Report No. 15081 Uganda Financial Sector Strategy Update CONFIDENTIAL (In Two Volumes) Volume II: Main Report 15081 VOL. 2 November 10, 1995 Public and Private Enterprise Division Eastern Africa Division Africa Region FOR OFFICIAL USE ONLY L !oY Document of the World Bank This document has restricted distribution and may be used by recipients only in the preformance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. List of Acronyms AfDB African Development Bank BOU Bank of Uganda CDC Commonwealth Development Corporation COOP Cooperative Bank DFCU Development Finance Company of Uganda DEG Deutsche Investitions und Entwicklungsgesellschaft EADB East Africa Development Bank EEC European Economic Community EFMP Economic Financial Management Project EIB European Investment Bank HFCU Housing finance Company of Uganda HVTS High value transfer system IDA International Development Association IFC International Financial Corporation IMC International Monetary Control IMF International Monetary Fund MICR Magnetic Ink Character Recognition MOF Ministry of Finance NBFI Nonbank Financial Institutions NGO Nongovernmental organization NIC National Insurance Company NH&CC National Housing and Construction Corporation NPART Nonperforming Assets Recovery Trust NPC National Payment Council NSSF National Social Security Fund OCR Optical Character Recognition POSB Post Office Savings Banks PWICO Pan World Insurance Company RTGS Real-time gross settlement system UCB Uganda Commercial Bank UCC Uganda Computer Center UDB Uganda Development Bank UDC Uganda Development Corporation ULC Uganda Leasing Company UMA Uganda Manufacturers Association UBA Uganda Bankers' Association UPTC Uganda Post and Telecommunications Corporation URA Uganda Revenue Authority USAID United States Agency for International Develonment VSAT Very small aperture terminal Table of Contents Chapter 1 - M acroecnomic Setting for Uganda's Financial Sector..................................... B ackg ro u nd................................................................................................................... Impact of Economic Liberalization on the Financial Sector............................................. 6 The Financial Sector, Resource M obilization, and the Real Economy ............................. 8 Challenges for the Future............................................................................................. 14 Chapter 2 - Reforming the Commercial Banking System............................................... 15 B a ck g ro u n d ................................................................................................................. 16 Issues and Options....................................................................................................... 23 Reform Strategy .......................................................................................................... 28 A ctio n P la n ................................................................................................................. 3 5 Chapter 3 - Reforming Uganda Commercial Bank.......................................................... 37 B ackg rou n d ................................................................................................................. 3 7 Issues and Options....................................................................................................... 39 Reform Strategy .......................................................................................................... 44 A ctio n P lan ................................................................................................................. 4 5 Chapter 4 - Financial Discipline and the Problem of Bad Debt ....................................... 46 B a ck g ro u n d ................................................................................................................. 4 6 Issues and Options....................................................................................................... 47 Reform Strategy .......................................................................................................... 51 A ctio n P lan ................................................................................................................. 5 2 Chapter 5 - Reforming the Bank of Uganda .................................................................... 54 Introduction and Background ....................................................................................... 54 Issues and Options....................................................................................................... 54 Summary Action Plan.................................................................................................. 64 Strengthening the BOU ................................................................................................ 65 Chapter 6 - Payment Systems ......................................................................................... 66 Background................................................................................................................. 66 Options for Improving Uganda's Payment Systems ...................................................... 69 Creating Infrastructure for the Development of Fmiancial M arkets................................71 Leveraging Government Receipts and Payments for Financial Deepening...................... 73 Improving Currency Use.............................................................................................. 74 Providing Payment Services in Underbanked Areas ...................................................... 74 S trateg y ....................................................................................................................... 7 5 A ctio n P lan .................................................................................................................. 7 8 Chapter 7 - Rural Financial M arkets............................................................................... 80 B a ck g ro u n d ................................................................................................................. 8 0 Issues and Options....................................................................................................... 86 Reform Strategy .......................................................................................................... 89 Chapter 8 - Interbank, Treasury Bill and Foreign Exchange Markets ........................... 92 Background................................................................................................................. 92 Interbank Shilling M arket ............................................................................................ 92 A ctio n P lan ............................................................................................................... 10 8 Chapter 9 -Long-Term Finance..................................................................................... 109 Development Banks................................................................................................... 109 Reform Strategy for Developm ent Banks.................................................................... 11 Other Long-Term Financial Institutions...................................................................... 12 In su ran ce ................................................................................................................... 13 Reform Strategy for Insurance................................................................................... 114 National Social Security Fund.................................................................................... 115 Reform Strategy for N SSF......................................................................................... 115 Chapter 10 - M acroeconom ic Im plications of Reform................................................... 17 F isca l Im p a ct............................................................................................................. 1 17 M onetary Impact....................................................................................................... 123 Impact of Real Economy............................................................................................ 125 A n n e x 1 ...........................................................................................................................a - A n n ex 2 ...........................................................................................................................a -2 List of Tables Table 1.1 - Investment and savings as a percentage of GDP at market prices Table 1.2 - Ugandan macroeconomic indicators, fiscal 1987-94 Table 1.3 - Uganda's money multiplier, 1991-94 Table 1.4 - Relative size of financial intermediation Table 1.5 - Lending allocation Table 2.1 - Uganda's commercial banks and their deposit base in December 1994 Table 2.2 - Level of loan loss provisions of Ugandan commercial banks, 1994 Table 2.3 - Commercial banks net worth positions, 1994 Table 2.4 - Ratio of key performance indicators to total assets Table 3.1 - UCB's operating performance Table 3.2 - Classification of UCB's loan portfolio Table 3.3 - UCB's balance sheet, December 31, 1994 Table 8.1 - Undersubscription in Uganda's T-bill auctions Table 8.2 - Bank participation in 91 day T-bill auctions List of Graphs Graph 1.1 - Uganda: falling inflation Graph 1.2 - Uganda: sources of monetary growth Graph 1.3 - Uganda: real interest rates Graph 1.4 - Uganda: sources of base money growth Graph 1.5 - Uganda: commercial banks advances (U Sh billions at 1989 prices) Graph 1.6 - Uganda: commercial banks advances (average prices as % of GDP) Graph 8.1 - Indicative interbank transactions vs BOU finance Graph 8.2 - Bank T-bill holdings, December 1994 Graph 8.3 - Excess reserves, end of the month Graph 8.4 - Average excess reserve positions, September 1993 - January 1995 Graph 8.5 - Daily excess reserves Graph 8.6 - Spread between maximum and minimum accepted bids: 91 day T-bills Graph 8.7 - Yield curve 25.01.95 Graph 8.8 - T-bill ownership Graph 8.9 - 91 day T-bill yield and excess reserves Graph 8.10 - 91 day T-bill Yield and average maximum and minimum deposit rates Graph 8.11 - 91 day T-bill yield and average maximum and minimum lending rates Graph 8.12 - Excess reserves and cumulative BOU foreign exchange purchases List of Boxes Box 2.1 - UCB's market share Box 2.2 - One way to assess bank performance Box 2.3 - Capital adequacy target for problem banks 1. Macroeconomic Setting For Uganda's Financial Sector 1.1 Looking into the macroeconomic background for Uganda's recent disinflation and the slowed growth of its money supply, it seems clear that these heartening trends are the result of economic liberalization accompanied by the government's new and vigorous fiscal policy. With restrictions largely removed from the country's financial markets, external current account and foreign exchange dealings, there are now few obstacles to capital account transactions. This liberalization has, moreover, been managed with less upset to stabilization than other countries have experienced, creating new target and control variables for the conduct of policy, such as base money targeting and interest rate policy. 1.2 Even the weaknesses of Uganda's financial sector has not prevented the continuous rehabilitation and expansion of the country's economy, which has achieved modest growth in the provision of finance for key sectors. Yet the extent of bad debt in the balance sheets of Ugandan banks-suggesting a significant misallocation of credit-is staggering, and there is clear evidence, moreover, that credit is unavailable to many potential borrowers, often because banks aware of poor credit discipline are cautious about borrowers' security. 1.3 For all the financial system's faults, it has managed to serve the economy during a sustained period of growth. Uganda's present challenge lies in overcoming the crisis in the financial sector and to create the fruitful symbiosis between finance and the real economy. Background 1.4 The troubles of the 1970s and early 1980s produced a severe contraction of Uganda's monetary economy, a decline in financial intermediation, and a loss of financial depth. This in turn led to the failure of many Nonbank Financial Institutions (NBFIs) and the concentration of financial services in the hands of a few commercial banks, of which the largest was the government owned Uganda Commercial Bank (UCB). These years also saw the politicization of commercial life and an undermining of the normal legal restraints and disciplines of business life. Fraud became rampant. The use of such credit instruments as checks declined, and Uganda largely reverted to a cash economy. Confidence in the banking system was further eroded by the currency conversion of 1987, when 30 percent of the value of balances was confiscated. The economy became severely demonetized, with the M2-GDP ratio falling to below 7 percent in fiscal 1987. 1.5 Since 1986 there has been a determined effort by the government to revive the economy and restore normal business practices, including the revival of noncash payment instruments and the restoration of sound banking practice. To promote economic revival, therefore, the economy has been substantially liberalized and the exchange rate floated. 1.6 These newly introduced economic policies have been very successful in encouraging economic growth. With political stability and regulatory liberalization in Uganda, the economy grew by an average of 5.8 percent from fiscal 1987 to 1994, and it is projected to expand by 8.5 percent in fiscal 1995. The growth has been broad-based. Sectors that had reverted in earlier times to subsistence have re-emerged into the cash economy. Export production has increased. Domestic trade has flourished, and some formal sector enterprises have been rehabilitated. Savings and Investment 1.7 This expansion was achieved, moreover, with little new investment in productive enterprise. Total investment expenditure is presently less than 15 percent of GDP, and estimated private investment stands at only about 6 percent of GDP (Table 1.1). Table 1.1. Investment and savings as a percentage of GDP at market prices Investment and Savings Fiscal Year 1987 1988 1989 1990 1991 1992 1993 1994 Gross domestic investment 9.3 10.3 10.3 11.9 14.8 15.2 14.1 12.9 Fixed capital formation Private 6.7 6.5 7.9 7.8 8.8 8.3 6.4 5.6 Public 2.5 3.7 2.6 4.1 5.7 7.0 7.5 7.2 Total 9.2 10.2 10.5 12.0 14.5 15.3 13.9 12.8 Gross domestic savings Public -0.3 0.0 2.0 0.7 -1.6 1.2 -1.9 0.3 Total -0.8 0.0 1.2 1.0 -0.3 -0.5 -1.2 1.1 Gross national savings Public 0.9 1.5 3.5 1.3 0.1 6.1 5.7 7.7 Total 0.1 1.0 2.1 1.0 0.4 1.2 4.6 7.1 Source: Statistics Department, MFEP and staff estimates. 1.8 Uganda's political troubles seriously affected the peoples' capacity for private saving and the country's ability to mobilize equity finance. Domestic savings, both public and private, are therefore negligible, limitmig the need for financial intermediation to convert domestically mobilized resources into investment. Given the state of domestic savings, moreover, the overwhelming bulk of investment financing comes from foreign savings, with public investment financed essentially by donors' grants and loans. Yet 1990 saw the beginning of a strong inflow of private transfers from abroad, which now exceed US$ 300 million a year. These transfers account for the recent upturn in private national savings in Uganda and for the continuing growth of private deposits in financial institutions. 1.9 Uganda's political and economic troubles in the past have also reduced the number and kind of assets which lenders are willing to accept as collateral. Borrowers, lenders, and regulators have left a combined insolvency, very poor credit discipline, and ineffective legal recourse against default. Economic stabilization 1.10 During the past few years of economic recovery, Uganda has suffered episodes of excessive monetization of the fiscal deficit. The economy has also been subject to such shocks as the halving of coffee export prices (in fiscal 1990), their subsequent meteoric rise (in 1994), and drought (in 1992). Aggravated by the decline in the value of coffee exports, fiscal deficits ballooned from 4.2 percent of GDP in 1986 to 14 percent in 1992. Although the deficit was 2 largely covered by external financing until 1989, there was also sufficient monetization to stimulate a high rate of inflation, which reached a peak of 168 percent in fiscal 1988. Because of its low M2-GDP ratio and the large budget deficits, Uganda's economy has been very inflation prone (Table 1.2). Table 1.2. Ugandan macroeconomic indicators, fiscal 1987-94 (in U Sh billion) 1987 1988 1989 1990 1991 1992 1993 1994 GDP real growth % (mp) 3.8 8.6 6.3 6.1 4.7 3.2 8.5 5.4 M2 growth % 91.1 211.9 124.3 57.0 46.7 53.5 42.0 33.3 Inflation Rate % (CPI) 126.5 168.0 130.6 45.4 25.6 41.8 26.0 10.3 (US$ Million) Exports (GNFS) 406.0 323.9 304.0 245.7 198.9 195.1 206.2 333.1 Imports (GNFS) 600.0 682.0 712.0 675.8 670.5 581.6 753.4 893.4 Source: Bank of Uganda GNFS: Goods and Non-Factor Services. Graph 1.1 Inflation and M2 Growth 1.11 In the 1990s Uganda achieved 250 Rates stabilization by means of fiscal consolidation. 200 Although the effects of the 1991 drought delayed the impact of these measures, by 1994, the 150 government had repaid its debt to the banking 100 system and built up a positive net credit balance. Annual increases in M2 have fallen from a peak 50 of212 percent in fiscal 1988 to 33 percent in 0 1 N fiscal 19943-when drought and monetization of 1986 1988 1990 1992 1994 an increased budgetary deficit brought a 0-% change M2 ---Inflation rate temporary upsurge in inflation. Graph 1. 1 shows tar upsurg1988,ech inarinflato Grea 11 s Source: Bank of Uganda, Quarterly Economic Report. that since 1988, each year's rate of increase in the Composite Consumer Price Index has fallen in close correlation with the falling rate of M2 expansion. Graph 1.2 Sources of Monetary Growth 1.12 Today, Uganda's prime source of 200 money supply growth (other than the effects of 150 Oclaims currency revaluation), however, is not increases 100 oP. in the banking system's net domestic assets but rather the country's rising reserves of foreign Hg Claims exchange. Graph 1.2 shows how-after falling D0 onGot until fiscal 1992-net foreign assets have risen -so strongly. There has, for instance, been a steady a Foreign increase in lending to the private sector. In the -loo Assets last six years lending to the government has -150 Fiscal Year only been a serious source of monetary growth . . Source: BOU, IMF, and Staff estmimates. in 1991-92. Smce then, fiscal operations have offset external expansionary pressures. 1.13 Economic stabilization revived investor confidence in Uganda's currency, measures have been taken to repress check fraud that have revived the transactions demand for bank deposits. 3 These two factors together-have raised the demand for money and increased the M2-GDP ratio to nearly 10 percent. The money supply's falling rate growth, combined with the fall in the velocity of circulation, have also combined to bring down inflation sharply. From its peak of 190 percent in fiscal 1988, the GDP deflator at factor cost fell to only 6 percent in 1994. 1.14 One consequence of mastering 25 Graph 1.3 Real Interest Rates inflation has been that real interest rates- which were strongly negative in the late 20 1980s-have since turned positive. As shown in Graph 1.3, except during the 10 1991/1992 inflationary hiccup, money 0 market and lending rates have been 0 88 90 92 96 consistently positive since 1989 in Uganda, 10 with only deposit rates remaining slightly -15 negative in real terms. ---Sa. dep % --Lendin %N 1.15 Threats to stability: the coffee ---T-Bill 91 Day boom. In fiscal 1995, Uganda's foreign Source: BOU, Quarterly Economic Report. exchange earnings from coffee exports increased rapidly following the trebling of world coffee prices from the middle of 1993 to the middle of 1994. The increase in export receipts in fiscal 1995 over those of 1994 expected to be of the order of US$ 390 million) will be equivalent to approximately 8.7 percent of 1994's GDP at market prices-sufficient to generate strong inflationary pressures if not counteracted or "sterilized" in some way. 1.16 Yet it appears that inflationary pressures in Uganda continue to abate in the current fiscal year. The composite consumer price index in April 1995 was only 4.5 percent-higher than in April 1994-and the underlying annual rate of inflation (abstracted from food price fluctuations) remained at approximately 10 percent. 1.17 With prudent economic policies and behavior, Uganda has managed to accommodate the coffee boom without a return to hyperinflation. Imports have expanded by 35 percent in current dollars (from US$ 718 million in fiscal 1994 to a projected US$ 966 million in 1995). Import growth alone, therefore, should absorb 63 percent of the US$ 390-million increase in export receipts. 1.18 The government has also made strenuous efforts to sterilize the residual effect of export growth on domestic expenditure. A new coffee tax in effect as of October 1994 is projected to yield US$ 33 million worth of revenue in the current fiscal year and around US$ 60 million in 1996. The government has committed to save all receipts from this windfall tax, as well as aid receipts above a level set in consultation with the International Monetary Fund (IMF). This would increase the government's credit balances with the banking system, offsetting the monetary impact of increases in the foreign exchange reserve. Uganda's government has been admirably steadfast in maintaining fiscal discipline. 1.19 Uganda's economy is now engaged in a slow, steady process to increase its financial depth. The growing demand for transactions and precautionary balances has allowed the money supply to increase somewhat without inflationary consequences. A 1 percent increase in the M2-GDP ratio that is likely to occur in fiscal 1995 will absorb the equivalent of US$ 55 to 60 million of incremental export receipts. 1.20 There has also been an appreciation of the Uganda shilling, which is expected to be valued some 19 percent higher in terms of the US dollar in fiscal 1995 than it was in fiscal 1994, 4 offsetting some of the effects of rising foreign exchange reserves on domestic monetary assets. However, this currency appreciation is likely to have an adverse impact on Uganda's exports 1.21 Finally, expenditure of a share of the increased receipts from coffee exports in the domestic economy has had a positive multiplier effect on real output-particularly in trade, services, and manufacturing. The consequent rise in real incomes brings with it a noninflationary rise in the demand for money balances. 1.22 This combination of leakage into imports, fiscal sterilization, declining velocity of circulation, rising demand for real balances, and exchange rate appreciation explains the benign impact of the coffee boom on inflation in Uganda. The period of greatest inflationary danger from the export windfall has now passed. Coffee prices have stabilized and are expected to decline over the medium term. 1.23 Yet from the point of view of monetary control it is important to note that, throughout this episode-the contribution to stabilization from the use of liquidity management instruments, such as the sale of Treasury bills and the manipulation of interest rate policy-has been very minor. Treasury bills are largely held by commercial banks. The stock of Treasury bills outstanding with nonbanks remained broadly static at U Sh 12 billion between June and December 1994, and that with banks rose only slightly from U Sh 32.7 billion to U Sh 35.5 million. The rise between December 1994 and March 1995 in the stock of Treasury bills of U Sh 5 billion with non-banks and U Sh 10 billion with banks was still less than the rise in banks' excess reserves of U Sh 19 billion over the same period. Nominal bank lending rates in Uganda have remained largely uninfluenced by the volatile yield on Treasury bills, changing little over the first three quarters of fiscal 1995. 1.24 Threats to stability: weakness of the financial sector. Another threat to macroeconomic stability in Uganda is the weakness of the financial sector. At the start of the 1990s Uganda Commercial Bank increased its branch network in a flurry of undisciplined expansion, incurring large operating losses. Banks short of liquidity frequently incurred overdrafts from the Bank of Uganda, undermining the country's monetary policy. Uganda's largest commercial bank, which at the time had 60 percent of all deposits in the banking system, was in growing need of liquidity assistance. 1.25 This danger was overcome by dint of tighter supervisory controls, a change of management at UCB and the closure of part of its branch network. By 1994 the system's chronic liquidity shortage had turned into a state of excess liquidity, and the UCB had fully reconstituted its statutory reserves. As of December 31, 1994, the UCB had cash and reserves of U Sh 27.5 billion-over 10 percent of its total assets. In April 1995, all of Uganda's commercial banks together held net reserves of U Sh 69 billion, over 20 percent of their deposit liabilities, as compared with only 14 percent in April 1994. Any threat to economic stability posed by illiquidity in the banking system had been removed. 1.26 Yet despite this success, the UCB, and several other banks, are insolvent and probably have been technically for some years. So long as depositors retain confidence in the integrity of their banks, this poses no immediate threat to the country's economic stability. So far the only bank to be liquidated was small, and its depositors were fully compensated from public funds without serious fiscal consequences. In testimony to continuing depositor confidence in the financial system, moreover, depositors in the main banks continue to rise. But Uganda's need for large-scale recapitalization of banks remains. 5 Issues for the future 1.27 Uganda's economy has grown despite low levels of investment and a very low rate of domestic savings. This is due largely the positive supply response to political and economic stabilization, and liberalization and the restoration of incentives. But sustaining this growth in the future will require far greater mobilization of domestic resources and a financial sector able to allocate the resources efficiently. 1.28 The country's macroeconomic stabilization has helped to restore business confidence and remonetize the economy. The government's fiscal action has been particularly vigorous at times, as at present, when private sector expenditure has been boosted by windfall export receipts. But as demand for real balances rise relative to GDP, the scope for seigniorage will increase if the economy remains stable. Impact of Economic Liberalization on the Financial Sector 1.29 In 1990 Uganda's economy was still extensively controlled, and foreign exchange was allocated administratively by the Bank of Uganda (BOU). Imports were not permitted without a license. Credit from commercial banks was controlled and to a significant extent directed by the Bank of Uganda. Nonbank financial institutions were required to hold defined quantities of government securities, and the BOU dictated interest rates on deposits and advances. 1.30 All of these controls have now been relaxed, if not fully abrogated. In 1990 foreign exchange bureaus for parallel-market trading were permitted. Foreign exchange auctions were introduced in 1992 but were replaced by an interbank market in foreign exchange in 1993, when the exchange rates were unified. In principle, Ugandan residents' access to foreign exchange is free only for current account purposes, but in practice there is no serious attempt to prevent access to the market, for capital account purposes as well. The Bank of Uganda intervenes sparingly in the interbank market as in September 1994, and then only to stabilize the rate. It nevertheless retains the option of intervention as one possible means of monetary control. 1.31 In the financial sector, the BOU has ceased entirely to control interest rates and the direction and volume of bank lending. Since 1992 BOU has used such instruments of indirect monetary control, as the auction of Treasury bills and the varying of bank reserve requirements. Neither of these two instruments, however, has been particularly effective or sensitive in influencing the volume of bank lending to the private sector. 1.32 The BOU does not, for instance, make much use of its power to vary bank reserve requirements because of the precarious profitability of so many banks, and because most banks have chosen to hold excess liquidity. The auction market for Treasury bills, moreover, is dominated by the commercial banks yet remains thin and illiquid. Interest rates at auction are volatile and have no immediate (and only limited short-term) impact on bank lending rates. 1.33 The BOU, in agreement with the IMF, is charged with adhering to a base-money time path designed to preserve the current low rate of inflation. With the abandonment of direct controls and with the present weakness of indirect monetary controls, Uganda is poorly placed to fine-tune the growth of its base money-particularly with the rapid rise in officially-owned net foreign assets. Instead, tight controls have been imposed on government expenditure to avoid increases in fiscal receipts above programmed levels in government credit balances with the BOU. Graph 1.4 shows Uganda's large swings in the aggregates that determine its base money-particularly in net foreign assets and in BOU advances to cover government liabilities. 6 1.34 Reasons for the recent explosion in IGraph 1.4 SourcesofBeMoy mutual claims by both the government and the 200 Growth Bank of Uganda are unclear. Explanation must I 150 'rm await the forthcoming audit of the BOU's loan items book. 0 1.35 The erratic course of base money on PS expansion raises the question of how effective 0 -.Claims base money control can be while Uganda is trying -0on Gvti to contain inflation and damp inflationary IMNet expectations. With the exception of a short lag, Foreign inflation has been stably related to the broad -150 End Fiscal Year Assets money supply. But because of fluctuations in ouUe F a both the cash-to-deposit ratio and banks' reserve ratio, the money multiplier has been unstable, and has recently declined. Even if it had been stable, Table 1.3. Uganda's money multiplier, 1991-94 Year Aso June December 1991 2.06 2.11 1992 1.75 1.69 1993 2.08 2.07 1994 1.94 1.91 Uganda's strong economic growth, falling velocity of circulation, and exchange rate appreciation would have contributed-alongside the government's policy of monetary restraint-to the process of disinflation. 1.36 The money multiplier relating the volume of M2 to base money is governed by the formula (l+c)/(c+r), where "c" is the ratio of cash in circulation to deposits in commercial banks and "r" is the ratio of banks' reserves to their deposit liabilities. In Uganda, both "c" and "r" have increased, but erratically so (Table 1.3). 1.37 Uganda has been fortunate to manage economic liberalization without suffering worse inflationary consequences. It is commonly the case that exchange, trade, and financial sector liberalization cause a drop in the domestic demand for money relative to GDP because the greater efficiency of liberalized markets and wider access to credit allow for lower balances than were previously required to finance both imports and domestic transactions. Savers also have wider access to currency substitutes after liberalization than before, either in the form of foreign currency balances or of securities. The scope for seigniorage also decreases when markets are liberalized. Unless the authorities reduce their reliance on seigniorage at the same time as they introduce measures of liberalization inflation is likely to increase. 1.38 Uganda's ability to implement a wide-ranging liberalization program at a time of rapid monetary growth without increasing inflation was probably the result of the country's very low demand for money at the start of the period, which gave little scope for a further decrease. Investor confidence in the country's prospects for growth, moreover, were strong enough to cause an increase in the demand for money in the face of forces that could otherwise have caused demand to decline. 1.39 Economic liberalization has posed less of a threat to macroeconomic stability in Uganda than elsewhere because of the country's lack of financial depth and the structural weaknesses of its 7 financial institutions. But as the financial sector overcomes the problem of bad debt and becomes stronger and more entrepreneurial, the instruments of monetary policy must also be strengthened. They must be sensitively used, moreover, to offset structural shifts in the demand for money and to supply needed credit. The Financial Sector, Resource Mobilization, and the Real Economy 1.40 Uganda's financial system was originally founded to facilitate the marketing of agricultural produce, particularly of coffee and cotton for export. In the 1920s and 1930s small producers of these crops were able to expand their output considerably thanks to crop financing, that is, the provision of credit through the marketing chain. Producers were paid promptly in advance of export sales and were able to obtain some inputs on credit. Foreign purchasers of Uganda's exports effected payment and often prepayment, through commercial banks with strong offshore connections. At the same time banks developed their role in financing imports by providing traders with working capital. 1.41 These original patterns of commercial banking activity, which took physical stocks of commodities or bills promising payment or delivery for security, are still plain in the operation of Uganda's commercial banking system of today. In December 1994, some 64 percent of commercial bank advances to the private sector were assigned to finance crops, trade, or services. 1.42 Before its troubles of the 1970s and early 1980s, Uganda's small economy had a sophisticated financial network, complete with nonbank savings and housing finance institutions as well as commercial banks serving both external and internal trade and the growing manufacturing sector. To provide term loans, new development finance institutions were created. Relative size offinancial intermediation 1.43 Since domestic savings in Uganda were minimal and the M2-GDP ratio was abnormally low, resources for investment were essentially mobilized from abroad. Flows of funds (changes in assets and liabilities) among the key agents in the Ugandan economy for fiscal 1991-94, illustrates the relative sizes of sources of financing and of financial intermediation. The data are drawn from the consolidated accounts of the commercial banks and the BOU, supplemented by fiscal and balance of payments data. Some flows are omitted for lack of data, such as those involving NBFIs or tiansactions between households and private enterprises. Clearly domestic transactions are dwarfed by external flows in both the public and the private sector. But while commercial banks have been building up their net asset positions with the BOU and with the external sector, their ability (or willingness) to extend advances to the private sector has diminished. Yet apart from the growing voluntary reserves of commercial banks, there is no obvious evidence that lending to the private sector by banks is being crowded out through the sale of government securities. Sales to the private sector are small, and sales to banks are roughly matched by government deposits with the commercial banks. 1.44 The Bank of Uganda's role is essentially that of banker to the government, and to the commercial banks (and thus manager of the money supply) and manager of the official foreign exchange reserves. 1.45 In fiscal 1994, when the country's net fiscal position strengthened rapidly, the government's deposits with the BOU and BOU's advances to the government burgeoned. One explanation for this sudden growth is that the BOU sold more and more Treasury bills. The government, however, had no need of this financing and put the proceeds into ever larger deposits 8 Table 1.4. Relative size offinancial intermediation: flow offunds in the Ugandan economy 1991-1994 (U Sh Billion) Source of Funds Fiscal Recipient Year Government BOU Commercial Private Foreign Banks Sector (A). (B) (C) (D) (E) Government 1992 -13.7 10.5 (1) 1993 44.4 3.9 1994 273.5 21.4 BOU 1992 36.1 -3.5 -0.4 50.5 (2) 1993 -24.0 1.0 -11.0 50.5 1994 227.3 0.3 4.4 75.3 Commercial Banks 1992 6.1 30.9 25.6 35.2 (3) 1993 2.9 14.6 46.5 45.6 1994 19.4 20.5 39.9 8.1 Private Sector 1992 4.6 70.0 n.a. n.a. (4) 1993 -6.5 85.6 1994 6.0 75.3 IMF 1992 192.7 (5) 1993 17.8 1994 -54.1 Other Foreign 1992 196.5 -1.5 7.3 130.6 (6) 1993 177.0 -2.8 15.6 289.6 1994 217.5 -0.5 7.8 332.9 Key to cells (Cl): Changes in government deposits in commercial banks (A2): Changes in BOU advances to government (C2): Changes in BOU advances to commercial banks (A3): Commercial Banks' advances to government and purchases of (C4): Changes in private sector deposits in commercial banks securities (A4): Non-bank financing of fiscal deficit (C6): Changes in foreign liabilities of commercial banks (A6): External financing of fiscal deficit (D2): Changes in BOU advances to the private sector, e.g. for crop financing (Bl): Changes in government deposits in BOU (D3): Changes in commercial banks' advances to the private sector (B3): Changes in Commercial banks' deposits in BOU (D6): Private inward transfers from abroad (BS): Changes in BOU liabilities to IMF (drawings less repurchases) (E2): Changes in BOU's foreign assets (B6): Changes in other external liabilities of BOU (E3): Changes in commercial banks' foreign assets at the BOU. But since the stock of Treasury Bills outstanding (U Sh 45 billion as of June 1994) only increased by U Sh 20 billion in fiscal 1994, other explanations must be sought. 1.46 Table 1.4 shows a healthy development of relations between the BOU and commercial banks. The banks have been building their formerly depleted reserve deposits at the BOU and have been borrowing little. As a result, their liquidity position has strengthened considerably and Uganda's banks now have substantial excess liquidity. Role ofBank of Uganda in financial intermediation 1.47 The BOU has also acted as an apex lending organization for development and crop finance. Donor agencies have used it-either directly or through commercial and development banks-to on-lend assistance for directed lending purposes, such as the USAID-financed program for the Rehabilitation of Productive Enterprises started in 1984. 9 1.48 Under the 1993 International Development Agency (IDA) credit for the Industry and Trade Credit Refinancing Fund, eligible commercial bank advances to businesses could be refinanced through the BOU. Until 1990 the BOU used commercial banks as its agents for its annual crop (essentially coffee) marketing finance programs. The BOU is now no longer actively involved in crop finance, since private sector coffee exporters generally can obtain prefinancing from abroad. Negative net flows from the BOU to the private sector also indicates the absence of new development finance activity. Weakness of commercial banks 1.49 Several of Uganda's commercial banks are insolvent and most are run inefficiently and at high cost. Intermediation margins, therefore, are currently above 13 percent, with a cost of funds to banks in the range of 3 to 4 percent and lending rates upward of 16 percent. Administration costs are high, banks carry high unremunerated excess reserves, and new regulations, require banks to keep large provisions against bad and doubtful loans. 