Catherine Pattillo, Anne Gulde, Kevin Carey, Smita Wagh, and Jakob Christensen
INTERNATIONAL MONETARY FUND
Financial sectors in low-income sub-Saharan Africa (SSA) are among the world's least developed. In fact, assets in most low-income African countries are smaller than those held by a single medium-sized bank in an industrial country. The absence of deep, efficient financial markets seriously challenges policy making, hinders poverty alleviation, and constrains growth. This book argues that building efficient and sound financial sectors in SSA countries will improve Africa's economic prospects. Based on a review of the key features of financial systems, it discusses the main obstacles and challenges that financial structures pose for SSA economies and recommends steps that could address major shortcomings in implementing the reform agenda.
AdamChristopher and DavidBevan2004“Aid and the Supply Side: Public Investment, Export Performance and Dutch Disease in Low Income Countries,”Department of Economics Discussion Paper Series (Oxford: University of Oxford).
Adam, Christopher, and DavidBevan,2004, “Aid and the Supply Side: Public Investment, Export Performance and Dutch Disease in Low Income Countries,” (Oxford: University of Oxford).)| false
AleemIrfan and LouisKasekende2001“Reforming Finance in a Low-Income Country: Uganda” in Financial Liberalization: How Far How Fast?ed. byGerardCaprioPatrickHonohan and Joseph E.Stiglitz (New York: Cambridge University Press).
Aleem, Irfan, and LouisKasekende,2001, ed. byGerardCaprio,PatrickHonohan, and Joseph E.Stiglitz (New York: Cambridge University Press).)| false
AlvessonMagnus and AlfredoTorrez2003“Central Bank Operations and Macroeconomic Stabilization in Angola,”in Angola: Selected Issues and Statistical Appendix IMF Country Report No. 03/292 (Washington: International Monetary Fund).
Alvesson, Magnus, and AlfredoTorrez,2003, “Central Bank Operations and Macroeconomic Stabilization in Angola,” (Washington: International Monetary Fund).)| false
BossoneBiagioPatrickHonohan and MillardF. Long2002“Policy for Small Financial Systems” in Financial Sector Policy for Developing Countries: A Readered. byG.CaprioP.Honohan andD.Vittas (New York: Oxford University Press, for the World Bank).
Bossone, Biagio,PatrickHonohan, and MillardF. Long,2002, , ed. byG.Caprio,P.Honohan, andD.Vittas (New York: Oxford University Press, for the World Bank).)| false
BuffieEdwardChristopherAdamStephenO’Connell and CatherinePattillo2004“Exchange Rate Policy and the Management of Official and Private Capital Flows in Africa,”IMF Staff PapersVol. 51 (Special Issue) pp. 126-60.
Buffie, Edward,ChristopherAdam,StephenO’Connell, and CatherinePattillo,2004, “Exchange Rate Policy and the Management of Official and Private Capital Flows in Africa,”, Vol. 51 (Special Issue), pp. 126-60.)| false
CaprioGerard and PatrickHonohan2005“Starting over Safely: Rebuilding Banking Systems,”Chapter 7 inFinancial Crises: Lessons from the Past Preparation for the Futureed. byG.CaprioJ.Hanson and R.Litan (Washington: Brookings Institution).
Caprio, Gerard, and PatrickHonohan,2005, “Starting over Safely: Rebuilding Banking Systems,”Chapter 7 in, ed. byG.Caprio,J.Hanson, and R.Litan (Washington: Brookings Institution).)| false
GershensonDmitriy2004“Measures to Enhance Financial Intermediation in Lesotho,”in Lesotho: Selected Issues and Statistical Appendix IMF Country Report No. 04/23 (Washington: International Monetary Fund).
Gershenson, Dmitriy,2004, “Measures to Enhance Financial Intermediation in Lesotho,” (Washington: International Monetary Fund).)| false
International Monetary Fund1996“The Use of Reserve Requirements in Monetary Control: Operational Features and Country Practices” MAE Operational Paper 96/01 (WashingtonMonetary and Exchange Affairs Department).
International Monetary Fund, 1996, (Washington, Monetary and Exchange Affairs Department).)| false
International Monetary Fund2006Senegal: Third and Fourth Reviews Under the Three-Year Arrangement Under the Poverty Reduction and Growth Facility and Request for Waiver of Performance CriteriaIMF Country Report No. 06/127 (Washington).
