"US Foreign Policy Regarding the Spread of Diseases in Nigeria"
Rinchin Lama
Empirical Analysis of Interest Rate Spread in Nigeria Muhammad Auwalu Haruna, Ahmadu Bello University, Kaduna, Nigeria
Abstract: Recent researches in Sub-Saharan Africa have indicated that interest rate spreads have re- mained high and are even increasing in most countries in spite of financial liberalization. This study, using alternative definitions of interest rate spread and time series dataset, empirically investigated the factors that influence the level of interest rate spread in Nigeria by means of regression analysis. The analysis showed that two determinants: bank-specific operating expenses and the industry-specific need for consistent growth of shareholders’ net worth explain the high spread in interest rates. Another finding is ISw3 (a definition of interest rate spread which includes fees and commissions) best represents intermediation cost. The findings suggest that the banking industry is not efficient or competitive, and as long as the banks can make profits without being efficient, the observed disconnection between the growth of the financial sector and that of real sector will undermine long term growth of the economy. It is, therefore, recommended that alternative sources of financial intermediation of non-bank type including capital market should be developed to support the real sector.
Keywords: Post Liberalization, Interest Rate Spread, Banking Efficiency, JEL Classification, E43, G14, G21
Introduction
FINANCIAL LIBERALIZATION IS expectedto improve the efficiencyof financialsystems in developing nations through reduced cost of intermediation from freeinginterest rates. Interest rate spread (hereafter, the spread), therefore, became increasingly the focus of research and policy attention in developing countries. However, studies
in Latin America, the Caribbean and Africa show that this expectation is not met (see for instance Barajas, Steiner, and Salazar, 1999; Randall, 1998; Ngugi, 2001; Chirwa and Mlachila, 2004 and Hesse, 2007). Such environments should be of particular concern for developing and transition countries where financial systems are largely bank-based, with limited alternative avenues for financial intermediation. The lack of convergence of interest rate spreads in developing countries toward those
observed in developed countries after financial liberalization may be connected to the rigidity of banks and banking behaviors especially in terms of market power from unchanged oper- ating structures. Other sources of rigidity may include increased loan provisioning from in- creased high-risk assets’ investment in pursuit of larger market share; high non-financial (operating) expenses; and effects of macroeconomic instability or the policy environment. Hesse (2007) argued that even the boom (before the bust) experienced by the Nigerian
banking sector after liberalization was largely due to proliferation of new banks driven by attractive arbitrage opportunities in the foreign exchange market. Banks seem only willing to be increasing lending rates while holding the deposit rates low in order to maintain high spread. He pointed out that on comparative basis with pre-liberalization period, financial intermediation did not improve but declined after financial liberalization began in 1986.
The International Journal of Interdisciplinary Social Sciences Volume 6, Issue 11, 2012, http://www.SocialSciences-Journal.com, ISSN 1833-1882 © Common Ground, Muhammad Auwalu Haruna, All Rights Reserved, Permissions: [email protected]
Persistence of high spread can be a major constraint to economic development. It is thus important that Nigerian policy makers succeed in reversing the trend in order to improve the effectiveness and efficiencies in financial intermediation. To choose the right policies, they need evidence about the key determinants of the spread. The main objective of this study therefore, is to evaluate the determinants of the spread in Nigeria. For analytical pur- poses, the determinants are classified into three broad groups: bank-specific, market/industry- specific and macroeconomic variables. The major contributions from the work are two. Firstly, the spread is defined with a structural consideration that suits the peculiarities of the Nigerian banking practice where increased intermediation costs are hidden through fees and commissions. Secondly, the empirical specification involved decomposing the selected banks’ audited financial statements in generating the ex-post spreads used. Ex-post rate being historical should offer more information than a theoretical ex-ante. The rest of the research is organized as follows: section 2 provides a brief overview of
the relevant literature including context analysis of the Nigerian banking industry; the method of analysis is considered in section 3; section 4 presents and discusses the results, and section 5 contains conclusion and recommendations.
Literature Review This section briefly discussed theoretical and empirical propositions regarding interest rate and intermediation costs and concepts of interest rate spreads. Finally it presents the post- liberalization developments in the Nigerian banking system.
