount rate is used to adjust for this preference.11 As we have seen in the Managerial Challenge for this chapter, the choice of the appropriate discount rate to evaluate public investments is crit- ical to the conclusions of any cost-benefit analysis. Projects that may appear to be

11A review of discounting and present-value concepts is provided in Appendix A.

Chapter 17: Long-Term Investment Analysis 663

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

justified at a low discount rate, say 5 percent, may seem to be a gross misallocation of resources at a higher rate, such as 15 percent. The choice of a discount rate is likely to have a profound impact on the type of projects to be accepted. A low rate favors invest- ments with long lives, most of which will be of the durable “bricks and mortar” variety, whereas a high rate favors those whose benefits become available soon after the initial investment.

The discount rate performs the function of allocating resources between the public and private sectors; and an appropriate discount rate should be chosen to properly indi- cate when resources should be transferred from one sector to another. Simply then, if resources can earn 10 percent in the private sector, they should not be transferred to the public sector unless they can earn something greater than 10 percent on the invested resources. The correct discount rate for the evaluation of a government project is the percentage rate of return that the resources utilized would otherwise provide in the private sector.

COST-EFFECTIVENESS ANALYSIS Although cost-benefit analysis may be applied in a wide range of areas, in many types of government activity it is simply not feasible because of the problems of measuring the value of program outputs. For instance, program analyses in the fields of national

Example Costs and Benefits of a Toyota Automobile Plant in Kentucky Toyota built an assembly plant near Lexington, Kentucky, that is able to produce 200,000 automobiles annually. To get Toyota to locate the plant in Kentucky, the state agreed to invest approximately $325 million over a 20-year period. These ex- penditures include the following:

• Land and site preparation $ 33 million • Local highway construction 47 million • Employee training center and education of workers 65 million • Education of Japanese workers and families 5 million • Interest on economic development bonds 167 million

The returns to the state over the 20-year period are estimated at $632 million in income, sales, and payroll taxes from Toyota, its suppliers, and related businesses.

These numbers yield an internal rate of return of 25 percent, according to a University of Kentucky research team. Because the state’s economic resources are limited, one must consider whether these resources should have been invested in other projects that would have generated even higher rates of return. However, as Brinton Milward, director of the university’s Center for Business and Economic Research, explains, “Could you have put these funds into improvements in educa- tion and transportation and come up with a better benefit-cost ratio? My guess is no. Manufacturing has a pretty high multiplier” (in terms of the repeated turnover of money in the form of jobs and sales).

664 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

security, health and safety, and income redistribution are more frequently conducted using the cost-effectiveness framework than the cost-benefit one. The cost-effectiveness analysis begins with the premise that some identified program outputs are useful, and it proceeds to explore (1) how these outputs may be most efficiently achieved or (2) what the costs are of achieving various levels of the pre-specified output.

Cost-effectiveness analysis is widely applied in Department of Defense program studies. The benefits of most defense activities may be thought of as providing levels of deterrence. For many years, for example, the strategic nuclear bomber force provided a virtually impossible-to-quantify level of deterrent benefit against first-strike nuclear attack.

Least-Cost Studies The most frequent type of cost-effectiveness analysis is least-cost studies. As might be ex- pected, the emphasis of these studies is to identify the least expensive way of generating some quantity of an output. For instance, a city might decide that it wishes to reduce by 20 percent the number of burglaries occurring each year within its jurisdictional limits. One approach could be to expand the size of the police force, increase the number of foot patrol officers, and increase the number of squad cars on the streets at any one time. Another possibility might be to require builders to install security bars on the win- dows of all new homes and to provide cash or tax incentives for current homeowners to improve their personal security systems. A third alternative might be a community drive supporting Operation Identification, where individuals place permanent identifying marks on their belongings to make “fencing” of this merchandise more difficult. Combi- nations of these programs are also possible. Each of these alternatives is evaluated in terms of the expenditure required to achieve the desired objective—a 20 percent reduc- tion in burglaries.

Objective-Level Studies A second type of cost-effectiveness analysis is objective-level studies. These studies at- tempt to estimate the costs of achieving several alternative performance levels of the same objective. This approach may be illustrated with the case of reducing automobile emission levels. Table 17.6 provides some hypothetical data relating to various emission-control standards. Although these estimates in Table 17.6 may be realistic for an internal combustion reciprocating engine, they may far overstate actual costs if alternative technology like a hybrid-electric engine were assumed.

TABLE 17.6 HYPOTHETICAL DATA RELATING TO THE COST OF

ACHIEVING VARIOUS LEVELS OF AUTO EMISSION REDUCTIONS

PERCENTAGE OF 2008 EMISSION LEVELS

COSTS ($ MILLIONS—INCLUDING FUEL CONSUMPTION, MORE

FREQUENT MAINTENANCE, AND ADDED NEW-CAR COSTS)

90 $ 200

70 250

40 500

20 2,500

10 7,500

5 38,000

1 140,000

cost-effectiveness analysis An analytical tool designed to assist public decision makers in their resource allocation decisions when benefits cannot be easily measured in dollar terms, but costs can be monetarily quantified.

Chapter 17: Long-Term Investment Analysis 665

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Table 17.6 does illustrate that as the level of objective achievement increases, the associated costs frequently increase at a much more rapid rate. This information may help the decision maker to reach more rational decisions. For example, the $2.5 billion expenditure needed to reduce emissions to 20 percent of their 2008 levels may be reasonable. What is less clear is whether an additional 19 percent (from 20 percent to 1 percent) emissions reduction is worth the required incremental expenditure of $137.5 billion ($140 billion less $2.5 billion).

SUMMARY

� A capital expenditure is a current outlay of funds that is expected to provide a flow of future cash benefits.

� The capital expenditure decision process should consist of the following steps: generating alterna- tive investment proposals, estimating cash flows, evaluating and choosing the projects to undertake, and reviewing the projects after implementation.

� The internal rate of return (IRR) is the discount rate that equates the present value of the net cash flows from the project with the net investment. An investment project should be accepted (rejected) if its internal rate of return is greater than or equal to (less than) the firm’s required rate of return (i.e., cost of capital).

� The net present value (NPV) of an investment is the present value of the net cash flows from the project, discounted at the firm’s required rate of return (i.e., cost of capital), minus the project’s net investment. An investment project should be accepted (rejected) if its net present value is greater than or equal to (less than) zero.

� The cost of capital is the cost of funds that are supplied to the firm. It is influenced by the riski- ness of the firm, both in terms of its capital struc- ture and its investment strategy.

� The after-tax cost of debt (issued at par) is equal to the coupon rate multiplied by 1 minus the firm’s marginal tax rate.

� The cost of equity can be estimated using a number of different approaches, including the dividend val- uation model and the capital asset pricing model.

� The weighted cost of capital is calculated by weighting the costs of specific sources of funds, such as debt and equity, by the proportions of

each of the capital components in the firm’s long- range target capital structure.

� Cost-benefit analysis is the public sector counter- part of capital budgeting techniques used for re- source allocation decisions.

� Cost-benefit analysis involves the following steps:

1. Determining the program objectives 2. Enumerating the alternative means of achieving

the objectives, subject to the legal, political, technological, budgetary, and other constraints that limit the scope of action

3. Evaluating all primary, secondary, and intangi- ble benefits and costs associated with each alternative

4. Discounting the benefits and costs using a social discount rate to arrive at an overall measure of the desirability of each alternative (e.g., benefit- cost ratio)

5. Choosing (or recommending) the best alterna- tive based on the overall measure of desirabil- ity, the relative magnitude of the unquantifiable intangibles, and sensitivity analysis.

� Because of the measurement problems arising from the intangible impacts and economic externalities of many public programs, cost-benefit analysis is most useful in comparing projects with similar ob- jectives and similar magnitudes of intangibles and externalities.

� In cases where it is not feasible to place dollar val- ues on final program outputs, cost-effectiveness analysis may be used. Cost-effectiveness analysis assumes a priori that the program objectives are worth achieving and focuses on the least-cost method of achieving them.

666 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Exercises 1. A firm has the opportunity to invest in a project having an initial outlay of $20,000. Net cash inflows (before depreciation and taxes) are expected to be $5,000 per year for five years. The firm uses the straight-line depreciation method with a zero salvage value and has a (marginal) income tax rate of 40 percent. The firm’s cost of capital is 12 percent. a. Compute the internal rate of return and the net present value. b. Should the firm accept or reject the project?

2. A machine that costs $12,000 is expected to operate for 10 years. The estimated salvage value at the end of 10 years is $0. The machine is expected to save the company $2,331 per year before taxes and depreciation. The company depreciates its assets on a straight-line basis and has a marginal tax rate of 40 percent. The firm’s cost of capital is 14 percent. Based on the internal rate of return criterion, should this machine be purchased?

3. A company is planning to invest $75,000 (before taxes) in a personnel training program. The $75,000 outlay will be charged off as an expense by the firm this year (year 0). The returns estimated from the program in the form of greater pro- ductivity and a reduction in employee turnover are as follows (on an after-tax basis):

Years 1–10: $7,500 per year Years 11–20: $22,500 per year

The company has estimated its cost of capital to be 15 percent. Assume that the entire $75,000 is paid at time zero (the beginning of the project). The marginal tax rate for the firm is 40 percent.

Based on the net present value criterion, should the firm undertake the training program?

4. Alliance Manufacturing Company is considering the purchase of a new automated drill press to replace an older one. The machine now in operation has a book value of zero and a salvage value of zero. However, it is in good working condi- tion with an expected life of 10 additional years. The new drill press is more effi- cient than the existing one and, if installed, will provide an estimated cost savings (in labor, materials, and maintenance) of $6,000 per year. The new machine costs $25,000 delivered and installed. It has an estimated useful life of 10 years, and a salvage value of $1,000 at the end of this period. The firm’s cost of capital is 14 percent and its marginal income tax rate is 40 percent. The firm uses the straight-line depreciation method. a. What is the net cash flow in year 0 (that is, initial outlay)? b. What are the net cash flows after taxes in each of the next 10 years? c. What is the net present value of the investment? d. Should Alliance replace its existing drill press?

5. The Charlotte Bobcats, a professional basketball team, has been offered the oppor- tunity to purchase the contract of an aging superstar basketball player from an- other team. The general manager of the Bobcats wants to analyze the offer as a capital budgeting problem. The Bobcats would have to pay the other team $800,000 to obtain the superstar. Being somewhat old, the basketball player is ex- pected to be able to play for only four more years. The general manager figures

Answers to the exercises in blue can be found in

Appendix D at the back of the book.

Chapter 17: Long-Term Investment Analysis 667

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

that attendance, and hence revenues, would increase substantially if the Bobcats obtained the superstar. He estimates that incremental returns (additional ticket revenues less the superstar’s salary) would be as follows over the four-year period:

YEAR INCREMENTAL RETURNS

1 $450,000

2 350,000

3 275,000

4 200,000

The general manager has been told by the owners of the team that any capital expenditures must yield at least 12 percent after taxes. The firm’s (marginal) in- come tax rate is 40 percent. Furthermore, a check of the tax regulations indicates that the team can depreciate the $800,000 initial expenditure over the four-year period. a. Calculate the internal rate of return and the net present value to determine

the desirability of this investment. b. Should the Bobcats sign the superstar?

6. Panhandle Industries, Inc. currently pays an annual common stock dividend of $2.20 per share. The company’s dividend has grown steadily over the past 10 years at 8 percent per year; this growth trend is expected to continue for the fore- seeable future. The company’s present dividend payout ratio, also expected to continue, is 40 percent. In addition, the stock presently sells at eight times current earnings—that is, its “multiple” is 8.

Calculate the company’s cost of equity capital using the dividend capitalization model approach.

7. The Gordon Company currently pays an annual common stock dividend of $4.00 per share. Its dividend payments have been growing at a steady rate of 6 percent per year, and this rate of growth is expected to continue for the foreseeable future. Gordon’s common stock is currently selling for $65.25 per share. The company can sell additional shares of common stock after flotation costs at a net price of $60.50 per share.

Based on the dividend capitalization model, determine the cost of a. Internal equity (retained earnings) b. External equity (new common stock)

8. The Williams Company has a present capital structure (that it considers optimal) consisting of 30 percent long-term debt and 70 percent common equity. The company plans to finance next year’s capital budget with additional long-term debt and retained earnings. New debt can be issued at a coupon interest rate of 10 percent. The cost of retained earnings (internal equity) is estimated at 15 per- cent. The company’s marginal tax rate is 40 percent.

Calculate the company’s weighted cost of capital for the coming year. 9. The state of Glottamora has $100 million remaining in its budget for the current

year. One alternative is to give Glottamorans a one-time tax rebate. Alternatively, two proposals have been made for state expenditures of these funds.

668 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The first proposed project is to invest in a new power plant, costing $100 mil- lion and having an expected useful life of 20 years. Projected benefits accruing from this project are as follows:

YEARS BENEFITS PER YEAR

($ MILLIONS)

1–5 $ 0

6–20 20

The second alternative is to undertake a job retraining program, also costing $100 million and generating the following benefits:

YEARS BENEFITS PER YEAR

($ MILLIONS)

1–5 $20

6–10 14

11–20 4

The state Power Department argues that a 5 percent discount factor should be used in evaluating the projects, because that is the government’s borrowing rate. The Human Resources Department suggests using a 12 percent rate, because that more nearly equals society’s true opportunity rate. a. What is implied by the various departments’ desires to use different discount

rates? b. Evaluate the projects using both the 5 percent and the 12 percent rates. c. What rate do you believe to be more appropriate? d. Make a choice between the projects and the tax-refund alternative. Why did

you choose the alternative you did?

10. The Department of Transportation wishes to choose between two alternative accident prevention programs. It has identified three benefits to be gained from such programs:

• Reduced property damage, both to the vehicles involved in an accident and to other property (e.g., real estate that may be damaged at the scene of an accident)

• Reduced injuries • Reduced fatalities

The department’s experts are willing to provide dollar estimates of property damage savings that are expected to accrue from any program, but they will only estimate the number of injuries and fatalities that may be averted.

The first program is relatively moderate in its costs and will be concentrated in a large city. It involves upgrading traffic signals, improving road markers, and repav- ing some potholed streets. Because of the concentration and value of property in the city, savings from reduced property damage are expected to be substantial. Like- wise, a moderate number of traffic-related deaths and injuries could be avoided.

The second program is more ambitious. It involves straightening long sections of dangerous rural roads and installing improved guardrails. Although the prop- erty damage savings are expected to be small in relation to total cost, the reduc- tion in traffic-related deaths and injuries should be substantial.

Chapter 17: Long-Term Investment Analysis 669

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The following table summarizes the expected costs and payoffs of the two programs:

YEAR 1 2 3 4 TOTAL

Alternative #1

Cost ($000) 200 200 100 50 550

Reduced property damage ($000) 50 100 250 100 500

Lives saved 60 40 35 25 160

Injuries prevented 500 425 300 150 1,375

Alternative #2

Cost ($000) 700 1,800 1,100 700 4,300

Reduced property damage ($000) 150 225 475 300 1,150

Lives saved 50 75 100 125 350

Injuries prevented 800 850 900 900 3,450

Assume that a 10 percent discount rate is appropriate for evaluating government programs: a. Calculate the net present costs of the two programs. b. Generate any other tables that you may find useful in choosing between the

programs. c. Can you arrive at any unambiguous choice between the two alternatives?

What factors are likely to weigh on the ultimate choice made?

Case Exercises COST-BENEFIT ANALYSIS12

The Michigan State Fairgrounds is centrally located in the Detroit Standard Metropol- itan Statistical Area (SMSA), which consists of Wayne, Oakland, and Macomb coun- ties. The population within the SMSA numbered 4,197,931 persons in 2000, more than 47 percent of the state’s total population. More than 59 percent and 75 percent of the state’s population reside within 60 and 100 miles, respectively, of the fair- grounds site. The site is located near an efficient freeway system that connects many areas of the state. The State Fairgrounds is operated by the Agriculture Department and is in a deplorable state of disrepair. Costs have exceeded revenues by a substantial margin every year in the recent past. A redevelopment program has been proposed for the fairgrounds that would serve several purposes:

1. Revitalization of the fairgrounds would prevent further economic deterioration of the existing facilities, increase attendance and consequently revenues, and perhaps make the fairgrounds an economically viable entity.

2. A further benefit to be realized would be an economic stimulus to the area re- sulting from increased employment from the initial construction program, as well as increased revenues realized from the additional business that the pro- posed new facilities would generate.

3. Finally, aesthetic value could be realized from the upgrading and redevelopment of what is currently a marginal area of the city.

12Adapted from an unpublished paper by Eric Hartshom of Wayne State University, “Cost-Benefit Analysis Concerning the Proposed Redevelopment Program for the Michigan State Fairgrounds.”

670 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The redevelopment program would consist of the overall rehabilitation of the grounds and buildings as well as the construction of several income-producing buildings, includ- ing a hotel and convention facility and a dog-racing track (providing dog racing is legal- ized in Michigan and the fairgrounds can obtain the necessary license). Either a new coliseum would be constructed or the present one redesigned and refurbished. The cost of the redevelopment program would be $20 million. Construction would take three years with 50 percent of the cost incurred in year 0, 30 percent in year 1, and 20 percent in year 2. The redevelopment program would require funding by the state and federal governments. The following estimated benefits would be derived from the project:

1. Initial construction benefits. Previous studies showed that 38 worker-years of em- ployment are derived from each $1 million in construction. Assuming an hourly rate of $6, 40 hours per week, and 50 weeks per year, and relating this calculation to the $20 million cost of the redevelopment program, results of $9,120,000 in economic benefit would be derived through increased employment. Like the con- struction costs, these benefits would be spread over three years ($4.560 million in year 0, $2.736 million in year 1, and $1.824 million in year 2).

2. Coliseum. An appropriate coliseum facility could generate an additional $500,000 annually (years 3–20) from shows and events not currently available in the Detroit area.

3. Increased state fair attendance. With improved facilities (such as those planned in the redevelopment program), annual attendance at the state fair is expected to in- crease from 700,000 presently to 1,000,000 people. Assuming present per capita expenditures ($3.33) at the Michigan State Fair, the increased attendance would result in an additional $1 million in revenue annually (years 3–20).

4. Convention and hotel facility. It is estimated that a 200-room hotel, convention, and dining facility located at the fairgrounds would generate nearly $1.5 million in additional revenue annually (years 3–20).

5. Dog-racing track. It is estimated that an average dog-racing facility will produce $1.5 million in revenue annually. However, it must be realized that dog racing is simi- lar to horse racing, and it is expected that a portion of the revenues generated by a dog-racing track would be realized owing to a transfer of funds from local horse- racing facilities. Because this transfer of funds should not be considered in the analysis, it would be assumed that one-third of the dog-racing revenues will result from the re- distribution of funds from local horse-racing tracks. Consequently, only $1 million in annual revenues (years 3–20) will be attributed to the proposed dog-racing track.

TYPE OF COST OR BENEFIT YEAR(S)

ANNUAL BENEFIT (+) OR COST (−) ($ MILLION)

Construction outlay 0 $−10.000

Construction outlay 1 −6.000

Construction outlay 2 −4.000

Increased employment 0 +4.560

Increased employment 1 +2.736

Increased employment 2 +1.824

Coliseum 3–20 +0.500

State Fair attendance 3–20 +1.000

Convention and hotel facility 3–20 +1.500

Dog-racing track 3–20 +1.000

Chapter 17: Long-Term Investment Analysis 671

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The costs and benefits of the proposed redevelopment are summarized in the table. Assume that a 10 percent interest rate is appropriate for discounting the costs and benefits of the proposed project.

Questions 1. Determine the benefit-cost ratio (defined as the ratio of discounted benefits to

costs) for the proposed fairgrounds development. 2. Based on this analysis, should the redevelopment program be undertaken? 3. List some of the secondary benefits and costs, as well as intangibles, associated

with the project. (In calculating the benefits of the fairgrounds redevelopment program, increased employment opportunities were included.)

4. What assumption about employment in the Detroit area must be made in associ- ating these benefits with the project?

5. Recalculate the benefit-cost ratio, assuming that these benefits are not included in the analysis. How does this result affect the desirability of the project? (In calcu- lating the benefits of the fairgrounds redevelopment project, it was assumed that $1.5 million in additional annual revenue would be generated from the conven- tion and hotel facility.)

6. What assumption is being made about the effects of this facility on other hotel and convention facilities? Is this assumption realistic?

7. Suppose that only $500,000 of the facility’s annual revenues can be attributed to “new” convention and hotel business. Recalculate the benefit-cost ratio under this assumption (also exclude employment benefits). How does this calculation affect the desirability of the project?

8. Suppose that the fairgrounds is unable to obtain a license to operate a dog-racing track. Assume that construction costs are reduced by 15 percent if a dog-racing track is not built. Recompute the benefit-cost ratio under this assumption (also exclude employment and convention facility benefits). How does this calculation affect the desirability of the proposed redevelopment project?

INDUSTRIAL DEVELOPMENT TAX RELIEF AND INCENTIVES Tax relief competition between states seeking high-paying industrial jobs threatens to overpay for any conceivable net benefits. In 1993, Alabama paid more than $300 million in highway, rail, sewer, and other infrastructure investments to obtain a $300 million Mercedes plant with 1,500 jobs. From 2006–2009, North Carolina built a new runway at the Triad Airport for $130 million and provided job training and tax breaks worth an- other $142.3 million to obtain a $300 million FedEx hub.

Questions 1. Assess the likely benefits of such a plant or hub and how one should go about

analyzing them. 2. What form might a report to the Industrial Development Commission take? Out-

line the requisite components.

672 Part 5: Organizational Architecture and Regulation

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

A APPENDIX

The Time Value of Money

INTRODUCTION Many economic decisions involve benefits and costs that are expected to occur at different future points in time. For example, the construction of a new office complex requires an im- mediate outlay of cash and results in a stream of expected cash inflows (benefits) over many future years. To determine if the expected future cash inflows are sufficient to justify the initial outlay, we must have a way to compare cash flows occurring at different points in time. Also, recall from Chapter 1 that the value of a firm is equal to the discounted (or present) value of all expected returns. These future returns are discounted at a rate of return that is consistent with the risk of the expected future returns. When future returns are more certain, the dis- count rate used is lower, resulting in a higher present value of the firm, all other things being equal. Conversely, when future returns are riskier or more uncertain, they are discounted at a higher rate, resulting in a lower present value of the firm, all other things being equal.

An explicit solution to the problem of comparing the benefits and costs of economic transactions that occur at different points in time requires answers to the following kinds of questions: Is $1 to be received one year from today worth less than $1 in hand today? If so, why is it worth less? How much less is it worth?

The answers to these questions depend on the alternative uses available for the dollar between today and one year from today. Suppose the dollar can be invested in a guaranteed savings account paying a 6 percent annual rate of return (interest rate). The $1 invested today will return $1(1.06) = $1.06 one year from today. To receive exactly $1 one year from today, only $1/(1.06) = $0.943 would have to be invested in the account today. Given the opportunity to invest at a 6 percent rate of return, we see that $1 to be received one year from today is indeed worth less than $1 in hand today, its worth being only $0.943. Thus, the existence of opportunities to invest the dollar at positive rates of return makes $1 to be received at any future point in time worth less than $1 in hand today.1 This is what is meant by the time value of money. The investor’s required rate of return is called the discount rate.

PRESENT VALUE OF A SINGLE PAYMENT We can generalize this result for any future series of cash flows and any interest rate. Assume that the opportunity exists to invest at a compound rate of r percent per annum. Then the present value (value today) of $1 to be received at the end of year n, discounted at r percent, is

PV0 = 1

ð1 + rÞn [A.1]

1In this analysis we are abstracting from price level considerations. Changes in the level of prices (the value of the dollar in terms of the quantity of goods and services it will buy) can also affect the worth of the dollar. In theory, future price increases (or decreases) that are anticipated by the market will be reflected in the interest rate.

A-1

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

The term 1/(1 + r)n is often called a Present Value Interest Factor, or PVIFr,n. Table B.4 in Appendix B contains PVIF values for various interest rates, r, and periods in the future, n.

Example Present Value If an opportunity exists to invest at a compound rate of return of 12 percent, then the present value of $1 to be received four years (n = 4) from today is

PV0 = 1

ð1 + :12Þ4 = PVIF12%,4ð Þ = $1ð0:6355Þ = $0:6355

As we see in Table A.1, investing $0.6355 today at an interest rate of 12 percent per annum will give $1 at the end of four years.

Alternatively, the PVIF factors from Table B.4 in Appendix B could be used to find the present value of $1 expected to be received in four years (n = 4), assuming an interest rate of 12 percent (r = 12%), as follows:

PV0 = $1ðPVIF12%,4Þ = $1ð0:63552Þ = $0:6355

TABLE A.1 PRESENT VALUE OF $1 TO BE RECEIVED

AT THE END OF FOUR YEARS

YEAR RETURN RECEIVED AT END OF YEAR

VALUE OF INVESTMENT AT END OF YEAR

0 (present) — $.6355 ← Initial amount invested

1 .6355(.12) = $.0762 .6355 + .0762 = .7117

2 .7117(.12) = .0854 .7117 + .0854 = .7971

3 .7971(.12) = .0957 .7971 + .0957 = .8928

4 .8928(.12) = .1072 .8928 + .1072 = 1.000 x

Example Present Value of a Deferred Bequest What is the present value of an expected bequest of $2 million to your university if the expected remaining life span of the donor is eight years and the university uses an interest rate of 9 percent to evaluate gifts of this type?

PV0 = $2,000,000ðPVIF9%,8Þ = $2,000,000ð0:50187Þ = $1,003,740

Your university would be indifferent between receiving $1,003,740 today or $2 million in eight years.

A-2 Appendix A: The Time Value of Money

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Solving for the Interest or Growth Rate PVIF can also be used to solve for interest rates. For example, suppose you wish to bor- row $5,000 today from an associate. The associate is willing to loan you the money if you promise to pay back $6,802 four years from today. The compound interest rate your associate is charging can be determined as follows:

PV0 = $6,802ðPVIFr,4Þ $5,000 = $6,802ðPVIFr,4Þ

PVIFr,4 = $5,000 $6,802

= 0:735

Reading across the Period 04 row in Table B.4, 0.735 (rounded to three places for simplicity) is found in the 8% column. Thus, the effective interest rate on the loan is 8 percent per year, compounded annually.

PRESENT VALUE OF A SERIES OF EQUAL PAYMENTS (ANNUITY) The present value of a series of equal $1 payments to be received at the end of each of the next n years (an annuity), discounted at a rate of r percent, is

PV0 = 1

ð1 + rÞ1 + 1

ð1 + rÞ2 + … + 1

ð1 + rÞn

PV0 = Xn t = 1

1

ð1 + rÞt [A.2]

Example Calculation of Earnings Growth Rates for Hanamaker Paper Another common application of the use of PVIF factors from Table B.4 is the calcu- lation of the compound rate of growth of an earnings or dividend stream. For exam- ple, Hanamaker Paper Company had earnings per share of $2.56 in 2001. Security analysts have forecasted 2006 earnings per share to be $6.37. What is the expected compound annual rate of growth in Hanamaker Paper Company’s earnings per share? We can use the PVIF factors from Table B.4 to solve this problem, as follows:

$2:56 = $6:37ðPVIFr,5Þ PVIFr,5 = 0:40188

Looking across the Period 05 row in Table B.4 we find a PVIF equal to 0.40188 under the 20% column. Thus, the compound annual growth rate of earnings for Hanamaker Paper Company is 20 percent. (Interpolation can be used for PVIF va- lues between the values found in the tables. In practice, financial calculators are normally used for these types of calculations.)

Appendix A: The Time Value of Money A-3

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

For example, the present value of $1 to be received at the end of each of the next four years, discounted at 12 percent, is

PV0 = X4 t = 1

1

ð1 + :12Þt

= 1

ð1 + :12Þ1 + 1

ð1 + :12Þ2 + 1

ð1 + :12Þ3 + 1

ð1 + :12Þ4 = 0:89286 + 0:79719 + 0:71178 + 0:63552 = $3:0374

As shown in Table A.2, investing $3.0374 today at 12 percent will return exactly $1 at the end of each of the next four years, with nothing remaining in the account at the end of the fourth year. Again, rather than perform the present value calculations (Equation A.2), we can use a table to look up the values we need. Table B.5 in Appendix B contains the present values at various interest rates of $1 to be received at the end of each year for various periods of time. The values in Table B.5 are called Present Value Interest Factors for Annuities, or PVIFAr,n, where r is the interest rate per period and n is the number of periods (normally years).

Using the PVIFA factors from Table B.5, the present value of an annuity (PVAN0) can be computed as

PVAN0 = PMTðPVIFAr,nÞ [A.3] where PMT = the annuity amount to be received each period.

