Poli Essay

1

Introduction

In late-century Africa, things fell apart. By way of illustra-tion, consider Figure 1.1, which lists civil wars in African countries from 1970 to 1995, as judged by the World Bank.

As time passes, the list grows. Angola, Chad, Namibia,

Nigeria, and Sudan enter the 1970s war-torn; in the mid-1970s,

Sudan exits the list, but Equatorial Guinea and Zimbabwe join

it; by 1980, Zimbabwe departs from the ranks of the war-torn,

but is replaced by Mozambique, Nigeria, and Uganda. The

pattern – a few dropping off, a larger number entering in –

continues into the early 1990s. Only one country that was con-

flict ridden in 1990 becomes peaceful by 1992, while eleven

others crowd into the ranks of Africa’s failed states.

Humanitarians, policymakers, and scholars: Each de-

mands to know why political order gave way to political con-

flict in late-century Africa. Stunned by the images and realities

of political disorder, I join them in search of answers. In so

doing, I – a political scientist – turn to theories of the state and

3

https://doi.org/10.1017/CBO9781316423974.001 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

year 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95

Burundi

Chad

Congo

Djibouti

Ethiopia

Kenya

Liberia

Mali

Mozambique Namibia

Nigeria Rwanda

Senegal

Sierra Leone Somalia Sudan

Uganda Congo

Zimbabwe

Figure 1.1. Civil wars, Africa 1970–1995. Source: World Bank (Sambanis 2002).

4

https://doi.org/10.1017/CBO 9781316423974.001

to the Cam bridge Core term

s of use, available at https://w w

w .cam

bridge.org/core/term s.

D ow

nloaded from https://w

w w

.cam bridge.org/core. Access paid by the U

C San D iego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

locate the sources of political disorder midst the factors that

lead states to break down.

I anchor this book in the work of Weber (1958) and view

coercion as the distinctive property of politics. As will become

clear in the next chapter, I depart from Weber – and his “struc-

turalist” descendants1 – by turning to the theory of games.

Driven by the realities of Africa, I view political order as

problematic: In light of the evidence Africa offers, political

order cannot be treated as a given. Rather, I argue, it results

when rulers – whom I characterize as “specialists in violence” –

choose to employ the means of coercion to protect the creation

of wealth rather than to prey upon it and when private citizens

choose to set weapons aside and to devote their time instead

to the production of wealth and to the enjoyment of leisure.2

When these choices constitute an equilibrium, then, I say,

political order forms a state.3

To address the collapse of political order in late-century

Africa, I therefore return to theory – the theory of the state – and

to theorizing – the theory of games. I do so because proceeding

in this fashion points out the conditions under which political

order can persist – or fail. I devote Chapter 2 to an informal

1 Evans, P., T. Skocpol, and D. Rueschmeyer (1985), Bringing the State Back In, Cambridge, U.K.: Cambridge University Press provides perhaps the best-known example.

2 I am drawing on Bates, R. H., A. Greif, et al. (2002), “Organizing Violence,” Journal of Conflict Resolution 46(5): 599–628.

3 The ambiguous phrasing is intended.

5

https://doi.org/10.1017/CBO9781316423974.001 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

derivation of those conditions. In the remaining chapters, I

turn from deduction to empirics and explore the extent to

which these conditions were to be found, or were absent, in

late-century Africa. The evidence leads me to conclude that

in the 1980s and 1990s, each of three key variables departed

from the levels necessary to induce governments and citizens

to choose in ways that would yield political order.

The Literature

Following the outbreak of conflict in Serbia, Somalia, Rwanda,

and elsewhere, the study of political violence has once again

become central to the study of politics. Familiar to many, for

example, would be the attempts by Collier and Hoeffler (2004)

and Fearon and Laitin (2003) to comprehend the origins of civil

wars. Also familiar would be studies of the impact of ethnic-

ity (Fearon and Laitin 2003), democracy (Hegre, Gates et al.

2001; Hegre 2003), and natural-resource endowments (e.g.,

Ross 2004). In my attempts to comprehend why things fell

apart in late-century Africa, I draw upon these writings. But I

also take issue with them, for virtually all share common prop-

erties from which I seek to depart.

Consider, for example, the assumption that civil war can be

best treated as the outcome of an insurgency. When thinking

about the origins of political disorder in Africa, I can find no

way of analyzing the origins of insurrection without starting

6

https://doi.org/10.1017/CBO9781316423974.001 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

with the behavior of governments. The conditions that led

to the breakdown of order in Africa include the authoritarian

nature of its states and their rulers’ penchant for predation. By

rendering their people insecure, they provoked insurgencies.

While both insurrectionaries and incumbents must necessar-

ily feature in the analysis of political disorder, in this instance it

makes sense not to focus exclusively on the rebels but to stress

as well the behavior of those whom they seek to drive from

power.

Recent contributions exhibit a second common feature:

the methods that they employ. Utilizing cross-national data,

they apply statistical procedures to isolate and measure the

relationship of particular variables with the onset and duration

of civil wars. I, too, make use of cross-national data; but rather

than collecting data for all countries in the globe, I restrict my

efforts to Africa. I do so in part because Africa provides an

unsettling range of opportunities to explore state failure and

because political disorder is so important a determinant of the

welfare of the continent. I also do so because I find it necessary

to draw upon my intuition. To employ that intuition, I need

first to inform it, be it by immersing myself in the field or in

qualitative accounts set down by observers. I have therefore

made use of a selected set of cases – those from the continent

of Africa – and my knowledge of their politics.4

4 The use of a subset of countries also eases the search for exogenous vari- ables, and thus causal analysis. For example, given the small size of Africa’s

7

https://doi.org/10.1017/CBO9781316423974.001 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

Lastly, if only because they are based on the analysis of

cross-national data, contemporary studies exhibit a third

property: Their conclusions take the form of “findings.” These

findings are based upon relationships between a selection of

key variables and the outbreak or duration of civil wars. Collier

and Hoeffler (2004), for example, stress the importance of

“opportunities,” that is, chances to secure economic rewards

and to finance political organizations. Noting that the magni-

tude of primary product exports, the costs of recruiting, and

access to funding from diasporas relate to the likelihood of

civil war, they conclude that “economic viability appears to be

the predominant systematic explanation of rebellion” (p. 563).

Fearon and Laitin (2003), by contrast, conclude that “capa-

bilities” play the major role: “We agree that financing is one

determinant of the viability of insurgency,” they write (p. 76).

But they place major emphasis on “state administrative, mil-

itary, and police capabilities” (p. 76), measures of which bear

significant relationships to the outbreak of civil wars in their

global set of data.

In this work, I proceed in a different fashion. I start by

first capturing the logic that gives rise to political order. While

I, too, test hypotheses about the origins of disorder, I derive

economies, I can treat global economic shocks as exogenous – something that yields inferential leverage when seeking to measure the impact of economic forces on state failure.

8

https://doi.org/10.1017/CBO9781316423974.001 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

these hypotheses from a theory. By adopting a more deductive

approach, I depart from the work of my predecessors.

Key Topics

Energized by such works as Kaplan’s “The Coming Anarchy”

(1994), students of Africa have focused on the relationship

between ethnic diversity and political conflict. At least since

the time that William Easterly and Ross Levine penned “Africa’s

Growth Tragedy” (1997), empirically minded social scientists

have sought to capture the impact of ethnicity on the eco-

nomic performance of Africa’s states. Interestingly, however,

they have found it difficult to uncover systematic evidence of

the relationship between measures of ethnicity and the likeli-

hood of political disorder.5

In this study I, too, find little evidence of a systematic rela-

tionship. And yet, the qualitative accounts – be they of the

killing fields of Darfur or of the tenuous peace in Nigeria – con-

tinue to stress the central importance of ethnicity to political

life in Africa. In response, I argue that ethnic diversity does

not cause violence; rather, ethnicity and violence are joint

5 For a discussion, see Bates, R. H., and I. Yackolev (2002), Ethnicity in Africa, in The Role of Social Capital In Development, edited by C. Grootaert and T. van Bastelaer, New York: Cambridge University Press; and Fearon, J., and D. Laitin (2003), “Ethnicity, Insurgency and Civil War,” American Political Science Review 97(1): 75–90.

9

https://doi.org/10.1017/CBO9781316423974.001 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

products of state failure. Their relationship is contingent: It

occurs when political order erodes and politicians forge polit-

ical organizations in the midst of political conflict.

The political significance of resource wealth has also

attracted much attention. Analyzing their data on civil wars,

Collier and Hoeffler (2004) report that “dependence upon pri-

mary commodity exports” constituted “a particularly power-

ful risk factor” for the outbreak of civil war (p. 593). Africa

is, of course, noted for its bounteous natural endowments of

petroleum, timber, metals, and gemstones. And scholars and

policymakers have documented the close ties between the dia-

mond industry and UNITA (National Union for the Total Inde-

pendence of Angola) in Angola (Fowler 2000), the smuggling

of gemstones and the financing of rebels in Sierra Leone (Reno

2000), and the mining of coltan and the sites of rebellion in

eastern Zaire (present-day Democratic Republic of the Congo)

(Kakwenzire and Kamukama 2000).

And yet, using Collier and Hoeffler’s (2004) own data,

Fearon (2005) has demonstrated that their findings are frag-

ile, depending in part on decisions about how to measure

and classify cases. In this study, too, I fail to find a signifi-

cant relationship between the value of natural resources and

the likelihood of state failure.6 Once again, then, there arises

6 For both Fearon (2005) and myself (this work), only the value of petroleum deposts is related to political disorder. Even here the relationship is fragile, however.

10

https://doi.org/10.1017/CBO9781316423974.001 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

a disparity between the evidence from cross-national regres-

sions and that from qualitative accounts. I shall argue that the

disparity suggests that the exploitation of natural resources

for war finance is a correlate rather than a cause of political

disorder.

A third factor plays a major role in the literature: democ-

ratization. Qualitative accounts, such as those of Mansfield

and Snyder (Mansfield and Snyder 1995; Snyder 2000) sug-

gest that democratization produces political instability and

leads to the mobilization of what Zakaria (1997) calls “illib-

eral” political forces. Careful empirical researchers, such as

Hegre (Hegre, Gates et al. 2001; Hegre 2004), confirm that new

democracies and intermediate regimes – those lying some-

where between stable authoritarian and consolidated demo-

cratic governments7 – exhibit significantly higher rates of civil

war. As demonstrated by Geddes (2003), many of these inter-

mediate regimes are the product of the “third wave” of democ-

ratization (Huntington 1991) and the collapse of communist

regimes and are therefore themselves new and vulnerable to

disorder.

In the 1980s and 1990s, many of Africa’s governments

reformed. Regimes that once had banned the formation

of political parties now faced challenges at the polls from

7 Using Polity coding. Available online at: http://www.cidcm.umd.edu/ polity/.

11

https://doi.org/10.1017/CBO9781316423974.001 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

candidates backed by an organized political opposition. And

in the late 1980s and early 1990s, militias assembled, states

failed, and Africa faced rising levels of political disorder. The

experience of Africa thus appears to conform to what the liter-

ature has recorded: Electoral competition and state failure go

together.

In analyzing the impact of political reform, I employ two

measures: the movement from military to civilian rule and the

shift from no- or one- to multiparty systems. In discussions of

democracy, the followers of Schumpeter (1950) argue for the

sufficiency of party competition; those of Dahl (1971) contend

that party competition is necessary but not sufficient. Without

an accompanying bundle of political and civil rights, the latter

argue, contested elections are not of themselves evidence of

democratic politics. In debates over the relationship between

party systems and democracy, I concur with the followers of

Dahl. When addressing political reform, I pay no attention to

the number of political parties, their relative vote shares, or

the conditions under which the opposition is allowed to cam-

paign. I therefore address not the relationship between democ-

racy and political conflict but rather the relationship between

political reform and political disorder.

Lastly, there are those who emphasize the impact of pov-

erty. That poverty and conflict should go together is treated

as noncontroversial, as if disorder were simply an expected

12

https://doi.org/10.1017/CBO9781316423974.001 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

corollary of the lack of economic development.8 But consider:

If, as many argue, lower per capita incomes imply lower wages

and therefore lower costs of rebellion, so too do they imply

fewer gains from predation; income thus cancels out the ratio

between the costs and benefits. From the theoretical point of

view, moreover, there is simply little that can be said about the

relationship between the average level of income – or, for that

matter, poverty – and incentives for violence. As I will argue

in Chapter 2, for our purposes, discussions of private income

can be set aside; for the logic of political order suggests that

the focus be placed not on private income but rather on public

revenues. Economic shocks will indeed play a major role in this

analysis, but the focus will be on their impact on the revenues

of states, not on the incomes of individuals.9 In this work, when

I measure the impact of income per capita, I treat it as a control

variable, rather than as a variable of theoretical interest.

In Chapter 2, I parse the logic of political order. I recount the

theory informally, portraying the interaction between govern-

ments and citizens and among citizens as well. Presented as a

8 Indeed, see Sambanis, N., and H. Hegre (2006), “Sensitivity Analysis of Empirical Results on Civil War Onset,” The Journal of Conflict Resolution 50(4): 508–35. The authors point to per capita income as one of the very few variables that bears a robust relationship with civic violence.

9 See the arguments in Hirshleifer, J. (1995), Theorizing About Conflict, in Handbook of Defense Economics, edited by K. Hartley and T. Sandler, New York: Elsevier.

13

https://doi.org/10.1017/CBO9781316423974.001 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

fable, the argument is based upon rigorous foundations and

points to the conditions under which governments choose to

engage in predation and citizens choose to take up arms.10

Chapters 3 through 5 set out the conditions that prevailed

prior to the collapse of political order. They document the

social and political configurations that were in place at the

time of the impact of the economic and political shocks that

dismantled the state in Africa. In Chapter 6, states fracture

and political disorder engulfs nations in Africa. Chapter 7

concludes.

10 The informed reader will note the parallels between my analysis and that of Azam, J.-P., and A. Mesnard (2003), “Civil War and the Social Contract,” Public Choice 115(3–4): 455–75; Snyder, R., and R. Bhavani (2005), “Diamonds, Blood and Taxes: A Revenue-Centered Framework for Ex- plaining Political Order,” The Journal of Conflict Resolution 49(4): 563– 97; and Magaloni, B. (2006), Voting for Autocracy, New York: Cambridge University Press.

14

https://doi.org/10.1017/CBO9781316423974.001 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

2

From Fable to Fact

Idevote this chapter to the exposition of a fable.1 Whilediminutive, it is incisive: It captures the incentives that drive the choices that lead to the failure of states. It is also

suggestive, for it points to the conditions under which polit-

ical order should, or should not, prevail. After expositing this

fable, I determine whether it is also informative. It can be

so only insofar as the forces that animate its central char-

acters find their parallel in late-century Africa. I devote the

last portions of the chapter to arguing that they do and that

the story communicated by the fable can therefore bear the

weight of the tragedy that befell the continent. The fable can

be used – with help – to explore the foundations of political

disorder.

1 A rigorous presentation appeared as Bates, R. H., A. Greif, et al. (2002), “Organizing Violence,” The Journal of Conflict Resolution 46(5): 599– 628.

15

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

A Fable

Consider the following scenario: A community is peopled by a

“specialist in violence” and two groups of citizens. Headed by

powerful patrons, the groups can act in a unified manner.2 The

specialist in violence earns his living from the use of force; he

either seizes the wealth of others or pockets funds they pay for

their protection. Sheltered behind their patrons, the citizens

generate incomes by engaging in productive labor; but they

too can be mobilized either to seize the income of others – or

to defend their incomes from seizure. The three personages in

this drama repeatedly interact over time. The question is: Can

political order prevail in such a setting?

The answer is: Yes. Under certain circumstances, the spe-

cialist will chose to use his control of the means of violence to

protect rather than to despoil private property. And the groups

of citizens will chose to devote their time and energies to labor

and leisure and forswear the use of arms, while rewarding the

specialist in violence for protecting them against raids by oth-

ers. In addition, under certain well-specified conditions, these

choices will persist in equilibrium, rendering political order a

state.

The primary reason for this outcome is that the players

interact over time. The specialist in violence and political

2 That is, they have solved the collective action problem.

16

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

From Fable to Fact

organizations can therefore condition their future choices on

present behavior; that is, they can make threats and inflict pun-

ishments and thus shape the behavior of others. Should one

group raid or withhold tax payments, the specialist can retal-

iate by changing from guardian to predator. And should the

specialist opportunistically seize the wealth of the member of a

group, his defection would trigger punishment by that citizen’s

confederates: They can withhold tax payments or mobilize for

fighting. If not sufficiently paid for the provision of security,

the specialist in violence can pay himself: he can turn from

guardian to warlord. And if preyed upon or left undefended,

then the citizens can furnish their own protection; they can

take up arms.

When both the specialist and the citizens turn to pun-

ishment, political order breaks down. People become inse-

cure. They also become poor; having to reallocate resources

to defense, they have fewer resources to devote to produc-

tive activity. The resultant loss of security and prosperity stays

the hand of a specialist in violence who might be tempted to

engage in predation or of a group that might be tempted to

forcefully seize the goods of another or withhold tax payments,

thus triggering political disorder.

To better grasp the incentives that animate this story, focus

on the choices open to the specialist in violence, as commu-

nicated in Figure 2.1. In this figure, the vertical axis repre-

sents monetary gains or losses. The further above zero, the

17

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

Payoffs

+

0

-

Time

Payoffs on the equilibrium path

Payoffs from defection and subsequent punishment

Figure 2.1. Payoffs from strategy choices.

greater the payoffs; the further below, the greater the losses.

The horizontal axis designates time, with the more immediate

payoffs occurring near the origin and the more distant ones

further to the right. The dotted line represents the flow of pay-

offs that result from tax payments; the flow is steady, mod-

erate, and positive in value. The dashed line represents the

18

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

From Fable to Fact

flow of payoffs that result from predation. Predation yields an

immediate benefit: The dashed line leaps above the dotted

line, indicating that the income from predation significantly

exceeds that from tax payments. But that one period spike

then gives way to a stream of losses, as illustrated by the plunge

below the zero point that separates gains from losses. Insofar

as a decision maker is forward looking, the losses that accrue

in the punishment phase caste a shadow over the returns from

defection and so temper any wish to engage in predation.

If summed over time, each line – that representing the

returns to taxation and that the returns to predation – yields an

expected payoff. What would determine their magnitudes? In

particular, what would determine whether the value of the vari-

able path, generated by predation, will be more or less attrac-

tive than that of the steady path, generated from tax payments?

The factors that determine the relative magnitude of these pay-

offs determine whether the specialist in violence will adhere

to the path of play and continue to behave as guardian or veer

from that path, engage in predation, and trigger the re-arming

of the citizenry and subsequent disorder.

The Conditions of Political Order

One factor is the level of tax revenue. If too low, the benefits of

predation may be tempting despite the subsequent costs.3 A

3 But they may also be if too high. See the discussion in Bates, R. H., A. Greif, et al. (2002), “Organizing Violence.”

19

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

second is the magnitudes of the rewards that predation might

yield. If sufficiently bounteous, the specialist in violence might

choose to deviate despite the losses. A third is the special-

ist’s rate of discount. A specialist in violence who is impatient,

greedy, or insecure will discount the future payoffs that accrue

along the path of play; and she will also discount the penal-

ties that follow an opportunistic deviation. She may therefore

find the prospect of predation more attractive than if she were

patient, prosperous, or secure.

The fable thus suggests that the possibility of political order

rests on the value of three variables: the level of public revenues,

the rewards from predation, and the specialist’s rate of dis-

count. The interplay of these forces helps to determine whether

governments safeguard or prey upon the wealth of the land;

whether groups of citizens take up arms; and whether there is

political order – or state failure.

The tale may be engaging; elsewhere it has been shown

to be logically consistent (Bates, Greif et al. 2002). But it is

informative only insofar as it captures and incorporates key

features of Africa’s political landscape. Only insofar as it does

so will it offer insight into the tribulations of that continent.

Features of Late-Century Politics

Recall that the scenario was populated by a specialist in vio-

lence and by citizens who could, should they choose, take up

20

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

From Fable to Fact

0� 5�

10� 15� 20� 25� 30� 35� 40�

1970-� 74 �

1�975�-� 79�

1980-� 84�

1985�-� 89�

1990-� 95�

Percentage�

Figure 2.2. Percentage country years in which country ruled by mili-

tary head of state.

arms. Now note a characteristic feature of late-century poli-

tics in Africa: A significant portion of Africa’s states were ruled

by their military. Turning to Figure 2.2, we find that from the

beginning of the 1970s to the end of the 1980s, in more than

30 percent of the observations, Africa’s heads of state came

from the armed forces.4 In the 1990s, U.S. president William

Clinton and British prime minister Tony Blair heralded the

emergence of a “new generation” of African rulers – Yoweri

Museveni in Uganda, Paul Kagame in Rwanda, Meles Zenawi

in Ethiopia, and Isaias Afwerki in Eritrea – while failing to men-

tion that each had come to power as the head of an armed

insurgency. In many states, then, power came from the barrel

of a gun (Ottaway 1999).5

4 For details of the sample, see Table A.1 in the Appendix. 5 Lest readers regard the link between coercion and politics to be distinc-

tive of politics in Africa, they might first recall the note sent by the father

21

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

Not only were heads of states specialists in violence, the cit-

izens, too, frequently took up arms. By way of illustration, con-

sider the case of Chad. At the beginning of our sample period,

1970, Francois Tombalbaya, head of the Parti Progressive Tcha-

dien (PPT), was president of Chad. Tombalbaya belonged to

the Sara, an agriculturalist people in the southern portions of

the country; the eastern and northern portions were popu-

lated by pastoralist peoples. As Tombalbaya consolidated his

rule, he posted administrators from the south to govern these

other regions. There they imposed policies designed to propa-

gate Sara culture and imposed new taxes on cattle. In response,

the pastoralists mounted protests, fomented riots, and formed

militias: the Front for the Liberation of Chad (FLT) in the east

and the Front for National Liberation (FROLINAT) in the north.

It was only by calling for military assistance from France that

Tombalbaya remained in power.6

of Frederick the Great to the young man’s tutors: “[I]n the highest mea- sure . . . instill in my son a true love of the military . . . and impress on him that nothing in the world can give a prince such fame and honor as the sword and that he would be the most despicable creature on earth if he did not revere it and seek glory from it. . . . ” (p. 18 of Asprey, R. B. (1986), Frederick the Great, New York: Ticknor and Fields). Recall, too, the rueful words of the dying Louis IV: “I have loved war too much.” (http://encarta.msm.com).

6 For accounts, see Buijtenhuijs, R. (1989), Chad, in Contemporary West African States, edited by D. B. Cruise O’Brien, J. Dunn, and R. Rath- bone, Cambridge, U.K.: Cambridge University Press; May, R. (2003), Internal Dimensions of Warfare in Chad, in Readings in African Poli- tics, edited by T. Young, Oxford: James Currey; Lemarchand, R. (1981), “Chad: The Roots of Chaos,” Current History (December); Nolutshungu,

22

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

From Fable to Fact

Whereas the militarization of Chad marks the opening of

the sample period, conflict between militias in Congo (Braz-

zaville) marks its end. In 1992, a southerner, Pascal Lissouba,

became president of Congo(B); in the run up to the next pres-

idential election, the strongman and former president, Denis

Sassou-Nguesso, declared his candidacy. As political tensions

mounted, each politician mobilized a private army: the Cobras,

who supported Sassou-Nguesso, and the Zulus, who sup-

ported Pascal Lissouba. Kindled in the provincial towns, fight-

ing between these groups erupted in the capital where the

mayor, Bernard Kolelas, had organized his own militia, the

Njinjas. Combat between these militias lay waste to one of

the major cities of French-speaking Africa.7

As seen in Figure 2.3, over the course of the sample period

1970–1995, reports of the formation of militias became more

common. With increasing frequency, citizens took up arms

and states lost their monopoly over the means of violence.

The scenario depicted at the outset of this chapter

thus incorporates two major features of the politics of late

S. C. (1996), Limits of Anarchy, Charlottesville: University Press of Virginia; and Azam, J.-P. (2007), The Political Geography of Redis- tribution, Chap. 6 in The Political Economy of Economic Growth in Africa, 1960–2000: An Analytic Survey, edited by B. Ndulu, P. Collier, R. H. Bates, and S. O’Connell, Cambridge, U.K.: Cambridge University Press.

7 One of the best accounts appears in Bazenguissa-Ganga, R. (2003), The Spread of Political Violence in Congo-Brazzaville, in Readings in African Politics, edited by T. Young, Oxford: James Currey.

23

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

.1

.2

.3

.4

.5

1970 1975 1980 1985 1990 1995 year

95% CI Fitted values

Figure 2.3. Reports of militias by year, percent of observations.

twentieth-century Africa: rule by specialists in violence and

the militarization of civic society. In accounting for political

disorder, it pointed to three key variables: the level of public

revenues received by governments; the magnitude of tempta-

tions they face, as determined by the rewards for predation;

and the relative weight placed upon them. A moment’s reflec-

tion leads to the recognition of the possible significance of

these variables for the politics of late-century Africa.

Revenues

In the 1970s, a sharp increase in the price of oil triggered global

recession. The increased price of energy led to higher costs of

production in the advanced industrial economies, resulting in

24

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

From Fable to Fact

the laying off of labor and a lowering of incomes. For Africa,

the result was a decrease in the demand for exports.

In Africa, as in many other developing regions, taxes on

trade constitute one of the most important sources of public

revenue. As the value of exports from Africa declined, so too

did the taxes collected by Africa’s governments. In the latter

decades of the twentieth century, then, while Africa’s people

faced a “growth tragedy” (Easterly and Levine 1997), its states

faced a crisis of public revenues. The break in the global econ-

omy was sharp and unanticipated; and the recovery of pub-

lic finance required comprehensive and protracted restructur-

ing, involving changes not only in tax rates but also in policies

toward trade and industrial development.

The economic forces at play in late-century Africa thus

aligned with the conditions in the fable, reducing the revenues

of governments. Within the framework of the fable, the decline

in public revenues represents a decline in the rewards from

public service. In the face of such a reduction, those who con-

trol the means of violence find the income derived from the

protection of civilians declining relative to the returns from

predation. By the logic of the fable, they would therefore be

more likely to turn to predation. Rather than providing secu-

rity, those who controlled the state would become a source of

insecurity, as they sought to extract revenue from the wealth

of their citizens.

25

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

Discount Rate

In the fable, if the government becomes more impatient or

insecure, the rewards that accrue to those who act as guardians

decline in value; so, too, the penalties that would be imposed

were they to revert to predation. As the “shadow of the future”8

thus dissipates, the level of temptation rises: Immediate ben-

efits weigh more heavily than future losses, and incumbents

may become more predatory, provoking state failure.

Returning to the empirical record, in the late 1980s, Africa

underwent a period of political reform. With the end of the

Cold War, the “third wave” of democratization9 swept across

the continent and governments that in the 1980s had been

immune to political challenges now faced organized polit-

ical opponents. As seen in Figure 2.4, whereas from the

early 1970s to the mid-1980s, more than 80% of the country-

year observations contained no- or one-party systems, by the

mid 1990s, more than 50% experienced multiparty systems.