1.50 Almost 40 percent of Uganda's bank advances are nonperforming. While deposit and lending rates, once controlled by the BOU, were liberalized in 1992, banks need high lending rates to cover their costs and losses on bad debt. 1.51 The question remains as to whether or not these weaknesses in the banking system are likely to pose a threat to the country's economic performance. The insolvency of UCB and of several other Ugandan banks, for instance, makes it impossible for these institutions to intermediate efficiently between depositors and borrowers and thus ensure that credit is extended to those borrowers who can use it most efficiently. 1.52 Borrowers who fail to repay are assumed (though it is not a certainty) to have used the funds for activities that are not economically viable, or to have been absorbed by enterprises that proved nonviable. 1.53 But it is not easy to tell how far sound borrowers were discouraged by high rates and how many were replaced by riskier borrowers with fewer financing options who are more liable to default. The data reveal no obvious pattern of speculative borrowing, making it likely that-in Uganda's persistent default culture inherited from earlier times-borrowers have been insensitive to real changes in the interest rate, so that credit allocation has been little affected. The main effect of high rates would therefore have been to raise the incidence of default. 1.54 On the face of it, Uganda's real economy suffered little from the weakness of the country's financial system. Expansion continues virtually uninterrupted despite the (inevitable) liquidity difficulties of firms. Banks have slowly improved their service to the economy, with deposits and bank lending growing faster than nominal GDP, increasing from 2.8 percent of GDP in fiscal 1990 to 4.3 percent in fiscal 1994. 1.55 Many small enterprises still have no access to bank lending. However, bad debt is still rampant and domestic savings are small. In addition, the farm sector, which has received much directed lending, has been largely unable to service its debt. Sectoral allocation of commercial bank lending 1.56 To date Uganda's producers and traders have had sufficient local currency liquidity to finance a steady growth of output of over 5 percent a year. Imports of consumables have also 10 increased rapidly. In fiscal 1995, it is expected that nonproject imports will increase by a third to US$ 780 million. (Table 1.5) Table 1.5. Lending allocation (U Sh billion) Sector 1989 1990 1991 1992 1993 1994 Dec. June Dec. June Dec. June Dec. June Dec. June Dec. Crop marketing 6.3 9.0 12.1 16.1 23.8 25.3 27.7 30.0 35.5 40.9 51.2 Agric. Product. 4.1 6.8 8.3 9.8 3.9 2.5 5.1 7.6 6.7 6.0 6.7 Manufacturing 4.9 5.1 7.1 9.0 12.3 17.8 22.1 26.3 32.4 40.2 46.1 Transport 4.1 4.5 4.9 6.1 5.7 6.7 7.3 7.9 9.5 9.1 8.2 Trade/Services 12.7 17.6 23.0 30.2 34.7 44.8 60.0 71.1 77.6 85.4 96.2 Building 2.5 4.4 4.3 5.7 7.7 10.1 10.5 10.9 11.5 12.7 18.0 TOTAL (incl. nes)34.7 44.6 59.8 77.0 88.4 104.8 129.4 154.1 175.4 194.6 228.8 Source: BOU data on commercial bank lending by sector compiled from banks' returns. 1.57 The provision of trade finance in Uganda requires a flexible and effective capacity for financial intermediation. In the last five years, financing for trade and commerce has grown rapidly in real terms, from U Sh 11 billion in December 1989 to U Sh 33 billion in December 1994 (in September 1989 prices)-an average of 25 percent a year. (Graph 1.5).' Some new banks (such as Greenland Bank ) have specialized in this branch of activity, and the loan portfolios of most Ugandan banks apart from UCB give pride of place to the financing of external and domestic trade. 1.58 Commercial banks have increased the Graph 1.5 Sectoral Distribution of Bank volume of their lending to manufacturing and crop 30.00 Credit marketing more rapidly. In terms of September 25.00 1989 prices, lending for manufacturing rose from 20.00- U Sh 4 billion in December 1989 to U Sh 17 billion 0 in December 1994, while that for crop marketing -s.00 rose from U Sh 5 billion to U Sh 18 billion-that 10.00 is, in excess of 30 percent a year. Within food 5.00 manufacturing, the beverage and tobacco. 0.00 subsectors receive the largest share of bank credit. 1990 1991 1992 1993 1994 1995 Ugandan banks have, however, lost an important -r- Cpnanc -- ae/ Prd us share of the coffee market to competition from Source: BOU data on commercial bank lending by sector compiled external sources of credit. by bank returns. 'Outstanding bank credits have been deflated by the Composite Consumer Price Index. 11 1.59 The data thus indicate that Uganda's commercial banks have been able to support economic expansion with rapid growth in the volume of their advances where they see profitable and manageable lending opportunities. (Graph 1.6). Bank lending has, however, always been selective. As banks strive to overcome the persistent problem of bad debt through more rigorous credit assessments, they have become increasingly cautious in their lending. 1.60 Uganda's farm sector uses credit essentially for marketing. While traders need credit to procure Graph 1.6 Commercial Bank Advances produce from producers, the producers themselves 4.5 (% GDP) are very low users of crop production credits. Much 4 3.5 -- - of the credit provided for production, moreover, has 3 come in the form of external assistance to the 2.5 - government, which has on-lent the aid in the form of 2 directed credits administered by commercial banks 1.5 - and development finance institutions. One example I is the Rural Farmers' Scheme administered by the o i UCB. In the past, lending operations involving funds 4 en thought to come from the government have been 6 F C characterized by poor credit discipline and low levels Source: BOU, Quarterly Economic Report. of recovery. Agricultural land has also proved to be poor security for commercial lending because transactions in land are subject to approval by local communities, and there is no ready market for land. 1.61 Until now, Uganda's agricultural sector has generally used few inputs. In the aftermath of the civil disturbance, moreover, the priority has been to restore internal communications, payment mechanisms, and trade. Now that there has been substantial progress on these fronts, more progressive farmers are seeking investment for farm improvement and to pay for the more intensive use of inputs. But for this they will need wider and more sustained access to credit-either provided directly by the banks or made available indirectly through the suppliers of agricultural inputs and equipment. Shortcomings in the provision offinancial services 1.62 Major commodity export houses, and the other blue chip trading and manufacturing companies in Uganda, experience no rationing of working capital from commercial banks. Uganda's banks, which currently have excess liquidity, are only too happy to lend to borrowers who are good credit risks. The principal complaint of borrowers is that of the high real interest rates on bank loans and overdrafts, which is exacerbated by the high and easy returns banks can make in the treasury bill market. 1.63 Compared with those of other countries, Uganda's financial services are expensive and inefficient. The average cost of credit to firms in the 1995 Ugandan Manufacturers' Association Survey (UMA) 2 was 22 percent a year. As firms were interviewed at a time when inflation in Uganda was of the order of 7 percent a year; the real interest rate which firms were paying was about 15 percent. 2 Private Enterprise Survey undertaken for the World Bank's Private Sector Assessment mission by the Uganda Manufacturers' Association Consultancy and Information Services. 12 1.64 The high cost of credit from domestic lenders, moreover, has induced traders who have access to foreign sources of finance-such as private coffee exporters-to rely heavily on the export prefinancing extended by their principals and customers abroad. From fiscal 1993 when coffee exporting was liberalized, external financing has increased significantly. Today almost two- thirds of coffee marketing finance is externally sourced. The foreign departments of commercial banks in Uganda have often been the agents for this foreign currency trade financing. 1.65 Other commercial borrowers in Uganda are generally limited in their ability to borrow from banks by the poor quality of their collateral. It is common in all countries that newly founded small enterprises without a track record have little access to bank credit. Most new ventures depend heavily on equity to attract start-up financing, and even venture capital financiers expect business owners to provide a substantial share of the initial costs. Indeed, the Ugandan Manufacturers' Association survey of Uganda's private sector completed in February 1995, found that the share of own capital and family equity invested in total long-term capital ranged from 68 percent for large firms to 79 percent for microenterprises, which is not uncommon for developing countries. 1.66 However, in Uganda, years of civil disturbance, a shortage of equity, and the dearth of assets banks will accept as collateral for loans pose significant problems for businesses seeking term credit. The major specialized term lending institution (UCB) is insolvent and has fewer external lines of credit to on-lend now than in the past. The resources available to other development finance institutions are also few, although growing. But even with new institutions offering venture capital and lease financing, the enterprises sampled in the Ugandan Manufacturers' Association survey, cited limited access to credit as second only to the incidence of taxation as a source of concern. Only 39 percent of the firms in the sample were users of commercial bank loans. 1.67 Two further factors presently inhibit the use of bank credit by Uganda's small and medium-sized enterprises. First, the processing of loan applications is often very slow. In 1994 for instance, in order to improve the quality of its lending, the UCB introduced rigorous procedures for verifying the security-whether mortgages or personal guarantees-taken as collateral for loans. New loans now have to be cleared with the department responsible for verification before they become effective. The average delay for firms polled in the UMA sample was sixteen weeks, which is particularly excessive given that most firms were seeking overdrafts or short-term loans for working capital. Lengthy delays, moreover, cause prospective borrowers to miss many business opportunities. 1.68 Reacting to loan losses incurred because of deficient or impaired collateral in past years, bank lenders also now ask for collateral well in excess of the value of the advances made. Banks seek to cover the transaction costs involved in realizing the value of security in the event of default. The current rule in UCB is that loans should not exceed 60 percent of the value of assets taken as security. Therefore, until legal processes for loan recovery are streamlined, borrowers are likely to continue to face demands for loan security they cannot meet. 1.69 Finally, there is a serious weakness in the foreign exchange services available in Uganda. The (now numerous) foreign exchange bureaus-operated mainly by commercial banks-conduct only spot transactions and provide no forward cover. The main reason for the absence of forward contracts in foreign exchange is the high rate of interest in the domestic market makes the cost of delivering foreign exchange in the future prohibitively expensive. Buyers and sellers of foreign exchange have no means of protecting their earnings against exchange rate fluctuations in Uganda. 13 Issues 1.70 Uganda needs to move quickly to widen access to credit and to lower its cost. But this will require progress on the problem of bad debt, including improvements in loan security, tighter credit discipline, and better legal recourse. Once loan recoveries improve, the costs of intermediation should fall, making commercial borrowing less prohibitive. Better loan security would also make banks more willing to lend without relaxing necessary prudential controls. 1.71 It is important that more funds need to be made available for intermediation through the financial sector. This will require increased domestic saving and financial deepening to increase the modest supply of funds currently available through Uganda's financial markets. More competition to supply these funds would lend to a greater range of better quality financial services. Challenges for the Future 1.72 Despite the weaknesses of its financial system, Uganda has achieved sustained growth and attained a substantial degree of macroeconomic stability despite profound external shocks. With the economy slowly becoming remonitized (although starting from a very low level), there is no longer any threat of systemic collapse within the financial sector. Indeed, confidence in Uganda's currency, financial institutions, and financial instruments has modestly revived. With annual remittance inflows of approximately US$ 300 million, capital flight has been reversed. The dangers of macroeconomic destabilization arising from uncontrolled lending by commercial banks or from bank insolvency following financial liberalization, have so far also been avoided. 1.73 But Uganda's financial situation is precarious. Impending institutional change within the sector could undermine the fragile confidence so recently established. To resolve the problem of bad debt, banks will probably consolidate their assets and eschew expansion until their balance sheets are restored to health, restricting funds available for immediate economic growth. But in the longer term Uganda's real economy can only gain from having a solidly based, competitive financial sector that can provide a wider range of services at reasonable cost. Only then will the current institutional impediments to the provision of credit be overcome. 1.74 Uganda's financial sector challenges for the future are therefore: o To rectify the current financial sector imbalances between assets and debts. o To manage the reform of the financial sector without an abrupt contraction of the current level of financial service provision. o To introduce competition into a rejuvenated financial sector. o To lower the supply cost of financial services by drastically reducing loan losses and by managing liquidity more efficiently through financial market operations. o To deploy the range of interventions needed-in the financial sector, the professions, the judiciary, and the government-to dispel the present widespread practice of debt default and inculcate the credit discipline that is required for normal business. 14 2. Reforming the Commercial Banking System 2.1 Uganda's commercial banking system is at the heart of the country's financial sector. As has been outlined in chapter 1, improvements in the performance of the commercial banking system are essential to support the anticipated growth of private sector enterprise. This chapter describes the key features of the existing commercial banking system, identifying weaknesses and problems, including those associated with banking supervision. It highlights the main issues to be addressed, together with reform options. A recommended strategy is set out to achieve a sustainable improvement in banking performance, with an action plan to summarize and guide its implementation. The strategic importance of the Uganda Commercial Bank (UCB), as part of the commercial banking system, is analyzed here, but the specific issues surrounding its reform are reserved for discussion in the following chapter. 2.2 The analysis Table 2.1. Uganda's commercial banks and their deposit base reveals that the (U Sh millions) commercial banking Majority Ownership Deposit Base Percentage Percentage (December of Total of Total sector is facing serious 1994) (December (December difficulties, including 1993) 1994) problems with poor Public Sector performance (and high UCB 130,397 47.2 38.5 costs), actual or likely Cooperative 54,198 5.9 7.1 insolvency, and Total 154,595 53.1 45.6 inadequate Private foreign implementation of Stanbic 20,503 7.0 6.0 banking regulations. Barclays 38,774 11.5 11.4 There is an urgent need Baroda 35,055 9.2 10.3 to implement a strong Standard 25,697 6.9 7.5 todcplement arogm Tropical Africa 4,636 0.9 1.3 Total 124,665 35.5 36.8 for its reform. That Private domestic program should adopt Nile 16,487 4.2 4.8 specific actions within Gold Trust 3,117 1.1 .9 four themes: (i) remove Greenland 18,808 1.6 5.5 government ownership Centenary 6,589 1.6 1.9 (including privatization Sembule 7,109 2.0 2.1 of UCB) and direct International Credit 2,222 0.2 .1 operational involvement Orient 4,526 0.5 1.3 in the commercial Kigezi 346 0.1 .1 banking sector; (ii) Total 59,204 11.3 17.4 maintain confidence in the system by facilitating Total bank deposits in 338,464 100 100 prompt turnarounds of Uganda problem banks, or Source: BOU, Quarterly Economic Report. otherwise managing their orderly exit from the system; (iii) create an improved banking environment, through enhanced market structures and regulations, in which market forces can properly encourage competition and discipline amongst private sector banks; and (iv) enhance the standards of bank supervision, and actively enforce measures to ensure compliance with market regulations. 15 Background 2.3 Market overview In December 1994 Uganda had fifteen licensed, operating commercial banks which could be categorized as public sector, foreign, and domestic private banks by their ownership structures. (Table 2.1). 2.4 Clearly UCB dominates the sector, accounting for 39 percent of total deposits, although taken together, foreign banks account for another 37 percent of deposits. Uganda's eight private domestic banks, on the other hand, are relatively small, with an average deposit base of around U Sh 7,400 million (US$ 7.8 million). 2.5 On the lending side, the market position is less clear-cut. UCB once again dominates the market in terms of gross loans, with its U Sh 106 billion portfolio accounting for 44 percent of the system total. But UCB's very high provisions for bad loans reduce its net loan position to U Sh 30 billion, only 20 percent of the equivalent system aggregate. UCB's position in the lending market with good-risk customers in the future is unlikely to be as large as past indicators of gross loans would suggest (Table 2.2). Table 2.2. Level of loan loss provisions of Ugandan commercial banks, 1994 (U Sh millions, as of end December, 1994) Bank Gross loans and Loan loss Loans net of advances provisions provisions Public UCB 106,972 76,456 30,516 Cooperative 22,310 3,251 19,059 Total 129,282 79,707 49,575 Private foreign Stanbic 3,053 316 12,737 Barclays 24,645 3,742 20,903 Baroda 25,439 1,491 23,948 Standard 10,828 561 10,267 Tropical Africa 2,858 1,474 1,384 Total 76,823 7,584 69,239 Private Domestic Nile 10,885 569 10,316 Gold Trust 2,210 69 2,141 Greenland 10,028 274 9,754 Centenary 3,900 1,084 2,816 Sembule 5,870 854 5,016 International Credit 1,702 84 1,618 Orient 925 - 925 Kigezi 179 2 177 Total 35,699 2,936 32,763 Total 241,803 90,225 151,578 Source: BOU and staff estimates 2.6 In total, Uganda's banking system supports 146 branches but has total nation-wide assets of only U Sh 558 billion (approximately US$ 620 million) and total deposits of U Sh 338 billion. This amounts to average total assets of only U Sh 37 billion per bank and U Sh 3.8 billion per branch. Indeed, paid-in capital requirements for the fifteen operating banks total only U Sh 13.6 billion or 2.4 percent of total assets. 16 2.7 UCB is 100 percent government-owned and-despite having the slowest growth in deposits of all Uganda's banks in 1994-still dominates the industry (Box 2.1). The influence of government ownership extends beyond UCB, since it owns substantial portions of most of the foreign banks, including 50 percent of Tropical Africa (formerly the Libyan bank), and 49 percent of Barclays, Stanbic, and Baroda. The government does not own any Standard Chartered Bank shares. Box 2.1 UCB's market share * 33 percent of total assets * 39 percent of total deposits * 44 percent of total gross loans * 58 percent of total branches * 62 percent of Kampala's branches 2.8 The Cooperative Bank's shareholding formally belongs to a set of cooperatives, however it maintains close ties with government. Its mode of operation, moreover, is still firmly rooted in public sector traditions. 2.9 There is a sharp contrast between banking services in Kampala and the rest of the country. Kampala itself is probably overbanked, with eighteen banks (including four recently licensed banks that anticipate operating in Kampala) that have forty-two branches operating (often unprofitably) in a small market. 2.10 Foreign banks, with a 37 percent market share of deposits, deal primarily with Europeans and multinationals. Outside Kampala, Baroda and Greenland, which have a strong customer base in the Asian community, are established in one or two locations. But the vast majority of rural business falls to UCB and the Cooperative Bank. In fact UCB-in addition to its agencies, many of which are losing money-maintains roughly forty branches in areas where there is no other commercial bank branch and about eighteen rural branches near Cooperative Bank branches. 2.11 Yet despite problems with its performance, the banking system in recent years has expanded, reflecting Uganda's strong macroeconomic performance as a whole. Total bank deposits grew by some 17 percent over the last eighteen months, although individual bank performance has been mixed. While UCB has achieved a 14 percent growth of deposits, Stanbic's deposits have fallen by 16 percent, and several of the private domestic banks (including Greenland and International Credit) have increased deposits by more than 150 percent. Even though many banks are small (in terms of total assets) and several have experienced declines in their deposit base, there have been no mergers and only the exit of Teefe Bank from the marketplace. For bank entry, licensing requirements are administered by the BOU's banking supervision department. Four new banks (based in Kampala) have been licensed within the last year, reportedly over the objections of the BOU. With their tax free status, relatively large capitalization, and fortunate timing (arriving now that the banking environment is improving), these four new banks will almost certainly exacerbate the problems of several of Uganda's private banks. Performance of commercial banks 2.12 On the basis of the data set compiled during several missions to Uganda, the two most obvious trends which surface are: (1) the growing insolvency of the banking system by taking stock of the poor lending record of Uganda's commercial banks, and (2) the growing reluctance on the part of the banking system to lend to the private sector. Although deposits have grown from U Sh 188 to U Sh 338 billion between the end of 1992 to end 1994, aggregate net worth has, in fact, fallen from a positive U Sh 27 billion to negative U Sh 60 billion over the same period. This 17 deterioration in net worth is primarily driven by the provisions made for non-performing loans which have accumulated over time and do not necessarily reflect poor lending practices in the past few years. The fraction of advances which were categorized as non-performing rose from 13.4 percent at the end of 1992, to over 45 percent at the end of 1994. With stricter loan loss provisions and the uncertainty concerning how the government intends to resolve the problem bank issue, banks have become increasingly reluctant to make new loans. At an aggregate level, the loan/deposit ratio has fallen from 81 percent at the end of 1992, to less than 73 percent by the end of 1994. The corresponding increase in the fraction of liquid assets to total assets, has been from 35.5 percent to almost 45 percent. 2.13 The trends shown above can be disaggregated by the ownership structure of Uganda's commercial banks. In general, foreign banks are becoming less important in the mobilization of total deposits, with their share falling from 43 percent at the end of 1992, to less than 37 percent at the end of 1994. Private domestic banks have largely taken the slack, increasing their share from 10 percent to almost 18 percent in this period. The negative aggregate core capital of Uganda's banking system, which decreased in absolute terms from negative U Sh 19.5 billion to negative 108.5 billion (from end 1992 to end 1994), was primarily driven by public sector banks which posted values of negative U Sh 24.2 billion and negative U Sh 117.4 billion during the same period. The loan loss provisions made in the specific case of UCB, account for this significant decrease in net worth. In terms of outstanding loans, public sector banks reduced their overall exposure from 69 percent of the total loans outstanding, to only 53 percent in the period 1992- 1994. Private domestic banks have, on the other hand, increased their market share from 6.5 percent to almost 15 percent in this period. Despite the growing importance of private domestic banks in deposit mobilization, their loan to deposit ratio fell from 71 percent to almost 60 percent during 1994 alone. As expected, the relatively stronger and more conservative foreign banks posted a much smaller fall in this ratio, from 66.8 to 61.6 percent during 1994. 2.14 Uganda's banks have performed poorly in recent years. In 1994, for instance, UCB recorded losses of U Sh 42,071 million, which constituted 95 percent of the total system-wide loss of U Sh 44,579 million. Half of the remaining banks (seven out of fourteen) recorded losses. Both foreign banks and private domestic banks (as a group) lost money with total after-tax losses reaching the equivalent of some 9 percent of total assets. 2.15 It should be noted that operating results in 1994 have been skewed somewhat by the introduction of tougher loan provisions. It is therefore likely that the annual bad debt expense during the year was artificially inflated by the need to catch up on earlier underprovisioning. Nonetheless, many banks including UCB, have reported losses in earlier years. The latest results only confirm poor system-wide profitability and eroding capital. 2.16 The effect of these losses and of bad loan experience that have weakened many banks' balance sheets has meant that the system as a whole now reports a negative net worth of nearly 11 percent of total assets (Table 2.3). Again, although the magnitude of UCB's negative net worth dwarfs that of other banks, significant weaknesses can also be seen in other institutions. 18 2.17 The BOU Table 2.3. Commercial banks net worth positions and ratings Banking (U Sh millions, as ofDecember 1994) Supervision Department h Bank Core Capital Total Net BOU Rating Devie hastm worth devised a system Public Sector for ranking and UCB -82,431 -109,211 unsatisfactory evaluating banks Cooperative -655 -8,226 unsatisfactory based on indicators Total -83,086 -117,437 of both condition and trend, which Private Foreign was used to analyze Stanbic 7,603 2,424 satisfactory Uganda's fifteen Barclays 2,373 -495 unsatisfactory banks. This system Baroda 2,194 2,177 fair has particularly Standard 8,216 4,381 satisfactory helped to identify Tropical Africa 2,238 -2,250 unsatisfactory problem banks. Total 18,148 6,237 Private domestic Nile 1,937 544 unsatisfactory 2.18 On the Gold Trust 234 234 satisfactory basis of BOU's Greenland 793 1,041 satisfactory Centenary -1,025 -1,123 unsatisfactory offsite CAMEL Sembule 182 -136 unsatisfactory rating system, as of International Credit 1,474 674 fair March 31, 1995, Orient 1,290 1,290 satisfactory eight banks were Kigezi 124 124 unsatisfactory rated as Total 5,009 2,648 unsatisfactory, two were designated as Total 5 9, 9 2 9 -108,553 fair, and only the Source: BOU, Quarterly Economic Report. remaining five were deemed to have maintained a satisfactory standard of performance. The poorly performing banks included both public sector banks, four domestic private banks, and even two foreign banks. As stated in the offsite analysis, unsatisfactory banks are characterized as problem banks which need intensive and effective supervision to avoid potential crisis or total failure. CAMEL is an acronym for five major components of bank performance-capital, asset quality, management, earnings, and liquidity. Bank supervisors who use the CAMEL rating system give a bank a numerical rating (typically from one to five) for each of the five CAMEL components as well as a composite rating that indicates the banks' overall financial condition. 19 Box 2.2 One way to assess bank performance Status Trend a Core capital as a percent of risk-adjusted a Return on average total assets (profit) over the assets. last twelve months. o Net worth as a percentage of risk adjusted a Change in percentage of nonperforming loans. assets. o Percentage of non-performing loans. o Change in market share of deposits over last three months. o Cost of funds. o Exposure to latent loan migration cost. o Relative quality of management. o New paid-in capital and recent events. o Liquidity considerations. o Compliance with concentration of credit and insider lending requirements. 2.19 Liquidity. With banks now flush with excess reserves (at the BOU) competition for deposits is diminishing. Loan-deposit ratios have dropped to a new overall banking system average low of 71.4 percent as of December 1994, with liquidity substantially exceeding the 20 percent requirement. A major reason for this move was UCB's dramatic improvement in liquidity. The Cooperative Bank, although still illiquid, is improving, and now has a loan-deposit ratio of 92 percent. Loan-deposit ratios for both aggregate foreign banks (now 61.6 percent) and private domestic banks (now 60.3 percent) dropped sharply over the past year. Already the need and desire for increased liquidity is imposing a large cost on banks, as evidenced by their relatively low 13.7 percent interest income on interest-bearing funds (before deducting suspended interest) despite interest rates on lending which were substantially higher than the levels during much of 1994. 2.20 Bad loans. Comparing gross loans with the much smaller figure for loans net of provisions, it is clear that non-performing assets dominate Ugandan banks' lending experience. In December 1994, the system as a whole had over 50 percent (U Sh 122 billion) of its portfolio in non-performing loans the equivalent of a little under 3 percent of GDP. UCB had almost 80 percent of its gross loans and advances in the non-performing category. Even when UCB is excluded, around 28 percent of loans are nonperforming. Even the best-performing foreign bank had 13.2 percent of its portfolio in nonperforming loans, which would be viewed as a problem bank in most developed markets. 2.21 Credit discipline in Uganda has been exceptionally poor, with some borrowers pledging defective security for their loans, and banks being unwilling or unable to use the legal infrastructure to collect defaulted debt. There are also examples of significant insider lending, particularly among the smaller private banks, and political lending in other cases. The Sembule Group, for example, which owns in excess of 50 percent of the shares of Sembule Bank, accounts for between 35 to 40 percent of that bank's total loans. 2.22 One response to UCB's bad loan problem has been the establishment of a Nonperforming Assets Recovery Trust (NPART). Established by law in late 1993, NPART's Board of Trustees is appointed from the private sector and its small management team is recruited internationally. The law requires a one-time transfer to NPART of all UCB loan assets that are nonperforming (effective January 31, 1995). The timing and precise terms (including the writing down of loan value) of the transfer are anticipated to be finalized in July 1995, but until then, the loans remain the responsibility of UCB. NPART will take responsibility for these loans and-using all available legal instruments, including special provisions outlined in the law that established NPART-collect the maximum possible amount of debt owed by Uganda's defaulting borrowers. 20 2.23 Intermediation costs. Ugandan's high proportion of bad loans and resultant substantial loan-loss provision costs-together with labor-intensive work practices, inefficient operational procedures, and high administrative overheads-has led to exceptionally high intermediation costs across the sector as a whole (Table 2.3). The second column of Table 2.4 contains equivalent ratios considered to be good performance for a medium-sized bank in a developed economy. Although the two sets of figures are not directly comparable (the market situation in a less developed country like Uganda is very different from that in a developed banking market),Table 2.4 nonetheless provides a benchmark for assessing bank performance. Table 2.4 Ratio of key performance indicators to total assets (as of December 1994) (percentage) Performance Expense Uganda Commercial banks Developed market comparator Operating Expenses Staff 5.3 2.0 Other 4.9 1.5 Total 10.2 3.5 Loan loss provisions 8.1 0.5 2.24 Banks' combined noninterest expenses in Uganda are around 10 percent of their total asset value, as compared to 3.5 percent that is typical in a developed market setting. Staff expenses account for about half of this cost. In 1993 total non-interest bank operating expenses in Uganda were 10.5 percent, a fraction higher than in 1994 (Table 2.4). Five years ago, however, bank operating expenses were far higher. 2.25 In Uganda's relatively segmented and uncompetitive market, banks have been able to pass on high operating costs to customers in the form of unusually high interest spreads. Over the past year the commercial banking system earned an average 13.7 percent on interest-bearing assets (before suspended interest). Its interest expenses (as a proportion of interest bearing liabilities), moreover, were a low 3.7 percent for borrowed funds. The gross interest spread was therefore 10.0 percent during 1994, more than twice that found in a well-functioning developed system, where the equivalent spread would typically be less than 5 percent. 2.26 The administrative inefficiency and high bad debt expense of commercial banks in Uganda is being borne by borrowers and depositors, although even these large spreads are unable to cover the costs of commercial banking (a spread of 10 percent is inadequate to cover an 18.3 percent cost of financial intermediation). One blue chip banking client in Uganda, for example, recently paid the strongest foreign bank a spread of about 12 percent over its average cost of funds for overdrafts (about 9 percent in real terms) and was required to place 100 percent in interest-bearing assets for opening letters of credit. Clearly, if Uganda's banks cannot accommodate them with better deals than this, prime clients will soon rely more heavily on external financing. The Ugandan banking system's recent loss of substantial coffee crop financing presages increasing losses of business to external competition if spreads for better customers are not reduced. 2.27 Depositors, too, are being penalized in that they are being paid largely negative real interest rates. In 1994 the average bank cost of borrowed funds in Uganda was a negative 1.5 percent in real terms. Yet even though a number of banks have discouraged deposits, there has been no corresponding disintermediation from the banking system. Indeed, over the twelve months 2 The gross interest spread was used to measure Uganda's rates, but other measures also indicate a wide gap between the interest rates that banks effectively charged borrowers and those paid to bank depositors. 21 ending September 1994, bank deposits grew at a rate of 39.9 percent as compared to the 43 percent growth in M2. For the three months ending December 31, 1994, however, bank deposits grew at a far slower annual rate of 24 percent. 2.28 Uganda's banks have been able to maintain these high interest spreads because the market is sufficiently segmented that there are few competitors for important parts of the business. Furthermore, deposits and a significant part of loan demand does not represent profitable business and are therefore not particularly attractive. In addition, banks are forced to pursue high margin, low-growth strategies because of system-wide losses and serious capital inadequacies, as banks have practiced little marginal cost pricing. Adverse selection has also probably come into play, with better-risk borrowers tending to refrain from seeking new loans. The banks therefore offer only higher-cost loans to the remaining, lower-quality borrowers. Overall, despite the presence of a number of banks operating in Uganda, effective competition has been weak, so that high-quality borrowers are forced to pay for much of the administrative and bad-debt cost, even though these costs are incrementally low for them as individual clients. Supervision and regulation 2.29 Through its Banking Supervision Department, the BOU plays the lead role in monitoring the activities of commercial banks. It also develops the regulatory framework for the banking system. On the whole Uganda's banking system is relatively well regulated. The BOU Banking Supervision Department is doing an excellent job of monitoring compliance with regulations as well as monitoring the financial condition of Uganda's banks. The BOU has an off-site supervision function that deploys its limited number of inspection staff in a coordinated and targeted fashion. 2.30 But while the present supervision techniques have succeeded in ensuring that BOU is relatively well-informed about conditions in the banking sector, it has not been effective in preventing difficulties from arising nor in resolving problems once they have been identified. The scale of the problems now being faced by UCB and other major banks suggests that a revised and enhanced approach to supervision is necessary. In regard to regulation, technical improvements are required in such key areas as licensing, on site supervision, auditing, and in the coverage and funding of deposit insurance. 2.3 The Financial Institutions Act of May 1993 added a mandatory deposit insurance scheme covering individual deposits up to U Sh 3 million. This scheme is estimated to cover 40 to 50 percent of total deposits and 90 percent of deposit accounts by number. But while commercial banks have paid in their U Sh 568 million initial payment, the Government of Uganda (GOU) has not paid in its required U Sh 2 billion share which was due in March 1995. The scheme has therefore not yet been activated although banks are obligated to pay in 0.2 percent of their deposits each year, amounting to roughly U Sh 700 million in 1995. Assuming that both the GOU and commercial bank shares are paid in, the fund would have about U Sh 3.3 billion to cover insured deposits which are estimated at U Sh 21 to 23 billion for major problem banks (excluding UCB). BOU is hesitating to take tough action to close several major problem banks, in part because it cannot afford the potential cost of indemnifying depositors. 2.32 In addition, because of substantial historical emigration, deaths through disease, and an impaired higher education system, Uganda has an inadequate number of qualified bankers. Most private domestic banks are too small to provide in-house training facilities, and the inability to find good managers therefore poses a serious problem. Limited managerial capacity and inadequate control systems have undoubtedly contributed to these banks' poor performance, and added to the high proportion of non-performing loans in their portfolios. 