TABLE A.2 PRESENT VALUE OF $1 TO BE RECEIVED AT THE END OF EACH

OF THE NEXT FOUR YEARS

YEAR

RETURN RECEIVED AT END OF YEAR

AMOUNT WITHDRAWN AT END OF YEAR

VALUE OF INVESTMENT AT END OF YEAR

0 (present) — — $3.0374 ← Initial amount invested

1 $3.0374(.12) = $.3645 $1.00 $3.0374 + .3645 – 1.00 = 2.4019

2 2.4019(.12) = .2882 1.00 2.4019 + .2882 – 1.00 = 1.6901

3 1.6901(.12) = .2028 1.00 1.6901 + .2028 – 1.00 = .8929

4 .8929(.12) = .1071 1.00 .8929 + .1071 – 1.00 = .0000

Example Present Value of an Annuity You have recently purchased the winning ticket in the Florida lottery and have won $30 million, to be paid in equal $3 million increments (PMT) at the end of each of the next ten years. What are your winnings worth to you today using an interest rate of 8 percent? The PVIFA factors from Table B.5 can be used to solve this problem as follows:

PVAN0 = $3,000,000ðPVIFA8%,10Þ = $3,000,000ð6:7101Þ = $20,130,300

Thus, your $30 million winnings are worth only $20,130,300 to you today.

A-4 Appendix A: The Time Value of Money

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Solving for the Interest Rate PVIFA factors also can be used to solve for the rate of return expected from an invest- ment. This rate of return is often referred to as the internal rate of return from an invest- ment. Suppose the Big Spring Tool Company purchases a machine for $100,000. This machine is expected to generate annual cash flows of $23,740 to the firm over the next five years. What is the expected rate of return from this investment?

Using Equation A.3 we can determine the expected rate of return in this example as follows:

PVAN0 = PMTðPVIFAr,5Þ $100,000 = $23,740ðPVIFAr,5Þ PVIFAr,5 = 4:2123

From the Period 05 row in Table B.5, we see that a PVIFA of 4.2123 occurs in the 6% column. Hence, this investment offers a 6 percent expected (internal) rate of return.

PRESENT VALUE OF A SERIES OF UNEQUAL PAYMENTS The present value of a series of unequal payments (PMTt, t = 1,… , n) to be received at the end of each of the next n years, discounted at a rate of r percent, is

PV0 = Xn t=1

PMTt ð1 + rÞt

= Xn t=1

PMTtðPVIFr,tÞ [A.4]

The PVIFr,t values are the interest factors from Table B.4. Thus, the present value of a series of unequal payments is equal to the sum of the present value of the individual payments.

Example Project Evaluation for Intel Intel Corporation is evaluating an investment in a new chip-manufacturing facility. The facility is expected to have a useful life of five years and yield the following cash flow stream after the initial investment outlay:

End of Year Cash Flow t PMTt

1 +$1,000,000

2 + 1,500,000

3 − 500,000

4 + 2,000,000

5 + 1,000,000

The negative cash flow in Year 3 arises because of the expected need to install pol- lution control equipment during that year. The present value of this series of

Appendix A: The Time Value of Money A-5

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

unequal payments can be computed using PVIF factors from Table B.4 and assum- ing a 10 percent interest (required) rate on the investment:

PV = $1,000,000ðPVIF10%,1Þ + $1,500,000ðPVIF10%,2Þ −$500,000ðPVIF10%,3Þ + $2,000,000ðPVIF10%,4Þ +$1,000,000ðPVIF10%,5Þ

= $1,000,000ð0:90909Þ + $1,500,000ð0:82645Þ −$500,000ð0:75131Þ + $2,000,000ð0:68301Þ +$1,000,000ð0:62092Þ

= $3,760,050

The present value of these cash flows ($3,760,050) should be compared to the re- quired initial cash outlay to determine whether to invest in the new manufacturing facility.

A-6 Appendix A: The Time Value of Money

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

B APPENDIX

Tables

TABLE B.1 VALUES OF THE STANDARD NORMAL DISTRIBUTION

FUNCTION*

Z 0 1 2 3 4 5 6 7 8 9

�3. .0013 .0010 .0007 .0005 .0003 .0002 .0002 .0001 .0001 .0000 �2.9 .0019 .0018 .0017 .0017 .0016 .0016 .0015 .0015 .0014 .0014 �2.8 .0026 .0025 .0024 .0023 .0023 .0022 .0021 .0021 .0020 .0019 �2.7 .0035 .0034 .0033 .0032 .0031 .0030 .0029 .0028 .0027 .0026 �2.6 .0047 .0045 .0044 .0043 .0041 .0040 .0039 .0038 .0037 .0036 �2.5 .0062 .0060 .0059 .0057 .0055 .0054 .0052 .0051 .0049 .0048 �2.4 .0082 .0080 .0078 .0075 .0073 .0071 .0069 .0068 .0066 .0064 �2.3 .0107 .0104 .0102 .0099 .0096 .0094 .0091 .0089 .0087 .0084 �2.2 .0139 .0136 .0132 .0129 .0126 .0122 .0119 .0116 .0113 .0110 �2.1 .0179 .0174 .0170 .0166 .0162 .0158 .0154 .0150 .0146 .0143 �2.0 .0228 .0222 .0217 .0212 .0207 .0202 .0197 .0192 .0188 .0183 �1.9 .0287 .0281 .0274 .0268 .0262 .0256 .0250 .0244 .0238 .0233 �1.8 .0359 .0352 .0344 .0336 .0329 .0322 .0314 .0307 .0300 .0294 �1.7 .0446 .0436 .0427 .0418 .0409 .0401 .0392 .0384 .0375 .0367 �1.6 .0548 .0537 .0526 .0516 .0505 .0495 .0485 .0475 .0465 .0455 �1.5 .0668 .0655 .0643 .0630 .0618 .0606 .0594 .0582 .0570 .0559 �1.4 .0808 .0793 .0778 .0764 .0749 .0735 .0722 .0708 .0694 .0681 �1.3 .0988 .0951 .0934 .0918 .0901 .0885 .0869 .0853 .0838 .0823 �1.2 .1151 .1131 .1112 .1093 .1075 .1056 .1038 .1020 .1003 .0985 �1.1 .1357 .1335 .1314 .1292 .1271 .1251 .1230 .1210 .1190 .1170 �1.0 .1587 .1562 .1539 .1515 .1492 .1469 .1446 .1423 .1401 .1379 � .9 .1841 .1814 .1788 .1762 .1736 .1711 .1685 .1660 .1635 .1611 � .8 .2119 .2090 .2061 .2033 .2005 .1977 .1949 .1922 .1894 .1867 � .7 .2420 .2389 .2358 .2327 .2297 .2266 .2236 .2206 .2177 .2148 � .6 .2743 .2709 .2676 .2643 .2611 .2578 .2546 .2514 .2483 .2451 � .5 .3085 .3050 .3015 .2981 .2946 .2912 .2877 .2843 .2810 .2776 � .4 .3446 .3409 .3372 .3336 .3300 .3264 .3228 .3192 .3156 .3121 � .3 .3821 .3783 .3745 .3707 .3669 .3632 .3594 .3557 .3520 .3483 � .2 .4207 .4168 .4129 .4090 .4052 .4013 .3974 .3936 .3897 .3859 � .1 .4602 .4562 .4522 .4483 .4443 .4404 .4364 .4325 .4286 .4247 � .0 .5000 .4960 .4920 .4880 .4840 .4801 .4761 .4721 .4681 .4641

*Note: Table values give the probability of a value occurring which is less than z standard deviations from the mean. Note 1: If a random variable X is not “standard,” its values must be “standardized”: z � (X � �)/�. That is:

Note 2: For z � �4, N(z ) � 0 to 4 decimal places; for z � 4, N(z ) � 1 to 4 decimal places.

P(X � x) � N ax � � � b

B-1

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE B.1 VALUES OF THE STANDARD NORMAL DISTRIBUTION

FUNCTION (CONTINUED)

Z 0 1 2 3 4 5 6 7 8 9

.0 .5000 .5040 .5080 .5120 .5160 .5199 .5239 .5279 .5319 .5359

.1 .5398 .5438 .5478 .5517 .5557 .5596 .5636 .5675 .5714 .5753

.2 .5793 .5832 .5871 .5910 .5948 .5987 .6026 .6064 .6103 .6141

.3 .6179 .6217 .6255 .6293 .6331 .6368 .6406 .6443 .6480 .6517

.4 .6554 .6591 .6628 .6664 .6700 .6736 .6772 .6808 .6844 .6879

.5 .6915 .6950 .6985 .7019 .7054 .7088 .7123 .7157 .7190 .7224

.6 .7257 .7291 .7324 .7357 .7389 .7422 .7454 .7486 .7517 .7549

.7 .7580 .7611 .7642 .7673 .7703 .7734 .7764 .7794 .7823 .7852

.8 .7881 .7910 .7939 .7967 .7995 .8023 .8051 .8078 .8106 .8133

.9 .8159 .8186 .8212 .8238 .8264 .8289 .8315 .8340 .8365 .8389 1.0 .8413 .8438 .8461 .8485 .8508 .8531 .8554 .8577 .8599 .8621 1.1 .8643 .8665 .8686 .8708 .8729 .8749 .8770 .8790 .8810 .8830 1.2 .8849 .8869 .8888 .8907 .8925 .8944 .8962 .8980 .8997 .9015 1.3 .9032 .9049 .9066 .9082 .9099 .9115 .9131 .9147 .9162 .9177 1.4 .9192 .9207 .9222 .9236 .9251 .9265 .9278 .9292 .9306 .9319 1.5 .9332 .9345 .9357 .9370 .9382 .9394 .9406 .9418 .9430 .9441 1.6 .9452 .9463 .9474 .9484 .9495 .9505 .9515 .9525 .9535 .9545 1.7 .9554 .9564 .9573 .9582 .9591 .9599 .9608 .9616 .9625 .9633 1.8 .9641 .9648 .9656 .9664 .9671 .9678 .9686 .9693 .9700 .9706 1.9 .9713 .9719 .9726 .9732 .9738 .9744 .9750 .9756 .9762 .9767 2.0 .9772 .9778 .9783 .9788 .9793 .9798 .9803 .9808 .9812 .9817 2.1 .9821 .9826 .9830 .9834 .9838 .9842 .9846 .9850 .9854 .9857 2.2 .9861 .9864 .9868 .9871 .9874 .9878 .9881 .9884 .9887 .9890 2.3 .9893 .9896 .9898 .9901 .9904 .9906 .9909 .9911 .9913 .9916 2.4 .9918 .9920 .9922 .9925 .9927 .9929 .9931 .9932 .9934 .9936 2.5 .9938 .9940 .9941 .9943 .9945 .9946 .9948 .9949 .9951 .9952 2.6 .9953 .9955 .9956 .9957 .9959 .9960 .9961 .9962 .9963 .9964 2.7 .9965 .9966 .9967 .9968 .9969 .9970 .9971 .9972 .9973 .9974 2.8 .9974 .9975 .9976 .9977 .9977 .9978 .9979 .9979 .9980 .9981 2.9 .9981 .9982 .9982 .9983 .9984 .9984 .9985 .9985 .9986 .9986 3. .9987 .9990 .9993 .9995 .9997 .9998 .9998 .9999 .9999 1.0000

Source: Statistical Analysis: With Business and Economic Applications, by Ya-lun Chou. Copyright © 1969 by Holt, Rinehart and Winston, Inc. Reprinted by permission of Holt, Rinehart and Winston, Inc.

B-2 Appendix B: Tables

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE B.2 TABLE OF “STUDENTS” DISTRIBUTION—VALUE OF t*

Degrees Probability of

Freedom 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.05 0.02 0.01 0.001

1 0.158 0.325 0.510 0.727 1.000 1.376 1.963 3.078 6.314 12.706 31.821 63.657 636.619 2 0.142 0.289 0.445 0.617 0.816 1.061 1.386 1.886 2.920 4.303 6.965 9.925 31.598 3 0.137 0.277 0.424 0.584 0.765 0.978 1.250 1.638 2.353 3.182 4.541 5.841 12.924 4 0.134 0.271 0.414 0.569 0.741 0.941 1.190 1.533 2.132 2.776 3.747 4.604 8.610 5 0.132 0.267 0.408 0.559 0.727 0.920 1.156 1.476 2.015 2.571 3.365 4.032 6.869

6 0.131 0.265 0.404 0.553 0.718 0.906 1.134 1.440 1.943 2.447 3.143 3.707 5.959 7 0.130 0.263 0.402 0.549 0.711 0.896 1.119 1.415 1.895 2.365 2.998 3.499 5.408 8 0.130 0.262 0.399 0.546 0.706 0.889 1.108 1.397 1.860 2.306 2.896 3.355 5.041 9 0.129 0.261 0.398 0.543 0.703 0.883 1.100 1.383 1.833 2.262 2.821 3.250 4.781

10 0.129 0.260 0.397 0.542 0.700 0.879 1.093 1.372 1.812 2.228 2.764 3.169 4.587

11 0.129 0.260 0.396 0.540 0.697 0.876 1.088 1.363 1.796 2.201 2.718 3.106 4.437 12 0.128 0.259 0.395 0.539 0.695 0.873 1.083 1.356 1.782 2.179 2.681 3.055 4.318 13 0.128 0.259 0.394 0.538 0.694 0.870 1.079 1.350 1.771 2.160 2.650 3.012 4.221 14 0.128 0.258 0.393 0.537 0.692 0.868 1.076 1.345 1.761 2.145 2.624 2.977 4.140 15 0.128 0.258 0.393 0.536 0.691 0.866 1.074 1.341 1.753 2.131 2.602 2.947 4.073

16 0.128 0.258 0.392 0.535 0.690 0.865 1.071 1.337 1.746 2.120 2.583 2.921 4.015 17 0.128 0.257 0.392 0.534 0.689 0.863 1.069 1.333 1.740 2.110 2.567 2.898 3.965 18 0.127 0.257 0.392 0.534 0.688 0.862 1.067 1.330 1.734 2.101 2.552 2.878 3.922 19 0.127 0.257 0.391 0.533 0.688 0.861 1.066 1.328 1.729 2.093 2.539 2.861 3.883 20 0.127 0.257 0.391 0.533 0.687 0.860 1.064 1.325 1.725 2.086 2.528 2.845 3.850

21 0.127 0.257 0.391 0.532 0.686 0.859 1.063 1.323 1.721 2.080 2.518 2.831 3.819 22 0.127 0.256 0.390 0.532 0.686 0.858 1.061 1.321 1.717 2.074 2.508 2.819 3.792 23 0.127 0.256 0.390 0.532 0.685 0.858 1.060 1.319 1.714 2.069 2.500 2.807 3.767 24 0.127 0.256 0.390 0.531 0.685 0.857 1.059 1.318 1.711 2.064 2.492 2.797 3.745 25 0.127 0.256 0.390 0.531 0.684 0.856 1.058 1.316 1.708 2.060 2.485 2.787 3.725

26 0.127 0.256 0.390 0.531 0.684 0.856 1.058 1.315 1.706 2.056 2.479 2.779 3.707 27 0.127 0.256 0.389 0.531 0.684 0.855 1.057 1.314 1.703 2.052 2.473 2.771 3.690 28 0.127 0.256 0.389 0.530 0.683 0.855 1.056 1.313 1.701 2.048 2.467 2.763 3.674 29 0.127 0.256 0.389 0.530 0.683 0.854 1.055 1.311 1.699 2.045 2.462 2.756 3.659 30 0.127 0.256 0.389 0.530 0.683 0.854 1.055 1.310 1.697 2.042 2.457 2.750 3.646

40 0.126 0.255 0.388 0.529 0.681 0.851 1.050 1.303 1.684 2.021 2.423 2.704 3.551 60 0.126 0.254 0.387 0.527 0.679 0.848 1.046 1.296 1.671 2.000 2.390 2.660 3.460

120 0.126 0.254 0.386 0.526 0.677 0.845 1.041 1.289 1.658 1.980 2.358 2.617 3.373 � 0.126 0.253 0.385 0.524 0.674 0.842 1.036 1.282 1.645 1.960 2.326 2.576 3.291

*Note: Probabilities given are for two-tailed tests. For example, a probability of 0.05 allows for 0.025 in one tail of the distribution and 0.025 in the other. Table 2 is taken from Table III of Fisher and Yates: Statistical Tables for Biological, Agricultural and Medical Research, published by Longman Group, Ltd., London (previously published by Oliver and Boyd, Edinburgh), and by permission of the authors and publishers.

Appendix B: Tables B-3

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

T A B L E

B .3

T H E

F -D

IS T R IB

U T IO

N —

U P P E R

5 P E R C E N T

B R E A K P O IN

T S

� 1

1 2

3 4

5 6

7 8

9 10

12 15

20 24

30 40

60 12

0 �

� 2 1

16 1.

4 19

9. 5

21 5.

7 22

4. 6

23 0.

2 23

4. 0

23 6.

8 23

8. 9

24 0.

5 24

1. 9

24 3.

9 24

5. 9

24 8.

0 24

9. 1

25 0.

1 25

1. 1

25 2.

2 25

3. 3

25 4.

3 2

18 .5

7 19

.0 0

19 .1

6 19

.2 5

19 .3

0 19

.3 3

19 .3

5 19

.3 7

19 .3

8 19

.4 0

19 .4

1 19

.4 3

19 .4

5 19

.4 5

19 .4

6 19

.4 7

19 .4

8 19

.4 9

19 .5

0 3

10 .1

3 9.

55 9.

28 9.

12 9.

01 8.

94 8.

89 8.

85 8.

81 8.

79 8.

74 8.

70 8.

66 8.

64 8.

62 8.

59 8.

57 8.

55 8.

53 4

7. 71

6. 94

6. 59

6. 39

6. 26

6. 16

6. 09

6. 04

6. 00

5. 96

5. 91

5. 86

5. 80

5. 77

5. 75

5. 72

5. 69

5. 66

5. 63

5 6.

61 5.

79 5.

41 5.

19 5.

05 4.

95 4.

88 4.

82 4.

77 4.

74 4.

68 4.

62 4.

56 4.

53 4.

50 4.

46 4.

43 4.

40 4.

36 6

5. 99

5. 14

4. 76

4. 53

4. 39

4. 28

4. 21

4. 15

4. 10

4. 06

4. 00

3. 94

3. 87

3. 84

3. 81

3. 77

3. 74

3. 70

3. 67

7 5.

59 4.

74 4.

35 4.

12 3.

97 3.

87 3.

79 3.

73 3.

68 3.

64 3.

57 3.

51 3.

44 3.

41 3.

38 3.

34 3.

30 3.

27 3.

23 8

5. 32

4. 46

4. 07

3. 84

3. 69

3. 58

3. 50

3. 44

3. 39

3. 35

3. 28

3. 22

3. 15

3. 12

3. 08

3. 04

3. 01

2. 97

2. 93

9 5.

12 4.

26 3.

86 3.

63 3.

48 3.

37 3.

29 3.

23 3.

18 3.

14 3.

07 3.

01 2.

94 2.

90 2.

86 2.

83 2.

79 2.

75 2.

71

10 4.

96 4.

10 3.

71 3.

48 3.

33 3.

22 3.

14 3.

07 3.

02 2.

98 2.

91 2.

85 2.

77 2.

74 2.

70 2.

66 2.

62 2.

58 2.

54 11

4. 84

3. 98

3. 59

3. 36

3. 20

3. 09

3. 01

2. 95

2. 90

2. 85

2. 79

2. 72

2. 65

2. 61

2. 57

2. 53

2. 49

2. 45

2. 40

12 4.

75 3.

89 3.

49 3.

26 3.

11 3.

00 2.

91 2.

85 2.

80 2.

75 2.

69 2.

62 2.

54 2.

51 2.

47 2.

43 2.

38 2.

34 2.

30 13

4. 67

3. 81

3. 41

3. 18

3. 03

2. 92

2. 83

2. 77

2. 71

2. 67

2. 60

2. 53

2. 46

2. 42

2. 38

2. 34

2. 30

2. 25

2. 21

14 4.

60 3.

74 3.

34 3.

11 2.

96 2.

85 2.

76 2.

70 2.

65 2.

60 2.

53 2.

46 2.

39 2.

35 2.

31 2.

27 2.

22 2.

18 2.

13

15 4.

54 3.

68 3.

29 3.

06 2.

90 2.

79 2.

71 2.

64 2.

59 2.

54 2.

48 2.

40 2.

33 2.

29 2.

25 2.

20 2.

16 2.

11 2.

07 16

4. 49

3. 63

3. 24

3. 01

2. 85

2. 74

2. 66

2. 59

2. 54

2. 49

2. 42

2. 35

2. 28

2. 24

2. 19

2. 15

2. 11

2. 06

2. 01

17 4.

45 3.

59 3.

20 2.

96 2.

81 2.

70 2.

61 2.

55 2.

49 2.

45 2.

38 2.

31 2.

23 2.

19 2.

15 2.

10 2.

06 2.

01 1.

96 18

4. 41

3. 55

3. 16

2. 93

2. 77

2. 66

2. 58

2. 51

2. 46

2. 41

2. 34

2. 27

2. 19

2. 15

2. 11

2. 06

2. 02

1. 97

1. 92

19 4.

38 3.

52 3.

13 2.

90 2.

74 2.

63 2.

54 2.

48 2.

42 2.

38 2.

31 2.

23 2.

16 2.

11 2.

07 2.

03 1.

98 1.

93 1.

88

20 4.

35 3.

49 3.

10 2.

87 2.

71 2.

60 2.

51 2.

45 2.

39 2.

35 2.

28 2.

20 2.

12 2.

08 2.

04 1.

99 1.

95 1.

90 1.

84 21

4. 32

3. 47

3. 07

2. 84

2. 68

2. 57

2. 49

2. 42

2. 37

2. 32

2. 25

2. 18

2. 10

2. 05

2. 01

1. 96

1. 92

1. 87

1. 81

22 4.

30 3.

44 3.

05 2.

82 2.

66 2.

55 2.

46 2.

40 2.

34 2.

30 2.

23 2.

15 2.

07 2.

03 1.

98 1.

94 1.

89 1.

84 1.

78 23

4. 28

3. 42

3. 03

2. 80

2. 64

2. 53

2. 44

2. 37

2. 32

2. 27

2. 20

2. 13

2. 05

2. 01

1. 96

1. 91

1. 86

1. 81

1. 76

24 4.

26 3.

40 3.

01 2.

78 2.

62 2.

51 2.

42 2.

36 2.

30 2.

25 2.

18 2.

11 2.

03 1.

98 1.

94 1.

89 1.

84 1.

79 1.

73

25 4.

24 3.

39 2.

99 2.

76 2.

60 2.

49 2.

40 2.

34 2.

28 2.

24 2.

16 2.

09 2.

01 1.

96 1.

92 1.

87 1.

82 1.

77 1.

71 26

4. 23

3. 37

2. 98

2. 74

2. 59

2. 47

2. 39

2. 32

2. 27

2. 22

2. 15

2. 07

1. 99

1. 95

1. 90

1. 85

1. 80

1. 75

1. 69

27 4.

21 3.

35 2.

96 2.

73 2.

57 2.

46 2.

37 2.

31 2.

25 2.

20 2.

13 2.

06 1.

97 1.

93 1.

88 1.

84 1.

79 1.

73 1.

67 28

4. 20

3. 34

2. 95

2. 71

2. 56

2. 45

2. 36

2. 29

2. 24

2. 19

2. 12

2. 04

1. 96

1. 91

1. 87

1. 82

1. 77

1. 71

1. 65

29 4.

18 3.

33 2.

93 2.

70 2.

55 2.

43 2.

35 2.

28 2.

22 2.

18 2.

10 2.

03 1.

94 1.

90 1.

85 1.

81 1.

75 1.

70 1.

64

30 4.

17 3.

32 2.

92 2.

69 2.

53 2.

42 2.

33 2.

27 2.

21 2.

16 2.

09 2.

01 1.

93 1.

89 1.

84 1.

79 1.

74 1.

68 1.

62 40

4. 08

3. 23

2. 84

2. 61

2. 45

2. 34

2. 25

2. 18

2. 12

2. 08

2. 00

1. 92

1. 84

1. 79

1. 74

1. 69

1. 64

1. 58

1. 51

60 4.

00 3.

15 2.

76 2.

53 2.

37 2.

25 2.

17 2.

10 2.

04 1.

99 1.

92 1.

84 1.

75 1.

70 1.

65 1.

59 1.

53 1.

47 1.

39 12

0 3.

92 3.

07 2.

68 2.

45 2.

29 2.

17 2.

09 2.

02 1.

96 1.

91 1.

83 1.

75 1.

66 1.

61 1.

55 1.

50 1.

43 1.

35 1.

25 �

3. 84

3. 00

2. 60

2. 37

2. 21

2. 10

2. 01

1. 94

1. 88

1. 83

1. 75

1. 67

1. 57

1. 52

1. 46

1. 39

1. 32

1. 22

1. 00

B-4 Appendix B: Tables

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

T A B L E

B .3

T H E

F -D

IS T R IB

U T IO

N —

U P P E R

5 P E R C E N T

P O IN

T S

(C O N T IN

U E D )

So ur

ce :E

.S . P

ea rs

on a

nd H

.O . H

ar tle

y, B

io m

et rik

a Ta

bl es

fo r S

ta tis

tic ia

ns ,V

ol . 1

, T ab

le 1

8 w

ith p

er m

iss io

n.

� 1

1 2

3 4

5 6

7 8

9 10

12 15

20 24

30 40

60 12

0 �

� 2 1

40 52

49 99

.5 54

03 56

25 57

64 58

59 59

28 59

82 60

22 60

56 61

06 61

57 62

09 62

35 62

61 62

87 63

13 63

39 63

66 2

98 .5

0 99

.0 0

99 .1

7 99

.2 5

99 .3

0 99

.3 3

99 .3

6 99

.3 7

99 .3

9 99

.4 0

99 .4

2 99

.4 3

99 .4

5 99

.4 6

99 .4

7 99

.4 7

99 .4

8 99

.4 9

99 .5

0 3

34 .1

2 30

.8 2

29 .4

6 28

.7 1

28 .2

4 27

.9 1

27 .6

7 27

.4 9

27 .3

5 27

.2 3

27 .0

5 26

.8 7

26 .6

9 26

.6 0

26 .5

0 26

.4 1

26 .3

2 26

.2 2

26 .1

3 4

21 .2

0 18

.0 0

16 .6

9 15

.9 8

15 .5

2 15

.2 1

14 .9

8 14

.8 0

14 .6

6 14

.5 5

14 .3

7 14

.2 0

14 .0

2 13

.9 3

13 .8

4 13

.7 5

13 .6

5 13

.5 6

13 .4

6

5 16

.2 6

13 .2

7 12

.0 6

11 .3

9 10

.9 7

10 .6

7 10

.4 6

10 .2

9 10

.1 6

10 .0

5 9.

89 9.

72 9.

55 9.

47 9.

38 9.

29 9.

20 9.

11 9.

02 6

13 .7

5 10

.9 2

9. 78

9. 15

8. 75

8. 47

8. 26

8. 10

7. 98

7. 87

7. 72

7. 56

7. 40

7. 31

7. 23

7. 14

7. 06

6. 97

6. 88

7 12

.2 5

9. 55

8. 45

7. 85

7. 46

7. 19

6. 99

6. 84

6. 72

6. 62

6. 47

6. 31

6. 16

6. 07

5. 99

5. 91

5. 82

5. 74

5. 65

8 11

.2 6

8. 65

7. 59

7. 01

6. 63

6. 37

6. 18

6. 03

5. 91

5. 81

5. 67

5. 52

5. 36

5. 28

5. 20

5. 12

5. 03

4. 95

4. 86

9 10

.5 6

8. 02

6. 99

6. 42

6. 06

5. 80

5. 61

5. 47

5. 35

5. 26

5. 11

4. 96

4. 81

4. 73

4. 65

4. 57

4. 48

4. 40

4. 31

10 10

.0 4

7. 56

6. 55

5. 99

5. 64

5. 39

5. 20

5. 06

4. 94

4. 85

4. 71

4. 56

4. 41

4. 33

4. 25

4. 17

4. 08

4. 00

3. 91

11 9.

65 7.

21 6.

22 5.

67 5.

32 5.

07 4.

89 4.

74 4.

63 4.

54 4.

40 4.

25 4.

10 4.

02 3.

94 3.

86 3.

78 3.

69 3.

60 12

9. 33

6. 93

5. 95

5. 41

5. 06

4. 82

4. 64

4. 50

4. 39

4. 30

4. 16

4. 01

3. 86

3. 78

3. 70

3. 62

3. 54

3. 45

3. 36

13 9.

07 6.

70 5.

74 5.

21 4.

86 4.

62 4.

44 4.

30 4.

19 4.

10 3.

96 3.

82 3.

66 3.

59 3.

51 3.

43 3.

34 3.

25 3.

17 14

8. 86

6. 51

5. 56

5. 04

4. 69

4. 46

4. 28

4. 14

4. 03

3. 94

3. 80

3. 66

3. 51

3. 43

3. 35

3. 27

3. 18

3. 09

3. 00

15 8.