With the shift to multiparty politics, those who presided over

Africa’s authoritarian governments faced an unanticipated

increase in the level of political risk. Few had prepared them-

selves to compete at the polls; some surely would have chosen

8 The phrase comes from Axelrod, R. (1985), The Evolution of Cooperation, New York: Basic Books.

9 Huntington, S. P. (1991), The Third Wave, Norman, OK: Oklahoma Uni- versity Press.

26

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

From Fable to Fact

0

10

20

30

40

50

60

1970- 74

1975- 79

1980- 84

1985- 89

1990- 95

Percentage of Country Years

No-Party

One-Party

Multiparty

Figure 2.4. Political competition over time.

to govern with more restraint had they known that they might

someday be forced from political office and shorn of the pro-

tection it afforded. Incumbents became less secure. And by the

logic of the fable, they would therefore find the modest rewards

that accrue to political guardians less attractive, and the fear

of future punishment less daunting, increasing the temptation

to engage in predation.

Resources

To a degree that exceeds any other region of the world, the

economies of Africa are based on the production of precious

minerals, gemstones, petroleum, and other precious com-

modities. These resources pose a constant temptation to those

with military power. Were they to shift from guardian to preda-

tor, their future prosperity would nonetheless be ensured,

underpinned by the income generated by natural resources.

27

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Introduction

Consider the case of Nigeria, where, in the words of Bill

Dudley (1982, p. 92): “[T]he oil boom was a disaster . . . ” – one

made worse by military rule. As Dudley states:

[T]he effect of the oil boom was to convert the military polit-

ical decision-makers . . . into a new property-owning, rentier

class working in close and direct collaboration with foreign

business interests with the sole aim of expropriating the sur-

pluses derived from oil for their private and personal benefit

(Dudley 1982, p. 116).

Consider, too, the Sudan or Chad, following the discov-

ery of oil. In both, incumbent regimes turned to repression,

the one harrying the Dinka and the other the Sara. Resource

wealth thus appears to shape the behavior of elites. In the

face of dwindling public resources or insecure political futures,

given the availability of wealth from appropriable resources,

they could greet with equanimity a future of political disorder.

Those immersed in environments richly endowed by nature

would therefore be willing to take actions that rendered others

insecure, thus triggering state failure.

Conclusion

The logic of the fable highlights the importance of public

revenues, democratization, and natural resources and the

manner in which they impinge upon the possibility of political

order. As we have seen, the elements that affect political order

28

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

From Fable to Fact

in the fable parallel political forces that shaped the politics of

the continent in the later decades of the twentieth century.

While many who have studied Africa have emphasized

the political importance of economic collapse, the “resource

curse,” and the relationship between political competition and

political conflict, this account focuses on the logic that system-

atically links these forces to the political incentives that under-

lie state failure. Being abstract, the logic is also adaptable; it

can play out in a variety of forms. Consider the nature of the

groups that may – or may not – transmute into militias. In

one setting, they may be the youth wings of political parties;

in another, regional coalitions; and in a third, ethnic groups.

The same applies to the specialists in violence. In some set-

tings, the military rule; clearly the military specialize in the

use of violence. In other instances, it is civilians who gov-

ern. Even a civilian head of state presides over police, public

prosecutors, and a prison system; by bringing them to bear

upon citizens, he too can transform the state into an instru-

ment for predation. In still other instances, the civil service

assumes the role of a specialist in violence, using its command

of the bureaucracy to redistribute income from the citizens

to themselves. Different actors can thus fulfill the major roles

in the fable, but their parts are inscribed in a common script.

By the choices they make, they animate the sources of political

order, induce state failure, thereby enacting the tragedy that

engulfed late-century Africa.

29

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

https://doi.org/10.1017/CBO9781316423974.002 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Part Two

Sowing the Seeds

https://doi.org/10.1017/CBO9781316423974 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

https://doi.org/10.1017/CBO9781316423974 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

3

Political Legacies

By convention, 1960 marks the year of independence inAfrica.1 Shortly after independence, Africa’s new states faced two withering critiques, one mounted by Franz Fanon

(1963) and a second by Rene Dumont (1962). Although their

indictments overlap, Fanon’s targeted their politics whereas

Dumont’s focused on their policies. In this chapter, I ana-

lyze the nature of post-independence politics, emphasizing

in particular the nature of political institutions. In Chapter 4,

I address the policies chosen by Africa’s governments in the

post-independence era.

As reported in Chapter 2, by the late 1970s, in more than

eighty percent of the country years,2 opposition parties failed

to challenge incumbent heads of state, most often because

it was illegal for them to do so (Figure 2.4), and in roughly

1 Of the forty-six states in our sample set of countries, only six had achieved independence prior to 1960; in 1960 alone, fifteen became sovereign.

2 The sample covers a panel of forty-six countries over twenty years. A single observation therefore constitutes a country year, e.g., Zimbabwe in the year 1970. Thus the origins of this awkward term.

33

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

one-third of the country years, military officers served as

heads of state (Figure 2.2). The political institutions of post-

independence Africa were thus authoritarian. For late-century

Africa, the consequence was an increased likelihood of politi-

cal disorder.

Throughout this chapter, I repeatedly draw illustrations

from Zambia’s political history. Box 3.1 provides a synopsis,

to which the reader may refer while seeking to master the sev-

eral narratives. Map 3.1 outlines the boundaries of Zambia’s

provinces, whose political leaders jockeyed for top positions

in the ruling party and national government.

The Incumbent’s Dilemma

When colonial regimes departed from Africa, they orchestrated

their retreat by holding elections and exiting midst the polit-

ical din. While competitors for office championed the cause

of independence and denounced the evils of colonialism, a

notable feature of their campaigns was the stress they placed

on seizing the “fruits of independence.”

In a careful study of the city of Abidjan, Michael Cohen

(1974) explores the use of power in Cote d’Ivoire. Rural back-

ers of the ruling party, he noted, used their political connec-

tions to move from provincial towns to the national capital

(Cohen 1974). Some had been appointed to the boards of state-

owned corporations, which produced “palm oil, hardwood,

34

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Political Legacies

Box 3.1. Political highlights, post-independence Zambia

– Zambia achieved independence in 1964, with the UNIP (the United National Independence Party) as the governing party and ANC (the African National Congress) as the opposition.

– High office in the governing party translated into high posts in the gov- ernment. Kenneth David Kaunda, president of UNIP, became presi- dent of Zambia as well, and Reuben Kamanga, a politician from the Eastern Province and vice president of UNIP, served also as vice president of Zambia.

– In 1967, UNIP held internal party elections. A Bemba-speaking bloc captured a majority of the seats in the Central Committee of the ruling party and Simon Kapwepwe, a Bemba-speaker from Northern Province, displaced Reuben Kamanga as vice president.

– In the subsequent general election, Barotse Province (also known as Western Province) joined the Central and Southern provinces in support of ANC.

– In 1969, the president dissolved the quarrelsome Central Committee of UNIP, Eastern Province politicians resumed their posts, and Reuben Kamanga returned as vice president.

– In 1971, the Bemba-speaking politicians, led by Simon Kapwepwe, defected from UNIP, the ruling party, and joined the opposition.

rubber . . . and construction equipment” (ibid., pp. 24–5). Oth-

ers received prized plots of land in the low-density town-

ships, where they built homes, and in the high-density areas,

where they constructed new enterprises. As they worked their

way up the political hierarchy, Cohen writes, the backers of

35

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

Map 3.1. Provinces of Zambia. Note: Ndola is the capital of the

Copperbelt. Source: www.answers.com/topic/ZM-Provinces.png.

the ruling party achieved even more desirable addresses.

“[A]dministrative and political control of urban land conces-

sions . . . turns out to be an extraordinarily sensitive measure

of political status within the ruling class,” he writes: “Admin-

istrative appointments or promotions are often accompanied

by approval of an individual’s application for land. . . . ” (ibid.,

pp. 44–5). Cohen concludes with a depiction of a housing pyra-

mid, in which the “ministers live in luxurious European-style

villas” (p. 47) while their subordinates dwelt in “smaller but

36

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Political Legacies

very luxurious homes in Cocody,” a prosperous suburb (p. 48).

To the powerful, he writes, went the rewards: “[T]he winning

coalition” used its power to achieve “wealth and position . . . ”

(p. 6).

Cote d’Ivoire achieved independence in 1960; Zimbabwe,

two decades later. As documented by Norma Kriger (2003),

freedom fighters, political organizers, and rank-and-file mem-

bers of Zimbabwe’s ruling party began agitating for the rewards

of independence. Under intense political pressure from the

ruling party, the Ministry of Home Affairs hired 3,500 freedom

fighters; the Ministry of Local Government, 2,600 more. The

Ministry of Health had to sign on 2,000 and the Central Intelli-

gence Organization more than 1,000 (ibid., p. 178). Once they

secured jobs, Kriger writes, the militants agitated for additional

benefits: compensation for losses incurred during the strug-

gle for independence, pensions, loans, and land. The political

movement that seized the state thus subsequently “built a vio-

lent and extractive political order” (ibid., p. 5), as the victors

continued to agitate for the fruits of independence.

The pattern has been documented for socialist Zambia

(Szeftel 1978) as well as capitalist Nigeria (Schatz 1977). As

described by Dumont (1962) and Fanon (1963), independence

represented the capture of the state by local political elites who

then used power to accumulate wealth.

The ambitions of the elites was equaled by the aspirations

of the electorate. Thus Barkan, in his study of elections in

37

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

post-independence Kenya (1976, 1986); Hayward and Kandeh,

in their study of Sierra Leone (1987); and Hayward, when he

turned to the study of Ghana (1976), report that constituents

viewed politicians as their agents whose job it was to bring

material benefits to the local community – jobs, loans, or cash.

Those in Kenya, Barkan notes, stoked the fires of political ambi-

tion, inciting candidates to bid for political support by con-

tributing funds for the construction of local projects (Barkan

1976). The result, as Allen writes of Benin, was “the exchange

of blocs of votes . . . for valued goods. . . . ” (1989, p. 22). Com-

petitive elections came to resemble a political marketplace, in

which votes were exchanged for material benefits.

Analysis

In 1983, Gerald Kramer (1983) explored the nature of political

competition in a world in which incumbents and challengers

compete by distributing material benefits.3 In his analysis, the

voters value private consumption and party labels and the

politicians control a fixed stock of material goods. In the com-

petition for votes, the incumbents move first: They distribute

benefits in a way designed to return to office, while preserving

as large a portion as possible for their own consumption. Once

3 See also Groseclose, T., and J. M. Snyder (1996), “Buying Supermajorities,” American Political Science Review 90(2): 303–15.

38

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Political Legacies

the governing party has proposed its allocation, the opposition

then responds with a counteroffer. In this competition, Kramer

asks, how will incumbents and challengers behave? How will

they play the game?

When seeking to unseat the incumbent and to do so at least

expense, Kramer argues, the challenger will bid for the support

of those who may be disadvantaged under the incumbent’s

rule. By offering slightly more than what the incumbent has

provided, the challenger can capture their votes and weaken

the incumbent’s coalition. He can then devote the rest of his

resources to obtaining the additional votes necessary to secure

a majority. The costs of this strategy will of course be higher the

greater the degree to which the voters identify with the party

in power.

Anticipating the strategy of the challenger, Kramer argues,

the incumbent’s best strategy will be to distribute benefits

widely. Should he fail to give a segment of the electorate bene-

fits equal to those enjoyed by others, then he simply will have

lowered the costs to the challenger of assembling a sufficient

number of votes to unseat him.4 The incumbent will therefore

distribute his resources uniformly across all members of the

electorate.

4 In addition, if he spends more on one segment of the electorate than upon others, he could lower his own costs – and increase the resources that he could retain for his own consumption – by reducing the differential.

39

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

Turning to the voters, Kramer advances the argument an

additional step by asking: What if they were to behave stra-

tegically? What if they were to back political parties instru-

mentally, rather than out of an unreasoned sense of loyalty?

Behaving strategically, Kramer argues, the voters, in pursuit of

private benefits, would reduce their level of party loyalty. The

incumbent can purchase the votes of those who strongly iden-

tify with the ruling party relatively cheaply; the support of those

less loyal would command a higher price. As Kramer demon-

strates, when the voters learn to play the system to their advan-

tage they will then extract all the benefits on offer. Thus the

incumbent’s dilemma: Pursuing power to accumulate wealth,

they find themselves having to surrender their ill-gotten gains

to retain political office.

Did not the history of political competition in Zambia lend

support to Kramer’s argument, it would be easy to dismiss his

analysis as overstylized, abstract, and therefore divorced from

the realities of African politics.

The Example of Zambia

When Zambia became independent in 1964, it was governed

by UNIP (the United National Independence Party), which

had won majorities in all but Central and Southern provinces,

where the opposition ANC (the Africa National Congress)

held sway. In local council elections, legislative elections,

40

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Political Legacies

bi-elections, and general elections, the governing party relent-

lessly targeted the opposition’s bailiwick. In each round of the

elections, it flooded the two provinces with organizers, provid-

ing them with housing, running money, and access to petrol

from the government’s stores. Most relevant for this discus-

sion was the theme of the government’s campaigns: “It pays to

belong to UNIP.” From the government’s point of view, those

who supported the opposition had merely increased the price

of their political loyalty. By refurbishing schools, grading roads,

and distributing public monies through local development

agencies, the government vigorously bid for votes from the

heartland of the opposition.

Naturally, political leaders in other regions deciphered the

lesson to be drawn from the government’s efforts. Most rele-

vant is the response of those from Luapula, a province long

loyal to the governing party. While politicians from the North-

ern Province dominated the Central Committee and therefore

the cabinet as well, the government built a well-surfaced road,

a railway, and an oil pipeline through Northern Province to the

coast. Political leaders from Luapula Province began to feel

that their colleagues from Northern Province were reaping a

disproportion of the benefits from holding office. By lowering

the level of their loyalty to UNIP, the politicians from Luapula

reasoned, they could increase the price of their support for

the incumbent regime and secure a larger share of the spoils

(Bates 1976). The flirtation of the “Luapulaists” with defection

41

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

adumbrated later revolts, as other regional blocs listed their

grievances and maneuvered to extract benefits from those in

power.

From the government’s point of view, the costs of retaining

office had risen. Threatened with additional provincial defec-

tions and thus with the loss of power, the president empan-

elled a commission to explore the electoral rules; he charged

the commission with enquiring into the merits of single-party

rule. As documented by Larmer (2006), the commission

solemnly convened hearings in each and every region. Having

heard testimony in favor and against the abolition of opposi-

tion parties, it sensibly performed the task for which it had in

fact been convened: It recommended that Zambia become a

one-party state.

While the case does not map as clearly onto the matrix of

Kramer’s model as does that of Zambia, the post-indepen-

dence politics of Benin suggests similar forces at play. “Re-

sources,” Allen writes (1989), “were necessarily limited, but ex-

pansion and retention of support implied an ever-increasing

pressure for allocation of resources. . . . ” (p. 25). The compe-

tition for support led to a twenty percent increase in public

employment and a forty percent increase in public expendi-

ture – all in the first five years of independence. But then the

government encountered a critical constraint: the unwilling-

ness of the central bank, which was controlled by France, to

underwrite further increases in spending. By the late 1960s,

42

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Political Legacies

Allen writes, it had become apparent to all that the system

based on the competitive supply of “pork” could no longer be

sustained (ibid.). The governing elite then put an end to elec-

toral competition.

As in Zambia, in Benin – and elsewhere – incumbents

formed single-party regimes. In other instances, and espe-

cially under military rulers, the incumbents formed no-party

systems. In the single-party regime, the cabinets were dom-

inated by top officials from the ruling party; in the no-party

system, the presidents formed cabinets as if picking a per-

sonal staff. In either case, in response to the crisis of clien-

telism, in Allen’s phrasing (Allen 1989), or to the high costs

of securing wealth from power, in the language of this study,

incumbents changed the structure of the political game. They

created authoritarian governments.

The New Political Game

Even after the banning of party competition, competitive polit-

ical forces remained, but they played out within the regime. It

was the head of state, rather than the voter, who now became

the object of competitive bidding, as minor apparatchiks jock-

eyed for recognition and competed for political favor and,

while doing so, marked down the price of their political loy-

alty. Political sycophancy replaced constituency service as the

best strategy for those with ambitions for office.

43

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

Given the new structure of political competition, it was the

supplier rather than the demander of political favors who now

held the advantage. In the game of authoritarian politics, the

head of state controlled both access to material benefits and

control of the means of coercion. And it was to the chief exec-

utive that wealth and power now flowed.

In most African states, major financial institutions fell

under the control of the chief executive. Allen (1989) notes

that presidents in Francophone West Africa kept the ministry

of planning in their portfolios, not because they were com-

mitted to the formulation of development plans but rather

because these ministries received, and disbursed, foreign aid;

by controlling them, the president controlled a major source

of foreign exchange. In the case of Benin, he noted, the foreign

aid channeled through this ministry totaled $600 million in

1980–83 and “thus matched the size of the recurrent budget”

(Allen 1989, p. 52). In countries outside of the Francophone

zone, the president often controlled the central bank. Accord-

ing to Erwin Blumenthal,5 the national bank of Zaire main-

tained such accounts in Brussels, Paris, London, and New York

registered in the name of the national president (Blumenthal

1982).

5 Blumenthal had been dispatched by the International Monetary Fund to restructure and manage the finances of Zaire.

44

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Political Legacies

In addition to the financial bureaucracy, the president con-

trolled the means of coercion. Policing remains a national, not

a local, activity throughout most of Africa. The office of the

president oversaw the ministry of interior. The attorney gen-

eral, the official prosecutor for the state; the special branch;

and the prison system – in most countries, these agencies

lodged within the office of the president. In addition, the pres-

ident controlled special military forces, many organized to

suppress internal opposition rather than to defend against

external threats. Examples would include Robert Mugabe’s

Fifth Brigade, which unleashed a reign of terror in opposi-

tion areas within five years after independence, or Kwame

Nkrumah’s President’s Own Guard Regiment (POGR), some-

times referred to as his “private army” (Meredith 2005, p. 19).

Consider, too, the military units that reported to the president

of Zaire. Among them numbered:

A Civil Guard, commanded by his brother-in-law, Kpama

Baramoto;

A Special Research and Surveillance Brigade, commanded

by General Blaise Bolozi, also related to the President by

marriage;

A Special Action Forces, a paramilitary unit, commanded

by Honore Ngabanda Nzambo-ku-Atumba, a close aide

of Joseph Desire Mobutu and his chief of intelligence;

45

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

and a Special Presidential Division, by all accounts the

most effective unit of them all, commanded by Gen-

eral Nzimbi Ngabale, also a “close relative” (Nzongola-

Ntalaja 2002, p. 154).

With control over wealth and the means of coercion,

authoritarian regimes were able to play a game that differed

from that played in the era of multiparty politics. As a mo-

nopoly supplier of political favors, the president could indi-

vidually tailor his political offers. Thus Kenneth Kaunda could

secure the loyalty of Mainza Chona at lower cost, given the

latter’s lack of a strong political base, than he could Simon

Kapwepe, who enjoyed a large following. Or Joseph Desire

Mobutu could recruit Barthelemy Bisegimana to serve as his

chief of staff at low cost, given the latter’s ambiguous standing

as a “citizen” of Rwandan extraction, but had to tolerate the

barbs and indulge (some of) the whims of Étienne Tshiesekedi

with his strong local backing (Nzongola-Ntalaja 2004). And

rather than having to allocate resources in a universalistic and

egalitarian manner, the chief executive could employ them to

assemble a team of just sufficient political weight for winning.

With control over the means of coercion, the president was

positioned to make take-it-or-leave-it offers; with control over

bounteous benefits and fearsome sanctions, he could prevent

efforts by others to collude. The winning coalition would there-

fore not be egalitarian and universalistic, but rather unequal

46

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Political Legacies

and minimum winning (Baron and Ferejohn 1989). And by

assembling a ruling coalition of small size, the president could

divert a larger portion of the “national pie” to his own bank

account.

The Shrinking Political Arena

In post-independence Africa, most states became authoritar-

ian (see Figures 2.2 and 2.4): Rather than having to distribute

benefits in a universalistic manner, incumbents could now

allocate them more narrowly, thereby retaining a greater por-

tion for themselves.

Once thus reconfigured, the political order appeared in-

creasingly to narrow; in the words of Kasfir (1976), in the 1970s,

it was “shrinking” in size.6 In search of resources to consume

and to expend in the pursuit of power, elites continued to en-

gage in extraction; their taxes were levied universally. But by

channeling benefits to those whom they favored, the elites

could offset the costs they inflicted upon those in whose loyalty

they sought to invest. The value of the (net) benefits would

increase as the number of clients declined, thus generating

incentives for the insiders to narrow the definition of what it

meant to be loyal.7 Incentives thus dictated a logic of exclusion.

6 The phrase is taken from Kasfir, N. (1976), The Shrinking Political Arena, Berkeley and Los Angeles: University of California Press.

7 This analysis draws upon Adam, C. S., and S. A. O’Connell (1999), “Aid, Taxation, and Development in Sub-Saharan Africa,” Economics and

47

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

The criterion for exclusion varied. Politicians often played

the nationality card: They thereby sought to exclude foreigners

from employment, as in Cote d’Ivoire (Cohen 1974), or from

ownership of land, as in Zaire (Lemarchand 2003; Nzongola-

Ntalaja 2004). In Zambia and Cote d’Ivoire, they invoked

national origins to discredit presidential candidates – Kenneth

Kaunda and Alassane Öutarra, respectively – arguing that they

had been born to immigrant parents.

Politicians also sought to restrict the benefits provided by

government to members of the ruling party. In single-party

states, those who were not members could not aspire to pub-

lic office or to a position in the public portion of the economy.

In Sierra Leone, Kpundeh records, clause 139 (3) of the cons-

titution of the ruling party provided that “no one can be

appointed or continue to be a permanent secretary ‘unless he

is a member of the recognized party’” that is, of the All People’s

Congress (APC), the governing party (Kpundeh 1995, p. 65). So,

too, in Zaire: When drafting the 1973 regulations for the ser-

vice, the civil service commissioner stated “special emphasis,

among the conditions required for recruitment, is placed on

party militancy and Zairian nationality” (Gould 1980, p. 67).

And in Senegal, Boone writes, “licenses were granted to the

Politics 11(3): 225–54; and Bueno de Mesquita, B., A. Smith, et al. (2003), The Logic of Political Survival, Cambridge, MA: The MIT Press. See also Kasara, K. (2007), “Tax Me If You Can,” American Political Science Review 101(1): 159–72.

48

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Political Legacies

bons militants of the UPS [the Senegalese Progressive Union,

the ruling political party]” (Boone 1990, as quoted in Tangri

1999, p. 75).

To achieve a deeper familiarity with the meaning of single-

party rule, I turn once again to the case of Zambia. After Pres-

ident Kaunda reinstated the Eastern Province politicians to

their posts in UNIP’s Central Committee (see Box 3.1), sev-

eral Bemba-speaking leaders defected and formed an oppo-

sition party. The government responded by filing trumped-

up charges of murder and assault and detained the dissident

leaders.

When Simon Kapwepwe, the Bemba-speaking vice presi-

dent, also defected from UNIP, the government realized that

it stood to lose political support in the Luapula, Northern, and

Copperbelt provinces – all dominated by Bemba-speakers –

and so it could be left in control of fewer than one-half of

the provinces in Zambia. The government therefore sought to

manipulate the electoral process. Seats in Parliament, it was

ruled, belonged to the party, not the person; and when a mem-

ber crossed the floor, her seat then became vacant, neces-

sitating a bi-election, which it contested vigorously and vio-

lently, supporting its candidates with the resources of the state.

The government also reverted to repression. When those who

defected from the ruling party sought reelection to Parliament,

they found their permits for meetings denied, their campaign

posters defaced, and their supporters intimidated by gangs

49

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

of youths and squads of police. In January 1972, one such

gang assaulted Simon Kapwepwe. In February, the government

banned his party and rounded up and imprisoned more

than 100 of its leaders. Many were beaten, some were tortur-

ed, and, as already noted, Zambia became a one-party state

(Gertzel, Baylies et al. 1984; Larmer 2006).

Following the end of multiparty politics, membership in the

ruling party became a form of citizenship. In 1970, the Provin-

cial Conference of UNIP resolved that “the UNIP membership

card should be made a legal document for the purpose of iden-

tification and holders of the card should be given preferential

treatment over non-holders in such spheres as employment,

promotions, markets, loans, business, housing and all socio-

economic activities” (Larmer 2006, pp. 36–7). Ordinary people

could not board public transport, cross bridges or pontoons,

or transact in public markets without producing a party card.

Those with educations and finances could not hold director-

ships or posts in state industries, qualify for bursaries or loans,

or secure the kinds of positions to which they aspired: ones

with a housing allowance, a limousine, and opportunities for

travel abroad. By tightly circumscribing the range of poten-

tial political beneficiaries, Zambia’s political elite more tightly

restricted access to economic opportunities.

In some instances, the logic that drove the politics of exclu-

sion appears to have culminated in the formation of a truly

50

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Political Legacies

miniscule elite. In Rwanda, for example, President Juvenel

Habyarimana, close family members, and senior members of

the family of his wife dominated the financial ministries, the

security services, and the ruling party. That the word akasu, or

small house, came to refer to this group underlines the diminu-

tive size of the inner circle (Prunier 1998). In Kenya, Jomo

Kenyatta, his sons, his wives, and their relatives were referred

to as the “royal family.” Burundi was governed by a small group

from Bururi; Zaire by the “Ngbandi” clique from Equateur; and

Togo by the Kabye from Kara in the north.

To comprehend the capacity of such small groups to remain

in power, it is useful to recall that the security services in

Kenya were headed by the president’s in-law; that Equateur,

Bururi, and Kara provided the military elite in Zaire, Burundi,

and Togo, respectively; and that the akasu headed a security

apparatus that in April 1994 proved capable of killing 800,000

Rwandans.

The restructuring of African political institutions thus trig-

gered a logic of exclusion, resulting in political privilege and

economic inequality. Implicit in these transformations lay as

well the strengthening of incentives for political elites to deal

in private rather than public goods.

Consider a district of 20,000 people, each expecting “his”

or “her” politician to provide one dollar in benefits. The crea-

tion of one public good, producing a dollar’s worth of benefits

51

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

for each resident, would be likely to cost less than the placing

of a dollar in the pocket of each resident. In general, as the

number of persons who claim benefits from the political elite

rises, the cost advantage to politicians of providing benefits in

the form of public rather than private goods increases as well.