22 Issues and options 2.33 There are several ways in which to view the commercial banking issues which are now being faced. From a policy perspective, there are at least three broad dimensions to be considered: * Dealing with UCB. UCB is government-owned, operates in a public sector tradition, has an unsound financial structure, continues to lose money, and its dominant market position distorts competition. It plays such a large role in the system that whilst it must be considered as part of an overall strategy, specific measures are needed to focus on its individual problems. * Resolving difficulties with other problem banks, and improving overall performance. Some foreign and private domestic banks appear to be in serious financial difficulties. Their unsoundness poses a threat to the credibility of, and confidence in, the banking system. Beyond this, banks customers are not being well served by widespread poor performance and high costs. * Preventing a recurrence of the present problems. Ensure that bank supervision not only monitors performance but enforces compliance with regulations. 2.34 In developing a coherent strategy to address these broad issues, account must be taken of the linkages between them. The sections below describe specific concerns which fall within the broad issues, and present some outline options from which the strategy should emerge. Given its importance, the discussion of the specifics of UCB is postponed to the following chapter. Financial stability and problem banks 2.35 As mentioned above, eight banks, accounting for 67.5 percent of deposits (as of December 1994), have been classified by the Bank of Uganda's Supervision Department as problem banks. Reasons for such poor banking system performance include the traditional lack of competitiveness in Uganda's banking system, which has led to extremely high noninterest expenses (10.2 percent of average total assets for the system as a whole and 7 percent for foreign banks). Now that Uganda is more stable politically and economically and, hence, a better credit risk, the system faces increasing competition from banks abroad. Uganda's domestic banks have suffered from the liberalization of the economy that has further facilitated external competition. In 1994, provisions for bad debt plus suspended interest were a high 8.1 percent of Uganda's average total bank assets, which only added to the problem of having poor credit discipline and has led to high bad debt expenses. Finally, Uganda's financial sector is plagued by chronically poor bank management. 2.36 Financial instability on this scale undermines confidence in the system as a whole. A collapse of one or more banks, furthermore, could lead to significant effects on the real economy. Depositors could face substantial losses, and existing and potential investors, businesses and other customers could refrain from using the banking system. Measures are certainly required to deal with the underlying causes of this distress. But what options are available in the short term to tackle the problem? 2.37 As the owner of UCB, the government is well-placed to enforce change. * Government could restructure UCB but maintain public ownership, although efforts made along these lines over the last three years have not been successful. * Government could privatize UCB with the aim of encouraging the new owners to inject capital and introduce new market discipline into the restructured organization. * Doing nothing about Uganda's other problem banks would mean refraining from directed action in favor of encouraging existing owners to improve the health of their 23 banks by injecting new capital. But such an approach is not likely to meet with success and could even cause the problem to worsen. o At the other end of the spectrum, banking regulations could be strictly enforced, which could well result in the closure of one or more banks. While this would send a powerful signal for financial discipline within the sector, it could be costly to enforce this regulation in terms of compensating depositors. It might also prompt customers to lose confidence in Ugandan banks. o An intermediate approach would be to try to enforce or engineer rescues, mergers, or other measures designed to turn around problem banks within a specified timetable. 2.38 Costs will inevitably be incurred in improving the financial position of UCB and other problem banks whether through full or partial recapitalization or by closing banks and compensating their depositors. The cost will fall on the government, existing owners, new investors, or depositors; depending on the arrangements made for each case. Ownership and competition 2.39 While ownership is not in and of itself the main determinant of bank performance, publicly owned banks have generally performed poorly in Uganda. Public ownership reinforced a bureaucratic rather than a market-based approach to doing business. Financial discipline has been weak, and there has been little accountability for management. Publicly run banking operations have therefore tended to be high in cost and low in quality. Public ownership has also led to direct government interference, particularly in UCB which has encouraged lending not justified by commercial criteria. Finally, the negative influence on UCB has spread, weakening the performance of other banks in the system. 2.40 Rather than focusing on its regulatory role, which would be consistent with its current policy in other sectors of the economy, the government should withdraw from banking ownership and from direct involvement in the banking business. Privatizing UCB would be a major step forward. Government is currently exploring ways to dispose of its shareholding interests in several banks. In particular it is negotiating with existing shareholders (who have pre-emption) about the possibility of wider ownership through sales of these shares to the Ugandan public. 2.41 Beyond the question of bank ownership, several factors affect competition in Uganda's financial sector. Many banks, for instance, are too small to compete with UCB, which has the advantage of size and furthermore, parastatals, government departments, and even municipalities have been instructed (at least informally) to bank with UCB. Entry into the financial sector has also been too liberal, accommodating new banks with weak track records and questionable management capabilities. Furthermore, tax privileges were granted under the 1991 Investment Code to new foreign banks, giving them a competitive advantage at a critical time when existing banks are struggling. Finally, it has proved difficult to arrange the purchase or merger of problem banks so that weak institutions are not weeded out. The inadequate scope and strength of the deposit insurance scheme, moreover, discouraged the forced closure of failing banks which would presently have to be done at the expense of depositors. Bad loans, high costs, and liquidity 2.42 The prevalence of nonperforming loans in all the major Ugandan banks' portfolios reflects not only Uganda's weak bank management and procedures, but more generally its poor credit disciplinary environment. This issue is of overwhelming importance because it contributes largely to high intermediation cost, bank's reluctance to aggressively increase their market share, and the banks' poor financial condition. 24 2.43 Although it is tempting to narrow interest spreads by regulating interest rates for lending and borrowing, such direct interference has not worked in the past since it introduces distortions into the market and fails to address underlying problems. High interest spreads reflect such underlying difficulties as the prevalence of bad loans in the system and will narrow only when these fundamental issues are resolved. 2.44 The current environment of bad loans, extremely poor credit discipline, and the recent toughening of loan classification requirements all contribute to banks' growing reluctance to lend. Taken together with significant inefficiencies in the country's money, interbank markets, and payment system, this reluctance has produced high liquidity positions in the banks. A problem is emerging, and could grow more serious if these trends continue, whereby the supply of safe instruments available for investing excess liquidity is far less than the demand for these instruments from the banking system. A response is required in terms of providing an adequate supply of short-term and/or liquid interest bearing instruments to the banking system. BOU should ensure that existing institutional and policy constraints to the supply and use of such instruments are eliminated (see chapter 8, on money market development). Banking regulation and supervision 2.45 Regulations and supervision both need improvement particularly in the areas of licensing, auditing requirements, provisioning, insider lending, loan concentration, and capital adequacy. 2.46 Licensing. Although Uganda's regulations on financial sector licensing are adequate, they are not well implemented. The BOU is given power under the 1991 Act to make decisions on licenses, but the criteria against which it makes its final judgments are not always clear and well- defined. The Ministry of Finance (MOF), moreover, tends to form its own opinion which is often different from that of the BOU, and has used its influence to promote the awarding of new licenses. Of the five new financial institution licenses issued in 1994, for instance, several went to applicants with questionable credentials. The integrity of the system can only be assured if the BOU is left unfettered to enforce the regulations. 2.47 Auditing requirements. The BOU has issued sensible instructions to auditors, plans to keep a list of acceptable auditors, and has instituted a tripartite review (beginning with the 1994 statements) for each bank and its auditors before accounts can be accepted. Audits are to focus on loan classification and provisioning. Accounts are due within four months of the fiscal year end. 2.48 Until now, there has been no required format for balance sheets, income statements, or notes to financial statements. Individual audit reports have varied greatly, making it difficult to compare different banks or each bank's BOU reporting against its financial statements. Indeed, many statements are misleading. 2.49 Ugandan banks' accounting standards and capability are also highly variable. IDA's institution capacity building operations, therefore, include support for the development of accounting standards and services in Uganda. 2.50 The most serious issue involves the lack of congruence between provisions and suspended interest in accord with BOU regulatory requirements. Agreements with the Uganda Revenue Authority (URA) allow the banks to count, as a tax deductible expense, only one-third of the bad debt expense associated with the 1993 year-end portfolio annually during 1994, 1995, and 1996 (in addition to 100 percent of the bad debt expense associated with new lending). The Uganda Investment Authority (UIA) has reportedly indicated that any bad debt expensed beyond those amounts during 1994 and 1995 will permanently lose its tax deduction. BOU has agreed that banks should include in their financial statements only one-third of that original provision during 25 1994, a cumulative two-thirds of it during 1995, with the full provision reflected only in the 1996 accounts and thereafter. This represents a de facto, but not a de jure, amendment of the provisioning requirement by allowing it to be implemented gradually over a three year period. The result is that audited financial statements are wrong and highly misleading. Indeed, for several banks, the statements show positive core capital and profits when the internal BOU supervision reports and these banks' reporting to BOU show significant negative core capital and substantial losses. The unfortunate effect is that audited financial statements are leading many banks and their owners to the false conclusion that their financial position is not as serious as indicated by the BOU. 2.51 Provisioning requirements. The BOU has imposed detailed, relatively sophisticated asset classification and provisioning policy on Ugandan banks, which, moreover, is being effectively enforced through an aggressive inspection program. The policy conforms reasonably with international norms in all but three ways. 2.52 Aging requirements are lenient by international standards with loans being classified as "substandard" after they are in default for six months, as "doubtful" after one year, and as "loss" after two years. Uganda's regulations gives the BOU an option, however, of cutting these elapsed times to three, six, and twelve months, respectively. 2.53 Under international norms, provisions are taken net of the estimated realizable value of collateral. Uganda's provisioning requirement is 100 percent of capitalized or accrued interest plus 20 percent of all loans classed as substandard, 50 percent of loans classed as doubtful, and 100 percent of loans classed as loss. These provisions against loans apply even if a bank has strong claims on collateral that exceeds the value of the loan. This departure from international practice, however, may be justified in view of doubts as to the reliability of the value of collateral given the dysfunctional legal infrastructure for debt collection in Uganda. Banks have the right to apply to the BOU for exceptional treatment of collateral in individual situations, but in fact the BOU has rarely given permission for such exemptions. 2.54 At present, loans classified as impaired in Uganda can only be reclassified as better performing if the bank collects an amount exceeding the maximum default in interest and principal, regardless of how much a loan has been restructured, rescheduled, or honored in terms of rescheduled payment schedules, etc. 2.55 Although Uganda could do more to change its regulations to conform with international best practice for provisioning requirements, or to institute more active supervision of their enforcement, it is not clear that existing deficiencies in provisioning policy have had a significantly negative impact. Arguably, since the aging requirements are relatively lenient and the size and enforcement compliance of provisioning requirements are relatively tight, they do, in fact, offset each other. Given all the other changes the banks now face and their relatively weak financial condition, the most pragmatic option maybe to enforce any changes in provisioning policy at present and to put off taking coordinated action until a later stage. 2.56 Insider lending. Insiders are defined as those who control 25 percent or more of the stock, directors, and executive officers (down to the deputy general manager level). The aggregate insider lending limit in Uganda is 25 percent of the core capital with loans to executive officers limited to the lesser of 5 percent of core capital or U Sh 50 million. Existing excesses as of September 1993 should be reduced by half of the excess over core capital per year until compliance is reached. For example, a loan amounting to 100 percent of core capital (that is, 75 percent above the maximum limitation) should be reduced in volume such that the loan is only 62.5 percent of core capital by year-end 1995. 26 2.57 Detailed reporting on insider lending is received promptly on a quarterly basis and is effectively monitored through inspections. Yet there appear to be insider lending violations at a number of banks, including an extremely serious violation at one major problem bank and a lesser- but significant-violation at another. 2.58 It is noteworthy that banks have largely failed to comply with requests to reduce insider loans in part, because no meaningful penalties are being assessed. Some banks , moreover, have minimal or negative core capital. A literal interpretation of Uganda's concentration limit requirements suggests that these banks should offer no insider loans, and indeed, no loans of any kind. Since insider lending undermines the public's confidence in the strength and integrity of the country's banking system, Uganda should seek to control it: * The BOU should tighten its enforcement of insider lending regulations by imposing well-defined penalties against transgressors. * The BOU could also offer short-term transitional arrangement during which defaulting banks could make adjustments. 2.59 Concentration limits. According to current regulations, loans are required not to exceed 25 percent of core capital to one credit risk (although in the transition period existing loans are required not to exceed 100 percent of core capital through December 1995 or 50 percent of core capital through December 1996). While these limits and transition periods seem appropriate, some firms are still borrowing far in excess of the 25 percent individual limit. It appears that compliance is now generally improving, with one of the most egregious cases now reduced to 64 percent of core capital. In certain instances, infringements of concentration limits do not appear to be harmful at present. For example, one case involves good credit risk and loans from a relatively well- functioning, foreign-owned bank. Still, in the main, Uganda needs to enforce the regulations without delay to ensure the stability of the system over the long term and open up the large-client market to greater competition. 2.60 Capital adequacy requirements. The Basle Committee guidelines suggest that Uganda's banks maintain 4 percent of their risk-adjusted assets in core capital. Requirements for banks' total net worth as of December 31, 1996, are 8 percent of risk-adjusted assets in core capital plus a minimum capital requirement of U Sh 500 million for local banks and U Sh 1 billion for foreign banks. Interim targets for banks that do not presently meet the 1996 target have not been spelled out, except to the extent that they are mentioned in performance contracts for problem banks. A number of banks are not in compliance with the 1996 requirement. This financial weakness poses the risk of banks being forced to close with inadequate capital to cover their outstanding obligations. 2.61 Until recently, the BOU had placed more emphasis on the minimum absolute capital requirement for banks than on such prudential requirements as and how much capital is on hand, measured as a percentage of risk-adjusted assets. A brief history of Uganda's financial regulations illustrates why even profitable private banks have had trouble meeting this minimum paid-in capital requirement. 2.62 Uganda raised its capital requirement one step from U Sh 20 million to U Sh 500 million, even for existing banks. For banks established before 1987, initial capital was paid in at the rate of U Sh 14 to the dollar as compared with the present rate of U Sh 900 to the dollar. The real value of the initial capital, therefore, has been effectively wiped out by the devaluation. Banks, moreover, were subjected to a 30 percent wealth tax early in their history, including a tax on their paid-in capital. Finally, because there is no meaningful capital market in Uganda, options for raising new paid-in capital are few-particularly if a banks is loss-making. All of these conditions 27 make Uganda's minimum capital adequacy requirements difficult to meet. To enforce the capital adequacy requirements in a practical way: o Uganda needs to set clear interim capital adequacy targets for banks (through 1995 and 1996) as a path to meeting the full conditions. o With respect to the minimum core capital requirement, a temporary (two to three years) dispensation could be granted to banks which meet capital adequacy requirements (see paragraph above on capital adequacy requirement). 2.63 Other regulations and possible solutions. Uganda's liquidity and cash reserve requirements seem fully satisfactory, and all but one bank are presently in full compliance. Requirements imposed on banks to report on their operations are also essentially sound. Within BOU, however, some refinements could be made to improve inspection and supervision reporting: o Limited resources for inspection suggest that less attention should be given to inspecting the strongest banks to allow for more frequent inspections of the weaker banks. As of now, several problem banks have not been re-inspected for some time. o The off-site monthly bank reporting formats, while generally comprehensive, need refinement. Reform Strategy 2.64 Uganda's financial sector reform strategy needs both to tackle immediately pressing difficulties-such as the threat of bank collapse and the looming crisis of public confidence in the system-to establish a structural foundation that will ensure the system's long-term health. Over the long term, the system will need far greater financial discipline and more vigorous competition within the sector. The following strategy recommends that Uganda: o Remove government ownership and direct operational involvement in the commercial banking sector. o Maintain confidence in the system by facilitating prompt turnarounds of problem banks, or otherwise managing their orderly exit from the system. o Create an improved banking environment, in which market forces can properly encourage competition and discipline amongst private sector banks, by supporting enhanced management capacity and a more effective regulatory framework. o Enhance the standards of bank supervision, and actively enforce measures to ensure compliance with market regulations. Remove government from bank ownership and operations 2.65 Since government ownership is overwhelmingly concentrated in UCB, privatizing UCB should be the first priority for action. Rapid privatization would send a powerful signal of intent and encourage the private ownership of banks and provision of banking services in Uganda. 2.66 As mentioned above, the government is currently negotiating to divest its minority holdings in foreign banks to private Ugandan shareholders with the aim of widening ownership. The existing shareholders have pre-emptive rights which they prefer to exercise. Once this problem is resolved and divestiture accomplished, funds from such divestiture could be placed in the deposit insurance fund or used to increase UCB's paid-in capital. New owners of share blocks (be they Ugandan or foreign) should then be able to re-focus Ugandan banking operations to better serve new market niches. 2.67 The removal of Government ownership will automatically reduce the opportunity and incentive for Government to interfere in banking operations. Government should refrain from using any other inappropriate channels to influence the commercial banks. In particular, the 28 Ministry of Finance should not seek to influence operational decisions made by BOU, but rather, should clearly guarantee the operational independence of BOU. Where the Government believes there is a policy interest to be served by the commercial banking system, it should establish transparent mechanisms to achieve stated objectives. For example, if the Government believes that small (uneconomic) rural bank branches must be kept open as part of the country's "public utility infrastructure", then an explicit subsidy-in the form of a service level contract-should be offered through a competitive tender to banks (including a privatized UCB and the Cooperative Bank) to provide those services. The Government should discontinue its existing practice which effectively underwrites the poorly identified losses made by a sole supplier (UCB) of rural banking services. 2.68 In those cases where a policy interest is to be served by the commercial banking system, the government should establish transparent mechanisms designed to achieve stated objectives. For example, if the government wishes to keep small and uneconomic rural bank branches open as part of the country's rural branching infrastructure, an explicit subsidy should be offered to provide those services. Such subsidies should be evaluated as part of a broader review of options to provide rural financial services (see chapter seven). By contrast, the government should discontinue its existing practice of underwriting losses that are both poorly identified and made by the sole supplier of rural banking services, that is, the UCB. Maintain public confidence in the system 2.69 The key elements of action under this theme are: * Resolve the threat of financial instability posed by the problem banks. * Provide appropriate protection to depositors over the long term through a well- functioning deposit insurance scheme and over the shorter-term by additional government support. * Reduce banks' burden of bad loans, in part by changing public attitudes regarding the abuse of credit. 2.70 A number of steps are needed to tighten market regulations. Core capital requirements, for instance, need to be applied to well-run and problem banks alike. Steps to assure solvency over the long term should be integral to any sector strategy designed to maintain confidence in the system. 2.71 To improve financial discipline and minimize the costs of bank failure and liquidation, the legal environment for enforcing the transfer and liquidation of collateral and other assets pledged by defaulting borrowers needs to be improved. The Non-Performing Asset Recovery Trust (NPART) was established by the NPART Act for the specific goal of accelerating and strengthening the recovery of funds from UCB's non-performing loans. The government decree which was the basis for the creation of NPART should make it possible to take stricter action against defaulting borrowers. At the same time, the transfer of non-performing loans from UCB will facilitate the privatization of UCB. NPART will take over all nonperforming loans (as of September 30, 1994) and the Government intends to replace them with government bonds so as to cover UCB's negative net worth prior to its sale. It is recommended that NPART commence operations as soon as possible and the management of NPART should begin to assess the volume of recoveries which can be realized 'out of court' and identify those cases which require litigation using the special tribunals to which NPART will have access. In the process of transferring UCB's non-performing loans to NPART, it is recommended that there must be an equivalent value transfer of government bonds to UCB to insure that UCB's net worth is not adversely affected by the transfer. Full recapitalization of UCB would have to wait till the actual negotiations with the potential buyer. 29 2.72 The urgency for NPART's commencement of operations should also be viewed in terms of the official signal it will send to other defaulting borrowers, concerning the resolve of the government to improve the degree of financial discipline in Uganda. Furthermore, decisive action in this front will also facilitate GOU's ability to assess the foreign assistance needed to implement the successful sale of UCB. Legal and other issues raised in the transfer of UCB's non-performing assets to NPART and discrepancies in asset valuation (what loans are to be transferred and when), need to be addressed immediately. 2.73 As mentioned above, by the end of March, 1995, eight banks in Uganda were categorized as problem banks by the BOU. Since then, one foreign bank has been removed from the group on account of improved performance. Of the remaining seven, COOP is being restructured with the direct advice and assistance of USAID, whereas UCB is treated separately in this report. Of the remaining five banks which all appear to be insolvent with negative core capital, BOU has intervened by taking over the management of two banks (Nile and Sembule), and has initiated a detailed portfolio audit of the remaining three to diagnose weaknesses in their balance sheets. The issue of how to handle the problem banks should be dealt under two categories: (i) banks whose management has been taken over by BOU, and (ii) banks in which the BOU has still not directly intervened. 2.74 In the case where insolvent banks have already been taken over (Nile and Sembule), unless existing shareholders are willing to inject new capital into the bank, their ownership rights should be revoked. By taking responsibility of an insolvent bank where the owners have effectively lost their equity stake, the government becomes de-facto owner and must view its responsibility in this way. If existing shareholders are unwilling to recapitalize these banks, BOU has only two alternatives: liquidation or keeping them open through injection of funds as part of a deal with new owners. If BOU/MOF wants to minimize direct costs, paying depositors up to the limit defined by deposit insurance (U Sh 3 million), and then liquidating the bank is the only option.3 Alternatively, for political or other considerations, if the Government wishes to keep these indigenous banks open and is willing to inject new funds (in the form of cash, bonds, or other guarantees) to cover the full amount of negative worth in these institutions, there are two options available: (i) look for a new owner who is willing to inject new capital into the bank, and (ii) look for an appropriate healthy bank with which to negotiate a merger. Whichever option the Government decides upon, it needs to take into account that other banks in which it may intervene in the future will expect comparable treatment. 2.75 As for banks in which BOU has not yet intervened, the necessary fact finding work (through portfolio audits, on-site inspections, etc.) needs to be completed before intervention. Furthermore, the BOU in consultation with the MOF, should also decide on an acceptable plan of how to handle problem banks. What is crucial, again, is that this plan of action should be decided The costs of liquidation are direct and indirect. Direct costs include: (1) the cost to the Government of compensating depositors - be it partial coverage as stated in the current insurance scheme, or full coverage, and (2) costs to be borne by large depositors in the event of partial coverage. One must note, that the direct costs to the Government may be negligible in the case of partial coverage when revenues from liquidating bank assets are also taken into account. The indirect costs of liquidation are: (1) the loss of domestic confidence in the banking system especially if insurance coverage is partial, and (2) the possibility that international agencies which have lost deposits will become averse to future lending to Uganda. The benefits of liquidation are: (1) it will send a very strong signal to existing banks, telling them that poor performance will be credibly punished with liquidation, and (2) the loss of small banks will not hurt Uganda's banking network since it is accepted that urban Uganda is overbanked. 3 This is more likely in the case of Uganda, where the existing deposit insurance scheme only covers deposits up to Ush 3 million. Since the two banks have been taken over by the authorities but are nevertheless still operating, any expectation of liquidation would drive existing depositors to hedge against possible losses either by withdrawing deposits over Ush 3 million or by breaking up larger deposits into smaller 'insured' accounts. Furthermore, new depositors would not open 'large' deposits accounts. In effect, the longer the period of indecision, the larger the government's liability. 30 upon before the intervention so as to keep the actual period of intervention to a bare minimum. In the case of Nile and Sembule, the government's indecision concerning the fate of the two banks after the actual intervention - while allowing them to accept new deposits - has simply increased the government's exposure and will most likely increase the fiscal burden till that point when the government does in fact decide.4 Failure to impose sanctions on poorly performing banks will only increase the likelihood of direct intervention, whereas indecision after intervention is a costly mistake. 2.76 The funds available in the deposit insurance scheme in Uganda are seriously inadequate to fulfill their role. Commercial banks in Uganda are currently required to place 0.2% of their deposit base in the fund. Despite the inadequacy of this requirement, the fund was only U Sh 568 million strong since the Government has yet to place its share of U Sh 2 billion into the fund. Given the insolvency on the banking system, even if the government were to pay its share into the fund, the deposit insurance scheme needs to be restructured to act as a proper buffer against insolvency or liquidation of problem banks.5 At present, only deposits under U Sh 3 million are covered by the insurance scheme in the event of bank failure. 2.77 Beyond the immediate problem that the deposit insurance scheme is currently under- funded, it is important to review the design of the scheme in relation to the two longer-term goals which need to be met by any such scheme: (i) maintaining domestic confidence in the banking system, and (ii) avoiding moral hazard in bank lending on the part of bank managers. The first issue is especially relevant in Uganda given the shallowness of its financial system. Since Uganda has one of the lowest M2/GDP ratios in the world, closure of banks accompanied by partial coverage of deposits will not only carry the risk of disintermediation, but could encourage the flight of financial savings to overseas banks. The second goal is relevant in terms of insuring that the deposit insurance scheme is designed such that bank lending practices are socially optimal. Since commercial banks in Uganda vary considerably in their financial health, having all banks pay a fixed premium on the insurance scheme will encourage weaker commercial banks toward adopting risky lending practices which in turn will exacerbate the weakness of the financial system. There is a medium-term need to design an insurance scheme where premiums are linked to some kind of risk rating system which reflects a bank's overall financial health. Furthermore, there should be a strong and credible signal that in the event of bank failure, the incumbent bank management will be replaced. Only on the basis of a well designed deposit insurance scheme which gives the right set of incentives to bank managers, will Uganda be able to develop a healthy and sustainable banking system. 2.78 The recommended approach is therefore to deal privately with the problem banks by providing each with individual instructions as to their obligations over the short and long-term if they are to meet prudential standards and restore their financial stability. Those instructions would include specific targets which the BOU would have to aggressively enforce. In broad terms, instructions to the problem banks would: 4This is more likely in the case of Uganda, where the existing deposit insurance scheme only covers deposits up to Ush 3 million. Since the two banks which have been taken over by the authorities are nevertheless still operating, any expectation of liquidation would drive existing depositors to hedge against possible losses either by withdrawing deposits over U Sh 3 million or by breaking up larger deposits into smaller 'insured' accounts. Furthermore, new depositors would not open 'large' deposits accounts. In effect, the longer the period of indecision, the larger the government's liability. Based on the current instability in the banking system, a reasonable short-run funding target would be 5 percent of deposits. This is based on the insured deposits which need to be paid off if the three most problematic banks were to be closed immediately. 31 o Encourage the banks owners to inject new resources to strengthen the banks' financial positions and to introduce new management disciplines to improve their operating performance. o Offer practical help to support these rescue efforts. o Set a time limit for agreed upon measures to demonstrate results. o Help the government develop contingency measure so that orderly and immediate closure could be enforced for those banks that remain in the problem category at the end of the period. o Provide specific actions based on a rapid portfolio audit for each of the problem banks which the BOU has recently arranged through independent internal auditors; such an audit would provide a more accurate assessment of each banks' lending exposure, insider lending, net worth, and capital adequacy. 2.79 Typically, problem banks would be granted up to one year to demonstrate a significant resolution of their difficulties. Instructions to them would set interim target so that their performance could be monitored during this period. Targets would be set for the injection of new capital to meet core capital requirements, the elimination of insider trading; and diversification of credit. o After deducting BOU's required provisioning, interim core capital minimums could be set at: 2 percent as of December 31,1995, 3 percent as of June 30, 1996, 4 percent as of December 31, 1996. o Banks that fail to meet the interim capital requirements should be strongly considered for immediate sale or closure. Banks not meeting the full 4 percent requirement as of December 31, 1996, should be potentially subject to closure. o If, due to extenuating circumstances, such a bank is not closed , financial penalties of, say, 1 percent of the shortfall per month should be assessed. 2.80 Financial penalties should be imposed on banks failing to meet interim targets. Practical help financed from technical assistance would include technical help to problem banks. At least one and probably two experts at bank work-outs, banking management, and financial engineering should be attached to the BOU Banking Supervision Department as soon as practicable. These specialists should work only on the major problem banks, tackling one or two simultaneously in an intensive manner. One expert could be made available to work full-time within one bank for a number of months, possibly taking over the bank's management on behalf of BOU. This would have the added advantage of preventing management from siphoning off funds in periods of distress. A second expert could assist BOU in the detailed analysis of problems and creation of work-out arrangements for other troubled banks. 2.81 A public announcement of these actions could anchor a wider campaign to strengthen customer confidence in Uganda's banking system. Further work is required, however, to develop an effective response for the problem of unprotected depositors. Uganda needs a plan showing how insured depositors' claims could be financed should one or more problem banks be closed since even when the deposit insurance scheme is paid up, it is not fully funded. Government will therefore have to specify the extent to which unprotected depositors should take a loss or be protected by government support. To assist problem banks and restore confidence in Uganda's financial system, several actions must be undertaken to improve the debt recovery environment: o Establishing a credit information bureau. 