International Monetary Fund, 2006, Senegal: Third and Fourth Reviews Under the Three-Year Arrangement Under the Poverty Reduction and Growth Facility and Request for Waiver of Performance Criteria, (Washington).)| false
Lafourcade, Anne-Lucie,JenniferIsern,PatriciaMwangi, and MatthewBrown,2005, (Washington: MIX Market). Available via the Internet: http:// www.mixmarket.org/medialibrary/mixmarket/Africa_Data_study.pdf.)| false
MehranH.P.UgoliniJ.-P.BriffautG.IdenT.LybekS.Swaray and P.Hayward1998Financial Sector Development in Sub-Saharan African CountriesIMF Occasional Paper No. 169 (Washington: International Monetary Fund).
Mehran, H.,P.Ugolini,J.-P.Briffaut,G.Iden,T.Lybek,S.Swaray, and P.Hayward,1998, Financial Sector Development in Sub-Saharan African Countries, (Washington: International Monetary Fund).)| false
MiccoAlejandroUgoPanizza and MonicaYanez2004“Bank Ownership and Performance,”IDB Research Department Working Paper No. 518 (Washington: Inter-American Development Bank). Available via the Internet: http://www.iadb.org/res/files/data_app_mpy.xls.
Micco, Alejandro,UgoPanizza, and MonicaYanez,2004, “Bank Ownership and Performance,” (Washington: Inter-American Development Bank). Available via the Internet: http://www.iadb.org/res/files/data_app_mpy.xls.)| false
NassarKoffie Ben2003“Recent Issues in the Implementation of Monetary Policy,”in Tanzania: Selected Issues and Statistical Appendix IMF Country Report No. 03/02 (Washington: International Monetary Fund).
Nassar, Koffie Ben,2003, “Recent Issues in the Implementation of Monetary Policy,” (Washington: International Monetary Fund).)| false
1Except for South Africa, Nigeria, and Kenya, the combined financial size of the the remaining countries is below $6 billion.
2This section draws on a wide variety of data sources, including recent country-level Financial Sector Profiles prepared by the IMF’s African Department. Country-level studies of financial markets conducted in the IMF’s African Department were also used (Uganda: Peiris, 2005; Botswana: Kim, 2004; Lesotho: Gershenson, 2004; and Kenya: Powell, 2003).
3Table 1 uses Bankscope country-level aggregate data, while the other banking sector tables are calculated from Bankscope bank-level data. The primary data source for Bankscope is the Fitch ratings database, which rates banks as actual or prospective borrowers from capital markets. Bankscope covers a sample of banks in each country. For SSA, 34 countries are covered, and within these countries, 381 banks compared with the 453 that were counted in the IMF African Department’s Financial Sector Profiles (2005). Because the unit of observation in the bank-level data is the bank, bank-level data are in effect weighted toward countries with more banks.
4Data from the IMF’s Financial Sector Assessment Programs (FSAPs) in SSA indicate that life insurance penetration indicators, measured by premiums/GDP, are very low in SSA—at 1—2 percent of GDP (except in South Africa). This is based on data from 1997 to 2002.
5Given differences in definition, NPLs may not be fully comparable across countries. The definition of NPLs is more stringent than “problem loans” (strictly based on the timing of overdue payments) reported in Table 2. These also take expected ability to repay into account.
6Trends in NPLs can be difficult to interpret, since a rising NPL trend may reflect better reporting mechanisms or tighter supervisory requirements. Large client exposure can make NPLs volatile from year to year. In addition, provisioning mechanisms may differ across countries, and the underlying collateral (if any) for NPLs will be a major determinant of their final impact on bank balance sheets.
7Based on Slack (2003), who surveys collection and dissemination of financial soundness indicators in 100 countries.
8Banking entry or activity restrictions cannot fully account for highly concentrated banking systems in SSA. While there is no clear difference between SSA and comparator groups on most of these restriction measures, the share of entry applications denied is somewhat higher in SSA (Barth, Caprio, and Levine, 2006).
9The interest margin measures the difference between interest earned on assets and interest paid on liabilities. More efficient banking systems will be able to have lower interest margins. Banks with high operating costs must earn high interest income to cover these costs, which is why it is used as a measure of inefficiency. However, the margin could also be high because of monopolistic profits or low because of risk aversion or interest rate controls, so it is not solely a measure of sectoral efficiency.