As Africa’s political elite restricted the scope of those entitled

to the benefits of independence, this advantage declined. The

shrinking of the political arena thus led to a reduction in the

incentives for those who sought positions of power to reward

their followers with public goods. Private benefits drove out

public goods as the coin of the political realm.8

Conclusion

In this chapter, I have argued that searching for wealth and

power, political elites reconfigured African political institu-

tions, transforming them from multi- to single- or no-party sys-

tems or replacing civilian governments with military regimes.

They also narrowed the range of those entitled to political ben-

efits. Rather than political independence serving the collec-

tive welfare, then, it instead conferred narrowly circumscribed

privileges upon those who won out in the competition for polit-

ical office.

8 This analysis builds upon Adam, C. S., and S. A. O’Connell (1999), “Aid, Taxation, and Development in Sub-Saharan Africa,” Economics and Poli- tics 11(3): 225–54; and Bueno de Mesquita, B., A. Smith, et al. (2003), The Logic of Political Survival, Cambridge MA: The MIT Press.

52

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Political Legacies

During the struggle for independence, Africa’s citizens

had embraced politics. In response to the political realities

about them, however, in the post-independence era, they

increasingly came to view their leaders as a source of insecurity

and the state as a source of threat rather than of well-being.

53

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

https://doi.org/10.1017/CBO9781316423974.003 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

4

Policy Choices

Focusing on the determinants of economic growth in thepost-independence period, researchers from the Africa Economic Research Consortium (AERC) isolated a set of “anti-

growth” syndromes: styles of policymaking that reduce the rate

at which national economies could grow (Ndulu, Collier et al.

2007). Most common is the combination of policies that they

designate as “control regimes,” which led to:

1. A closed economy.

2. The distortion of key prices in the macroeconomy.

3. The promotion and regulation of industries.

4. The regulation of markets.

In this chapter, I shall describe these policies and discuss

their origins and their consequences. Control regimes are eco-

nomically costly, and I shall explain why incumbents nonethe-

less retained them, even after their costs were known. The

reason, I argue, is that the policies generated political ben-

efits for Africa’s authoritarian regimes. They provided elites

55

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

with sources of income and furnished means for transform-

ing even declining economies into political organizations,

enabling politicians to recruit political dependents, willing to

fight – if necessary – to keep them in power. While yielding

political advantages, however, these policies contributed to

the subsequent collapse of Africa’s states.

The Content of Control Regimes

As reported by the AERC researchers, governments that adopt

control regimes regulate trade, manipulate the interest and

exchange rates, and develop close ties with urban-based

industries.

The Control of Trade

In the post-independence period, governments imposed tar-

iffs and quantitative controls on a wide range of industrial pro-

ducts. To sell their goods in Africa’s markets, foreign firms then

had to “jump over” these barriers and to invest in the plant and

equipment that would enable them to produce and thus mar-

ket their goods locally. These policies most frequently targeted

the goods most commonly consumed by the residents of poor

societies: processed foods, beverages, textiles, shoes, blankets,

kerosene, and other consumer products. In at least one case,

56

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Policy Choices

Zambia, the government severely restricted the importation of

automobiles; and for a brief and inglorious moment, automo-

biles were produced in Livingstone, a small urban center on

the southern border of the country (Elliott 1971).

Macroeconomic Policies

Given the low level of industrialization, investors wishing to

establish new firms had to import plant and equipment from

abroad. To lower the costs of such investments, governments

restructured financial markets. Creating banks that targeted

“commerce,” “industry,” or more broadly “development,” they

made available loans at low rates of interest to those seeking

to invest in projects to which they accorded a high priority.

Outside of the Franc zone, they issued their own currencies.

Many then employed their control over the banking system to

set the rate at which this currency could be exchanged for cur-

rencies from abroad. By overvaluing their currency, they set

the exchange rate to the advantage of importers: Because they

could purchase foreign “dollars” more cheaply, those seeking

to invest in local industry could then import plant and equip-

ment at lower cost. Trade barriers having already been set in

place, their goods remained protected against foreign compe-

tition, whose products would have gained a price advantage

as a result of the revaluation of the local currency.

57

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

Industrial Regulation

Governments that implemented control regimes also imple-

mented regulatory policies that enhanced the profitability

of firms. By licensing, they discouraged entry and protected

established producers. When governments themselves owned

firms, the governments were certain to prevent new firms

from competing with established producers; public enter-

prises then remained as the monopoly suppliers of their prod-

ucts. Moreover, because governments subsidized the costs of

capital, many firms adopted capital-intensive technologies;

they then tended to operate most profitably when producing

near full capacity. Because the protected markets of Africa were

small, the result was the creation of highly concentrated indus-

tries, with but one or two large firms in each, with firms operat-

ing at low capacity and therefore at high cost. But because the

noncompetitive structure of the domestic market conferred

on firms the power to set prices, they could remain privately

profitable, even while highly inefficient.

The Incidence of Costs and Benefits

When governments artificially increased the value of their cur-

rencies, the benefits that they conferred upon the importers

of capital equipment were matched by the costs they inflicted

58

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Policy Choices

on exporters. When foreign dollars converted into fewer cedi

(the unit of currency of Ghana), or other African currencies,

exporters experienced a reduction in their incomes. In Africa’s

agrarian economies, most exporters were farmers, who pro-

duced coffee, cocoa, sugar, cotton, sisal, and other crops for

foreign markets. When governments artificially increased the

value of their currencies, they may have protected the prof-

its of industrial firms by imposing tariffs and quantitative

restrictions on imports, but they rarely offered similar pro-

tection to farmers. Producers of rice in West Africa therefore

found themselves competing in local markets with imports

from Louisiana, and producers of cassava in Central Africa

faced competition from bakers advantaged by the low costs of

imported wheat. Trade policies were thus biased against the

exporters of cash crops and the producers of food crops as

well.

When governments regulated urban industries, they pro-

tected the profits of urban firms: By limiting competition,

they granted them the power to set prices to their advantage.

When governments regulated agriculture, they conferred mar-

ket power on consumers: They created monopsonies for the

purchase of both export and food crops. Governments pur-

chased the cash crops at a low domestic price, sold them at

the prices prevailing in international markets, and deposited

the difference in the public treasury. They purchased the food

59

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

crops at prices set to ensure that soldiers, bureaucrats, and

urban workers would be assured of low cost food.

In the 1960s, the majority of Africa’s population lived in

the rural areas and agriculture constituted the largest single

industry. The policies thus favored the interests of a minority

over those of the vast majority of the population in most states.

As noted by Dumont (1966), “In May 1961 a number of farmers

north of Brazzaville said to me: ‘Independence isn’t for us; it’s

only for the city people’” (p. 17). It was precisely this property

of post-independence policies that Dumont condemned.

Control regimes thus benefited the urban and industrial

sector; indeed, given the aspiration for industrial development

that motivated many policymakers, this was their intent. But

they did so at the expense of the great majority of Africa’s pop-

ulation – those who lived in the rural areas – and the greatest

of Africa’s industries – agriculture.

For these policies to persist, opposition to them had to be

demobilized. The most likely opponents would be farmers;

and because they constituted a political majority, the farmers

were dangerous. The political commitment to control regimes

could persist, then, only insofar as political challengers

lacked an incentive to pursue electoral majorities. Authoritar-

ian institutions thus underpinned the imposition of control

regimes.

The relationship between political institutions and public

policies is captured by the data in Figures 4.1 and 4.2.

60

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Policy Choices

0

10

20

30

40

50

60

70

Military Civilian All

P er

ce n

t O

b se

rv at

io n

s

Percent Control Regimes

Figure 4.1. Control regimes and military government.

0

10

20

30

40

50

60

70

No-Party One-Party Multiparty All

P er

ce n

t O

b se

rv at

io n

s

Percent Control Regimes

Figure 4.2. Control regimes and party system.

As indicated in Figure 4.1, military governments were far

more likely than civilian ones to adopt control regimes. From

the period 1970–1995, in more than 50% of the country years,

if the data registered the presence of military regimes, they

registered the presence of control regimes as well. As shown

in Figure 4.2, one- and no-party regimes were also more likely

to adopt control regimes, with more than 50% of the observa-

tions that centered on no- or single- party systems exhibiting

control regimes as well, as compared with but 30% of those

with multiparty systems. In this study, I label as authoritarian

governments that are headed by soldiers rather than civilians

61

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

and those civilian regimes that have banned the formation of

opposition parties. The data thus suggest an elective affinity

between authoritarian politics and interventionist policies in

the post-independence period.1

Economic Costs and Political Benefits

Given that agriculture is the largest single industry in most

African countries, it is not surprising that the economies of

countries that imposed control regimes appear to have grown

more slowly than others. The AERC research team measured

the economic impact of four major “anti-growth” syndromes

(Ndulu, Collier et al. 2007). The first was state failure. Next

came “inter-temporal redistribution,” which most commonly

occurred when governments would consume rather than save

the proceeds of resource booms. A third was ethnic or regional

redistribution, when governments became the political agents

of subnational minorities. The fourth was the adoption of con-

trol regimes.

Controlling for a variety of factors that might affect growth –

the growth rate of trading partners, for example – and

1 See, too, the Appendix. Turn as well to Chapters 4 and 11 of the first vol- ume of Ndulu, B., P. Collier, et al. (2007), The Political Economy of Economic Growth in Africa, 1960–2000, 2 vols, Cambridge, U.K.: Cambridge Univer- sity Press.

62

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Policy Choices

correcting for the impact of the growth rate on the choice

of policies, the researchers confirmed what most would have

expected: that state failure was the most damaging to growth.

When states failed in late-century Africa, growth rates fell

between 1.8% and 1.9% per annum. More surprising, perhaps,

was that they found that the imposition of a control regime

led to a loss of roughly 1.6 percentage points per annum in the

growth rate, thus rivaling the impact of state failure. State fail-

ure was relatively rare, occurring in just more than 10% of the

country years, 1970–1995. The imposition of control regimes,

however, was most decidedly not: They appear in more than

60% of the country year observations in the late 1970s and

early 1980s. The adoption of control regimes thus imposed

high costs on Africa’s economies.

If the policies harmed the economic interests of most

Africans and lowered the growth rate of national economies,

then why were they chosen? And, once chosen, why did they

remain in place? The answer, I argue, is that the policies served

political rather than economic interests.The interventionist

style of policymaking enabled governments to target benefits

to important constituencies, thus – in the short term, at least –

promoting political order. And by transforming industries and

markets into political organizations, it enabled governments

to spin webs of political obligation and thus forge the political

machines that kept them in power.

63

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

Political Benefits

In West Africa, the richer regions lie in the southern portions,

which are heavily forested. The Sahelian regions – dry and with

uncertain rainfall – have little to offer the international econ-

omy2; lying inland from the coast, what little they have to offer

has necessarily to be shipped at high cost, leaving few prof-

its for producers. Throughout West Africa, then, there exists a

disparity between the economies of the coast and interior.

Illustrative is the case of Nigeria. At the time of indepen-

dence, the economy’s most important exports – cocoa, palm

oil, and other agricultural products – flowed from the south.

Not only did the south have prosperous farmers, but it also was

home to the merchants, bankers, and lawyers who provided the

services for the export industries. While the north was not rich,

it was powerful. It contained more than one-half of Nigeria’s

population. It was relatively homogeneous: The great major-

ity of its people followed Islam and considered themselves to

be Hausa-Fulani. And its emirates provided a means for orga-

nizing its people. While relatively poor, then, the region could

marshal formidable political forces.

By dint of the north’s large size and degree of organization,

following independence, its politicians assumed control of the

executive branch of the federal government. Pursuing a policy

2 The exception is cotton, which long has faced high tariff barriers in global markets.

64

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Policy Choices

of import substituting industrialization (Helleiner 1966; Little,

Scitovsky et al. 1970; Schatz 1977), the government promoted

the formation of domestic industries, many of which it con-

vinced to locate in the north. It also placed a disproportion of

its development projects in the region. The costs of these ini-

tiatives fell largely upon the south, whose consumers paid the

higher prices that resulted from tariff protection and whose

farmers paid the taxes that financed public investments.

The south’s reaction is captured in a statement issued in

1964 by the government of the eastern region:

Take a look at what they [the North] have done. . . . Kainji Dam

Project – about £150 of our money when completed – all in the

North. . . . Bornu Railway Extension – about £75 million of our

money when completed – all in the North. . . . Military training

and ammunition factories and installations are based in the

North, thereby using your money to train Northerners to fight

Southerners. . . . Building of a road to link the dam site and the

Sokoto cement works – £7 million, when completed – all in the

North. . . . Total on all these projects about £262 million (italics

in original Gboyega 1997, p. 161).

The regional tensions that marked the politics of Nigeria

found their parallel elsewhere in Africa. In Togo, General

Gnassingbé Eyadéma held power for thirty-eight years; him-

self from the north, Eyadéma used the powers of the state to

extract the wealth of the south for the benefit of his family, the

military, and his region. In Ghana and Cote d’Ivoire, it was the

65

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

south that tended to control the national government; but by

choosing northerners for the vice presidency (as in the case of

Ghana) or by intervening in the economy and channeling pub-

lic funds into the interior (as in Cote d’Ivoire), political leaders

sought to ease regional tensions.3

Turning to East Africa, Uganda, too, was marked by north-

ern poverty and southern prosperity. In Tanzania, the pro-

ducers of export crops cluster at the higher elevations, where

they enjoy bounteous and reliable rainfall and moderate

temperatures; the dry climate and arid lands at lower eleva-

tions sustain subsistence production. Both Milton Obote in

Uganda and Julius Nyerere in Tanzania built their political

base in the poorer regions, propounded socialist principles,

and imposed control regimes in an effort to redistribute the

wealth of the prosperous regions to the semi-arid zones.

While these examples are suggestive, the evidence from

Zambia is more compelling. It enables one to observe “in real

time,” as it were, the process by which regional tensions shaped

policy choices. Within a half decade of independence, the

United National Independence Party – UNIP, the governing

party – was wracked by conflict between regional blocs of

3 For an analysis of similar tensions in Cameroon, see Bayart, J.-F. (1989), Cameroon, in Contemporary West African States, edited by John Dunn, Donal B. Cruise O’Brien, and Richard Rathbone, London: Oxford Uni- versity Press; and Levine, V. T. (1986), Leadership and Regime Changes in Perspective, in The Political Economy of Cameroon, edited by M. G. Schatzberg and I. W. Zartman, New York: Praeger.

66

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Policy Choices

politicians, each of whom sought to seize high offices in the

ruling party and thus in the government as well. Politicians

from the Bemba-speaking districts made up one faction. They

came from the Northern and Luapula provinces, and also

from Copperbelt Province, the mineral rich region to which

Bemba speakers had long migrated in search of employment

(see Map 3.1). Another faction consisted of politicians from

the Nyanja-speaking provinces: Eastern Province and Lusaka

(which includes the city of Lusaka, the national capital). To

the side stood the leaders from Barotse, or Western Province,

in the southwest; the Central and Southern provinces, largely

controlled by the opposition party; and the North-Western

Province, which was sparsely inhabited.

At the time of independence, 1964, politicians from the

Eastern Province held the vice presidency and the largest sin-

gle bloc of seats in the Central Committee (see Box 3.1). Three

years later, the Bemba-speaking politicians coalesced with

those from Central and Southern provinces to seize the vice

presidency and capture the Central Committee. The result was

a political crisis within the ruling party, as the losers sought to

lay claim to the offices they once had held, and to the govern-

ment posts, with their attendant perquisites, that went with

them. In response to this crisis, the president, Kenneth Kaunda,

introduced the first of what became known as “economic

reforms.” In this instance, the reforms involved the takeover of

foreign-owned companies. The distribution and management

67

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

of their assets provided a resource to compensate those who

lost out in the competition for power.

In the general elections that followed those within the rul-

ing party, Barotse, whose leaders had allied with the Nyanja-

speaking bloc, defected to ANC (the Africa National Congress),

thus joining the Central and Southern provinces in the ranks

of the opposition. To consolidate support in the six provinces

that remained loyal, President Kaunda struck once again, this

time nationalizing the copper industry. In 1969, the newly

elected vice president of UNIP resigned, sparking rumors of

the withdrawal of the Bemba-speaking bloc from the govern-

ing party. The president responded with yet another economic

“reform,” adding the banking and insurance industries to the

government’s portfolio (see Elliott 1971; Szeftel 1978; Burdette

1988).

In the years following independence, regional conflict thus

punctuated the politics of Zambia. In response to the open-

ing of each political fissure, the president extended the scope

of the government’s control of the economy. By nationalizing

firms, gaining control over key sectors, and building a regu-

latory apparatus about publicly owned firms,4 the president

multiplied the political resources at his command. He posted

4 The firms were known as INDECO (Industrial Development Corporation), FINDECO (Finance and Development Corporation), and MINDECO (Min- ing Development Corporation), all under ZIMCO (Zambia Industrial and Mining Corporation).

68

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Policy Choices

politicians to the boards of each firm in which the government

held an interest. And he staffed the bureaucracy that superin-

tended each group of firms with directors chosen from within

the ruling party.

Interventionist policies, the example of Zambia suggests,

confer upon governments the resources with which to ame-

liorate political tensions, many arising from conflicts between

regions. Control regimes may be expensive, then; they may

reduce the rate of economic growth of the national econ-

omy. But their costs appear to represent the costs of forging

viable political bargains. In Africa, political order is expensive

to maintain.

The Maintenance of Authoritarian Regimes

If conflicts between regional political delegations help to

account for the adoption of control regimes, the question still

remains: Why, once chosen, did they remain in place?

A major reason for the retention of these policies, I would

argue, is that they were economically rewarding for those in

power – and politically useful as well. They provided the liga-

ments that bound together Africa’s authoritarian regimes.

As we have seen, when imposing control regimes, govern-

ments often pegged their currency at a value higher than that

which would have been generated by a competitive market.

When doing so, they created an excess demand for foreign

69

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

“dollars.” Those in possession of local currency could then

purchase foreign currencies at a price that lay below that pre-

vailing in currency markets. Because the demand for foreign

exchange exceeded the supply at official prices, the foreign

currencies then had to be rationed: Whoever controlled their

allocation was now in a position to reward his political follow-

ers, his family, and his friends. Control over the central bank

creates the opportunity to increase one’s wealth and to build

a political network.

By appreciating the value of the domestic currency, gov-

ernments that intervened in currency markets increased the

demand for imports. When the currency was set at an artifi-

cially high level, those who secured it at the official price could

purchase foreign goods more cheaply. By importing those

goods and selling them in the domestic market, they could

then pocket in local currency the benefit created by the gov-

ernment’s manipulation of the exchange rate. But because the

currency was set at an artificially high level, those who export

earned less in foreign markets: Each dollar earned abroad gen-

erated less local income. With the demand for imports increas-

ing and the incentives to export decreasing, governments that

set the value of their currency too high soon began to incur

trade deficits. To stem these deficits, they began to regulate

imports. They blocked the importation of “luxuries” to facili-

tate the continued importation of “essential” goods, banning

the import of liquor, for example, to enable the purchase of

70

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Policy Choices

medicines or banning the importation of motorcars to safe-

guard the purchase of tractors.

Governments that sought to control the value of their

money in international markets thus soon found themselves

regulating the flow of trade as well. One result was the growth

of a bureaucracy to ration access to foreign exchange and the

importation of foreign goods. Another was the growth of politi-

cal machines, as those who controlled and enforced trade regu-

lations conferred the right of access to foreign markets, thereby

creating clients: people who owed them their economic for-

tunes and whose political loyalty they could therefore expect

in return.

The intersection of the borders between Rwanda, Uganda,

and eastern Zaire (present-day Congo) provides an apt illustra-

tion. In 1973, General Juvenal Habyarimana deposed Gregoire

Kayibanda as president of Rwanda. Already commander of

the armed forces, the new president sought control over the

economic bureaucracy as well, including – and perhaps espe-

cially – the central bank.

To the west of Rwanda lie some of the most productive

lands of Zaire: temperate, well watered, and endowed with

rich, volcanic soils. There grows some of the best coffee pro-

duced in East Africa. Near the border also lie deposits of gold

and some of the last major herds of elephants in Africa. Able to

purchase hard currencies at advantageous rates, those upon

whom Habyarimana conferred access to foreign exchange at

71

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

the official rate were at an advantage in the scramble for the

riches of eastern Zaire. The president targeted his largesse on

his family, his wife’s family, and his subordinates in his mili-

tary, many of whom, like Habyarimana and his wife, came from

the northern districts of the country. Funded by members of

Rwanda’s inner circle, traders and businessmen crossed the

border to Zaire and purchased coffee, ivory, and gold. Trans-

porting these goods to the coast, they returned laden with

luxuries: liquor, automobiles, appliances, and expensive cloth-

ing, to be sold in shops, boutiques, and showrooms owned

by friends of the president.5 Each member of the president’s

circle then assembled his own political retinue from among

those to whom they had extended favors: the granting of for-

eign exchange, the “right” to market shoes or liquor purchased

abroad, or to sell their coffee to private buyers instead of to the

state monopoly. The politicians thus cast webs of political obli-

gation about the informal markets to which the government’s

interventionist policies gave rise.

Recall that the ruling elite controlled not only the economic

agencies but also the security services. The political ties that

ramified about the regulated economy were forged not only

from selective benefits but also from targeted sanctions. Many

5 Meredith, M. (2005), The State of Africa: A History of Fifty Years of Independence, London: Free Press. Interviews by the author, Rwanda, 2000.

72

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Policy Choices

of the most valuable sources of wealth were illegal: Riches were

gleaned in the “shadow economy.” Those who secured their

income illegally were liable to seizure, prosecution, and deten-

tion – or worse, and their vulnerability grew in proportion to

their bank accounts. Given that they had violated the law, their

prospects – financial and political – lay at the discretion of those

who controlled the coercive apparatus of the state.

Conclusion

In this chapter, I have argued that there is an elective affin-

ity between political institutions and policy choices in post-

independence Africa. The banning of opposition parties and

the end of multiparty politics enabled political elites to adopt

and retain economic policies that harmed farmers, even

though in many states the rural producers formed a major-

ity of the population. Because politicians competed for the

favor of the state house rather than for the backing of citizens,

the numerical supremacy of Africa’s rural population posed no

threat to those in power and so failed to alter their choice of

policies.

Because of the incidence of the costs, control regimes effec-

tively constituted a tax on agriculture. If only because agricul-

ture represented the single largest industry in most of Africa’s

economies, the policies thereby lowered the continent’s rate of

73

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

economic growth. That these policies nonetheless remained

in place reflects the political advantages that they conferred:

resources that authoritarian elites could employ to ameliorate

political tensions, to recruit political clients, and to build polit-

ical machines, and thereby remain in power.

74

https://doi.org/10.1017/CBO9781316423974.004 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

5

Subnational Tensions

Beneath the political surface of Africa’s authoritarianregimes, there were forces at work that sowed the polit- ical landscape with multitudinous opportunities for conflict.

The economies of Africa’s rural communities rendered them

politically expansionary, and therefore generated competing

claims for land. So long as political order reigned at the national

level, and so long as the incumbent regimes could marshal

the resources with which to purchase or to compel political

restraint, the resultant conflicts could be contained. When

states began to fail, however, local conflicts then acquired

national significance. They offered opportunities to politicians

The argument in this chapter should be viewed as a contribution to the study of Africa’s “political geography,” as pioneered by Herbst, J. (2000), States and Power in Africa, Princeton, NJ: Princeton University Press; Boone, C. (2003), Political Topographies of the African State: Rural Authority and Institutional Choice, Cambridge, U.K.: Cambridge University Press; and Azam, J.-P. (2007), The Political Geography of Redistribution, Chapter 6 in The Political Economy of Economic Growth in Africa, 1960–2000, edited by B. Ndulu, P. Collier, R. H. Bates, and S. O’Connell, Cambridge, U.K.: Cambridge University Press.

75

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

seeking to consolidate political followings, and as national

elites were drawn to parochial disputes, Africa’s rural citizens,

in search of political champions, flocked about them. When

political order declined in late-century Africa, it therefore did

so precipitously. Competition between local communities thus

increased the costs of governing by authoritarian regimes and

the pace with which they subsequently collapsed.

Rural Dynamics

To apprehend the forces at play, consider a family and its

choice of where to settle.1 The family will naturally choose to

farm the highest-quality land, where its efforts will result in

the greatest return. Alternatively, by working such lands, it

can secure sufficient food to feed itself at least effort. Now let

another family arrive and the population increase. This fam-

ily must choose between being the second family to settle on

the highest-quality land or the first to settle on the land of the

next-best quality. Where it settles depends upon the relative

magnitude of the output that it can secure in the two loca-

tions. For purposes of argument, assume that the differential

in land quality is such that this second family, as did the first,

1 This analysis follows Ricardo, D. (1821), On the Principles of Political Econ- omy and Taxation, London: John Murray.

76

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Subnational Tensions

secures a higher return to its labor in the highest-quality land.

It will then choose to locate adjacent the first family. As the

cycle repeats itself over time, a settlement will therefore grow

in the lands of higher quality.

A second pattern will also emerge, however: a trickle of set-

tlers to the periphery. Because of diminishing returns, as more

families crowd onto the lands of high quality, the increment

in production that results from each additional unit of labor

declines. New arrivals will therefore eventually find it more

attractive to be the first to settle on the lands of lesser quality

rather than to be the last to settle on lands of superior quality.

There therefore begins a process of dispersal in the settlement

pattern.

Arable and Pastoral Production

In Africa, as elsewhere, agriculture involves more than the

planting and harvesting of crops. It also involves the breed-

ing and herding of livestock, and this activity too induces the

shifting of population to the periphery. When livestock graze,

they make extensive use of land. As the core becomes more

densely settled, land becomes scarce; it therefore increases in

value. To conserve on the use of this resource, farmers there-

fore tend to shift their livestock to less densely settled areas. In

addition, when grazing, cattle, goats, and sheep may wander

77

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

into the fields, and pastoralism can therefore lower the return

from arable production. Farmers therefore seek to separate

the two activities, shifting their livestock from the core to the

periphery. In the absence of “mixed farming,”2 the two activi-

ties – arable and pastoral production – will be most productive

if managed apart.

Investment

Families combine persons of different genders, ages, and gen-

erations, and this property furnishes an additional reason for

territorial expansion: the opportunity to invest and thereby

escape a future dominated by diminishing returns.

As time passes and population increases, without technical

change, each additional unit of labor adds less to the total prod-

uct. If labor is paid its marginal product, then wages fall. Even

were the total product to be divided equally, insofar as output

increases more slowly than does population, per capita con-

sumption will fall. In either case, the society becomes poorer

with the passage of time.

In the face of diminishing returns, out-migration offers an

escape from poverty. Because migration is costly, it is likely to

2 This is the change that marked the commercial revolution in European agriculture. See, for example, Timmer, C. P. (1969), “The Turnip, the New Husbandry, and the English Agricultural Revolution,” Quarterly Journal of Economics 83: 375–96.