32 * Sending information on defaulting borrowers, whose loans will have been transferred to NPART to all banks. * Establishing a commercial court to process decrees, execute loans, and process foreclosures efficiently. * Making sure that government pays its overdue bills to suppliers, particularly those who have been unable to pay their nonperforming loans with banks. * Allowing commercial banks to transfer nonperforming loans for collection at a fee within the next three years to NPART. This should, however, be contingent on an adequate recovery performance of NPART on UCB's nonperforming loans. Improve the banking environment 2.82 To improve its banking environment, Uganda will have to build the sector's management capacity, encourage competition, and enhance the regulatory framework for banks. Banking management capacity can be enhanced over the long term by introducing training facilities and programs for commercial bank staff and making these facilities equally available to private and public banks. Specifically, a training fund could be established at the Bankers' Institute to support short-term courses, in-house training advisers, and a scholarship fund for overseas training for selected bank officers. The fund would only support programs that could at least partially recover their cost. Three simple and direct measures could be adopted to encourage competition: * Parastatal and other public sector entities should be instructed that they are free to bank with any nonproblem bank of their choice, based on which bank best fits their service and financial needs. * Consideration should be given to more strictly enforce limits on maximum credit to one borrower, even when that borrower is very sound, so that a wider range of banks can bid for large-client business, which is currently concentrated in UCB and the larger foreign banks. * Licensing criteria should be tightened to place greater emphasis on bank owners' and managers' resources and reputations, quality of business plan, etc. This would stem the flow of poorly qualified entrants into the financial sector. Newly approved applications should also be canceled if applicants fail to comply with the new requirements. 2.83 Financial soundness also requires better enforcement of existing regulations as well as structural changes to ensure that the financial sector has access to accurate, useful financial information. Bank audit reports should be prepared in accordance with the new, BOU-provided format for accounts from 1994 onwards. Enhance the standards of bank supervision 2.84 The BOU should enforce insider lending requirements for all banks but especially for problem banks far more strictly. Specific instructions should be given to problem banks without delay. For each bank, BOU should: * Prepare a beginning list of insider loans in ex.-s of requirements. * Prepare an annual amount, a set percentage of core capital, as the repayment target for each of these individual loans. 33 o Send out a reminder now about the penalties for noncompliance with the interim requirement as of December 31, 1995, giving a 30-day grace period from the date of the letter for compliance. o Assess a penalty of, say, 2 percent a month on insider loans that exceed the interim requirement until the bank is in full compliance. In the case of banks with extremely little core capital, a transitional limit of 1 percent of risk-adjusted assets can be substituted for the 25 percent of core capital usually required. .-j Enforcement of interim requirements. While no changes should be made in loan concentration limits the interim requirements should be strongly enforced. To force the larger and better-risk borrowers to deal with more than one bank, to enhance competition in the marketplace, and to reward those banks that are most soundly capitalized, BOU should not allow any exceptions to this policy. 2.86 In situations where specific exceptions have already been approved, those exceptions should be capped at the present lending level, and a gradual program should be agreed upon to move the bank toward compliance. The BOU should send a clarifying letter: o Stating that no new facilities should be granted in excess of 25 percent of a bank's core capital, or 1 percent of risk-adjusted assets in the case of banks with little core capital. o Listing all existing loans that exceed the limit. o Reiterating the transitional loan requirements associated for year-end 1994 and 1995. o Announcing the 1 percent a month penalty on that portion of loans that exceeds the transitional limitation, to being 60 days from the date of the letter. 2.87 BOU inspection reports. The BOU inspection reports also need to be improved: o A summary at the front of each inspection report should list all demands to change dates for compliance. o A cover letter should ask the bank to agree to make these changes and to call for monthly reporting on compliance. o The summary should list all previous demands for changes and the present status of compliance for each. For a more detailed description of the analysis from this data source, see Annex 1. 2.88 Other actions. Other actions in the supervisory regime could include: o Assisting recapitalization by allowing banks to count grant income as extraordinary income and core capital. o Moving toward international standards for loan-loss provisioning. Changing the Uganda Revenue Authority agreement to make all provisions deductible in the accounts for 1994 and thereafter to require full provisioning in the statements from that date forward. 34 Action Plan for Reforming Uganda's Commercial Banking System S6TERM (WIIN 1 YEAR LONGERTERM WITHIN 3 YEARS) Strategy element: Removing government ownership and operational involvement * Privatize UCB. * Divest government's minority holdings in other banks. * Cease MOF or other government agency * Remove implicit subsidies to banks interference in day-to-day banking operations. (particularly UCB) and undertake a study to review options for effectively providing rural banking services. * Guarantee BOU's operational independence. Maintain public confidence in the system * Transfer bad UCB loans to a fully operational NPART. * Ensure privatized UCB meets capital adequacy * Keep otherwise sound banks open if they have requirements. positive core capital and show a trend towards profitability. * Complete a rapid portfolio audit of the problem banks using independent external audits * Set and confirm measurable, time-bound * Impose financial penalties on problem banks * Immediately following any substantial failure targets for requirements (such as core capital, for any noncompliance with instructions. to comply with the performance contract, insider lending and concentration limits) for enforce the orderly closure of any remaining each problem bank. Monitor compliance problem banks. monthly and enforce sanctions aggressively. * Attach bank work-out experts to BOU and the * Develop detailed contingency plans for the problem banks to assist with management. orderly closure or merger of problem banks. * Government to pay its share of the mandatory * If feasible, establish a credit information deposit insurance payments in full. Scheme bureau within the BOU or the Bankers' must collect 1995 deposit insurance, which Institute. has been stated at 0.2 percent of deposit base 35 SH4OT TRM(WFU 6 ONTHS) MlM TM (WTHIN 1 YEA1R) T,OAP) 1P (Wfi%W3 9 A149 o Government to pay all overdue bills so that 0 Establish a commercial court for efficient suppliers can settle their own bad debt. processing of decrees and the execution of loans and foreclosures. To solve the issue ofproblem banks o Allow commercial banks other than BOU to engage NPART to collect their debt for a fee. 0 Take immediate action to liquidate or recapitalize intervened banks. To improve the banking environment o Instruct parastatals and other public sector a Establish and operate a fund to support new bodies to bank with any nonproblem bank of training facilities and programs to train bank their choice. managers. o Strictly enforce credit concentration limits. o Ensure that short-term or liquid interest- bearing instruments are available to sop up excess bank funds. o Tighten licensing criteria and revoke recently issued licenses in the event of noncompliance. o Establish interim core capital requirements, a Enforce permanent core capital requirements. counting grant income as core capital when in line with international practice o Adopt improved standards for bank audits and their presentation. To supervise the financial system carefully o Enforce insider lending rules with financial o Watch problem and unsatisfactory banks a Move towards international standards for loan- penalties for noncompliance. carefully, taking early action where needed. loss provisioning. o Improve BOU off-site supervision and inspection reports. 36 3. Reforming the Uganda Commercial Bank 3.1 Given UCB's dominant role in Uganda's banking system, its fragile financial position threatens the stability of Uganda's financial system as a whole. Hence, reforming UCB must take top priority in the financial sector strategy. Background 3.2 UCB's operating performance has been dismal over the last four years, with cumulative losses through September 30, 1994 amounting to U Sh 95 billion (US$ 107 million) (Table 3.1). The principal factors contributing to this level of losses were the need to make loan loss provisions of U Sh 68 billion, and the fact that net interest income was down 60 percent from fiscal 1991 levels because of UCB's substantial volume of nonperforming loans. In 1994 net interest income was U Sh 18 billion (or 13.3 percent of the bank's year-end assets). The decline in noninterest expense, however, reflected restructuring efforts which included closing fifty-two branches, converting an additional fifty-three branches to agencies, and cutting the number of staff in half to 1,663. Interim unaudited operating results for the four months ending January 31, 1995 reflect a loss of U Sh 1.4 billion before any loan loss provision (Uganda Commercial Bank 1995). 3.3 It should be cautioned that these results may not accurately indicate full-year profit and loss because of the inclusion of extraordinary expenses for retrenchment costs and supplementary Table 3.1. UCB's operating performance (U Sh millions) Income and Expense As of September 30 1991 1992 1993 1994 Income Net Interest 9,124 8,503 3,999 3,800 Noninterest 6,035 8,286 6,976 8,100 Expense Noninterest -12,604 -23,865 -26,526 -18,500 Total Preprovision operating income 2,555 -7,076 -15,551 -6,600 Expense Loan Loss provision -4,972 -17,150 -19,135 -27,200 Total Pretax ordinary income -2,417 -24,226 -34,686 -33,800 Source: Financial Sector Adjustment Credit Implementation Secretariat-17/11/94 Note: All figures audited except those for 1994. pay to avoid a strike in December, incomplete records of accruals of interest income and expense at some branches, and the inclusion in income of bad debt recoveries that will not recur in the future - when nonperforming loans are transferred to the Nonperforming Assets Recovery Trust (NPART). Raw interest rate spreads (between interest income and expense) are running in the region of 18 percent (with interest-earning assets of 20 to 24 percent, interest-bearing liabilities of 2 to 3 percent). Given the 64 percent demand deposit component of total deposits, it is likely that net interest margins are, if anything, above 18 percent. Because of UCB's difficulties in generating quality loan assets, however, actually realizing such net yields remains elusive. High noninterest operating expenses also continue, although reduced substantially from previous years. 37 Solvency 3.4 Stockholder equity declined steadily over the last four years reaching a deficit of U Sh 82 billion (US$ 93 million) by December 31, 1994. Excluding supplementary capital has the effect of worsening deficit equity by U Sh 24 billion to U Sh 106 billion (US$ 119 million). This exclusion arises from a revaluation reserve, which, in accordance with the conventions of the Basle agreement, may not exceed half of the total capital. While the need for recapitalization is evident, the amount and methods needed to accomplish this require further discussion. Client base and loan portfolio 3.5 Several features stand out in describing UCB's client base. In terms of inflows: o UCB handles around 40 percent of Uganda's total market deposits. o Of UCB's total deposits (about U Sh 130 billion), some 32 percent (U Sh 40 billion) are held by government or parastatals, with parastatals accounting for the largest share. o Total demand deposits of U Sh 84 billion are concentrated among the bank's top 100 clients, who hold about 40 percent of the total. Savings accounts are much more dispersed, with the top 100 clients accounting for only 2 percent of the total. o Some U Sh 22 billion of administered (government-owned) funds are held at UCB. o Kampala is regionally important to UCB, because 53 percent of the bank's total deposits are held there. In market terms, however, UCB is even more dominant in rural areas, where the bank is frequently the sole provider of services. 3.6 In terms of outflows: o Although the market share of agricultural crop financing has declined recently from 75 percent to 63 percent (off-shore sources are now providing significant resources for crop financing, especially for coffee production), it is still an important market for the UCB. o Government and parastatal advances (U Sh 100 million) accounted for approximately one-tenth of 1 percent of UCB's total loans for the year through December 1994. 3.7 Ernst and Young reviewed UCB's loan portfolio as of December 31, 1993. The report they issued in August 1994 revealed that over 80 percent of the portfolio was impaired. As much as 60 percent, moreover, could be classified as a loss (Table 3.2). In response to UCB's bad-loan problem, the government established NPART by law in late 1993. NPART's Board of Trustees was appointed from the private sector, while a small management team was recruited internationally. In a one-time transaction, all UCB loan assets that were nonperforming as of January 31, 1995 were to be transferred to NPART, which would then take responsibility to collect on defaulted debt. Nonperforming loan assets were also to be the subject of an updated loan portfolio review. But as it now stands, the process of transferal was delayed and the terms of the transfer have yet to be resolved. If the need to safeguard security and loan collection go unattended during this uncertain period of transition, the UCB runs substantial risk of incurring further bad debt. 38 Table 3.2. Classification of UCB's loan portfolio (U Sh millions) Loan type Loan classification Good Watch Sub-Standard Doubtful Loss Group percent Mortgage 423 220 470 2,177 3,334 7 Development 979 246 619 3,083 26,164 33 Commercial 11,590 4,617 4,132 8,744 27,262 60 Total 12,992 5,083 5,221 14,004 56,760 100 Class percentage 14 5 6 15 60 Source: Ernst and Young, Review of UCB's Loan Portfolio as of December 31, 1993. UCB restructuring 3.8 In the face of poor performance the UCB has begun closing branches. In fiscal 1992, the UCB had 188 branches. During fiscal 1993, these were reduced to 169 units. In the last fiscal year, 32 branches were either closed outright or merged with other units, and 53 branches were converted to agency status offering limited service. Presently, the UCB has 84 full branches operating and 53 agencies. 3.9 Under the on-going restructuring program, a consulting firm was appointed to provide experts in key managerial positions within the bank, including that of chief operating officer. This team of four has helped the UCB board deliberate the closing of eight more branches and transfer their deposits, advances, and clients to other units. Deposits in the branches to be closed amount to U Sh 5 billion, and it is hoped that attrition can be held to 10 percent. Operational costs at these branches approximate U Sh 577 million, of which staff costs account for 71 percent and the annual lease expense is less than 2.5 percent. The UCB board is also considering the closure of all fifty- three agencies, whose total annual costs amount to U Sh 1 billion, of which U Sh 762 billion are nonstaff expense. The UCB anticipates that this combination of branch and agency reductions will realize savings of U Sh 1 billion before any consideration of lease terminations or reducing any staff redundancies. The remaining staff expense associated with the closed branches is U Sh 700 million and the lease expense is U Sh 14 million. The number of staff at the main office once dedicated to supporting g the closed units has yet to be identified. The UCB's national network of brai.ches will continue although much is reduced in size. Issues and Options 3.10 Whatever route is taken to resolve the problems of the UCB, it is imperative that further losses from bad loans be contained and that no additional claims be made on government. The capital requirements of approximately U Sh 80 billion should be limited to unfunded commitments to the maximum degree possible. This is the base amount that would have to be provided by the government in cash or bonds if the loans were called. In fact, if illiquid fixed assets were placed under the auction hammer, total costs might approach U Sh 130 billion (the level of deposits as of December 1994). Economic costs associated with the loss of financial services would be even greater and more disruptive. Having to make cash available to meet depositor claims is obviously not a desirable state of affairs. 3.11 Wider policy issues involved in resolving UCB's problems include market concentration, provision of financial services in rural areas, and Ugandan ownership within the banking sector. 39 Market concentration 3.12 UCB dominates Uganda's commercial banking sector. Its overbearing market presence has stifled competition. There is therefore a concern that-should it be privatized as is-the new private owners would use its size to further limit competition in the marketplace. 3.13 Three factors mitigate such concern. UCB's market share is in some ways not as substantial as it might first seem, since the tremendous impairment of its loan portfolio curtails its influence on new lending. The vast majority of its customers are bad debtors to whom new loans should not be extended and whose old loans will be transferred to NPART. UCB's share has also been steadily eroding. By the time a new and healthy private bank is established (a process which recent experience illustrates includes downsizing, and branch closures), the UCB's market share will be reduced even further. The BOU, moreover, is planning to strengthen its supervision and regulation of the banking sector, with a specific focus of attacking insider lending and enforcing credit concentration limits. An improved regulatory framework would further reduce the ability of a privatized UCB to abuse its dominant position. Ruralfinancial services 3.14 The privatization of UCB or any other solution envisioned should proceed independently of measures taken to maintain basic financial services in rural areas. Indeed there are strong grounds for urging the government to tender a competitive contract for the provision of such services. Even if UCB won the contract, the contract would specify government's financial commitment, and the level and quality of services expected. Ugandan bank ownership 3.15 Privatizing UCB could result in majority foreign ownership (perhaps even 100 percent), thereby weakening the involvement of Ugandans in their own banking system. On balance, however, it is not clear what other option is to be preferred. A prospective sale of shares to the Ugandan public is not tenable before the UCB is restored to sustained profitability and has a sound balance sheet. Restricting conditions of sale to insist on the maintenance of a substantial share of government ownership is not only likely to deter prospective bidders but also defeats the purpose of the sale-to introduce new management and capital and to remove political influence from the sector. 3.16 Uganda's most attractive option is therefore to open the sale of UCB without restriction. This would open the possibility that once a restructured UCB returned to profitability, shares could be sold to the Ugandan public. 40 3.17 Uganda, therefore,-has three plausible options for the sale or restructuring of UCB: * Liquidating or selling selected assets, liabilities, and branches to qualified purchasers. * Deferring privatization until after restructuring. * Privatizing immediately through sale of sufficient assets and liabilities to qualified private owners to achieve the desired level of corporate governance to return the UCB to sustained profitability. Liquidation 3.18 In developed economies, the problems of UCB would be dealt with expeditiously through liquidation, although it is not likely that the bulk of the shares would be in the hands of the government. Stockholders, moreover, would have recognized losses at a much earlier date, regulatory authorities would have withdrawn the banks licenses, and the institution would have been forced to close. In consequence, its assets, liabilities, and branches would be offered for sale. To the extent there was an insufficiency of assets to cover deposit liabilities and as stockholders equity was depleted, a deposit liability scheme, if it existed, would have been triggered. 3.19 In economies with deposit insurance schemes that are inadequately funded, sick banks might be merged with healthy institutions. In several instances, the government has been forced to absorb some portion of the obligation to meet depositor claims. 3.20 An offering of selected assets and liabilities to qualified purchasers is essentially a modified plan of liquidation. Such an approach has the advantage of letting the market determine the most attractive assets, liabilities, and branches, and place a value on them. But this is a difficult way to dispose of the bank, which-in a developing economy-is fraught with peril. Only the best assets will be selected, and in Uganda, few of these will be located outside Kampala. Uganda's government would then be faced with the need to continue to provide banking services across the nation, but with very much weakened institutional structures. 3.21 The drastic step of liquidating the UCB would be likely to damage confidence in the entire banking system severely. It would also effectively decimate the core infrastructure of the commercial banking system, which has served Uganda's enterprises and private clients over many years. While it is impossible to enumerate these indirect costs, Government would bear substantial direct financial costs in the event of liquidation. 3.22 UCB's unaudited balance sheet as at end December 1994, reported a shortfall of assets over liabilities of approximately U Sh 82 billion (Table 3.3). Assuming that all depositors were paid in full, this would be the minimum charge falling to the Government in the event of liquidation. The actual shortfall in the event of liquidation, however, would be greater than U Sh 82 billion, because neither net loans, fixed assets, nor other assets would be realized at 100 percent of their book value. 41 Table 3.3. Uganda Commercial Bank (as ofDecember 31, 1994) (U Sh million) Assets Liabilities Cash 8,582 Demand Deposits 83,890 Reserve a/c 22,091 Savings Deposits 30,908 Due from: Time Deposits 15,599 domestic banks 118 Total Deposits 130,397 foreign banks 28,468 Due to Uganda banks 630 Investments: Due to Foreign. Banks & fx' 31,561 Government 3 Administered Funds 22,478 Other 524 Borrowing at Central Bank 2,902 Bills Payable-U Sh 17,737 Bank of Uganda schemes 1,652 Advances & Discounts Loans 45,959 Other Liabilities 28,174 Overdrafts 18,617 Other Provisions 2,253 Administered loans 12,410 Non Performing Assets 32,759 Capital: Gross Loans 109,745 Paid up 6,463 Less Loan Loss Provisions: Accumulated Losses -115,068 Specific -76,300 Revaluation Reserves 26,779 General -156 Other Reserves 450 Net Loans 33,289 Total Capital -81,376 Fixed assets 30,481 Earnings Y-T-D -1,055 Other assets 28,494 Total Assets 153,702 Total Liabilities 153,702 Source: Form BS 100--Monthly Statement ofAssets and Liabilities for Commercial Banks in Uganda as provided to Bank of Uganda Research and Policy Function-December 31, 1994. 3.23 It is difficult to forecast what proceeds could be anticipated from loans of an institution such as the UCB in liquidation. Given such a distress sale in the present conditions of the Ugandan market, a working assumption could be made that the realizable value of these assets would fall somewhere between 25 to 40 percent of their book value. In such a case, the shortfall could expand to U Sh 123 billion or even U Sh 133 billion. Even this estimate does not reflect the final cost, since there would be management charges associated with the undertaking of the liquidation process and UCB's redundancy expenses could equal as much as six months of its current total salaries and benefits. Deferred privatization 3.24 This is a strategy that the government has in effect attempted to implement over the past two years. As experience has shown, restructuring has proved difficult within the UCB's existing organizational and ownership framework. It does not seem possible to overturn the public sector mentality of the UCB within a reasonable period of time. A fundamental difficulty with this approach is that it attempts to second-guess what the new owner will want changed in the details of restructuring. 'Represents due to banks and "other liabilities payable in foreign currency" and of this amount U Sh. 18,890 million is captioned "Foreign Exchange accounts". The asset account in an equivalent amount is similarly captioned. 42 3.25 Recent experience also suggests that deferring privatization will increase the ultimate cost to government as compared with immediate action. Notwithstanding restructuring, UCB's operating losses continue to mount, and the bad loan portfolio situation may worsen further. Latest reports suggest that operating losses are presently running at around U Sh 300,000 a month, implying a cost from this source alone of some US$ 4 million each year. If the factors driving bad debts are not tackled, yet higher costs could be incurred while awaiting privatization, and even more loan-loss provision would be needed. Immediate privatization 3.26 Whether or not this option is ultimately viable and attractive can only be determined by a serious approach to the marketplace in the form of a well-prepared offer for sale. Targeted potential purchasers should include established bankers overseas who may be considering expansion into the East African market. The working assumption is that UCB's institutional infrastructure, dominant market position in a growing country, and existing deposit base will prove attractive to potential purchasers despite the difficulties to be faced. The injection of private capital and expertise could help preserve UCB, which is the core of Uganda's banking sector, and could turn its performance around both in terms of financial standing and quality of banking services. 3.27 The cost of immediate privatization would be substantial. But this option would avoid all-or at least the worst-of the disruption and undermining of confidence that would accompany the closure and liquidation of UCB. Furthermore, the sale (privatization) option offers the opportunity to dispose of deposit liabilities at a premium and to avoid bearing the full burden of redundancy costs. For all of these reasons, the cost of privatization should be substantially below the forecasted cost of liquidation. 3.28 Testing the privatization of the UCB in the market could, of course, result in a full or partial sale of its shares and assets. Selling UCB as a consolidated entity would be the most attractive option for government, since it would provide the clearest route to end direct government involvement in the sector and would let the private sector resolve the problems associated with UCB. But would a purchaser be willing to acquire all UCB shares from the government? The answer will depend upon buyers' assessment of the financial market potential in Uganda, and on the assets and liabilities that remain with the UCB at the time it is to be privatized. 3.29 UCB's loan assets and fixed assets are the two key asset categories at issue. Net loan assets, on an unaudited basis, were already down to U Sh. 33 billion as of December 31, 1994. Ernst & Young's next loan portfolio review (as of January 31, 1995) will form the basis for determining what loan assets are to be transferred (at net book value) to NPART around the end of June 1995. 3.30 In principle, purchasers are unlikely to have an interest in the UCB loan assets and only limited interest in the bank's fixed assets. Yet in 1994 these two asset categories make up about half of the bank's total assets (excluding the category of "other assets"). By identifying those assets to be transferred to NPART, the Ernst & Young review is likely to reduce the UCB's loan assets component still further. Should the review indicate,that the UCB's loan assets will be a manageable number after the transfer to NPART has been accomplished, the prospects for a sale of UCB as a corporate entity will be enhanced. A prospective purchaser will be more willing to discuss discounts on a smaller, healthier loan portfolio. Fixed asset valuations will continue to be a negotiation issue but perhaps less of a problem after a purchaser has indicated which branches should be included in the package. 43 3.31 In privatization, there is also the question of staff redundancies, since prospective purchasers are not likely to want to employ existing staff who-given the history of UCB-will be considered tainted. A privatization effort, therefore, will try to identify those interested parties, with the requisite banking experience, who are willing to assume deposit liabilities, selected assets, and payroll obligations for as many employees as possible. But it may well be easier to privatize UCB by carving out assets and liabilities to be disposed of through sale, while retaining those to be disposed of through liquidation. Should loan assets become an inconsequential item (as a result of transfers to NPART), the sale of UCB as a corporate entity may be more readily achieved. But in any case, fixed asset valuations would still have to be negotiated. Reform Strategy 3.32 Any review of Uganda's possible options for UCB will point to the relative merits of immediate privatization. Indeed, government has committed itself to privatize UCB without delay. Consistent with this objective, it has been recommended that a majority ownership position in the bank, or in a large share of its assets and liabilities, be offered to qualified investors at the earliest opportunity. 3.33 It is critical that this be done expeditiously. UCB is experiencing severe operating difficulties. Fraud could well be occurring at present and will doubtless by increase during any continued period of uncertainty. Staff morale will decline as a result of continued losses and further branch and agency closings-even when closing such units is appropriate and commercially justifiable. In such an environment, employees will be under intense pressure to act in their own self interest and not necessarily do what is best for the UCB. 3.34 At the outset a new owner-manager could carry up to 100 percent of the assets and liabilities until sustainable profit levels are achieved. The prospective owner-manager would need the full ownership position to allow new governance to be implemented unfettered. In due course, when profits are stabilized, shares could be offered periodically to Ugandan citizens and institutions, consistent with the development of Ugandan capital markets. It is anticipated that substantial local ownership will result, but that ultimate beneficial ownership will remain at 51 percent or more with the prospective qualified investor. 3.35 A merchant bank should be selected immediately to prepare data, draw up an offering dociunent, contact qualified investors, and help compare and negotiate proposals. This process should be completed within three to four months. 3.36 A rural branch network must be maintained to provide financial services to support development in rural Uganda. A study should therefore be initiated to identify the minimum number of locations needed to provide these services (such as fund transfer and deposit taking), the institutional arrangement most appropriate for providing them, and whether or not a special government incentive scheme is needed to ensure that the services will be provided. If so, the study will investigate the cost of providing such incentives and how such a government intervention could be administered commercially. Following the successful sale of the more commercially viable parts of UCB, the remaining branches, including assets and liabilities would be assessed in light of study's recommendations. Relevant branches, with their assets and liabilities, would be incorporated into the proposed framework for the provision of rural financial services, while those not required for this purpose would be liquidated. 3.37 This study should not be viewed as a precondition for the appointment of the merchant bank. Nor should it interfere with the proposed sale of UCB. The continuous, large, monthly losses presently being incurred by UCB require that it be sold as soon as possible. 44 3.38 In the event a merchant bank cannot attract a purchaser (or purchasers) for substantial portions of UCB, the privatization advisers should have gained enough market experience to implement a revised approach for their optimal disposition. A review of progress achieved by the merchant bank will be undertaken at the end of the three to four month period. Should the privatization process become prolonged, additional measures to stem losses at UCB and continue down-sizing will need to be considered. Action Plan 3.39 The steps in launching the initiative to privatize the UCB include: * Appoint Merchant Bank to prepare data and documentation to offer UCB for sale to qualified interested buyers. * Appoint consultants and initiate study on agreed terms of reference on the provision of 2 financial services to rural areas. * Implement interim remedial measures at UCB: (i) ensure (through appropriate measures taken by UCB's board/senior management) that safeguards are in place to prevent assets stripping, (ii) restrict new lending and growth in deposits, (iii) transfer impaired loans (based on portfolio as it stood on January 31, 1995) to NPART-UCB to be held responsible for maintaining and safeguarding all documentation on impaired loans (including documentation on collateral and related securities) until they are transferred to NPART, (iv) UCB to be provided with treasury bonds bearing market rates of interest equivalent to net market value of loans transferred to NPART (as determined by Ernst & Young study which is nearing completion). * Disposal of any unprofitable branches agreed to by UCB's board and the government of Uganda * Complete necessary legal/institutional steps to allow sale of UCB which is a strategic parastatal (by September, 1995). * Invite bids from merchant banks to sell UCB in whole or in parts (i.e., offer for sale by October 31, 1995). * Complete evaluation of bids and make sale decision, or revise strategy (for disposing of UCB) if no sale is achieved by January 31, 1996. * Make recapitalization decision on UCB (including the form of recapitalization) consequent on negotiation with purchaser of UCB. Complete privatization process by transfer of ownership to purchasers by February 29, 1996. * Based on results of rural finance study, ensure operation of UCB bank branches essential for providing financial services to rural areas that are either not sold or closed by new owners, simultaneously with transfer of UCB to new owners. 2 This study will look at a minimum level of financial services, where they are to be provided, and types of intervention to ensure their provision (see Chapter 7 for details). 45 4. Financa Disciplline and the Problem of Bad Debt 4.1 The Ugandan financial system has had a long-standing lack of financial discipline on the part of borrowers, which has been aggravated by the country's history of financial and economic mismanagement. With both lending institutions and borrowers regularly disregarding the basic tenets of financial discipline, resolving Uganda's problem of endemic bad debt will require coordinated effort on the part of government, banks, courts, regulators, and borrowers. Background 4.2 Before the early 1970s the incidence of bad debt in Uganda was modest. Thereafter- during the period of economic mismanagement-banks were forced to set aside prudential criteria and make loans on the basis of political directives. In the late 1970s and early 1980s, it became common practice for Ugandan borrowers to treat loans as grants, or at least to regard payment terms as negotiable. With the decline in public sector real wage and salary levels, moreover, bank loans were considered as (initially costless) supplementary rewards for public service. As corruption in certain echelons of the public sector spread, it also became possible for borrowers with poor security to obtain bank loans by corrupt means. 4.3 The extent of bad debt in Uganda was magnified in the 1980s by donors' well-intentioned provision of substantial economic rehabilitation assistance, which was on-lent by commercial and development banks to the private sector. Inadequately supervised, these loans were made at a time when the practice of default had become accepted and widespread. At the same time the Ugandan economy suffered from both macroeconomic instability and civil conflict. The value of some borrowers' security and some projects themselves were destroyed. The rate of recovery on development loans administered by Ugandan financial institutions reached an all-time low. 4.4 Yet in the late 1980s and early 1990s, the impact of almost universal nonrecovery on banks' profitability was masked by high rates of inflation. Because loans were not indexed, the real value of their principal was rapidly eroded by inflation. The loss of principal through borrower default was thus relatively unimportant. Banks were able to remain profitable (in a cost accounting sense) thanks to the high nominal rates of interest charged on new advances. 4.5 By the mid-1 990s, however, the monetary authorities have stabilized Uganda's economy and the price levels. The loss of loan principal through borrower default is now a serious threat to banks' profitability. 4.6 To curb the fiscal deficit and restrain public expenditure, and thereby to achieve macroeconomic stabilization, has required vigorous government action. Strict cash limits have been imposed on the annual outlays of ministries and government agencies. Yet these limits have tended to perpetuate the practice (common during the unstable 1970s and 1980s) of delaying payment to government suppliers and contractors, who in turn pass on their cash flow problems to banks by defaulting on their bank loans and advances. 4.7 By the end of September 1994, therefore, over half of all commercial bank loans in Uganda were nonperforming. Largely as a result of having to make high provisions for bad and doubtful debts, nine out of the country's fifteen commercial banks were reporting net income losses. Over the twelve months ending in September 1994, aggregate losses by commercial banks in Uganda amounted to U Sh 50.1 billion or 10.7 percent of the sectors end-of-period total assets. 46 4.8 The largest commercial bank, UCB, with almost 40 percent of total deposits, has an exceptionally high incidence of nonperforming loans in its portfolio. A review by Ernst & Young of its end-1993 portfolio of loans in excess of U Sh 3 million found that, by value of the portfolio, 86 percent were nonperforming and 60 percent should be classified as losses. In view of the revised accounts for the year ending September 1993, UCB's provisions should therefore be increased from 55 percent of the value of the portfolio to 69 percent (see Chapter 3 above). 4.9 UCB's experience has been broadly paralleled by those Ugandan banks who lend outside of a narrow spectrum of blue-chip borrowers. In fiscal 1994, even the foreign banks as a group had over 22 percent of their portfolio in nonperforming loans. 