Financial Liberalization and Bank Interest Rates Banks as financial intermediaries are very important as they play a key role in transforming deposits into financial assets by serving as the link between the deficit and surplus sectors of the economy. In so doing, they screen borrowers and monitor their activities in financial systems characterized by incomplete and asymmetric information. Of course this must be achieved at some cost to both the depositors and borrowers. Nonetheless, banks’ operating efficiency is quite crucial in ensuring the success of financial liberalization. So if they are repressed as proposed by McKinnon-Shaw’s hypothesis, the desired impact they should have in economic growth may be impaired. A sizable body of literature has indicated the importance of financial liberalization in fa-
cilitating economic development and growth. However, there is no complete agreement on the McKinnon-Shaw paradigm that the removal of financial repression through freeing interest rates and removal of credit ceilings/rationing increase the prospects of economic growth and development. Examples of the proponents of the hypothesis (as cited by Chirwa and Mlachila, 2004) are Agarwala, 1983; Khan and Senhadji, 2000; Khatkhate, 1988; King and Levine, 1993; and Levine, 1997. Whereas Taylor, 1983 and van Wijnbergen, 1983 have argued that high interest rates could be inimical to economic growth by reducing demand for bank credit. In spite of this divergence in the literature, the conventional view remains that absence
of financial repression can lead to higher growth by enhancing financial intermediation. Banking efficiency is often characterized by the level of interest rate spreads, the difference between lending and deposit rates. Financial systems in developing countries typically show significantly high and persistent spreads (Barajas et al., 1999; Chirwa and Mlachila, 2004;
2
THE INTERNATIONAL JOURNAL OF INTERDISCIPLINARY SOCIAL SCIENCES
and Hess, 2007). The expectation is that freeing interest rates and the barriers to entry into the financial system would lead to greater competition and lower profit margins of financial institutions, captured through low interest spreads. Another point of divergence is what constitutes the spread. In the next sub-section, some
conceptual definitions are considered with a case made for chosen definitions for this research work.
Interest Rate Spread: Conceptual Definitions Conceptually, interest margin is different from spread in bank performance analysis. Net interest margin (NIM) is the strict difference between the lending and deposit rates:
NIM = Lr - Dr (1) Where Lr = lending rate; Dr = deposit rate.
However, payments for services in the intermediation process like loan screening and mon- itoring, savings processing and management, payment services; and information asymmetry are other relevant costs between the interest rate paid to savers and the interest rate charged to borrowers. Adding these costs as a wedge expressed as ∑n i=1(Ci) to the interest margin, we arrive at the interest rate spread:
IS = Lr - Dr + ∑ n i=1(Ci) (2)
Where IS = the spread, Ci = i th cost of services in the intermediation process, n = total
number of relevant costs. As such the larger the banking inefficiencies as measured by ∑n
i=1(Ci), the higher the spread will be; and the higher will both be the fall in demand for and the benefits of financial intermediation. From the perspective of dealership model risk consideration, equation 2 is expressed dif-
ferently as banks are viewed as risk-averse in both loan and deposit markets. The spread is captured as fees charged for intermediation service on both deposit mobilization and lending:
P L = P + α P D = P – β (3)
where P is the bank’s opinion of the price of loan or deposit, and (α) and (β) are respective charges for provision of intermediation services. From (3) the spread is defined as:
IS = (α + β) (4)
This means ∑n i=1(Ci) = (α + β); and ∑ n i=1(Ci) can therefore be decompose into α and β.
To further measure the true spread as cost of intermediation, one-off and/or revolving fees and commissions are included in some models. Adding these fees and commissions (denoted as fj) to equation 4, we have:
IS = (α + β) + ∑m j =1(f j) (5)
3
MUHAMMAD AUWALU HARUNA
Inclusion of fees and commissions gives the actual full cost to customers in a lending situation, especially in inefficient markets like Nigeria where banks establish processes to circumvent interest rates control. In this regard, Brock and Rojas-Suarez’s (2000) narrow and broad definitions of the spread are represented by equations 4 and 5 respectively. The practice in Nigeria is characterized by equation 5 where the real costs of intermediation
are embedded in revolving fees and commissions to achieve two things. Firstly, pay less to depositors by showing commensurate low lending rate. Secondly, due to information asymmetry, the full cost of lending is screened from the regulatory authorities; hence low cost of borrowing statistics will continue to be reported. Furthermore, according to Ho and Saunders (1981) the spread is defined by market micro-
structure of the banking sector itself and the macro policy environment within which the banks operate. They differentiate between a pure spread and an actual spread. The pure spread which is captured only under perfect condition, is a microstructure phenomenon, and is a function of the bank’s risk management, the size of its transactions, interest rate elasticity and interest rate variability. Actual spread incorporates the pure spread and macroeconomic variables like monetary and fiscal policy activities. Thus conceptually, the spread as defined by equation 5 is most adequate for this work since it can be specified with bank-specific, market-specific and macroeconomic variables.