68 6.

36 5.

42 4.

89 4.

56 4.

32 4.

14 4.

00 3.

89 3.

80 3.

67 3.

52 3.

37 3.

29 3.

21 3.

13 3.

05 2.

96 2.

87 16

8. 53

6. 23

5. 29

4. 77

4. 44

4. 20

4. 03

3. 89

3. 78

3. 69

3. 55

3. 41

3. 26

3. 18

3. 10

3. 02

2. 93

2. 84

2. 75

17 8.

40 6.

11 5.

18 4.

67 4.

34 4.

10 3.

93 3.

79 3.

68 3.

59 3.

46 3.

31 3.

16 3.

08 3.

00 2.

92 2.

83 2.

75 2.

65 18

8. 29

6. 01

5. 09

4. 58

4. 25

4. 01

3. 84

3. 71

3. 60

3. 51

3. 37

3. 23

3. 08

3. 00

2. 92

2. 84

2. 75

2. 66

2. 57

19 8.

18 5.

93 5.

01 4.

50 4.

17 3.

94 3.

77 3.

63 3.

52 3.

43 3.

30 3.

15 3.

00 2.

92 2.

84 2.

76 2.

67 2.

58 2.

49

20 8.

10 5.

85 4.

94 4.

43 4.

10 3.

87 3.

70 3.

56 3.

46 3.

37 3.

23 3.

09 2.

94 2.

86 2.

78 2.

69 2.

61 2.

52 2.

42 21

8. 02

5. 78

4. 87

4. 37

4. 04

3. 81

3. 64

3. 51

3. 40

3. 31

3. 17

3. 03

2. 88

2. 80

2. 72

2. 64

2. 55

2. 46

2. 36

22 7.

95 5.

72 4.

82 4.

31 3.

99 3.

76 3.

59 3.

45 3.

35 3.

26 3.

12 2.

98 2.

83 2.

75 2.

67 2.

58 2.

50 2.

40 2.

31 23

7. 88

5. 66

4. 76

4. 26

3. 94

3. 71

3. 54

3. 41

3. 30

3. 21

3. 07

2. 93

2. 78

2. 70

2. 62

2. 54

2. 45

2. 35

2. 26

24 7.

82 5.

61 4.

72 4.

22 3.

90 3.

67 3.

50 3.

36 3.

26 3.

17 3.

03 2.

89 2.

74 2.

66 2.

58 2.

49 2.

40 2.

31 2.

21

25 7.

77 5.

57 4.

68 4.

18 3.

85 3.

63 3.

46 3.

32 3.

22 3.

13 2.

99 2.

85 2.

70 2.

62 2.

54 2.

45 2.

36 2.

27 2.

17 26

7. 72

5. 53

4. 64

4. 14

3. 82

3. 59

3. 42

3. 29

3. 18

3. 09

2. 96

2. 81

2. 66

2. 58

2. 50

2. 42

2. 33

2. 23

2. 13

27 7.

68 5.

49 4.

60 4.

11 3.

78 3.

56 3.

39 3.

26 3.

15 3.

06 2.

93 2.

78 2.

63 2.

55 2.

47 2.

38 2.

29 2.

20 2.

10 28

7. 64

5. 45

4. 57

4. 07

3. 75

3. 53

3. 36

3. 23

3. 12

3. 03

2. 90

2. 75

2. 60

2. 52

2. 44

2. 35

2. 26

2. 17

2. 06

29 7.

60 5.

42 4.

54 4.

04 3.

73 3.

50 3.

33 3.

20 3.

09 3.

00 2.

87 2.

73 2.

57 2.

49 2.

41 2.

33 2.

23 2.

14 2.

03

30 7.

56 5.

39 4.

51 4.

02 3.

70 3.

47 3.

30 3.

17 3.

07 2.

98 2.

84 2.

70 2.

55 2.

47 2.

39 2.

30 2.

21 2.

11 2.

01 40

7. 31

5. 18

4. 31

3. 83

3. 51

3. 29

3. 12

2. 99

2. 89

2. 80

2. 66

2. 52

2. 37

2. 29

2. 20

2. 11

2. 02

1. 92

1. 80

60 7.

08 4.

98 4.

13 3.

65 3.

34 3.

12 2.

95 2.

82 2.

72 2.

63 2.

50 2.

35 2.

20 2.

12 2.

03 1.

94 1.

84 1.

73 1.

60 12

0 6.

85 4.

79 3.

95 3.

48 3.

17 2.

96 2.

79 2.

66 2.

56 2.

47 2.

34 2.

19 2.

03 1.

95 1.

86 1.

76 1.

66 1.

53 1.

38 �

6. 63

4. 61

3. 78

3. 32

3. 02

2. 80

2. 64

2. 51

2. 41

2. 32

2. 18

2. 04

1. 88

1. 79

1. 70

1. 59

1. 47

1. 32

1. 00

Appendix B: Tables B-5

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE B.4 PRESENT VALUE OF $1 (PVIF )

Period 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% Period

01 .99010 .98039 .97007 .96154 .95233 .94340 .93458 .92593 .91743 .90909 01 02 .98030 .96117 .94260 .92456 .90703 .89000 .87344 .85734 .84168 .82645 02 03 .97059 .94232 .91514 .88900 .86384 .83962 .81639 .79383 .77228 .75131 03 04 .96098 .92385 .88849 .85480 .82270 .79209 .76290 .73503 .70883 .68301 04 05 .95147 .90573 .86261 .82193 .78353 .74726 .71299 .68058 .64993 .62092 05

06 .94204 .88797 .83748 .79031 .74622 .70496 .66634 .63017 .59627 .56447 06 07 .93272 .87056 .81309 .75992 .71063 .66506 .62275 .58349 .54705 .51316 07 08 .92348 .85349 .78941 .73069 .67684 .62741 .58201 .54027 .50189 .46651 08 09 .91434 .83675 .76642 .70259 .64461 .59190 .54393 .50025 .46043 .42410 09 10 .90529 .82035 .74409 .67556 .61391 .55839 .50835 .46319 .42241 .38554 10

11 .89632 .80426 .72242 .64958 .58468 .52679 .47509 .42888 .38753 .35049 11 12 .88745 .78849 .70138 .62460 .55684 .49697 .44401 .39711 .35553 .31683 12 13 .87866 .77303 .68095 .60057 .53032 .46884 .41496 .36770 .32618 .28966 13 14 .86996 .75787 .66112 .57747 .50507 .44230 .38782 .34046 .29925 .26333 14 15 .86135 .74301 .64186 .55526 .48102 .41726 .36245 .31524 .27454 .23939 15

16 .85282 .72845 .62317 .53391 .45811 .39365 .33873 .29189 .25187 .21763 16 17 .84436 .71416 .60502 .51337 .43630 .37136 .31657 .27027 .23107 .19784 17 18 .83602 .70016 .58739 .49363 .41552 .35034 .29586 .25025 .21199 .17986 18 19 .82774 .68643 .57029 .47464 .39573 .33051 .27651 .23171 .19449 .16354 19 20 .81954 .67297 .55367 .45639 .37689 .31180 .25842 .21455 .17843 .14864 20

21 .81143 .65978 .53755 .44883 .35894 .29415 .24151 .19866 .16370 .13513 21 22 .80340 .64684 .52189 .42195 .34185 .27750 .22571 .18394 .15018 .12285 22 23 .79544 .63414 .50669 .40573 .32557 .26180 .21095 .17031 .13778 .11168 23 24 .78757 .62172 .49193 .39012 .31007 .24698 .19715 .15770 .12640 .10153 24 25 .77977 .60953 .47760 .37512 .29530 .23300 .18425 .14602 .11597 .09230 25

B-6 Appendix B: Tables

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE B.4 PRESENT VALUE OF $1 (PVIF) (CONTINUED)

Period 11% 12% 13% 14% 15% 16% 17% 18% 19% 20% Period

01 .90090 .89286 .88496 .87719 .86957 .86207 .85470 .84746 .84043 .83333 01 02 .81162 .79719 .78315 .76947 .75614 .74316 .73051 .71818 .70616 .69444 02 03 .73119 .71178 .69305 .67497 .65752 .64066 .62437 .60863 .59342 .57870 03 04 .65873 .63552 .61332 .59208 .57175 .55229 .53365 .51579 .49867 .48225 04 05 .59345 .56743 .54276 .51937 .49718 .47611 .45611 .43711 .41905 .40188 05

06 .53464 .50663 .48032 .45559 .43233 .41044 .38984 .37043 .35214 .33490 06 07 .48166 .45235 .42506 .39964 .37594 .35383 .33320 .31392 .29592 .27908 07 08 .43393 .40388 .37616 .35056 .32690 .30503 .28478 .26604 .24867 .23257 08 09 .39092 .36061 .33288 .30751 .28426 .26295 .24340 .22546 .20897 .19381 09 10 .35218 .32197 .29459 .26974 .24718 .22668 .20804 .19106 .17560 .16151 10

11 .31728 .28748 .26070 .23662 .21494 .19542 .17781 .16192 .14756 .13459 11 12 .28584 .25667 .23071 .20756 .18691 .16846 .15197 .13722 .12400 .11216 12 13 .25751 .22917 .20416 .18207 .16253 .14523 .12989 .11629 .10420 .09346 13 14 .23199 .20462 .18068 .15971 .14133 .12520 .11102 .09855 .08757 .07789 14 15 .20900 .18270 .15989 .14010 .12289 .10793 .09489 .08352 .07359 .06491 15

16 .18829 .16312 .14150 .12289 .10686 .09304 .08110 .07073 .06184 .05409 16 17 .16963 .14564 .12522 .10780 .09293 .08021 .06932 .05998 .05196 .04507 17 18 .15282 .13004 .11081 .09456 .08080 .06914 .05925 .05083 .04367 .03756 18 19 .13768 .11611 .09806 .08295 .07026 .05961 .05064 .04308 .03669 .03130 19 20 .12403 .10367 .08678 .07276 .06110 .05139 .04328 .03651 .03084 .02608 20

21 .11174 .09256 .07680 .06383 .05313 .04430 .03699 .03094 .02591 .02174 21 22 .10067 .08264 .06796 .05599 .04620 .03819 .03162 .02622 .02178 .01811 22 23 .09069 .07379 .06014 .04911 .04017 .03292 .02702 .02222 .01830 .01509 23 24 .08170 .06588 .05322 .04308 .03493 .02838 .02310 .01883 .01538 .01258 24 25 .07361 .05882 .04710 .03779 .03038 .02447 .01974 .01596 .01292 .01048 25

Appendix B: Tables B-7

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE B.5 PRESENT VALUE OF AN ANNUITY OF $1 (PVIFA )

Period 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% Period

01 .9901 .9804 .9709 .9615 .9524 .9434 .9346 .9259 .9174 .9091 01 02 1.9704 1.9416 1.9135 1.8861 1.8594 1.8334 1.8080 1.7833 1.7591 1.7355 02 03 2.9410 2.8839 2.8286 2.7751 2.7233 2.6730 2.6243 2.5771 2.5313 2.4868 03 04 3.9020 3.8077 3.7171 3.6299 3.5459 3.4651 3.3872 3.3121 3.2397 3.1699 04 05 4.8535 4.7134 4.5797 4.4518 4.3295 4.2123 4.1002 3.9927 3.8896 3.7908 05

06 5.7955 5.6014 5.4172 5.2421 5.0757 4.9173 4.7665 4.6229 4.4859 4.3553 06 07 6.7282 6.4720 6.2302 6.0020 5.7863 5.5824 5.3893 5.2064 5.0329 4.8684 07 08 7.6517 7.3254 7.0196 6.7327 6.4632 6.2093 5.9713 5.7466 5.5348 5.3349 08 09 8.5661 8.1622 7.7861 7.4353 7.1078 6.8017 6.5152 6.2469 5.9852 5.7590 09 10 9.4714 8.9825 8.7302 8.1109 7.7217 7.3601 7.0236 6.7101 6.4176 6.1446 10

11 10.3677 9.7868 9.2526 8.7604 8.3064 7.8868 7.4987 7.1389 6.8052 6.4951 11 12 11.2552 10.5753 9.9589 9.3850 8.8632 8.3838 7.9427 7.5361 7.1601 6.8137 12 13 12.1338 11.3483 10.6349 9.9856 9.3935 8.8527 8.3576 7.9038 7.4869 7.1034 13 14 13.0088 12.1062 11.2960 10.5631 9.8986 9.2950 8.7454 8.2442 7.7860 7.3667 14 15 13.8651 12.8492 11.9379 11.1183 10.3796 9.7122 9.1079 8.5595 8.0607 7.6061 15

16 14.7180 13.5777 12.5610 11.6522 10.8377 10.1059 9.4466 8.8514 8.3126 7.8237 16 17 15.5624 14.2918 13.1660 12.1656 11.2740 10.4772 9.7632 9.1216 8.5435 8.0215 17 18 16.3984 14.9920 13.7534 12.6592 11.6895 10.8276 10.0591 9.3719 8.7556 8.2014 18 19 17.2201 15.2684 14.3237 13.1339 12.0853 11.1581 10.3356 9.6036 8.9501 8.3649 19 20 18.0457 16.3514 14.8774 13.5903 12.4622 11.4699 10.5940 9.8181 9.1285 8.5136 20

21 18.8571 17.0111 15.4149 14.0291 12.8211 11.7640 10.8355 10.0168 9.2922 8.6487 21 22 19.6605 17.6581 15.9368 14.4511 13.1630 12.0416 11.0612 10.2007 9.4424 8.7715 22 23 20.4559 18.2921 16.4435 14.8568 13.4885 12.3033 11.2722 10.3710 9.5802 8.8832 23 24 21.2435 18.9139 16.9355 15.2469 13.7986 12.5503 11.4693 10.5287 9.7066 8.9847 24 25 22.0233 19.5234 17.4181 15.6220 14.9039 12.7833 11.6536 10.6748 9.8226 9.0770 25

B-8 Appendix B: Tables

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE B.5 PRESENT VALUE OF AN ANNUITY OF $1 (PVIFA ) (CONTINUED)

Period 11% 12% 13% 14% 15% 16% 17% 18% 19% 20% Period

01 .9009 .8929 .8850 .8772 .8696 .8621 .8547 .8475 .8403 .8333 01 02 1.7125 1.6901 1.6681 1.6467 1.6257 1.6052 1.5852 1.5656 1.5465 1.5278 02 03 2.4437 2.4018 2.3612 2.3216 2.2832 2.2459 2.2096 2.1743 2.1399 2.1065 03 04 3.1024 3.0373 2.9745 2.9137 2.8550 2.7982 2.7432 2.6901 2.6386 2.5887 04 05 3.6959 3.6048 3.5172 3.4331 3.3522 3.2743 3.1993 3.1272 3.0576 2.9906 05

06 4.2305 4.1114 3.9976 3.8887 3.7845 3.6847 3.5892 3.4976 3.4098 3.3255 06 07 4.7122 4.5638 4.4226 4.2883 4.1604 4.0386 3.9224 3.8115 3.7057 3.6046 07 08 5.1461 4.9676 4.7988 4.6389 4.4873 4.3436 4.2072 4.0776 3.9544 3.8372 08 09 5.5370 5.3282 5.1317 4.9464 4.7716 4.6065 4.4506 4.3030 4.1633 4.0310 09 10 5.8892 5.6502 5.4262 5.2161 5.0188 4.8332 4.6586 4.4941 4.3389 4.1925 10

11 6.2065 5.9377 5.6869 5.4527 5.2337 5.0286 4.8364 4.6560 4.4865 4.3271 11 12 6.4924 6.1944 5.9176 5.6603 5.4206 5.1971 4.9884 4.7932 4.6105 4.4392 12 13 6.7499 6.4235 6.1218 5.8424 5.5831 5.3423 5.1183 4.9095 4.7147 4.5327 13 14 6.9819 6.6282 6.3025 6.0021 5.7245 5.4675 5.2293 5.0081 4.8023 4.6106 14 15 7.1909 6.8109 6.4624 6.1422 5.8474 5.5755 5.3242 5.0916 4.8759 4.6755 15

16 7.3792 6.9740 6.6039 6.2651 5.9542 5.6685 5.4053 5.1624 4.9377 4.7296 16 17 7.5488 7.1196 6.7291 6.3729 6.0472 5.7487 5.4746 5.2223 4.9897 4.7746 17 18 7.7016 7.2497 6.8389 6.4674 6.1280 5.8178 5.5339 5.2732 5.0333 4.8122 18 19 7.8393 7.3650 6.9380 6.5504 6.1982 5.8775 5.5845 5.3176 5.0700 4.8435 19 20 7.9633 7.4694 7.0248 6.6231 6.2593 5.9288 5.6278 5.3527 5.1009 4.8696 20

21 8.0751 7.5620 7.1016 6.6870 6.3125 5.9731 5.6648 5.3837 5.1268 4.8913 21 22 8.1757 7.6446 7.1695 6.7429 6.3587 6.0113 5.6964 5.4099 5.1486 4.9094 22 23 8.2664 7.7184 7.2297 6.7921 6.3988 6.0442 5.7234 5.4321 5.1668 4.9245 23 24 8.3481 7.7843 7.2829 6.8351 6.4338 6.0726 5.7465 5.4509 5.1822 4.9371 24 25 8.4217 7.8431 7.3300 6.8729 6.4641 6.0971 5.7662 5.4669 5.1951 4.9476 25

Appendix B: Tables B-9

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE B.6 DURBIN-WATSON STATISTIC FOR 2.5% SIGNIFICANCE (ONE-TAIL)

OR 5.0% SIGNIFICANCE (TWO-TAIL)

m � 1 m � 2 m � 3 m � 4 m � 5

n dL dU dL dU dL dU dL dU dL dU

15 0.95 1.23 0.83 1.40 0.71 1.61 0.59 1.84 0.48 2.09 16 0.98 1.24 0.86 1.40 0.75 1.59 0.64 1.80 0.53 2.03 17 1.01 1.25 0.90 1.40 0.79 1.58 0.68 1.77 0.57 1.98 18 1.03 1.26 0.93 1.40 0.82 1.56 0.72 1.74 0.62 1.93 19 1.06 1.28 0.96 1.41 0.86 1.55 0.76 1.73 0.66 1.90 20 1.08 1.28 0.99 1.41 0.89 1.55 0.79 1.72 0.70 1.87 21 1.10 1.30 1.01 1.41 0.92 1.54 0.83 1.69 0.73 1.84 22 1.12 1.31 1.04 1.42 0.95 1.54 0.86 1.68 0.77 1.82 23 1.14 1.32 1.06 1.42 0.97 1.54 0.89 1.67 0.80 1.80 24 1.16 1.33 1.08 1.43 1.00 1.54 0.91 1.66 0.83 1.79 25 1.18 1.34 1.10 1.43 1.02 1.54 0.94 1.65 0.86 1.77 26 1.19 1.35 1.12 1.44 1.04 1.54 0.96 1.65 0.88 1.76 27 1.21 1.36 1.13 1.44 1.06 1.54 0.99 1.64 0.91 1.75 28 1.22 1.37 1.15 1.45 1.08 1.54 1.01 1.64 0.93 1.74 29 1.24 1.38 1.17 1.45 1.10 1.54 1.03 1.63 0.96 1.73 30 1.25 1.38 1.18 1.46 1.12 1.54 1.05 1.63 0.98 1.73 31 1.26 1.39 1.20 1.47 1.13 1.55 1.07 1.63 1.00 1.72 32 1.27 1.40 1.21 1.47 1.15 1.55 1.08 1.63 1.02 1.71 33 1.28 1.41 1.22 1.48 1.16 1.55 1.10 1.63 1.04 1.71 34 1.29 1.41 1.24 1.48 1.17 1.55 1.12 1.63 1.06 1.70 35 1.30 1.42 1.25 1.48 1.19 1.55 1.13 1.63 1.07 1.70 36 1.31 1.43 1.26 1.49 1.20 1.56 1.15 1.63 1.09 1.70 37 1.32 1.43 1.27 1.49 1.21 1.56 1.16 1.62 1.10 1.70 38 1.33 1.44 1.28 1.50 1.23 1.56 1.17 1.62 1.12 1.70 39 1.34 1.44 1.29 1.50 1.24 1.56 1.19 1.63 1.13 1.69 40 1.35 1.45 1.30 1.51 1.25 1.57 1.20 1.63 1.15 1.69 45 1.39 1.48 1.34 1.53 1.30 1.58 1.25 1.63 1.21 1.69 50 1.42 1.50 1.38 1.54 1.34 1.59 1.30 1.64 1.26 1.69 55 1.45 1.52 1.41 1.56 1.37 1.60 1.33 1.64 1.30 1.69 60 1.47 1.54 1.44 1.57 1.40 1.61 1.37 1.65 1.33 1.69 65 1.49 1.55 1.46 1.59 1.43 1.63 1.40 1.66 1.36 1.69 70 1.51 1.57 1.48 1.60 1.45 1.63 1.42 1.66 1.39 1.70 75 1.53 1.58 1.50 1.61 1.47 1.64 1.45 1.67 1.42 1.70 80 1.54 1.59 1.52 1.63 1.49 1.65 1.47 1.67 1.44 1.70 85 1.56 1.60 1.53 1.63 1.51 1.66 1.49 1.68 1.46 1.71 90 1.57 1.61 1.55 1.64 1.53 1.66 1.50 1.69 1.48 1.71 95 1.58 1.62 1.56 1.65 1.54 1.67 1.52 1.69 1.50 1.71

100 1.59 1.63 1.57 1.65 1.55 1.67 1.53 1.70 1.51 1.72

m � number of independent variables n � number of observations Source: From J. Durbin and G. S. Watson, “Testing for Serial Correlation in Least-Squares Regression,” Biometrika, Vol. 38 (1951); 159–177. With the permission of the authors and the Trustees of Biometrika.

B-10 Appendix B: Tables

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

TABLE B.7 CRITICAL VALUES FOR THE DICKEY-FULLER TEST

Sample Size

25 50 100 �

F ratio 7.24 6.73 6.49 6.25 AR(1) model 2.16 2.08 2.03 2.00 AR(1) model with constant 0.72 0.66 0.63 0.60 AR(1) model with constant and time trend �0.15 �0.15 �0.28 �0.33

D. Dickey and W. Fuller, “Likelihood Ratio Tests for Autoregressive Time Series with A Unit Root,” Econometrica, 49, 1981.

Appendix B: Tables B-11

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

C APPENDIX

Differential CalculusTechniques in Management

Decision analysis involves determining the action that best achieves a desired goal or ob- jective. It means finding the action that optimizes (that is, maximizes or minimizes) the value of an objective function. For example, we may be interested in determining the output level that maximizes profits. In a production problem, the goal may be to find the combination of inputs that minimizes the cost of producing a desired level of output. In a capital budgeting problem, the objective may be to select those projects that maxi- mize the net present value of the investments chosen. Many techniques are available for solving optimization problems such as these. This appendix focuses on the use of differ- ential calculus.

RELATIONSHIP BETWEEN MARGINAL ANALYSIS AND DIFFERENTIAL CALCULUS In Chapter 2, marginal analysis was introduced as one of the fundamental concepts of microeconomic decision making. In the marginal analysis framework, resource allocation decisions are made by comparing the marginal benefits of a proposed change in the level of an activity with the marginal costs of that change. The proposed change should be made as long as the marginal benefits exceed the marginal costs. By following this basic rule, resources can be allocated efficiently and profits or shareholder wealth can be maximized.

Initially, let us assume that the objective we are seeking to optimize, Y, can be ex- pressed algebraically as a function of one decision variable, X.

Y = f ðXÞ [C.1] Recall that marginal profit is defined as the change in profit resulting from a one-unit

change in output. In general, the marginal value of any variable Y, which is a function of another variable X, is defined as the change in the value of Y resulting from a one-unit change in X. The marginal value of Y, My, can be calculated from the change in Y, ΔY, that occurs as the result of a given change in X, ΔX:

My = ΔY ΔX

[C.2]

When calculated with this expression, different estimates for the amount ΔY may be obtained, depending on the size of the incremental change in X that we use in the com- putation. The true marginal value of a function is obtained from Equation C.2 when ΔX is made as small as possible. If ΔX can be thought of as a continuous (rather than a discrete)

C-1

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

variable that can take on fractional values,2 then in calculating My by Equation C.2, we can let ΔX approach zero.

In concept, differential calculus takes this approach. The derivative, or, more pre- cisely, first derivative,3 dY/dX, of a function is defined as the limit of the ratio ΔY/ΔX as ΔX approaches zero; that is,

dY dX

= limit ΔX−→0

ΔY ΔX

[C.3]

Graphically, the first derivative of a function represents the slope of the curve at a given point on the curve. The definition of a derivative as the limit of the change in Y (that is, ΔY) as ΔX approaches zero is illustrated in Figure C.1(a).

Suppose we are interested in the derivative of the Y = f(X) function at the point X0. The derivative dY/dX measures the slope of the tangent line ECD. An estimate of this slope, albeit a poor estimate, can be obtained by calculating the marginal value of Y over the interval X0 to X2. Using Equation C.2, a value of

M0y = ΔY ΔX

= Y2 − Y0 X2 − X0

MANAGERIAL CHALLENGE A Skeleton in the Stealth Bomber’s Closet1

In 1990, the U.S. Air Force publicly unveiled its newest long-range strategic bomber, the B-2 or “Stealth” bomber. This plane is characterized by a unique flying wing design engineered to evade detection by enemy radar. The plane has been controversial because of its high cost. However, a lesser known controversy relates to its fundamental design.

The flying wing design originated from a secret study that concluded that a plane’s maximum range could be achieved if virtually all the volume were contained in the wing. A complex mathematical appendix was at- tached to the study.

However, Professor of Engineering Joseph Foa dis- covered that a fundamental error had been made in the initial report. It turned out that the original re- searchers had taken the first derivative of a complex equation and found that it had two solutions. The

original researchers mistakenly concluded that the all- wing design was the one that maximized range, when, in fact, it minimized range.

In this chapter we introduce some of the same opti- mization techniques applied to the Stealth bomber proj- ect. We develop tools designed to maximize profits or minimize costs. Fortunately, the mathematical functions we deal with in this chapter and throughout the book are much simpler than those that confronted the origi- nal “flying wing” engineers. We introduce techniques that can be used to check whether a function, such as profits or costs, is being minimized or maximized at a particular level of output.

1Based on W. Biddle, “Skeleton Alleged in the Stealth Bomber’s Closet,” Science (May 12, 1989), pp. 650–651.

2For example, if X is a continuous variable measured in feet, pounds, and so on, then ΔX can in theory take on fractional values such as 0.5, 0.10, 0.05, 0.001, 0.0001 feet or pounds. When X is a continuous variable, ΔX can be made as small as desired.

derivative A measure of the marginal effect of a change in one variable on the value of a function. Graphically, it represents the slope of the function at a given point.

3It is also possible to compute second, third, fourth, and so on, derivatives. Second derivatives are discussed later in this appendix.

C-2 Appendix C: Differential Calculus Techniques in Management

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

is obtained for the slope of the CA line. Now let us calculate the marginal value of Y using a smaller interval, for example, X0 to X1. The slope of the CB line, which is equal to

M00y = ΔY ΔX

= Y1 − Y0 X1 − X0

gives a much better estimate of the true marginal value as represented by the slope of the ECD tangent line. Thus we see that the smaller the ΔX value, the better the estimate of the slope of the curve. Letting ΔX approach zero allows us to find the slope of the Y = f(X) curve at point C. As shown in Figure C.1(b), the slope of the ECD tangent line (and the Y = f(X) function at point C) is measured by the change in Y, or rise, ΔY, divided by the change in X, or run, ΔX.

FIGURE C.1 First Derivative of a Function

Y

Y2

Y = f(X) Y1

Y0

X0 X1 X2 X

(a) Marginal change in Y = f(X) as ΔX approaches 0

C

E

D

B

A

Y

Y = f(X)

Y0

X0 X

(b) Measurement of the slope Y = f(X) at point C

C

E

D

ΔY

ΔX

Appendix C: Differential Calculus Techniques in Management C-3

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Process of Differentiation The process of differentiation—that is, finding the derivative of a function—involves de- termining the limiting value of the ratio ΔY/ΔX as ΔX approaches zero. Before offering some general rules for finding the derivative of a function, we illustrate with an example the algebraic process used to obtain the derivative without the aid of these general rules. The specific rules that simplify this process are presented in the following section.