78

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Subnational Tensions

be the younger rather than the older generation that migrates,

for the younger generation can amortize the costs of migra-

tion over a longer stream of earnings. To treat migration as a

choice made solely by the younger generation, however, is to

fail to recognize other incentives at play. The elders, too, are

subject to the consequences of diminishing returns; as popu-

lation grows, they, too, experience a fall in the wage rate and

in average income. As elders, they may be less likely to emi-

grate. But they, too, would benefit from the out-migration of

the young, as their departure would reduce the quantity of

labor and therefore raise the earnings of workers in the core.

Diminishing returns thus creates an incentive for the elders

to invest in the out-migration of the young. The search for an

escape from diminishing returns strengthens the incentives to

invest in expansion in Africa’s rural economies.

Variations in Form

Thus far I have emphasized the economics of territorial expan-

sion. It is important to address the politics as well. In doing

so, I draw the conventional distinction between decentralized

and centralized societies in Africa (Fortes and Evans-Pritchard

1987). In decentralized societies, politics is dominated by fam-

ily heads; there is no chief and no bureaucracy. In centralized

societies, there exists a chief executive, a retinue of palace offi-

cials, and bureaucrats who levy taxes and make war. I note as

79

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

well a third kind of political system, one in which there is no

bureaucracy but the society is spanned by formal institutions

called age grades. Despite the variation in the way the societies

are structured, each type can be regarded as offering an alter-

native political solution to a common problem: the need for

intergenerational contracts that will promote the peopling of

the periphery.

The Family

Migration requires infusions of capital. If a farmer, the junior

member will have to be supported until he claims, clears, and

cultivates a piece of land. If a pastoralist, he will need to be given

stock with which to build a herd. In either case, repayment is

deferred. In some instances, the returns to such investments

take the form of increased land holdings and a lowering of risk,

as the family estate comes to ramify across different ecological

zones; in others, it yields a flow of milk, curds, and hides from

flocks consigned to the young for safekeeping, or of cattle with

which to pay bride price and increase the size and prestige

of the lineage. In either case, the elder investors incur costs

today; the young recipients later repay; and there arises a flow

of resources back to senior members of the family.

The transformation of the family into a means of invest-

ment confronts a major dilemma, however. The transactions

80

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Subnational Tensions

are separated temporally: The costs fall upon the elders in

present time while repayment must of necessity be delayed.

In addition, the parties to the transaction are separated by

space. The elders cannot monitor the efforts of the young; they

cannot assess the validity of excuses for non-repayment, such

as the loss of livestock to disease or of crops to grazing wildlife.

The potential for opportunism is therefore high, weakening

the incentives to invest.

In African societies, the politics of gerontocracy provide one

solution to this dilemma. The solution takes the form of the

conferral upon the elders of resources and sanctions sufficient

to enable them to counter the attractions of defection by the

young.

In many African societies, only those who are married and

have fathered children of their own can hold seats in polit-

ical councils, take part in policy debates, and lay claim to

prestigious honors. And often it is the elders who control the

resources required for the payment of bride wealth. Because of

polygamy, they also control a large portion of the stock of mar-

riageable women, and thus the opportunities for the young

men to find suitable brides. The elders’ control over the possi-

bility of marriage therefore yields them power over the political

prospects of the young (see Meillassoux 1981).

If an elder rules that certain rituals have not been prop-

erly observed or that certain ceremonies have been improperly

81

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

performed, then a marriage – or a birth – might not be legit-

imate. Those whose standing in the family is thus rendered

uncertain may therefore lose access to the property or to the

political offices controlled by their lineage. That the elders

interpret family law therefore places them in a position to gov-

ern the allocation of both wealth and power.

Gerontocratic political institutions thus shape the incen-

tives that govern the conduct of the young. Whereas they may

prefer to avoid their obligations, given the power of the elders,

the young are unlikely to choose to do so. Within a political

gerontocracy, the elders control sufficient sanctions to make

it in the interests of the young to keep their pledges. Knowing

that the young will not defect, the elders are therefore will-

ing to invest; they are willing to sponsor the movement to the

frontier. Political structures thus shape economic incentives

in ways that strengthen the forces of territorial expansion in

rural Africa.3

3 See Fortes, M. (1958), Introduction, inThe Developmental Cycle in Domes- tic Groups, edited by J. Goody, Cambridge, U.K.: Cambridge Univer- sity Press; Kenyatta, J. (1953), Facing Mount Kenya, London: Secker and Warburg; Sahlins, M. D. (1961), “The Segmentary Lineage: An Organization of Predatory Expansion,” American Anthropologist 63: 322–45; Sahlins, M. D. (1968), Tribesmen, Englewood Cliffs, NJ: Prentice-Hall; Sahlins, M. D. (1971), Tribal Economies, in Economic Development and Social Change, edited by G. Dalton, Garden City, NY: Natural History Press for the Amer- ican Museum of Natural History: 43–61; and Bates, R. H. (1989), Beyond the Miracle of the Market, Cambridge, U.K: Cambridge University Press.

82

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Subnational Tensions

Age Grades

In some societies, relationships between generations are

explicitly marked by the presence of age grades. In such soci-

eties, youths pass through a series of stages before being

allowed to marry and assume senior positions in the tribe.

Toward the end of their “probationary period,” they serve as

warriors. One of their tasks is to provide defense; a second is

to conquer, seizing cattle and appropriating land. Each age set

adds an increment to the total population of the tribe; each is

expected to add as well to its productive holdings. As stated by

Waller and Sobania, writing of the Masai:

[T]raditions of nineteenth-century expansion and warfare are

structured to link successive stages in their occupation of

Maasailand and control of its resources to the progression of

age-sets. The advances made by one set are consolidated and

exploited by their successors, land resources of stock, graz-

ing and water captured are utilized by elders. In this [process

of] . . . individual maturation, the continuous flow of age-sets,

and community growth and expansion are woven together. . . .

(Waller and Sobania 1994, p. 58).

States

Those who study the origins of political centralization in Africa

often stress the role of conflict: The lineages that can conquer

83

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

or subvert are those that furnish kings (see, for example,

Wrigley 1996). They also stress the willingness of followers to

obey, that is, to cede power and wealth to the ruling lineage.

In search of the factors that shape the level of deference, it is

useful to return once again to the role of land, and in particular

to the significance of differences in its productivity.

As stressed by Carneiro (1970) and Reyna (1990), when land

is of uniform quality, those who feel oppressed can resist sim-

ply by exiting. As stressed by Turner (1957), the existence of

this option limits the power of headmen and promotes polit-

ical schism rather than political centralization. When there

is a differential in the productivity of the best and next-best

lands, however, then political centralization becomes possible.

When the high-quality lands are circumscribed by unproduc-

tive ones, people will be reluctant to exit, even though coerced

or taxed. Thus it is that states formed in the highlands of the

Sahara, where the rains fall midst the desert, but rarely in the

savannahs, where the uniform productivity of the land made

exit a viable strategy (Vansina 1966). Thus, too, the location of

states in the richly endowed river valleys, where alluvial soils

and abundant moisture promises returns far greater than those

in adjacent territories.

Not only do communities in such favored settings tend to be

more highly centralized; they also tend to be more densely pop-

ulated (Stevenson 1968). The price of land is therefore high rel-

ative to that of labor. As a result, the monarch can accumulate

84

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Subnational Tensions

power. He can do so by exchanging the protection of land rights

for political services, such as the payment of taxes or the levy-

ing of conscripts.

Like their decentralized counterparts, centralized king-

doms tended to expand. Rather than dispatching youths to

settle lands on the periphery, in states, monarchs recruited

them into the military and sent them to conquer new terri-

tories. The occupation of newly seized territories decreased

the pressure of population on the lands of the core; it there-

fore increased the wage rate in the center. And it brought an

influx of wealth from assets captured by the military: taxes

from traders and miners, as in Ashanti (Wilks 1975); on ports,

as in Dahomey (Polanyi 1991) and Uganda (Wrigley 1996); and

on ivory, as in Central Africa (Vansina 1966). By the forceful

seizure of resources abroad the military added to the stock of

wealth at home. Centralized societies thereby secured higher

incomes for their members through expansion and conquest.

Impact on Contemporary Politics

Because of the “imperial peace,” traditional states now rarely

mobilize for conquest or warfare in Africa; nor, in most cases,

do age-grade societies continue to keep their youths under

arms. Nonetheless, past conquests by monarchs and warriors

created territorial disputes that reverberate to this day and so

shape contemporary politics. And even in the present, families

85

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

organize the out-migration of junior kin, generating conflicts

between “strangers” and “sons of the soil.” Beneath the sur-

face of national politics there thus lie the political tensions

produced by the dynamics of agrarian societies.4

Kenya

The Kikuyu of Kenya exemplify the decentralized mode of

expansion. In the nineteenth century, the Kikuyu resided on

the slopes of Mt. Kenya, where the soils were rich, the tem-

peratures moderate, and where rains fell both in spring and

the autumn, enabling the production of two crops a year. As

described in detail by Ensminger and Leakey (Leakey 1977), as

their numbers rose, families opened up new territory, moving

to lower-lying lands at the base of Mt. Kenya. With yet further

increases in population, the Kikuyu spread outward. Young

people, entrusted with the family herds, were among the first

to be dispatched to the frontier; they were soon followed by

young couples who planted gardens. Crossing the mountains

of Aberdare range, they settled along the upper margins of the

escarpment bordering the Rift Valley (Mbithi and Barnes 1975).

Doing so, they penetrated into contested terrain: lands grazed

by the pastoralists – the Masai to the north and south and the

Kalenjin-speakers to the west.

4 The argument just offered can be viewed as providing “micro- foundations” for Fearon’s findings regarding the origins of ethnic warps. See Fearon (2004).

86

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Subnational Tensions

In “normal” years, the pastoralists tend to graze on the

valley floor; during the dry season, they drive their herds to

higher elevations. In periods of drought, they enter forested

areas along the rim of the valley, which offers browse that could

replace the grasses. For their part, the Kikuyu had found the

lands unsettled; when the pastoralists and their herds took

refuge in the wooded fringe, the Kikuyu then felt that they had

been invaded. By contrast, from the pastoralists’ point of view,

the Kikuyu had reduced their options for dealing with the risks

of nature.

In the 1990s, the conflicts in the Rift Valley moved from

the local to the national political agenda. Following violent

demonstrations at home and mounting pressures from abroad

(Hempstone 1997), President Daniel arap Moi agreed in 1991

to an end to single-party rule. The pastoralists composed the

political base of the Kenya African National Union (KANU), the

governing party; the Kikuyu steadfastly backed the political

opposition. Campaigning for votes in the Rift Valley, oppo-

sition politicians backed the cause of the Kikuyu settlers;

the incumbents backed the communities whose lands they

had “invaded.” “Majimboism” – meaning federalism – became

a code word for this dispute. Were federalism to be adopted,

the more numerous pastoralists would have gained control

over the provincial government, leading to the extinguish-

ing of Kikuyu land rights in the Rift Valley – and to ethnic

cleansing.

87

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

In the midst of growing insecurity, ambitious elites hurried

to build competing political organizations. In the five months

that passed from the end of single-party rule to the time of

multiparty elections, the Rift Valley became a breeding house

for the formation of armed militias, as politicians sought to

build reputations for being able to defend rights to land.5

Ethiopia

In northwestern Kenya and eastern Uganda, young Karama-

jong, Pokot, and Samburu, carrying AK-47s rather than spears,

pillage the cattle of their neighbors and compel agricultural-

ists to allow their cattle to graze on their fields (Fratkin, Roth

et al. 1994; Jalata 2005). Further north lie the Oromo who, by

tradition, initiate a new age set of warriors every eight years.

5 Daily Nation, Constituency Review: Laikipia District, July 16, 2002, pp. 11–14; Rutten, M. (2001), “Fresh Killings”: The Njoro and Laikipia Violence in the 1997 Kenyan Elections Aftermath, in Out for the Count: The 1997 General Elections and Prospects for Violence in Kenya, edited by M. Rutten, A. Mazrui, and F. Gignon, Kampala, Uganda: Fountain Pub- lishers; Mwakikagile, G. (2001), Ethnic Politics in Kenya and Nigeria, Huntington, NY: Nova Science Publishers; Kimenyi, M. S., and N. Ndung’u (2005), Sporadic Ethnic Violence: Why Has Kenya Not Experienced a Full Blown Civil War? in Understanding Civil War: Evidence and Analysis, Vol- ume 1 (Africa), edited by P. Collier and N. Sambanis, Washington, DC: The World Bank; Finance Magazine, “Kalenjin Liberation Army,” September 15, 1992, pp. 20–6; National Council of Churches of Kenya (1992), The Cursed Arrow, Nairobi: NCCK; Republic of Kenya, Parliamentary Select Committee (1992), Report of The Parliamentary Select Committee to Inves- tigate Ethnic Clashes in Western and Other Parts of Kenya, Nairobi: Kenya Parliament.

88

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Subnational Tensions

Over the centuries, their youthful fighters have helped the

Oromo to spread from their homeland in Borena throughout

the Ethiopian lowlands, occupying the northeastern territo-

ries near the Red Sea and the southwestern regions bordering

Kenya.

Until the revolution of 1974, the Ethiopian state rested

on foundations forged from traditional states on the high-

lands. The highland kingdoms embodied the “high culture”

of Ethiopia. They defended the elaborate ecclesiastical hier-

archy of the Coptic Church, lived off incomes extracted from

peasants, and participated in the culture of the imperial court.

The Oromo, by contrast, embodied the “low” culture. While

a large number are Christian, few Oromo staffed the hierar-

chy of the church. Their economy is based on pastoralism, not

farming, and their society is egalitarian, not hierarchical.

Propelled by the expansionary dynamics of the age-grade

system, the Oromo peopled the margins of the Ethiopian state,

territorially and ideologically. They became the object of cam-

paigns mounted by the hegemonic center. In the times of the

empire, they were forced to convert to Christianity and to tithe

to the church; following the revolution, they were forced to sur-

render their lands to a socialist state. They have been subject to

forceful occupation by clients of the national government. In

the era of the empire, the center granted court favorites lands

on the frontier and the right to enserf the Oromo who occu-

pied them; following the revolution, the central government

89

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

stationed military units in the periphery and commandeered

food and livestock to feed its soldiers and bureaucrats. Seek-

ing economic development, the present government invests

in the growing of cotton, sugar, and wheat. More often than

not, it locates its projects not in the highly populated center

but rather in the less crowded periphery, resulting in the loss of

land and water rights for the Oromo (de Waal 1991; Salih and

Markakis 1998; Lewis 2001; Marcus 2002).

The politics of Ethiopia is thus marked by conflicts between

a dynamic and expansionary society, which has extended its

territory and claims to land, and a state system that champions

what it regards as the interests of the center. Powerful issues of

culture underlie these conflicts. Central, too, are disputes over

land.

Uganda

When colonizing East Africa, the British had found it bet-

ter to work through rather than to displace the kingdom of

Buganda.6 Conferring upon it the status of a protectorate, they

employed its administration and police to govern other por-

tions of Uganda. And they rewarded the Baganda for their ser-

vices by acceding to their territorial claims, which included

6 Uganda is the country, Buganda the territory of the Baganda, one of the tribes that dwell in Uganda. By the same construction, Bunyoro is the kingdom of the Banyoro people.

90

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Subnational Tensions

ownership of several “counties” that Buganda had seized from

Bunyoro, a neighboring kingdom. When Milton Obote sought

Uganda’s independence from Britain, he found it useful to

ally with the Baganda, and he allied his UPC (Uganda Peo-

ple’s Congress) with the KY (Kabaka Yekka), the court party of

the Kabaka, their paramount chief. Implicit in the agreement

was his government’s support for the land claims of Buganda.

By championing the cause of a party to this dispute, junior

politicians could build political ties with a political kingdom,

secure a powerful political ally, and thereby accelerate their rise

to political prominence. More senior politicians, particularly

those within the upper ranks of the UPC, also took advantage

of the dispute between the two kingdoms. By threatening to

champion the cause of the Bunyoro, they could threaten to

alienate the KY, thus destabilizing the Obote regime. In this

manner, they sought to extort favors from the central govern-

ment. The conflict between two of the most powerful states

in Uganda thus destabilized the Obote government, driving

Uganda close to state failure – and to single-party rule (Kasfir

1976; Kasozi 1994; Hansen and Twaddle 1995; Kabwegyere

1995; Khadiagala 1995).

Conclusion

Africa’s peoples, like the rest of us, desire higher incomes. In

the absence of technical change, the law of diminishing returns

91

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Sowing the Seeds

ensures that the growth of population results in immiseration

rather than prosperity. To elude the power of that law, people

flee to the periphery. In the conditions that prevail in rural

Africa, the search for prosperity thus fuels territorial expansion

and competing claims to land. In times of political disorder,

these local conflicts can accelerate the failure of states.

The dynamics depicted in this chapter are not unique to

Africa, of course. They resemble those that shape political con-

flict in South Asia, although there the protagonists are char-

acterized as “strangers” and “sons of the soil” (Weiner 1978;

Brass 1985) rather than as “tribes,” as in the literature on Africa.

They find their parallel in pre-industrial Europe as well, espe-

cially at the time of the migration of the Germans and Goths

(Bartlett 1993). That the migrants were known as jovenes –

or youths – highlights the role that generational succession

played in the political dynamics of these societies. The feudal

order that emerged in response to these invasions was based

on the exchange of protection of property for political service

(Bloch 1970). That this exchange characterizes political con-

tracts in much of Africa highlights the broader significance of

the dynamics discussed in this chapter.

In the chapter that follows, I turn to the outbreak of political

disorder in late-century Africa. As adumbrated in the fable of

Chapter 2, it was triggered by elite predation – something ren-

dered more likely because of the lowering value of the resources

at the elite’s command, their rising level of political insecurity,

92

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Subnational Tensions

and the high levels of temptation they faced, given Africa’s

resource endowments. While triggered at the elite level, polit-

ical disorder was marked by the rapid spread of insecurity to

the local level, as popular movements rapidly formed and their

members took up arms. The nature of Africa’s societies helps to

account for the speed with which political disorder cascaded

from the center to the periphery of Africa’s states.

93

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

https://doi.org/10.1017/CBO9781316423974.005 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Part Three

Things Fall Apart

https://doi.org/10.1017/CBO9781316423974 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

https://doi.org/10.1017/CBO9781316423974 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

6

Things Fall Apart

This chapter gathers together the threads of the argu-ment. It highlights the impact of changes in key vari- ables – the level of public revenues and the elite’s rate of dis-

count – arguing that sharp, exogenous shocks helped to drive

their value into ranges that threatened the underpinnings of

political order. That these changes took place in an environ-

ment richly endowed by nature meant that the payoffs to the

incumbent elites from defection could rapidly become more

attractive than those to good governance. In the context of

Africa’s resource endowments, the value of these variables

needed to alter but little before predation became more attrac-

tive than stewardship, thus leading to choices that triggered

state failure.

The changes in the values of these variables resulted in

part from the impact of previous choices: the forging of

The title purposely echoes Achebe, C. (1975), Things Fall Apart, New York: Fawcett Crest.

97

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

authoritarian political institutions and the choice of control

regimes. It also resulted from sharp external shocks, the first

economic recession, resulting from the rise of energy prices,

and the second political, resulting from the geo-political

realignment that followed the end of the Cold War.

The Decline of Public Revenues

In late-century Africa, governments faced a decline in public

revenues, resulting from past policy choices, changes in the

global economy, and the predatory behavior of political elites.

The Untaxed Economy

Emizet (1998) notes the web of regulations and controls that

Zaire (present-day Congo) imposed upon the producers of pri-

mary products. “The goal of these institutional arrangements

was to expropriate economic surplus . . . ,” he writes (Emizet

1998, p. 105). But, he notes,

Citizens . . . reacted to the existing institutional arrangements

by exiting the official economy, especially in coffee growing

and gold regions of Kivu, Upper Congo (Haut Congo) and

Lower Congo (Bas Congo), as well as in the diamond regions of

Eastern Kasai (Kasai Oriental). . . . The central bank reported

that these activities in the second half of the 1970s cost the

government an annual average of 15 percent equivalent in tax

revenues. (Emizet 1998, pp. 105–6)

98

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

In 1982 Zaire had exported 2,000 kg of gold to Belgium;

neighboring Burundi had exported less than 1,000. By 1990, it

was Zaire that exported less than 1,000 kg of gold to Belgium

and Burundi that exported 2,000. Evidence from the dia-

mond industry also suggests high levels of smuggling, with the

amount exported illegally being “50 to 100 percent of recorded

exports” (Emizet 1998, p. 122). The regulation and taxation of

economic activity thus led to the flight of the real economy

from the reach of the government.

The Global Economy

In response to sharp increases in energy prices and the costs

of capital, in the early 1980s, the level of unemployment in

the advanced industrial (OECD; Organisation for Economic

Co-operation and Development) nations rose by 50% and the

rate of economic growth fell to less than 1%. The demand for

imports therefore plummeted and the value of Africa’s exports

declined. So too, did the revenues generated by taxes on trade,

the single largest source of public revenues for most of Africa’s

governments (see Figure 6.1).

As producers of oil, several African states in fact gained

from the rise in petroleum prices; producers of coffee and

cocoa also benefited from a late-century price rise, resulting

from a sharp drop in exports from Latin America. The gov-

ernments of the nations that thus prospered launched new

99

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart 25

30 35

40 45

P er

ce nt

G ov

er nm

en t R

ev en

ue s

1970 1975 1980 1985 1990 1995 year

95% CI Fitted values

Revenues from Trade by Year

-. 01

0 .0

1 .0

2 R

at e

of G

ro w

th

1970 1975 1980 1985 1990 1995 year

95% CI Fitted values

Growth of Government Revenues by Year

Figure 6.1. Government revenues.

projects, but following the later return of petroleum prices to

normal levels, they then found themselves burdened by the

costs of these ventures. Many then borrowed, finding willing

lenders among banks now flush with deposits from the oil-

producing states. When the commodity booms receded, these

governments were then faced with the costs of servicing their

debts. As had the governments of nations whose export earn-

ings had declined, governments in nations that initially ben-

efited from changes in the global economy therefore found

themselves financially strapped.

The late twentieth century marked a time of fiscal crisis for

the state in Africa.

100

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

Predation

As stated by Sahr John Kpundeh (1995, p. 24), “during the

period 1983–1986, it was difficult to distinguish the Sierra

Leone government from a private enterprise. . . . ” Resisting any

attempt to form an independent central bank, Siaka Stevens, its

president, pegged the national currency at an artificially high

level and began rationing access to it. He allocated a major por-

tion to the National Trading Company, to which he assigned the

exclusive rights to import of nearly 100 commodities. As joint

owner of the company, Stevens shared in its monopoly prof-

its. Had foreign exchange been allocated by the market rather

than by discretion, its sale would have swelled the coffers of

the state rather than the bank account of its president.1

Even more dramatic was Stevens’s plundering of the dia-

mond industry (Reno 1995). Sierra Leone’s diamond deposits

lay in a region that supported the Sierra Leone People’s Party

(SLPP), the political opposition, and were worked by a private

corporation, the Sierra Leone Selection Trust. As a member

of De Beers, the international diamond cartel, Selection Trust

tightly regulated diamond production so as to underpin prices

1 See also Reno, W. (1995), Corruption and State Politics in Sierra Leone, Cambridge, U.K.: Cambridge University Press; and Reno, W. (2003), Sierra Leone: Warfare in a Post-State Society, in State Failure and State Weakness in a Time of Terror, edited by R. I. Rotberg, Cambridge, MA, and Washing- ton, DC: The World Peace Foundation/Brookings Institution: 71–100.

101

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

in the global market. The taxes it paid constituted a major por-

tion of the public revenues of Sierra Leone.

Siaka Stevens reconstituted Selection Trust as the National

Diamond Mining Company, however. Ostensibly representing

the nationalization of the industry, the restructuring instead

represented its privatization: Stevens and his cronies domi-

nated both the board and management. By dismantling the

controls imposed by Selection Trust, Stevens permitted the

working of the alluvial deposits by private individuals, tak-

ing care to allocate licenses to political loyalists. Those who

entered diamond production formed political colonies in the

heartland of the opposition. Serving as local units of the ruling

party, they helped to convert – or to intimidate – those about

them into supporting the government in power.

Stevens thus benefited financially and politically from the

transformation of the diamond industry; the state lost out.

Indicative of the magnitude of the diversion of funds is the

magnitude of the decline of reported diamond production,

which fell from 595,000 carats in 1980 to 48,000 in 1988 (Smillee,

Giberie et al. 2000). Also indicative is the decline in tax pay-

ments, which fell from $200 million in 1968 to $100 million in

1987 (Musah 2000).

As indicated in Figure 6.2, the share of central government

revenues in Sierra Leone’s gross domestic product eroded,

falling to less than 5% at one point in the 1990s.

102

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart 0�

5� 10

� 15

� 20

R ev

en ue

s as

P er

ce nt

o f G

D P

19�75� 19�80� 19�85� 19�90� year�

Figure 6.2. Fall of government revenues, Sierra Leone.

The Impact of Declining Revenues

The decline in public revenues adversely impacted the

incomes of public employees. Returning once again to Sierra

Leone, Sahr Kpundeh provides a vivid example. Interview-

ing the Freetown Commissioner of Taxes in the md-1980s,

he “was shown his pay stub. . . . If he buys a bag of rice . . . to

feed his family [or] pays . . . for transportation to and from work

every day, his expenses exceed his earnings” (Kpundeh 1995,

p. 67). The commissioner therefore worked fewer hours in his

public office and more in the private economy. Janet Mac-

Gaffey reports similar findings for Kinshasa in 1986. Employees

103

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

simply could not survive on the salaries paid them by the gov-

ernment, she concludes (MacGaffey 1991).

Given the erosion of public sector salaries, the quality

of public services declined. Teachers abandoned their class-

rooms, nurses left clinics untended, and offices stood empty

while public servants turned to private trade in search of

income. In addition, the level of corruption rose. In economies

in which the government regulated prices, goods disappeared

from the shelves; those in charge would sell the product to

those willing to pay the market as opposed to the official

price. The same was true in post offices, where stamps might

be scarce at the window but be available on the street; or in

medical or veterinary offices, where pharmaceuticals might

be in short supply but available in private clinics. In schools,

children found themselves paying for supplies that once were

freely provided; in hospitals, patients found it necessary to

“tip” to secure a towel, a washcloth, or a bed pan.

A bureaucracy that had been created to facilitate the lives

of the citizens began instead to undermine their welfare. Its

members began to feed themselves by consuming the time

and money of those they once had served.

The most visible of those endowed with the power to coerce

was, of course, the military. Their salaries, too, eroded or

fell into arrears. Their uniforms became tattered, the qual-

ity of food declined in their mess halls, and their equipment

104

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

malfunctioned and, for want of funds, could not be repaired.

In his account of the events leading up to the attempted coup

in Kenya in 1982, James Dianga (2002) stresses the lack of such

basics as proper clothing, palatable food, and affordable hous-

ing. The soldier “has signed a contract with the State,” he argues

(p. 48). The soldier will defend the state; the state will ensure a

decent life for the soldier. But with the “decline in the supply of

uniforms,” Dianga writes, soldiers began “to wonder why the

contract was not being honored” (Dianga 2002, p. 49).