4.10 As the larger of Uganda's two development finance institutions, the Uganda Development Bank's (UDB's) net worth is negative. Cumulative provisions to December 1994 amount to 53 percent of gross loans advanced and receivable (U Sh 27.5 billion, excluding accrued interest). The UDB's (unaudited) financial loss in 1994 of U Sh 2.6 billion on gross income from investments of U Sh 4.9 billion is indicative of extensive provisioning against recent loans and of the high proportion of the overall loan portfolio that is non-performing. The much smaller Development Finance Corporation of Uganda, by contrast, counts only 20 percent of its loan and equity portfolio as impaired. 4.11 The problem of bad debt has restricted enterprises' access to financial services and increased the cost of borrowing in Uganda. 4.12 Banks must maintain a high level of provisions, which severely reduces their capital and impairs their ability to provide financial services. A sample survey of 265 enterprises at the end of 1994 found that only 39 percent of them used bank loans. Even large-scale firms in the sample made only limited use of bank credit: only 42 percent of them had overdraft facilities. Since lack of working capital is an important constraint on business expansion, difficulty in obtaining adequate short-term financing undermines the chances of new ventures. Established businesses, too, have to tie up large amounts of their capital in credit to their customers. In the survey, 39 percent of firms reported offering credit to their customers, including 12 percent to new customers. 4.13 Ugandan banks, moreover, are now striving to restore their profitability by charging high real interest rates to borrowers. The average interest rate paid by the 265 enterprises in the sample survey was 22 percent, and in the year to October 1994, the composite consumer price index rose by nearly 6 percent. The real interest rate paid by these enterprises was thus of the order of 16 percent so that many firms were deterred from borrowing by the high cost of credit. 4.14 In addition, in the absence of good information about credit risks and given the difficulties encountered by lenders in pursuing delinquent borrowers for the recovery of their loans lenders are now cautious about extending new loans. The bankers' guideline is that loans should not exceed 60 percent of the value of securities taken as collateral, but in practice, much more collateral is often demanded. In the sample of 265, those firms that had taken out bank loans reported offering an average of 234 percent loan value as collateral. Many firms, moreover, were unable to offer collateral of the amount and quality bankers required. Issues and Options 4.15 The bad debt problem has highlighted a series of issues requiring attention by bank management, the judiciary, the administration of justice, the government, and relevant professional associations. Uganda has to overcome its tradition of disregarding financial discipline, which will 47 require nothing less than a concerted effort on all sides under determined political leadership. Happily, Uganda has already undertaken a number of important steps toward this end. Banks' lending practice 4.16 In several banks, credit appraisal has hitherto been lax. In particular, banks have not always respected the statutory limits on lending to directors or to other insiders. The security accepted as collateral was too often not verified for adequacy and authenticity before loans were granted. Borrowers' valuations were taken at face value. Title deeds were not verified with the Land Office. Personal guarantees and powers of attorney enabling borrowers to offer other peoples' property were accepted as security and later found to be defective and unenforceable. (Powers of attorney generally expire at the contractual term of loans and are thus valueless for recovering overdue loans). Outside the main cities, moreover, mortgages were taken out on properties not readily marketable, both for reasons of customary law and because of communal disapproval. 4.17 In the 1993 review of larger loans in the UCB portfolio, most securities were found not to be owned by the borrower and to have been charged against powers of attorney from registered owners. The UCB survey also found lax record-keeping. Some securities taken for delinquent loans could not readily be traced. The UCB had failed to establish firm repayments schedules and to monitor borrowers' performance against them. 4.18 Ugandan banks rarely use receivership as a means of recovering debts, though the concept of receivership is recognized in the Companies Act and Mortgage Decree of 1974. Ugandan law is virtually identical to that of Kenya where banks often appoint receivers over the assets of defaulting companies. The power to appoint receivers arises from debentures creating floating charges over all the assets of a borrower if these form part of loan agreements with banks. 4.19 Ugandan banks' almost universal practice is to take as security a charge on immovable property. In the few cases where banks have appointed receivers, debtors have obtained court injunctions delaying their control over the debtor company's assets, which were damaged or removed. Receivership, moreover, is expensive and only worth using when large loans are involved and the borrower has adequate realizable assets. If more commonly practiced in Uganda, it would make a limited but still worthwhile contribution to banks' ability to recover loans. 4.20 Compounding these problems, some banks have made a practice of granting noncommercial loans on the basis of political directives. While not a serious problem in terms of banks' overall balance sheets, this practice contributes to the banks' general disregard for the basic laws of lending, After years of lax practice, there is now a need to retrain bank staff under the direction of professional managers whose goal it is to restore the viability and integrity of the institutions they run. 4.21 Ugandan banks also now recognize that sound lending requires rigorous appraisal of the character, financial and economic track record, and prospective cash flow of borrowers. The UCB has recently begun to systematically check the authenticity of documentation offered as loan security. 4.22 Uganda's banks need a reliable source of financial information about prospective borrowers to weed out those with a record of delinquency. Both banks and nonbank development finance institutions need confidential, case-by-case access to a computerized data base of bad debt experiences. Such an information exchange could be set up under the aegis of the Bankers' Association or the BOU's Bank Supervision Department, which receives such information in the 48 course of its sector inspections. Delinquent borrowers whose loans from UCB have been transferred to the Nonperforming Assets Recovery Trust (NPART) should also be included in the data base. Law enforcement and disputes settlement 4.23 Lenders in Uganda have serious problems bringing legal action against defaulting borrowers. Court orders for foreclosure or the seizure of assets given in security are frequently delayed for months because of the congestion of cases before the courts. Defendants, moreover, are too readily granted adjournments of their cases by judges, whether because of inappropriate influence or from difficulty of scheduling court cases. 4.24 All civil cases involving sums in dispute over U Sh 5 million are heard in the High Court. But while the High Court can more or less cope with the flow of new cases now coming before it, it has been unable to reduce its backlog of some 5000 cases, some going back to 1986. At present a serious litigant may wait six months for an initial one-day hearing. Subsequent hearings, if needed, are likely to take place at monthly intervals thereafter. Defendants can therefore easily cause delays either by failing to appear in court or by contriving to have papers mislaid. 4.25 When judgment has been given against a defaulting debtor, further delays can occur in loan recovery if debtors take out ex parte injunctions preventing the sale of their assets or protesting at the conditions of the sale or the values realized. Poorly publicized auctions or private treaty sales arranged by court brokers, furthermore, may yield very low values. 4.26 In the past it has also been difficult to obtain legal resolution of government's failure to pay suppliers, which has subsequently led to bad debt. Suppliers who sued the government could at best obtain a declaratory order requiring the government to pay. 4.27 This situation has now begun to improve. Some problem banks have voluntarily tightened their lending procedures and are pressing for recovery of nonperforming loans through legal processes. The government also decided recently to increase the capacity of the High Court to handle commercial cases and to tackle some of the procedural and administrative sources of delay in the administration of justice. 4.28 Under the Mortgage Decree of 1974, two procedures are open to banks for the recovery of nonperforming loans. If they have clear title to assets pledged in security sufficient to recover debts owed to them, banks may-after giving sixty days notice-take possession of the borrowers' assets. In other cases creditors may apply to the courts for recovery of their loans. Some 80 percent of the UCB's nonperforming loans are presently the subject of court action. 4.29 Both of these procedures, however, are uncertain and frustrating for creditors. Taking possession of assets can be frustrated by court injunctions obtained by delinquent borrowers. Even when court orders are given in favor of creditors, further problems can arise in realizing loan securities should debtors dispute the values realized at public auctions conducted by court brokers The courts have upheld claims that assets may be undersold. 4.30 In recognition of these problems the decree establishing NPART provides for the creation of a tribunal where decisions for the liquidation of assets pledged by delinquent borrowers whose debts are transferred for recovery to NPART can be reached expeditiously. Some 2,500 cases will be transferred to the tribunal, which will be presided over by a High Court judge. Its decisions, moreover, cannot be held up by disputes over the validity of the public auction of seized assets. 49 The planning of the NPART tribunal, however, is not yet far advanced. The High Court is not formally aware of it, nor of its intended modus operandi. 4.31 Following an interim recommendation of the Law Reform Commission, however, the Ugandan authorities have set up a Commercial Division of the High Court and appointed to it seven experienced judges. The High Court has also decided that it will no longer allow ex parte injunctions preventing the sale of debtors' assets. Henceforth, injunctions will have to be served on plaintiffs with the result that creditors are likely to suffer only minor delay while the case is being considered by a judge. 4.32 The recovery of debts through public auctions would be further facilitated if the court brokers, who are appointed by the registrar of the High Court and who conduct auctions, were more closely supervised. Auctions should be well publicized, and-to limit the scope for collusive bidding-creditors should be certain to make their own valuation of the assets to be sold. 4.33 The Law Reform Commission, which is due to complete its report in June 1995, is expected to recommend that more commercial disputes be heard before magistrate courts (which are at present unable to hear cases involving sums greater than U Sh 5 million). This may not be of particular help to banks, however, since poorly paid magistrates generally lack expertise in commercial cases and are more open to influence by defendants. 4.34 The Law Reform Commission is also expected to recommend the institution of pre-trial conferences of all parties involved to establish the facts and expedite proceedings in court. 4.35 The reform of the High Court will lead to the more expeditious and effective processing of debt recovery cases, only if it has more staff and more adequate physical facilities for hearing more cases. Under the World Bank's Capacity Building Project, provision is being made for the training of judicial, registry, and secretarial staff. Danish aid is helping to finance the construction of new law courts. By 1996, these initiatives will have overcome some of the physical, organizational, and administrative obstacles that have aggravated judicial delays. With improved management information, the Ministry of Justice intends to assess all cases brought by different plaintiffs against the same defendant simultaneously. There would be additional advantage in a comprehensive review of all outstanding High Court cases, which would be prioritized to make the most efficient use of Uganda's present judicial facilities. 4.36 Under the Capacity Building Project, the Ugandan Companies Act is also being reviewed. Although the Act covers insolvency, it is insufficiently specific about the powers and responsibilities of receivers. For this reason there have been very few cases of receivership in Uganda brought by commercial creditors, and most of those were brought by government ministries as part of the process of liquidating insolvent parastatals. Receivership remains the virtual monopoly of one firm of accountants. If the duties and powers of court-appointed receivers were more clearly delineated in legislation, and if the procedures of administrative receivership were defined, the scene would be set for the more orderly satisfaction of creditors' claims on insolvent borrowers. 4.37 At present, private companies in Uganda are not required to file annual accounts. It would help banks to assess the credit-worthiness of potential borrowers if an amendment were made to the Companies Act requiring all companies to file accounts. 50 Professional standards 4.38 Uneven standards of professional etiquette and probity in the professions of accountancy, surveying, valuing, arbitration, and even law in Uganda contribute to the difficulty of enforcing creditors' claims on delinquent debtors. In some cases, such as accountancy, professional bodies have only very recently been set up and have yet to exercise their influence on professional standards. Of particular concern to banks is the practice of certifying company accounts that are known to be inaccurate and of making exaggerated valuations of landed property and other assets used as collateral for loans. Action by Government 4.39 Several administrative and legislative aspects of Uganda's bad debt problem will require attention by government. The government has therefore recently increased the capacity of the courts to handle commercial litigation and to force central government departments to reduce their domestic payment arrears. 4.40 Further desirable administrative actions by government include: * Ceasing to pressure financial institutions to make loans against their better commercial judgment. * Ensuring that public sector agencies and enterprises pay suppliers and contractors promptly within contractual deadlines. To achieve this, central government administrations could be required to pay their bills before the end of the fiscal year in which the bills were presented. * Encouraging Uganda's financial sector professional bodies (representing accountants, lawyers, values, surveyors, and arbitrators) to enforce the highest professional standards among their members. * Enforcing the provisions of the Companies Act concerning trading by firms while insolvent, the preparation and auditing of accounts, and the filing of accounts with the government. 4.41 The basic legal framework provided by Uganda's Companies Act is satisfactory, although somewhat dated. Not many legislative changes, therefore, are needed, and these are of lesser urgency than are reforms in administration and enforcement. But to modernize the law with respect to commercial disputes and the recovery of bad debt, Uganda could: * Amend the Companies Act to reflect the best available contemporary legislation world- wide (perhaps along lines currently under consideration in Kenya), for defining the duties and powers of receivers. * Reconsider the Mortgage Decree of 1974 in the light of the NPART Statute, Section 11 (7a), which gives clearer guidance on when the values of properties realized at public auction can be considered valid for the settlement of debts. Reform Strategy 4.42 Because inappropriate commercial, professional, and administrative practices are ingrained in Ugandan financial dealings, it will not be easy or quick to overcome the country's endemic problems with bad debt. Tackling the problem will require strong political leadership and decisive action on the part of the professions to improve standards in commercial life. Lenders, 51 too, must act more rigorously to improve the quality of their loan books, and courts must ensure more effective enforcement of contracts. 4.43 If encouraged by political leadership, cultural change will come in due course. Priorities for action in the short-term must therefore lie with matters susceptible to administrative change. These include: o Bank owners and management, encouraged and policed by the Bank Supervision Department of the BOU, must aspire to the highest standards of loan appraisal, evaluation of securities, and follow-up. All lenders-development finance institutions as well as banks-should be much less tolerant of nonperforming loans and more vigorous in collections and recovery. o Government must not allow any officially sponsored lending scheme where a high level of loan recovery is unlikely and which would therefore officially sanction lack of financial discipline. o Government should allocate support for significantly more staff, training, and physical facilities for the adjudication of commercial cases in the courts. o The administration of justice should be streamlined and outstanding cases systematically reviewed and prioritized for more effective scheduling of court hearings. Courts should be less tolerant of delays and alert to spurious pretexts for adjournments or for injunctions for the stay of court orders. Action Plan 4.44 Given this strategy for financial sector reform, Uganda's banks will wish to undertake the following actions immediately: o Institute training for staff in the mechanics of loan appraisal and enforce standardized appraisal guidelines, which will include assessing borrowers' characters, financial history, and prospective cash flow. o Appraise land titles and the value of securities rigorously. o Institute bonus incentives for good loan appraisal and follow-up practice. o Resist all pressures to lend for noncommercial reasons. o Adhere strictly to BOU on limits on insider lending. o Exchange credit information about loan defaulters among members of the Banker's Association and development finance institutions. 4.45 For the longer term, banks may wish to: o Make more use of the Companies Act provision allowing lenders to take floating charges and debentures over the assets of borrowers so as to make full loan recovery more likely. 4.46 The BOU will wish to institute the following measures to improve bank supervision: o As part of its inspection activities, the BOU should require problem commercial banks to improve their management of lending operations, their credit appraisal and follow- up procedures, and their conformity with insider lending limits. o The BOU should actively encourage Uganda's banks to set up and use a credit reference bureau. 4.47 Uganda's courts will wish to undertake the following measures immediately: 52 * Define powers and procedures of the NPART tribunal and the Commercial Division of the High Court, to provide fair, expeditious adjudication of debt recovery cases. * Institute courts to be less indulgent in granting adjournments of court proceedings, delays, and suspensions of court judgments in debt collection cases. * Penalize delays in conforming with court decisions more severely by increasing the rate of interest due on sums payable the further the date of payment from the date of the court judgment. * Periodically review the courts' backlog of debt collection case and establish priorities and a timetable for hearing them as expeditiously as possible. 4.48 For the longer term, Uganda's courts may wish to: * Impose harsher penalties for deliberate acts of deception by borrowers, such as offering invalid guarantees and securities for loans. 4.49 For its part, Uganda's government will wish to: * Pay suppliers promptly and require that parastatals and local government do likewise. * Cease to pressure banks to make loans for noncommercial reasons. * Encourage professional bodies to set and enforce high professional standards for their membership. 4.50 For the longer term, government may wish to: * Curtail by law, appeals to the courts against the decisions of arbitrators. * Update company law-perhaps along the lines of proposals now under consideration in Kenya)-to define the powers and duties of receivers more clearly. * Amend the 1974 Mortgage Decree to prevent disputes over the validity of the supposed "market value" of assets sold at auction. * Amend the law concerning land title to make rural land more easily marketed and therefore more acceptable as collateral for bank loans. 53 5. Reforming the Bank of Uganda Introduction and Background 5.1 The two World Bank financial sector missions found that over the past five years the Ugandan authorities, particularly the Bank of Uganda (BOU), with World Bank assistance had directed considerable effort to strengthening the role, capacity, and authority of the central bank. 5.2 There had clearly been substantial improvement in the performance of the BOU at all levels since the 1990 Financial Sector Review. Sound legislation supporting the BOU's operations is now in place, and the BOU now plays a larger and more effective role in monetary, foreign exchange, and supervisory policy. There has also been a substantial move toward market-oriented implementation of monetary and foreign exchange policy. 5.3 Within the BOU many operational arrangements were streamlined. The top management structure was tightened and a number of deficiencies in BOU's operations were rectified. Audited annual accounts of the BOU are now available on a timely basis, and a more soundly based comprehensive computerization program was established (under the EFMP project). Most importantly, the BOU recently instituted a forward-looking strategic plan to make it more efficient and cost-effective. In December 1994, 500 staff left the BOU whose staff now numbers 1,200 and further sizable reductions in staff are to come. 5.4 But the World Bank Mission found that much more needs to be done if the BOU is to fulfill its role as Uganda's Central Bank, to help in the formulation and implementation of economic policy, and to facilitate Uganda's financial development. 5.5 The following areas were identified by the Mission as requiring close attention: o The BOU's untenable financial position in terms of both its balance sheet and its income. o The BOU's management culture, their general performance, and the means available to change them. o The need for a time-table for instituting strategic change along the lines recently proposed by BOU's consultants. o The need to process applications for new bank licenses in accordance with the recently enacted Financial Institutions Statute. o The need to institute procedures for dealing more firmly and with greater dispatch with problem banks, together with a review of the supervisory authority of the BOU. o The need to review Ugandan banks' deposit insurance scheme, and the BOU's development finance operations. Issues and Options The untenable financial position of the BOU 5.6 As of June 30, 1994, the BOU was technically insolvent, and the situation has deteriorated further since then. Losses from BOU operations continue and are compounded by sizable foreign exchange losses following appreciation of the exchange rate. As it now stands, foreign exchange 54 liabilities booked in the BOU's balance sheet exceed its holding of foreign exchange. In brief, according to its balance sheet (Table 5.1): * The BOU incurred a deficit of income over expenditure of almost U Sh 3.3 billion in fiscal 1994. In the September quarter of 1994, there was a further deficit of almost U Sh 1 billion. * The BOU recorded a foreign currency translation loss of U Sh 37.3 billion in fiscal 1994, and from July through September 1994, there was a further loss of almost U Sh 1 billion. * The BOU had a negative net worth of U Sh 11.5 billion in its 1993 balance sheet. By June 30, 1994, net worth had decreased to U Sh 23.2 billion. * Even after substantial changes directed at cost-cutting and improvements in efficiency, budget forecasts predict continuing deficits of income over expenditure. 5.7 It is evident that the Table 5.1. BOU balance sheet (June 1993, June 1994) BOU's negative net worth is (U Sh million) far larger than that shown Net worth 1993 1994 on the balance sheet (U Sh million) (U Sh million) primarily because the true Assets value of the loan portfolio is Tangible fixed assets 11,324 13,774 substantially overstated. Investments 2,816 494 The problem is somewhat External assets 362,581 403,807 different from that Loans and advances 538,839 760,422 encountered by many other Other assets 9,433 9,246 central banks which have Total 924,993 1,187,743 Liabilities made large scale loans to Currency in Circulation 108,902 147,765 the private sector that could Deposits 505,131 815,266 not be recovered. By Other creditors 8,046 3,834 contrast, the BOU's loan Foreign liabilities 226,088 190,987 portfolio consists entirely of DR allocation 49,185 41,346 government obligations. Capital 15,050 15,050 The outstanding question is Reserves 12,591 (26,415) the extent to which the Total 924,993 1,187,743 govurnment accepts Source: Coopers & Lybrand responsibility for recorded BOU loans. 5.8 Loans to government plus those seen by the BOU as guaranteed by the government amount to just over U Sh 750 billion. This includes: * Direct loans to the government (570 U Sh billion). * Loans to parastatals guaranteed by the government (l lU Sh billion). * Loans recorded as having been made at the request of the government (13 U Sh billion). * Amounts due for servicing loans (including bearing foreign exchange losses) and accounts for the government. Of these loans, some 160 U Sh billion relates to managing the government's accounts with the IMF and a further U Sh 5 billion relates to borrowings from the European Economic Community (EEC), Russia, and barter trade. It has to be established that all these so called loans are accepted as such by the government and the Various government departments. 55 5.9 To the extent that the government and its various departments do not accept responsibility for specific loans and agree to appropriate repayment arrangements, the BOU will need to write them off, which will substantially increase its negative net worth. To determine the BOU's true financial position it is therefore recommended that-after the present internal review of the BOU's books has been completed-the bank's loan portfolio, and possibly its total balance sheet, be reviewed again by an independent external accountant to make sure that all deficiencies have been recognized. 5.10 But whatever the outcome of these reviews, a substantial recapitalization of the BOU is clearly necessary. A central bank must have a sound balance sheet. An unsound balance sheet, particularly one that indicates the central bank is insolvent, weakens its capacity to enforce prudential regulations on banks and financial institutions. Of even greater significance for the Ugandan authorities is the message it gives to overseas donors and lenders, with grave implications for the cost and availability of loans. 5.11 Present work in both the BOU and the Ministry of Finance seems to be driven primarily by concern over the imbalance between income and expenditures for both the BOU and the government. In fact, while the BOU's balance sheet difficulties could be resolved at no cost to the budget, resolution of its income difficulties would have an inevitable impact on the budget. The current debate in Uganda also centers on ways to make the BOU efficient and cost effective without weakening its authority or preventing it from fulfilling its responsibilities. The BOU balance sheet 5.12 Once the validity of the BOU's recorded loans have been established, it will be possible to calculate the amount of money needed for recapitalization. The aim of recapitalization is to make sure that the BOU has adequate capital and reserves to support its operations. To accomplish this: o An injection of U Sh 38 billion is needed to eliminate the negative reserves valued as of September 30, 1994. An additional sum will be needed to cover further loan write- offs which could amount to U Sh 260 billion and is unlikely to be less than U Sh 200 billion. o Minimum capital should be established for the BOU at U Sh 30 billion rather than at the present level of paid-up capital of only U Sh 15 billion. (Currently BOU's capital is 1.25 percent of liabilities, far less than the minimum U Sh 20 billion specified in the 1993 legislation). o The BOU's reserves must be made adequate to support monetary policy and foreign exchange activities conducted to further the economic policy interests of the country. These activities can result in large financial losses throughout the economy. The sums involved can be large as shown by the U Sh 11 billion loss from currency appreciation, (between July and September 1994) and the accumulated U Sh 130 billion foreign exchange loss in relation to IMF accounts. Considering the risks facing the BOU, therefore, reserves in the range U Sh 75 to 100 billion would be modest. 5.13 In summary, appropriate recapitalization of the BOU would require amounts of the following order: o Increasing capital by U Sh 15 billion shillings. o U Sh 25 billion to remove the present negative reserve. o U Sh 75/100 billion to establish minimum reserves to support BOU's monetary and foreign exchange policy operations. 56 * That part of the estimated U Sh 260 billion loan portfolio that is understood to be subject to some uncertainty as to authenticity; IMF management costs (U Sh 160 billion) form the major part of this. 5.14 These numbers are substantially larger than those currently being discussed between the BOU and the MOF but they must be recognized. On the information available to the Mission it is not clear why the BOU in its proposals to the MOF have not fully included the losses on management of IMF accounts set out in its published accounts. If budgetary compromise is needed it should be on the income received on capital rather than the amount of that capital. It would appear that the two issues at present are being confused. The following section deals with the income issue with particular emphasis on the budgetary implications. While it is far from the best solution, the balance sheet problems of the BOU can be solved with no impact on the Governments budget by funding it with zero interest bonds. The BOU's income position 5.15 In the present Ugandan economic situation, the only practical option for recapitalizing the BOU is to issue government securities. Given the sums involved, moreover, these securities must initially bear a rate of interest below market value. An acceptable rate must therefore be negotiated between the BOU and the government, balancing national budgetary difficulties against the BOU's need for income to fund operations and sustain reserves. 5.16 Over time, however, the rate of interest on BOU government securities must be raised to market rates, for only if market rates are paid will the BOU be able to fund future losses occasioned by Uganda's monetary policy, foreign exchange, and central banking operations generally. In any case, international accounting standards may soon require such assets to be valued at market value rather than at cost. 5.17 A program to move to market rates would be preferable to having to make further large, discreet capital injections to the BOU at times dictated by the movements of exchange and interest rates, budgetary needs, or the governments' overseas borrowing capacity. Ideally the process would build a stock of BOU treasury bills and possibly bonds of longer maturities, which the bank could then use to effect its monetary policy operations. It would therefore be preferable to raise below-market bonds to market rates in batches rather than to raise the rate on all bonds at the same time. Provided that the BOU is operating efficiently, market rates on bonds would not be a net cost to the government because additional earnings of BOU would be returned by way of dividends. Timing dislocations on the national budget could be largely covered by requiring that interim dividend payments be made before the end of the financial year. Extraordinary financial items 5.18 Central banks require substantial reserves to cover potentially large losses stemming from changes in foreign exchange and interest rates. Acquiring such reserves for the BOU will take some time, since current losses are much larger than the bank's current minimum reserve. It is common for countries to restore foreign exchange reserves to their central banks from the national budget. This course is recommended for Uganda and the BOU. 5.19 The BOU and government must resolve the question of who bears responsibility for the cost of servicing the IMF and related accounts. 5.20 The recapitalization process will resolve the foreign exchange losses to date. It does not provide for the future. There are three alternatives: 57 o The Government meets the costs directly as they occur. o The Government reimburses the BOU for all cost incurred. o the BOU meets the costs; in this case the required funds must be provided for in the calculation of the interest rate on the recapitalization bonds The cost to the budget would be much the same in all cases although the timing could be different. Accountability 5.21 While an independent central bank must operate in an efficient and cost effective manner, the BOU has yet to make it clear that it can do so. The BOU should therefore review its operations to make sure they are efficient and that charges are appropriate for services. 5.22 The issue of efficient and cost effective operations are covered in subsequent sections of this paper. Two specific areas are worthy of mention. First, the BOU's currency issue operations should, at least over time, be turned around as they currently are a cost to the bank and should be a source of income. Most central banks have substantial earnings from their currency operations. For the BOU the difficulty is the large volume of small denomination notes they print and circulate. Consideration should be given to establishing whether it would not be cost effective to replace low denomination notes by coins. Given the proposal to purchase note sorting machines to upgrade present arrangement, early review of the matter would seem to be timely since it could significantly impact the number of machines required. The second issue is charges for banking services provided to the government. At present no charges are payable when, as at present, the government has deposits in excess of loans. Charges only relate to net debit balances. The BOU should consider alternative means of charging that at least make a contribution towards the substantial costs incurred in providing banking services for the government. Bank licensing 5.23 In Uganda today, lack of an accepted procedure for licensing banks and dealing with problem banks has hampered the country's economic growth. It is clear that no country seeking to develop a financial system conducive to economic growth would have licensed some of the banks recently approved in Uganda. In some cases, the possibility of future failure is uncomfortably higi.. Some of the licenses represent a potential hazard to the banking system, since these banks have characteristics that have been associated with the diversion of resources away from government and the more productive sectors of the economy. 5.24 It appears that government has failed to appreciate the possible consequence of granting bank licenses too freely. The BOU should therefore provide an assessment of the sector to improve official understanding of the issues involved. 5.25 Theory. Ugandans hold a range of views regarding the benefits of licensing new banks. Some hold that, in a liberal society, there should be freedom of entry and that a certain number of failures is inevitable. That argument, however, does not take into account the cost of bank failures to society, which is far greater than the cost of failure in other industries. In no country is it sensible to permit entry to banking unless it is expected that the operation will be run with integrity and ultimately succeed. 5.26 Inappropriately licensed banks can lead to bank failures that undermine the public's confidence in the financial system even if depositors are reimbursed. Savings are then reduced and investment is misdirected. If depositors are reimbursed, moreover, government expenditure on other socially important areas must be reduced. 58 5.27 Integrity is all-important in banking because of the opportunities to facilitate tax avoidance or the laundering of illegal gains. Most countries, therefore, only license new banks that demonstrate integrity and a high chance of profitability. To do otherwise runs the risk of damaging the financial system as well as its ability to mobilize funds. 5.28 In Uganda, with its extremely fragile financial system and its numerous insolvent or weak banks, the risks posed by an overly liberal licensing policy are even greater. A strong case can therefore be made for declaring a moratorium on new banks for a year or so to help existing banks consolidate their position and recover from the present crisis. In the longer term, only applications that are soundly based, bring new products or will operate in an underbanked area should be accepted; and such proposals are likely to be rare. Investment could be increased far more by rehabilitating the existing banking system than by licensing new banks of the caliber displayed by recent applicants. 5.29 Processing procedures. Uganda's Financial Institutions Statute specifies that new banks must have a sound basis for licensing. Under the statute, applications must be made initially to the BOU. In the event of refusal to grant a license, the applicant can appeal to the Minister, who will then deal with the appeal in consultation with the central bank. Within a period prescribed by law, the BOU must satisfy itself as to: * The financial condition and history of the applicant. * The nature of the business of the applicant. * The competence and integrity of the proposal management. * The adequacy of the applicant's capital structure (minimum capital requirements are specified for local and overseas operators), earnings prospects, business and financial plans. * The convenience and needs of the community to be served. * Whether or not public interest would be served by granting the license. 5.30 It is important that the central bank be the prime party in the licensing process because of the importance of maintaining a financial system that is financially sound and operates with integrity and in the interests of the community. Because of their key role and their access to information about local and overseas applicants, central banks are usually in good position to make the appropriate judgments. Because these judgments are inevitably somewhat subjective they are better made at the central bank than at the political level. 5.31 It cannot be emphasized too strongly that licensing decisions must be made purely on financial considerations. Political considerations should play no part and the approach to licensing banks must always be cautious. Applicants must be assessed carefully against the criteria specified above in the statutes. 