Example Process of Differentiation: Profit Maximization at Illinois Power Suppose the profit, π, of Illinois Power can be represented as a function of the out- put level Q using the expression

π = −40 + 140Q − 10Q2 [C.4]

We wish to determine dπ/dQ by first finding the marginal profit expression Δπ/ΔQ and then taking the limit of this expression as ΔQ approaches zero. Let us begin by expressing the new level of profit (π + Δπ) that will result from an increase in output to (Q + ΔQ). From Equation C.4, we know that

π + Δπ = −40 + 140ðQ + ΔQÞ − 10ðQ + ΔQÞ2 [C.5] Expanding this expression and then doing some algebraic simplifying, we obtain

π + Δπ = −40 + 140Q + 140ΔQ − 10½Q2 + 2QΔQ + ðΔQÞ2� = −40 + 140Q − 10Q2 + 140ΔQ − 20QΔQ − 10ðΔQÞ2 [C.6]

Subtracting Equation C.4 from Equation C.6 yields

Δπ = 140ΔQ − 20QΔQ − 10ðΔQÞ2 [C.7] Forming the marginal profit ratio Δπ/ΔQ, and doing some canceling, we get

Δπ

ΔQ =

140ΔQ − 20QΔQ − 10ðΔQÞ2 ΔQ

= 140 − 20Q − 10ΔQ

[C.8]

Taking the limit of Equation C.8 as ΔQ approaches zero yields the expression for the derivative of Illinois Power’s profit function (Equation C.4).

dπ dQ

= limit ΔQ−→0

½140 − 20Q − 10ΔQ�

= 140 − 20Q

[C.9]

If we are interested in the derivative of the profit function at a particular value of Q, Equation C.9 can be evaluated for this value. For example, suppose we want to know the marginal profit, or slope of the profit function, at Q = 3 units. Substi- tuting Q = 3 in Equation C.9 yields

Marginal profit = dπ dQ

= 140 − 20ð3Þ = $80 per unit

C-4 Appendix C: Differential Calculus Techniques in Management

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Rules of Differentiation Fortunately, we do not need to go through this lengthy process every time we want the derivative of a function. A series of general rules, derived in a manner similar to the pro- cess just described, exists for differentiating various types of functions.

Constant Functions A constant function can be expressed as

Y = a [C.10]

where a is a constant (that is, Y is independent of X). The derivative of a constant func- tion is equal to zero:

dY dX

= 0 [C.11]

For example, consider the constant function

Y = 4

which is graphed in Figure C.2(a). Recall that the first derivative of a function (dY/dX) measures the slope of the function. Because this constant function is a horizontal straight line with zero slope, its derivative (dY/dX) is therefore equal to zero.

Power Functions A power function takes the form of

Y = aXb [C.12]

where a and b are constants. The derivative of a power function is equal to:

dY dX

= b · a · Xb−1 [C.13]

FIGURE C.2 Constant, Linear, and Quadratic Functions

ΔY = 2

ΔX = 1

Y

Y = 4

X

10

8

6

4

2

0 2 4 6 8

(a) Constant function

Y

X

10

8

6

4

2

0 2 4 6 8

(b) Linear function

Y

Y = X2

X

10

8

6

4

2

–4 –2 0 2 4

(c) Quadratic function

Y = 2X

Appendix C: Differential Calculus Techniques in Management C-5

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

A couple of examples are used to illustrate the application of this rule. First, consider the function

Y = 2X

which is graphed in Figure C.2(b). Note that the slope of this function is equal to 2 and is constant over the entire range of X values. Applying the power function rule to this example, where a = 2 and b = 1, yields

dY dX

= 1 · 2 · X1−1

= 2X0 = 2

Note that any variable raised to the zero power, e.g., X0, is equal to 1. Next, consider the function

Y = X2

which is graphed in Figure C.2(c). Note that the slope of this function varies depending on the value of X. Application of the power function rule to this example yields (a = 1, b = 2):

dY dX

= 2 · 1 · X2−1

= 2X

As we can see, this derivative (or slope) function is negative when X < 0, zero when X = 0, and positive when X > 0.

Sums of Functions Suppose a function Y = f(X) represents the sum of two (or more) separate functions, f1(X), f2(X), that is,

Y = f1ðXÞ + f2ðXÞ [C.14] The derivative of Y with respect to X is found by differentiating each of the separate

functions and then adding the results:

dY dX

= df1ðXÞ dX

+ df2ðXÞ dX

[C.15]

This result can be extended to finding the derivative of the sum of any number of functions.

Example Rules of Differentiation: Profit Maximization at Illinois Power (continued) As an example of the application of these rules, consider again the profit function for Illinois Power, given earlier in Equation C.4:

π = −40 + 140Q − 10Q2

In this example Q represents the X variable and π represents the Y variable; that is, π = f(Q). The function f(Q) is the sum of three separate functions: a constant

C-6 Appendix C: Differential Calculus Techniques in Management

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Product of Two Functions Suppose the variable Y is equal to the product of two separate functions f1(X) and f2(X):

Y = f1ðXÞ · f2ðXÞ [C.16] In this case the derivative of Y with respect to X is equal to the sum of the first func-

tion times the derivative of the second, plus the second function times the derivative of the first.

dY dX

= f1ðXÞ · df2ðXÞdX + f2ðXÞ · df1ðXÞ dX

[C.17]

For example, suppose we are interested in the derivative of the expression

Y = X2ð2X − 3Þ Let f1(X) = X

2 and f2(X) = (2X – 3). By the preceding rule (and the earlier rules for differentiating constant and power functions), we obtain

dY dX

= X2 · dY dX

½ð2X − 3Þ� + ð2X − 3Þ · dY dX

½X2�

= X2 · ð2 − 0Þ + ð2X − 3Þ · ð2XÞ = 2X2 + 4X2 − 6X

= 6X2 − 6X

= 6XðX − 1Þ

Quotient of Two Functions Suppose the variable Y is equal to the quotient of two separate functions f1(X) and f2(X):

Y = f1ðXÞ f2ðXÞ [C.18]

For such a relationship the derivative of Y with respect to X is obtained as follows:

dY dX

= f2ðXÞ · df1ðXÞdX − f1ðXÞ ·

df2ðXÞ dX

½f2ðXÞ�2 [C.19]

function, f1(Q) = −40, and two power functions, f2(Q) = 140Q and f3(Q) – 10Q 2.

Therefore, applying the differentiation rules yields

dπ dQ

= df1ðQÞ dQ

+ df2ðQÞ dQ

+ df3ðQÞ dQ

= 0 + 1 · 140 · Q1−1 + 2 · ð−10Þ · Q2−1

= 140 − 20Q

This result is the same as obtained in Equation C.9 by the differentiation process.

Appendix C: Differential Calculus Techniques in Management C-7

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

As an example, consider the problem of finding the derivative of the expression

Y = 10X2

5X − 1

Letting f1(X) = 10X 2 and f2(X) = 5X – 1, we have

dY dX

= ð5X − 1Þ · 20X − 10X2 · 5

ð5X − 1Þ2

= 100X2 − 20X − 50X2

ð5X − 1Þ2

= 50X2 − 20X

ð5X − 1Þ2

= 10Xð5X − 2Þ ð5X − 1Þ2

Functions of a Function (Chain Rule) Suppose Y is a function of the variable Z, Y = f1(Z); and Z is in turn a function of the variable X, Z = f2(X). The derivative of Y with respect to X can be determined by first finding dY/dZ and dZ/dX and then multi- plying the two expressions together:

dY dX

= dY dZ

· dZ dX

= df1ðZÞ dZ

· df2ðXÞ dX [C.20]

To illustrate the application of this rule, suppose we are interested in finding the derivative (with respect to X) of the function

Y = 10Z − 2Z2 − 3

where Z is related to X in the following way:4

Z = 2X2 − 1

First, we find (by the earlier differentiation rules)

dY dZ

= 10 − 4Z

dZ dX

= 4X

4Alternatively, one can substitute Z = 2X2 – 1 into Y = 10Z – 2Z2 – 3 and differentiate Y with respect to X.

C-8 Appendix C: Differential Calculus Techniques in Management

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

and then

dY dX

= ð10 − 4ZÞ · 4X

Substituting the expression for Z in terms of X into this equation yields

dY dX

= ½10 − 4ð2X2 − 1Þ� · 4X

= ð10 − 8X2 + 4Þ · 4X = 40X − 32X3 + 16X

= 56X − 32X3

= 8Xð7 − 4X2Þ These rules for differentiating functions are summarized in Table C.1.

APPLICATIONS OF DIFFERENTIAL CALCULUS TO OPTIMIZATION PROBLEMS The reason for studying the process of differentiation and the rules for differentiating functions is that these methods can be used to find optimal solutions to many kinds of maximization and minimization problems in managerial economics.

TABLE C.1 SUMMARY OF RULES FOR DIFFERENTIATING FUNCTIONS

FUNCTION DERIVATIVE

1. Constant Function

Y = a dY dX

= 0

2. Power Function

Y = aXb dY dX

= b · a · Xb−1

3. Sums of Functions

Y = f1(X) + f2(X) dY dX

= df1ðXÞ dX

+ df2ðXÞ dX

4. Product of Two Functions

Y = f1(X) · f2(X) dY dX

= f1ðXÞ · df2ðXÞdX + f2ðXÞ · df1ðXÞ dX

5. Quotient of Two Functions

Y = f1ðXÞ f2X

dY dX

= f2ðXÞ · df1ðXÞdX − f1ðXÞ ·

df2ðXÞ dX

½f2ðXÞ�2

6. Functions of a Function

Y = f1(Z), where Z = f2(X) dY dX

= dY dZ

· dZ dX

Appendix C: Differential Calculus Techniques in Management C-9

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Maximization Problem As you recall from the discussion of marginal analysis, a necessary (but not sufficient) condition for finding the maximum point on a curve (e.g., maximum profits) is that the marginal value or slope of the curve at this point must be equal to zero. We can now express this condition within the framework of differential calculus. Because the de- rivative of a function measures the slope or marginal value at any given point, an equiv- alent necessary condition for finding the maximum value of a function Y = f(X) is that the derivative dY/dX at this point must be equal to zero. This requirement is known as the first-order condition for locating one or more maximum or minimum points of an algebraic function.

Second Derivatives and the Second-Order Condition Setting the derivative of a function equal to zero and solving the resulting equation for the value of the decision variable does not guarantee that the point will be obtained at which the function takes on its maximum value. (Recall the Stealth bomber example at the start of this appendix.) The slope of a U-shaped function will also be equal to zero at its low point and the function will take on its minimum value at the given point. In other words, setting the derivative to zero is only a necessary condition for finding the maxi- mum value of a function; it is not a sufficient condition. Another condition, known as the second-order condition, is required to determine whether a point that has been de- termined from the first-order condition is either a maximum point or minimum point of the algebraic function.

This situation is illustrated in Figure C.4. At both points A and B the slope of the function (first derivative, dY/dX) is zero; however, only at point B does the function

Example First-Order Condition: Profit Maximization at Illinois Power (continued) Using the profit function (Equation C.4)

π = −40 + 140Q − 10Q2

discussed earlier, we can illustrate how to find the profit-maximizing output level Q by means of this condition. Setting the first derivative of this function (which was computed previously) to zero, we obtain

dπ dQ

= 140 − 20Q

0 = 140 − 20Q

Solving this equation for Q yields Q* = 7 units as the profit-maximizing output level. The profit and first derivative functions and optimal solution are shown in Figure C.3. As we can see, profits are maximized at the point where the function is neither increasing nor decreasing; in other words, where the slope (or first deriv- ative) is equal to zero.

first-order condition A test to locate one or more maximum or minimum points of an algebraic function.

second-order condition A test to determine whether a point that has been determined from the first-order condition is either a maximum point or a minimum point of the algebraic function.

C-10 Appendix C: Differential Calculus Techniques in Management

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

take on its maximum value. We note in Figure C.4 that the marginal value (slope) is continually decreasing in the neighborhood of the maximum value (point B) of the Y = f(X) function. First the slope is positive up to the point where dY/dX = 0, and thereafter the slope becomes negative. Thus we must determine whether the slope’s marginal value (slope of the slope) is declining. To test whether the marginal value is decreasing, take the derivative of the marginal value and determine whether it is neg- ative at the given point on the function. In effect, we need to find the derivative of the derivative—that is, the second derivative of the function—and then test whether it is less than zero. Formally, the second derivative of the function Y = f(X) is written as d2Y/dX2 and is found by applying the previously described differentiation rules to the first derivative. A maximum point is obtained if the second derivative is negative; that is, d2Y/dX2 < 0.

FIGURE C.3 Profit and First Derivative Functions

dπ dQ500

400

300

200

100

0 1 2 3 4 5 6 7

= 140 – 20Q = 0 = Marginal profit at Q = 7

Total profit (π)

π($)

8 9 10 11 1312 14 –10

Q units

π = – 40 + 140Q – 10Q2

= 140 – 20Q

200

150

100

50

–100

–150

–200

0 1 2 3 4 5 6 7 8 9 10 11 1312 14

–50 Q units

dπ dQ

($/unit)

dπ dQ

Appendix C: Differential Calculus Techniques in Management C-11

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

FIGURE C.4 Maximum and Minimum Values of a Function

= 0

Y

X

X

Y = f(X)

B

A

dX dY

= 0 dX dY

dX dY

dX dY

Example Second-Order Condition: Profit Maximization at Illinois Power (continued) In the profit-maximization example, the second derivative is obtained from the first derivative as follows:

dπ dQ

= 140 − 20Q

d2π dQ2

= 0 + 1 · ð−20Þ · Q1−1

= −20

Because d2π/dQ2 < 0, we know that a maximum-profit point has been obtained.

C-12 Appendix C: Differential Calculus Techniques in Management

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

An opposite condition holds for obtaining the point at which the function takes on a minimum value. Note again in Figure C.4 that the marginal value (slope) is continually increasing in the neighborhood of the minimum value (point A) of the Y = (X) function. First the slope is negative up to the point where dY/dX = 0, and thereafter the slope becomes positive. Therefore, we test to see whether d2Y/dX2 > 0 at the given point. A minimum point is obtained if the second derivative is positive; that is, d2Y/dX2 > 0.

Minimization Problem In some decision-making situations, cost minimization may be the objective. As in profit-maximization problems, differential calculus can be used to locate the optimal points.

Summarizing, we see that two conditions are required for locating a maximum or minimum value of a function using differential calculus. The first-order condition deter- mines the point(s) at which the first derivative dY/dX is equal to zero. After we obtain one or more points, a second-order condition is used to determine whether the function takes on a maximum or minimum value at the given point(s). The second derivative d2Y/dX2 indicates whether a given point is a maximum (d2Y/dX2 < 0) or a minimum (d2Y/dX2 > 0) value of the function.

Example Cost Minimization: KeySpan Energy Suppose we are interested in determining the output level that minimizes average total costs for KeySpan Energy, where the average total cost function might be ap- proximated by the following relationship (Q represents output):

C = 15 − 0:040Q + 0:000080Q2

Differentiating C with respect to Q gives

dC dQ

= −0:040 + 0:000160Q

Setting this derivative equal to zero and solving for Q yields

0 = 0:040 + 0:000160Q

Q* = 250

Taking the second derivative, we obtain

d2C dQ2

= +0:000160

Because the second derivative is positive, the output level of Q = 250 is indeed the value that minimizes average total costs.

Appendix C: Differential Calculus Techniques in Management C-13

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

PARTIAL DIFFERENTIATION AND MULTIVARIATE OPTIMIZATION Thus far in this appendix, the analysis has been limited to a criterion variable Y that can be expressed as a function of one decision variable X. However, many commonly used economic relationships contain two or more decision variables. For example, a demand function relates sales of a product or service to such variables as price, advertising, pro- motion expenses, price of substitutes, and income.

Partial Derivatives Consider a criterion variable Y that is a function of two decision variables X1 and X2.

5

Y = f ðX1, X2Þ [C.21] Let us now examine the change in Y that results from a given change in either X1 or X2. To isolate the marginal effect on Y from a given change in X1 (that is, ΔY/ΔX1), we must hold X2 constant. Similarly, if we wish to isolate the marginal effect on Y from a given change in X2 (that is, ΔY/ΔX2) the variable X1 must be held constant. A measure of the marginal effect of a change in any one variable on the change in Y, holding all other variables in the relationship constant, is obtained from the partial derivative of the func- tion. The partial derivative of Y with respect to X1 is written as ∂Y/∂X1 and is found by applying the previously described differentiation rules to the Y = f(X1, X2) function, where the variable X2 is treated as a constant. Similarly, the partial derivative of Y with respect to X2 is written as ∂Y/∂X2 and is found by applying the differentiation rules to the function, where the variable X1 is treated as a constant.

Example Partial Derivatives: Indiana Petroleum Company To illustrate the procedure for obtaining partial derivatives, let us consider the fol- lowing relationship in which the profit variable, π, is a function of the output level of two products (heating oil and gasoline) Q1 and Q2:

π = −60 + 140Q1 + 100Q2 − 10Q 2 1 − 8Q

2 2 − 6Q1Q2 [C.22]

Treating Q2 as a constant, the partial derivative of π with respect to Q1 is obtained:

∂π

∂Q1 = 0 + 140 + 0 + 2 · ð−10Þ · Q1 − 0 − 6Q2 = 140 − 20Q1 − 6Q2 [C.23]

Similarly, with Q1 treated as a constant, the partial derivative of π with respect to Q2 is equal to

∂π

∂Q1 = 0 + 0 + 100 − 0 + 2 · ð−8Þ · Q2 − 6Q1 = 100 − 16Q2 − 6Q1 [C.24]

5The following analysis is not limited to two decision variables. Relationships containing any number of vari- ables can be analyzed within this framework.

partial derivative A measure of the marginal effect of a change in one variable on the value of a multivariate function, while holding constant all other variables.

C-14 Appendix C: Differential Calculus Techniques in Management

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Maximization Problem The partial derivatives can be used to obtain the optimal solution to a maximization or minimization problem containing two or more X variables. Analogous to the first-order conditions discussed earlier for the one-variable case, we set each of the partial deriva- tives equal to zero and solve the resulting set of simultaneous equations for the optimal X values.

Example Partial Derivatives: Demand Function for Shield Toothpaste Partial derivatives can be useful in demand analysis, especially in quantitative stud- ies. Suppose the demand for Shield toothpaste is estimated as tubes per year,

Q = 14:6 + 2:2P + 7:4A [C.25]

where Q = quantity sold, P = selling price, and A = advertising campaigns, the partial derivatives of Q with respect to P and A are

∂Q ∂P

= −2:2 and ∂Q ∂A

= 7:4

To take another example, for the multiplicative exponential demand function

Q = 3:0P−:50A:25

The partial derivative of Q with respect to P is

∂Q ∂P

= 3:0A:25ð−:50P−:50−1Þ = −1:5P−1:50A:25

Similarly, the partial derivative of Q with respect to A is

∂Q ∂A

= 3:0P−:50ð:25A:25−1Þ = :75P−:50A−:75

Example Profit Maximization: Indiana Petroleum Company (continued) Suppose we are interested in determining the values of Q1 and Q2 that maximize the company’s profits given in Equation C.22. In this case, each of the two partial derivative functions (Equations C.23 and C.24) would be set equal to zero:

0 = 140 − 20Q1 − 6Q2 0 = 100 − 16Q2 − 6Q1

This system of equations can be solved for the profit-maximizing values of Q1 and Q2.

6 The optimal values are Q*1 = 5:77 units and Q*2 = 4:08 units. 7 The opti-

mal total profit is

π* = −60 + 140ð5:77Þ + 100ð4:08Þ + 10ð5:77Þ2 − 8ð4:08Þ2 − 6ð5:77Þð4:08Þ = 548:45 6The second-order conditions for obtaining a maximum or minimum in the multiple-variable case are somewhat complex. A discussion of these conditions can be found in most basic calculus texts. 7Exercise 10 at the end of this appendix requires the determination of these optimal values.

Appendix C: Differential Calculus Techniques in Management C-15

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

SUMMARY

� Marginal analysis is useful in making decisions about the expansion or contraction of an economic activity.

� Differential calculus, which bears a close relationship to marginal analysis, can be applied whenever an al- gebraic relationship can be specified between the de- cision variables and the objective or criterion variable.

� The first derivative measures the slope or rate of change of a function at a given point and is equal to the limiting value of the marginal function as the marginal value is calculated over smaller and smal- ler intervals, that is, as the interval approaches zero.

� Various rules are available (see Table C.1) for find- ing the derivative of specific types of functions.

� A necessary, but not sufficient, condition for find- ing the maximum or minimum points of a func- tion is that the first derivative be equal to zero, which is known as the first-order condition.

� A second-order condition is required to determine whether a given point is a maximum or minimum. The second derivative indicates that a given point is a maximum if the second derivative is less than zero or a minimum if the second derivative is greater than zero.

� The partial derivative of a multivariate function measures the marginal effect of a change in one variable on the value of the function, holding con- stant all other variables.

Exercises 1. Define Q as the level of output produced and sold, and suppose that a firm’s total revenue (TR) and total cost (TC) functions can be represented in tabular form as shown here.

OUTPUT (Q)

TOTAL REVENUE

(TR) TOTAL COST

(TC ) OUTPUT

(Q)

TOTAL REVENUE

(TR) TOTAL COST

(TC )

0 0 20 11 264 196

1 34 26 12 276 224

2 66 34 13 286 254

3 96 44 14 294 286

4 124 56 15 300 320

5 150 70 16 304 356

6 174 86 17 306 394

7 196 104 18 306 434

8 216 124 19 304 476

9 234 146 20 300 520

10 250 170

a. Compute the marginal revenue and average revenue functions. b. Compute the marginal cost and average cost functions. c. On a single graph, plot the total revenue, total cost, marginal revenue, and

marginal cost functions. d. Determine the output level in the graph that maximizes profits (Profit =

Total revenue – Total cost) by finding the point where marginal revenue equals marginal cost.

C-16 Appendix C: Differential Calculus Techniques in Management

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

e. Check your result in part (d) by finding the output level in the tables devel- oped in parts (a) and (b) that likewise satisfies the condition that marginal revenue equals marginal cost.

2. Consider again the total revenue and total cost functions shown in tabular form in the previous problem. a. Compute the total, marginal, and average profit functions. b. On a single graph, plot the total profit and marginal profit functions. c. Determine the output level in the graph and table where the total profit

function takes on its maximum value. d. How does the result in part (c) in this exercise compare with the result in

part (d) of the previous exercise? e. Determine total profits at the profit-maximizing output level.

3. Differentiate the following functions: a. TC = 50 + 100Q − 6Q2 + .5Q3

b. ATC = 50/Q + 100 − 6Q + .5Q2

c. MC = 100 − 12Q + 1.5Q2

d. Q = 50 − .75P e. Q = .40X1.50

4. Differentiate the following functions: a. Y = 2X3/(4X2 − 1) b. Y = 2X/(4X2 − 1) c. Y = 8Z2 − 4Z + 1, where Z = 2X2 − 1 (differentiate Y with respect to X)

5. Define Q to be the level of output produced and sold, and assume that the firm’s cost function is given by the relationship

TC = 20 + 5Q + Q2

Furthermore, assume that the demand for the output of the firm is a function of price P given by the relationship

Q = 25 − P

a. Define total profit as the difference between total revenue and total cost, and express in terms of Q the total profit function for the firm. (Note: Total rev- enue equals price per unit times the number of units sold.)

b. Determine the output level where total profits are maximized. c. Calculate total profits and selling price at the profit-maximizing output level. d. If fixed costs increase from $20 to $25 in the total cost relationship, deter-

mine the effects of such an increase on the profit-maximizing output level and total profits.

6. Use the cost and demand functions in Exercise 5 to calculate the following: a. Determine the marginal revenue and marginal cost functions. b. Show that, at the profit-maximizing output level determined in part (b) of

the previous exercise, marginal revenue equals marginal cost and illustrates the economic principle that profits are maximized at the output level where marginal revenue equals marginal cost.

Appendix C: Differential Calculus Techniques in Management C-17

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

7. Determine the partial derivatives with respect to all of the variables in the follow- ing functions: a. TC = 50 + 5Q1 + 10Q2 + .5Q1Q2 b. Q = 1.5L.60 K.50

c. QA = 2.5PA –1.30Y.20PB

.40

8. Bounds Inc. determined through regression analysis that its sales (S) are a func- tion of the amount of advertising (measured in units) in two different media. This relationship is given by the following equation (X = newspapers, Y = magazines):

SðX,YÞ = 200X + 100Y − 10X2 − 20Y2 + 20XY a. Find the level of newspaper and magazine advertising that maximizes the

firm’s sales. b. Calculate the firm’s sales at the optimal values of newspaper and magazine

advertising determined in part (a).

9. The Santa Fe Cookie Factory is considering an expansion of its retail piñon cookie business to other cities. The firm’s owners lack the funds needed to undertake the expansion on their own. They are considering a franchise arrangement for the new outlets. The company incurs variable costs of $6 for each pound of cookies sold. The fixed costs of operating a typical retail outlet are estimated to be $300,000 per year. The demand function facing each retail outlet is estimated to be

P = $50 − :001Q

where P is the price per pound of cookies and Q is the number of pounds of cookies sold. [Note: Total revenue equals price (P) times quantity (Q) sold.] a. What price, output, total revenue, total cost, and total profit level will each

profit-maximizing franchise experience? b. Assume that the parent company charges each franchisee a fee equal to

5 percent of total revenues, and recompute the values in part (a). c. The Santa Fe Cookie Factory is considering a combined fixed/variable fran-

chise fee structure. Under this arrangement, each franchisee would pay the parent company $25,000 plus 1 percent of total revenues. Recompute the values in part (a).

d. What franchise fee arrangement do you recommend that the Santa Fe Cookie Factory adopt? What are the advantages and disadvantages of each plan?

10. Show that the optimal solution to the set of simultaneous equations in the Indiana Petroleum example are Q*1 = 5:77 and Q*2 = 4:08.

C-18 Appendix C: Differential Calculus Techniques in Management

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

D APPENDIX

Check Answers to Selected End-of-Chapter Exercises

Chapter 1

Case Exercise – Designing a Managerial Incentives Contract 5. $1,200,000 6. $167 million versus $118.8 million

Chapter 2 3. Budget = $875 million 4. c. ν = 0.067

Chapter 3 2. 44% 5. P = $90 6. a. ED = −0.59 8. a. EX = 1.34.

Close substitutes. 9. Q2006 = 5,169

Q2007 = 3,953

Chapter 4 3. d. r2 = 0.885 9. a. Y0 = −14.7351 + 3.9214 Size + 3.5851

Rooms −0.1181 Age −2.8317 Garage

Case Exercise – Soft Drink Demand Estimation 2. ED = −3.38

Appendix 4A 2. a. Y0 = 1.210 + 0.838 Selling Expenses,

r2 = 0.93

4. a. (i) S0 = 247.644 + 0.3926 Advertising −0.7339 Price

(ii) Log(S0) = 2.4482 + 0.7296 Log Advertising −0.2406 Log Price

Chapter 5 3. b. Sum(Actual/Forecast)/6 = 636.6%/6

= 106.1%, thus + 6% 4. b. GNP = C + I + G = 635 + 120 + 200

= 955 7. b. Y02007 = 259.03 8. a. December 2007 = 468

Chapter 6 1. Both increase 3. Outsource abroad and buy foreign

assets 6. 50% decline. Relative purchasing

power parity

Chapter 7 3. b. 10 or 11 men 5. c. APX = 6X − 0.4X

2

7. a. 4.88%

Case Exercise – Production Function for Wilson Company 4. EK = 0.415, EL = 1.078

Appendix 7A 1. d. Q* = 43.231, X* = 10.902,

Y* = 11.278

D-1

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Chapter 8 2. b. ($90,000)

Case Exercise – Cost Analysis 1. $4.55

Appendix 8A 1. a. L* = 2.5 units

Chapter 9 2. a. Q* = $574.08 (million) 6. a. $30,000,000

Case Exercise – Cost Functions 5. Q* = 1,675

Case Exercise – Charter Airline Operating Decisions 3. Indirect Fixed

Cost = $23,900

Chapter 10 8. b. P* = $1,220 9. b. $900,000 on advertising

Chapter 11 2. c. Q* = 125 3. e. π* = $263,625 4. b. P* = $60 8. a. ROI = 14.2% 9. a. ROI = 12.98%

Chapter 12 2. a. P* = $145

Q*A = 30

5. a. P* = $9,666.70, Q* = 666.7 6. c. P* = $125, Q* = 50

Chapter 13 3. b. Dominant strategy for AMC

is to “Not Abide” 5. {$150, Match}, No 6. Least should pass. More should

always attack Most, and know- ing that, Most should always attack More. If they pass, Least will get a second opportunity to attack a once stronger but now weakened opponent.

8. {Late, Late} is one of two pure Nash equilibria.

Chapter 14 1. PUS = $80, POVERSEAS = $22.50 3. a. π = −20 + 96Q1 + 76Q2 − 2Q21

− Q22 11. 22 seats

Chapter 15 3. High interest rates, large prin-

cipal, long term, unsecured 4. Vertical integration if power

plant dependent on this type of coal. Otherwise, long-term supply contracts.

Case Exercise – Division of Investment Banking Fees in a Syndicate 1. Lead underwriter = $97 million

Syndicate Co-manager = 0 Syndicate Member 3 = $1 million Syndicate Member 4 = 0 Syndicate Member 5 = $2 million

Appendix 15A 4. Electricity, T-bills 5. $1.3 million.