10Other countries also are considering regional solutions—given economies of scale and cost sharing—but regulatory and supervisory problems outside of monetary unions are more difficult to resolve.
11Public credit registries and private credit bureaus have low coverage across all groups, though coverage seems to be rising more rapidly for the non-SSA low-middle-income group (the low-middle-income countries comparator group includes low-and lower-middle-income countries). The World Bank country income categories based on gross national income (GNI) per capita in 2004 classify low-income countries as those having GNI of $825 or less, and lower-middle-income countries as those having GNI of $3,255 or less.
12World Bank and IMF (2005) provides other examples using data from the Doing Business survey. For example, Nigeria has the most cumbersome regulations in the world for registering property (21 procedures, 27 percent of the property value in fees, and a registration period of 274 days). Such processes, similar to those in other SSA countries, help explain why adequate collateral is often a problem for borrowers.
13While to date less than half of SSA countries have participated in the IMF’s FSAP, for those that were assessed, compliance with some Basel Core Principles has been largely in line with results in other countries. However, compliance with the principle of independence of supervisors, and several principles related to prudential regulations, is relatively low, according to IMF staff calculations based on Financial System Stability Assesments.
14The effective reserve ratio is calculated as the ratio of statu-torily required reserves to the sum of demand and time savings and foreign currency deposits.
15Studies generally find them both ineffective and distor-tionary as a monetary policy tool (Gulde, 1995) and a hindrance to secondary market development.
16Transitioning to greater use of market-based instruments is constrained by the limited interbank market and weaknesses in central bank liquidity forecasting. Country studies conducted in the IMF’s African Department on these and related monetary policy issues include Angola (Alvesson and Torrez, 2003); The Gambia (Harjes, 2004); Nigeria (Gobat, 2003); and Tanzania (Nassar, 2003).
17The size of a branch network may not accurately depict physical access to bank branches because banks in many countries concentrate their branches in urban areas. Data on the rural-urban distribution of bank branches are not available.
18The limited geographical coverage in Africa could be a result of low population density, constraining bank incentives to serve sparsely populated areas. However, average population density is at par with the world average.
19Non-oil per capita income was used in the case of oil-producing countries, given that a large proportion of the population does not benefit from oil revenues. Beck, Demirguc-Kunt, and Peria (2005) find similar evidence for African and non-African countries in a sample of 91 countries. Illiteracy levels, which are closely related to per capita income, are also correlated with access in SSA.
20Based on World Bank Investment Climate Surveys in seven SSA countries: Eritrea, Ethiopia, Kenya, Senegal, Tanzania, Uganda, and Zambia.
21The World Bank Investment Climate Surveys also found that banks require high collateral—on average more than 170 percent of loan value.
22Real lending rates are highest in the WAEMU and CEMAC countries (based on maximum lending rates), averaging 17.5 and 15.5 percent, compared with 11 percent and 9 percent in low-income and middle-income non-CFA countries, respectively.
23An increasing share of government claims in total claims can be consistent with the trend toward fiscal deficit reduction in SSA, to the extent that higher bank financing compensated for financing through arrears and the central bank, which has been declining in SSA.
24Reserves and foreign assets also account for more asset growth in low-income SSA countries than claims on the private sector.
26Banks may be unable to lend if regulation creates an artificial floor on deposit rates and ceiling on lending rates, and limits the ability of commercial banks to reduce deposits or expand lending. Banks—particularly those with monopoly power in the loan market—may also be unwilling to lend when transaction costs are high and risk-adjusted returns low.
27Precautionary reasons for liquidity might include volatility in the deposit base, unavoidably high lending risks, or poorly developed interbank markets and similar structural factors.
28In the CEMAC region, the transmission mechanism was weak in both high and low (involuntary excess reserves) regimes, which was explained by the fact that involuntary excess liquidity (involuntary excess reserves) was relatively high across the whole sample period.
29Interest rate liberalization was associated with sharp increases in real interest rates in many countries. For the 15 countries with outstanding debt in both periods, the median ex post real interest rate rose nearly 10 percentage points between 1985—89 and 1995—2000; in the full non-CFA sample, median interest payments on domestic debt amounted to 15 percent of fiscal revenues in 1995—2000 (see Christensen, 2004). In addition, in the late 1990s, high real interest rates and rapidly mounting interest burdens discouraged the use of bond sales for monetary control in Uganda and Tanzania (Buffie and others, 2004).