As the value of their salaries declined, soldiers began to pay

themselves. Like doctors and nursing aides, they sold services

to which the citizens were formally entitled. Most commonly,

they regulated access to public thoroughfares. As Kasozi states

for Uganda in the mid-1980s:

Any soldier who needed money . . . would just pick an isolated,

strategic part of the road, put logs or chains across it, and

wait for unfortunate travellers. These twentieth-century high-

waymen would rob everyone of anything they fancied: cash,

watches, casette radios, clothes, and the like. (Kasozi 1994,

p. 152)

In Zaire, soldiers turned to looting. In the early 1990s,

Mobutu attempted to draw Étienne Tshisekedi, the leader of

the opposition, into his ruling clique. Tshisekedi sought not

only an illustrious title – that of prime minister – but also power

105

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

and therefore demanded control over the military and the

central bank. Mobutu conceded to Tshisekedi’s demands, but

he pressured him to direct the bank to pay off long-standing

debts: arrears built up with utility companies, bills submit-

ted by suppliers of petroleum, transport services, and gov-

ernment stores, and salaries owed civil servants and soldiers.

When Tshisekedi refused, Mobutu fired him and then ordered

the central bank to pay. The result was a flood of emissions.

Soldiers received their salaries in the form of new banknotes;

but local merchants, knowing the government to be bankrupt,

refused to accept them. Having been paid in scrip deemed to

be worthless, the armed forces responded by going on “looting

sprees” (Lemarchand 2003), p. 40.2 They demolished down-

town Kinshasa, the national capital, cordoning off commercial

blocks, chasing shopkeepers from their premises, smashing

windows, and carting off food, clothing, furniture, and appli-

ances. Similar disturbances broke out in Lubumbashi in 1991;

in Mbanzu-Ngungu, Goma, and Mbandika in 1992; and in

Kisangani, Goma, and Rutshuru in 1993. On the one hand,

these “pillages,” as they were called, signaled the paucity of

the resources with which to pay public servants; on the other

they heralded the breakdown of the state.

2 See also Pech, K. (2000), The Hand of War: Mercenaries in the Former Zaire 1996–97, in Mercenaries: An African Security Dilemma, edited by A.-F. Musah and J. K. Fayemi, London: Pluto Press; and Nzongola-Ntanlaja, G. (2002), The Congo from Leopold to Kabila, London: Zed Books.

106

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

Managing Regionalism

The decline of public revenues also made it more difficult

to manage regional tensions. In Cote d’Ivoire, for example,

political order rested on a series of pacts negotiated between

regional elites and the center (see Azam 1994; Boone 2003).

Southerners, and in particular the Akan, controlled the center.

Prominent in the periphery were the Senoufo, who possessed

a well-organized polity in the north. “The complaint of the

northerners,” Boone writes, was that “their region was impov-

erished and relegated to backward status in the national polit-

ical economy. . . . ” (2003, p. 263). To counter mounting discon-

tent, President Houphouët-Boigny launched a series of public

initiatives starting projects that led to the opening of paras-

tatal agencies, the construction of roads, and the founding

of cotton and livestock industries in the region. Channeling a

massive flow of benefits to the area dominated by the Senoufo,

the government recruited members of the ruling clans into the

agencies that managed these projects (Boone 2003, pp. 267ff).

Should revenues fall, however, the government would be

unable to fulfill the periphery’s demands. And indeed with

the end of the coffee boom of the late 1970s, those in the

center could no longer credibly pledge to target the north with

largesse (Rapley 1993). The north therefore began to organize

against the central government. After the death of Houphouët-

Boigny, the forces of the north gathered about Allasane

107

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

Öuttara; once prime minister, he now sought to become

president. Led by Laurent Gbagbo, southern politicians rallied

to check the rise of Öuttara, portraying him as a non-national

and therefore ineligible for high office. The courts agreed.

Following a coup by soldiers, whom the government had

failed to pay, the rival politicians transformed their political

organizations into armed militias. Cote d’Ivoire collapsed.3

When public revenues begin to decline, then, the likelihood

of political disorder increases. When poorly reimbursed, pub-

lic servants use political power to raise their own pay; they

become more predatory. Fiscal dearth also renders it more

difficult to induce those who are dissatisfied to continue to

participate in the political game, rather than withdraw from it;

regional tensions therefore rise, and with them, threats to the

integrity of the state. As a result, political order is threatened

by the conduct of the elite in the core and of politicians in the

periphery of Africa’s states.

Political Reform

The decline of public revenues not only triggered efforts by

public employees to pay themselves, it also incited popular

opposition to those in power. In response to the declining

3 For an incisive analysis, see Azam, J.-P. (2001), “The Redistributive State and Conflicts in Africa,” The Journal of Peace Research 38(4): 429–44.

108

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

quality of public services – and public life – in Africa, people

demanded political reform. Citizens called for changes in the

institutions that structured political life. They sought to render

government the servant of the citizen, rather than her master,

and viewed the introduction of multiparty politics and com-

petitive elections as the way of achieving that end. While the

impulse for reform originated within Africa itself, it also arose

among the continent’s creditors, as those who held its debt

sought ways to alter the policy choices of its regimes. By ren-

dering governments accountable to their people, they sought

to create incentives for them to choose policies that would

promote the growth of Africa’s economies and bring greater

prosperity to its people.

While the reforms were designed to secure political ac-

countability and economic prosperity, they also contributed

to political disorder. By raising the level of insecurity for those

in power, they strengthened the incentives for them to defect,

engaging in predation and thus provoking their citizens to take

up arms.

The Local Impulse

Oquaye (1980), writing about life in Ghana, recalls blackouts

because of the “breakdown of . . . electricity supply” (p. 38).

Children, he writes, had to drink “from filthying pools” . . . and

“septic tanks remained un-flushed,” raising the risk of disease

109

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

(ibid., p. 38). Shortages of petrol led to the breakdown of the

transport system, resulting in increased prices for food (ibid.).

Life in Sierra Leone traced a similar trajectory, as roads and

railways in the interior fell into disrepair (Richards 1995); pub-

licly owned companies shut down for lack of power and main-

tenance (ibid., p. 26). Most galling, Richards reports, was the

decline of the educational system, which deprived youths of

what their families had regarded as their “birthright” (ibid.,

p. 177): the chance to acquire skills, to improve their future

prospects, and to enhance the quality of their lives. Again

and again students rallied in protest against the low quality of

their schools. In some instances, their parents joined in these

demonstrations. And in reaction, the government dispatched

troops to beat, arrest, and detain those taking part (ibid.,

pp. 54ff).

Growing dissatisfaction with the quality of public services –

and punitive response to calls for their improvement – gen-

erated calls for political reform. Benin provides an apt illus-

tration. In 1975, the ruling party had endorsed “Marxist-

Leninism” and the government had expanded the range of its

services and the size of its civil service accordingly. By the late

1980s, however, the government lacked the resources to pay

its workers. The result was wave after wave of demonstrations

by government employees and increased indiscipline amongst

soldiers. While unable to meet the salaries of those it employed,

110

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

the political elites did manage to find ways to pay itself: In 1988,

the issuance of $500 million in unsecured loans to the presi-

dent and his cronies led to the collapse of three state-owned

banks. Such acts inspired further demonstrations, encouraged

and cheered on by ambitious challengers to the incumbent

regime.

Paralyzed by the mounting waves of protest, the president

of Benin, Mathieu Kerekou called for a “Conference Nationale

des Forces Vives . . . at which business, professional, religious,

labor, and political groups, together with the government,

would be given an opportunity to draw up a new constitutional

framework” (Meredith 2005, p. 388).4 Kerekou had expected to

dominate the proceedings of the conference, but he failed to

do so. Declaring themselves a sovereign assembly, the confer-

ees dissolved the government, appointed a new prime min-

ister, and laid down a schedule for new elections – elections

that Kerekou lost to Nicephone Soglo, the assembly’s preferred

candidate.

An intriguing feature of the reform movement in Africa was

the tendency for events in one country to respond to, or to

trigger, events in another. Benin’s national conference opened

February 19, 1990; February 25, a second opened in Congo.

4 See also Heilbrunn, J. (1993), “Social Origins of National Conferences in Benin and Togo,” Journal of Modern African Studies 31(2): 227–99.

111

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Table 6.1. The spread of political reform

Election Outcome: Incumbent

Country Conference Date Duration Month Free and Fair Ousted Retained

Benin Feb-90 1 week Feb-91 yes √

Mar-96 yes √

Congo Feb-91 3 months Aug-92 yes √

Gabon Mar-90 3 weeks Dec-93 no √

Mali Jul-91 2 weeks Apr-92 yes √

Niger Jul-91 6 weeks Feb-93 yes √

Burkina Faso Aug-91 2 months Dec-91 no √

Ghana Aug-91 7 months Dec-92 yes √

Togo Aug-91 1 month Aug-93 no √

Zaire Aug-91 1 year – –

Central African

Republic

Oct-91 2 months Aug-92 yes √

Chad Jan-93 3 months Jun-96 no √

112

https://doi.org/10.1017/CBO 9781316423974.006

to the Cam bridge Core term

s of use, available at https://w w

w .cam

bridge.org/core/term s.

D ow

nloaded from https://w

w w

.cam bridge.org/core. Access paid by the U

C San D iego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

The national conference of Benin closed on February 28, 1990;

on March 1, that of Gabon opened (Robinson 1994).

As shown in Table 6.1, following five of the first six national

conferences, the incumbent head of state was compelled to

leave office. The climax came in Zambia, where the national

conference called for multiparty elections. News of Kaunda’s

defeat in these elections (October 1991) resounded throughout

the continent: The forces of political reform had claimed one

of Africa’s “founding fathers.” On the one side, those still in

office found reason to revise upward their assessment of the

magnitude of the threat posed by those clamoring for political

reform. On the other, the reformers took heart, finding reason

to redouble their efforts.

External Forces

By the end of the 1970s, the international community was

fully aware of Africa’s economic plight. Emboldened by the

reformist mandate bestowed by its president, Robert Mc-

Namara, the World Bank had financed a dazzling array of

small-farmer and community-level projects. As recounted in

its official history, the World Bank’s own evaluations revealed

a distressingly low rate of return for its Africa projects: “More

than any other task the Bank had undertaken, its engage-

ment with Sub-Saharan Africa sapped the institution’s . . .

confidence,” it reports (Kapur 1997, p. 720). When seeking

113

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

reasons for the failure of its projects, the Bank found them

in “the policy environment.” In its famed “Berg Report,”5

the Bank documented the tendency of Africa’s governments

to adopt policies that distorted market prices and under-

mined economic incentives and so crippled growth and

development.

In addition to being a financer of projects, the World Bank

then became an advisor to governments. In pursuit of policy

change, it drew upon two sources of strength. The first was

expertise. Through publications, seminars, and the training

of public servants, the Bank sought to expose the economic

costs of prevailing policies and to offer alternatives. The sec-

ond was capital. In any given country at any given time, the

Bank would normally finance a multitude of projects, the can-

cellation of any one of which would go largely un-noticed by

the national government. To gain the attention of policymak-

ers, Please (1984) writes, the Bank therefore began to bundle

its projects into sectoral programs; more would then be at risk

were the Bank to suspend its lending. Sectoral programs soon

gave way to country programs and to conditionality, as the

Bank sought to strengthen further its leverage over policymak-

ers in debtor nations and to sharpen the incentives for policy

reform.

5 World Bank (1981), Accelerated Development in Sub-Saharan Africa: An Agenda for Action, Washington, DC: The World Bank.

114

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

As Africa’s creditors focused on the behavior of African

governments, they struggled with the question: Why would

these governments adopt policies that undermined economic

prosperity? Over time, a consensus emerged: that the behavior

of these governments reflected their lack of political account-

ability. Not being accountable, governments in Africa could

adopt policies that conferred concentrated benefits on the

elites while imposing widely distributed costs on others.

Increasingly, then, the World Bank focused not only on pol-

icy choice but also on political reform.6

Among the most active of those championing political

reform was Keith Jaycox, vice president of the World Bank. In

meeting after meeting, conference after conference, and inter-

view after interview, he called for the introduction of political

reforms. As reluctant as he may have been to call openly for

the introduction of democratic institutions, he left but little

doubt that Africa’s creditors would welcome the legalization

of opposition parties and the holding of competitive elections

for political office.

The economic crisis that alienated Africa’s citizens thus

impelled Africa’s creditors to champion political reform as

well. Lending further impetus to the two political currents

6 See, for example, World Bank (1989), Sub-Saharan Africa: From Crisis to Sustainable Growth, Washington, DC: The World Bank; and World Bank (1991), Governance and Development, Washington, DC: The World Bank.

115

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

was a second shock at the global level. With the collapse of

the Soviet Union, diplomats and security specialists who dur-

ing the Cold War had been disinclined to unseat authoritarian

regimes no longer had reason to object to efforts of economic

technocrats to displace them.7

The disintegration of the Soviet Union strengthened the

position of those inclined to bully rather than to cajole Africa’s

governments. By way of illustration, consider the demise of

Joseph Desire Mobutu, president of Zaire. In the midst of

the Cold War, foreign observers had averted their gaze from

Mobutus’s depredations, largely because of his support in

fighting “communism” in southern Africa. Following the fall

of the Soviet Union, the hands of those pushing for politi-

cal reform in Zaire were no longer stayed by those seeking

Mobutu’s political services. To receive further financial aid,

Mobutu – like other tyrants – had now to reform.

In the 1980s, changes in the international environment

thus amplified the impact of local political forces that had

been calling for reform, and the grasp of Africa’s authoritarians

on political power became less secure. Abandoned by foreign

patrons and facing increasing threats at home, incumbents

had increased reason to fear for their political futures. Their

time horizons therefore shortened. In the long run, repression

7 See the discussion in Dunning, T., “Conditioning the Effects of Aid: Cold War Politics, Donor Credibility, and Democracy in Africa,” International Organization 50(2): 409–23.

116

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

might increase the level of political disorder, but incumbents

had less reason to place great weight on the long run. That even

a founding father like Kenneth Kaunda could be turned out of

office wondrously focused their minds.

Elite Responses

In response to the call for multipartyism, wily cynics, like

Mobutu, sponsored the formation of political parties, rather

than banning their formation. Where plurality voting pre-

vailed, the incumbent could then prevail against a fragmented

opposition. Others, like Daniel arap Moi of Kenya, took more

sinister measures. The shift from Jomo Kenyatta to arap Moi

had entailed a shift from a political base centered in the Central

Province to one located in the Rift Valley and from an old guard,

largely Kikuyu, to a new guard, largely Kalenjin-speaking. The

rise of the reform movement imparted new energy to those

who had been marginalized. To counter their attacks on his

regime, Moi increasingly made use of the coercive powers at

his command. Invoking the Preservation of Public Security Act,

he jailed his political opponents. His security services were

implicated in the killing of a cleric, who was an outspoken pro-

ponent of political reform, and a civil servant, who appears to

have been too diligent in his enquiries into corruption. When

opposition politicians called for an end to the bullying tactics

of Moi and his henchmen, they, too, were arrested, tortured,

117

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

and detained (Hempstone 1997; Anguka 1998; Nnoli 1998;

Mwakikagile 2001; Meredith 2005).

Moi was not the only incumbent to bring the powers of

the state to bear upon his challengers. So, too, did Gnassingbé

Eyadema, the longtime president of Togo. Inspired by events

in neighboring Benin, parliamentarians in Togo had also called

for a national assembly; to the surprise of many, they suc-

ceeded in stripping the president of many of his powers, trans-

ferring them to the office of the prime minister – a figure whom

they, as legislators, would install in office. As befits a military

man, Eyadema fought back. His artillery shelled the palace of

the prime minister; his infantry trampled upon those who took

to the streets in protest; his police closed newspapers and jailed

professionals, party workers, and priests. Eyadema forcefully

repressed those who had challenged him and re-appropriated

the powers of the presidency (Heilbrunn 1997).

As intimidating as Moi or Eyadema might have been, nei-

ther matched the ferocity of the elites of Burundi or Rwanda.

While officially a one-party state, throughout the last decades

of the twentieth century, Burundi in the 1980s was ruled by

its army, and indeed by a small coterie of officers from the

province of Bururi. Pressured by donors abroad and, it would

appear, misperceiving his popularity at home, Pierre Buy-

oya, army major and president, agreed to legalize the forma-

tion of opposition political parties and to call for elections.

118

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

The Front pour la démocratie du Burundi (FRODEBU) consti-

tuted the largest challenge to the incumbent regime. Headed

by Melchior Ndadye, FRODEBU appealed to the majority Hutu

and defeated Buyoya, a Tutsi, in the 1993 elections, win-

ning close to two-thirds of the vote. The elections took place

July 10, 1993; on October 2, 1993, Ndadya was assassinated

and the military, slaying tens of thousands of Hutu, returned

to power (Lemarchand 1993; Ngaruko and Nkurunziza 2000;

Ould-Abdallah 2000).

The slaughter in Burundi resonated ominously with events

in neighboring Rwanda. Rwanda, itself divided between Hutu

and Tutsi, was also ruled by its military, clothed in the guise

of a political party, the Mouvement révolutionnaire national

pour le développement (MRND). But whereas Burundi’s polit-

ical elite was drawn largely from the minority Tutsi, that in

Rwanda came from the Hutu, the ethnic majority. More pre-

cisely, it came from the portion of the Hutu who originated

in Ruhengiri, a prefecture in the northwestern portion of the

country. The danger posed by reform in Rwanda, then, was not,

as in Burundi, revolution by those long suppressed; rather, it

was that the Hutu majority would split, with the “moderates”

aligning with the Rwanda Patriotic Front (RPF) – the milita-

rized political movement that championed the interests of the

Tutsi – to dislodge their northern brethren from power. And

indeed, as the process of political reform proceeded, such an

119

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

alignment became more likely. In the transition government,

negotiated under international auspices in Arusha, both the

MRND and the RPF gained eleven seats. Should the RPF draw

support from one of the minor parties included within the

transitional government – the Liberal, the socialist, or the

Christian Democratic parties – it could then form a govern-

ment. The allocation of posts generated at Arusha thus left

the hardliners insecure, and they determined to render any

alliance between Tutsi and Hutu infeasible.

In neighboring Burundi, the Tutsi-led military had returned

to power by assassinating the leadership of the opposition and

slaughtering their Hutu supporters. The incumbent regime in

Rwanda broadcast these facts widely and portrayed them as

foreshadowing the fate of the Hutu, should the RPF come to

power. They also launched massacres of their own. By attacking

Tutsi in the name of the Hutu, they rendered improbable the

forging of political alliances between the RPF and other politi-

cal parties and incredible the promises of good faith necessary

for their construction. In Rwanda, as in Burundi, attempts to

introduce political reform thus triggered vindictive reprisals

and incumbents inflicted terror and pain upon their citizens

in an effort to forestall the loss of power (Prunier 1998; Jones

1999; Jones 2001). Figure 6.3 suggests the level of co-variation

between political reform on the one hand and the militariza-

tion of civic society on the other.

120

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart .2

� .2

5� .3

� .3

5� .4

� Li

ke lih

oo d

of R

ep or

t o f M

ili tia

1970� 1975� 1980� 1985� 1990� 1995� year�

1. 6�

1. 8�

2� 2.

2� 2.

4� 1=

N o-

2 =

S in

gl e-

3 =

M ul

tip ar

ty �

1970� 1975� 1980� 1985� 1990� 1995� year�

Figure 6.3. Political reform and militarization.

Natural Resources

With the loss of public revenues, governments became more

predatory. With the loss of their political monopolies, they

became less secure. Recall, once again, the opening fable and

the third and last of the variables whose values define the

possibility of political order: the level of temptation. Because

of their rich endowment of natural resources, many govern-

ments in Africa were tempted to abandon their role as guardian

and to embrace the role of predator, employing the power

of the state to extract wealth from the continent’s natural

resources.

121

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

In the midst of economic decline, an alternative source

of income lay close at hand: the continent’s rich deposits of

petroleum, gemstones, and precious metals. To seize such

prizes might require the use of force; it might provoke resis-

tance, particularly in the region in which the resources lay. But

the short-term benefits would readily outweigh the long-term

costs, particularly at a time when governments were finding

increasing reason to discount their political futures.

Consider, for example, the Sudan. In 1962, politicians and

military from the south rebelled against the central govern-

ment, protesting its refusal to agree to a federal form of gov-

ernment and its forced incorporation of southern troops into

the national army. When Jafar Numeri seized the presidency

in 1969, he negotiated an end to the conflict. But in 1978, oil

was discovered in a region lying in the south. Numeri then

redrew the provincial boundaries of Sudan, effectively placing

the oil fields in the portion of the country controlled by the cen-

tral government. The south’s perception of Numeri abruptly

changed; once regarded as a guardian of their interests, he

now appeared a threat. The south soon took up arms again.8

8 See Johnson, D. H. (1995), The Sudan People’s Liberation Army and the Problem of Factionalism, in African Guerillas, edited by C. Clapham, Oxford, U.K.: James Currey; Johnson, D. H. (2003), The Root Causes of Sudan’s Civil War, Bloomington, IN: Indiana University Press; and de Waal, A., and A. H. A. Salam (2004), Islamism, State Power, and Jihad in Sudan, in Islamism and its Enemies in the Horn of Africa, edited by A. de Waal, Bloomington, IN: Indiana University Press.

122

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

Consider, too, the case of the Democratic Republic of Congo

(previously Zaire). A centralized state under Belgian rule, it

fragmented soon after independence. At the forefront of those

who sought to dismantle the state was Katanga, a region richly

endowed with copper, cobalt, and other minerals. Following

Katanga’s forceful reintegration into Congo, however, the prof-

its from the mines accrued to the central government – a gov-

ernment presided over by Joseph Desire Mobutu.9 Succumb-

ing to the temptations offered by these riches, Mobutu became

one of the wealthiest men in the world while presiding over the

disintegration of Zaire.

For a last example, turn to Angola. In the 1960s, a group

of intellectuals, some based in Lisbon and others in Angola’s

capital city, Luanda, formed the Popular Movement for the

9 See Gould, D. (1980), Bureaucratic Corruption and Underdevelopment in the Third World: The Case of Zaire, London: Pergamon Press; Blumenthal, E. (1982), “Zaire: Rapport sur sa Credibilite Financiere Internationale,” La Revue Nouvelle 77(November 11): 360–78; MacGaffey, J. (1991), The Real Economy of Zaire, Philadelphia: University of Pennsylvania Press; Weiss, H. (1995), Zaire: Collapsed Society, Surviving State, Future Policy, in Col- lapsed States, edited by I. W. Zartman, Boulder, CO: Lynne Rienner; Thom, W. G. (1999), “Congo-Zaire’s 1996–1997 Civil War in the Context of Evolv- ing Patterns of Military Conflict in Africa in the Era of Independence,” The Journal of Conflict Studies 19(2): 93–123; Otunnu, O. (2000), An Historical Analysis of the Invasion of the Rwanda Patriotic Army, in The Path of a Genocide: The Rwanda Crisis from Uganda to Zaire, edited by H. Adelman and A. Suhrke, London: Transaction Publishers; and Pech, K. (2000), The Hand of War: Mercenaries in the Former Zaire 1996–97, in Mercenaries: An African Security Dilemma, edited by A.-F. Musah and J. K. Fayemi, London: Pluto Press.

123

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

Liberation of Angola (MPLA). Part political party, part military

force, the MPLA were well positioned when the Portuguese

retreated from Africa. Its leaders quickly seized the national

capital, the central bureaucracy – and Angola’s oil fields. As

described by Birmingham (2002), Chabal (2002), and others

(Dietrich 2000; Meredith 2005), while the MPLA speaks of serv-

ing “the needs of the people,” it in fact channels little of Angola’s

oil wealth to them. The president has retreated to his palace;

the party elite to their villas; and their Mercedes and Land

Cruisers course through streets of Luanda, which is strewn with

garbage and broken glass and inhabited by maimed soldiers.

Other cases could be adduced, each suggesting the manner

in which the temptation to defect – that is, to employ the means

of violence to engage in predation – can overpower the incen-

tives to employ the means of violence to safeguard life and

property. In the midst of fiscal crisis, the temptation increased.

So great are the riches offered by Africa’s natural resources that,

in these instances at least, the rewards to be gained by seizing

them appear to have outweighed the prospects of living in the

midst of political disorder.

Reverberations

Public revenues declined, political elites became insecure,

and the temptation to engage in predation therefore rose in

124

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

Africa, amidst abundant opportunities to do so. In the face of

increased threats from above, citizens found reason to search

for patrons who could help them to safeguard life and prop-

erty. While inflicting widespread costs, disorder also offered

attractive prospects for those willing to invest in the build-

ing of political organizations.10 Among the strategies they

could employ, one stood out: the championing of claims to

land.

To illustrate the process, we return to the politics of east-

ern Zaire. During the colonial period, Rwandans settled in the

district of Masisi in the northern part of the region, attracted

by jobs in the coffee farms and mines of the area. Local chiefs

conferred land rights on the aliens in exchange for the payment

of tribute. By the 1990s, the Bahunde – the local population –

comprised a mere 15% of the population of Masisi and realized

that they now constituted a minority, disadvantaged politically

by their small numbers and economically by the appropriation

10 This is, of course, a central argument in the work of David Keen. See, for example, Keen, D. (1998), The Economic Functions of Violence in Civil War, Adelphi Paper 320, Oxford, U.K.: International Institute for Strategic Studies; Keen, D. (2000), Incentives and Disincentives for Vio- lence, in Greed and Grievance: Economic Agendas in Civil Wars, edited by M. Berdal and D. Malone, Boulder, CO: Lynne Rienner; and Keen, D. (2001), The Political Economy of War, in War and Underdevelopment: The Economic and Social Consequences of Conflict, Vol. 1, edited by F. Stewart, V. Fitzgerald, and Associates, Oxford, U.K.: Oxford University Press.

125

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

of their lands. They therefore began to organize. The Bahunde

demanded the dismissal of immigrants from local government

offices; some had become chiefs. They demanded a change in

the land laws, claiming their rights as “sons of the soil” for

the return of properties sold to “strangers.” And when Rwanda

disintegrated in the east and Zaire beneath them, they took up

arms. They grouped their local militias into a loose but armed

coalition, called Mayi-Mayi, and backed the fortunes of local

politicians who championed the expulsion of the Rwandan

immigrants from eastern Zaire (Pech 2000; Mamdani 2001;

Nzongola-Ntanlaja 2002; Lemarchand 2003).