Use open bidding, multiple- rounds, highest-wins-and-pays.

11. Apple’s expected profit is $1.5 million less from understatement.

Chapter 16 3. a. HHI before = 1,964. So, in

general, No, although offset- ting efficiency arguments may come into play as long as 1,984 is to the 1,800 standard.

6. b. π* = $450 million 11. Coordinate on Nash

equilibrium (Lucent Imitate, Motorola Develop) in joint venture with compensation of at least $1 billion to Motorola.

Chapter 17 2. IRR = 9.1%. So, No. 4. b. NCF10 = $5,560 5. a. IRR = 14.94%, NPV = $45,176 6. ke = 13.4% 7. b. k0e = 13% 8. ka = 12.3% 9. b. Power plant: NPV@12% =

−$22.71 million, NPV@5% = $62.65 million

Case Exercise – Cost- Benefit Analysis 1. B/C ratio = 1.90

D-2 Appendix D: Check Answers to Selected End-of-Chapter Exercises

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Glossary

absolute cost advantage A compari- son of nominal costs in two locations, companies, or economies.

adverse selection A limited choice of lower-quality alternatives attributable to asymmetric information.

agency costs Costs associated with resolving conflicts of interest among shareholders, managers, and lenders. Agency costs include the cost of monitoring and bonding performance, the cost of constructing contracts designed to minimize agency conflicts, and the loss in efficiency resulting from unresolved agent-principal conflicts.

allocative efficiency A measure of how closely production achieves the least-cost input mix or process, given the desired level of output.

antitrust laws A series of laws passed since 1890 to limit monopoly power and to maintain competition in most American industries.

appraisal An estimate of value by an independent expert.

arbitrage Buying cheap and selling elsewhere for an immediate profit.

asset specificity The difference in value between first-best and second- best use.

asymmetric information Unequal, dissimilar knowledge.

authorization level Capacity authorized for sale in lower margin segments.

autocorrelation An econometric problem characterized by the existence of a significant pattern in the successive values of the error terms in a linear regression model.

average product The ratio of total output to the amount of the variable input used in producing the output.

backwards induction Reasoning in reverse time sequence from later consequences back to earlier decisions.

barriers to exit Economic losses resulting from non-redeployable assets or contractual constraints upon business termination.

benchmarking A comparison of performance in similar jobs, firms, plants, divisions, and so forth.

benefit-cost ratio The ratio of the present value of the benefits from a project or program (discounted at the social discount rate) to the present value of the costs (similarly discounted).

best-reply response An action that maximizes self-interest from among feasible choices.

brand loyalty A customer sorting rule favorable to incumbents.

break-even analysis A technique used to examine the relationship among a firm’s sales, costs, and operating profits at various levels of output.

break-even sales change analysis A calculation of the percentage increase in unit sales required to justify a price discount, given the gross margin.

business risk The inherent variability or uncertainty of a firm’s operating earnings (earnings before interest and taxes).

capital assets A durable input that depreciates with use, time, and obsolescence.

capital budgeting The process of planning for and evaluating capital expenditures.

capital expenditure A cash outlay designed to generate a flow of future cash benefits over a period of time extending beyond one year.

cartels A formal or informal agreement among firms in an oligopolistic industry that influences such issues as prices, total industry output, market shares, and the division of profits.

ceteris paribus Latin for “all other things held constant.”

chain store paradox A prediction of always-accommodative behavior by incumbents facing entry threats.

G-1

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

class action suit A legal procedure for reducing the search and notification costs of filing a complaint.

closed-end leases with fixed residual values A credible commitment mechanism for limiting the depth of price promotions and the rate of planned obsolescence.

Coase theorem A prediction about the emergence of private voluntary bargaining in reciprocal externalities with low transaction costs.

Cobb-Douglas production function A particular type of mathematical model, known as a multiplicative exponential function, used to represent the relationship between the inputs and the output. coefficient of determination A measure of the proportion of total variation in the dependent variable that is explained by the independent variable(s).

coefficient of variation The ratio of the standard deviation to the expected value. A relative measure of risk.

cointegrated Stochastic series with a common order of integration and exhibiting an equilibrium relationship such that they do not permanently wander away from one another. common-value auction Auction where bidders have identical valuations when information is complete.

company audit A governance mechanism for separating random disturbances from variation in unobservable effort.

complementary goods Complements in consumption whose demand decreases when the price of the focal product rises. complementors Independent firms that enhance the focal firm’s value proposition.

concentrated market A relevant market with a majority of total sales occurring in the largest four firms.

congestion pricing A fee that reflects the true marginal cost imposed by demand in excess of capacity.

consolidated market A relevant market whose number of firms has declined through acquisition, merger, and buyouts.

conspicuous focal point An outcome that attracts mutual cooperation.

contestable markets An industry with exceptionally open entry and easy exit where incumbents are slow to react.

contingent payments A fee schedule conditional on the outcome of uncertain future events.

contracts Third-party enforceable agreements designed to facilitate deferred exchange.

contribution analysis A comparison of the additional operating profits to the direct fixed costs attributable to a decision.

contribution margin The difference between price and variable cost per unit.

cooperative game Game structures that allow coalition formation, side payments, and binding third-party enforceable agreements.

core competencies Technology-based expertise or knowledge on which a company can focus its strategy.

cost fixity A measure of fixed to total cost that is correlated with gross profit margins.

cost function A mathematical model, schedule, or graph that shows the cost (such as total, average, or marginal cost) of producing various quantities of output.

cost of capital The cost of funds that are supplied to a firm. The cost of capital is the minimum rate of return that must be earned on new investments undertaken by a firm.

cost-based strategy A business-level strategy that relies upon low-cost operations, marketing, or distribution.

cost-benefit analysis A resource- allocation model that can be used by public sector and not-for-profit organizations to evaluate programs or investments on the basis of the magnitude of the discounted costs and benefits.

cost-effectiveness analysis An analytical tool designed to assist public decision makers in their resource allocation decisions when benefits cannot be easily measured in dollar terms, but costs can be monetarily quantified.

credible commitment A promise that the promise-giver is worse off violating than fulfilling.

credible threat A conditional strategy the threat-maker is worse off ignoring than implementing.

cross price elasticity The ratio of the percentage change in the quantity demanded of Good A to the percentage change in the price of Good B, assuming that all other factors influencing demand remain unchanged.

cross-sectional data Series of observations taken on different observation units (for example, households, states, or countries) at the same point in time.

cyclical variations Major expansions and contractions in an economic series that usually are longer than a year in duration.

degree of operating leverage (DOL) The percentage change in a firm’s earnings before interest and taxes (EBIT) resulting from a given percentage change in sales or output.

demand function A relationship between quantity demanded and all the determinants of demand.

diseconomies of scale Rising long-run average costs as the level of output is increased.

dividend valuation model A model (or formula) stating that the value of a firm (i.e., shareholder wealth) is equal to the present value of the firm’s future dividend payments, discounted at the shareholder’s required rate of return. It provides one method of estimating a firm’s cost of equity capital.

dominant strategy An action rule that maximizes the decision maker’s welfare independent of the actions of other players.

G-2 Glossary

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

durable goods Goods that yield benefits to the owner over a number of future time periods.

Dutch auctions A descending-price auction.

dynamic pricing A price that varies over time based on the balance of demand and supply, often associated with Internet auctions.

economic profit The difference between total revenue and total economic cost. Economic cost includes a “normal” rate of return on the capital contributions of the firm’s partners.

economies of scope Economies that exist whenever the cost of producing two (or more) products jointly by one plant or firm is less than the cost of producing these products separately by different plants or firms.

efficient rationing A customer sorting rule in which high-willingness-to-pay customers absorb the capacity of low- price entrants.

endgame reasoning An analysis of the final decision in a sequential game.

engineering cost techniques A method of estimating cost functions by deriving the least-cost combination of labor, capital equipment, and raw materials required to produce various levels of output, using only industrial engineering information.

English auctions An ascending-price auction.

expectation damages A remedy for breach of contract designed to elicit efficient precaution and efficient reliance on promises.

expected value The weighted average of the possible outcomes where the weights are the probabilities of the respective outcomes.

experience goods Products and services whose quality is undetectable when purchased.

external diseconomy of scale An increase in unit costs reflecting higher input prices.

externality A spillover of benefits or costs from one production or utility function to another.

fixed costs The costs of inputs to the production process that are constant over the short run.

focal outcomes of interest Payoffs involved in an analysis of equilibrium strategy.

focus groups A market research technique employing close observation of discussion among target consumers.

Folk theorem A conclusion about cooperation in repeated Prisoner’s Dilemma.

forward sales contracts A consensual agreement to exchange goods delivered in the future for cash today, with no possibility of performance excuse.

fragmented market A relevant market whosemarket shares are uniformly small.

free trade area A group of nations that have agreed to reduce tariffs and other trade barriers.

frustration of purpose doctrine An illustration of the default rules of contract law that can result in excusal of contract promises.

full contingent claims contract An agreement about all possible future events.

full-cost pricing A method of determining prices that cover overhead and other indirect fixed costs, as well as the variable and direct fixed costs.

game theory A theory of inter- dependent decision making by the participants in a conflict-of-interest or opportunity-for-collaboration situation.

game tree A schematic diagram of a sequential game.

governance mechanisms Processes to detect, resolve, and reduce post- contractual opportunism.

grim trigger strategy A strategy involving infinitely long punishment schemes.

gross profit margin Revenue minus the sum of variable cost plus direct fixed

cost, also known as direct costs of goods sold in manufacturing.

Herfindahl-Hirschman Index A measure of market concentration equal to the sum of the squares of the market shares of the firms in a given industry.

heteroscedasticity An econometric problem characterized by the lack of a uniform variance of the error terms about the regression line.

hostage or bonding mechanisms A procedure for establishing trust by pledging valuable property contingent on your nonperformance of an agreement.

identification problem A difficulty encountered in empirically estimating a demand function by regression analysis. This problem arises from the simultaneous relationship between two functions, such as supply and demand.

incentive compatibility constraint An assurance of incentive alignment.

incentive-compatible revelation mechanism A procedure for eliciting true revelation of privately held information from agents with competing interests.

income elasticity The ratio of the percentage change in quantity demanded to the percentage change in income, assuming that all other factors influencing demand remain unchanged.

incomplete information Uncertain knowledge of payoffs, choices, and other factors.

incomplete information Uncertain knowledge of payoffs, choices, and so forth.

incremental contribution analysis An incremental managerial decision that analyzes the change in operating profits (revenue – variable costs – direct fixed costs) available to cover indirect fixed costs.

industry analysis Assessment of the strengths and weaknesses of a set of competitors or line of business.

infinitely repeated games A game that lasts forever.

Glossary G-3

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

information technology strategy A business-level strategy that relies on IT capabilities.

inputs A resource or factor of production, such as a raw material, labor skill, or piece of equipment that is employed in a production process.

internal economies of scale Declining long-run average costs as the rate of output for a product, plant, or firm is increased.

internal hedge A balance sheet offset or foreign payables offset to fluctuations in foreign receipts attributable to exchange rate risk.

internal rate of return (IRR) The discount rate that equates the present value of the stream of net cash flows from a project with the project’s net investment.

inverse intensity rationing A custo- mer sorting rule that assures that low-willingness-to-pay customers absorb the capacity of low-price entrants.

iterated dominant strategy An action rule that maximizes self-interest in light of the predictable dominant-strategy behavior of other players.

law of comparative advantage A principle defending free trade and specialization in accordance with lower relative cost. learning curve effect Declining unit cost runs attributable to greater cumulative volume.

lemons markets Asymmetric information exchange leads to the low-quality products and services driving out the higher-quality products and services. life cycle pricing Pricing that varies throughout the product life cycle.

linear incentives contract A linear combination of salary and profit sharing intended to align incentives. long run The period of time in which all the resources employed in a production process can be varied.

maquiladora A foreign-owned assembly plant in Mexico that imports and assembles duty-free components for export and allows owners to pay duty only on the “value added.”

marginal analysis A basis for making various economic decisions that analyzes the additional (marginal) benefits derived from a particular decision and compares them with the additional (marginal) costs incurred.

marginal cost The incremental increase in total cost that results from a one-unit increase in output.

marginal factor cost (MFCL ) The amount that an additional unit of the variable input adds to total cost.

marginal product The incremental change in total output that can be obtained from the use of one more unit of an input in the production process (while holding constant all other inputs).

marginal rate of technical substitution (MRTS ) The rate at which one input may be substituted for another input in producing a given quantity of output. marginal revenue The change in total revenue that results from a one-unit change in quantity demanded.

marginal revenue product (MRPL ) The amount that an additional unit of the variable production input adds to total revenue. Also known as marginal value added. marginal use value The additional value of the consumption of one more unit; the greater the utilization already, the lower the use value remaining. marginal utility The use value obtained from the last unit consumed.

market concentration ratio The percentage of total industry output produced by the 4, 8, 20, or 50 largest firms. maximin strategy A criterion for selecting actions that minimize absolute losses.

maximum sustainable yield (MSY) The largest production harvest that can be

produced by the resource stock as a perpetuity.

minimum efficient scale (MES) The smallest scale at which minimum costs per unit are attained.

mixed bundling Selling multiple products both separately and together for less than the sum of the separate prices.

mixed Nash equilibrium strategy A strategic equilibrium concept involving randomized behavior.

monopolistic competition A market structure very much like pure competition, with the major distinction being the existence of a differentiated product. monopoly A market structure characterized by one firm producing a highly differentiated product in a market with significant barriers to entry.

moral hazard problem A problem of post-contractual opportunism that arises from unverifiable or unobservable contract performance. multicollinearity An econometric problem characterized by a high degree of intercorrelation among some or all of the explanatory variables in a regression equation.

Nash equilibrium strategy An equili- brium concept for nondominant strategy games.

natural monopoly An industry in which maximum economic efficiency is obtained when the firm produces, distributes, and transmits all of the commodity or service produced in that industry. The production of natural monopolists is typically characterized by increasing returns to scale throughout the relevant range of output.

net present value (NPV) The present value of the stream of net cash flows resulting from a project, discounted at the required rate of return (cost of capital), minus the project’s net investment.

network effects An exception to the law of diminishing marginal returns that

G-4 Glossary

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

occurs when the installed base of a network product makes the efforts to acquire new customers increasingly more productive.

noncooperative games Game structures that prohibit collusion, side payments, and binding agreements enforced by third parties.

non-redeployable assets Assets whose value in second-best use is near zero.

non-redeployable reputational asset A reputation whose value is lost if sold or licensed.

non-redeployable specific assets Assets whose replacement cost basis for value is substantially greater than their liquidation value.

normal form of the game A represen- tation of payoffs in a simultaneous- play game.

oligopoly A market structure in which the number of firms is so small that the actions of any one firm are likely to have noticeable impacts on the performance of other firms in the industry.

operating leverage The use of assets having fixed costs (e.g., depreciation) in an effort to increase expected returns.

operating risk exposure A change in cash flows from foreign or domestic sales resulting from currency fluctuations.

opportunity costs The value of a resource in its next best alternative use. Opportunity cost represents the return or compensation that must be forgone as the result of the decision to employ the resource in a given economic activity.

optimal incentives contract An agreement about payoffs and penalties that creates appropriate incentives.

optimal mechanism design An efficient procedure that creates incentives to motivate the desired behavioral outcome.

optimal output for a given plant size Output rate that results in lowest average total cost for a given plant size.

optimal overbooking A marginal analysis technique for balancing the cost of idle capacity (spoilage) against the opportunity cost of unserved demand (spill).

optimal plant size Plant size that achieves minimum long-run average total cost.

optimal plant size for a given output rate Plant size that results in lowest average total cost for a given output.

overall production efficiency A measure of technical and allocative efficiency.

parallel imports The purchase of a foreign export product in one country to resell as an unauthorized import in another country.

participation constraint An assurance of ongoing involvement.

patent A legal government grant of monopoly power that prevents others from manufacturing or selling a patented article.

pecuniary externality A spillover that is reflected in prices and therefore results in no inefficiency.

pooling equilibrium A decision setting that elicits indistinguishable behavior.

post-contractual opportunistic behavior Actions that take advantage of a contract partner’s vulnerabilities and are not specifically prohibited by the terms.

present value The value today of a future amount of money or a series of future payments evaluated at the appropriate discount rate.

price discrimination The act of selling the same good or service, distributed in a single channel, at different prices to different buyers during the same period of time.

price elasticity of demand The ratio of the percentage change in quantity demanded to the percentage change in price, assuming that all other factors influencing demand remain unchanged. Also called own price elasticity.

price leadership A pricing strategy followed in many oligopolistic industries. One firm normally announces all new price changes. Either by an explicit or an implicit agreement, other firms in the industry regularly follow the pricing moves of the industry leader. price signaling A communication of price change plans, prohibited by antitrust law. price skimming A new-product pricing strategy that results in a high initial product price being reduced over time as demand at the higher price is satisfied. principal-agent problem An incentives conflict in delegating decision-making authority. private-value auction Auction where the bidders have different valuations when information is complete. probability The percentage chance that a particular outcome will occur. product differentiation strategy A business-level strategy that relies upon differences in products or processes affecting perceived customer value. production function A mathematical model, schedule (table), or graph that relates the maximum feasible quantity of output that can be produced from given amounts of various inputs. production isoquant An algebraic function or a geometric curve representing all the various combi- nations of two inputs that can be used in producing a given level of output. production process A fixed- proportions production relationship. prospect theory A basis for hypothesizing that the satisfaction from avoiding losses exceeds the anticipation of equal-value prospective gains.

protection level Capacity reserved for sale in higher margin segments.

public goods Goods that may be consumed by more than one person at the same time with little or no extra cost, and for which it is expensive or

Glossary G-5

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

impossible to exclude those who do not pay.

public utilities A group of firms, mostly in the electric power, natural gas, and communications industries, that are closely regulated by one or more government agencies. The agencies control entry into the business, set prices, establish product quality standards, and influence the total profits that may be earned by the firms.

pure competition A market structure characterized by a large number of buyers and sellers of a homogeneous (nondifferentiated) product. Entry and exit from the industry is costless, or nearly so. Information is freely available to all market participants, and there is no collusion among firms in the industry.

random rationing A customer sorting rule reflecting randomized buyer behavior.

real terms of trade Comparison of relative costs of production across economies.

reciprocal externality A spillover that results from competing incompatible uses.

relational contracts Promissory agreements of coordinated performance among owners of highly interdependent assets.

relative purchasing power parity A relationship between differential inflation rates and long-term trends in exchange rates.

relevant market A group of firms belonging to the same strategic group of competitors.

reliance relationships Long-term, mutually beneficial agreements, often informal.

reliant assets At least partially non- redeployable durable assets.

reservation prices The maximum price a customer will pay to reserve a product or service unto their own use.

returns to scale The proportionate increase in output that results from a

given proportionate increase in all the inputs employed in the production process.

revenue management A cross- functional order acceptance and refusal process.

risk A decision-making situation in which there is variability in the possible outcomes, and the probabilities of these outcomes can be specified by the decision maker.

sales penetration curve An S-shaped curve relating current market share to the probability of adoption by the next target customer, reflecting the presence of increasing returns.

search goods Products and services whose quality can be detected through market search.

seasonal effects Variations in a time series during a year that tend to appear regularly from year to year.

secular trends Long-run changes (growth or decline) in an economic time-series variable.

self-enforcing reliance relationship A non-contractual, mutually beneficial agreement. separating equilibrium A decision setting that elicits distinguishable behavior. sequential game A game with an explicit order of play. shareholder wealth A measure of the value of a firm. Shareholder wealth is equal to the value of a firm’s common stock, which, in turn, is equal to the present value of all future cash returns expected to be generated by the firm for the benefit of its owners. short run The period of time in which one (or more) of the resources employed in a production process is fixed or incapable of being varied. simultaneous game A strategy game in which players must choose their actions simultaneously. slippery slope A tendency for wars of attrition to generate mutual losses that worsen over time.

social discount rate The discount rate to be used when evaluating benefits and costs from public sector investments. speculation Buying cheap and selling later for a delayed profit (or loss). spill Confirmed orders that cannot be filled. spoilage Perishable output that goes unsold. spot market transactions An instantaneous one-time-only exchange of typically standardized goods between anonymous buyers and sellers. standard deviation A statistical measure of the dispersion or variability of possible outcomes. standard error of the estimate The standard deviation of the error term in a linear regression model. sterilized interventions Central bank transactions in the foreign exchange market accompanied by equal offsetting transactions in the government bond market, in an attempt to alter short-term interest rates without affecting the exchange rate.

stockouts Demand in excess of available capacity.

strategic holdouts A negotiator who makes unreasonable demands at the end of a unanimous consent process.

strategy game A decision-making situation with consciously interdependent behavior between two or more of the participants.

stratified lottery A randomized mechanism for allocating scarce capacity across demand segments.

subgame perfect equilibrium strategy An equilibrium concept for noncooperative sequential games. substitute goods Alternative products whose demand increases when the price of the focal product rises. sunk cost A cost incurred regardless of the alternative action chosen in a decision-making problem.

G-6 Glossary

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

supply curve A relationship between price and quantity supplied, holding other determinants of supply constant.

supply function A relationship between quantity supplied and all the determinants of supply.

survivor technique A method of estimating cost functions from the shares of industry output coming from each size class over time. Size classes whose shares of industry output are increasing (decreasing) over time are presumed to be relatively efficient (inefficient) and have lower (higher) average costs.

sustainable competitive advantages Difficult to imitate features of a company’s processes or products.

target return-on-investment pricing A method of pricing in which a target profit, defined as the (desired profit rate on investment × total gross operating assets) is allocated to each unit of output to arrive at a selling price.

technical efficiency A measure of how closely production achieves maximum potential output given the input mix or process.

threshold sales curve A level of advance sales that triggers reallocation of capacity.

time-series data A series of observations taken on an economic variable at various past points in time.

transaction risk exposure A change in cash flows resulting from contractual commitments to pay in or receive foreign currency.

translation risk exposure An accounting adjustment in the home currency value of foreign assets or liabilities.

trembling hand trigger strategy A punishment mechanism that forgives random mistakes and miscommunications.

two-person zero-sum game Game in which net gains for one player necessarily imply equal net losses for the other player.

unraveling problem A failure of cooperation in games of finite length.

value at risk The notional value of a transaction exposed to appreciation or depreciation because of exchange rate risk.

value proposition A statement of the specific source(s) of perceived value, the value driver(s), for customers in a target market.

value-in-use The difference between the value customers place on functions, cost savings, and relationships attribut- able to a product or service and the life cycle costs of acquiring, maintaining, and disposing of the product or service.

variable costs The costs of the variable inputs to the production process.

versioning A new product rollout strategy to encourage early adoption at higher prices.

vertical requirements contract A third- party enforceable agreement between stages of production in a product’s value chain.

Vickrey auction An incentive- compatible revelation mechanism for drawing out sealed bids equal to private value.

volume discount Reduced variable cost attributable to larger purchase orders.

winner’s curse Concern about overpaying as the highest bidder in an auction.

Glossary G-7

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Index

A ABC Network, 586 Abegglen, James, 292 Absolute cost advantage, 208 Absolute purchasing power parity, 195–196 Accord (automobile), 4–5, 13 Accounting cost, 276–277 Acquisition cost, 279 Adobe Systems, 560 Advanced Micro Devices (AMD), 383, 458–

459, 591, 618–619 Adverse selection and asset specificity, 376 brand-name reputations as hostages, 373– 374

and hostage or bonding mechanisms, 372–373

and notorious firm, 369–371 price premiums with non-redeployable assets, 374–376

and reliance relationships, 372–373 solutions to, 372–376

Advertising defined, 363, 371 elasticity, 86–87 intensity, 366–367 net value of, 367

Advertising agencies, 366, 560 Aerotek, 432–433 A-frame, 253 Agency conflict, 10 Agency costs, 11 Agency problems, 11–13 Airbus, 188, 309, 412 Aircraft industry break-even analysis in, 318 marginal costs, 309 market share, 412, 414

Aircraft leasing, 478–479

Aircraft production, 237 Airlines barometric price leadership, 430 contestable markets, 614 cost structure, 275–276 economies of scope in, 314–315 entry deterrence, 444–445 hubs, 384 low-cost discounters, 444–445 market share, 414, 533 price signaling, 462 pricing, 416

Airport, 384 Allocative efficiency, 252 Allstate Insurance, 339 Alpha chip technology, 639 Altria, 435 Aluminum space-frame vehicles, 293–294 Amazon.com, 336, 362 American Airlines, 26–27, 275–276, 339,

430, 445, 460, 538–539 American Insurance Group (AIG), 624–625 American West, 275 Amgen, Inc., 14 Amoco, 222 Analog Inc., 598 Analysis of variance, 112–114, 118 Andean Group, 206 Antidumping, 211 Antitrust laws Clayton Act, 615 Federal Trade Commission Act, 615 Hart-Scott-Rodino Antitrust Improve- ment Act, 615–616

Robinson-Patman Act, 615–616 Sherman Act, 614

Antitrust regulations and cross price elasticities, 87 mergers, 617–619

monopolization, 388, 619–620 price fixing, 462, 616 refusals to deal, 622 resale price maintenance agreements, 622–623

wholesale price discrimination, 620–622 APEC, 207 Apple Computer, 9, 214, 334–335, 386, 470,

499–500, 599–600, 603, 639 Apple stores, 9, 499 Appraisals, 597 ARAMCO, 422 Arbitrage, 191 Arc income elasticity, 84–85 Arc price elasticity of demand, 72 Archer-Daniels-Midland (ADM), 418 Argentina, 206 ASEAN, 207 Asking price, 28–29 Asset specificity, 376 Assets, dissolution of, 574–575 AstraZeneca, 345, 635 Asymmetric information, 17, 368, 584–587 Atlantic Cement Co., Inc., 630 Auctions

asymmetric information bidding games, 584–587

common-value, 587–588 Dutch, 584, 594–595 English, 584, 594–595 first-price sealed-bid, 594–595 incentive-compatible revelation mecha- nism, 592–594

information revelation in, 587–588 online, 582–583, 596 open bidding design, 588–589 optimal mechanism design, 580 price-value, 589–592 queue service rules, 580

I-1

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Auctions (continued ) revenue equivalence of alternative types, 594–596

second-highest sealed-bid, 592–594 types of, 583–584 underbidding in, 589–592 Vickrey, 592–594 winner’s curse, 584–587

Authorization level, 536 Auto inspections, 630 Autocorrelation, 126–128 Automobile assembly plants, 241, 664 dealerships, 70, 84, 374 demand, 116–117 excess production capacity, 489 rental, 67–68, 414–415, 516

Aventis, 635 Average cost function, 281–286, 312 Average fixed cost (AFC), 281–285 Average product, 236, 239–241 Average profit, 43–45 Average total cost (ATC), 281–285, 316 Avis, 414–415

B Backwards induction, 457, 466 Balance of payments, 222–223 Baltimore Orioles, 539 Bananas, 212–213 Bank mergers, 254 Bank of Japan, 190 Banks economies of scale, 291 economies of scope, 315 technical/allocative efficiency, 254

Bargaining, impediments to, 628–629 Barometric price leadership, 430 Barometric techniques, 154–155 Barriers to entry, 48, 343–346, 613 Barriers to exit, 350 Baseball, 257–259, 429, 539 Bayesian reputation effects, 459 Beautiful Mind, A (Nasar), 448 Beecham v. Europharm, 221 Beer industry, 439 Beetle (automobile), 79 Bell Sports, 493–494 Benchmarking, 559 Benefit-cost ratio, 659 Berkshire Hathaway, Inc., 8, 478 Bertelsmann Inc., 617 Best Buy, 471–472, 496 Best-reply response, 453 Big Mac index, 201 Biotech stocks, 55 Birkenstocks, 196 Block booking, 376 Blockbuster, Inc., 364–365

Blue Chip Economic Indicators, 162–163 Bluefin tuna stocks, 268 Blu-ray, 238 Boeing, 179, 188, 237, 309, 412 Bolivia, 206 Bonds, 50, 55 Boomer v. Atlantic Cement Co., Inc., 630 Boston Scientific, 502 Branch Banking and Trust, 254 Brand loyalty, 438, 491 Brand names, 373–374 Brazil, 206 Break-even analysis in aircraft industry, 318 algebraic method, 319–320 and composition of operating costs, 323– 324

vs. contribution analysis, 323 contribution margin, 320 defined, 317 graphical method, 318–319 and inconsistency of planning horizon, 324

limitations of, 323–324 linear chart, 319, 321–322 and multiple products, 324 and risk assessment, 326–327 and uncertainty, 324

Break-even sales change analysis, 348 Bristol-Myers-Squibb, 398, 525 British Petroleum, 222 British Telephone, 526 Brooke Group Ltd. vs. Brown and

Williamson Tobacco Company, 460 Brown and Williamson Tobacco Company,

460 Budget, 82 Budget constraint, 67 Buffett, Warren E., 8 Bundling, 518–521 Bureau of Alcohol, Tobacco, and Firearms

(BATF), 624 Business cycles, 154–155 Business risk, 326 Business trip, consumption choices on, 66–