Studies on Determinants of Interest Rate Spread The theories of the determinants of commercial banks’ interest rate spreads in the literature are classified into three broad categories: bank-specific, industry (market) specific or mac- roeconomic in nature. Bank-specific characteristics usually include the size of the bank, ownership pattern, loan portfolio quality, capital adequacy, overhead costs, operating ex- penses, and shares of liquid and fixed assets (Brock and Rojas-Suarez, 2000; Crowley, 2007; Demirguc-Kunt and Huizinga, 1999; Folawewo and Tennant, 2008; Gelos, 2006; Moore and Graigwell, 2000; Ngugi, 2001; Robinson, 2002; and Sologoub, 2006). The market-specific determinants include level of competition or market power, degree
of development of the banking sector, taxes and reserve requirements (Fry, 1995 and Elkayam, 1996). Cho (1988) observed that liberalization theory overlooks endogenous constraints like absence of functioning equity market which are critical to efficient allocation of resources by the banking sector. This impact is very obvious in Nigeria where Banks exhibit market power in both deposit and lending markets. Fry (1995) explained that absence of direct fin- ancial markets like the equity and bonds market leads to over reliance on debt finance; this over exposes the financial institutions thereby forcing them to absorb too much risk. Macroeconomic variables include inflation, growth of output, exchange rates and money
market real interest rates. The macroeconomic environment affects the performance of the banking sector to the extent of its influence on the ability of borrowers to timely honor the debt repayment obligation. An unstable macroeconomic environment exhibits a positive correlation between the lending rate and the nonperforming loan portfolio. Cukierman and Hercowitz (1990) attempt to explain the relationship between anticipated inflation and the degree of market power measured as the spread between the deposit and lending rates. They find that when the number of banking firms is oligopolistic, an increase in anticipated inflation leads to an increase in interest rate spread.
4
THE INTERNATIONAL JOURNAL OF INTERDISCIPLINARY SOCIAL SCIENCES
The 3 broad classifications are employed in this work. The a priori expectations of both signs and magnitudes are detailed on table 2.
Table 2: Definition of Determinants
Significance/A Priori ExpectationsDefinitionsVariablesClassifications Requires more spread to cover. It is expected to have direct effect on Spread.
Non-interest Exp/ Total Earning Assets
Operating Expenses(OE)
Firm-Specific Banks would tend to push this cost to customers. In ex-post analysis,
Provision for bad debt/Total loans & Advances
Loan Loss Provisions (LLP) LLP on the income statement de-
creases spread. Hence inverse rela- tionship is anticipated. Active intermediation indicates high IMED. Competitive environ-
Total Loans/Total Deposit Liabilities
Financial Intermediation (IMED)
Market-Specific ment decreases spread; hence an inverse relationship. Requires more spread to accumu- late. It is expected to have a posit- ive relationship with Spread.
Shareholders’ Funds/ Total Assets
Shareholders’ Networth (SHN)
Proxied by its annual average rate of growth/depreciation. It is expec- ted to have direct effect on Spread.
[(fxr) t–(fxr) t-1]/(fxr)t-1 where (fxr) = periodic exchange rate and t-1= annual time-lag.
ExchangeRate Depreciation (ERD)
Macroeconomic Proxy for marginal cost of funds; a bench mark for interest rate de-
Average Annual Treasury Bill rates
Treasury Bill (TRB)
cisions by banks. As a cost indicat- or, it should generate a positive re- lationship with spread. This is to capture business cycle effects. Inflation can also affect
[(CPI) t –(CPI) t- 1]/(CPI)t-1 where t-1= annual time-lag.
Annual Inflation Rate (IFL) spread if monetary shocks are not
passed wholly to deposits and lending rates, or adjustment occurs at different speed and time.
Interest Rate Management in Nigeria According to Soyibo (1997), the pre-Structural Adjustment Program (SAP) banking devel- opment in Nigeria can be captured in three phases namely; 1892–1952: the era of laissez faire banking, 1952–1958: the era of limited banking regulations and 1959–1985: the period of prudential Regulations and Control. The most significant development for this study is post enactment of the banking act of 1969 when interest rate commenced as a policy instru-
5
MUHAMMAD AUWALU HARUNA
ment by empowering the Central Bank of Nigeria (CBN) to regulate all interest rates through linkage to minimum rediscount rate (MRR). Though the period 1959–1985 coincided with the global trend of credit rationing and rates
control (Soyibo, 1997), a number of problems arose from the banking regulations in the early 1980s that questioned the entire mode and implementation of banking regulation. The worsening financial crisis necessitated the market deregulation exercise of 1985–1993. The reforms led to dramatic changes in the banking environment as evidenced by changes in operating structures in terms of both the industry and the macroeconomic environment (Nnanna, 2004). For instance, the number of commercial and merchant banks increased sharply from 41 in 1986 to 120 at the end of 1993 when the adjustment was completed. The pre-liberalization ownership structure also changed in favor of private ownership with foreign interest playing only a supporting role. However, the fact that the Nigerian financial system remained fragmented with increased
financial disintermediation after liberalization (Hesse, 2007 and Lewis and Stein, 1997), brings to question possible changes in both the bank market and internal characteristics. For instance many of the new banks were not interested in sourcing funds from depositors for real sector lending, but rather to make quick profits from the arbitrage and other rent-seeking activities often not legal. The regulatory authorities packaging the deregulation exercise also seem to have missed the banking internal structural misalignment by mainly focusing only on the need to increase economies of scale as assurance for a more stable banking system, and improved financial intermediation. Therefore, the need to study the internal characteristics of banks becomes more obvious since in spite of all the institutional and structural changes, interest rate spread remains high.