Further south in Kivu, the thread that tied the region to

Kinshasa, the national capital, had long run through the hands

of one Barthelemy Bisengimana, chief of staff for President

Joseph Desire Mobutu. Kivu was the home of the Banyamu-

lenge, a group that had migrated from Rwanda and Burundi

and taken residence in Zaire. Bisengimana was himself a Tutsi

and championed the rights of the Banyamulenge, defending

in particular their claims to citizenship and land. But when in

the 1980s nationality became a prerequisite for citizenship,

Bisengimana became vulnerable. Labeled a Rwandan and

therefore a foreigner, he was squeezed out of Mobutu’s inner

circle. And with the downfall of their advocate in Kinshasa,

the Banyamulenge, too, became vulnerable. After the fall of

Bisengimana, his enemies – many in search of wealth and

power in the frontier territories of eastern Zaire – revoked the

126

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

citizenship of the Banyamulenge. Because they now could

not vote, they could not secure representation on the local

councils. Gladly taking their place, others allocated to them-

selves the perquisites of office: licenses for vehicles, permits

for shops, and housing. More important still, given the density

of settlement in the region, by labeling the Banyamulenge for-

eigners, they deprived them of land rights and so themselves

gained access to one of the most valuable of resources in Kivu’s

agrarian economy (Nzongola-Ntanlaja 2002; Lemarchand

2003).

As aliens, the Banyamulenge were thus left exposed to the

whims of those with access to power. In response, they took up

arms. And when Rwanda invaded Zaire in 1996, their militias

joined in the crossing of Zaire, the entry into Kinshasa, the

toppling of Mobutu – and the dismembering of the Zairian

state.

Conclusion

In late-century Africa, external shocks and forces set in motion

by previous decisions led to the erosion of the fiscal founda-

tions of the state. In response to the declining quality of pub-

lic life, citizens called for political reform, Africa’s creditors

echoed their demands; and in the early 1990s, both were able

to slip the restraints formerly imposed by foreign powers, now

less motivated by concerns arising from the Cold War.

127

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

In the latter decades of the twentieth century, then, the

values of the variables that define the possibility of the state

altered. Order gave way to disorder, as elites attacked their

own citizens, the latter sought to provide their own security,

and states failed in late-century Africa.

128

https://doi.org/10.1017/CBO9781316423974.006 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

7

Conclusion

In the last decades of the twentieth century, the sinisterlyclownish garb of teenage killers in Liberia, the theatri- cal rage of mobs in Mogadishu, and the dignified suffering of

refugees in camps throughout Africa vividly underscored the

significance of political order. The power of these images cried

out for a response from humanitarians and policymakers. It

challenged scholars as well by posing that most innocent and

unsettling of questions: Why? Why in late twentieth-century

Africa did states fail and things fall apart?

To address these questions, I have retreated to the founda-

tions of my field, which focus on coercion and the properties

of the state. I have also re-immersed myself in the politics of

Africa. From the first came a theory; from the second, the evi-

dence with which to explore – and to test – its answers.

The realities of contemporary Africa compel us to realize

that political order is not a given; it is the product of decisions.

There is political order when citizens choose to turn away from

129

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

military activity and to devote their energies to productive

labor and when those who govern – specialists in violence –

choose to employ their power to protect rather than to prey

upon the wealth that their citizens create. Political order

becomes a state when these choices persist as an equilibrium.

The foundations of the state lie in the conditions that support

that equilibrium; so, too, then, must the origins of state failure.

The fable that framed this analysis highlights the condi-

tions that rendered possible political order. It also suggests the

importance of forces unleashed in the late twentieth century.

Changes in the global economy and economic mismanage-

ment at home resulted in fiscal dearth: The decline in public

revenues led to predation by those in positions of power and to

resistance by those whom they ruled. The fall of communism

permitted erstwhile patrons to abandon abusive incumbents

and enabled those who had protested the quality of gover-

nance to lay claim to the rights of political opposition. Loos-

ing support from abroad and facing new threats from within,

incumbents faced a sharp and unanticipated increase in the

level of political risk. And in many states, the political elite

dwelt in the midst of resources bestowed by nature. Those

in power could seize control of petroleum deposits or dia-

mond fields and be better off, even though bearing the costs

of fighting, than had they continued to subsist on the salaries

paid to those who served the public. It was within this ambi-

ence of temptation that the value of public finances and the

130

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Conclusion

time horizons of elites sharply altered. And it was within the

ambience of local tensions, arising from competition over land

rights and over the power to allocate them, that political dis-

order rapidly spread. The conditions that rendered political

order an equilibrium no longer prevailed and states collapsed

in late-century Africa.

Changing Perspectives

In advancing this argument, I depart in several ways from

the current literature on political violence. Rather than focus-

ing on the protest from below – as do Collier and Hoeffler

(2004), Fearon and Laitin (2003), Kalyvas (2006), Weinstein

(2007), and their predecessors, such as Popkin (1979) and

Scott (1976) – I explore its origins “at the top.” Rather than

probing the motives of rebels or the nature of their organiza-

tions, I instead ask: Why would governments adopt policies

that impoverish their citizens? Why would they “overextract”

wealth from their domains? Why would they alter the distri-

bution of income so grossly that it would become politically

unsustainable? By addressing such questions, I explored the

ways in which incumbent regimes prepared the field for the

forces of political disorder.

Not only do I thus change the point of entry, focusing on

the behavior of incumbents rather than insurgents, but I also

recast the role of the economic forces. In this work, I did not

131

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

focus on national income, as do Fearon and Laitin (2003), Col-

lier and Hoeffler (2004), and Sambanis and Hegre (2006)1; by

the same token, neither did I focus on the impact of poverty,

as do the contributors to the World Bank studies of civil war

(Collier, Hoeffler et al. 2003). Rather, I traced political disorder

to crises in public revenues.

Not only does this work thus depart from contemporary

treatments of the role of economic forces. It also offers new

perspectives on ethnicity, the resource curse, and democra-

tization, several of the central topics addressed in studies of

violence.

Ethnicity

The level of ethnic diversity is greater in the African conti-

nent than in other regions of the world.2 The level of disorder

is high. Many therefore hold ethnicity responsible for Africa’s

political conflicts. To this line of reasoning, I offer two alterna-

tives. The first flows from the inherently expansionary nature

of local societies in rural Africa. Because the search for eco-

nomic well-being underpins a strategy of territorial expan-

sion, groups file competing claims for land rights and political

1 But see Alexander, M. (2007), Is Poverty to Be Blamed for Civil Wars? Cam- bridge MA: Department of Government, Harvard University.

2 See the regional comparisons offered in Easterly, W., and R. Levine (1997), “Africa’s Growth Tragedy: Policies and Ethnic Divisions,” Quarterly Journal of Economics 112(4): 1203–50.

132

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Conclusion

fissures crisscross the nations of Africa. When states are sta-

ble, property rights are secure; when states begin to fail, cit-

izens turn to other sources for their protection. At times of

state failure, politicians can therefore marshal political follow-

ings and recruit armed militias by championing the defense of

land rights. In the midst of state failure, ethnicity may therefore

come to the fore. But by this reasoning, it is the product rather

than the source of political disorder.

Secondly, given that in most African countries some regions

are better endowed than others and that ethnic groups tend

to occupy distinct territories, demands for regional redistribu-

tion take on an ethnic coloring and regional conflicts assume

the guise of ethnic discord. Ethnic conflict is not a “clash of cul-

tures,” then, but rather a struggle over the regional allocation

of resources.

In discussing ethnicity, I have also noted – and stressed –

the disparity between the conclusions drawn from qualitative

accounts of political disorder and those drawn from cross-

national studies of the relationship between ethnicity and state

failure.3 The first emphasizes the significance of ethnicity; the

other, its failure to correlate with measures of political disorder.

3 See also the evidence that the scale of measurement employed in quanti- tative measures – that is, the use of national averages – fails to capture the variability of interest, which occurs at the subnational level. When such variability is captured in the measurements, then statistical estimates of the relationship between ethnic differences rise. See Murshed and Gates (2003) and Cederman and Girardin (2007).

133

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

Rather than arguing for the superiority of a particular method

of research, however, I choose instead to combine the two sets

of findings. Ethnic tensions do in fact relate to political con-

flict in Africa, I would argue, but they do so at times of state

failure.

The Resource Curse

Just as Africa is the continent most blessed with ethnic diver-

sity, so, too, is it the continent most blessed with natural

resource wealth: By one reckoning, 30% of Africa’s population

live in resource-rich economies, as opposed to 11% elsewhere

in the developing world.4 It is natural, then, that its politics is

frequently employed to illustrate the power of the “resource

curse”: the link between natural resource wealth and political

disorder (Collier 2000; Herbst 2000).

Just as observational data for the importance of ethnicity

is contradicted by statistical evidence, so, too, do qualitative

accounts of the role of precious metals and gemstones contra-

dict the quantitative findings. While Collier and Hoeffler (2004)

suggest a close link between the value of primary products

and civil wars, their findings have been called into question

4 Collier, P., and S. O’Connell (2007), Opportunities, Choices and Syn- dromes, Chapter 2 in The Political Economy of Economic Growth in Africa, 1960–2000, edited by B. Ndulu, P. Collier, R. H. Bates, and S. O’Connell, Cambridge, U.K.: Cambridge University Press.

134

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Conclusion

by Fearon (2005). They are also called into question by my

research, which, like that of Fearon (2005), finds only oil pro-

duction to be significantly related to the likelihood of politi-

cal disorder (see Appendix). But just as a combination of the

two kinds of evidence generates a deeper understanding of

the relationship between ethnicity and conflict, so, too, does

it teach us more about the political importance of natural

resources.

Qualitative accounts repeatedly link rebel movements to

the working of deposits of minerals, gemstones, and other

commodities. Statistical investigations largely find little by way

of a relationship between natural resource wealth and political

violence.5 The conflicting evidence suggests to me, at least, the

importance of the temporal course of political disorder. The

first step involves the disintegration of the state; the second,

the turmoil that follows. The quantitative evidence bears upon

the first; it indicates that states whose economies have been

richly endowed are no more likely to fail than are others. The

case materials pertain to the subsequent period of disorder. At

this stage rival forces seek to seize control over timber, metals,

5 See Fearon, J. D. (2005), “Primary Commodities Exports and Civil War,” Journal of Conflict Resolution 49(4): 483–507. See also Snyder, R., and R. Bhavani (2005), “Diamonds, Blood and Taxes: A Revenue-Centered Framework for Explaining Political Order,” The Journal of Conflict Resolu- tion 49(4): 563–597; and Snyder, R. (Forthcoming), “Does Lootable Wealth Breed Disorder? A Political Economy of Extraction Framework,” Compar- ative Political Studies.

135

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

and gemstones, and to employ the resources that flow from

their possession. Resource wealth and political conflict then

co-vary.

On the one hand, this revision stands as a critique: Implic-

itly it charges the earlier literature with having mistaken a

symptom of state failure for a cause. On the other, it stands

as a positive contribution, suggesting an important feature of

the consequences of state failure.

Democratization

In the broader literature on political conflict, scholars treat

political reform with caution. New democracies, they find, are

politically unstable; far more secure are authoritarian regimes

and “consolidated” democracies.6 By contrast, in the literature

on Africa, political reform is widely celebrated and democrati-

zation viewed as valuable, both inherently and instrumentally.

6 See Hegre, H., S. Gates, et al. (2001), “Toward a Democratic Civil Peace? Democracy, Political Change and Civil War, 1816–1992,” American Political Science Review 95(1): 33–48; Hegre, H. (2003), Disentangling Democracy and Development as Determinants of Armed Conflict, paper presented at the annual meeting of the International Studies Association, Portland, Oregon; Goldstone, J., R. Bates, et al. (2005), A Global Forecasting Model of Political Instability, McClean, VA: State Failure Task Force, SAIC; Bates, R., D. Epstein, et al. (2006), Political Instability of Task Force Report, Phase IV Findings, McLean VA: SAIC; and Epstein, D. L., R. Bates, et al. (2006), “Democratic Transitions,” American Journal of Political Science 59(3): 551–69.

136

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Conclusion

Clearly, this book can be read as supportive of the argu-

ments of those who are skeptical of the benefits of political

reform. By provoking a sharp, upward revision in the level of

political insecurity of incumbent regimes, I have argued, polit-

ical reform provoked political disorder. But further reflection

suggests an alternative reading. Recall that it was authoritar-

ianism that lay the foundations for state failure; multiparty

political systems would have been less likely to impose control

regimes as governments that advocated such policies could not

have retained the support of the political majority. Insofar as

authoritarian governments can champion policies that under-

mine their economies, political reform thus removed a major

source of political instability. Moreover, because the evidence

linking political reform to political disorder derives from less

than a decade of data, it may be misleading. We need further

evidence before we can determine whether the relationship

between political reform and political disorder reported here

represents the turbulence associated with transitional dynam-

ics or constitutes, as the skeptics would have it, the properties

of a new steady state.

State Failure

In the late twentieth century, the political foundations of Africa

were hit with shocks, both economic and political, and subject

137

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Things Fall Apart

to forces that eroded political order. Posed dispassionately,

Africa was subject to an experiment, as these forces pushed

the value of key variables into ranges in which the possibility

of political order became vanishingly small. It was the misfor-

tune of Africa’s peoples to be caught in a perfect storm – one

in which political fundamentals were so altered that the foun-

dations of the state lay nakedly revealed: a sight that was both

horrible – and instructive.

In closing, we return one last time to the fable and turn to

a portion that, until now, has remained un-read. The state has

collapsed. And in the midst of the disorder that then engulfs

the specialist in violence and the citizenry, the government

turns to predation while the citizens enlist behind champions

who offer protection in exchange for political services. People

now dwell in a world wherein the government has turned into

a warlord and where they themselves have picked up arms.

Following the logic delineated by Bates, Greif et al. (2002),

we can learn more about the subsequent fate of these people.

Among the insights we achieve is that in the midst of political

disorder, they must trade off between peace and prosperity.

When private individuals provide their own protection, one

way they can achieve security is by being poor: They can “deter”

attacks by having few possessions worth stealing. In the midst

of state failure, then, poverty becomes the price of security.

Cruelly, the opposite also follows: The price of prosperity is

being prepared to fight. In a world in which people provide

138

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Conclusion

their own protection, if they wish to accumulate wealth, they

must be prepared to defend it. They must be willing to pick up

arms.

Whereas those who live in states can enjoy both security

and prosperity, those who live where states have failed must

choose whether to be wealthy or secure; without being willing

to fight, they cannot be both. The formation of militias midst

diamond fields is thus emblematic of the way in which people

must live when states fail.

139

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

https://doi.org/10.1017/CBO9781316423974.007 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Part Four

Appendix

https://doi.org/10.1017/CBO9781316423974 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

https://doi.org/10.1017/CBO9781316423974 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Cross-National Regressions

This study has drawn on a combination of argument,narrative, and quantitative data. The narratives form the body of the manuscript and the formal arguments ani-

mated the opening fable, but until now the statistical evidence

has lurked in the background. It is time for it to step forward.

In this appendix, I first discuss the data and the inferential

challenges they posed. I then bring statistical analysis to bear

upon the three central phenomena addressed in the study:

policy choice, political reform, and political disorder.

The Data

The data form a time series, cross-sectional panel, drawn from

46 countries (see Table A.1) and 26 years (1970–1995).

Most states in Africa achieved independence in the early

1960s, and many initially were unable to gather and report

data of key interest to this study. I judged 1970 to offer a suitable

143

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

Table A.1. Countries in the forty-six-nation sample, 1970–1995

1. Angola 24. Madagascar

2. Benin 25. Malawi

3. Botswana 26. Mali

4. Burkina Faso 27. Mauritania

5. Burundi 28. Mauritius

6. Cameroon 29. Mozambique

7. Cape Verde 30. Namibia

8. Central Africa Republic 31. Niger

9. Chad 32. Nigeria

10. Comoros 33. Rwanda

11. Congo, Republic of 34. São Tomé and Principe

12. Cote d’Ivoire 35. Senegal

13. Djibouti 36. Seychelles

14. Equatorial Guinea 37. Sierra Leone

15. Ethiopia 38. Somalia

16. Gabon 39. Sudan

17. The Gambia 40. Swaziland

18. Ghana 41. Tanzania

19. Guinea 42. Togo

20. Guinea-Bissau 43. Uganda

21. Kenya 44. Dem. Rep. of the Congo

22. Lesotho 45. Zambia

23. Liberia 46. Zimbabwe

compromise between the depth of the panels and the prob-

lems posed by missing data, and 1970 therefore became the

initial year of the sample. Beginning this project in 1997, I ini-

tially adopted 1995 as the terminal year; the data come from

144

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

published sources and a two-year time lag between the time of

writing and the date of the most current observation seemed

the best that could be achieved. As no other end date would be

any less arbitrary and as attempts at currency would run afoul

pauses for analysis and publication, I have therefore stuck with

1995 as the cutoff date for the sample.1

The panel is composed of all independent countries in sub-

Saharan Africa, with the exception of South Africa. The mag-

nitude of the South African economy and the character of its

politics rendered it a major – and potentially highly influen-

tial – outlier.

Working with a talented team of graduate students, I

gathered economic and financial data from sources com-

monly employed by those building cross-national samples of

country-level data. Tables A.2, A.6, and A.9 describe the char-

acteristics of the measures employed in each portion of the

analysis and the sources from which they were taken.

General Overview

Shaping the strategy of estimation were the extent and inci-

dence of missing data, difficulties of measurement, and pat-

terns of dependence among the observations.

1 The data can be found at http://Africa.gov.harvard.edu. Macartan Humphreys built the original Web site; Maria Petrova updated it. As I write, I am continuing to update the data.

145

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

Missingness

Some countries in some years failed to report statistics. To

the extent that the problem originated from the institutional

weakness of Africa’s newly independent states, limiting the

depth of the panel provided a remedy. The problem also arose,

however, from the subsequent collapse of some of these states.

Because of the potential for selection bias, this source required

a more sophisticated response.

Until recently, scholars, when faced with missing data,

have reverted to “list-wise deletion”: If data for a variable

were missing, then the case was dropped from the data set.

Not only does the dropping of cases throw information away

and thus render estimates inefficient, but also, should the

data not be missing completely at random, then list-wise

deletion may render the estimates biased (King, Honeker et

al. 2001). For the reasons just discussed, in this instance,

the likelihood of bias approaches certitude. Employing the

methods championed by Rubin (1996) and implemented by

Schafer (1997), I therefore created multiple data sets that

incorporate values for missing observations that have been

statistically imputed using variables whose values could be

observed. I have posted the resultant data sets on my Web

site (http://people.iq.harvard.edu/∼Rbates/), along with the R-scripts employed to generate them.

146

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

Qualitative Dependent Variable

Having reviewed the problems confronting those who seek

accurate body counts, I simply could not bring myself to use

reported deaths as a measure of disorder.2 I attempted to cal-

culate the percent of a nation’s territory controlled by rebels,

but accurate estimates of this variable also proved difficult to

devise. Data on refugees and displaced persons appeared to

be much more precise; but the systematic reporting of these

data began late in the sample period and would therefore have

necessitated a severe truncation of the panel.

I sought to use the formation of militias as my indicator of

state failure. I could not use the number of militias as a depen-

dent variable, however: Given that militias often change their

names, I ran the risk of double-counting; and given that some

sought to hide their identity, I ran the risk of undercounting.

Nor could I make use of the number of combatants; public esti-

mates varied wildly, reflecting the incentives of the rebel side

to claim popular backing and of the government to deprecate

such claims.

2 For an illuminating discussion of the accuracy of battle death data in the Liberian civil war, see the appendix to Ellis, S. (1999), The Making of Anarchy, New York: New York University Press. For more general and technical discussions, see the papers made available through the House- holds in Conflict Network, which can be accessed on the World Wide Web (www.hicn.org).

147

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

In the end, I therefore chose to employ a categorical mea-

sure. I created an indicator that took on the value of “1” if there

was any report3 of an armed militia in a given country in a given

year and a “0” otherwise.

Patterns of Dependence

To correct for the impart of correlation across time periods

and within countries on the standard errors, I made use of

robust standard errors, clustering by country. The presence of

militias at one time might well affect the likelihood of their

being reported at another. I therefore adopted the techniques

devised by Beck, Katz et al. (1998), introducing cubic splines to

correct my estimates for the impact of temporal dependence

arising from the events themselves.

Possibilities also arise for interdependence among coun-

tries within the cross sections. I introduced “period dummies”

to control for the impact of shocks that might be common to

the whole sample set of countries: the rise of oil prices in the

1970s, for example, or the end of the Cold War in the late 1980s.

For each observation, I also computed4 the value of the depen-

dent variable in neighboring states, allowing me to control for

the possibility, say, that the likelihood of political reform or

3 Reports are from the sources listed in Table A.9. 4 Computations were conducted with the assistance of James Habyarimana.

148

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

disorder in a country in a given year was related to the extent

of political reform or disorder in its neighbors.

Discussion

Given the nature of the data, the challenge therefore became to

draw precise and unbiased estimates from multiple, imputed

sets of time series, cross-sectional data, with a binary depen-

dent variable, controlling for the sources of error discussed

above. Given the state of the art, the properties of the data set

limited my choice of models.

The following three sections present the estimates that I

have obtained.5

Policy Choices

I attributed the initial choice of control regimes in part

to demands for regional redistribution. Their persistence I

attributed to the authoritarian nature of governments, which

virtually disenfranchised those who bore the costs of these

policies. Table A.2 presents the variables employed in the anal-

ysis of policy choices. Table A.3 presents estimates from a

5 I was assisted in these labors by Matthew Hindman and Marc Alexander, plus others who labored ’round Gary King’s shop at Harvard University, Olivia Lau and Rebecca Nelson in particular. I owe special thanks to Jas Sekhon.

149

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Table A.2. Variables employed in analysis of policy choices

Standard Measure Mean Deviation Source

Dependent Variable Control regime 1 = Yes 0.51 0.500 Ndulu, Collier, et al. 2007

0 = No Syndrome-free 1 = Yes 0.251 0.434 Ndulu, Collier, et al. 2007

0 = No Independent Variables Authoritarian regime 1 if no- or single-party. 0.704 0.456 Keesings Contemporary Archives

0 otherwise Africa Confidential

Economist Intelligence Unit

Military government 1 if head of state is 0.441 0.497 Keesings Contemporary Archives

or ever has been Africa Confidential

military professional. Economist Intelligence Unit

0 otherwise

150

Privileged region 1 = yes 0.891 0.331 Harvard research team 0 = No

President from non- 1 = yes 0.570 0.495 Harvard research team privileged region 0 = No

Period1 1 if 1970–74

0 otherwise

Period2 1 if 1975–79

0 otherwise

Period3 1 if 1980–84

0 otherwise

Period4 1 if 1985–89

0 otherwise

Note: The variables “authoritarian regime?” and “military government?” have been lagged by one year. Ndulu, B., P. Collier, et al. 2007. The Political Economy of Economic Growth in Africa, 1960–2000. 2 vols. Cambridge, U.K.: Cambridge University Press.

151

https://w w

w .cam

bridge.org/core/term s. https://doi.org/10.1017/CBO

9781316423974.008 D

ow nloaded from

https://w w

w .cam

bridge.org/core. Access paid by the U C San D

iego Library, on 07 Jan 2020 at 19:42:33, subject to the Cam bridge Core term

s of use, available at

Table A.3. Correlates of policy regimes, pooled sample

Control Regime Syndrome-Free Policy Choices

Coefficient P > t Coefficient P > t Coefficient P > t Coefficient P > t (1) (2) (3) (4)

Authoritarian regime 0.214 0.636 −0.826 0.106 (0.473) (−1.614)

Military government 0.143 0.774 −1.390 0.016 (0.287) (−2.415)

Privileged region 2.198 0.082 2.207 0.084 −3.530 0.004 −3.547 0.000 (1.746) (1.735) (−2.877) (−3.503)

President from non- 0.982 0.028 0.971 0.036 −0.329 0.552 −0.185 0.748 privileged region (2.191) (2.100) (−0.596) (−0.322)

Period1 1.102 0.005 1.122 0.005 −0.586 0.131 −0.779 0.047 (2.825) (2.817) (−1.512) (−1.984)

Period2 1.609 0.000 1.646 0.000 −1.580 0.001 −1.819 0.000 (5.129) (5.150) (−3.278) (−3.907)

Period3 1.540 0.000 1.573 0.000 −1.796 0.000 −1.989 0.000 (5.048) (5.147) (−3.614) (−4.159)

Period4 1.148 0.000 1.171 0.000 −1.153 0.001 −1.235 0.001 (4.575) (4.650) (−3.311) (−3.386)

Constant −3.704 0.004 −3.631 0.004 3.561 0.006 3.530 0.000 (−2.909) (−2.907) (2.757) (3.526)

Observations 1150 1150 1150 1150

Note: t statistics in parentheses. Robust standard errors, grouped by country.

152

https://doi.org/10.1017/CBO 9781316423974.008

to the Cam bridge Core term

s of use, available at https://w w

w .cam

bridge.org/core/term s.

D ow

nloaded from https://w

w w

.cam bridge.org/core. Access paid by the U

C San D iego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

pooled set of data; given the qualitative nature of the depen-

dent variables (see Table A.2), the estimates are based upon a

logit model. When, I introduce country-specific fixed effects

(Table A.4), I employ a conditional logit model.

The dependent variables are policy choices: the choice

of “control regime” in the left-hand panel (equations 1 and

2) and “syndrome-free” policymaking on the right (equa-

tions 3 and 4).6 In both sets of equations, the period dummies

(Period 1 . . . Period 4) help to control for time-specific effects,

such as changes in the global economy or the interna-

tional balance of power. The reference category is 1990–95

(Period 5).

In Table A.3, the variable “privileged region” takes the value

1 when there exists major regional inequality in a country and

0 when there does not. Such inequalities can arise because

of differences in soil quality (in 53% of the cases in which

such inequalities were judged to exist), mineral deposits (in

47% of the cases), or a climate favorable to the production of

export crops (in 92% of the cases). As this variable is time invari-

ant, it could not be incorporated into fixed effects equations

(Table A.4). I therefore make use instead of a variable that

takes on the value 1 if the incumbent president is from a non-

privileged region and 0 if not. In Table A.4, the coefficients on

6 For details concerning the content of these policies, see Chapter 4.

153

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Table A.4. Correlates of policy regimes, conditional logit

Control Regime Syndrome-Free Policy Choices

Coefficient P > t Coefficient P > t Coefficient P > t Coefficient P > t (1) (2) (3) (4)

Authoritarian regime 0.434 0.268 −0.081 0.835 (1.113) (−0.209)

Military government 0.070 0.859 −0.30791 0.517 (0.177) (−0.652)

President from non- 0.823 0.051 2.024 0.037 −1.567 0.020 −1.5256 0.029 privileged region (1.972) (2.104) (−2.657) (−2.498)

Period1 2.060 0.000 2.024 0.000 −1.161 0.001 −1.1956 0.001 (6.029) (5.873) (−3.291) (−3.287)

Period2 3.981 0.000 4.047 0.000 −3.609 0.000 −3.6319 0.000 (9.153) (9.315) (−5.411) (−5.668)

Period3 4.300 0.000 4.310 0.000 −4.571 0.001 −4.5604 0.001 (9.128) (9.204) (−3.904) (−5.504)

Period4 2.604 0.000 2.657 0.000 −2.215 0.000 −2.1527 0.000 (7.262) (7.406) (−5.532) (−5.504)

Constant −3.704 0.004 −3.631 0.004 (−2.909) (−2.907)

Observations 700 700 675 675

Note: t statistics in parentheses.