30While these restrictions contain risks and prevent speculative activity, they should be balanced against the need for dealers to take open positions to provide liquidity to the market (Canales-Kriljenko, 2003).
31For most SSA countries, calculations of the Reinhart and Rogoff (2005) measure of ex post exchange rate flexibility indicates substantial intervention even for countries notionally committed to a floating exchange rate (Masson and Pattillo, 2005).
32IMF (2004) also points out that, to encourage banks to trade with each other in the interbank market, remuneration rates on reserves deposited with the central bank should be lower than the cost of borrowing from the central bank at the discount window.
33A recent study of IMF-supported programs for 83 countries measured financial sector conditionalities in three ways: intensity (number of financial sector conditions per program per year), hardness (share of prior actions and performance criteria in total program measures), and compliance (proportion of program measures implemented as scheduled) (Giustiniani and Kronenberg, 2005). Intensity has not risen as sharply in SSA as elsewhere in the world because there are no programs driven by financial crises. Nevertheless in the 27 SSA countries for which programs initiated in the early and mid-1990s could be compared to those in the late 1990s to 2001, financial sector conditionality increased by 60 percent. In SSA, “harder” types of banking sector conditionality have been increasing: this measure was higher than the global average in 2000—03. As in the rest of the world, compliance declined over 1995—2003, and was lower than in other areas of structural reform (Appendix 5, Figures A7, A8, and A9).
34The sector comprises NGOs, nonbank financial institutions, credit unions and cooperatives, rural banks, savings and postal financial institutions, and, in some cases, even commercial banks.
35This section is mainly based on a database of 167 MFIs in 37 SSA countries, which was created in 1998 by the CGAP for reporting MFIs in developing countries. While the database coverage is generally good, it is not exhaustive. In some cases it underestimates the true size of the MFI sector. The database distinguishes between three different types of MFIs: (1) regulated (banks, regulated NBFIs, regulated NGOs); (2) cooperative (financial cooperatives and credit unions); and (3) unregulated (other NGOs, NBFIs, MFI projects, and others). However, detailed soundness indicators for these institutions were available for only 27 SSA countries. We have supplemented the database information with credit union data for 15 SSA countries from the World Council of Credit Unions and postal savings banks data from the World Savings Banks Institute.
36For the 86 SSA MFIs that provided information continuously for 2001—03.
37This relationship only holds when credit unions are excluded from the sample, which is sensible; these institutions are often linked to larger enterprises.
38On average, these loans are somewhat bigger than in the Middle East and North Africa, East Asia, and South Asia, but significantly smaller than those offered in Eastern Europe, Latin America, and the Caribbean (Lafourcade and others, 2005).
39This is explained in part by the fact that there are many new MFIs with recently extended loans. The share of NPLs tends to increase over the life of a loan.
40These include general commercial law, corporate law and rules for joint ventures, laws on secured transactions (guarantees and collateral), debt enforcement law, bankruptcy law, arbitration law, accounting law, and contract laws for the carriage of goods by road. Harmonization is also under way for labor and consumer sales law.
41Capital account restrictions in SSA are complex. An average of indicators for controls on 13 types of capital transactions (where a value of 1 indicated a control) was equal to 0.8 for SSA in 1995–99 and 0.75 in 2000–04, compared with 0.71 and 0.7 for low-middle-income countries outside SSA for the same periods. The global averages for these periods are 0.66 and 0.63, respectively (data from IMF, 2005a).
1For example, until recently Kenya had six development financial intermediaries—two development banks and four DFIs, with significant overlap of functions and all with NPL rates over 50 percent. Malawi had an Industrial Development Bank (IDB for industry and agriculture), the Malawi Development Corporation (a holding company), the Agricultural Development and Marketing Corporation (smallholder agriculture), and the Small Enterprise Development Organization, in addition to a funding subsidiary of the IDB. All of these institutions were heavily dependent on donor funding.
2It has recently been privatized, though a minority government stake holding remains.
3Brownbridge and Harvey (1998) argue that the presence of a development bank often acted as an unintended safeguard for commercial banks, as the most severe lending distortions were concentrated at the development bank, mitigating the need for direct government intervention in the operations of commercial banks.
You are not logged in and do not have access to this content. Please login or, to subscribe to IMF eLibrary, please click here