68 Business-to-business (B2B) transactions, 528 Buyers, power of, 346–347

C Cable television industry, 313, 343, 567 Cadbury Schweppes, 351 Caduet (drug), 525 California, deregulation of electricity in,

230–232 California Public Utility Commission, 231–

232 Call centers, 219 Camry (automobile), 4–5, 13

Canada, 206, 218 Cap and trade, 610–611, 632–633 Capacity expansion, 4–5 Capacity reallocation, 534–536 Capital assets, 277 Capital budgeting basic framework, 647 defined, 647 estimating cash flows, 649–650 evaluating and choosing investment projects, 650–653

generating capital investment projects, 648 and marginal analysis, 42 process, 647–653

Capital elasticity, 257 Capital expenditure, 647 Capital investment projects, 648 Capoten (drug), 398, 525 Carbon dioxide (CO2) emissions, 95–97 Carburetors, 208–209 Carnival Cruise Lines, 451–452, 460–462 Cartels allocation of restricted output, 421–426 analysis, 426–429 collusive agreements, 417–418 defined, 417 OPEC, 422 price-output determination, 421 profit maximization, 421–426

Cash flow estimation, 649–650 “Cash for Clunkers” program, 83 CBS, 586 Celera Genomics, 637 Cell phones, 479 Cellophane, 87 Cereal industry, 367, 414 Chain store paradox, 457–459 Chanel No. 5, 394 Chase Econometric Associates, 162 ChemChina, 186 Chevron Corp., 305 Chevy Volt (automobile), 43, 69 Chicago Board of Trade, 610 Chicago Pacific Corporation, 605 Chief executive officers (CEOs), compensa-

tion of, 11–12 China exports, 185–186 foreign investments in, 185–186 gross domestic product, 176, 185–186 gross national product, 177 growth rate, 191 Hangzhou region, 187 imports from United States, 175–178 infrastructure investments, 187 international trade, 185 property rights, 186–187 Shanghai region, 187 trading partners, 185

I-2 Index

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

China International Trust and Investment Corp., 184

China Mobile, 186 Chinese yuan, 176 Chrysler Corp., 69, 648–649 Cifunsa SA, 219 Cigarette taxes, 62–63 Circuit City, 496 Cisco, 476–477, 560 Claritin (drug), 398–399 Clarke-Groves incentive-compatible revela-

tion mechanism, 601–602 Class action suits, 628–629 Clayton Act, 615 Clean Air Act of 1990 (CAA), 2, 16, 95, 610,

632 Climate change, 95–97 Closed-end leases with fixed residual values,

480 Coal power plant, 3–4 Coase, Ronald, 610 Coase theorem, 628–629 Coasian bargaining, 626–628 Cobb-Douglas production function, 233,

255–256 Coca-Cola Company, 179, 215, 346, 348,

435, 453–454 Coefficient of variation, 53–54 Coffee price elasticity for, 80 pricing agreement, 423

Coincident indicators, 154–157 Cointegrated stochastic series, 166 Colgate-Palmolive, 185 Collusion cost structures, 419–421 number and size distribution of sellers, 419–421

percentage of external output, 420 product heterogeneity, 419–421 size and frequency of orders, 419–420 threat of retaliation, 420

Colombia, 206 Color television, 138–139 Common-value auctions, 587–588 Community banks, 291 Compact fluorescent light (CFL) bulbs,

47–48 Company audit, 559 Compaq Computer, 334, 336 Comparative advantage, 207–209 Competitive markets adverse selection, 369–371 incomplete vs. asymmetric information, 368

lemons markets, 368, 371–372 notorious firm, 369–371 search goods vs. experience goods, 368–369

Competitive strategy, 336–337 Complementary goods, 32–33, 87, 110–111 Complementors, 343 Computer memory chips, 208–209 Computer price index, 361 Computer Tree, 9 Concentrated market, 341 Conference Board, 155, 158 Congestion pricing, 513–514 Consensus forecasts, 162–163 Consolidated market, 341 Conspicuous focal point, 461 Constant rate of growth trends, 145–146 Constant-cost industry, 360 Consumer behavior and real income effect, 65–66 and substitution effect, 66–69 surveys, 159 targeting, 69

Consumer expenditure plans, 159 Consumer focus groups, 98 Consumer price index (CPI), 38 Consumer Product Safety Commission

(CPSC), 624 Containers (shipping), 581 Contestable markets, 353, 493–494, 614 Continental Airlines, 275, 430, 526–527 Contingent payments, 597–598 Contracts commercial, 549–552 in cooperative games, 547–549 defined, 549 forward sales contracts, 552 frustration of purpose doctrine, 550–551 full contingent claims, 553 functions of, 549–552 governance mechanisms, 553–556 incentive-compatible, 604 incentives, 561–564 incomplete information, 553 linear incentives, 563 moral hazard problem, 554–555 optimal incentives, 602–604 post-contractual opportunistic behavior, 553

spot market transactions, 551–552 vertical requirements, 549

Contribution analysis vs. break-even analysis, 323 and composition of operating costs, 323– 324

defined, 323 and inconsistency of planning horizon, 324

limitations of, 323–324 and multiple products, 324 and uncertainty, 324

Contribution margins, 320, 349, 365, 393 Cooperative game, 449, 547–549

Coors, 439 Copenhagen Summit on Climate Change, 97 Copper prices, 360 Copyright, 383–384 Cordis, 502 Corporate bonds, 55 Corporate governance, 553–556 Corporate restructuring, 11 Correlated macroeconomic shocks, 217 Correlation analysis, 99 Correlation coefficient, 111–112 Cost analysis

in airlines, 275–276 depreciation cost measurement, 277–279 inventory valuation, 279–280 overview, 275 sunk cost of underutilized facilities, 280– 281

Cost fixity, 347 Cost functions

average, 281–286 defined, 281, 306 estimating, 306–317 long-run, 286–287, 310 short-run, 281–285, 310–311

Cost inflation, 195 Cost leadership strategy, 340 Cost of capital

debt capital, 654 defined, 653 external equity capital, 656 internal equity capital, 655–656 weighted, 657

Cost overruns, 599–600 Cost-based strategy, 339 Cost-benefit analysis

accept-reject decisions, 658–659 constraints, 661–662 defined, 19, 658 of direct benefits, 662 of direct costs, 663 of indirect costs or benefits and intangi- bles, 663

objectives, 660–662 program-level analysis, 659 schematic of, 661 steps in, 660

Cost-effectiveness analysis defined, 664–665 least-cost studies, 665 objective-level studies, 665–666

Cost-output relationship controlling variables in, 307–308 empirical, 308–309 polynomial, 309

Costs accounting, 276–277 acquisition, 279 depreciation, 277–279

Index I-3

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Costs (continued ) direct, 307, 663 economic, 276–277 fixed, 281 indirect, 663 indirect fixed, 307 marginal, 283–285 measuring, 307 minimizing subject to output constraint, 249–251

opportunity, 277–279 overhead, 290 primary, 663 replacement, 279 revelation, 599 sunk, 280–281 transportation, 291 variable, 281

Couponing, 517–518 Cournot oligopoly model, 415–417 Credible commitments, 471, 476 Credible threats, 471 Credit rationing, 372 Credit threats, 471 Crocs shoes, 140 Cross price elasticity of demand and antitrust, 87 combined use with demand elasticities, 89–90

defined, 87 electricity use, 89 estimating, 98 negative, 87 positive, 87

Cross-functional revenue management, 531–532

Cross-sectional data, 141, 256 Crude oil cartel, 422 input costs, 358 price-output determination, 424–425 prices, 37–40, 272–273, 358 proven reserves, 427 refineries, 245–246 Saudi Arabian production, 39, 425–426 and U.S. trade deficit, 223 years of proven reserves, 273–274

Cruise lines, 451–452, 460–462 Cummins Engine Company, 34–35, 181–

183, 192, 223 Current status loans, 254 Customer loyalty programs, 69 Customer segmentation, 435 Customer sorting rules brand loyalty, 491 in contestable markets, 493–494 efficient rationing, 491 inverse intensity rationing, 491 random rationing, 491–492

sunk costs in decision making, 492 wars of attrition, 494–495

Cyclical variations, 141 CYPHER stent, 502

D Daimler-Benz, 34 Data collection, 99 Data processing jobs, 219 De Beers Consolidated Mines Ltd., 376,

383–384 Debt capital, cost of, 654 Debt securities, 55 Decision tree, 466 Decision-making management responsibilities in, 5–6 model, 5

Declining rate of growth trends, 146 Deferred stock, 11 Degree of operating leverage (DOL), 324–

326 Dell Computer, 9, 334, 336–337, 339–340,

385, 560, 573 Delta Airlines, 339 Demand combined effect of elasticities, 89 cross price elasticity of, 87–89 elastic, 75–76 and equilibrium market price, 27–29 and exchange rate, 34–35 factors and expected effect, 33 income elasticity of, 83 inelastic, 74–75 price elasticity of, 69–83 and real income effect, 65–66 shifts in, 33 and substitution effect, 66–69 unit elastic, 76

Demand and supply, 27–29 Demand curves defined, 31–32 for gasoline, 64–65 in oligopoly, 432–433 perfectly elastic, 74

Demand function defined, 32, 62 linear model, 100 multiplicative exponential model, 100–101 statistical estimation of, 99–101

Demand schedule. See Demand curves Dense wave divisional multiplexing

(DWDM), 138 Depreciation, 307 Depreciation cost measurement, 277–279,

307 Deregulation, 623–626 Designer jeans, 110–111 Deterministic trend analysis components of time series, 141

cross-sectional data, 141 seasonal variations, 146–147 secular trends, 143–146 time-series data, 141 time-series models, 142–143

Diamonds, 28, 55, 376, 384 Diamond-water paradox, 30–31 Diesel engines, 34–35 Digital camera market, 437 Digital Equipment Corp., 639 Digital moviehouse projectors, 478 Digital signal processor (DSP) chips, 598 Digital TV, 546–547 Direct benefits, 662 Direct cost of goods sold (DCGS), 395 Direct costs, 307, 663 Direct mail coupon, 84 Direct segmentation, 513–515 DIRECTV, 620 Discount coupons, 67–68 Discount rate, 663–664 Diseconomies of scale defined, 291 effects of, 293–295 external, 359

DISH Network, 620 Dish soap, 203–204 Disney World, 237, 383, 517 Dividend valuation model, 655–656 Doha Round, 82 Dominant price leadership, 430–432 Dominant strategy, 447, 452–453 Dooney & Burke, 373 Double-log transformation, 133 Double-the-difference price guarantees,

472–474 Drug stores, 570 Drugs, patented, 635 Duke Power, 3, 16 Dulles Toll Road, 512 Dummy variables, 147–148 Duopoly, 412 DuPont, 87, 183, 192 Durable goods, 68 Durbin-Watson statistic, 132–134 Dutch auction, 594–595 Dutch auctions, 584 Dynamic pricing, 528 Dynamic random access memory chips

(DRAMs), 213, 420

E eBay, 596 EchoStar, 620 Eckerd, 570 Econometric models advantages of, 159–160 defined, 159 multi-equation models, 160–162

I-4 Index

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

single-equation models, 160–161 of U.S. economy, 162

Econometrics, 99 Economic cost, 6, 276–277 Economic indicators, 154–157 Economic profit, 6, 277 Economies of scale in cable industry, 313 vs. economies of scope, 314 firm-level internal, 290–291 internal, 287 and international joint ventures, 568 learning curve effect, 288–289 mass customization, 288–289 money-center banks vs. community banks, 291

and monopolies, 384 overall effects of, 293–295 percentage of learning, 287–288 plant-level internal, 290 product-level internal, 287–288 volume discounts, 288

Economies of scope in banking industry, 315 defined, 314 vs. economies of scale, 314

Ecuador, 206 Effective exchange rate (EER), 201–202 Efficient rationing, 491 Eisner, Michael, 10 Elastic demand, 75–76 Electric utilities average cost function, 312 deregulation, 230–232 dynamic pricing, 515 electricity-use elasticities, 89 green power technologies, 230–232 long-run cost functions, 312 regulation of, 397–398 short-run cost functions, 311 sustainability, 2–4 underutilized, 287

Electrical industry, 463–464 Electronic commerce, 340–341 Eli Lilly, 14, 222, 345, 398–399, 525, 635 Embedded options, 15 Embedded real options, 15 Emerging market stocks, 55 Emerson Corporation, 289 EMI Group, 617 Emissions trading, 610–611 Employment Standards Administration, 624 Endgame reasoning, 456–457 Engineering cost techniques, 314–315 English auctions, 584, 594–595 Enron, 556 Enterprise Rent-A-Car, 414–415 Entry deterrence customer sorting rules, 491–495

excess capacity as credible threat, 488–489 and incumbent price commitment, 468, 469, 470

pre-commitment using non-redeployable assets, 488–491

Environmental Protection Agency (EPA), 2–3, 16, 610, 624, 632–633

Epogen (drug), 14 Equilibrium market price, 27–29 of gasoline, 36–40 and incremental costs, 30–31 and marginal utility, 30–31

Equimarginal criterion, 249–250 ESPN, 586 Euro, 188 European Commission (EC), 221–222 European Union, 214–216 vs. NAFTA, 219–220

Evista (drug), 345 Excess capacity, 488–490 Exchange rates and demand, 34–35 effective, 201–202 and import-export sales, 179–183 operating risk exposure, 181 transaction risk exposure, 180 translation risk exposure, 181

Excise taxes, 63 Executive compensation, 11–12 Executive performance pay, 12 Exhaustible natural resources, 270–272 Exide Batteries, 222 Expansion, 4–5 Expected value, 50 Experience goods, 368–370, 375 Exponential smoothing, 150–151 Exponential Valley Inc., 591 Exports, 191–193 External diseconomy of scale, 359 External equity capital, 656 Externalities Coase theorem, 628–629 Coasian bargaining, 626–628 defined, 626 impediments to bargaining, 629 pecuniary, 626 reciprocal, 626–628 resolution by regulatory directive, 629–630 resolution by sale of pollution rights, 632–

633 resolution by taxes and subsidies, 630–632

Exxon Mobil, 10, 39, 414–415

F Farmlands, 55 Fast-second advantage, 469–470 Federal Aviation Administration (FAA), 624 Federal Communications Commission

(FCC), 400, 585–586, 588–590, 624

Federal Energy Regulatory Commission (FERC), 624

Federal Reserve Bank of Philadelphia, 162–163 Federal Reserve System, 624 Federal Trade Commission Act, 615 Federal Trade Commission (FTC), 88, 620 FEMSA, 615 Fiber optic networks, 137–139 Firm-level internal economies of scale, 290–291 First-come, first-served rule, 581 First-mover advantage, 469–470 First-order exponential smoothing, 150–151 Five Forces strategic model

intensity of rivalrous tactics, 347–351 myth of market share, 351 power of buyers and suppliers, 346–347 threat of entry, 343–346 threat of substitutes, 342–343

Fixed costs, 281, 322 Fixed input, 234–235 FlexJets, 478 Focal outcomes of interest, 469 Focus groups, 98 Folk theorem, 455 Food and Drug Administration (FDA), 345,

624 Food Safety and Inspection Service, 624 Ford Expedition (sport utility vehicle), 476–

477 Ford Explorer (sport utility vehicle), 65 Ford Motor Co., 36, 70, 79, 139, 222, 237,

292–294, 366 Forecasting techniques

accuracy of models, 140 barometric techniques, 154–155 deterministic trend analysis, 141–147 econometric models, 159–163 input-output analysis, 166 selection criteria, 140 smoothing techniques, 147–153 stochastic time-series analysis, 163–166 survey and opinion-polling techniques, 155–159

Forecasts, hierarchy of, 139–140 Foreign exchange

arbitrage, 191 coordinated intervention, 189–190 and government transfers, 189–190 and import-export sales, 189 and inflation, 194–195 and interest rates, 194 long-term exchange rate trends, 191–195 managed float, 187 market for U.S. dollars as, 187–188 operating risk exposure, 181 and real growth rates, 191–194 risk, 180–181 short-term exchange rate fluctuations, 190–191

Index I-5

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Foreign exchange (continued ) speculation in, 191 and speculative demand, 189–190 sterilized interventions, 190 and transaction demand for currency, 189 transaction risk exposure, 180 translation risk exposure, 181

Forward market, 16 Forward sales contracts, 552 Fountain drink, 348 Fragmented market, 341 France, 219–220 Franchise, government-authorized, 384 Free trade, 82, 207–209 Free trade areas correlated macroeconomic shocks in, 217 defined, 216 European Union, 214–216 intraregional trade, 216 mobility of labor, 216–217 North American Free Trade Agreement, 214–216

optimal currency areas, 216 value at risk, 216

Freightliner, 218 Frequent buyers, 69 Frustration of purpose doctrine, 550–551 Fuel surcharges, 26–27 Full contingent claims contract, 553 Full-cost pricing, 525–526 Full-line pricing, 568–570 Futures market, 16

G Gain-share discounting, 347 Game theory analysis of, 446–447 best-reply response, 453 credible threats and commitments, 471 dominant strategy, 447, 452–453 managerial purpose of, 445–446 mechanisms of establishing credibility, 471

mixed Nash equilibrium strategy, 454 Nash equilibrium strategy, 453–454 and oligopolistic rivalry, 445–446 Prisoners’ Dilemma, 450–452

Game tree, 466 Game(s) components of, 447–448 cooperative, 449 iterated dominant strategy, 448 maximin strategy, 451 noncooperative, 449 normal form of, 447 repeated, 449 sequential, 451 simultaneous, 450–454 single-period, 449

strategy, 446–447 two-person, 449–450

Gas stations, 37, 358, 426 Gasoline, 36–40 and crude oil input costs, 358 demand schedule, 64–65 price components, 426–427 price elasticity of demand for, 71 prices, 358 U.S. household demand for, 64

Gateway Inc., 340 Genentech, 388 General Agreement on Tariffs and Trade

(GATT), 82 General Electric, 12, 464 General Mills, 367, 436 General Motors, 13, 79, 167, 199, 210, 253,

322, 476 Generic drugs, 398–399 Germany, 219–220 Gillette Co., 179, 438 Global Crossing Inc., 102, 137–138 Global warming, 95–97 Goods complementary, 32–33, 87, 110–111 durable, 68 experience, 368–370, 375 import-export traded, 34 search, 368–369 substitute, 32, 82, 110–111

Governance mechanisms, 553–556 Governmental protection of business, 633–634 Grand Cherokee (sport utility vehicle), 648 Gray markets, 220–222 Green power technologies, 230–232 Greenhouse gases, 644–646 Grim trigger strategy, 455 Gross domestic product (GDP), 139, 175–

177, 191–193, 202–203 Gross point margin, 394–396 Guttenberg, Johannes, 28 Guttenberg Bible, 28

H Hadley v. Baxendale, 550 Hamilton Beach/Proctor-Silex, Inc., 650–651 Handbag market, 373 Hanes Corp., 349 Hangzhou, China, 187 Hardwoods, 29 Harsanyi, John, 448 Hart-Scott-Rodino Antitrust Improvement

Act, 615–616 Health care reform, 62–64 Health maintenance organizations (HMOs),

346 Hedges balance sheet, 181, 227–228 covered, 227–228

financial, 180, 227–228 internal, 212, 227–228 operating, 227–228

Heinz Ketchup, 215 Herfindahl-Hirschman Index (HHI), 617–

618, 620 Hershey Foods, 652 Hertz, 414–415 Heteroscedasticity, 126–127 Hewitt Associate LLC, 183 Hewlett-Packard, 183–184, 334, 385, 412, 639 High-definition DVD, 238 Highway Safety Act of 1996, 79 Hitachi Ltd., 292 Home prices, 360 Hon Industries, 351 Honda Motors, 4–5, 210, 490 Hostage mechanism, 372–373, 471, 475–476 Hughes-JVC, 477 Human Genome Sciences, 637 Hydrochlorofluorocarbon (HCFC), 644 Hynix, 420 Hyundai, 489

I IBM Corp., 180, 183, 185, 290, 334, 336, 386,

388, 470, 477, 603, 634 Immelt, Jeff, 12 Import controls, 209–211 Import-export sales, 179–183 Import-export traded goods, 34 Incentive compatibility constraint, 561 Incentive-compatible revelation mechanism Clarke-Groves, 601–602 in contracts, 604 cost revelation, 599 defined, 592, 598

Income elasticity of demand arc, 84–85 defined, 83 electricity use, 89 estimates, 86 interpreting, 85–86 point, 85 use with demand elasticities, 89–90

Income-superior products, 83 Incomplete information, 368, 553 Increasing returns, 213–214 and information services, 239 with network effects, 237–239, 384–388 and total product function, 240

Increasing-cost industry, 360 Incremental contribution analysis, 526–527 Independent software vendors (ISVs), 386 Index of Leading Economic Indicators, 158 Indirect benefits, 663 Indirect costs, 663 Indirect fixed costs, 307 Indirect segmentation, 515

I-6 Index

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Individual demand, 31–32 Industry analysis, 337 Industry standards, 463 Inelastic demand, 74, 75 Infant industries, 211 Infineon Technologies, 420 Infinitely repeated games, 456 Inflation, 194–195 Information services, 239 Information technology, 305–306 Information technology strategy, 339–341 Ingram Book Group, 362 Innovation theory of profit, 7 Innovations, 437–439 Input-output analysis, 166 Inputs defined, 232–233 fixed, 234–235 optimal combination of, 248–249 variable, 234–235

Insurance policies, 371–372 Intangibles, 663 Intel Corp., 213, 290, 382–383, 458–459,

598, 618–619 Intensity of rivalry, 347–351 Interest rates, 158, 194 Interlink Surgical Steel, 438 Internal economies of scale, 287 Internal equity capital, cost of, 655–656 Internal hedge, 212 Internal rate of return (IRR) defined, 651 vs. net present value, 653

International Monetary Fund, 176 International trade absolute cost advantage, 208 comparative advantage, 207–209 free trade, 207–209 free trade areas, 214–217 gray markets, 220–222 import controls, 209–211 increasing returns, 213–214 internal hedge, 212 knockoffs, 220–222 network externalities, 214 parallel importing, 220–222 real terms of trade, 208–209 regional trading blocs, 204–207 shares of, 204–205 strategic trade policy, 211–213 tariffs, 209–211 trade deficit, 222–223

Internet, 138–139, 527–529 Internet auctions, 582–583, 596–598 Internet Explorer, 238, 344, 388 Interstate Commerce Commission (ICC),

624 Intraregional trade, 216 Inventory, 159

Inventory valuation, 279–280 Inverse demand function, 100 Inverse intensity rationing, 491 Investigational new drugs (INDs), 14 Investment net cash flow, 49 net present value, 46–47

Investment analysis capital budgeting, 647–653 capital expenditure, 647 cost of capital, 653–657 cost-benefit analysis, 658–663 cost-effective analysis, 664–666 discount rate, 663–664

Investment opportunity curve, 647 Isocost lines, 248–249 Iterated dominant strategy, 448 iTunes, 617

J Japan bilateral trade with United States, 208– 209

gross domestic product, 176 Japanese corporations, 292 J.D. Powers, 13 Jeans, 110–111 Jet fuel surcharges, 26–27, 276 JetBlue, 445 Jiangsu Changzhov Pharmaceutical, 222 Jobs, Steve, 9, 335 Joint ventures, 568, 599, 603 Jojoba beans, 29 J.P. Morgan, 526

K Kahnemen, Daniel, 568 Kaldor-Hicks criterion, 660 Keebler’s, 412 Kellogg’s, 179, 367, 396, 437 Knockoffs, 220–222 Knowledge capital, 636 Kodak, 10, 437, 460, 566 Kohlberg Kravis Roberts & Co., 10 Krispy Kreme, 14

L La Quinta Motels, 232 Labor costs, 36 mobility of, 216–217

Lagging indicators, 154–157 Last-come, first-served rule, 581 Law of comparative advantage, 207–209 Law of diminishing marginal returns, 236–

237 Leading indicators, 154–158 Learning curve effect, 288–289 Lease prices, 480

Least-cost studies, 665 Lee jeans, 110–111 Leegin Creative Leather Products, 623 Lemons markets, 367, 371–372 Leveraged buyouts (LBO), 11 Lexus (automobile), 69 Licensing, 633–639 Life cycle pricing, 523–525 Limit pricing, 396–399, 524 Linear incentives contract, 563 Linear model, 100 Linear regression model

autocorrelation, 126–128 heteroscedasticity, 126–127 identification problem, 130–132 multicollinearity, 130 specification and measurement errors, 129

Linear regression relation, 103 Linear trends, 144–145 Long run, 235, 358–360 Long-run average cost (LRAC), 364 Long-run cost functions

in electricity generation, 312 optical capacity utilization, 286–287 statistical estimation of, 310

Long-run costs with Cobb-Douglas function constant returns, 302 decreasing returns, 301–302 equations, 301–302 increasing returns, 303–304

Long-Term Capital Management (LTCM), 53

Lotteries, 582 Low-cost airlines, 444–445 Lucent Technologies, 637–639 Lufthansa Airfreight, 182

M Macroeconomic forecasting, 158–159 Macroeconomic shocks, 217 Major league baseball, 257–259, 429 Managed float, 176, 187 Management, decision-making in, 5–6 Managerial contracting

alternative hiring arrangements, 557–558 moral hazard problem, 558–561 optimal incentives contracts, 561–564

Managerial economics, 4 Managerial efficiency of profit, 7 Manufacturer’s suggested retail price

(MSRP), 622–623 Manufacturing industries

Cobb-Douglas function, 256 cross-sectional analysis of, 256 production elasticities, 257

Maquiladora, 218 Marginal analysis

and capital budgeting, 42

Index I-7

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Marginal analysis (continued ) defined, 41 Mini Cooper vs. Chevy Volt, 43 and resource allocation, 41 Sara Lee Corp., 42 Tenneco shipyard, 42

Marginal costs, 283–285 of advertising, 365 of capital curve, 647 defined, 31 and marginal benefits, 41

Marginal factor cost, 242–243 Marginal private cost (MPC), 631 Marginal product, 235–236, 239–241 Marginal profit, 43–45 Marginal profit contribution, 365 Marginal rate of technical substitution, 245–

247 Marginal return, 41 Marginal revenue, 74, 76–78, 355–356, 364–

365 Marginal revenue product, 242–243 Marginal use utility, 30 Marginal use value, 30 Market barriers to entry, 613 competitive, 368–372 conduct, 612 contestable, 353, 493–494, 614 performance, 611–612 structure, 612–613

Market concentration ratio, 617 Market experiments, 99 Market share myth of, 351 and oligopoly, 411–415

Market structure contestable markets, 353 defined, 361 monopolistic competition, 354 monopoly, 353–354 oligopoly, 355 pure competition, 352–353

Marketing research consumer focus groups, 98 market experiments, 99 marketing experiments, 98

Marriott Corp., 437 Marshall, Alfred, 30 Mass customization, 288–289 Mass transit, 95–97 Matsushita Electrical Industrial Company,

238, 292 Mavica (digital camera), 437 Maximin strategy, 451 Maximum sustainable yield (MSY), 269–270 Maytag, 605 McDonald’s, 178–179, 192, 520–521 McGraw-Hill survey, 158–159

Measurement errors, 129 Medtronic, 502 Memory chips, 208–209, 213, 603 Mercedes-Benz, 181–182 Merck, 183, 221, 306 Merck vs. Primecrown, 221 MERCOSUR, 206–208 Mergers, 619 Herfindahl-Hirschman Index, 617–618 market concentration ratio, 617

Mexico import-export trade, 206 maquiladora, 218

Microprocessors, 382–383, 458–459 Microsoft Corp., 179, 185, 214, 238, 334,

344, 385–386, 388, 477, 546–547, 561, 567, 616, 621

Miller Brewing Co., 615 Mine Safety and Health Administration, 624 Mini Cooper (automobile), 43, 476–477 Mini-mills, 212 Minimum efficient scale (MES), 293–295 Minivans, 69 Mixed bundling, 520–521 Mixed Nash equilibrium strategy, 454 Modelo, 615 Monday Night Football, 586 Moneyball (Lewis), 257–259 Money-center banks, 291 Monopolistic competition defined, 354 long run, 362–363 price-output determination under, 361– 363

short run, 362 Monopoly antitrust regulations, 619–620 capacity investments, 396 characteristics of, 353–354 contribution margin, 393 control of critical resources, 384 and copyright, 383–384 credible commitments, 476 defined, 353, 383 economies of scale, 384 and government-authorized franchise, 384

gross point margin, 394–396 increasing returns from network effects, 384–388

limit pricing, 396–397 natural, 401–402 and patent, 383–384 price and output determination in, 388– 392

regulated, 397–400 sources of market power, 383–388 value proposition, 393–394

Monopoly theory of profit, 7

Monster wafers, 290 Morgenstern, Oskar, 448–449 Moroney, John, 256 Motorola, 599–600, 603, 637–639 Movie theatres, 477 Moving averages, 148–151 Multicollinearity, 130 Multi-equation models, 160–162 Multiple linear regression model computer programs, 115 defined, 114 forecasting with, 115 inferences about population regression coefficients, 115–116

population regression coefficients, 115 Multiple regression analysis, 308 Multiple-product pricing, 506–507 Multiplicative exponential model, 100–101 Music CDs, 222