154

https://doi.org/10.1017/CBO 9781316423974.008

to the Cam bridge Core term

s of use, available at https://w w

w .cam

bridge.org/core/term s.

D ow

nloaded from https://w

w w

.cam bridge.org/core. Access paid by the U

C San D iego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

the variable therefore indicate the changes in the likelihood of

choosing a particular policy regime when a president from a

non-privileged region enters office.

“Authoritarianism” takes on the value 1 when no- or single-

party systems are in place. The variable “military government”

takes on the value of 1 when the head of state is, or was, a

professional soldier, and the coefficient indicates the differ-

ence in the likelihood of the respective policy being chosen

by a military as opposed to civilian regime. In Table A.3, the

coefficients therefore indicate the difference in the likelihood

of a given policy being chosen when a government is “author-

itarian” or military as opposed to when a government is not.

In Table A.4, the coefficients indicate the changes in the like-

lihood of a given policy choice when associated with changes

in the type of government.

The estimates in Table A.3 indicate that countries char-

acterized by regional inequality are significantly more likely

to adopt control regimes and significantly less likely to

adopt “syndrome-free” or market-oriented economic poli-

cies. The coefficients on the period dummies confirm that

control regimes were abandoned following the end of the

Cold War and that syndrome-free policymaking became more

common.

In Table A.4, the temporal dummies remain highly signif-

icant when country-specific fixed effects are introduced into

the estimates. They strongly underscore the impact of “global”

155

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Table A.5. First differences – determinants of policy choice, pooled sample

Change in Probability of Adopting Control Regime

Change in Probability of Adopting Syndrome-Free Policies

Percentage Change in Probability

95% Confidence

Interval

Percentage Change in Probability

95% Confidence

Interval

Percentage Change in Probability

95% Confidence

Interval

Percentage Change in Probability

95% Confidence

Interval

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

Authoritarian regime 0.031 0.009 0.059 −0.067 −0.114 0.034 Military government 0.023 0.197 0.31 −0.135 −0.212 −0.075 Privileged region 0.233 0.18 0.289 0.254 0.003 0.051 −0.515 −0.581 −0.452 −0.500 −0.574 −0.444 Period1 0.053 0.02 0.095 0.065 0.025 0.112 −0.012 −0.034 −0.009 −0.025 −0.057 −0.003 Period2 0.125 0.069 0.195 0.15 0.09 0.227 −0.100 −0.174 −0.047 −0.132 −0.215 −0.068 Period3 0.138 0.077 0.218 0.159 0.094 0.24 −0.132 −0.225 −0.066 −0.153 −0.254 −0.08 Period4 0.092 0.044 0.154 0.11 0.057 0.174 −0.069 −0.137 −0.022 −0.07 −0.15 −0.03

Note: In each instance, the coeffiicent indicates the percentage change in probability of the choice of policy regime resulting from a movement from 0 to 1 in the value of the independent variable.

156

https://doi.org/10.1017/CBO 9781316423974.008

to the Cam bridge Core term

s of use, available at https://w w

w .cam

bridge.org/core/term s.

D ow

nloaded from https://w

w w

.cam bridge.org/core. Access paid by the U

C San D iego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

forces on the choice of policy regime. The estimates confirm

that when presidents from a non-privileged region assume

power, the likelihood of adopting a control regime signifi-

cantly increases. The coefficients on the type of government

(Table A.3) indicate that authoritarian governments and mil-

itary rgimes are less likely to adopt syndrome-free styles of

policymaking.

Table A.5 provides data on the magnitude of the relation-

ships between the independent and dependent variables and

is based upon the coefficients in equation 1 of Table A.4. The

estimates are produced by varying each variable from 0 to 1

while holding all others at their modal values (in this instance,

0).7 Countries with a privileged region are roughly 25% more

likely to impose a control regime and 50% less likely to lack

syndrome-free policies. The late 1970s to early 1980s emerges

as the period in which governments were most likely to adopt

interventionist policies; countries were roughly 12 to 15% more

likely to adopt control regimes and 13 to 15% less likely to adopt

market friendly economic measures in the late 1970s and early

1980s by comparison with the 1990s. Authoritarian govern-

ments were roughly 3% more likely to adopt control regimes

and 7% less likely to adopt syndrome-free policy regimes; mil-

itary governments, 2% and 14%, respectively.

7 The first two numbers in a given row indicate the lower and upper bounds of the 95% confidence interval of these estimated magnitude of the response of the dependent variable.

157

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Table A.6. Variables pertaining to political reform

Standard Measure Mean Deviation Source

Dependent Variable Multiparty system 1 = Yes 0.208 0.406 Data collected by research team from

0 = No- or single-party Keesings Contemporary Archives Africa Confidential

Reform 1 = Yes: Civilian government 0.637 0.481 Economist Intelligence Unit 0 = Professional soldier head of state

Independent Variables Income Log of GDP per capita (PPP) 1173.38 974.723 Penn World Tables Mark 5.6

Urban population Percent of population living 25.849 13.476 World Development Indicators

Literacy Percent of adult population 58.917 19.921 World Development Indicators

that is illiterate

Modernization Factor score derived from −0.018 0.022 principal components

analysis of INCOME,

LITERACY, and URBAN

POPULATION

Petroleum Value of exports per capita 87.010 14.331 Data collected by research team

in constant U.S. dollars (’000) from commercial sources

158

Trade taxes Percentage central government revenues 34.23 0.659 World Development Indicators

from taxes on trade

Business cycle Weighted average growth rate of G7 0.021 0.019 Data created by research team

economies using data from

Penn World Tables Mark 5.6

Aid dependence Foreign aid percent of central government 53.643 68.934 World Development Indicators

expenditure

Duration

No-party system Length of time in years 2.405 4.685 Data collected by research team from

of duration of political system Keesings Contemporary Archives

Africa Confidential

Economist Intelligence Unit

One-party system ditto 3.415 5.551 ditto

Multiparty system ditto 1.271 3.934 ditto

Neighbor average Average level of reform among neighboring 2.741 1.112 Data collected by research team from

states, where 0 = no-party system, Keesings Contemporary Archives 1 = single-party system, and Africa Confidential 3 = multiparty system Economist Intelligence Unit

1990–95 1 if 1990–95

0 otherwise

Note: All independent variables lagged by one year.

159

https://doi.org/10.1017/CBO 9781316423974.008

D ow

nloaded from https://w

w w

.cam bridge.org/core. Access paid by the U

C San D iego Library, on 07 Jan 2020 at 19:42:33, subject to the Cam

bridge Core term s of use, available at https://w

w w

.cam bridge.org/core/term

s.

Appendix

Political Reform

The fiscal crisis of the states of late-century Africa, I have

argued, led to a declining quality of public services and to

increased predation by pubic officials, both of which sparked

popular resentment and increasing demands for political

reform. Within Africa, political reform was contagious, dissem-

inating rapidly across national boundaries and into neighbor-

ing states. It was lent impetus from abroad, as Africa’s cred-

itors chided its authoritarian regimes. The power of finance

rose sharply when the interests of Western foreign policy-

makers aligned with those of the fiscal technocrats with the

end of the Cold War.

Table A.6 presents the variables employed to test this line of

argument. Reform is marked by the movement from military

to civilian regimes and from no- and single-party to multiparty

systems among the latter. To analyze the determinants of such

movements, I employ a conditional logit model (Table A.7).

As a result of this specification, several countries drop out

of portions of the analysis, some – like Botswana – because they

remained multiparty systems throughout the sample period;

others – such as Swaziland – because they never reformed;

some – such as the Democratic Republic of Congo – because

they were “always” governed by a professional soldier; and still

others – such as Tanzania – because they consistently remained

under civilian rule.

160

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Table A.7. Correlates of political reform, conditional logit

Adoption of a Multiparty System Adoption of a Civilian Regime

Coefficient P > t Coefficient P > t Coefficient P > t Coefficient P > t (1) (2) (3) (4)

Income 0.000 0.344 0.001 0.042

(0.956) (2.222)

Urban population 0.071 0.116 −0.020 0.585 (1.648) (−0.554)

Literacy −0.003 0.915 −0.004 0.867 (−0.109) (−0.172)

Modernization 0.218 0.807 0.393 0.557

(0.252) −0.597 Petroleum 0.001 0.363 0.001 0.405 −0.002 0.171 0.000 0.825

(0.909) (0.832) (−1.393) (−0.221) Trade taxes 0.018 0.211 0.017 0.192 −0.033 0.012 −0.029 0.045

(1.308) (1.343) (−2.767) (−2.204)

(continued )

161

https://doi.org/10.1017/CBO 9781316423974.008

to the Cam bridge Core term

s of use, available at https://w w

w .cam

bridge.org/core/term s.

D ow

nloaded from https://w

w w

.cam bridge.org/core. Access paid by the U

C San D iego Library, on 07 Jan 2020 at 19:42:33, subject

Table A.7 (continued )

Adoption of a Multiparty System Adoption of a Civilian Regime

Coefficient P > t Coefficient P > t Coefficient P > t Coefficient P > t (1) (2) (3) (4)

Business cycle −14.782 0.048 −14.523 0.048 0.359 0.952 0.838 0.888 (−1.987) (−1.983) (0.060) −0.141

Aid dependence −0.002 0.712 0.000 0.996 −0.008 0.234 −0.010 0.051 (−0.369) (0.004) (−1.246) (−1.992)

Duration

No-party system −0.124 0.010 −0.094 0.033 −0.246 0 −0.256 0.000 (−2.588) (−2.132) (−5.023) (−5.641)

Single-party system −0.100 0.004 −0.082 0.016 0.104 0.018 0.100 0.024 (−2.898) (−2.407) (2.396) (2.277)

Neighbor average 0.328 0.054 0.358 0.025 0.545 0.006 0.596 0.000

(1.954) (2.250) (2.939) (3.647)

1990–1995 1.938 0.003 2.292 0.000 0.741 0.074 0.643 0.094

(3.579) (5.388) (1.810) (1.686)

Number of observations 980 728 728 675

Note: t statistics in parentheses. Estimates derived from a conditional logistic model, grouped by country.

162

https://doi.org/10.1017/CBO 9781316423974.008

to the Cam bridge Core term

s of use, available at https://w w

w .cam

bridge.org/core/term s.

D ow

nloaded from https://w

w w

.cam bridge.org/core. Access paid by the U

C San D iego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

As seen in Table A.7, the analysis suggests that political

institutions in Africa exhibit historisis. As indicated by the

coefficients on the duration variable, the longer a country has

been subject to a no- or single-party system, the less likely it

is to change to a multiparty system. The estimates also sug-

gest a relationship between political reform and fiscal dearth.

In the case of civilian regimes, declines in the growth rate

of the advanced industrial nations (“business cycle”) signifi-

cantly and negatively correlate with increases in the likelihood

of changing to a multiparty system; in the case of military gov-

ernments, declines in “trade taxes” significantly and negatively

relate to the likelihood of converting to a civilian form of gov-

ernment.

The coefficient on “urban population” suggests the role

played by urban dwellers in the movement to multiparty

systems; that on “income,” the role played by the middle and

upper classes in overturning military regimes. That protest

diffused across political boundaries is confirmed by the pos-

itive and significant coefficient on “neighbor average,” which

provides a measure of the degree of political liberalization in

neighboring states (see Table A.6).

Interestingly, aid dependence appears not to bear a statis-

tically detectable relationship with the likelihood of change to

a multiparty system; and in the movement to civilian regimes,

the strength of its relationship varies depending upon the vari-

ables included or dropped from the analysis. More robust is

163

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Table A.8. Marginal effects, political reform

Change in Probability

95% Confidence Interval

Income From min to max −0.057 −0.666 0.031 296 6,965

Urban population From min to max 0.105 0 0.87 2.4 81.7

Literacy From min to max −0.049 0.543 0 17.1 100

Petroleum 0.08 −0.052 0.858 From min to max 0 6,574

Trade taxes 0.044 0 0.443 From min to max .022 76.51

Business cycle −0.029 −0.294 0 From min to max −0.018 .052

Aid dependence From min to max −0.025 −0.352 0.034 0 513

Duration of party system No-party system From min to max −0.109 −0.782 0 0 25

One-party system From min to max −0.101 −0.731 0 0 25

Neighbor average From min to max 0.024 −0.005 0.295 0 6

1990–95 From 0.048 0 0.383 0 1

Note: Estimates derived from simulations of a probit that included country dummies. Consult Fernandez-Val. (2004), “Estimation of Structural Parameters and Marginal Effects in Binary Choice Panel Data Models with Fixed Effects,” Cambridge, MA: Department of Economics, MIT.

164

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

the relationship between the time dummy and the likelihood

of reform: the coefficient on the variable “1990–95,” which may

capture the impact of the end of the Cold War, is positive in sign

and statistically significant.

Table A.8 reports the magnitudes of the coefficients of equa-

tion 1 of Table A.7. The coefficient on the measure of dura-

tion under single-party rule suggests that those long subject

to single-party rule, such as Tanzania, would be roughly 10%

less likely to reform than would others, such as Kenya, whose

experience with single-party rule was more limited. The data

in Table A.8 also suggest that countries such as Rwanda or

Burundi, where urban dwellers constitute less than 5% of the

population, were 10% less likely to adopt multiparty rule than

were those, such as Botswana or Cape Verde, where they com-

posed over 60% of the population.

Political Disorder

The analysis of political disorder focused on the impact of three

forces. The first was public revenues: When deprived of suffi-

cient payments for the provision of governance, I argued, elites

would use their power to pay themselves. They would veer from

the equilibrium path, in the words of the fable, and behave

in ways that would render citizens insecure. The second was

political reform: When compelled to allow political opponents

to organize in an effort to displace them, political elites, feeling

165

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Table A.9. Variables pertaining to political disorder

Standard Measure Mean Deviation Source

Dependent Variable

Militias 1 = Yes 0.247 0.431 Data collected by research team from 0 = No Keesings Contemporary Archives

Africa Confidential Economist Intelligence Unit

Independent Variables Revenues Central government’s revenues as

percent of GDP 18.489 0.382 World Development Indicators

Petroleum Value of exports per capita in constant U.S. dollars

87.010 14.331 Data collected by research team from commercial sources

Party System No-party system 1 = Yes 0.355 0.015 Data collected by research team from

0 = No Keesings Contemporary Archives One-party system ditto 0.445 0.015 Africa Confidential

Economist Intelligence Unit Multiparty system ditto 0.205 0.013

Duration No-party system Length of time in years 2.405 4.685 Data collected by research team from

of duration of political system Keesings Contemporary Archives Africa Confidential Economist Intelligence Unit

One-party system ditto 3.415 5.551 ditto Multiparty system ditto 1.271 3.934 ditto

166

Neighbor average Number of neighboring states 1.514 1.953 Data collected by research team from reporting militias or civil or international wars

Keesings Contemporary Archives Africa Confidential Economist Intelligence Unit

Privileged region 1 = yes 0.891 0.331 Harvard research team 0 = no

President from non- privileged region

1 = yes 0.570 1.953 ditto

0 = no Period1 1 if 1970–74

0 otherwise Period2 1 if 1975–79

0 otherwise Period3 1 if 1980–84

0 otherwise Period4 1 if 1985–89

0 otherwise Time since last report Count 5.993 6.235 Instrumental Variables Business cycle Weighted average growth rate of

G7 economies 0.021 0.019 Data created by research team using data from

Penn World Tables Mark 5.6 Trade taxes Percentage central government

revenues from taxes on trade 34.210 16.32 World Development Indicators

Revenues Central government’s revenues as percent of GDP

World Development Indicators

Two-year lag 19.272 9.871 Three-year lag 19.100 9.618

Note: All independent variables lagged one year, unless otherwise noted.

167

https://doi.org/10.1017/CBO 9781316423974.008

D ow

nloaded from https://w

w w

.cam bridge.org/core. Access paid by the U

C San D iego Library, on 07 Jan 2020 at 19:42:33, subject to the Cam

bridge Core term s of use, available at https://w

w w

.cam bridge.org/core/term

s.

Appendix

insecure, would increase the rate at which they would discount

the future, thus fearing less the disorder that might be provoked

by their predatory actions. The last was resource abundance:

in an environment richly endowed with natural resources, the

potential economic losses following state failure could readily

be offset by using the power of the government to appropriate

the deposits. In such a setting, political elites would more read-

ily succumb to the temptation to deviate from the equilibrium

path and to change from guardians to predators.

Table A.9 defines and describes the variable employed to

test this argument and lists the sources from which they were

taken. Table A.10 presents the coefficients derived from two

models: A logit model applied to the pooled sample and a

conditional logit model, which incorporates country-specific

effects. I attempted to locate instruments that would enable

me to correct for the impact of political disorder on public rev-

enues, but I failed to do so. Table A.11 records first differences,

calculated from the first equation in Table A.10, which provide

measures of the magnitude of the coefficients.

The coefficients on “revenues” are of the sign anticipated

but significant only in the first model. The difference in the

estimates from the two models suggests that it is the quantity

rather than changes in the quantity of public revenues that

counts. Note the data in Table A.11 regarding the magnitude

of the coefficients. In the case of “revenues,” the data suggest

that a shift from the level of revenues garnered in Sierra Leone

168

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Table A.10. Changes in the probability of militias

Pooled Model Conditional Logit

Coefficient (1) P > t

Coefficient (2) P > t

Revenues −0.039 0.061 −0.017 0.556 (−2.04) (−0.609)

Petroleum 0.000 0.743 −0.001 0.727 (−0.328) (−0.349)

No-party system −0.459 0.246 −0.788 0.107 (−1.166) (−1.616)

One-party system −1.017 0.045 −2.166 0.000 (−2.004) (−4.071)

Duration No-party system 0.096 0.006 −0.019 0.649

(2.789) (−0.455) One-party system 0.038 0.238 0.146 0.000

(1.179) (3.642) Multiparty system −0.024 0.600 0.065 0.357

(−0.524) (0.920) Privileged region 1.403 0.014

(2.491) President from non- −0.567 0.093 −0.807 0.043

privileged region (−1.679) (−2.027) Neighbor average 0.092 0.161 −0.052 0.552

(1.401) (−0.595) Period1 −1.504 0.000 −1.611 0.002

(−3.526) (−3.169) Period2 −0.914 0.009 −1.398 0.000

(2.615) (−3.565) Period3 −0.039 0.883 −0.076 0.819

(−0.148) (−0.228) Period4 −0.183 0.505 0.015 0.959

(−0.667) (0.051) Time since last report −0.188 0.000 −0.017 0.497

(−5.089) (−0.68) Constant −0.385 0.559

(−0.584) Number observations 1048 813

Note: t statistics in parentheses. In equations 1 and 3, robust standard errors, clustered by country.

169

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Table A.11. Magnitude of effects

Pooled Model (First Differences)

Percent Change in Probability

95% Confidence Interval

Revenues −0.314 −0.634 −0.089 From min to max 3.5 53.5

Petroleum −0.121 −0.897 0.403 From min to max 0 6547.48

No-party system −0.070 −0.209 0.010 From 0 to 1

One-party system −0.111 −0.315 −0.007 From 0 to 1

Duration No-party system 0.486 0.171 0.734 From min to max 0 25 One-party system 0.182 −0.056 0.548 From min to max 0 25 Multi-party system −0.054 −0.351 0.312 From min to max 0 25

Privileged region 0.231 −0.006 0.428 From 0 to 1

President from non- privileged region

−0.072 −0.232 0.006

From 0 to 1 Neighbor average 0.191 −0.047 0.518

From 0 to 1 Period1 −0.137 −0.353 −0.030

From 0 to 1 Period2 −0.103 −0.268 −0.017

From 0 to 1 Period3 0.005 −0.071 0.103

From 0 to 1 Period4 −0.023 −0.114 0.054

From 0 to 1 Time since last report −0.382 −0.708 −0.127

From min to max 0 25

170

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

in the mid-1980s to that in Botswana in the same period would

be associated with a 30% reduction in the likelihood of state

failure. Given that political disorder is likely to depress the

level of public revenues, these estimates are likely to be biased

downward.

The coefficients for “petroleum,” a measure of natural

resource endowments, offer little evidence of a relationship

between either differences or changes in the level of the value

of natural resources and the likelihood of reports of militias.

Insofar as the coefficient on “privileged region” (equation 1)

incorporates the effect of natural resource endowments, its

sign and significance does offer evidence in favor of such a rela-

tionship.7 The coefficient on “president from non-privileged

region” is negative and significant in both equations, suggest-

ing that political tensions decline when the poorer region gains

control of the state. As seen in Table A.11, countries contain-

ing a “privileged region” are 23% more likely to experience

state failure. Having a president from a non-privileged region

reduces the likelihood by 7 percentage points; the estimate is

imprecise, however.

The two sets of findings – that having to do with regional

differences and that with the origins of heads of state – warrant

7 The variable “privileged region” should be viewed as an imperfect first step toward capturing within country variation. Better is the rapidly increasing use of geographic information systems, as by Buhaug, H. and G. Gates, “The Geography of Civil War,” Journal of Peace Research 39(4): 417–33.

171

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

additional discussion. Acemoglu and Robinson (2001) and

Azam and Mesnard (2003) view politics as centering on redis-

tribution. Economic inequality – class, in the case of Acemoglu

and Robinson and region, in the case of Azam and Mesnard –

animates threats and the use of power. In response to threats

and to forestall violence, those with wealth may offer to share

it; but in the absence of credible means to commit themselves

to fulfill such promises, such offers may well be discounted.

What makes such promises credible, they argue, is the restruc-

turing of institutions. In Acemoglu and Robinson (2001), the

restructuring takes the form of empowering the poor, as by

broadening the franchise or increasing the powers of popular

assemblies. In the context of Africa, the lodging of executive

power in the hands of poorer regions may play an analogous

role to empowerment of the lower classes in industrial states.

The greater credibility of pledges to use the power of the state

to redistribute the wealth of the nation may help to account

for the negative relationship between our measure of disor-

der and the holding of the presidency by the poorer regions of

the nation; by the same token, the negative relationship may

provide evidence in support of their arguments.

The last of the “theoretical” variables is the party sys-

tem. A change to a competitive party system, I have argued,

leads to an increase in the level of political risk. Most rele-

vant, then, are the coefficients of Table A.10 that are based

on within-country changes; both the coefficients on the

172

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

“no-party” and “one-party” are negative, and the latter is highly

significant. The estimates of the magnitude of the coefficients

(Table A.11) are based on equation 1; because they pertain to

differences between rather than to changes in levels, they can

have but little bearing on this discussion.

The control variables also yield coefficients of interest.

Those of “Period1” and “Period2” (1970–74 and 1975–79,

respectively) are significantly lower than those of “Period5”

(1990–95, the reference category), thus offering evidence of

an increase in the likelihood of disorder over time. The esti-

mates in Table A.11 indicate that reports of militias were 10

to 14% less likely in the 1970s than in the 1990s. As seen in

the coefficient to “neighbor average,” countries whose neigh-

bors experienced more violence were more likely to experience

disorder, although the coefficient is not statistically significant.

The data in Table A.11 suggest the magnitude of the effect: A

country such as Mauritius, being an island and therefore iso-

lated from unruly neighbors, would be roughly 20% less likely

to experience political disorder than, say, Zaire in 1994, which

bordered six neighbors engulfed in political conflict.

The estimates remain robust to the inclusion of additional

control variables. These include the standard “modernization”

variables: measures of education, income, and urbanization.

They also include measures of shocks: economic shocks, such

as terms of trade or growth shocks; climatic shocks, such as

droughts; and political shocks, such as national elections. They

173

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Appendix

include as well such standard controls as population, to cap-

ture the size of the country; and such non-standard controls, as

a measure of the extent to which ethnic groups sprawl across

national borders.

In conclusion, I would draw attention to the “duration” vari-

ables. In any given year and for any given party system, these

variables indicate the number of years that the system has been

in place. When the duration variables are excluded from the

models, the coefficients on the party system dummies fail to

behave in a systematic manner. When they are included, the

coefficients behave in ways that are meaningful and are highly

robust to the inclusion of other variables and to the choice

of model. Why should this be the case? The pattern may be a

statistical artifact, and therefore of little significance, or it may

suggest something about the nature of political institutions

and therefore be important. I continue to be puzzled – and

tantalized – by this finding, but I leave the issue unresolved

while drawing it to the attention of others.

174

https://doi.org/10.1017/CBO9781316423974.008 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:33, subject

Bibliography

Acemoglu, D., and J. A. Robinson (2001). “A Theory of Political Transitions.” American Economic Review 91(4): 938–63.

Achebe, C. (1975). Things Fall Apart. New York: Fawcett Crest. Adam, C. S., and S. A. O’Connell (1999). “Aid, Taxation, and Development in

Sub-Saharan Africa.” Economics and Politics 11(3): 225–54. Alexander, M. (2007). Is Poverty to Be Blamed for Civil Wars? Cambridge, MA:

Paper submitted to gov. 2782, Spring Term. Department of Government, Harvard University.

Allen, C. (1989). Benin. In Benin, The Congo, Burkina Faso: Economics, Politics, and Society. Edited by C. Allen, M. S. Radu, K. Somerville, and J. Baxter. London and New York: Pinter Publishers: 1–137.

Anguka, J. (1998). Absolute Power: The Ouko Murder Mystery. London: Pen Press.

Asprey, R. B. (1986). Frederick the Great. New York: Ticknor and Fields. Axelrod, R. (1985). The Evolution of Cooperation. New York: Basic Books. Azam, J.-P. (1994). “How to Pay for Peace? A Theoretical Framework with Ref-

erence to African Countries,” Public Choice 83(1–2): 173–84. ———. (2001). “The Redistributive State and Conflicts in Africa.” The Journal

of Peace Research 38(4): 429–44. ———. (2007). The Political Geography of Redistribution. Chapter 6 in The

Political Economy of Economic Growth in Africa, 1960–2000: An Analytic Survey. Edited by B. Ndulu, P. Collier, R. H. Bates, and S. O’Connell. Cam- bridge, U.K.: Cambridge University Press.

Azam, J.-P., and A. Mesnard (2003). “Civil War and the Social Contract.” Public Choice 115(3–4): 455–75.

Banks, A. (1999). Cross-National Times Series Data Archive. Binghamton, NY: Center for Social Analysis, State University of New York at Binghamton.

175

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Bibliography

Barkan, J. D. (1976). “Further Reassessment of ‘Conventional Wisdom’: Polit- ical Knowledge and Voting Behavior in Rural Kenya.” American Political Science Review 70(2): 452–55.

———. (1986). Politics and the Peasantry in Kenya. Nairobi: Institute for Devel- opment Studies, University of Kenya.

Baron, D., and J. Ferejohn (1989). “Bargaining in Legislatures.” American Polit- ical Science Review 83(4): 1181–1206.