Methodology and Data Average deposit and lending rates published by the CBN are on ex-ante basis. However for meaningful post liberalization analysis, the spread was generated ex-post from the financial statements of the sampled banks. As a variant of Chirwa and Mlachila (2004), the study used time series analysis instead of panel data analysis. This is considered adequate because of the level of homogeneity of the Nigerian banking firms. Similarly we differ from their definition and composition of some spreads as highlighted below.
Narrow Definitions 1. ISn1 = NIM = (Interest received – Interest paid)/Loans 2. ISn2 = (Interest received/loans) − (Interest paid/deposits); unlike Chirwa and Mlachila (2004) that defined ISn2= (interest received on loans only/loans) − (interest paid on deposits only/deposits). In Nigeria the core of interest paid and received are loan related. As such it is more specific as a measure of loan cost since there may be no significant difference in the two approaches in Nigeria. 3. ISn3= (interest plus commission received/loans) − (interest plus commission paid/deposits);
6
THE INTERNATIONAL JOURNAL OF INTERDISCIPLINARY SOCIAL SCIENCES
Broad Definitions With these definitions we are considering loan specific basis using earning assets and interest bearing liabilities in place of total assets and total liabilities, respectively. 4. ISw1 = (interest received − interest paid)/total earning assets; 5. ISw2 = (interest received/total earning assets) − (interest paid/ interest bearing liabilities); 6. ISw3= (interest plus commission received/total earning assets) − (interest plus commission paid/ interest bearing liabilities); 7. ISBM = average prime lending rate-average deposit rate. The seventh is a bench-mark spread that is directly calculated from the published average deposits and savings rates against both prime and maximum lending rates.
Table 1: All Interest Rate Spreads
B-MARKMAX.AVER. ISw3ISw2ISw1ISn3ISn2NIMSPREADRATERATE 6.64%-2.86%2.99%28.25%14.18%1.78%2.40%12.00%9.60%1986 6.34%-3.30%3.84%31.95%16.17%3.03%4.18%19.20%15.02%1987 6.08%-3.19%3.68%30.63%15.50%2.90%4.00%17.60%13.60%1988 7.35%-3.87%4.46%37.13%18.79%3.52%4.70%24.60%19.90%1989 8.84%-7.27%8.38%69.80%35.33%3.05%6.90%27.70%20.80%1990 9.13%-4.80%5.53%46.05%29.91%2.01%3.34%20.80%17.46%1991 7.35%-3.86%4.44%37.03%24.05%8.84%10.84%31.20%20.36%1992 12.31%-6.47%7.45%62.08%40.32%7.74%15.21%38.32%23.11%1993 7.76%-4.08%4.70%39.14%25.42%4.88%6.30%21.00%14.70%1994 9.01%-4.73%5.45%45.40%29.48%5.66%7.17%20.79%13.62%1995 10.02%-5.26%6.06%50.53%32.82%6.30%7.90%20.86%12.96%1996 10.04%-5.41%6.25%40.51%25.82%6.39%16.48%23.32%6.84%1997 11.42%-6.57%6.81%45.64%29.95%7.37%12.09%21.34%9.25%1998 9.46%-4.23%4.67%40.98%23.69%5.11%18.22%29.70%11.48%1999 11.91%-7.14%7.85%66.35%46.89%8.47%12.10%21.55%9.45%2000 11.29%-6.48%6.87%63.32%43.11%7.54%12.09%21.34%9.25%2001 12.07%-7.40%7.84%59.35%41.44%8.60%15.48%29.70%14.22%2002 11.65%-6.88%7.23%57.51%39.04%8.01%11.79%22.47%10.68%2003 11.26%-6.45%6.61%44.31%28.63%7.24%10.12%20.62%10.51%2004 11.49%-5.78%5.67%37.59%21.94%6.58%10.80%19.47%8.67%2005
SOURCE: Compiled from the banks’ financial statements and CBN annual reports.
7
MUHAMMAD AUWALU HARUNA
Looking at table2, a very interesting indicator of banking real cost in manifest in ISw2 where it is all throughout negative due to absence of fee incomes. Periodic gaps between ISw2 and ISw3 are between 12 and 17%. This situation is a good indicator of hiding costs through fees.
Population and Sampling At the time bank consolidation commenced in 2005, there were 24 banks in Nigeria classified as either “New Generation” (NGB) or “Old Generation” (OGB) based on their age and level of efficiency. Perception of efficiency levels between the OGB and the NGB are different. As such in order to avoid sample concentration or bias, 13 sample points taken were stratified into 6 NGB1 and 6 OGB2 with FSB International (the thirteenth) as their hybrid3.