5.32 Looking beyond the moratorium, Uganda's Minister of Finance and the BOU now need to work together to devise an approach for handling bank license applications. To this end the BOU should prepare: * A report analyzing the possible impact on investment and the stability of the financial sectors, as well as possible financial consequences for the government of granting additional bank licenses. The report should also assess whether or not the licensing of new banks is likely to increase the number of bank failures in Uganda. * A detailed description of what is required of applicants to satisfy each criterion specified by statute. (Such documentation is common in other countries). The Minister of Finance and the Uganda government must then accept that assessment to avoid the risk of jeopardizing the country's financial stability. 59 5.33 The BOU must also process applications expeditiously, at the outside within the legal time constraints, keeping both the Minister of Finance and the applicant informed of progress. The BOU might also wish to reconsider the extent to which its Board is involved in the licensing process. As a matter of routine (say monthly), the BOU should provide the Minister with a brief report on applications outstanding and their status. Experience overseas shows that direct appeals to the Minister frequently comes from those applicants that need the closest scrutiny. Enhancing banking legislation 5.34 To strengthen the statute and reduce the risk of abuse the BOU should review its minimum capital requirement for private banks. A bare minimum reform measure would be to require that the minimum specified never be breached. Any new banks would therefore need to show initial capital to cover the specified minimum and all developmental costs until the banks reached profitability. 5.35 The BOU also needs to review the justification for granting bank licenses to a single or dominant owner. The experiences of many countries has shown granting banking licenses to operators who are essentially the financing arm of an associated business does little to improve the banking system and is fraught with danger. Unfortunately some such licenses have been granted in Uganda. 5.36 Uganda's laws should be amended to give the BOU the power to prevent certain individuals from being directors of a bank. Directors' primary responsibility is to protect the position of depositors. This responsibility, moreover, ranks ahead of any responsibility to shareholders. Problem banks 5.37 Problem banks must always be handled quickly because failure to do so inevitably increases losses and reduces the chance of rehabilitation. But by normal central banking standards, BOU action with regard to problem banks in Uganda has been neither adequate nor timely. 5.38 It must be acknowledged that the task facing the BOU is extremely difficult. The insolvent banks account for 60 percent of the system's total deposits. In such a situation many of the normal courses of action (such as sale to or merger with another bank) are not available. 5.39 Despite the closure of Teefe bank, the further use of the closure option is limited by the lack of funds needed to indemnify depositors. The BOU believes that depositors must be paid when banks close, both to preserve confidence in the financial system and because the authorities must bear some of the responsibility for the banks' difficulties. Banks in Uganda have been licensed too readily for some time. They have been inadequately supervised. The government's failure to pay its own accounts on time, moreover, exacerbates banks' difficulties. 5.40 While the BOU's capacities for on-site inspection and off-site monitoring are now quite satisfactory, it does not have staff skilled in the management of a bank. The option of appointing a manager to take over the running of a failing bank or at least to monitor its management is therefore also not open. 5.41 Virtually the only course open to the BOU for the moment is to establish performance agreements with banks. Several such agreements have been established in the past, and five are currently being prepared. It is admittedly difficult to establish agreements when a bank is weak and shareholders lack the funds or willingness to inject new capital. But even allowing for this, it 60 would appear that the BOU has been slow to act insufficiently firm when dealing with problem banks. It is therefore suggested that: * The modalities for depositor protection be settled between the BOU and the government. * The BOU seek technical assistance to help monitor and manage problem banks. * The BOU take firmer and more timely action. * The BOU plan what actions it should take in the future. 5.42 Protection ofdepositors. The BOU and the Ministry of Finance must concur on the policy and modalities covering the reimbursement of depositors in failed banks. In the Ugandan situation it would be unwise not to reimburse at least small depositors. The limit of U Sh 3 million specified in the deposit insurance arrangements would seem to be a reasonable cut-off point except for individuals with loans, in which case a net position should be established. 5.43 The major difficulty is that the Deposit Insurance Scheme presently has funds aggregating U Sh 568 million against deposits at high risk that approach U Sh 20 billion. The Scheme could not, therefore, cover the contingent liability it was designed for. Some months ago the BOU reimbursed depositors of a small bank that closed and could not provide the needed funds. The budget was therefore the only possible source to indemnify depositors. An early decision should be made as to whether or not the government is willing to provide funds for this purpose so as to open the way for appropriate action. If not, Uganda needs a plan to handle the consequences of allowing depositors to lose some of their funds after a bank failure. 5.44 Technical assistance. Uganda should recruit one or two banking experts, possibly under technical assistance arrangements, as quickly as possible to assist with the necessary detailed oversight and management of the country's five privately owned problem banks, which presently account for 10 percent of deposits. It must be made clear to banks that the BOU has the authority to introduce a new manager when existing management fails to rehabilitate a bank. 5.45 Firmer and more timely action. The BOU should closely examine its operations with regard to problem banks. Even service agreements have not been arranged as quickly as needed. Such agreements need to be firm and to have tight timetables and regular targets. While such agreements may not be able to retrieve the situation in all cases, chances for success diminish with delay. 5.46 Planning. The BOU should develop formal procedures to be followed for the handling of problem banks. Having a plan in place should make BOU action easier and speedier in the future. BOU managerial and operational capacity 5.47 Major improvements have already been made in the BOU's management, organizational, and operational arrangements. A more streamlined structure centered on six executive directors is in place. Departments are grouped on a sound functional basis. Capacity has been enhanced in many areas, by training and the use of well-qualified experts. For example, the inspection and monitoring, computerization and the maintenance of the BOU's accounts. 5.48 The strategic plan developed by external consultants contains a blueprint for improving the skills and performance of all BOU staff. It first seeks to introduce a results-oriented approach to the BOU, improve the BOU's ability to manage change, and introduce a work ethic based on productivity, accountability, and integrity. 61 5.49 Major difficulties stem from the fact that at present responsibility and accountability within the BOU is not well-defined. To a large degree, moreover, management skills are not well developed so that staff are generally reluctant to make decisions. Most have only a limited view of the BOU's aims as a whole or even in their particular areas of responsibility. Again, even senior BOU staff have little experience with current central banking procedures. Many adhere blindly to past Ugandan practices or are unduly influenced by casual contacts with outside advisors and consultants, who themselves have been of variable quality. 5.50 Improving and implementing the strategic plan. While the strategic plan developed by external consultants (ARA) and presently being reviewed by the BOU is generally a good management tool, it fails to define many new central bank responsibilities. Proposed business plans for various units mainly reflect present activities of the BOU. The strategic plan therefore fails to outline adequately the kinds of activities departments should develop in the future. 5.51 The strategic plan defines departmental objectives purely in terms of commercial business objectives. Such objectives, however, are not always paramount. The BOU, as Uganda's central bank, which frequently carries out national economic policy should not be operating solely to achieve profits. 5.52 An experienced central banker must review the BOU's overall strategic plan and the job descriptions for senior management and other critical positions at an early stage. 5.53 The strategic plan has already been the catalyst for substantial reductions in BOU staff. Under the voluntary program 500 staff members left by the end of 1994. Given the choice, the BOU would have kept only sixteen of those that left. Further staff reductions are being planned on a selected basis. Present expectation is that in less than a year staff numbers will have fallen from the present 1,700 to about 800. This reduction in staff has been a major advance for the BOU. Further reductions must have close regard for the on-going staffing needs of the bank to ensure that effective workers are retained. The emphasis should be on removing ineffective staff, staff with inappropriate skills, and those not willing to meet upgraded BOU work standards. Further staff rationalization requires sound job descriptions detailing the tasks to be performed over the next year or two. BOU job specifications should reflect, for example, tasks related to the development of financial markets and instruments. 5.54 Early emphasis should be placed on upgrading management, both in its organization and skills, particularly at the most senior levels of the BOU. As yet BOU job descriptions, responsibilities, and areas of authority have not been formally developed. It is also imperative that the BOU develop a training program for its senior management in both management techniques and accepted central banking practice. This program should include secondments to other central banks and the judicious use of experienced consultants and technical assistants known to be well- versed in central banking. It is suggested that the training program recently developed is not extensive enough and should be expanded. 5.55 The BOU must develop a time-bound action program to implement the strategic plan once it has been accepted. With most reforms having to do with improving management and organization already considered, the BOU's major remaining capacity issue will be how to obtain qualified bank workers and how to establish an environment conducive to high-quality work. To acquire a fully motivated staff, the BOU will have to pay close attention to training, recruitment, and organization. 62 Deposit insurance 5.56 Uganda's present deposit insurance arrangements are inadequate, so that budget funds would be needed to repay depositors in the event of bank closures. Yet Uganda's scheme was only recently established, and deposit insurance begun only when major banks were shown to be insolvent. This is much like establishing a housing insurance scheme at the time of an earthquake-neither is likely to have the resources needed. Inevitably in such a case, the obligations of the deposit insurance scheme will have to be met with funds from the government's budget. 5.57 Uganda's deposit insurance scheme will have inadequate funds for some time to come. An early review should therefore consider: * Whether it is realistic for the fund to cover deposits at insolvent banks. A strong case can be made to exclude them. * What premiums would be equitable and realistic. It is inequitable and inappropriate for sound banks to be made to support insolvent banks or poor performers by means of equal premiums. There are clear financial reasons why weak banks, on the other hand, cannot pay premiums that accurately reflect the risks their losses pose to their depositors. 5.58 The BOU must develop a plan for the establishment over time of a sound deposit insurance scheme and estimate the possible cost to the government of covering bank deposits until it is achieved. Close consideration must therefore be given to the level of deposits to be covered. Authority of the BOU 5.59 The extent of independent authority enjoyed by the BOU has been a contentious issue for a number of years. Since the 1990 Financial Sector Review, the BOU has appropriately assumed greater authority over the implementation of monetary and foreign exchange policy and over the supervision of banks. 5.60 Unfortunately, there are some recent signs that the authority of the BOU is currently being reduced. Licensing banks provides one such example. Beyond that it would seem that the current debate on recapitalization and the associated restoration of the income of the BOU to adequate leve's has been focusing heavily on external views on the detailed operations of the BOU. 5.61 Uganda's financial sector runs the very real risk of collapse if recapitalization is delayed while details of the strategy plan are worked out. The BOU's ability to intervene, furthermore, is hampered by the Ministry of Finance bias that treasury bills be issued only to fulfill their own funding requirements. One solution is for the BOU to issue its own paper-an option precluded by its own shaky financial position. It is recommended that government take immediate steps to make sure that the authority and independence of the BOU are not undermined. Development finance 5.62 BOU's role in development finance has never been clearly defined. At times in the past the BOU has been more heavily involved than was appropriate and has incurred significant losses. In recent years its participation has been limited on account of a specific emphasis on not carrying credit risks. Since central bankers are traditionally poor assessors of credit risk, this situation should continue. 63 5.63 There is some indication, however, that the BOU's involvement in development finance is increasing. Indeed, the strategic plan calls for the BOU to play a more active role. We recommend rather that the BOU maintain its present modest role with regard to development finance. Should the BOU become more involved with development finance, it should restrict its activities to facilitating and monitoring investment projects. Under no circumstances should the BOU undertake responsibility for credit risk. Summary Plan of Action 5.64 The following are the main actions recommended for Uganda to improve the performance of its central bank. The BOU'sfinancial position o The BOU should complete the examination of its own loan portfolio and have it reviewed again by an independent external accountant or auditor. o Government should assess what level of reserves is required to cover the BOU's foreign exchange and monetary policy operations. o Government should investigate ways to recapitalize the BOU, such as issuing bonds at rates agreed upon by the BOU and the Ministry of Finance. o An amendment to the BOU statute should establish arrangements for funding of any substantial foreign exchange losses in the future. o The BOU should review its own operations to establish appropriate reimbursement for' services provided (such as banking for the government) and to make sure that its operations are cost-effective. Prudential control of the financial system o The Government and BOU should publicly announce a moratorium on new licenses for two years with prespecified exceptions (including new banks wishing to operate in underbanked rural areas outside Kampala). o The government should establish and document for public information, the procedures required for processing applications for banking licenses. These must take into account the legal criteria for eligibility and specify what is required to satisfy the BOU with regard to section six of the Financial Institutions Statute. o The BOU should seek the services of people qualified to handle problem banks, including taking over the bank's management where necessary. o The BOU should undertake an analysis and disseminate the results of the impact of new banks on the financial and economic system, particular with regard to the level of investment and stability and performance of financial markets. o The BOU and Ministry of Finance should establish a policy for reimbursing depositors in failed banks. o The BOU should review the deposit insurance scheme and estimate potential claims on the budget to reimbursing depositors in failed banks. o The BOU should revamp its procedures to achieve more from the use of performance agreements with banks. (See chapter 2). o The BOU should develop a formal procedure for handling problem banks. 64 * The Financial Institutions Statute should be amended to establish appropriate minimum capital requirements for banks, to give the BOU the power to preclude single or dominant owners and reject certain individuals as directors of a bank. Strengthening the BOU * A reputable central banker should review the BOU's strategic plan. * This central banker should also review the job specifications for senior positions at the BOU. * The BOU should immediately develop and implement a program to upgrade senior management organization and skills. * The BOU should develop and implement a time-bound action plan for the full implementation of the agreed upon strategic plan. 65 6. Payment Systems 6.1 "Payment and settlement systems," notes the BIS 1994 Annual Report, "are to economic activity what roads are to traffic." Uganda's present lack of fast, secure, and efficient payment systems, hampers her economic development. People who do not trust financial instruments (such as checks) or who finds them inconvenient because of high charges and excessive delay conduct their business in cash. This leads to high costs for currency handling and printing, and depresses the money multiplier. On the other hand, the substantial government receipts and payments that pass through the banking system are subject to a system of verification and transfer that is cumbersome, costly, and difficult to monitor. Yet the technology and know-how now exist to improve the efficiency of Uganda's payments systems for relatively little cost. Background 6.2 The profile of payment systems in Uganda today must be viewed against the macrofinancial and historical background of the economy. The devastation of war and strife between 1972-86 and the shock of devaluing the currency so that U Sh 100 in 1972 was worth U Sh 7 in 1987 pushed the economy almost completely back to barter. In 1989 the currency/M2 ratio was almost 50 percent, one of the highest in the world. 6.3 But by 1994, this ratio had declined to 36 percent. Financial deepening (M2/GDP) increased from 6.7 percent to 9 percent between 1989 and 1994. However, Uganda's ratio of M2/GDP is still one of the lowest in the world. In that same period, demand deposits went from U Sh 25 billion, to U Sh 145 billion. Again, however, in relative terms, these amounts are quite small. 6.4 Clearly, although Uganda's financial sector started with a very small base, it has recently recorded encouraging patterns of growth. There is evidence that if the financial infrastructure of payments, clearing, and settlement systems is strengthened, financial deepening and transactions velocity can be accelerated still further, spurring sustainable economic growth. Institutional infrastructure and demand 6.5 Over the past seven years, the institutional infrastructure for delivering payments services (mainly bank branches) in Uganda has grown from 93 to 145. It is, however, still quite deficient both in terms of number and spatial distribution. The population per bank branch in Uganda is 115,000 as compared to about 15,000 for a highly populated country like India. The distribution of bank branches, moreover, is skewed along the country's east-west axis. Kampala alone has 42 branches, while thirteen districts have only one bank branch, that of the financially fragile UCB. The fragility of the two state-run institutions that dominate the branch network-the UCB has eighty-four branches and the Cooperative Bank has twenty-three-is also cause for concern. (At end 1994, UCB had an adjusted net worth of negative U Sh 81.82 billion and the Cooperative Bank had one of negative U Sh 1.1 billion.) Against such an institutional background it is important that inter-institutional arrangements (such as clearing houses) and communications technology be strengthened to optimize institutional availability. 6.6 On the positive side, the demand for payments services through the institutional infrastructure in Uganda is stable and growing. A BOU survey of all fifteen banks in the country showed that, even though banks are not providing attractive schemes for term savings, the demand 66 for transactions accounts is large. These banks together had 2,661 term deposit accounts, 155,783 current accounts, and 505,902 savings accounts. The fact that there was no withdrawal of deposits when banks slashed their average interest rates on savings accounts from 14 percent to 2 percent in fiscal 1994, reflects the country's lack of alternative avenues for saving as well as the absolute need for such accounts for purposes of transactions and safety. Even with no rise in interest rates there would be significant value added were these accounts to provide checking services or credit and debit cards and their number and usage were to increase. Instruments 6.7 Since Uganda's economy has been cash-based for so long, the country has developed little diversification or innovation in its payment instrument. Currency, not surprisingly, remains the predominant way to pay for transactions. 6.8 In Uganda, the amount of currency in circulation grew five times in as many years (going from U Sh 25 billion in 1989 to U Sh 135 billion in 1994). In industrialized economies, currency still accounts for 5 to 10 percent of the money supply, suggesting that currency will play an important role in Uganda's economy for a long time to come. For this reason, its quality and availability must be improved. 6.9 Currency printing and distribution costs-incurred by the BOU-have fluctuated wildly. The ratio of value printed to cost of printing in fiscal 1991 was 1.98; in 1994, it was 189.2; and in 1995, it is expected to be 4.76. Clearly, it must be rationalized. 6.10 Given the high costs of printing and distributing currency and its immobility as a medium of exchange, there also is an urgent need to bring in more modem payment instruments. One such instrument-the use of checks-is already growing steadily in both volume and value. During the last four years the number of checks cleared in Kampala went roughly from 1,900 to 3,000 a day, and their collective value rose from U Sh 4.04 billion to U Sh 13.05 billion a day-in spite of the fact that checks are issued only on current accounts, are frequently forged or backed by insufficient funds (and are therefore distrusted by the public), and clear slowly, especially between cities. Risks and legal infrastructure 6.11 In order to encourage the growth of noncash modes of payment it is important that Uganda's laws clearly define the rights and responsibilities of payer, payee, collecting banker, paying banker, clearing house, and the adjudication process in case of dispute. There is at present no such law in Uganda. The only penalty for bouncing a check appears in section 264(b) of the Penal Code, which mandates imprisonment or a fine of up to ten times the value of the check. The impact of such stringent sanctions on check usage is difficult to access. Another Bills of Exchange law dates from the time of British rule and is not quite relevant for today's financial transactions. 6.12 Until recently, the BOU was covering adverse clearing positions by automatic overdrafts so that systemic risks arising from the liquidity constraints of participant banks were not a feature of the Kampala Clearing House. Since that inefficient arrangement (from the angle of monetary control) was suspended in December 1994, however, four cases of gridlock have occurred. Under pressure from banks, therefore, the rule was modified, effective February 13, 1995 to give a member bank with adverse clearing balance until 4:00 P.M. to borrow from the market, after which time it can avail itself of an overnight credit facility. This, however, does not entirely remove the problem of poor liquidity and ensuing systemic gridlock, and better measures for managing settlement risk are still needed. 67 Technology and communications 6.13 A survey on the use of technology in Ugandan banks revealed that almost all use computers to some extent and that some banks use them extensively. In these banks, almost all staff are exposed to the use of computers. The BOU's computerization plan, which covers its eight cash and clearing centers, offers a good opportunity for upgrading the technology of funds-transfer in Uganda. In addition, the Uganda Post and Telecommunications Corporation (UPTC) has licensed four private vendors for the provision of satellite communication services. Since UPTC is to be split and telecoms privatized in fiscal 1996, it is now possible to commission a private telecom carrier. Government banking needs 6.14 Uganda's national budget for fiscal 1995 estimates a total expenditure of U Sh 512 billion, of which 75 percent is recurrent expenditure. The wage bill requires U Sh 125 billion and other recurrent costs take up U Sh 256 billion. Almost all of these payments are made by the issuance of BOU checks to such relevant authorities as education officials, who distribute school teachers' pay. But the disbursement of funds from this point onwards is inefficient from the view-point of the final recipient, the financial system, and the government. A BOU check does not fetch an immediate credit but has to go back to the BOU for counter-checking that it is, in fact, an authorized government payment. Any advantage for the financial system of payroll credit is also lost, since most employees immediately withdraw cash. This makes operations both unwieldy and unsecure, and the government has difficulty tracking cash flows even two to three weeks after disbursement. 6.15 The UCB handles all revenue receipts, collecting an estimated 80 percent of the government's total revenue of about U Sh 480 billion. Assuming uniform flows, the UCB collects about U Sh 32 billion a month. This collection is transferred to the BOU at the end of the month, giving UCB an average float of U Sh 16 billion at all times. If government distributed this float among several banks, it could leverage a lot more services for itself and bank clients. Financial markets 6.16 Uganda's financial sector is small and consists primarily of commercial banks. But a more varied financial market could grow with stronger financial infrastructure. The country's money markets, consisting presently of the inter-bank market, are driven by contingencies of reserve requirements and seasonal demands for cash rather than by more complex funds- management techniques designed to balance liquidity and return in optimum combinations on a daily basis. A facility to make daily trading and quotations possible would almost certainly induce a more active money market in Uganda. 6.17 Since the change in clearing house rules in February, the treasury bills (T-bills) market is likely to experience additional demand. Banks whose adverse clearing balances bar them from the money market can now borrow from the BOU against the lodgment of treasury bills. But this concession is not likely to stimulate more than random spikes in the trading pattern. The dematerialization of scrips and required book-entry of transfers would also be helpful. A depository system can work as an interim measure until legal enactments recognizing electronic data as evidence in law are put in place. 68 Organization and human resources 6.18 The relative progress in familiarity with computers among employees in the banking sector in Uganda is encouraging. But the understanding of payment systems issues, especially their importance in enhancing the efficiency of the financial sector and necessity of risk-containment measures, is little understood even among decision-makers. Development assistance in this area is necessary, as is the need for a body (committee or council) that can address payment systems issues comprehensively. This council, consisting mainly of banks and the BOU, would address a far broader canvas of issues than the Clearing House Committee. It would have ownership of technical, legal, and operational reforms. Options for Improving Uganda's Payment Systems 6.19 There are three major issues that need to be addressed to encourage the habit of writing checks and to make sure that checks be cleared quickly and securely. Greater use of checks would hasten financial deepening. How to minimize settlement risk 6.20 Settlement risk is the risk that a bank with an adverse clearing on any given day would not have sufficient reserves to cover its net deficit and still meet its cash reserve requirements nor would it be able to borrow from the market to cover its liquidity gap. In such a case, the only option is to reverse the insolvent bank's entries, which could affect other banks facing liquidity crises, since they would get no credit on checks presented on the deficit bank. Settlement risk can be eliminated only when the central bank steps in to provide temporary overdrafts covering the net deficit of the member bank. This, however, leads to uncontrolled expansion in money supply, and is therefore the least preferred solution. 6.21 Up to December 1994, the BOU had been providing automatic overdraft to cover adverse clearing balances. While there was no settlement risk, therefore, Uganda had almost no control over monetary expansion. From December 1, 1994, when the BOU stopped granting overdrafts, a member bank has to either borrow from the market or face suspension from the clearing house the next day. Because this rule waz too stringent, four cases of gridlock had occurred by February 1995, and the rule was relaxed. A member bank must now find funds by 4 :00 P.M. (clearing takes place between 10:00 A.M. and noon), after which it is allowed to borrow overnight from the BOU against the lodgment of treasury bills. Failure to do the latter would lead to exclusion from clearing the next day, and then review of membership. 6.22 The arrangement now in force, while somewhat more flexible, is neither fool-proof nor secure. Given the poor solvency, liquidity, and rampant insider lending in some of the smaller banks, a clear moral hazard issue is at hand. If a bank cleared large checks issued by inside firms and then revealed a massive liquidity deficit at the clearing house, it could precipitate a systemic crisis. To avoid that, the BOU-or depositors in the final reckoning-would have to pick up the bill. 6.23 Various methods for minimizing settlement risk have been adopted in different parts of the world: only major banks with benchmark levels of net worth and assets may be admitted as members of the clearing house. Other banks must seek a correspondent relationship with approved settlement banks. Based on periodic reviews of the volumes presented for clearing, each member bank will be required to maintain collateral with the clearing house equivalent to a certain ratio of 69 average daily volumes. Settlement lags must be reduced to the minimum level possible. Member banks must work out multilateral and bilateral net debit caps in relation to each other. Where banks estimate breach of these limits, they would have to postpone, at least partially, the clearing of instruments. Liquidity-pooling and loss-sharing arrangements would have to be devised among settlement banks (where membership is restricted) to avoid settlement failure by decoupling temporary illiquidity problems from more serious problems of insolvency. How to speed up clearance of checks 6.24 The clearing process extends from the lodgment of checks by clients to sorting at the recipient bank, presentation at the clearing house, settlement among banks, and finally credit to the client's account. The faster this process is the more attractive the check becomes as a payment instrument. The usual benchmark is the US Expedited Funds Availability Act of 1987 which obligates banks to make funds available to clients on the second day following lodgment (written in the literature as D+2) in the case of local checks and five days from lodgment (D+5) in the case of out-station checks. In many advanced economies, this benchmark is exceeded. 6.25 In Kampala and eight other clearing centers in the country, local checks are cleared in three days following lodgment (D+3), an improvement as of December 1, 1994, before which it was D+4. Yet intercity checks can take anywhere from a week to twenty days to clear, except between Kampala and Jinja, where a courier arrangement makes it possible in D+3. 6.26 Several options are possible for speeding up the clearing process for local checks: o Automate sorting of checks by using MICR (Magnetic Ink Character Recognition) technology. All banks would use checks with code numbers printed at the bottom in magnetic ink. Special machines (called reader-sorters) could then read and sort these checks in banks. o Computerize settlement so that the information read off the MICR checks (or entered manually in the case of ordinary checks), would be stored on magnetic media and transferred to a computer at the clearing house, which could then calculate the settlement obligations of the banks. o Truncate the journey of the physical check at the receiving branch by means of OCR (Optical Character Recognition). MICR information would be read off at the branch where the check was presented. The digitized information would then be transmitted to the bank's clearing center either on magnetic media or over the line. The clearing center would convey all such information to the clearing house in a similar manner, where computerized settlement would take place. 6.27 In the case of intercity checks also, Uganda has several options: o Once all of a bank's branches are computerized and linked by telecommunications lines, it is possible to present checks drawn on an out-station branch of that bank in local clearing, (after balance confirmation) over the line. The physical instrument can be dispatched later. o Once clearing houses are computerized and linked by telecommunication lines, truncated information can be sent to the clearing house where the check was drawn, and the item included in the local settlement. o Where clearing houses are linked by satellite communication channels and scanners are available, images of the checks can be transmitted to the concerned center, 70 instruments verified by the branch on which the check is drawn, and presented in local clearing. 6.28 The solution chosen by Uganda will largely depend on the country's relatively low volume of checks (3,000 a day in Kampala; 150 local and 30 intercity in Jinja; 100 local and 20 intercity in Masaka) and the availability of suitable technology. Check use 6.29 In economies that can afford more advanced technologies, plastic and electronic instruments have substantially replaced such paper instruments as checks, drafts, and payment orders. When volumes are large and the physical delivery, handling, and sorting of checks become a major problem, financial intermediaries try to popularize plastic and electronic forms of payment. Many developing countries are now debating whether to skip the check phase and go straight for plastics and electronics, or to traverse the usual payment system evolutionary route. 6.30 Most Ugandan banks, still struggling to reach respectable levels of size and capital, are not likely to invest large sums in technology to introduce more modem payment instruments in the near future. The rate of growth in savings and financial deepening of the economy are also still low, suggesting that encouragement is needed to increase the use of checks as a payment instrument for the time being. Factors that discourage check-writing need to be minimized and positive features to be capitalized upon. 6.31 The factors that presently inhibit check usage in Uganda include: fear of frauds and forgeries, high service charges by banks, the non-issuance of checkbooks for savings bank accounts, and slow clearance for intercity checks. Steps that could be taken to promote check use include requiring checks to be bilingual and readable by MICR and to require that the withdrawal of funds from banks beyond a certain level (say U Sh 100,000) not be made in cash. Creating Infrastructure for the Development of Financial Markets 6.32 At present Uganda does not have an explicit and transparent legal framework that defines the rights and responsibilities of contracting parties to a financial transaction nor does it provide a mechanism for its effective enforcement. It is also not advanced in the use of information technology. 6.33 In the case of payment, clearing, and settlement systems, Uganda needs two major laws: a comprehensive enactment on negotiable instruments and an enactment on the admissibility of electronically stored and transmitted information as evidence in law. Models of both these laws are available in developed economies, as well as substantial literature on the experience with such laws. The availability of worldwide expertise and experience in the development of such laws presents a major opportunity for developing countries. Uganda would do well to take advantage of this opportunity. 6.34 Information technology could also help Uganda to expand and deepen its financial markets-particularly for payment, clearing, and settlement systems-since the price of such technology has declined rapidly. The following opportunities are clearly visible: * Establishment of a real-time high-value funds transfer system (to minimize settlement risk, enable banks to better manage and fine-tune their liquidity, and make the transfer of funds immediate and secure) would support the growth and efficiency of Uganda's money and securities markets by enabling immediate settlement of that leg of transactions in these markets that relates to transfer of funds. 71 * Establishment of an intercity communications network between Kampala and the eight other centers where the BOU has opened cash and clearing centers (to clear intercity checks in two days, provide real-time information and accounting of cash balances that would enable businesses and banks to locate outside Kampala, conduct simultaneous treasury-bill auctions at up-country centers, and transmit management information, data, and images among banks for an appropriate fee). * Establishment of a depository and book-entry system for securities (initially treasury- bills). 