N Nabisco, 412 Nader, Ralph, 79 Nash, John, 448–449 Nash equilibrium strategy, 453 National Association of Purchasing Agents

survey, 159 National Bureau of Economic Research, 154 National Football League (NFL), 160–161 National Highway Traffic Safety Adminis-

tration (NHTSA), 624 National Industrial Conference Board, 158 National Labor Relations Board (NLRB), 624 Natural gas companies, 399–400 Natural monopolies, 401–402 Natureview Farms Yogurt, 393–395 NBC, 586 NEC, 458–459 Negative correlation, 126–127 Nestlé, 82 Net cash flow (NCF), 49–50, 649 Net income after tax (NIAT), 649–650 Net income before tax (NIBT), 649 Net investment, 649 Net present value (NPV) defined, 45, 652 determining, 46–47 vs. internal rate of return, 653 positive, sources of, 48 and risk, 48–49

NetJets, 477–478 Netscape, 238, 388, 621 Network effects, 237–239, 343, 384–388 Network externalities, 214 Niche pricing, 524–525 Nitrous oxide (NOX) emission, 2–3 Nokia, 409–410 Noncooperative game, 449 Noncredible commitments, 474

I-8 Index

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Nonlinear regression models double-log transformation, 133 polynomial transformation, 134 reciprocal transformation, 133–134 semilogarithmic transformation, 133

Nonproduction worker elasticity, 257 Non-redeployable assets, 374–376, 488–491 Non-redeployable reputational asset, 471 Non-redeployable specific assets, 564–565 Normal form of the game, 447 Normal probability distribution, 51–53 North American Free Trade Agreement

(NAFTA), 214–216 and automobile asssembly, 36 vs. European Union, 219–220 exports, 217–218 members of, 206 non-tariff trade barriers, 218 and price elasticity of demand, 82 single currency in, 217

Northwest Airlines, 462 Not-for-profit (NFP) organizations characteristics of, 18 efficiency objective in, 19 objectives, 18 services, 18

Notorious firm, 369–371 Novasc (drug), 525 Nuclear Regulatory Commisssion (NCC),

624

O Objective product differentiation, 346 Objective-level studies, 665–666 Occupational Safety and Health Adminis-

tration (OSHA), 624 Offce of the Comptroller of Currency, 624 Offer price, 28–29 Office Depot, 88, 345 Office Max, 345 Office supply products, 88 Office supply retailers, 88, 345 OfficeMax, 88 Ole Musk, 394 Oligopoly cartels, 417–429 collusion factors, 419–421 Cournot model, 415–417 defined, 355 and extreme brand loyalty, 438 and game theory, 445–446 interdependencies in, 415–417 kinked demand curve model, 434 market shares, 413–414 market structures in, 411 overview, 411 price leadership, 429–433 price wars, avoiding, 434–439 relative market shares in, 411–415

segmented, 438 in the United States, 411–415

O.M. Scott & Sons, 11 Online auctions, 582–583, 596–598 Online sales, 528–529 OPEC, 272–273, 358 Opel, 199 Open bidding design, 588–589 Operating cost, composition of, 323–324 Operating leverage, 324–326 Operating risk exposure, 181 Operating system, 344 Opinion-polling techniques, 155–160 overview, 155

Opportunity cost, 6, 277–279, 649 Optimal capacity utilization optimal output for a given plant size, 286–287

optimal plant size, 287 optimal plant size for a given output rate, 287

Optimal differential pricing algebraic approach, 505–506 demand estimation by market segment, 503

graphical approach, 504–505 multiple-product pricing, 506–507 and price elasticity of demand, 507–508

Optimal incentives contracts, 602–604 Optimal input level, 243 Optimal mechanism design, 580 Optimal overbooking, 537–540 Optimal plant size, 287 Optimal scale of operation, 311–314 Oracle Corp, 388 Orange County, 290 Ordinary least-squares (OLS), 164 Organization of Petroleum Exporting

Countries (OPEC), 422 Organizational form, 564–570 Outputs, maximization of, 251 Outsourcing, 179, 183–184 Overall production efficiency, 253 Overhead costs, 290 Oysters industry, 270 Ozone depletion, 644–646

P Pacific Gas and Electric, 230–232 Palm, Inc., 387, 637 PalmPilot, 387 Pampers, 179 Panama Canal, 183 Paraguay, 206 Parallel imports, 220–222 Participation constraint, 561 Partnerships cost overruns, 599–600 cost revelation in, 599

dissolution of assets in, 574–575 Patents, 383–384, 398, 525, 633–639 PCS PrimeCo., 587–589 Pecuniary benefits, 663 Pecuniary externalities, 626 Penetration pricing, 523 PeopleSoft, 388 Pepperidge Farms, 412 PepsiCo, 177, 348, 453–454, 455, 453–454 Perceived product differentiation, 346 Perfect price discrimination (PPD), 523 Perfectly elastic demand curve, 74 Performance-based pay, 12 Performing arts, 18 Performing loans, 254 Permits, 633 Personal communications service (PCS),

587–589 Personal computer industry, 334–335, 361,

382, 387 Personal digital assistants (PDAs), 637 Personify, 522 Peru, 206 PetroChina, 186 Petroleum refineries, 245–246 Pfizer, 525 Pharmaceutical industry, 14 Philip Morris, 435 Pickens, T. Boone, 39 Pillsbury, 517 Pittsburgh National, 254 Pizza Hut, 70 Planned obsolescence, 476–477 Plant and equipment expenditure plans,

158–159 Plant-level internal economies of scale, 290 Point price elasticity of demand, 73 Pollution abatement, 2–4 Pollution rights, 632–633 Polynomial transformation, 134 Pooling equilibrium, 563 Population regression coefficients, 103–105,

108–110 Port Authority Transit (PAT), 97 Porter, Michael, 337, 341 Positive correlation, 126–127 Post Cereals, 436 Post-contractual opportunistic behavior,

553 Post-purchase discounting risk, 477 Power plants, 2–4 Pre-commitments, 488–491 Predatory pricing, 460 Premier (smokeless cigarette), 10 Present value, 46 Present value interest factor (PVIF), 46–47 Price change

real income effect of, 65–66 substitution effect of, 66–69

Index I-9

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Price competition, 348 Price cost margin (PCM), 348–349 Price discrimination, 521–523, 620–622 Price elasticity of demand in absolute values, 74 arc, 72 in automobile dealerships, 70 for coffee, 80 combined use with demand elasticities, 89–90

defined, 70 and differential pricing, 507–508 electricity use, 89 empirical, 81 factors, 80–83 for gasoline, 71 marginal revenue and, 74, 76–77 and percentage of consumer’s budget, 82 in pizza restaurants, 70 point, 73 and positioning of income-superior pro- ducts, 83

and price levels for monopolists, 391–392 and revenues, 73, 78–80 and substitute goods, 82 and time period of adjustment, 83

Price fixing, 418, 616 Price leadership barometric, 430 defined, 429–430 dominant, 430–432

Price premiums, 374–376 Price signaling, 462 Price skimming, 524 Price variation, 215–216 Price wars avoiding, 434–439 customer segmentation with revenue management, 435

and innovation, 437–439 and market growth, 435 nonprice tactics in, 439 reference prices and framing effects, 436– 437

Priceline, 596 Price-matching guarantees, 461–462 Price-output determination algebraic approach, 390–391 graphical approach, 389–390 for monopolist, 388–392 under monopolistic competition, 361–363 of natural monopoly, 401 and price elasticity of demand, 391–392 under pure competition, 355–360 spreadsheet approach, 388–392

Price-value auction, 589–592 Pricing bundling, 518–521 congestion, 513–514

couponing, 517–518 differential, in target market segments, 512–523

and direct segmentation, 513–515 dynamic, 528 full-cost, 525–526 incremental contribution analysis, 526– 527

on the Internet, 527–529 life cycle pricing, 523–525 limit, 524 niche, 524–525 optimal differential, 503–511 penetration, 523 price discrimination, 521–523 price skimming, 524 reservation prices, 518 target return-on-investment, 526 two-part tariffs, 515–517 value-based, 500–503 value-in-use, 501

Pricing agreement, 460 Prilosec (drug), 635 Primary benefits, 662 Primary costs, 663 Principal-agent model, alternative manage-

rial contracts, 557–558 Principal-agent problem, 10, 561 Printers, 412 Prisoners’ Dilemma, 450–452 Bayesian reputation effects, 459 chain store paradox, 457–459 endgame reasoning, 456–457 folk theorem, 455 grim trigger strategy, 455 industry standards, 463–464 infinitely repeated games, 456 mutual forbearance and cooperation in, 458

price-matching guarantees, 461–462 trembling hand trigger strategy, 455–456

unraveling problem, 456 winning strategies, 459–461

Prius (automobile), 69 Proactive price discrimination, 534 Probability distributions, 49–50 Problem, identification of, 5 Process rays, 251–252 Procter & Gamble, 222, 366, 560 Product differentiation strategy, 338, 346 Product life cycle, 523–525 Production capacity, 322 Production function in automobile assembly plant, 241 average product, 236, 239–241 Cobb-Douglas, 233, 255–256 defined, 232 fixed input, 234–235

increasing returns with network effects, 237–239

inputs, 232–233 law of diminishing marginal returns, 236– 237

long run, 235 for major league baseball, 257–259 marginal, 235–236, 239–241 marginal rate of technical substitution, 245–247

with multiple variable inputs, 243–248 with one variable input, 235–241 production isoquant, 243–244, 247 short run, 234–235 total product, 239–241 variable input, 234–235

Production isoquant, 243–244, 247 Production process allocative efficiency, 252 defined, 251 fixed proportions, 250–252 measuring efficiency of, 252–253 overall production efficiency, 253 and process rays, 251–252 technical efficiency, 253

Production worker elasticity, 257 Product-level internal economies of scale,

287–288 Profit cartels, 421–426 and executive compensation, 11–12 innovation theory of, 7 managerial efficiency of, 7 maximizing, 390–391 monopoly theory of, 7 vs. revenue maximization, 388–392 risk-bearing theory of, 7 temporary disequilibrium theory of, 7

Property rights, 186 Prospect theory, 568–570 Protection level, 536 Protective tariffs, 209–211 Prozac (drug), 14, 222, 398–399, 525, 635 Public goods, 18 Public sector, 18 Public Service Company of New Mexico,

400 Public transportation, 95–97 Public utilities, 397 Publishing companies, 77 Puget Sound Energy, 232 Purchasing power parity absolute, 195–196 Big Mac index of, 201 qualifications of, 198–200 relative, 197–198 trade-weighted exchange rate index, 201– 203

use of, 200–201

I-10 Index

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

as yardstick of comparative growth, 196– 197

Pure competition characteristics of, 352–353 contestable markets, 353 long run, 358–360 price-output determination under, 355– 360

profit maximization under, 357 short run, 355–357

Purina, 517

Q QUALCOMM, 638 Quanta Computer, 185 Queue service rules, 580

R Railroads, 627 Random fluctuations, 141 Random rationing, 491–492 Ratio to trend method, 146–147 Rawlings Sporting Goods, 338 Razor market, 438 Ready-to-eat (RTE) cereal industry, 367, 436 Real income effect, 65–66 Real option, 15 Real terms of trade, 208–209 Reciprocal externalities, 626–628 Reciprocal transformation, 133–134 Recontracting costs, 17 Recreational facilities, 18 Red Hat, 639 Reference prices, 436–437 Refineries, 245–246 Refusals to deal, 622 Refuse collection and disposal, 290 Regression analysis, 99 Regulated monopolies communications companies, 400 electric power companies, 397–399 natural gas companies, 399–400 public utilities, 397 rationale for regulation, 400–402

Regulatory agencies, 624 Relational contracts, 565–567 Relative purchasing power parity, 196–197 Relevant market, 341 Reliance relationships, 372–373 Reliant assets, 565 Renewable resources, production economics

of, 267–270 Rental cars, 69 Repeated games, 449, 453–454 Replacement cost, 279 Replacement guarantees credible commitments, 476 defined, 473 hostage mechanism, 475–476

lease prices and anticipated risks, 480 planned obsolescence, 476–477 post-purchase discounting risk, 477

Required return, 54–55 Resale price maintenance agreements, 622–

623 Resale value, 476–477 Research and development (R&D), 305–306 Reservation prices, 518 Residual claimants, 17 Resource allocation, 9, 41 Restricted stock, 11 Restructuring, 11 Retail banking, 291 Retail gas stations, 37, 414–415 Returns to scale decreasing, 255 defined, 253 increasing, 255 measuring, 254–255

Revco, 570 Revenue maximizing, 388–389 and price elasticity, 73 and price elasticity of demand, 78–80

Revenue equivalence theorem, 594–596 Revenue management in baseball, 539 capacity reallocation, 534–536 cross-functional, 531–532 decisions, 533–540 defined, 531 optimal overlooking in, 537–540 proactive price discrimination in, 534 sources of sustainable price premiums, 531–533

spill, 529–530 spoilage, 529–530 stockouts, 529–530

Risk assessment, 326–327 and coefficient of variation, 53–54 defined, 49 and expected value, 50 measurement of, 49–54 and net present value, 48–49 and normal probability distribution, 51– 53

premium, 54–55 and probability distributions, 49–50 and required return, 54–55 and standard deviation, 51

Risk-bearing theory of profit, 7 Risk-free return, 54–55 Rivalry, intensity of, 347–351 R.J. Reynolds, 435 RJR Nabisco, 10 Robinson-Patman Act, 462, 615–616, 620–

622

Root mean square error (RMSE), 140 Royal Caribbean, 451–452, 460–462

S Sales force polling, 159 Sales forecasting, 159, 167 Sales penetration curve, 384 Sample regression line, 103–104 Samsung, 383, 420, 489 Sara Lee Corp., 42 Saturn Corporation, 13 Saudi Arabia

crude oil pricing, 272–273, 424–425 crude oil production, 272–273, 425–426

Scala S.p.A., 203–204 Schering-Plough Corp., 388, 398–399 Schick-Wilkinson Sword, 438 Schwinn, 566 Sea-Land/Maersk, 581 Search goods, 368–369 Seasonal effects, 141

dummy variables, 147–148 ratio to trend method, 146–147

Second-highest sealed-bid auctions, 592– 594

Secular trends constant rate of growth trends, 145–146 declining rate of growth trends, 146 defined, 141 linear trends, 144–145 types of, 143

Securities and Exchange Commission (SEC), 624

Seiko, 90 Self-enforcing reliance relationship, 602–604 Selling

and optimal advertising intensity, 366– 367

optimal level of, 363–365 and promotional expenses, 363–365

Selten, Reinhard, 448–449 Semiconductor industry, 213, 458–459, 568 Semilogarithmic transformation, 133 Sensitivity analysis, 5 Separating equilibrium, 564 Sequential coordination game, 465–466 Sequential game

analyzing, 464–467 business rivalry as, 467–489 defined, 451 game tree, 466 sequential coordination game, 465–466 subgame perfect equilibrium strategy, 466–467

7-Eleven Japan, 339–341 Sewell, Carl, 374 Sewell Cadillac, 374 Shanghai, China, 187 Shanghai Zhong Qi Pharmaceutical, 222

Index I-11

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Shapiro, Carl, 477 Shareholder value and asymmetric information, 17 and complete markets, 16 and recontracting costs, 17 residual claimants, 17

Shareholder wealth maximizing, 13–17 measuring, 8–9 and resource allocation, 9

Shareholders, and executive compensation, 11–12

Sheet steel, 212–213 Shell, 39 Sherman Act, 88, 614 Sherwin-Williams Company, 99–100, 105–

106, 112–114, 129 Shipping, 581 Shipyard, 41 Short run, 234–235, 355–357, 362 Short-run cost functions, 281–285 statistical estimation of, 310–311

Siemens, 429, 603 Silicon wafers, 290 Simmons Mattress Company, 99 Simple linear regression model, 101–106 assumptions, 102–103

Simultaneous game, 451 Simultaneous games, 450–454 Single-equation models, 160–161 Single-period game, 449 Slippery slope, 495–497 Smokestack scrubbers, 3 Smoothing techniques barometric techniques, 154–155 first-order exponential smoothing, 151–153 moving averages, 148–151 overview, 147–148

Social discount rate, 658 Soft drink industry, 348, 453–454 Sony BMG Music Corp., 617 Sony Corp., 238, 437–439 Sound Warehouse, 472–473 South America, 206 Southern California Edison, 231 Southern Company, 3–4, 8, 15 Southland Corp., 339–340 Southwest Airlines, 275, 336, 339, 444–445 Specification errors, 129 Speculation, 191 Sperm whale oil, 29 Spill, 529–530 Spoilage, 529–530 Sport Obermeyer, 530, 537 Spot market transactions, 551–552 Spur Industries v. Del Webb Development,

626 Standard deviation, 51 Standard error of the estimate, 107–108

Staples, 88, 345 Starbuck’s Coffee, 16 State excise taxes, 63 State Grid Corporation, 186 Steel, 212–213 Steel ingot, 316 Steel production, 316 Stents, 502 Sterilized interventions, 190 Stochastic disturbance, 103 Stochastic time-series analysis, 163–166 Stock options, 11, 559–560, 562 Stockouts, 529, 537 Stocks, 55 Strategic holdout, 629 Strategic positioning, 336–337 Strategic trade policy, 211–213 Strategies cost-based, 339 information technology, 339–341 production differentiation, 338

Strategy game, 446–447 Strategy process, 336 Stratified lotteries, 582 Stride Rite, 623 StubHub.com, 582–583 Subgame perfect equilibrium strategy, 466–

467 Subsidies, 630–632 Substitute goods, 32, 82, 87, 110–111, 342–

343 Substitution effect, 66–69 Sulfur dioxide (SOX) emissions, 2–3 Sun Microsystems, 192, 388, 470, 621 Sunk cost, 280–281, 492 Supermarket pricing, 506–507 Superstores, 88 Suppliers, power of, 346–347 Supply and demand, 27–29 and equilibrium market price, 27–29 factors and expected effect, 35

Supply curves, 36 Supply function, 35 Survey of Current Business, 158–159 Surveys consumer expenditure plans, 159 inventory changes and sales expectations, 159

macroeconomic activity, 159–160 overview, 155 plant and equipment expenditure plans, 158–159

sales forecasting, 159 Survivor technique, 316–317 Survivor (television show), 446–447 Sustainability, 2 Sustainable competitive advantages, 337 Sylvania, 87, 566

T Taco Bell, 324 Taiwan Instrument Co., 508–512 Target return-on-investment pricing, 526 Tariffs, 209–211 Taxes, 630–632 Taxi medallions, licensing of, 479 Technical efficiency, 253 Telecommunications industry, 387, 400 Temporary disequilibrium theory of profit, 7 Tenneco, 41 Tesco, 221 Texaco, 16 Theme parks, 383 Theory of Games and Evolution, The (Von

Neumann/Morgenstern), 448 Thomson, 429 Threat of substitutes, 342–343 Threshold sales curve, 535 Ticketmaster, 582–583 Time depreciation, 307 Time series components of, 141 cyclical variations, 141 data, 141 defined, 126 models, 141–142 random fluctuations, 141 seasonal effects, 141, 146–147 secular trends, 141, 143–146 smoothing techniques, 147–153

Time Warner, 313, 566 Time Warner Cable, 518–520 Times Mirror Co., 633 Timken, 305–306 Tires, 212–213 Tobacco tax, 62–63 Toll roads, 514 Toshiba, 238, 603 Total cost, 364 Total product, 239–241 Total profit, 43–45 Total revenue, 78, 364 Townsend-Greenspan, 162 Toyota Motors, 4–5, 179, 199, 476, 664 Tradable pollution allowances (TPAs), 2, 16 Trade deficit, 222–223 Trade secrets, 636–639 Trade-weighted exchange rate index,

201–203 Trading blocs, 204–207 Transaction risk exposure, 180–181 Translation risk exposure, 181 Transportation costs, 291 Treasury bills, 54–55, 158 Treasury bond, 158 Trembling hand trigger strategy, 455–456 Tversky, Amos, 568 Two-part tariffs, 515–517

I-12 Index

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Two-person game, 449–450 Two-person zero-sum game, 450 Tying agreements, 621

U Underbidding, 589–592 Unit costs, 316 Unit elastic demand, 76 United Airlines, 339, 384 United Auto Worker (UAW), 36 United States bilateral trade with Japan, 208–209 destination of exports, 220 energy policy, 272–273 exports, 191–193 exports to China, 175–178 gross domestic product, 175–177 oil reserves, 272 oligopoly in, 411–415 trade deficit, 222–223 trading bloc, 206 trading partners, 218

United Steel Workers (USW), 212 Universal Music Group, 617 Universal Studios, 383 Unraveling problem, 456 Uruguay, 206 US Airways, 50, 275–276, 339, 350–351, 445 U.S. Army Corps of Engineers, 241 U.S. dollar depreciation, 223 equilibrium price, 189 foreign exchange market for, 187–191

U.S. Treasury bills, 54–55, 593 Use depreciation, 307 UUNet, 138

V Value at risk, 216 Value proposition, 335, 393–394 Value-based pricing, 501 Value-Mart, 148 Variable costs, 213–214, 281 Variable identification, 99 Variable input defined, 234 long run, 235 marginal factor cost, 242–243 marginal revenue product, 242–243 multiple, 243–248 optimal input level, 243 short run, 234–235

Varian, Hal, 477 Verizon, 468 Versioning, 477 Vertical integration, 571–573 Vertical requirements contracts, 549 VF Corporation, 110–111 Vickrey, William, 592 Vickrey auction, 592–594 Virtual Vineyards, 522 Volkswagen, 79 volume discounts, 288 Voluntary import restraints (VIRs), 210 Volvo, 34 Von Neumann, John, 448

W Walker Corporation, 150–151, 153 Walmart, 88, 179, 218, 237 Walt Disney, 10 Warner Brothers, 238 Warner Music Group, 617

Wars of attrition, 494–495 Watches, 90 Web browsers, 344, 388, 621 WebTV, 546–547, 567 Weighted cost of capital, 657 Whale oil, 29 Wharton Econometric Forecasting Associ-

ates, 162 Whirlpool, 605 Whitman’s Sampler, 394 Whole Foods Inc., 620 Wholesale price discrimination, 620–622 Wide-bodied jets, 309, 414 Wild Oats Markets, 620 Willingness to pay (WTP), 631 Windows operating system, 238, 344, 477,

616, 621 Winner’s curse, 584–587 Winning strategies, 459–461 Wireless Co., 587–589 World Bank, 211 World Trade Organization (WTO),

211 WorldCom, 102, 138, 556

X Xerox, 337, 388

Y Yield management. See Revenue

management Yogurt, 82 You Ke (drug), 222

Z Zantac (drug), 525

Index I-13

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

NOTES

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

NOTES

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

NOTES

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

NOTES

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

NOTES

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

NOTES

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

NOTES

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

NOTES

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Examples, Managerial Challenges, and International Perspectives, in Managerial Economics: Applications, Strategy, and Tactics, 12e

PART I: INTRODUCTION 1

Chapter 1: Introduction and Goals of the Firm 2 Managerial Challenge: How to Achieve Sustainability: Southern Company 2

Capacity Expansion at Honda, N.A., and Toyota Motors, N.A. 4 Shareholder Wealth Maximization at Berkshire Hathaway 8 Resource-Allocation Decisions and Shareholder Wealth: Apple Computer 9

Agency Costs and Corporate Restructuring: O.M. Scott & Sons 11 Executive Performance Pay: General Electric 12 What Went Right/What Went Wrong: Saturn Corporation 13 What Went Right/What Went Wrong: Eli Lilly Depressed

by Loss of Prozac Patent 14 Amgen’s Potential Profitability Is Realized 14 Real Option Value Attributable to Fuel-Switching Technology at Southern Company 15

Tradable Pollution Permits at Duke Power 16 Case: Designing a Managerial Incentives Contract 21 Case: Shareholder Value of Wind Power at Hydro Co.: RE < C 23

Chapter 2: Fundamental Economic Concepts 26 Managerial Challenge: Why Charge $25 per Bag on Airline Flights? 26

Discovery of Jojoba Bean Causes a Collapse of Whale Oil Lubricant Prices 29

Exchange Rate Impacts on Demand: Cummins Engine Company 34

NAFTA and the Reduced Labor Costs of Ford Assembly Plants in Detroit 36

Speculation Sends Crude Oil Input Price on a Roller-Coaster Ride at ExxonMobil and Shell 40

Tenneco Shipyard Marginal Analysis 41 Marginal Analysis and Capital Budgeting Decisions: Sara Lee Corporation 42

Marginal Analysis of Driving a Mini Cooper versus a Chevy Volt 43

Changing a Lightbulb Saves $40 and May Save the Planet 47 Probability Distributions and Risk: US Airways Bonds 50 What Went Right/What Went Wrong: Long-Term Capital

Management (LTCM) 53 Relative Risk Measurement: Arrow Tool Company 54 Risk-Return Trade-Offs in Stocks, Bonds, Farmland, and Diamonds 55

Case: Revenue Management at American Airlines 58

PART II: DEMAND AND FORECASTING 61

Chapter 3: Demand Analysis 62 Managerial Challenge: Health Care Reform and Cigarette Taxes 62

Consumption Choices on a Business Trip to San Francisco 66 What Went Right/What Went Wrong: Chevy Volt 69 Pizza Hut and Ford Dealers Respond to Deficient Demand 70 Price Elasticity at Various Price Points along a Linear Demand Curve for Gasoline 71

Content Providers Press Publishing Companies to Lower Prices 77

VW’s Invasion of North America 79

Price Elasticity Estimates for Coffee Vary by Price 80 Free Trade and the Price Elasticity of Demand: Nestlé Yogurt 82

Targeting a Direct Mail Coupon at a Ford Dealership 84 Income Elasticities: Empirical Estimates 86 Why Pay More for Fax Paper and Staples at Staples? 88 Price and Income Effects: The Seiko Company 90 Case: Polo Golf Shirt Pricing 93

Chapter 4: Estimating Demand 95 Managerial Challenge: Global Warming and the Demand for Public Transportation 95

Estimating Cross Elasticity: Simmons Mattress Company 98 Variable Identification and Data Collection: Sherwin-Williams Company 99

Linear, Not Exponential, Sales at Global Crossing Inc. 102 Estimating Regression Parameters: Sherwin-Williams Company 105

Are Designer Jeans and Lee Jeans Complements or Substitutes? 110

The Estimated Demand for New Automobiles 116 Case: Soft Drink Demand Estimation 124

Appendix 4A: Problems in Applying the Linear Regression Model 126

Constant Elasticity Demand: Pepsi 133

Chapter 5: Business and Economic Forecasting 137

Managerial Challenge: Excess Fiber Optic Capacity at Global Crossing Inc. 137

What Went Right/What Went Wrong: Crocs Shoes 140 Linear Trend Forecasting: Prizer Creamery 144 Seasonally Adjusted Forecasts: Prizer Creamery 146 Dummy Variables and Seasonal Adjustments: Value-Mart Company 148

Moving Average Forecasts: Walker Corporation 150 Exponential Smoothing: Walker Corporation 153 Leading Indicators Change 158 Single-Equation Forecasts: The Demand for Game-Day Attendance in the NFL 160

Long-Term Sales Forecasting by General Motors in Overseas Markets 167

Case: Cruise Ship Arrivals in Alaska 172 Case: Lumber Price Forecast 173

Chapter 6: Managing in the Global Economy 175 Managerial Challenge: Financial Crisis Crushes U.S. Household Consumption and Business Investment: Will Exports to China Provide the Way Out? 175

What Went Right/What Went Wrong: Export Market Pricing at Toyota 179

Collapse of Export and Domestic Sales at Cummins Engine 181

European Slowdown Decreases DuPont Exports 192 Birkenstocks for Sale Cheap! 196 What Went Right/What Went Wrong: GM, Toyota,

and the Celica GT-S Coupe 199

This symbol denotes an International Perspectives example.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Are Dirty Dishes Mobile? Dixan, Joy, Dawn, and Generic Patti Scala from Scala S.P.A. 203

President Bush’s 2002 Steel Tariffs and President Obama’s 2009 Tire Tariffs: Justified Sanctions or Hypocritical Protectionism? 212

Intel’s Chip Market Access Improves in Japan 213 Microsoft and Apple: A Role for Protective Tariffs? 214 If Europe, What about One Currency for NAFTA? 217 Household Iron Manufacturer in Mexico Becomes Major Engine Block Supplier to Detroit: Cifunsa SA 219