The Model Specification Most models of the determinants of bank interest rate spreads are often based on the frame- work of a bank as a profit- or wealth-maximizing firm; that is seeking to maximize profits defined by a feasible set of assets and liabilities whose per unit prices and costs are set by the bank. This approach views banks as risk-adverse dealers in both the loan and deposit markets where loan requests and deposit generation are at random and unsynchronized. Thus by incorporating various aspects of the competitive process and scale economies, these models provide the basis for the empirical testing of the spread in a manner consistent with the Structure Conduct Performance (S-C-P) and efficient market hypotheses. In this regard, the econometric model to be estimated is:
Y = βX + U (6)
Where Y is vector of the spreads as defined, β is a vector of parameters, U is a stochastic error term, and X is a vector of the explanatory variables as detailed on table 2. Unlike Enendu (2003) who analyzed ex-ante commercial bank spreads in Nigeria, this
study looked at ex-post spread which is likely to be more relevant given the incongruity between the state of the Nigerian real sectors and the independent growth of the banking sector. Following Beck and Fuchs (2004) and Hesse (2007), an accounting decomposition of the spread was conducted first to generate the ex-post spreads before the econometric analysis. Many researchers decomposed the spread in straight nominal terms; however to avoid heteroscedasticity, ratios instead of the Naira value for the bank balance sheet accounts are used in estimation of the model. In the model, it is hypothesized that the spread is a function of the three (3) broad classifications of the determinants tabulated in table 2.
1 Access Bank, Diamond Bank, GTBank, Zenith Bank, Intercontinental and Oceanic Bank 2 First Bank, Union Bank, UBA, Afribank, WEMA, and Inland Bank (now First Inland). 3 Federal Savings Bank was an old establishment and a fringe player that assumed a full modern commercial bank role after liberalization.
8
THE INTERNATIONAL JOURNAL OF INTERDISCIPLINARY SOCIAL SCIENCES
Empirical Results This section starts with trend analysis before the main regression analysis. Table 1 gives the various interest rate spreads defined and as decomposed on the ex-post narrow and wide measures. The accompanying determinants ratios are contained on table 3.
Table 3: Determinants of Interest Rate Spread
IFLTBRERDSHNIMEDOELLP 5.40%8.50%48.10%18.43%26.44%4.08%15.90%1986 10.15%11.75%71.68%18.25%24.92%4.18%15.09%1987 56.07%11.75%11.56%17.50%23.89%4.01%14.47%1988 50.50%17.50%31.54%16.02%21.72%4.86%17.53%1989 7.52%17.50%13.48%21.25%22.04%7.26%32.97%1990 12.86%15.00%18.18%21.94%26.42%5.45%21.75%1991 44.58%21.00%41.27%17.64%29.29%4.77%17.49%1992 57.17%26.90%18.92%26.22%33.22%4.65%29.32%1993 57.03%12.50%3.33%16.53%20.95%6.67%18.49%1994 72.81%12.50%75.86%19.17%23.14%7.74%21.44%1995 29.19%12.25%-14.29%21.34%25.76%8.62%23.87%1996 8.25%12.00%0.00%25.09%30.79%5.65%17.47%1997 10.16%12.95%0.00%22.09%29.17%6.84%17.31%1998 6.75%17.00%75.00%21.89%28.91%7.28%20.62%1999 6.79%12.00%0.00%22.47%23.74%9.16%23.11%2000 18.79%12.95%40.00%20.78%23.23%7.26%18.39%2001 14.84%18.88%-9.48%19.75%26.14%7.38%24.71%2002 12.29%15.02%-12.49%18.84%25.65%7.38%20.45%2003 15.00%14.21%-7.00%19.46%31.12%7.11%16.98%2004 17.80%7.00%0.56%22.23%36.42%6.67%10.21%2005
SOURCE: Compiled from the banks’ financial statements and CBN annual reports.
Trend Analysis of the Ex-post Spread Looking at figures 1 and 2, there is a widening gap between the narrow and wide definitions of interest rate spread (i.e. Sn 3-Sn 2 and Sw 3-Sw 2). This is very important in understanding the structure of banking charges as experienced by the Nigerian banking system where in- creases in real lending rates are hidden under fees. For banks to sustain high lending charges, also suggests dearth of non-bank sources of financing. Further evidence is from the nature
9
MUHAMMAD AUWALU HARUNA
of Sw2, whose scatter points are negative indicating loss in financial intermediation when fees are not captured as part of the spread.
In spite of the fact that both the implicit lending rates and the average deposit rates have slightly come down over the years, overall the NIM is above the pre-liberalization period, indicating a fall in efficiency level.
10
THE INTERNATIONAL JOURNAL OF INTERDISCIPLINARY SOCIAL SCIENCES
Figure 3 gives trends of the determinants of the spread within the liberalization period.
The market micro structure in Nigeria is likely to offer more explanation for spreads’ stick- iness because the banking system is booming when the real sector (reflecting the interplay of macroeconomic tools) is in the best state static, showing little or no growth.