6.35 Real-time, high-value, finds transfer system. To establish a real-time, high-value funds transfer system, Uganda could set up a private communication carrier in Kampala linking the BOU with all banks. This would mean setting up an electronic exchange, terrestrial lines, and a minicomputer at the BOU-perhaps as a common system with the one being proposed for the BOU under EFMP. It would also require necessary software, training of personnel, and personal computers (with modems) at all the banks, many of which already have some computer technology. Besides capital costs, there would also be recurrent costs for maintenance of the exchange and salaries for two or three staff members employed for the purpose. The advantages of this set-up would be high levels of security, and fault-tolerant operation. The system, moreover, would be owned and run by the BOU and the banks together. 6.36 Alternatively, banks could lease dedicated lines from UPTC to link the equipment mentioned above. Each bank that wanted to hook up to the HVTS (high value transfer system) would have to lease a line at a cost of approximately US$ 2, 000 a year. That, however, would be the only recurrent cost. Capital costs would include only the minicomputer (and could be the same as proposed under EFMP) and the software at the BOU. The advantage of this option is its low cost, but system reliability would depend on UPTC service. Security could be built in with necessary codes and encryption, but some possibility of bugs and computer viruses would persist. Ownership of this network would be distributed. 6.37 Intercity communication system. To establish a communication system linking the BOU in Kampala to the eight up-country cash and clearing centers, the following equipment would have to be installed at each of the eight centers and Kampala: two personal computers, a laser printer, a fax machine, and a scanner. Including software, the total cost for all centers would be about US$ 150.000. 6.38 For the communication carrier, Uganda could use existing UPTC telephone lines. These, however, are highly unreliable, with up-time in centers like Arua sometimes less than 50 percent. 6.39 Alternatively, it could lease dedicated lines from UPTC. But lease charges and maintenance costs are estimated at US$ 320,000 a year, and reliability and security would be uncertain. 6.40 Radio communications could be used with high frequency radio transceivers and data communication equipment. But terrain deflection of radio waves requires the installation of transceivers at short distances and performance can be affected by bad weather. 6.41 Finally, Uganda could use satellite communication with a VSAT (very small aperture terminal). The advantage of VSAT is its flexibility and versatility. It can carry data, voice, and image simultaneously-which would require more than one network using telecom lines or radio communication. Four private vendors have been licensed to sell satellite communication channels in Uganda. The vendors would invest in building a switching hub at the cost of US$ 1.5 to 2.0 million. These costs would be recovered by sale of VSAT dishes and annual operating costs. For 72 servicing nine centers, the capital costs would be about US$ 200,000, with annual operating expenses of about US$ 75,000. 6.42 The VSAT proposal is the preferred solution for reasons of reliability, security, and relatively low cost. It has already been proposed under the EFMP. 6.43 Securities, depository, and book-entry system. To set up a depository and book-entry system for securities, Uganda could introduce a limited system for treasury-bills and other government securities only or a larger system that could also take care of the large-value holdings in the equity and debt markets that are likely to come up in the near future. It is too early to think of screen-based trading systems. 6.44 The capital cost for a large system would not exceed US$ 50,000, but Uganda would also have to hire a systems manager and train staff. The system would be housed at the BOU. Leveraging Government Receipts and Payments for Financial Deepening 6.45 The aggregate annual amount of government receipts and payments that pass through the banking system is of the order of U Sh 800 billion. Compared to the total assets of the banking system at (U Sh 460 billion), this figure is large even after recognizing that assets are a stock figure and receipts and payments represent a flow figure. Should these funds continue to circulate as deposits within the banking system, their effect on financial deepening and the money multiplier will be immense-with, of course, the requirement of more active monetary management to offset the inflationary impact of government spending. 6.46 Factors that limit the circulation of government funds in the banking sector include the fact that most recurrent government payments (such as wages) are withdrawn from banks in cash. Cash has a tendency to circulate outside the banking system until it becomes worn out physically. Ugandans withdraw cash because confidence in checks is low and banking services are largely unavailable except in major centers. 6.47 On the receipts side, the Uganda Revenue Authority itself does not accept checks toward payment of government dues but rather asks for deposit by cash or draft. Drafts being expensive, people are forced back into dealing in cash. Checks are not accepted, however, because if a check bounces, it is too cumbersome to chase down the defaulter. This cycle perpetuates the use of cash in Uganda's economy. 6.48 The treasury department maintains that majority of the government payroll is met by direct credit to bank accounts, but the actual process is cumbersome. The Uganda Computer Center (UCC) draws up the payroll by the twelfth of each month. The payroll is then sent to respective ministries for confirmation. The ministries confirm it and raise vouchers to print the requisite numbers of checks. The UCC then prints the checks and sends them back to the government departments where they are signed by authorized personnel. Each government agency then receives the checks along with a list of employees' banks and account numbers. Checks are deposited first with the agency's banks, generally the UCB. They are then sent for clearing to the BOU Kampala, and only thereafter credited to the appropriate institutional or individual accounts. 6.49 While the development of better services to bank customers will ultimately depend on competition within that market, the government can use the massive float it allows the banking system to both extract better services for customers and to deepen the financial sector: The Uganda Revenue Authority (URA), or the Treasury, could periodically auction government revenue collection to banks. While allowing a limited number of days of 73 float, government can ask banks to accept checks toward tax payment. The bank would first issue a temporary receipt acknowledging deposit of a check, and after the check goes through clearing, the actual receipt for tax collection can be issued. o All parastatals and government bodies (such as electricity, water, and sanitation agencies) could be required to provide the option of direct debit to customers. But these organizations would then have to give an assurance that their billing was at least 95 percent accurate. Once it is automated, the clearing house could perform this direct payment function. o For direct credit of payroll, the UCC should receive all changes to payroll from each government agency by the twentieth of each month. It could then generate a final list of payees' bank account numbers, get the list authorized by the Ministry of Finance and hand it over to the BOU. Two days before pay day, the BOU could debit the government and credit the employees' and institutions' banks at that day's clearing, handing over banchwise lists of payees to each bank. The banks would then transmit the lists to their branches, and employees' accounts would be credited on pay day. Improving Currency Use 6.50 Given the cash dominant nature of the Ugandan economy, the broad strategy has to aim at reduction in cash usage. However, since cash will always be used to some extent, it is important that the quality and availability of cash is improved in as cost-effective a manner as possible: o Except in fiscal 1994 when the U Sh 5000 note was introduced, the cost of currency printing has been high. To normalize currency costs at reasonably low levels, it is advisable to phase out such small-denomination notes as the U Sh 10, 20, and 50. The denomination which has maximum circulation, (probably the U Sh 1000), should also be issued as a coin, which would increase its life many times over. o The incineration of soiled and mutilated notes is done at Jinja. But the incinerator is old and polluting and has now broken down. Presently the notes are being destroyed at the rate of twenty-five bags a day in a make-shift incinerator near the river at Jinja, but there is an accumulated backlog of 2,000 bags. The BOU should immediately purchase two or three large shredders (one as back-up), and switch to shredding of notes. Shredded paper can be sold to paper-pulping units. o To improve the availability of currency in up-country centers, cash and clearing centers need to become linked through a communication network. A few more centers could be opened if necessary. Then radio communications or terrestrial lines could be used to link the centers with bank branches. o The BOU has already worked out trigger levels for cash replenishment at the cash and clearing centers. These could be refined to account for seasonality, which is quite pronounced, by the use of demand-forecasting models. Providing Payment Services in Underbanked Areas 6.51 There is a definite issue regarding the adequacy of payment, banking, and credit-delivery systems in Uganda which has been exacerbated by the closure of over half of the UCB branches over the past two years, mostly in the rural areas. Even granting fresh bank licenses would not necessarily lead to the opening of fresh branches in rural areas. And while credit-delivery could be improved by strengthening such local institutions as credit unions and (Nongovernmental organizations) NGOs, the provision of services for savings and remittances remains critical; 74 especially as a recent household survey revealed that 22 percent of rural incomes derive from remittances. 6.52 The only other institution that has substantial nationwide presence is the post office. There are 312 post office branches but of these, only 126 contain Post Office Savings Banks (POSB). The infrastructure available at these offices, moreover, is rudimentary, and the staff would need training. Also, while the UPTC falls under the purview of the Ministry of Works, Transport, and Communications, the POSB system is owned by the Ministry of Finance but uses post office premises and employees. The POSB system itself has only 35 employees all located in Kampala. 6.53 In order to extend banking services through the Post Office (however inadequate its infrastructure at present), the government may consider the following changes: * A hundred strategically located post office branches could be linked through VSAT to provide postal giro services (direct fund transfers among account holders). With necessary modifications, the maintenance of accounts and transfer of funds could be modeled on the Swedish giro system. * Either the POSB could manage postal giro by maintaining the central accounts at Kampala and paying a fee for the use of post office premises and personnel at the other centers, or the POSB could be merged with the Post Office, which would then provide the giro services on an integrated basis. * The estimated total expenditure for the one-hundred postal giro offices, including capital costs and training, would be US$ 4 million. The viability of the scheme would depend on its ability to generate sufficient business to recover capital costs (over, say, twenty years), and to cover annual operating costs. As a rough estimate, average deposits of U Sh 3.5 to 4.0 billion yielding a return of 10 percent are required to break even. Given the fact that small banks with just one or two branches are already able to mobilize deposits of U Sh 10 billion, this target does not appear difficult. Strategy 6.54 In light of the possible solutions for problems with Uganda's payment system discussed above, the following measures can be recommended as part of Uganda's overall financial sector strategy. To reduce settlement risk 6.55 Given the degree of insolvency and illiquidity throughout the banking system, especially in the case of some of the smaller banks, the risk that a bank will not be able to fund its net debit clearing balance on any day (settlement risk) is presently very high. Settlement failure could also easily snow-ball into system-wide collapse. Control of risk at the clearing house should therefore take top priority. Of the risk-control measures followed in various parts of the world, the following combination is recommended for Uganda: * Limit membership in the clearing house in a two-stage process. First, allow only banks with a benchmark asset base (of say, U Sh 20 billion) to be admitted as direct members. Other banks can set up correspondent relationships with members. At the end of a year, allow only those members to remain who meet Uganda's new core capital requirements. Oblige new banks to start with correspondent relationships. 75 o Require each member bank to deposit eligible collateral at the clearing house (eligibility to be decided by the Clearing House Committee) equivalent to its highest net debit clearing balance during the previous year. o Reduce settlement lag by shortening the clearing cycle and moving high-value transfers to a real-time gross settlement system. To reduce clearing time 6.56 To contain risks associated with lags in the clearance and settlement of checks, and reduce time taken in clearing checks, Uganda should institute the following measures: o Computerize settlement work at the Kampala Clearing House and, in stages, at other centers. o Allow clearing to take place at 4:00 P.M. each day, and return clearing at 8.30 A.M. the following morning. Given the relatively small volume of checks in Kampala (roughly 200 per bank per day), and even more so in the eight other centers, this will enable banks to credit their clients within one day, or at most two, from the date of lodgment of the check for collection. o Introduce check-imaging technology and communications (with adequate security safeguards) to reduce the time taken to clear intercity checks at BOU's eight cash and clearing centers. To encourage the use of checks 6.57 The use of checks should be encouraged to help deepen the financial sector.. The Government may want to consider the following: o Require banks to issue check-books for savings accounts provided that the balance in the savings account remains above U Sh 50,000. o Discontinue such bank charges as ledger and check fees, which are presently set by the Uganda Bankers' Association (UBA) and approved by the BOU. The cost of bank inefficiency on the assets side of their balance sheet should not be transferred to the liabilities side. o Require banks to make checks bilingual, and as volumes increase, require that checks be MICR-coated for automated sorting and reading. o Legislate and enforce procedures for prompt action in the event of check fraud to reduce public distrust. To establish a real-time gross settlement system (RTGS) for the transfer of high-value funds among banks 6.58 Real time implies immediate settlement of transactions. Gross settlement is the concomitant debiting and crediting of accounts by transaction, rather than the present system of aggregated, or net, settlement of all transactions at the end of the day. The benchmark for high value could be fixed by mutual consultation among banks and the BOU. A suggested benchmark would be U Sh 0.5 billion. A RTGS would significantly reduce settlement risk, increase banks' ability to fine-tune liquidity, allow for real-time monitoring by the central bank, and support the development of Uganda's money and securities market. 6.59 A RTGS system need not necessarily be high-tech or expensive. One low-cost solution is to install a minicomputer at the BOU and to link it to personal computers at different banks by 76 dedicated telecom lines leased from UPTC. The capital cost for computer equipment at the BOU would not exceed US $50,000, with the added major expense of application software. Banks that already have personal computers would only need a US $2(0 modem. Annual lease charges per dedicated line from UPTC are estimated at US $2,000. To establish payment infrastructure for the development offinancial markets 6.60 The infrastructure of payment, clearing, and settlement systems are the bedrock of financial sector development. Three components of this infrastructure must be expeditiously enacted or installed: * Pass an integrated law on negotiable instruments that defines the rights and obligations of payer, payee, collecting bank, paying bank, correspondent bank, clearing house, rapid dispute redressal mechanism, deterrents for frauds and forgeries, and definition of electronically stored and transmitted data that can be considered as evidence in law. Several models of such laws are available, and an expert could be commissioned to put together the features of each that are the most desirable for the Ugandan financial system. * Establish a communication system linking Kampala BOU and it's eight cash and clearing centers through VSAT (very small aperture terminal). Satellite communication is the most efficient and cost-effective communications option available. Already proposed under EFMP as a BOU network, it could be slightly expanded and modified to serve as a carrier for other banks too, on payment of a small fee. * Set up securities depository and book-entry system. This would essentially consist of a minicomputer system at a cost of not more than US$ 50,000, the training of staff, and employment of a systems manager. It could also be used for initial deposit and accounting of private equity and debt issues for the proposed stock market. To leverage government receipts and payment 6.61 It is estimated that 80 percent of government's total revenue (about U Sh 480 billion) is collected by the UCB. This is a figure roughly equal to the total assets of the banking system. If we assume uniform flows and the fact that collections are transferred to government's account at BOU at the end of the month, government receipts cause an average daily float of U Sh 16 billion in the banking system. To obtain the maximum leverage out of this float in terms of financial services, both for itself or for the general public, government should invite bids from banks for collection of its revenue receipts. 6.62 On the expenditure side 75 percent of government's budget of U Sh 512 billion consists of recurrent expenditure ( wage bill, U Sh 125 billion; other recurrent U Sh 256 billion). Almost all of these payments are made by issuance of BOU checks to the relevant authorities (such as to education officials to distribute school teachers' pay). But from this point onwards, all BOU checks go back for verification, resulting in delays of one %..:ek to three weeks and funds are general withdrawn immediately in cash, vitiating the effectiveness of the money multiplier. It is therefore important that check usage be made more attractive by the measures suggested above and better safeguards be effected before issuing government checks. 77 To fill in the gaps in the nationwide payment infrastructure 6.63 Although 22 percent of rural income presently derives from remittances, the reach of bank branches serving this vital need is far from sufficient. Government therefore needs to flag this issue for strategic decision making. If banking sector reform is not expected to expand the network of bank branches, and ways for supporting such intermediaries as credit unions are unclear, it would be advisable to revitalize one-hundred critical branches of the Post Office Savings Banks. A postal giro system based on the Swedish pattern merits serious consideration. As a rough estimate, the equipment and training needed to set up the postal giro in a hundred critical locations would cost US $4 million-relatively little when compared to what would be required to recapitalize sick public sector banks. To set up the training and organization needed for payment system development 6.64 It is necessary to set up a National Payments Council (NPC) made up of representatives from banks, other financial institutions, and the BOU to oversee the modernization and reform efforts relating to payment systems in Uganda. The NPC could subdivide its areas of endeavor into technical, legal, and regulatory, and operational reform. 6.65 Expert personnel in the areas of payment systems (including clearing house, the legal aspects of finance, information, and technological training) will be needed to implement measures suggested above as well as a well-thought-out program for training banking sector staff. Technical assistance from the World Bank could be useful to Uganda for these purposes. Action Plan Urgent actions (suggested completion by September 1995) o Re-define the rules for clearing-house membership, obtain collateral from settlement banks, and change clearing hours to reduce clearing and settlement lag. o Appoint a payment systems expert to coordinate and advise on clearing-house reorganization, work with consultants to draft new laws regarding negotiable instruments, and seek expert advice on the functional specifications needed for a real- time, gross settlement system and communications carrier. o Have the BOU in consultation with the Uganda Bankers Association constitute a NPC and vest in it the authority to oversee the modernization of the payment system. o Draw up a plan for training selected banking personnel in the use of computers for banking applications that involve communications and networking software. o Invite bids from legal consultants and appoint a firm to draw up draft legislation regarding negotiable instruments, including proposed action against frauds and forgeries and to permit the use of electronically stored and transmitted data as evidence in law. o Government should hold informal consultations with various banks to assess their position with regard to the collection of government revenue receipts and the services they are willing to offer in return. o The BOU should draw up functional specifications for computerization of settlement work at the clearing house and for the communications infrastructure linking Kampala and the eight other centers. 78 * The BOU should finalize funding arrangement for computerization of settlement work at clearing houses and for the main system (at Kampala) for the communications network. Actions in the short term (suggested completion by April 1996) * BOU in consultation with UBA, should instruct banks to issue check-books on savings accounts above the threshold level, to reduce or eliminate ledger and check fees, and to introduce other measures to encourage the use of checks. * Technical specifications for communications carrier systems, to be installed at clearing houses and at BOU, should be set, bids should be invited, and systems should be purchased and installed. * The BOU should negotiate contracts with vendors of satellite communication channels to install VSAT dishes and other communications equipment. * The NPC should make arrangements with banks that wish to be linked through a RTGS system, purchase a minicomputer and install it at the BOU, load software, and negotiate the lease of dedicated telephone lines with UPTC. * A bill on negotiable instruments should be drafted, examined, discussed, and passed into law through the prevailing legislative process. * Technical assistance should be sought to define functional and technical specifications for a securities depository and book-entry system, which should then be examined and amended as needed and-once found acceptable (particularly with regard to equity and debt issues)-passed into law. * Experts in the areas of information and technology training should be appointed to train selected personnel in communications software and other areas with banking applications. * Government should invite first bids for revenue collection by banks and award a contract for the work for the coming fiscal year. * Government should make a final decision regarding the revival of the POSB (post office savings bank) for the installation of a postal giro. If government decides to restore the system, it must begin negotiations to arrange funding for the project, and invite bids from project management consultants. Actions in the medium term (suggested completion by June 1997) * RTGS trial runs should be completed, the system modified accordingly, and live runs begun. * Systems for a securities depository and book-entry system should be procured, a manager appointed and personnel trained. (Trial-runs may go beyond June 30, 1997). * Trial runs of the communications network between BOU and Kampala and its eight cash and clearing centers should be completed, the network modified accordingly, and live runs begun. * Training of bank personnel in the use of computers and communications software for banking applications should continue. * Government should review bank performance in revenue collection and those with better records may bid for collection contracts for the next three years. * Implement the postal giro (if so decided). 79 7. Rural Financial Markets 7.1 The vast majority of Uganda's general population and most of its poor work in agriculture. Farmers frequently need money to buy essential inputs, to produce and market their products, and to weather the long lead-time separating planting and sale. Yet few Ugandan smallholders have access to credit. Over-extended and burdened by bad debt, UCB has recently had to close many rural branches, making it all the more critical for government to take immediate measures to determine what level of coverage the rural community needs and to expand the depth of the country's rural financial markets. Background 7.2 Agriculture is the mainstay of the Ugandan economy. In 1991 it accounted for 51 percent of GDP and over 90 percent of exports. Eighty-nine percent of the population is rural, and 80 percent of the employed are engaged in agriculture. Agricultural output comes almost exclusively from about 2.5 million smallholders-80 percent of whom have less than 2 hectares each. Only tea and sugar are grown on large estates, which total 40,000 hectares. 7.3 Food crop production dominates the agricultural sector, contributing 71 percent of agricultural GDP. Livestock products contribute another 17 percent (average 1989 to 1991). Other subsectors are small. Export crop production is only 5 percent of agricultural GDP; the fisheries subsector accounts for 4 percent; and forestry for 3 percent. The degree of monetization also varies across subsectors. Only a third of food crop production is marketed, as compared with two-thirds of livestock production, and all of the export crop output. 7.4 The household characteristics of rural dwellers are important in discussing rural finance, since most rural enterprises, farms, stores, and processing operations, are family-owned and operated. Financial aspects of such family-run operations are not separable from other family budgeting decisions. Financing rural enterprise development then becomes a question of funding the family budget. To assess this market properly the rural family should be viewed as an enterprise in itself, rather than dealing only with the specific business venture under consideration. 7.5 The National Household Budget Survey (1989-90) found that: o The average rural household consisted of 5.6 persons, compared with 4.5 for urban dwellers. o The average yearly expenditure per capita in rural areas was about US$ 104, compared with US$ 218 per capita in urban areas. o In rural areas, 71 percent of total expenditure went for food (including beverages and tobacco) compared with 51 percent in urban areas. 7.6 The range of possible options for delivering the financial services for rural enterprise development is greatly affected by the availability of opportunities for employment and revenue generation outside the farm. Few farm families rely on farming for all of their income. The rural household survey found that, amongst the better-off rural dwellers, some 27 percent of total income came from off-farm activities, both employment and enterprise management. For poorer rural dwellers, about 22 percent of their total income came from transfers and remittances. Improvements in the efficiency with which these remittances are transferred would therefore be of direct benefit to the rural poor. As of now, although local remittances are handled at least partially by the informal network, long-distance remittances are dependent on the formal financial system 80 for transfer. A general increase in investment and economic activity in rural towns and villages could improve the farmers' access to employment and should also improve rural financing mechanisms. Improvements in transport and information services could have as much effect on a farm family's cash flow as a direct loan. Ruralfinancial markets 7.7 In considering customers in rural financial markets, it is useful to distinguish between farm and off-farm enterprises. Farming enterprises include both large commercial enterprises and small (generally family-owned and operated) concerns. Off-farm enterprises also include large commercial ventures and small (usually family-owned) enterprises with less than ten employees. 7.8 Because the lag between the purchase of inputs and sale of output can be from four months (for annual crops) to seven years (for some tree crops), the rural community has to rely on external financing for its normal production needs. Normally this finance comes from equity sources. For very small enterprises, it generally finances consumption of the farm fanily during the growing season. Larger farms and off-farm enterprises will typically have access to the formal banking system and be able to interact with the financial system according to standard commercial procedures (such as opening deposit accounts well over the minimum size and providing collateral for loans). The funding that flows to and from such larger enterprises will be a function of business opportunities in the region. 7.9 Specialized financial products. It is in the interaction between the financial system and smaller farms and off-farm enterprises, where innovation may be required. Since these enterprises and their owners rarely meet standard minimum access requirements into the formal financial system (minimum deposit size, collateral for loans, a bank record for loans), temporary external intervention may be justified to introduce and test new techniques. 7.10 The market for the delivery of financial services to small and micro enterprises in rural areas and rural towns is large. Small enterprises are the largest source of employment in Uganda, albeit at low levels of marginal productivity and income. They are responsible for all of agricultural GDP and a large share of nonagricultural GDP in rural areas. Most of Uganda's poor, moreover live in rural areas. Measures to improve the sustainability, efficiency, and outreach, of the rural financial system could substantially improve income prospects for the rural poor as well as contribute to Uganda's overall agricultural growth. 7.11 The small size, dispersion, and the lack of public information about the ownership, assets, and operations, make it both costly and risky to deal with them commercially. Yet work in this field has shown that such enterprises are willing to pay full price for the financial services that fit their needs. An informal financial network almost always springs up where the formal system is not accessible to small business. Formal institutions that emulated the products, delivery, and monitoring systems of the informal system could charge full price for their services and eventually be profitable. By the same token, nonfinancial institutions able to go between the formal financial sector and rural client enterprises, to provide the desired services with the desired flexibility are able to recover the full cost of intermediation. 7.12 Covariance. An important feature of lending to,farm enterprises is the seasonal nature of the financing requirements. All customers need money at the same time, and all make deposits at the same time. The success of the farm enterprise in any given season is also linked to the weather, especially in Uganda, where irrigation is negligible. If the rains are inadequate, all farmers and their creditors are in cash-flow difficulties together. To minimize the negative effect of 81 such correlations, financial-institutions active in farm finance should diversify their assets across various types of enterprise, and across agro-ecological regions. 7.13 Access. Small borrowers will have a need for deposit and loan instruments as well as fund-transfer services. But the size of each transaction is likely to be small, making efficient, low- cost processing and administration procedures within the financial institution doubly important. Another important feature of small, family-enterprise finance is the premium put on ease and guarantee of access to loan and savings facilities. To be attractive to this kind of customer, who is often poor enough to be severally affected by reductions in consumption levels, quick turn-around must be a key feature of any financial instrument. But if these features are available, rural customers are willing to pay for them. It has been shown in various parts of the world that viable rural financial institutions can be created based on the principles of easy acres, speed and low cost. 7.14 Intermediation outside financial institutions. Due to high transaction and monitoring costs, most farmers do not rely on direct transactions with banks to fund production. They are reached mainly by finance available to the commercial enterprises they deal with in the course of business: crop-purchasing enterprises, and tradespeople providing consumer goods and agricultural inputs. These may include marketing and processing concerns trading in agricultural commodities, cooperative unions and primary societies, farmers groups, and associations of different kinds. Thus the coffee trade, although centralized in a few large coffee export enterprises in Kampala, affects a large share of Uganda's rural population through the purchases of their agents. A few large loans to the Kampala based exporter will result in a flood of liquidity into thousands of rural households that supply coffee. Purchase mechanisms for tobacco, cotton, and tea have a similar effect, although their outreach is less broad and their finance and purchase mechanisms are different. 7.15 A series of small financial institutions operating on the fringes of the commercial system (such as rural savings and credit cooperatives, and various non-governmental organizations) also take deposits and provide loans. But fund transfer services are provided exclusively through the commercial branch-banking network. Ruralfinance services and institutions 7.16 The main financial products or services required by rural Uganda include financing for crop purchase and processing, crop production and rural enterprise, trade and commerce as well as deposit taking (mainly demand and short-term savings deposits), fund transfer services, letters of credit, and other fee based activities. But even with such services available, the bulk of the finance actually used for the purchase of agricultural produce would still be obtained through large contracts negotiated in Kampala, where the large enterprises that process, market, and export Uganda's industrial crops have their head offices. The branch networks of the relevant commercial banks would be important, primarily to facilitate transfer of funds. 7.17 Commercial bank deposits and advances. Deposit mobilization in Uganda's rural areas has been low. In 1993, deposits from UCB's non-Kampala branches represented some 37 percent of its deposit base, and UCB has the largest active branch network in Uganda. The other banks with a significant -but much smaller-rural branch network, the Cooperative Bank, raised some 45 percent of its deposits in branches outside Kampala. Overall, rural deposits made up perhaps 10 to 15 percent of the country-wide total of U Sh 301 billion in mid-1994. While the amount of deposit mobilization by the 540-odd rural savings and credit cooperatives is not known, it is likely to show up eventually as deposits in the commercial system. If deposit rates were raised and 82 appropriate instruments put in place, significant increase in rural deposits could very probably be mobilized. 7.18 Most of the commercial bank credit used in rural areas is actually provided through large loans negotiated with processing and marketing firms in Kampala. Coffee traders or unions, for instance, negotiate bank loans in Kampala to support their domestic operations through to point of export. But smaller production loans to farmers for the purchase of inputs and labor not linked to large trading and processing companies are generally negotiated at branches. This is especially true for smallholder farmers. Some 30 percent of all advances from the commercial system goes into agriculture-related activities, divided more or less equally across production, processing, and marketing (table 7.1). 7.19 On June 30, 1994 loans outstanding for agriculture and crop finance totaled some U Sh 67 billion. UCB had 63 percent of this business and Cooperative Bank some 12 percent, Barclays and Baroda the only other two banks with significant interests in the sector, held some 15 percent of the total advances to agriculture. 7.20 Financing in rural areas goes beyond any narrow definition of agriculture. The majority of the loans at rural branch banks, for instance, were for trade finance. If only 10 percent of these loans were negotiated at rural branches, an additional U Sh 5.3 billion would be available to enterprises in rural areas, bringing formal rural lending to U Sh 72 billion. 7.21 The largest single business in rural finance is the financing of coffee purchases, processing, and export marketing. UCB had the lion's share of the business until fiscal 1994 when coffee buyers started turning to off-shore banks. Of an estimated U Sh 75 billion obtained from financial institutions in 1994 for crop purchase.and marketing (mainly coffee), perhaps 45 percent was provided from non-Ugandan banks. Coffee traders appear to have moved overseas to obtain finance, driven by the high real interest rates charged by domestic banks. While the Cooperative Bank continues to finance the purchases of the Cooperative Unions, their share of the market is declining. Because prices for coffee have risen so dramatically at all points in the marketing chain, the actual amount of financing obtained from the domestic financial system has not yet decreased that dramatically. But as coffee prices drop the competition from off-shore sources will have a more powerful effect on the demand for funds in the local market, perhaps causing a decline in margins. 7.22 The very large margins (due to very high lending rates) currently prevalent in rural markets should, if they remain, enable even fairly ineffective banks to lend profitably, while supporting expansion of the branch network. It is unlikely, however, that such high real lending rates (18 to 22 percent in an inflationary environment of 7 to 8 percent) will persist. The only banks likely to escape the downward pressure on rates are specialty banks, such as the Centenary Rural Development Bank or the Cooperative Bank. 7.23 Centenary has been developing savings and loan products that respond closely to the preferences of the low-income segment of the market in Kampala and in rural towns. The response has been fairly successful. Since it is almost alone in this niche, Centenary should be able to charge above-market rates for its services for a while. 