EU Ban on Some Parallel Imports Pleases European but not U.S. and Japanese Manufacturers 221

What Went Right/What Went Wrong: Ford Motor Co. and Exide Batteries: Are Country Managers Here to Stay? 222

Case: Predicting the Long-Term Trends in Value of the U.S. Dollar and Euro 226

Case: Debating the Pros and Cons of NAFTA 226

Appendix 6A: Foreign Exchange Risk Management 227

Internal Hedge from BMW Operations on I-85 in South Carolina 227

Toyota and Honda Buy U.S. Assembly Capacity 228 Cummins Engine Goes Short 228

PART III: PRODUCTION AND COST 229

Chapter 7: Production Economics 230 Managerial Challenge: Green Power Initiatives Examined: What Went Wrong in California’s Deregulation of Electricity? 230

An Illustrative Production Function: Deep Creek Mining Company 233

What Went Right/What Went Wrong: Factory Bottlenecks at a Boeing Assembly Plant 237

Increasing Returns at Sony Blu-ray and Microsoft Windows 238

Three Stages of Production on the Camaro Assembly Line 241

Just What Exactly Is a Refinery, and Why Won’t Anyone Build One? 245

Isocost Determination: Deep Creek Mining Company 248 Cost Minimization: Deep Creek Mining Company 251 GM’s A-Frame Supplier Achieves 99.998 Percent Technical Efficiency 253

Technical and Allocative Efficiency in Commercial Banks at BB&T 254

Moneyball: A Production Function for Major League Baseball 257

Case: The Production Function for Wilson Company 263

Appendix 7A: Maximization of Production Output Subject to a Cost Constraint 265

Appendix 7B: Production Economics of Renewable and Exhaustible Natural Resources 267

Oyster Seedbed Replenishment on Chesapeake Bay 270 Saudi Arabian Oil Minister Plays a Waiting Game 272

Chapter 8: Cost Analysis 275 Managerial Challenge: US Airways Cost Structure 275 Opportunity Costs at Bentley Clothing Store 278 Inventory Valuation at Westside Plumbing and Heating 279

Short-Run Cost Functions: Deep Creek Mining Company 282 The Average Cost per Kilowatt Hour in Underutilized Power Plants 287

Mass Customization and the Learning Curve 288 Percentage of Learning: Emerson Corporation 289 IBM and Intel Fabricate Monster Silicon Wafers 290 Refuse Collection and Disposal in Orange County 290 Economies of Scale: Superscale Money-Center versus Community Banks 291

Flexibility and Operating Efficiency: Ford Motor Company’s Flat Rock Plant 292

How Japanese Companies Deal with the Problems of Size 292

Aluminum-Intensive Vehicle Lowers the Minimum Efficient Scale at Ford 293

Case: Cost Analysis 298

Appendix 8A: Long-Run Costs with a Cobb-Douglas Production Function, Advanced Material 301

Chapter 9: Applications of Cost Theory 305 Managerial Challenge: How Exactly Have Computerization and Information Technology Lowered Costs at Chevron, Timken, and Merck? 305

What Went Right/What Went Wrong: Boeing: The Rising Marginal Cost of Wide-Bodies 309

Short-Run Cost Function for Multi-Product Food Processing 310

Short-Run Cost Functions: Electricity Generation 311 Long-Run Cost Functions: Electricity Generation 312 Scale Economies in the Traditional Cable Industry: Time-Warner 313

Economies of Scope in the Banking Industry 315 The Survivor Technique: Steel Production 316 Boeing 777 Exceeds Break-Even Sales Volume 318 Break-Even Analysis: Allegan Manufacturing Company 320 Fixed Costs and Production Capacity at General Motors 322 Taco Bell Chihuahua Drives Sales 324 Operating Leverage: Allegan Manufacturing Company 325 Business Risk Assessment: Allegan Manufacturing Company 327 Case: Cost Functions 330 Case: Charter Airline Operating Decisions 331

PART IV: PRICING AND OUTPUT DECISIONS: STRATEGY AND TACTICS 333

Chapter 10: Prices, Output, and Strategy: Pure and Monopolistic Competition 334

Managerial Challenge: Resurrecting Apple 334 What Went Right/What Went Wrong: Xerox 337 Rawlings Sporting Goods Waves Off the Swoosh Sign 338 Think Small to Grow Big: Southwest Airlines 339 Dell’s Cost Leadership in PC Assembly 340 The E-Commerce of Lunch at 7-Elevens in Japan 340 The Relevant Market for Web Browsers: Microsoft’s Internet Explorer 344

Potential Entry at Office Depot/Staples 345 Eli Lilly Poses a Threat of Potential Entry for AstraZeneca 345 Objective versus Perceived Product Differentiation: Xerox 346 Price Competition at the Soda Fountain: PepsiCo Inc. 348 Contribution Margins at Hanes Discourage Discounting 349 Intensity of Rivalry at US Airways 350

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Profit Maximization under Pure Competition (Short Run): Adobe Corporation 357

Gasoline Price Rises to Record Levels Reflecting a Spike in Crude Oil Input Costs 358

Copper Price Rise by 400 Percent Contributes to Housing Bubble 360

What Went Right/What Went Wrong: The Dynamics of Competition at Amazon.com 362

Long-Run Price and Output Determination: Blockbuster, Inc. 364

Optimal Advertising at Parkway Ford 365 Ford and P&G Tie Ad Agency Pay to Sales 366 Optimal Advertising Intensity at Kellogg’s and General Mills 367

Credible Product Replacement Claims: Dooney & Bourke 373 Customers for Life at Sewell Cadillac 374 Hostage Exchange with Efficient Uncut Diamond Sorting at De Beers 376

Case: Blockbuster, Netflix and Redbox Compete for Movie Rentals 380

Case: Saving Sony Music 381

Chapter 11: Price and Output Determination: Monopoly and Dominant Firms 382

Managerial Challenge: Dominant Microprocessor Company Intel Adapts to Next Trend 382

The Mickey Mouse Monopoly: Disney 383 Impermanent Control of a Denver Airport Hub: United Airlines 384

What Went Right at Microsoft but Wrong at Apple Computer 386

What Went Right/What Went Wrong: Pilot Error at Palm 387 Profit versus Revenue Maximization for Polo Golf Shirts 388 Profit Maximization for a Theme Park Restaurant 390 Price Elasticity and Price Levels for Monopolists 392 Markups and Contribution Margins on Chanel No. 5, OleMusk, and Whitman’s Sampler 394

Components of the Gross Margin at Kellogg Co. 396 Using Limit Pricing to Hamper the Sales of Generic Drugs 398 What Went Right/What Went Wrong: The Public Service

Company of New Mexico 400 Case: Differential Pricing of Pharmaceuticals: The HIV/AIDS Crisis 406

Chapter 12: Price and Output Determination: Oligopoly 409

Managerial Challenge: Are Nokia’s Margins on Cell Phones Collapsing? 409

Hewlett-Packard’s Dominance in Printers 412 Auto Rental and Retail Gasoline Firms Consolidate: Enterprise Rent-A-Car, and Exxon/Mobil 414

Airline Pricing: The Pittsburgh Market 416 The Cournot Oligopoly Solution: Siemens and Lucent-Alcatel 416

How Ocean Shipping Conferences Have Affected Shipping Rates 419

DRAM Chipmakers Pay Enormous Fines for Forming a Global Cartel 420

The OPEC Cartel 422 Coffee Pricing Agreement Dissolves amidst Dilemma 423 Exhaustible Natural Resources: Saudi Arabia Plays a Waiting Game 424

What Drives the Cost of $3 per Gallon Gasoline? 426

Revenue-Sharing in Major League Baseball 429 Barometric Price Leadership: American Airlines and Continental Airlines 430

Price Leadership: Aerotek 432 Price Wars at General Mills and Post 436 What Went Right/What Went Wrong: Good-Better-Best

Product Strategy at Kodak and Marriott 437 What Went Right at Interlink Surgical Steel and Gillette? 438 Nonprice Tactics in a Price War: Coors 439 Case: Cell Phones Displace Mobile Phone Satellite Networks 442

Chapter 13: Best-Practice Tactics: Game Theory 444 Managerial Challenge: Large-Scale Entry Deterrence of Low-Cost Discounters: Southwest, People Express, Value Jet, Kiwi, and JetBlue 444

The Nobel Prize Goes to Three Game Theorists 449 Solving the Chain Store Paradox: Semiconductor Pricing at Intel, NEC, and AMD 458

Brown and Williamson’s Reputation for Predating 460 Signaling a Punishment Scheme: Northwest 462 Business Gaming at Verizon 468 Technology Leader or Fast-Second: IBM 470 Double-the-Difference Price Guarantees: Best Buy 472 Noncredible Commitments: Burlington Industries 474 Resale Value of a Mini Cooper 476 Leasing Digital Moviehouse Projectors: Hughes-JVC 478 NetJets Fractional Ownership Plans for Learjet and Gulfstream Aircraft and Lexus 478

The Licensing of Taxi Medallions and Cell Phones 479 The Superjumbo Dilemma 485 Case: The Superjumbo Dilemma 485

Appendix 13A: Entry Deterrence and Accommodation Games 488

Excess Capacity in the World Car Market: Samsung and Hyundai 489

Contestable Market in Bicycle Helmets: Bell Sports 493 Circuit City Driven over the Brink 496

Chapter 14: Pricing Techniques and Analysis 499 Managerial Challenge: Pricing of Apple Computers: Market Share versus Current Profitability 499

Coated Coronary Stents Reduce the Cost of Later Surgeries 502 Supermarket Pricing 507 Differential Pricing at Taiwan Instrument Co. 508 Congestion Charges on Los Angeles, San Diego, Houston, Minneapolis, Denver, and Dulles Toll Roads 514

Dynamic Pricing for Electricity 515 What Went Right/What Went Wrong: Two-Part Pricing

at Disney World 517 What Went Right/What Went Wrong: Price-Sensitive

Customers Redeem 517 McDonald’s Introduces Mixed Bundles as “Extra Value Meals” 520

eBusiness Clickstreams Allow Price Discrimination: Personify and Virtual Vineyards 522

Loss of Patent Protection Limits the Price of Prozac: Eli Lilly 525

Niche Pricing at Pfizer 525 Full-Cost Pricing Results in the Loss of a Big Contract at J.P. Morgan: British Telephone 526

Incremental Contribution Analysis at Continental Airlines 526

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

Spill and Spoilage at Sport Obermeyer 530 The Optimal Probability of a Stockout at Sport Obermeyer 537

Pinpoint Booking Accuracy at American Airlines 538 Revenue Management in Baseball: The Baltimore Orioles 539

PART V: ORGANIZATIONAL ARCHITECTURE AND REGULATION 545

Chapter 15: Contracting, Governance, and Organizational Form 546

Managerial Challenge: Controlling the Vertical: Ultimate TV 546

Crankshaft Delivery Delay Causes Plant Closing 550 The Enforcement and Excusal of Contract Promises: The Extraordinary Case of 9/11 552

What Went Right/What Went Wrong: Moral Hazard and Holdup at Enron and WorldCom 556

Indexed Stock Options at Adobe Systems, Dell, and Cisco 560 P&G Pays Ad Executives Based on Their Performance 560 What Went Right/What Went Wrong: Why Have

Restricted Stock Grants Replaced Executive Stock Options at Microsoft? 562

Schwinn and Sylvania Dealers Offered Exclusive Territories 566 Kodak and Time Warner Form a Digital Photography Alliance 566

What Went Right/What Went Wrong: Cable Allies Refuse to Adopt Microsoft’s WebTV as an Industry Standard 567

Economies of Scale and International Joint Ventures in Chip Making 568

Full-Line Forcing in Pens, Aspirin, and Multivitamins at Revco and Eckerd 570

What Went Right/What Went Wrong: Dell Replaces Vertical Integration with Virtual Integration 573

Case: Designing a Managerial Incentive Contract 578 Case: The Division of Investment Banking Fees in Syndicate 578

Appendix 15A: Auction Design and Information Economics 580

Containerized Shipping at Sea-Land/Maersk 581 Online Auctions and Stratified Lotteries at Ticketmaster 582 Winner’s Curse at ESPN 586 Open Bidding Simultaneous Auction of PCS Spectrum Rights 589

Exponential Valley Inc. Auctions a Chip Patent 591 Second-Highest Sealed-Bid Auction: U.S. Treasury Bills 593 Internet Auction Design Becomes a Big E-Business Debate: eBay versus Priceline 596

Intel and Analog Inc. Form Partnership to Develop DSP Chip 598

Joint Venture in Memory Chips: IBM, Siemens, and Toshiba 603

Whirlpool’s Joint Venture in Appliances Improves upon Maytag’s Outright Purchase of Hoover 605

Case: Spectrum Auction 608 Case: Debugging Computer Software: Intel 608

Chapter 16: Government Regulation 610 Managerial Challenge: Cap and Trade, Deregulation, and the Coase Theorem 610

Why City-Pair Airlines Are Not Contestable Markets 614 Why Miller Beer Is So Hard to Find in Mexico 615 California’s Class Action Suit against Microsoft Settled for $1.1 Billion 616

Intel and AMD Fight It Out in Microprocessors 618 Trustbusters Reappear: DISH Network-DIRECTV Merger Disapproved 620

Potentially Anticompetitive Practices: Microsoft’s Tying Arrangements 621

RPM at Stride Rite and Leegin Creative Leather Products 623 What Went Right/What Went Wrong: The Need for a

Regulated Clearinghouse to Control Counterparty Risk at AIG 625

Coase’s Railroad 627 Mandatory Auto Inspections 629 Boomer v. Atlantic Cement Co., Inc., 26 N.Y. 2d 219 630 Plant Expansion Requires Purchase of Open-Market Pollution Rights: Times Mirror Co. 633

What Went Right/What Went Wrong: Delayed Release at Aventis 635

What Went Right/What Went Wrong: Technology Licenses Cost Palm Its Lead in PDAs 637

Competing Business Plans at Celera Genomics and Human Genome Sciences 637

What Went Right/What Went Wrong: Motorola: What They Didn’t Know Hurt Them 638

Case: Microsoft Tying Arrangements 643 Case: Music Recording Industry Blocked from Consolidating 643

Chapter 17: Long-Term Investment Analysis 644 Managerial Challenge: Multigenerational Effects of Ozone Depletion and Greenhouse Gases 644

Capital Expenditures at Chrysler: The Grand Cherokee 648 Cash-Flow Estimation: Hamilton Beach/Proctor-Silex, Inc. 650 Calculation of Internal Rate of Return: Hamilton Beach/Proctor-Silex 651

Net Present Value Calculation: Hershey Foods 652 Cost of Debt Capital: AT&T 654 Cost of Internal Equity Capital: Fresno Company 656 Cost of External Equity Capital: Fresno Company 656 Weighted Cost of Capital: Columbia Gas Company 657 Costs and Benefits of a Toyota Automobile Plant in Kentucky 664

Case: Cost-Benefit Analysis 670 Case: Industrial Development Tax Relief and Incentives 672

This symbol denotes an International Perspectives example.

Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.Copyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.

  • Cover Page
  • Title Page
  • Copyright Page
  • Dedication
  • Preface
  • About the Authors
  • Brief TABLE OF CONTENTS
  • Contents
  • PART I INTRODUCTION
    • 1 Introduction and Goals of the Firm
      • Chapter Preview
      • Managerial Challenge: How to Achieve Sustainability: Southern Company
      • What is Managerial Economics?
      • The Decision-Making Model
      • The Role of Profits
      • Objective of the Firm
      • Separation of Ownership and Control: The Principal-Agent Problem
      • What Went Right / What Went Wrong: Saturn Corporation
      • Implications of Shareholder Wealth Maximization
      • What Went Right/What Went Wrong: Eli Lilly Depressed by Loss of Prozac Patent
      • Summary
      • Exercises
      • Case Exercise: Designing a Managerial Incentives Contract
      • Case Exercise: Shareholder Value of Wind Power at Hydro Co.: RE < C
    • 2 Fundamental Economic Concepts
      • Chapter Preview
      • Managerial Challenge: Why Charge $25 per Bag on Airline Flights?
      • Demand and Supply: A Review
      • Marginal Analysis
      • The Net Present Value Concept
      • Meaning and Measurement of Risk
      • What Went Right/What Went Wrong: Long-Term Capital Management (LTCM)
      • Risk and Required Return
      • Summary
      • Exercises
      • Case Exercise: Revenue Management at American Airlines
  • PART II DEMAND AND FORECASTING
    • 3 Demand Analysis
      • Chapter Preview
      • Managerial Challenge: Health Care Reform and Cigarette Taxes
      • Demand Relationships
      • What Went Right / What Went Wrong: Chevy Volt
      • The Price Elasticity of Demand
      • International Perspectives: Free Trade and the Price Elasticity of Demand: Nestlé Yogurt
      • The Income Elasticity of Demand
      • Cross Elasticity of Demand
      • The Combined Effect of Demand Elasticities
      • Summary
      • Exercises
      • Case Exercise: Polo Golf Shirt Pricing
    • 4 Estimating Demand
      • Chapter Preview
      • Managerial Challenge: Global Warming and the Demand for Public Transportation
      • Estimating Demand Using Marketing Research Techniques
      • Statistical Estimation of the Demand Function
      • A Simple Linear Regression Model
      • Using the Regression Equation to Make Predictions
      • Multiple Linear Regression Model
      • Summary
      • Exercises
      • Case Exercise: Soft Drink Demand Estimation
    • 4A Problems in Applying the Linear Regression Model
      • Introduction
      • Nonlinear Regression Models
      • Summary
      • Exercises
    • 5 Business and Economic Forecasting
      • Chapter Preview
      • Managerial Challenge: Excess Fiber Optic Capacity at Global Crossing Inc.
      • The Significance of Forecasting
      • Selecting a Forecasting Technique
      • What Went Right / What Went Wrong: Crocs Shoes
      • Alternative Forecasting Techniques
      • Deterministic Trend Analysis
      • Smoothing Techniques
      • Barometric Techniques
      • Survey and Opinion-Polling Techniques
      • Econometric Models
      • Stochastic Time-Series Analysis
      • Forecasting with Input-Output Tables
      • International Perspectives: Long-Term Sales Forecasting by General Motors in Overseas Markets
      • Summary
      • Exercises
      • Case Exercise: Cruise Ship Arrivals in Alaska
      • Case Exercise: Lumber Price Forecast
    • 6 Managing in the Global Economy
      • Chapter Preview
      • Managerial Challenge: Financial Crisis Crushes U.S. Household Consumption and Business Investment: Will Exports to China Provide the Way Out?
      • Introduction
      • What Went Right / What Went Wrong: Export Market Pricing at Toyota
      • Import-Export Sales and Exchange Rates
      • International Perspectives: Collapse of Export and Domestic Sales at Cummins Engine
      • Outsourcing
      • China Trade Blossoms
      • The Market for U.S. Dollars as Foreign Exchange
      • Determinants of Long-Run Trends in Exchange Rates
      • Purchasing Power Parity
      • What Went Right / What Went Wrong: GM, Toyota, and the Celica GT-S Coupe
      • International Trade: A Managerial Perspective
      • Free Trade Areas: The European Union and NAFTA
      • Largest U.S. Trading Partners: The Role of NAFTA
      • What Went Right / What Went Wrong: Ford Motor Co. and Exide Batteries: Are Country Managers Here to Stay?
      • Perspectives on the U.S. Trade Deficit
      • Summary
      • Exercises
      • Case Exercise: Predicting the Long-Term Trends in Value of the U.S. Dollar and Euro
      • Case Exercise: Elaborate the Debate on NAFTA
    • 6A Foreign Exchange Risk Management
  • PART III PRODUCTION AND COST
    • 7 Production Economics
      • Chapter Preview
      • Managerial Challenge: Green Power Initiatives Examined: What Went Wrong in California’s Deregulation of Electricity?
      • The Production Function
      • Production Functions with One Variable Input
      • What Went Right / What Went Wrong: Factory Bottlenecks at a Boeing Assembly Plant
      • Determining the Optimal Use of the Variable Input
      • Production Functions with Multiple Variable Inputs
      • Determining the Optimal Combination of Inputs
      • A Fixed Proportions Optimal Production Process
      • Measuring the Efficiency of a Production Process
      • Returns to Scale
      • Summary
      • Exercises
      • Case Exercise: The Production Function for Wilson Company
    • 7A Maximization of Production Output Subject to a Cost Constraint
      • Exercise
    • 7B Production Economics of Renewable and Exhaustible Natural Resources
      • Renewable Resources
      • Exhaustible Natural Resources
      • Exercises
    • 8 Cost Analysis
      • Chapter Preview
      • Managerial Challenge: US Airways Cost Structure
      • The Meaning and Measurement of Cost
      • Short-Run Cost Functions
      • Long-Run Cost Functions
      • Economies and Diseconomies of Scale
      • International Perspectives: How Japanese Companies Deal with the Problems of Size
      • Summary
      • Exercises
      • Case Exercise: Cost Analysis
    • 8A Long-Run Costs with a Cobb-Douglas Production Function
      • Exercises
    • 9 Applications of Cost Theory
      • Chapter Preview
      • Managerial Challenge: How Exactly Have Computerization and Information Technology Lowered Costs at Chevron, Timken, and Merck?
      • Estimating Cost Functions
      • What Went Right / What Went Wrong: Boeing: The Rising Marginal Cost of Wide-Bodies
      • Break-Even Analysis
      • Summary
      • Exercises
      • Case Exercise: Cost Functions
      • Case Exercise: Charter Airline Operating Decisions
  • PART IV PRICING AND OUTPUT DECISIONS: STRATEGY AND TACTICS
    • 10 Prices, Output, and Strategy: Pure and Monopolistic Competition
      • Chapter Preview
      • Managerial Challenge: Resurrecting Apple
      • Introduction
      • Competitive Strategy
      • What Went Right / What Went Wrong: Xerox
      • Porter’s Five Forces Strategic Framework
      • A Continuum of Market Structures
      • Price-Output Determination under Pure Competition
      • Price-Output Determination under Monopolistic Competition
      • What Went Right / What Went Wrong: The Dynamics of Competition at Amazon.com
      • Selling and Promotional Expenses
      • Competitive Markets under Asymmetric Information
      • Solutions to the Adverse Selection Problem
      • Summary
      • Exercises
      • Case Exercise: Blockbuster, Netflix, and Redbox Compete for Movie Rentals
      • Case Exercise: Saving Sony Music
    • 11 Price and Output Determination: Monopoly and Dominant Firms
      • Chapter Preview
      • Managerial Challenge: Dominant Microprocessor Company Intel Adapts to Next Trend
      • Monopoly Defined
      • Sources of Market Power for a Monopolist
      • What Went Right / What Went Wrong: Pilot Error at Palm
      • Price and Output Determination for a Monopolist
      • The Optimal Markup, Contribution Margin, and Contribution Margin Percentage
      • Regulated Monopolies
      • What Went Right / What Went Wrong: The Public Service Company of New Mexico
      • The Economic Rationale for Regulation
      • Summary
      • Exercises
      • Case Exercise: Differential Pricing of Pharmaceuticals: The HIV / AIDS Crisis
    • 12 Price and Output Determination: Oligopoly
      • Chapter Preview
      • Managerial Challenge: Are Nokia’s Margins on Cell Phones Collapsing?
      • Oligopolistic Market Structures
      • Interdependencies in Oligopolistic Industries
      • Cartels and Other Forms of Collusion
      • International Perspectives: The OPEC Cartel
      • Price Leadership
      • The Kinked Demand Curve Model
      • Avoiding Price Wars
      • What Went Right / What Went Wrong: Good-Better-Best Product Strategy at Kodak and Marriott
      • Summary
      • Exercises
      • Case Exercise: Cell Phones Displace Mobile Phone Satellite Networks
    • 13 Best-Practice Tactics: Game Theory
      • Chapter Preview
      • Managerial Challenge: Large-Scale Entry Deterrence of Low-Cost Discounters: Southwest, People Express, Value Jet, Kiwi, and JetBlue
      • Oligopolistic Rivalry and Game Theory
      • Analyzing Simultaneous Games
      • The Escape from Prisoner’s Dilemma
      • Analyzing Sequential Games
      • Business Rivalry as a Self-Enforcing Sequential Game
      • Credible Threats and Commitments
      • Mechanisms for Establishing Credibility
      • Replacement Guarantees
      • Summary
      • Exercises
      • Case Exercise: International Perspectives: The Superjumbo Dilemma
    • 13A Entry Deterrence and Accommodation Games
      • Excess Capacity as a Credible Threat
      • Pre-commitments Using Non-Redeployable Assets
      • Customer Sorting Rules
      • Tactical Insights about Slippery Slopes
      • Summary
      • Exercises
    • 14 Pricing Techniques and Analysis
      • Chapter Preview
      • Managerial Challenge: Pricing of Apple Computers: Market Share versus Current Profitability
      • A Conceptual Framework for Proactive, Systematic-Analytical, Value-Based Pricing
      • Optimal Differential Price Levels
      • Differential Pricing in Target Market Segments
      • What Went Right / What Went Wrong :Two-Part Pricing at Disney World
      • What Went Right / What Went Wrong:Price-Sensitive Customers Redeem
      • Pricing in Practice
      • The Practice of Revenue Management, Advanced Material
      • Summary
      • Exercises
  • PART 5 VORGANIZATIONAL ARCHITECTURE AND REGULATION
    • 15 Contracting, Governance, and Organizational Form
      • Chapter Preview
      • Managerial Challenge: Controlling the Vertical: Ultimate TV
      • Introduction
      • The Role of Contracting in Cooperative Games
      • Corporate Governance and the Problem of Moral Hazard
      • What Went Right / What Went Wrong: Moral Hazard and Holdup at Enron and WorldCom
      • The Principal-Agent Model
      • What Went Right / What Went Wrong: Why Have Restricted Stock Grants Replaced Executive Stock Options at Microsoft?
      • Choosing the Efficient Organizational Form
      • What Went Right / What Went Wrong: Cable Allies Refuse to Adopt Microsoft’s WebTV as an Industry Standard
      • International Perspectives: Economies of Scale and International Joint Ventures in Chip Making
      • Vertical Integration
      • What Went Right / What Went Wrong: Dell Replaces Vertical Integration with Virtual Integration
      • Summary
      • Exercises
      • Case Exercise: Borders Books and Amazon.com Decide to Do BusinessTogether
      • Case Exercise: Designing a Managerial Incentive Contract
      • Case Exercise: The Division of Investment Banking Fees in a Syndicate
    • 15A Auction Design and Information Economics
      • Optimal Mechanism Design
      • First-Come, First-Served versus Last-Come, First-Served
      • Auctions
      • Incentive-Compatible Revelation Mechanisms
      • International Perspectives: Joint Venture in Memory Chips: IBM, Siemens, and Toshiba
      • International Perspectives: Whirlpool’s Joint Venture in Appliances Improves upon Maytag’s Outright Purchase of Hoover
      • Summary
      • Exercises
      • Case Exercise: Spectrum Auction
      • Case Exercise: Debugging Computer Software: Intel
    • 16 Government Regulation
      • Chapter Preview
      • Managerial Challenge: Cap and Trade, Deregulation, and the Coase Theorem
      • The Regulation of Market Structure and Conduct
      • Antitrust Regulation Statutes and Their Enforcement
      • Antitrust Prohibition of Selected Business Decisions
      • Command and Control Regulatory Constraints: An Economic Analysis
      • What Went Right / What Went Wrong: The Need for a Regulated Clearing house to Control Counterparty Risk at AIG
      • Regulation of Externalities
      • Governmental Protection of Business
      • The Optimal Deployment Decision: To License or Not
      • What Went Right / What Went Wrong: Delayed Release at Aventis
      • What Went Right / What Went Wrong: Technology Licenses Cost Palm Its Leadin PDAs
      • What Went Right / What Went Wrong: Motorola: What They Didn’t Know Hurt Them
      • Summary
      • Exercises
      • Case Exercise: Microsoft Tying Arrangements
      • Case Exercise: Music Recording Industry Blocked from Consolidating
    • 17 Long-Term Investment Analysis
      • Chapter Preview
      • Managerial Challenge: Multigenerational Effects of Ozone Depletion and Greenhouse Gases
      • The Nature of Capital Expenditure Decisions
      • A Basic Framework for Capital Budgeting
      • The Capital Budgeting Process
      • Estimating the Firm’s Cost of Capital
      • Cost-Benefit Analysis
      • Steps in Cost-Benefit Analysis
      • Objectives and Constraints in Cost-Benefit Analysis
      • Analysis and Valuation of Benefits and Costs
      • The Appropriate Rate of Discount
      • Cost-Effectiveness Analysis
      • Summary
      • Exercises
      • Case Exercise: Cost-Benefit Analysis
      • Case Exercise: Industrial Development Tax Relief and Incentives
  • APPENDICES
    • A The Time Value of Money
    • B Tables
    • C Differential Calculus Techniques in Management
    • D Check Answers to Selected End-of-Chapter Exercises
    • Glossary
    • Index
    • Notes