The Regression Analysis Empirical studies on time series data assumes that the data is stationary. However to avoid autocorrelation and a spurious results, stationarity testing is important. In examining the stochastic characteristic of each time series, a stationarity test using Augmented Dickey Fuller (ADF) test was carried out. All variables including the dependent variables are first difference stationary. As such the result on table 4 were arrived at after the first difference. From the different definitions of the spread enumerated above, ISw3 is identified to be
the most significant in establishing the causality with the chosen determinant variables4. Among the seven (7) variables5 on a bivariate analysis, two (2) macro variables (IFL and
TRB) and one (1) micro firm-level variable (LLP) appear statistically not significant in de- termining the level of interest rate spread within the period the data set is analyzed. In addition to failing both the t and F tests, they appear with less than 16% R2, an indication of poor representation. In addition to all these, IFL appear with a wrong sign. The other variables OE, SHN, ERD and IMED are statistically significant. SHN has the
strongest showing in terms of both statistical significance and explanatory powers followed by OE, ERD and IMED in that order. In terms of power to explain the variations of the re-
4 The results were achieved by a process of experimentation; that is the first stages in testing were running bivariate regressions to trace an indication of singly-significance, before inclusion of more variables to observe differentiated impacts. 5 OE=operating expenses LLP=loan loss provision IFL=inflation index IMED=intermediation index SHN=share- holders net worth TRB=treasury bills ERD=exchange rate depreciation index
11
MUHAMMAD AUWALU HARUNA
gressand, SHN alone accounts for 41%. None of the other variables however would account for less than 21%. This appears consistent with the expectations of this research. It is anticip- ated that the owners’ quest to grow their networth would be the main reason why interest rate spread in an imperfect market like Nigeria remains high.
Table 4: Regression Results
DW St.
F- stat.
A R- sq
R- sqd
CIFL (1)
TBR (1)
ERD (1)
SHN (1)
IMED (1)
OE (1)
LLP (1)
Variable
1.898.520.750.84(4.45)(0.00)0.29(0.02)0.41(0.01)1.02(0.22) Coeffi- cient
SW3
2.720.010.120.010.180.120.230.12 Std. Error
(1.64)(0.29)2.46(2.13)2.29(0.05)4.36(1.79) t-Statis- tic
1.666.750.690.81(0.11)(0.07)1.02(0.07)1.65(1.04)3.360.40 Coeffi- cient
SN3
17.780.080.770.061.160.791.530.80 Std. Error
(0.01)(0.87)1.32(1.17)1.42(1.32)2.200.50 t-Statis- tic
2.205.030.610.760.720.01(0.09)0.02(0.09)(0.01)(0.44)(0.02) Coeffi- cient
SW2
2.340.010.100.010.150.100.200.11 Std. Error
0.311.31(0.90)2.13(0.62)(0.09)(2.20)(0.22) t-Statis- tic
1.406.130.670.80(13.46)(0.02)1.54(0.08)2.09(1.24)4.26(0.73) Coeffi- cient
SN2
14.170.070.620.050.930.631.220.64 Std. Error
(0.95)(0.29)2.50(1.70)2.26(1.98)3.50(1.14) t-Statis- tic
2.326.800.690.81(0.43)(0.01)0.08(0.02)0.14(0.02)0.360.07 Coeffi- cient
SW1
2.120.010.090.010.140.090.180.10 Std. Error
(0.20)(1.18)0.83(2.30)1.01(0.23)1.950.69 t-Statis- tic
12
THE INTERNATIONAL JOURNAL OF INTERDISCIPLINARY SOCIAL SCIENCES
2.334.520.580.74(6.96)0.010.58(0.02)0.28(0.04)1.46(0.48) Coeffi- cient
NIM
3.780.020.160.010.250.170.320.17 Std. Error
(1.84)0.533.51(1.49)1.14(0.26)4.50(2.79) t-Statis- tic
Source: Author’s Computation using E-views 7.1 package
This is supported by the market power exhibited through a strong influence of IMED in the regression equation. To reinforce this central interest in growing self-interested private wealth, the speculated round tripping of foreign exchange appears significant. Furthermore, the banking system is important in analyzing the concept of principal-agent relationship as most of the banks are managed by agents. As such it is compensating for the Management to keep OE high consistent with growing shareholders’ wealth in order to maintain a good standard of living. The asymmetry between the booming banking industry and the major growth indicators of the real sector, buttresses the failure of the core macro variables TBR and IFL to appear with significant influence. LLP’s insignificant influence is another con- firmation of two issues; first findings of Lewis and Stein, 1997; Hesse, 2007 on financial disintermediation; and second the banks’ market power in both loan and deposit markets. Implicit rates appear to have factored in expected loan provisions. This will generate very little problem loans; hence concern about loan provision is secondary. The influence of these variables is further investigated by taking some levels of multivariate
analyses before testing the main model and its counterparts. ‘Pulling-effect’ was anticipated in terms of significance and explanatory power, as we increase multiple combinations. As such a set of forty-two (42) in two (2) regressors were tested. In all combinations, again OE, ERD and SHN appear significant with acceptable R2. Only IMED among the bivariate sig- nificant variables appear insignificant in combination with ERD. TRB improved in its ex- planatory power when combined with IMED, SHN and IFL by appearing significant; while IFL becomes significant only when combined with LLP. LLP’s significance appears only in combination with IMED. Of interest, no single group of variables- market, industry or macroeconomic, is significant