Barro, R., and J.-W. Lee (1993). “International Comparisons of Educational Attainment.” NBER Working Paper 4349, April.

Bartlett, R. (1993). The Making of Europe: Conquest, Colonization, and Cultural Change, 950–1350. Princeton, NJ: Princeton University Press.

Bates, R. H. (1976). Rural Responses to Industrialization. New Haven and London: Yale University Press.

———. (1989). Beyond the Miracle of the Market. Cambridge, U.K.: Cambridge University Press.

Bates, R. H., A. Greif, et al. (2002). “Organizing Violence.” Journal of Conflict Resolution 46(5): 599–628.

Bates, R. H., and I. Yackolev (2002). Ethnicity in Africa. In The Role of Social Capital in Development. Edited by C. Grootaert and T. van Bastelaer. New York: Cambridge University Press.

Bates, R., D. Epstein, et al. (2006). Political Instability of Task Force Report, Phase IV Findings. McLean, VA: SAIC.

Bayart, J.-F. (1989). Cameroon. In Contemporary West African States. Edited by Donal B. Cruise O’Brien, John Dunn, and Richard Rathbone. London: Oxford University Press.

Bazenguissa-Ganga, R. (2003). The Spread of Political Violence in Congo- Brazzaville. In Readings in African Politics. Edited by T. Young. Oxford: James Currey.

Beck, N., J. Katz, et al. (1998). “Taking Time Seriously: Time-Series-Cross- Section Aanalysis with a Binary Dependent Variable.” American Journal of Political Science 42(4): 1260–88.

Birmingham, D. (2002). Angola. In A History of Postcolonial Lusophone Africa. Edited by P. Chabal. London: Hurst & Company: 138–84.

Bloch, M. (1970). Feudal Society. Chicago: University of Chicago Press. Blumenthal, E. (1982). “Zaire: Rapport sur sa Credibilite Financiere Interna-

tionale.” La Revue Nouvelle 77(November 11): 360–78. Boone, C. (1990). Merchant Capital and the Roots of State Power. New York:

Cambridge University Press.

176

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Bibliography

———. (2003). Political Topographies of the African State: Rural Authority and Institutional Choice. Cambridge, U.K.: Cambridge University Press.

Brass, P., ed. (1985). Ethnic Groups and the State. London: Croom, Helm. Bueno de Mesquita, B., A. Smith, et al. (2003). The Logic of Political Survival.

Cambridge, MA: The MIT Press. Buhaug, H., and S. Gates (2002). “The Geography of Civil War.” Journal of Peace

Research 39(4): 417–33. Buhaug, H. and J. K. Rod (2006). “Local Determinants of African Civil Wars,

1970–2001.”Political Geography 25: 315–35. Buijtenhuijs, R. (1989). Chad. In Contemporary West African States. Edited by

John Dunn, Donal B. Cruise O’Brien, and Richard Rathbone. Cambridge, U.K.: Cambridge University Press.

Burdette, M. (1988). Zambia. Boulder, CO: Westview. Carneiro, R. L. (1970). “A Theory of the Origin of the State.” Science 169(3946):

733–38. Cederman, L.-E., and L. Girardin (2007). “Beyond Fractionalization: Mapping

Ethnicity on Nationalist Insurgencies.” American Political Science Review 101(1): 173–85.

Chabal, P. (2002). A History of Postcolonial Lusophone Africa. London: Hurst and Company.

Cohen, M. (1974). Urban Policy and Political Conflict in Africa: A Study of the Ivory Coast. Chicago: University of Chicago Press.

Collier, P. (2000). Doing Well Out of Civil War. In Greed and Grievance: Economic Agendas in Civil War. Edited by M. Berdal and D. Malone. Boulder, CO: Lynne Rienner.

Collier, P., A. Hoeffler, et al., eds. (2003). Breaking the Conflict Trap: Civil War and Development Policy. Washington, DC: The World Bank.

Collier, P., and A. Hoeffler (2004). “Greed and Grievance in Civil Wars.” Oxford Economic Papers, New Series 56(4): 563–95.

Collier, P., and S. O’Connell (2007). Opportunities, Choices and Syndromes. Chapter 2 in The Political Economy of Economic Growth in Africa, 1960– 2000. Edited by B. Ndulu, P. Collier, R. H. Bates, and S. O’Connell. Cambridge, U.K.: Cambridge University Press.

Dahl, R. (1971). Polyarchy. New Haven: Yale University Press. Daily Nation. “Constituency Review: Laikipia District.” July 16, 2002, pp. 11–

14. de Waal, A. (1991). Evil Days: Thirty Years of War and Famine in Ethiopia. New

York: Human Rights Watch.

177

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Bibliography

de Waal, A., and A. H. A. Salam (2004). Islamism, State Power, and Jihad in Sudan. In Islamism and Its Enemies in the Horn of Africa. Edited by A. de Waal. Bloomington, IN, Indiana University Press.

Dianga, J. W. (2002). Kenya, 1982: The Atttempted Coup. London: Penn Press. Dietrich, C. (2000). Inventory of Formal Diamond Miniing in Angola. In

Angola’s War Economy. Edited by J. Cilliers and C. Dietrich. Pretoria, South Africa: Institute for Security Studies.

Dumont, R. (1962). L’Afrique Noire est Mal Partie. Paris: Editions Seuil. ———. (1966) False Start in Africa. New York: Praeger. Dudley, B. (1982). An Introduction to Nigerian Government and Politics.

Bloomington, IN: Indiana University Press. Dunning, T. (2004). “Conditioning the Effects of Aid: Cold War Politics, Donor

Credibility, and Democracy in Africa.” International Organization 58(2): 409–23.

Easterly, W., and R. Levine (1997). “Africa’s Growth Tragedy: Policies and Ethnic Divisions.” Quarterly Journal of Economics 112(4): 1203–50.

Elliott, C., ed. (1971). Constraints on the Economic Development of Zambia. Nairobi: Oxford University Press.

Ellis, S. (1999). The Making of Anarchy. New York: New York University Press. Emizet, K. N. F. (1998). “Confronting Leaders at the Apex of the State: The

Growth of the Unofficial Economy of the Congo.” African Studies Review 41(1): 99–137.

Epstein, D. L., R. Bates, et al. (2006). “Democratic Transitions.” American Jour- nal of Political Science 59(3): 551–69.

Evans, P., T. Skocpol, and D. Rueschmeyer (1985). Bringing the State Back In. Cambridge, U.K.: Cambridge University Press.

Fanon, F. (1963). The Wretched of the Earth. New York: Presence Africaine. Fearon, J. D. (2004). “Why Do Some Civil Wars Last So Much Longer than

Others?” Journal of Peace Research 41(3): 275–301. ———. (2005). “Primary Commodities Exports and Civil War.” Journal of

Conflict Resolution 49(4): 483–507. Fearon, J. D., and D. Laitin (2003). “Ethnicity, Insurgency and Civil War.” Amer-

ican Political Science Review 97(1): 75–90. Finance Magazine. “Kalenjin Liberation Army.” Septemberr 15, 1992, pp. 20–6. Fortes, M. (1958). Introduction. In The Developmental Cycle in Domestic

Groups. Edited by J. Goody. Cambridge, U.K.: Cambridge University Press. Fortes, M., and E. E. Evans-Pritchard, eds. (1987). African Political Systems.

New York: KPI in association with the International African Institute.

178

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Bibliography

Fowler, R. R. (2000). “Report of the Panel of Experts on Violations of Secu- rity Council Sanctions Against UNITA.” New York: S. Council, The United Nations.

Fratkin, E., E. A. Roth, et al. (1994). Introduction. In African Pastoralist Sys- tems. Edited by E. Fratkin, E. A. Roth and K. A. Galvin. Boulder, CO: Lynne Rienner.

Gboyega, A. (1997). Nigeria: Conflict Unresolved. In Governance as Conflict Management. Edited by I. W. Zartman. Washington, DC: Brookings Institu- tion Press.

Geddes, B. (2003). Paradigms and Sand Castles. Ann Arbor: University of Michigan Press.

Gertzel, C., C. Baylies, et al., eds. (1984). The Dynamics of the One-Party State in Zambia. Manchester: U.K.: University of Manchester Press.

Goldstone, J., R. Bates, et al. (2005). A Global Forecasting Model of Political Instability. McClean, VA: State Failure Task Force, SAIC.

Gould, D. (1980). Bureaucratic Corruption and Underdevelopment in the Third World: The Case of Zaire. London: Pergamon Press.

Groseclose, T., and J. M. Snyder (1996). “Buying Supermajorities.” American Political Science Review 90(2): 303–15.

Hansen, H. B., and M. Twaddle (1995). Uganda: The Advent of No-Party Democracy. In Democracy and Political Change in Sub-Saharan Africa. Edited by J. Wiseman. London: Routledge.

Hayward, F. (1976). “A Reassessment of Conventional Wisdom about the Informed Public: National Political Information in Ghana.” American Polit- ical Science Review 70(2): 917–41.

Hayward, F. M., and J. D. Kandeh (1987). Perspectives on Twenty-Five Years of Elections in Sierra Leone. In Elections in Independent Africa. Edited by F. M. Hayward. Boulder, CO: Westview Press.

Hegre, H. (2003). Disentangling democracy and development as determinants of armed conflict. Paper presented at the annual meeting of the International Studies Association, Portland, Oregon.

———. (2004). “The Duration and Termination of Civil War.” Journal of Peace Research 41(3): 243–52.

Hegre, H., S. Gates, et al. (2001). “Toward a Democratic Civil Peace? Democ- racy, Political Change and Civil War, 1816–1992.” American Political Science Review 95(1): 33–48.

Heilbrunn, J. (1993). “Social Origins of National Conferences in Benin and Togo.” Journal of Modern African Studies 31(2): 277–99.

179

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Bibliography

———. (1997). Togo: The National Conference and Stalled Reform. In Political Reform in Francophone West Africa. Edited by J. F. Clark and D. E. Gardiner. Boulder, CO: Westview Press.

Helleiner, G. (1966). Peasant Agriculture, Government, and Economic Growth in Nigeria. Homewood, IL: R. D. Irwin.

Hempstone, S. (1997). Rogue Ambassador. Sewanee, TN: University of the South Press.

Herbst, J. (2000). “Economic Incentives, Natural Resources and Conflict in Afruca.” Journal of African Economies 9(3): 270–94.

Herbst, J. (2000). States and Power in Africa. Princeton, NJ: Princeton University Press.

Heston, A., R. Summers, et al. (2002, October). Penn World Tables Version 6.1. Philadelphia: Center for International Comparisons, University of Pennsyl- vania.

Hirshleifer, J. (2001). The Dark Side of the Force. Cambridge, U.K.: Cambridge University Press.

———. (1995). Theorizing about Conflict. In Handbook of Defense Economics. Edited by K. Hartley and T. Sandler. New York: Elsevier.

Huntington, S. P. (1991). The Third Wave. Norman, OK: Oklahoma University Press.

Jalata, A. (2005). Oromo and Ethiopia: State Formation and Environmental Conflict, 1868–2004. Trenton, NJ: Red Sea Press.

Johnson, D. H. (1995). The Sudan People’s Liberation Army and the Problem of Factionalism. In African Guerillas. Edited by C. Clapham. Oxford, U.K.: James Currey.

———. (2003). The Root Causes of Sudan’s Civil War. Bloomington, IN: Indiana University Press.

Jones, B. D. (1999). Civil War, the Peace Process, and Genocide in Rwanda. In Civil Wars in Africa. Edited by T. M. Ali and R. O. Matthews. Montreal: McGill-Queens Press.

———. (1999). Military Intervention in Rwanda’s Two Wars: Partisanship and Indifference. In Civil Wars, Insecurity, and Intervention. Edited by B. F. Walters and J. Snyder. New York: Columbia University Press.

———. (2001). Peacemaking in Rwanda: The Dynamics of Failure. Boulder, CO: Lynne Rienner.

Kabwegyere, T. (1995). The Politics of State Formation and Destruction in Uganda. Kampala, Fountain Publishers.

Kakwenzire, J., and D. Kamukama (2000). The Development and Consolidation of Extremist Forces in Rwanda. In The Path of a Genocide: The Rwanda

180

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Bibliography

Crisis from Uganda to Zaire. Edited by H. Adelman and A. Suhrke. London: Transaction Publishers.

Kalyvas, S. (2006). The Logic of Violence in Civil War. Cambridge, U.K.: Cam- bridge University Press.

Kaplan, R. D. (1994). “The Coming Anarchy.” The Atlantic Monthly 273(Febru- ary 2): 44–76.

Kasara, K. (2007). “Tax Me If You Can.” American Political Science Review 101(1): 159–72.

Kasfir, N. (1976). The Shrinking Political Arena. Berkeley and Los Angeles: University of California Press.

Kasozi, A. B. K. (1994). The Social Origins of Violence in Uganda, 1964–1985. Montreal: McGill-Queen’s University Press.

Keen, D. (1998). “The Economic Functions of Violence in Civil War.” Adelphi Paper 320. Oxford, U.K.: International Institute for Strategic Studies.

———. (2000). Incentives and Disincentives for Violence. In Greed and Grievance: Economic Agendas in Civil Wars. Edited by M. Berdal and D. Malone. Boulder, CO: Lynne Rienner.

———. (2001). The Political Economy of War. In War and Underdevelopment: The Economic and Social Consequences of Conflict, Vol. 1. Edited by F. Stew- art, V. Fitzgerald and Associates. Oxford, U.K.: Oxford University Press.

Kenyatta, J. (1953). Facing Mount Kenya. London: Secker and Warburg. Khadiagala, G. M. (1995). State Collapse and Reconstruction in Uganda. In

Collapsed States. Edited by I. W. Zartman. Boulder, CO: Lynne Rienner. Kimenyi, M. S., and N. Ndung’u (2005). Sporadic Ethnic Violence: Why Has

Kenya Not Experienced a Full Blown Civil War? In Understanding Civil War: Evidence and Analysis, Volume 1 (Africa). Edited by P. Collier and N. Sambanis. Washington, DC: The World Bank.

King, G., J. Honeker, et al. (2001). “Analyzing Incomplete Political Science Data: An Alternative Algorithm for Multiple Imputation.” American Political Science Review 95(1): 49–69.

Kpundeh, S. J. (1995). Politics and Corruption in Africa: A Case Study of Sierra Leone. New York: University Press of America.

Kramer, G. H. (1983). “Electoral Politics in the Zero-Sum Society.” Social Sci- ence Working Paper 472. Pasadena, CA: Division of Humanities and Social Science, California Institute of Technology.

Kriger, N. (2003). Guerrilla Veterans in Post-War Zimbabwe: Symbolic and Violent Politics, 1980–1987. Cambridge, U.K.: Cambridge University Press.

Larmer, M. (2006). “A Little Bit Like a Volcano”: The United Progressive Party and Resistance to One-Party Rule in Zambia, 1964–1980. Typescript.

181

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Bibliography

Leakey, L. S. B. (1977). The Southern Kikuyu before 1903. London: Academic Press.

Lemarchand, R. (1981). “Chad: The Roots of Chaos.” Current History 80(470): 414–18, 436–38.

———. (1993). Burundi in Camparative Perspective. In The Politics of Eth- nic Conflict Regulation. Edited by J. McGarry and B. O’Leary. London: Routledge.

———. (2003). The Democratic Republic of the Congo: From Failure to Potential Reconstruction. In State Failure and State Weakness in a Time of Terror. Edited by R. I. Rotberg. Cambridge, MA, and Washington, DC: The World Peace Foundation/Brookings Institution: 29–70.

Levine, V. T. (1986). Leadership and Regime Changes in Perspective. In The Political Economy of Cameroon. Edited by M. G. Schatzberg and I. W. Zartman. New York: Praeger.

Lewis, H. (2001). A Galla Monarchy. Madison, WI: University of Wisconsin Press.

Little, I. M. D., T. Scitovsky, et al. (1970). Industry and Trade in Some Developing Countries. Oxford, U.K.: Oxford University Press.

MacGaffey, J. (1991). The Real Economy of Zaire. Philadelphia: University of Pennsylvania Press.

Magaloni, B. (2006).Voting for Autocracy: Hegemonic Party Survival and Its Demise in Mexico. New York: Cambridge University Press.

Mamdani, M. (2001). When Victims Become Killers. Princeton, NJ: Princeton University Press.

Mansfield, E., and J. Snyder (1995). “Democratization and War.” Foreign Affairs 74(3): 79–97.

Marcus, H. G. (2002). A History of Ethiopia. Berkeley and Los Angeles: Univer- sity of Los Angeles Press.

Marshall, M. G., and K. Jaggers (2000). Polity IV Project: Political Regime Char- acteristics and Transitions, 1800–1999 Dataset Users’ Manual. College Park, MD: CIDCM, University of Maryland.

May, R. (2003). Internal Dimensions of Warfare in Chad. In Readings in African Politics. Edited by T. Young. Oxford: James Currey.

Mbithi, P. M., and C. Barnes (1975). The Spontaneous Settlement Problem in Kenya. Kampala, Uganda: East African Literature Bureau.

Meillassoux, C. (1981). Maidens, Meal, and Money. Cambridge, U.K.: Cam- bridge University Press.

Meredith, M. (2005). The State of Africa: A History of Fifty Years of Independence. London: Free Press.

182

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Bibliography

Murshed, S. M., and S. Gates (2003). Spatial-Horizontal Inequality and the Maoist Insurgency in Nepal. Paper commissioned by The Department for International Development, UK. London.

Musah, A.-F. (2000). A Country Under Siege: State Decay and Corporate Military Intervention in Sierra Leone. In Mercenaries: An African Security Dilemma. Edited by A.-F. Musah and J. K. Fayemi. London: Pluto Press.

Musah, A.-F., and J. K. Fayemi (2000). Africa in Search of Security. In Mercenaries: An African Security Dilemma. Edited by A.-F. Musah and J. K. Fayemi. London: Pluto Press.

Mwakikagile, G. (2001). Ethnic Politics in Kenya and Nigeria. Huntington, NY: Nova Science Publishers.

National Council of Churches of Kenya. (1992). The Cursed Arrow. Nairobi: NCCK.

Ndulu, B., P. Collier, et al. (2007). The Political Economy of Economic Growth in Africa, 1960–2000, 2 vols. Cambridge, U.K.: Cambridge University Press.

Ngaruko, F., and J. D. Nkurunziza (2000). “An Economic Interpretation of Conflict in Burundi.” Journal of African Economies 9(3): 370–409.

Nnoli, O. (1998). Ethnic Conflicts in Africa: A Comparative Analysis. In Ethnic Conflict in Africa. Edited by O. Nnoli. Dakar, Senegal: Codesria.

Nolutshungu, S. C. (1996). Limits of Anarchy. Charlottesville: University Press of Virginia.

Nzongola-Ntanlaja, G. (2002). The Congo from Leopold to Kabila. London: Zed Books.

———. (2004). The Politics of Citizenship in the DRC. Paper prepared for the annual International Conference. Edinburgh: University of Edinburgh.

Oquaye, M. (1980). Politics in Ghana, 1972–1979. Accra, Ghana: Tornado Pub- lications.

Ottaway, M. (1999). Africa’s New Leaders. Washington, DC: Carnegie Endow- ment for International Peace.

Otunnu, O. (2000). An Historical Analysis of the Invasion of the Rwanda Patriotic Army. In The Path of a Genocide: The Rwanda Crisis from Uganda to Zaire. Edited by H. Adelman and A. Suhrke. London: Transaction Publishers.

Ould-Abdallah, A. (2000). Burundi on the Brink, 1993–1995. Washington, DC: United States Institute of Peace Press.

Pech, K. (2000). The Hand of War: Mercenaries in the Former Zaire 1996–97. In Mercenaries: An African Security Dilemma. Edited by A.-F. Musah and J. K. Fayemi. London: Pluto Press.

Please, S. (1984). The Hobbled Giant. Boulder, CO: Westview.

183

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Bibliography

Polanyi, K. (1991). Dahomey and the Slave Trade. New York: AMS Press. Popkin, S. L. (1979). The Rational Peasant. Berkeley and Los Angeles: University

of California Press. Prunier, G. (1998). The Rwanda Crisis: History of a Genocide. London: Hurst. Rapley, J. (1993). Ivoirien Capitalism: African Entrepreneurs in Côte d’ Ivoire.

Boulder, CO: Lynne Rienner. Reno, W. (1995). Corruption and State Politics in Sierra Leone. Cambridge,

U.K.: Cambridge University Press. ———. (2000). Shadow States and the Political Economy of Civil Wars. In Greed

and Grievance: Economic Agendas in Civil Wars. Edited by M. Berdal and D. Malone. Boulder, CO: Lynne Riennner.

———. (2003). Sierra Leone: Warfare in a Post-State Society. In State Failure and State Weakness in a Time of Terror. Edited by R. I. Rotberg. Cambridge and Washington, DC: The World Peace Foundation/Brookings Institution: 71–100.

Republic of Kenya, Parliamentary Select Committee. (1992). Report of the Par- liamentary Select Committee to Investigate Ethnic Clashes in Western and Other Parts of Kenya. Nairobi: Kenya Parliament.

Reyna, S. P. (1990). Wars Without End: The Political Economy of a Precolonial African State. Hanover: University Press of New England for the University of New Hampshire.

Ricardo, D. (1821). On the Principles of Political Economy and Taxation. London: John Murray.

Richards, P. (1995). Rebellion in Liberia and Sierra Leone: A Crisis of Youth? In Crisis in Africa. Edited by O. Furley. London: I. B. Tauris.

Robinson, P. (1994). “The National Conference Phenomenon in Francophone Africa.” Comparative Studies in Society and History 36(3): 575–610.

Ross, M. L. (2004). “How Does Natural Resource Wealth Influence Civil Wars? Evidence from Thirteen Cases.” International Organization 49(4): 508–37.

Rubin, D. B. (1996). “Multiple Imputation after 18+ Years (with discussion).” Journal of the American Statistical Association 91(434): 473–89.

Rutten, M. (2001). “Fresh Killings”: The Njoro and Laikipia Violence in the 1997 Kenyan Elections Aftermath. In Out for the Count: The 1997 General Elections and Prospects for Violence in Kenya. Edited by M. Rutten, A. Mazrui, and F. Gignon. Kampala, Uganda: Fountain Publishers.

Sachs, J., and A. Warner (1995). “Economic Reform and the Process of Global Integration.” Brookings Papers on Economic Activity (1): 1–118.

Sahlins, M. D. (1961). “The Segmentary Lineage: An Organization of Predatory Expansion.” American Anthropologist 63(2): 322–45.

184

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Bibliography

———. (1968). Tribesmen. Englewood Cliffs, NJ: Prentice-Hall. ———. (1971). Tribal Economies. In Economic Development and Social

Change. Edited by G. Dalton. Garden City, NY: Natural History Press for the American Museum of Natural History: 43–61.

Salih, M. A. M., and J. Markakis. (1998). Ethnicity and State in Eastern Africa. Uppsala, Sweden: Nordiska Afrikainstutel.

Sambanis, N., and H. Hegre (2006). “Sensitivity Analysis of Empirical Results on Civil War Onset.” Journal of Conflict Resolution 50(4): 508–35.

Schafer, J. L. (1997). Imputation of Missing Covariates in Multivariate Linear Mixed Model. University Park, PA: Department of Statistics, The Pennsylva- nia State University.

Schatz, S. P. (1977). Nigerian Capitalism. Berkeley and Los Angeles: University of California Press.

Schumpeter, J. A. (1950). Capitalism, Socialism and Democracy. New York: Harper & Row.

Scott, J. C. (1976). The Moral Economy of the Peasant. New Haven and London: Yale University Press.

Smillee, I., L. Giberie, et al. (2000). The Heart of the Matter: Sierra Leone, Dia- monds, and Human Security. Ottowa: Partnership Africa-Canada.

Snyder, J. (2000). From Voting to Violence. New York: W. W. Norton. Snyder, R. (2006). “Does Lootable Wealth Breed Disorder? A Political Economy

of Extraction Framework.” Comparative Political Studies 39(8): 943–68. Snyder, R., and R. Bhavani (2005). “Diamonds, Blood and Taxes: A Revenue-

Centered Framework for Explaining Political Order.” The Journal of Conflict Resolution 49(4): 563–97.

Stevenson, R. F. (1968). Population and Political Systems in Tropical Africa. New York: Columbia University Press.

Szeftel, M. (1978). Conflict, Spoils, and Class Formation in Zambia. Ph.D. Dis- sertation. Manchester, U.K.: University of Manchester Press.

Tangri, R. (1999). The Politics of Patronage in Africa: Parastatls, Privatizaton, and Private Enterprise. Trenton, NJ: Africa World Press.

Thom, W. G. (1999). “Congo-Zaire’s 1996–1997 Civil War in the Context of Evolving Patterns of Military Conflict in Africa in the Era of Independence.” The Journal of Conflict Studies 19(2): 93–123.

Timmer, C. P. (1969). “The Turnip, the New Husbandry, and the English Agri- cultural Revolution.” Quarterly Journal of Economics 83(2): 375–96.

Turner, V. (1957). Schism and Continuity in an African Society. Manchester, U.K.: University of Manchester Press on behalf of the Rhodes-Livingstone Institute.

185

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject

Bibliography

Vansina, J. (1966). Kingdoms of the Savanna. Madison: University of Wisconsin Press.

Waller, R., and N. W. Sobania (1994). Historical Perspectives. In African Pas- toralist Systems. Edited by E. Fratkin, K. A. Galvin, and E. A. Roth. Boulder, CO: Lynne Rienner.

Weber, M. (1958). Politics as a Vocation. In From Max Weber. Edited by H. H. Gerth and C. W. Mills. New York: Oxford University Press.

Weiner, M. (1978). Sons of the Soil: Migration and Ethnic Conflict in India. Princeton, NJ: Princeton University Press.

Weinstein, J. (2007). Inside Rebellion. Cambridge, U.K.: Cambridge University Press.

Weiss, H. (1995). Zaire: Collapsed Society, Surviving State, Future Policy. In Collapsed States. Edited by I. W. Zartman. Boulder, CO: Lynne Rienner.

Wilks, I. (1975). Asante in the Nineteenth Century. Cambridge, U.K.: Cambridge University Press.

World Bank. (1981). Accelerated Development in Sub-Saharan Africa: An Agenda for Action. Washington, DC: The World Bank.

———. (1989). Sub-Saharan Africa: From Crisis to Sustainable Growth. Wash- ington, DC: The World Bank.

———. (1991). Governance and Development. Washington, DC: The World Bank.

———. (2006). World Development Indicators 2006. Washington, DC: The World Bank.

Wrigley, C. (1996). Kingship and the State, The Buganda Dynasty. New York and London: Cambridge University Press.

Zakaria, F. (1997). “The Rise of Illiberal Democracy.” Foreign Affairs 76(November–December): 22–43.

186

https://doi.org/10.1017/CBO9781316423974.009 to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. Downloaded from https://www.cambridge.org/core. Access paid by the UC San Diego Library, on 07 Jan 2020 at 19:42:32, subject