on standalone basis. In fact, even the F test of the macroeconomic model does not justify TRB, ERD and IFL combination. It is only the R2 that register improvement on the single ERD; whereas the R2 for the micro models are woeful. To observe variation improvement, a build up towards the complete model was done by
introducing one variable at a time and its entry selection is based on its strength in the bivariate regressions. Thus, we first take the three (3) most significant variables; that is SHN, OE and ERD. All test statistics improved with R2 reaching 70%. This integrates the main drivers of sticky interest rate spread to be growth of shareholders’ networth and the welfare of the staff, especially the management cadre. When IMED was integrated, together with ERD they become statistically insignificant though R2 improved. Adding TRB further en- hanced the R2 to about 80%, however only OE and IMED appear as significant variables. Including the sixth variable (IFL) marginally improved the R2 with no additional variable
that is statistically significant. The F statistic also deteriorated from 10.27 to 7.96 as further
13
MUHAMMAD AUWALU HARUNA
evidence of the weak statistical significance of (IFL) in the model. The complete model presents an R2 of 84% as the best fit with only IMED and IFL being insignificant. ERD and LLP also appear with a wrong sign. The summary of the results of the complete models of the other definitions are as follows:
ISBM weakens the explanatory power of the variables by further indicating ERD as statistically not significant. ISn1 further adds SHN as insignificant. ERD, IFL and LLP are insignificant with ISn2, however, with a better R
2. The only variable that is significant with ISn2 is OE; though the R
2 is the second best at 81% after the 84% R2 for ISw3 regression. ISw1 and ISw2 have both OE & ERD as the only significant variables with R2 at 81% and 76% respectively.
Conclusion and Recommendations
Conclusion The study empirically investigated the factors that determine the level of interest rate spread in Nigeria. The bank-specific character of operating expenses and the industry’s shareholders’ networth provide most explanation for the high interest rate spread. This is consistent with the independent nature of the high banking financial performance when the real sectors of the economy are disinvesting. Empirical evidence from this research points to the fact that commercial banks in Nigeria
exhibit market power in both deposit and loan markets that enable them to shift most risks and costs (inefficiency) to customers. To keep the spread wide, outside market rate charges are disguised as other fees and remain undetectable by the regulatory authorities. This has been the source of explanatory power of the ISw3 definition of the spread. In this regard even if the ex-ante spread is seen as converging, the effective interest rate spread from the ex-post view point will remain high. This means financial liberalization has not benefited the depositors and borrowers in Nigeria just like other areas studied within sub-Saharan Africa.
Recommendations The results show existence of a wide spread as if the market structure has not changed as the Nigerian banking system appears not to be supportive of the real sectors of the economy. This is attributable to the main factors causal to widening interest rate spread in Nigeria; i.e. bank internal characteristic of operating expenses, and market/industry characteristic of spiral growth of shareholders networth. Furthermore, whenever the main control variable is internal to the banks under supervision, macroeconomic stability becomes impossible. It is therefore recommended that firstly, deliberate efforts must be made to expand altern-
ative sources of financial intermediation of non-bank type. Small scale businesses who generally do not qualify for commercial banks financing are the growth area of the real sectors of the economy. As such Nigeria’s economic growth could be stable and sustained if this group are adequately financed and at the most optimum costs. Consolidation exercise aimed at supporting the big corporations seem at odds with this basic need.
14
THE INTERNATIONAL JOURNAL OF INTERDISCIPLINARY SOCIAL SCIENCES
Secondly, the development of the capital market is vital to enhance competitiveness. Second-tier products should be developed to avail the small players access long term finan- cing. Further efforts of fiscal and monetary policy actions should development of enhance the financial market as fiscal discipline is identified as a prerequisite for successful financial liberalization.
15
MUHAMMAD AUWALU HARUNA
References Barajas, A., Steiner, R., & Salazar, N. (1999, June). Interest Spreads in Banking in Columbia, 1974–96.
IMF Staff Papers 46(2). International Monetary Fund. Beck, T. A., & Fuchs, M. (2004). Structural Issues in the Kenyan Financial System: Improving Com-
petition and Access. World Bank Policy Research Working Paper 33–63. Brock, P. L., & Rojas-Suarez. (2000). Understanding the Behaviour of Bank Spreads in Latin America.
Development Economics 63(1), 113–34. Caprio, G., & Klingebiel, D. (2003). Episodes of Systemic and Borderline Financial Crises. World
Bank. Chirwa, E. W., & Mlachila, M. (2004). Financial Reforms and Interest Rate Spreads in the Commercial
Banking System in Malawi. IMF Staff Papers. Cho, Y. J. (1988). The Effect of Financial Liberalization on the Efficiency of Credit Allocation: Some
Evidence from South Korea. Journal of Development Economics 29, 101–110. Crowley, J. (2007). Interest Rate Spread in English-Speaking African Countries. IMF WP/07/101. In-
ternal Monetary Fund. Cukierman, A., & Hercowitz, Z. (1990). Oligopolistic Financial Intermediation, Inflation and Interest
Rate Spread. (2). The David Horowitz Institute for the Research of Developing Countries, Tel Aviv University.
Demirguc-Kunt, A., & Huizinga, H. (1999). Determinants of Commercial Bank Interest Margins and Profitability: Some Internationa Evidence. World Bank Economic Review 13(2) 379–408.
Elkayem, D. (1996). The Effect of Monetary Policy on Local Currency Segment in Israel 1986–1900. Bank of Israel Economic Review 68, 1–22.
Enendu, C. I. (2003). Determinants of Commercial Bank Interest Rate Spread in a Liberalized Financial System: Empirical Evidence from Nigeria (1989–2000). Economic and Financial Review 41(1).
Folawewo, A. O., & Tennant, D. (2008, July 9–11). Determinants of Interest Rate Spreads in Sub- Saharan African Countries: A Dynamic Panel Analysis. 13 th Annual African Econometrics Society Conference. Pretoria, Republic of South Africa.
Fry, M. J. (1995). Money, Interest and Banking in Economic Development (2nd Edition ed.). Baltimo and London: John Hopkins University Press.
Gelos, G. (2006). Banking Spreads in Latin America. IMF Working Paper WP/06/44. Washington: Internal Monetary Fund.
Hesse, H. (2007, June). Financial Intermediation in the Pre-Consolidated Banking Sector in Nigeria. PolicyResearch Working Paper Series, 4267. World Bank.
Ho, T., & Saunders, A. (1981). The Determinants of Bank Interest Margins: Theory and Evidence. (581–600, Ed.) Journal of Financial and Quantitative Analysis 4.
Lewis, P., & Stein, H. (1997). Shifting Fortunes: The Political Economy of Financial Liberalization in Nigeria. Journal of World Development 25, 5–22.
Moore, W., & Graigwell, R. (2000, November 2). Market Power and Interest Rate Spreads in the Caribbean. Kingston, Jamaica: XXXII Annual Monetary Studies Conference.
Ngugi, R. W. (2001, May). An Empirical Analysis of Interest Rate Spread in Kenya. AERD Research Paper (106). African Economic Research Consortium Paper Series.
Nnanna, O. J. (2004). Financial Sector Development and Economic Growth in Nigeria: An Empirical Investigation. CBN Economic and Fianacial Review 42(3). Central Bank of Nigeria.
Randall, R. (1998). Interest rate Spreads in the Eastern Caribbean. IMF Working Paper Series. Wash- ington: International Monetary Fund.
Robinson, J. W. (2002). Commercial Bank Interest Rate Spreads in Jamaica: Measurement, Trend and Prospects. RetrievedApril2010, fromwww.boj.org.jm/uploads/pdf/papers_pamphlets.pdf
Sologoub, D. (2006). The Determinants of bank Interest Margins and Profitability: Case of Ukraine. Retrieved April 2010, from www.bof.fi/bofit/seminar/bofcef06/sologub.pdf
16
THE INTERNATIONAL JOURNAL OF INTERDISCIPLINARY SOCIAL SCIENCES
Soyibo, A. (1997). Financial Liberalisation and Bank Restructuring in Sub-Saharan Africa: Some Lessons for Sequencing and Policy Design. AERC Paper. Nairobi, Kenya.
About the Author Muhammad Auwalu Haruna Muhammad Auwalu Haruna holds a B.Sc. in Economics (ABU 1983), a M.Sc in Banking & Finance (UI 1990), and a M.Sc. in Economics (ABU 2008). Haruna was also awarded the Lagos Business School’s case study based Advance Diploma in Management, 1993. Presently, Haruna is a Doctoral Student in Economics at Ahmadu Bello University, Zaria. Haruna is also a retired banker, whose banking experience spans over a period of fifteen (15) years, with a stint both in Merchant Banking and Commercial Banking. Haruna has acquired ad- equate training and exposure in Credit Analysis & Marketing, Treasury Management (local currency), General Management and Banking Operations (Pay & Receive - Local). Haruna has participated in Strategic Planning, other visionary agenda and processes including business re-focus and turn-around drive in the late 90s.
17
MUHAMMAD AUWALU HARUNA
Copyright of International Journal of Interdisciplinary Social Sciences is the property of Common Ground Publishing and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.