7.24 The Cooperative Bank has a close relationship with Uganda's network of rural primary cooperative societies and their unions. By working through the better primary societies and unions, it has been able to provide production and marketing finance more or less profitably for a large number of smallholder farmers. Because these farmers have few other options, it can charge above -market rates for this service. Funding of cooperative unions' trade requirement has provided 83 liquidity in rural areas and helped smallholders, but since this is a much more competitive business, interest charges are likely to be much more closely related to commercial rates. 7.25 Rural branch banking. The UCB has a commanding rural presence. At end 1994 it operated sixty branches and fifty-three associated agencies outside Kampala in Uganda's thirty- nine districts. Even UCB managers acknowledge that this network is too dispersed. A critical review of branch profitability and activity conducted by the UCB itself recommended the closure of the fifty-three agencies and three rural branches, which were found to duplicate existing installations. This would still leave at least one branch per district, a network that, if well managed, would provide a strong base for national banking activities. 7.26 The Cooperative Bank has twenty-one rural branches, a network that provides reasonable access to the whole country. While activity is small and the quality of the portfolio in these branches is not good, the Cooperative Bank feels that its network is about the right size to provide service to its corporate clients and the cooperative sector around the country. 7.27 The bank with the healthiest rural network is Bank of Baroda, which has six non-Kampala branches. Branch responsiveness to client needs appears good, loan monitoring procedures appear sound, and the ratio of advances to deposits is high, which-if the loans are good-should make for profitable operations. Centenary Rural Development Bank has a small network of seven branches outside Kampala. These branches are working hard under new management to consolidate their business and improve their monitoring and management systems. Centenary does not intend to expand coverage until the consolidation is complete and financial solvency has been attained. 7.28 Lending from each branch in the rural branch network is well below 50 percent of the deposits mobilized. This is normal in countries with reasonably well developed financial systems, where rural financial resources are mobilized primarily to finance higher-return, lower-risk enterprises in urban centers. Most of the UCB and Coop Bank branch network lend out a very low share of the deposits mobilized (39 percent for UCB in late 1993; 28 percent for Coop Bank ). While the advances-to-deposit ratio in non-Kampala branches is above 50 percent for the Bank of Baroda, its overall impact is less significant. 7.29 The use of the financial system for transfers and payments is an important function of the rural financial network, and fee income for such transfers generates a high proportion of branch income. In fact, monopoly branches with a minimal staff located in areas with reasonable economic activity could be justified on the basis of their potential fee income and deposit collection activities alone-assuming that the bank as a whole could place the funds. 7.30 Uganda's banks should therefore be able to operate a network of rural branch banks profitably. While this network would have to be substantially smaller than that currently administered by UCB, it should still be able to reach the main areas of rural economic activity with reasonable efficiency. Individual branches need to be assessed in terms of their contribution to the total network, in terms of deposit mobilization and fee-income generation. The transfer pricing for funds raised (a function of the returns the head office could obtain from funds mobilized) would be a key determinant of branch profitability. An efficient, well-connected network of branches would allow for country-wide fund transfers and financial services which are now available in Kenya and were available in Uganda before the nationalization of the financial system in the early 1970s. 84 Specialty banks 7.31 Uganda's rural financial services could be provided by financial institutions set up in towns and rural areas to target small borrowers with specially designed and financially sustainable lending and deposit-taking procedures. Around the world such financial institutions as the BRI in Indonesia, the Grameen Bank in Bangladesh, the BancoSol in Bolivia, and ADEMI in the Dominican Republic have succeeded in providing a range of financial services to poor clients and turned a profit. The methodology is based either on: (i) closely supervised small, short term loans to customers whose alternate source of funds is the (very high cost) informal market, or (ii) closely supervised loans to small groups of customers, who then administer the disbursement, and collection of these funds independently, with the group as a whole accountable for the loan to the financial institution. 7.32 By being responsive to such customers' needs, using straight-forward assessment and simplified disbursement procedures, quick turn-around on loan, and offering responsive savings instruments (with immediate access and no limits on size or the number of withdrawals), such banks have developed a loyal clientele among the rural poor. Run well and with adequate margins, such operations are commercially viable. Good administration keeps defaults low, and margins are set to cover the higher cost of a large number of small loans and small savings deposits. 7.33 The Centenary Rural Development Bank is a bank owned by sixteen Catholic dioceses in the West and Southwest of Uganda and a French development agency. It was started as a nonbank financial institution in 1985 to assist the Church with its development work. It was recently given a banking license and is now allowed to offer normal checking and savings accounts services. Centenary is active in Kampala and in five districts in the South and West of Uganda. It holds some 40,000 accounts, both savings and checking, and its assets in October 1994 were U Sh 7.2 billion, of which U Sh 3.2 billion was in advances. 7.34 Since 1991, the bank has been receiving technical assistance financed by the German Savings Banks Association. With this assistance, a new portfolio of some 890 loans (with US$1.7 million outstanding) has been developed targeted at small enterprises in small towns and rural areas. The average size of loans outstanding (US$ 1,900) is small, although not as small as some of the "micro" lending programs in other countries. Lending is not collateralized. By putting in place a strict monitoring and client assessment mechanism, the bank has been able to keep repayments up. Of the new portfolio, only 10 percent of the principal has been in arrears for over thirty days. Deposit-taking services have also been streamlined and fashioned to meet the needs of the small savers outside of Kampala who are neglected by other commercial banks. 7.35 The Centenary Rural Development Bank is currently insolvent. With a net worth of negative U Sh 1.3 billion, it is struggling to meet the requirements of the Financial Institutions Act by the deadline of December 1996. The institution is working off the consequences of many years of poor lending procedures. Various donors have agreed to contribute to its capital, and the village-based owners are also increasing their shares. But time will be needed for Centenary to reach solvency and profitability 7.36 Uganda's Cooperative Bank is another specialty bank that has traditionally been the main source of finance for rural primary cooperatives and their unions. During the 1970s and 1980s, the cooperative system owned and operated Uganda's main rural enterprises. The unions, which are second-level cooperatives owned by the primary societies, themselves owned and operated the country's main coffee factories and cotton gins. But during the economic chaos of the late 1970s and early 1980s, most of these enterprises slipped into insolvency and civil war under government- 85 fixed prices and margins and weak management. The decline of the cooperative system, moreover, carried their financing agent down with it. With a large proportion of its portfolio in default, the Cooperative Bank is presently insolvent. 7.37 During the 1990s Uganda changed its policy with regard to cooperatives. Under the 1991 Act, cooperatives were given increased autonomy and made more responsive to their owners. The Cooperative Bank was also readjusted to try to put it back on a viable financial footing. 7.38 As of June 30, 1994, the Cooperative Bank had assets of U Sh 30.5 billion with net advances of U Sh 13.8 billion. The institution is technically insolvent. The USAID has been providing technical assistance and funding in an effort to bring the organization back to solvency. The Cooperative Bank currently has technical assistance staff in key line-management positions and is following a well-defined course toward profitability. The USAID provides an equity contribution to the bank every time a new milestone is accomplished. 7.39 The current plan is for the Cooperative Bank to become the premier rural bank in Uganda by: * Developing its network of affiliate primary societies to serve as rural intermediaries and increase the outreach of the financial system into rural areas (an important government policy objective). * Diversifying outside of its traditional clientele to include large, commercially viable (noncooperative) ventures, thereby providing a more stable source of income and improving the quality and diversity of the balance sheet. * Attempting with the assistance of USAID, to develop a network of village banks to provide financial services to small rural enterprises with the backing, guidance, liquidity management and supervision of the Coop Bank's branches. This network would be modeled on Indonesia's successful BRI. 7.40 Success in establishing a village bank network would go far toward the development of a sustainable rural banking system capable of increasing financial outreach and providing services to the rural poor. Under the Coop Bank's strengthened management, progress toward solvency, deposit mobilization, and diversification appears to be proceeding well. 7.41 Uganda has other institutions (FINCA, Uganda Women's Finance Trust) that also target their loan and deposit-taking services to clients who would not otherwise interact with a financial institution. By doing this, they are bringing these clients into the cash economy, thereby increasing their liquidity and overall financial savings within the country. These organizations intend, eventually, to link up with a commercial bank so that the financial transactions they oversee will one day enter into the formal system. Most of these organizations, however, are unable to cover operating costs from their income and are dependent on donations from Nongovernmental organizations (NGOs), donors, or the GOU for continued operation. Issues and Options Government support for rural lines of credit 7.42 The Bank of Uganda recently completed a study of government attempts to support lending to farmers in thirteen cotton-growing districts. The study confirmed that, across the board, specialized lines of credit have been poorly administered and have achieved very poor recovery rates. None of these activities were shown to be financially viable and are sustained now only with continued injections of predominantly government funds to cover losses and operating costs. The 86 only exceptions are instances of reasonable wholesale loans to good cooperative primary societies and unions. The study shows conclusively that targeted lines of credit where banks do not carry the credit risk are unlikely to be well administered, since banks and borrowers alike regard government loans as disguised grants. Loans are therefore subject to lax approval criteria and efforts at collection are desultory. It appears that it is not lack of liquidity that slows bank lending in rural areas but a combination of poor loan appraisal and administration, the perception that funds from government are not subject to strict repayment standards, and unrealistic ideas on how to make cost-effective loans in rural areas. Use of these government lines of credit, moreover, has undermined efforts to place the financial system on a sound commercial footing. 7.43 It is therefore recommended that such lines of credit be avoided in future, unless there is conclusive evidence that shortages of loanable funds (especially funds for medium-term lending)- and not the availability of good quality borrowers and a bank's capacity to work with them-are hampering lending operations and business development. Such a line of credit is also permissible to support a one-time process of institutional change designed to address gaps in the financial system, or to provide services for a new category of high-priority clientele such as the rural poor. But even in cases such as these, access to government funds should be used sparingly, within the capital adequacy constraints of the intermediating bank, and only when mixed with funding from the financial intermediary. All government lines of credit programs should have strong economic incentives built in to encourage profitable administration and eventual phasing out of the programs. 7.44 The Entandikwa Program. A particular instance of a government-funded line of credit soon to begin operation is the Entandikwa Program. The objectives of this scheme are laudable. It will support small enterprise development in communities in rural Uganda, significantly reducing rural poverty. But this effort runs the risk of undermining the possibility to develop rural financial institutions that could provide similar support on a sustainable, self-financing basis. 7.45 The proposed credit scheme relies on the District and County Administration and Intermediary Agencies for project and beneficiary assessment, disbursement of funds, monitoring of project implementation, and loan collection. But the recent Rapid Appraisal of Rural Finance and Credit Schemes carried out by the Agriculture Policy Committee found that: * Administrative costs for the project are likely to be high (over 75 percent of the amount disbursed in one case). * Recovery rates are likely to be low (ranging from 20 to 80 percent). * Provision of financial services (including deposit taking, lending, and fund transfer) is best carried out by financial institutions specializing in servicing different types of business and clientele. 7.46 For the provision of financial services to the rural population to be sustainable, the institutions that provide these services have to be able to recover costs by charging their clients for them. The main charge applied to lending services is the interest rate. For example, the Centenary Bank, an institution that targets clients similar to those served by the Entandikwa scheme, has to charge 25 to 30 percent on its loans to break even, since administrative costs are 16 to 18 percent, loan-loss provisions amount to 6 to 8 percent, and the cost of mobilizing funds is 3 to 4 percent. Microenterprise lending programs around the world, including the Grameen Bank which has administrative costs of 16 percent of its loan portfolio, has a similar cost structure. By lending funds at 12 percent, the Entandikwa proposal runs the risk of undercutting the market for competing, potentially sustainable credit schemes. 7.47 Operating a program which is likely to have low recovery rates also weakens the climate for sustainable credit administration. Increasing the costs (default losses) financial institutions 87 have to cover and putting downward pressure on returns (interest rates) reduces any possibility of putting in place sustainable financial institutions. 7.48 It is recommended that the Entandikwa program go forward given the laudable objectives of the program and the government's commitment to it. In order that the program not affect the development of sustainable financial intermediation in rural Uganda, however: o The program should be converted into a community grant scheme. o The savings requirement should be converted into a matching contribution (the share defined) required from the individual or group before receipt of the grant for the proposed project. Specialty institutions and branch banking 7.49 Specialty financial institutions provide financial services to poor customers who nevertheless produce an extremely large share of the GDP. Fostering the development of branch banking networks and helping to develop innovate approaches to the delivery of financial services in difficult market niches is a legitimate public function. It is recommended that: o All institutions that desire to operate as banks be held to the requirements of the Financial Institutions Act. o Where government intervention is required or financial services have to be downgraded, restructuring should take into consideration the social objectives of specialty institutions. The goal is to design organizations that can deliver the desired services to the target clientele using procedures that make them financially viable. The Centenary Rural Development and Cooperative Banks 7.50 The financial performance of Centenary Bank and Coop. Bank would seem to be a reasonable basis for the continued development of these two institutions, which could play an important role in addressing the needs of small entrepreneurs, both rural and urban. Free access to the influence and information base of the Catholic Church (in the case of Centenary) or the cooperative system (in the case of Coop Bank) are assets that, if wisely used, could support wide- reaching operations in selected rural areas. 7.51 While the future of Centenary Bank and Cooperative Bank as separate institutions should be left to the decision of potential investors who are considering making contributions to their equity, the Government should take steps to ensure that, should these specialty banks have to be closed down, the portfolio of 'new' loans and the investment in training, systems development and banking expertise made in the personnel in both banks does not disappear, but rather continues to be used in financial institutions with similar objectives and target clientele. The country can only profit from the continued development of the kinds of commercially viable banking practices, targeted at the rural and urban poor, which both Centenary Bank and Coop Bank are pioneering in Uganda. The effect of the sale of UCB on the rural branch network 7.52 As mentioned above, the UCB dominates the rural financial market at present. The current reassessment and down-sizing of the UCB has already resulted in the closure of many of its agencies and branches outside of Kampala. These outposts, located outside of district capitals in far-flung villages, are frequently not profitable and would not have been maintained by any 88 commercial bank. The sale of UCB would result in the closure of still more branches, many in district capitals or large commercial centers outside of Kampala. 7.53 It has been argued, however, that a reasonably well-distributed network of branches is needed to support the development of a cash economy and foster enterprise development around the country. The restructuring of UCB could result in a reduction of its branch network to a level well below the minimum strategic requirement for such a network. Outside of UCB, the bank with the largest rural network is the Cooperative Bank. It may be the case that Cooperative Bank will expand into districts and markets vacated by UCB without the need for additional governmental intervention. 7.54 It is recommended that when the status of the UCB branch network is known-that is, concurrent to or immediately following the final disposition of UCB assets and liabilities-a study be undertaken of the rural financial market to determine: * A minimum set of locations where basic financial services (fund transfer, deposit taking) should be made available. * What institutional alternatives are most appropriate for providing these services. * Whether a special government subsidy or incentive scheme is necessary to ensure that these services are provided, and how such a government intervention should be administered to ensure that the network was commercially managed. 7.55 The conclusions of such a study would clearly depend on the expansion plans of the Cooperative Bank and other commercial financial institutions, the strength of interest in taking over assets and liabilities after UCB's withdrawal, and the need to encourage competition in the rural financial market. 7.56 In order not to prejudice the possibility of putting in place a rural bank capable of continuing to supply rural financial services in the absence of UCB, UCB branches not closed under the restructuring plan and not purchased by new owners should continue to operate pending the outcome of the rural banking sector study. 7.57 Following the successful sale of UCB, the remaining branches, (including assets and liabilities) would be assessed in light of the study's recommendations. Branches, with assets and liabilities determined to be strategically important would be incorporated into the proposed framework for the provision of rural financial services while those not required for this purpose would be liquidated. This study is not a precondition, however, and should in no way interfere with the proposed sale of UCB. The large monthly losses being incurred by the continuous operation of UCB require that it be sold as soon as possible. Reform Strategy 7.58 To be effective, Uganda's rural financial system must be well integrated into the national commercial financial network. Its health will ultimately depend on the health of the financial system at the center and depend on the national investment climate, as much as on the economic and social infrastructure of the region. 7.59 A viable and sustainable rural financial network would include: * A base national network of one or more commercial banks capable of providing the liquidity, deposit services, loan services, and currency transfer systems, check-clearing and payment mechanisms needed to support the development of a cash economy. 89 o A regional network of local financial institutions (not necessarily full-fledged banks) such as cooperatives, savings and credit unions, and village banks or associations that are tied into the commercial network yet have special ties or knowledge of smaller clients. o Specialized financial institutions that provide services to particular niches in the rural financial market, custom designing deposit and loan instruments, and maintaining information on targeted clientele. 7.60 Important preconditions for the success of a rural financial system includes: o Pricing financial instruments and services to cover costs. o Designing deposit-taking, and loan services offered by the local financial institutions to respond to the savings and borrowing needs of the particular client (including those who are poor or who operate small or micro-enterprises). o Putting a mechanism in place (possibly through the commercial bank network) for supervising the operations of the local financial institutions to prevent excessive default and maintain the integrity of the financial system. o Strengthening the legal means for banks and enterprises in rural areas to enforce contractual agreements at a reasonable cost and with the expectation of a fair judgment within a reasonable amount of time. 7.61 The proposed strategy stresses the development of specialized, sustainable financial institutions and relationships in rural areas. The strong assumption is that rural enterprise can and should bear the full cost of efficient financial intermediation. Measures should therefore be taken to ensure that such intermediation is competitive and that risks to institutions that engage in this activity are kept to a minimum. Rural financial markets 7.62 Government's objectives in supporting the development of rural financial markets should be to encourage the expansion of coverage and depth within the bounds of prudential responsibility and commercial viability. To support rural financial market development, government should provide a regulatory environment and the basic infrastructure needed to support investment in financial institutions, and the development of rural enterprises, agricultural and otherwise. These standard regulatory functions-well recognized and easily carried out in urban areas (especially in a financial market as concentrated as Uganda's)-becomes more expensive and difficult to put in place in rural areas. In view of the widely dispersed locations, and high cost of transportation and information transfer in a rural financial system, designing supervisory mechanisms that are cost- effective becomes crucial. Ruralfinancial institutions 7.63 To support the development of rural financial institutions, government should: o Provide the financial infrastructure needed for common use such as currency transfer systems, and an efficient check-clearing and payment mechanisms. Providing financial infrastructure may also mean ensuring the presence of a minimum network of financial institutions across the country. o Provide prudential regulation of deposit-taking institutions to avert negative consequences for both depositors and the financial system in the event of collapse of a bank. 90 * Intervene where there is some kind of market failure or where financial markets have not yet been developed to facilitate the design and introduction of new instruments, operational approaches, and institutions. 7.64 When financial institutions can be tailored to provide services to the rural poor and yet still return a profit, social returns will be high. Government should take special steps to: * Provide incentives for the development of techniques for bringing small depositors and borrowers into the financial market that are also commercially viable. * Provide long-term funds (on market terms) only where the financial system has not yet developed maturity transformation and where term funding is desirable; especially in agriculture and agro-industries because of the long gestation period of investments. * Develop a viable insurance market against the risks (such as drought or systemic default) that characterize rural enterprises: Rural enterprises 7.65 Given the high costs of interacting with a dispersed and poor rural population, it makes sense to use the information base and commercial interactions already developed by nonfinancial rural enterprises to provide financial services, particularly credit. The development of off-farm enterprises helps bring rural dwellers into the cash economy. Farmers, too, engage in a variety of enterprises related to the supply of agricultural inputs and consumer goods and to the demand for agricultural products for processing and resale. Measures to be taken by government to promote rural enterprise include: * Improving such public goods as roads, communications facilities, market infrastructure, electric power, and water supply. * Improving security of land tenure and smoothing transactions involving property rights in rural and small urban areas. * Improving the effectiveness of the judicial system to enforce contracts and settle ownership disputes. * Improving security against theft (a feature that adds greatly to the cost of banking in rural Uganda). Grant-funded assistance (with sunset provisions) could be used to improve the business concepts and administrative system, needed for these institutions to become profitable. If correctly designed and run, such institutions could be the link through which the financial and commercial markets meet the needs of the rural poor. 91 . fInterbank, Treasury BW And Foreign Exchange Markets Background 8.1 In the last few years Uganda has entered into the process of deregulating financial markets. Interest rates were deregulated in November 1992 and the foreign exchange market was unified in November 1993. As a part of this process the Bank of Uganda is implementing monetary control through indirect methods and auctioning Treasury Bills (T-bills) in order to keep the quantity of base money in line with a growth path determined by the reserve money program. 8.2 A deregulated financial sector and a central bank using Indirect Monetary Control (IMC) both require the presence of a liquid interbank market in the local currency and a liquid secondary market in T-bills. The banks need these markets to manage liquidity and interest rate risk while the central bank needs the markets to help transmit policy. 8.3 But at present the Ugandan interbank and T-bill markets are highly illiquid. This is having several undesirable effects on the financial system. Illiquidity is causing banks to hold very high levels of excess reserves, which raises the cost of intermediation. The high demand for excess reserve balances also reduces the demand for T-bills, leading to a thin primary T-bill market characterized by volatile interest rates. The volatile primary market T-bill rates, in turn, are causing uncertainty, leading some banks to reduce the maturity of deposits they are prepared to accept, which reduces their ability to lend long term. 8.4 The thinness of the interbank and T-bill markets is also weakening the interest rate transmission mechanism between the BOU auctioning of T-bills and bank lending and deposit rates. This relationship is further weakened by high cost structures in banks in Uganda. 8.5 The foreign exchange market is subject to lumpy flows occasioned by aid money and coffee exports. These lumpy flows, combined with the newness of the market, led to significant exchange rate volatility in 1994 that necessitated BOU intervention. 8.6 At present, Uganda is running a domestic budget deficit financed by overseas funding, and the fact that the deficit is funded offshore makes the Treasury reluctant to see T-bill sales rise. There is a need for close coordination between the BOU and the Treasury, particularly since offshore financing is potentially inflationary. Interbank Shilling Market 8.7 The interbank market in shillings is extremely limited. Although there appears to be a rising trend, no accurate data on the volume of transactions is available. A rising trend is consistent with the announced changes to Uganda's clearing-house rules, scheduled to take effect from December 1994. These changes are intended to make access to BOU credit extremely difficult, forcing the banks to trade among themselves for liquidity. However, the rising trend in interbank debt may not indicate an increase in interbank transactions (Graph 8.1). Conversations with the banks suggest a highly irregular interbank market, fluctuating between U Sh I and 3 billion. 8.8 The lack of liquidity in the interbank market means that banks rely on their own deposit bases to fund their clearing needs. Almost all Ugandan banks think in terms of retaining an amount of excess reserves on hand sufficient to cover their largest expected adverse clearing. In a more 92 developed market banks would keep a much smaller amount on hand and would rely on borrowing in the interbank market to meet larger needs. IndectiveIntebank Transactions vs BOU Finance 12 --------------------------- 10 \ .JD 6 I 2 0 9.93 11.93 t94 3.94 5.94 7.94 9.94 11.94 1.95 Note: Inteibank debt includes unsettled clearing claims &depo Graph 8. 1 8.9 At a very approximate estimate, Ugandan banks have a voluntary demand for U Sh 17 billion of excess reserves to meet clearing needs. Excess reserves earn nothing at the BOU. Based on the ninety-one-day and 182-day T-bill rates, the opportunity cost of maintaining these_voluntary balances is between U Sh I to 2.5 billion a year. This clearly raises the cost of intermediation in Uganda. A more liquid interbank market and greater liquidity in T-bills would enable banks to reduce the quantity of excess reserve balances required for clearing purposes. 8.10 The lack of liquidity in the interbank market has six main sources: * Banks do not have long-standing relationships with each other. * The market is segmented on the basis of credit quality. * Many banks do not hold T-bills which they could offer as security for interbank loans. * There is a general excess of liquidity in the market which has existed since mid 1994. * The demand for interbank funds is limited to clearing needs at present. * Bank treasury management is formative. 8.11 Before December 1994, when the clearing house rules were changed, banks did not need interbank relationships with each other. These relationships take time to form. Further, the quality of Ugandan banks varies considerably. Now that the banks need to form relationships, they are attempting to judge the credit worthiness of other banks, and are encountering problems in doing this. The quality of auditing in Uganda is low, and banks do not trust each others' accounts. Many banks, furthermore, have poor balance sheets and represent a high credit risk. No bank acting on prudent commercial criteria is going to lend to such a bank on an unsecured basis, if at all. 8.12 Secured interbank lending is beginning to develop. Stanbic has secured lending relationships with two local banks, and Standard Chartered is in the process of developing such relationships. UCB and Cooperative Bank are also prepared to lend on a secured basis. Yet as Graph 8.2 shows, six out of fifteen banks hold no T-bills. This presents a barrier to the growth of secured lending. 8.13 Many Ugandan banks do not hold T-bills even when they hold more excess reserves than they require for clearing (banks combined have held between U Sh 25 to 45 billion in excess reserves compared to U Sh 17 billion estimated demand) for a range of reasons having to do with the T-bill market. The main reasons are that T-bills are illiquid and, further, banks have difficulty in bidding for the full quantity they would like to. 93 8.14 A virtuous circle could be created to improve both the interbank and T-bill markets. Improving banks' ability to bid for T-bills would assist the development of a secured interbank market, and a more reliable interbank market would reduce demand for excess reserve balances, increasing demand for T-bills. Ban TB ill Haidings, Der 94 9000 - - - - - - - - - - - - - - - - - - 8000 7000 6000 5000 .~4000 S3000 2000 1000 Nile UCICB ile UCBKigezi COOPIc Orient rd Greefilrust Tropical Sembule Cerudet Barc rt n Graph 8. 2 8.15 Uganda recently adopted the policy that if a bank cannot meet its clearing obligations, it is shut out of the clearing house. This policy comes at a time when many banks are in a weak financial position and the interbank market cannot be expected to fund weak banks on an unsecured basis. If a weak bank holds no T-bills for collateral and cannot obtain unsecured funds to meet its clearing obligations, the implications of the policy are that the BOU must be prepared to enforce the bank's recapitalization, to enforce its orderly exit, or to lend to it. 8.16 A discount window for BOU advances secured over a banks' statutory reserves would provide liquidity to banks that hold no T-bills and cannot obtain unsecured funds. The rate on this window should be priced so as to discourage banks from using it as a source of funds. Effectively, it would be a window of last resort. 8.17 A rate based on a margin above the average maximum lending rate of all banks would achieve this objective. Currently this rate would be over 30 percent. Any bank making use of such a lart-resort funding facility clearly has significant problems, which require investigation. Frequent use of the window (more than once a quarter) should therefore initiate an investigation by the Bank Supervision Department resulting in action to rectify the problem. 8.18 The financial sector's lack of long-standing relationships and general air of mutual distrust have led some banks to be suspicious of any bank that approaches them for funding on a regular basis. Knowing some banks are weak but lacking more specific information, banks review requests for interbank funds as a sign that the requesting bank is in trouble. This way of thinking is antithetical to market development but is unfortunately understandable given the conditions in Uganda. The problem of extreme segmentation in the interbank market caused by the presence of very weak banks will only be fully solved by the banks concerned being made solvent or exiting the market. 8.19 Lack of interbank liquidity also reflects the general excess of liquidity in the market since mid-1994 (Graph 8.3). This surfeit of liquidity reduces the need for banks to trade with each other to meet their clearing needs. Since mid-1994, there has been between U Sh 10 to 25 billion more excess reserves than the banks require to manage their clearing in the current illiquid state of the market. This means that on most occasions banks can meet their clearing commitments out of their 94 own funds without approaching the interbank market. If the level of excess reserves were managed Excess Reserves, end of Month by the BOUso that it were much 4s --------------------------, closer to the transactions demand 40 j Total for excess reserves (estimated at U 3 Sh 17 billion), banks would have 25- EstVied to trade with each other much o _ _ .I Tr--s-c--o more frequently, and the interbank 10 market would become more liquid. 0 8.20 While managing the level 9.93 1.94 3.94 5.94 7.94 9.94 1.95 of excess reserves in relation to _. demand is a desirable objective, Graph 8. 3 the BOU should delay adopting this policy until the liquidity of T-bills and the interbank market improves somewhat. In particular the adoption of a target level for excess reserves close to estimated demand should be delayed. Adopting this policy too early would cause severe liquidity problems for the weaker banks, which could be difficult to resolve. 8.21 Banks can also use the interbank market as a permanent source of funds when they face high bankable demand for credit relative to their ability to mobilize deposits. In such cases, banks can turn to the wholesale market for funding. In Uganda, this motive for using the interbank market has not developed yet. 8.22 The treasury management skills of many Ugandan banks are weak never having developed under the old, unliberalized environment. Improving banks' ability to manage their liquidity and interest rate risks would not only improve the operation of the interbank market but would also reduce the level of risk throughout the banking sector. A basic course in treasury management could be arranged through the Banking Institute. 8.23 Training in treasury skills is particularly important for UCB. On average, UCB holds 41 percent of the banking system's excess reserves (Graph 8.4) Although Average Excess Reserve Positions. Sept. 93-Ja. 9 UCB is subject to 1s -------------------------------------- significant adverse clearings, it generally gains from clearing due to 8 its role as collection agent A for the Uganda Revenue Authority (URA) and its branch network. Conversations with UCB indicate that it has a -- - - - N-*2Ir - - - - - - - - uce particularly conservative Sernb rudeb G TUFop areeanrso a Ca strategy for managing Graph 8. 4 liquidity. It is important 95 that UCB be fully skilled in treasury management so that the reserves it has are available to circulate in the market on a commercially prudent basis. Poor treasury management at UCB could compromise the BOU's ability.to target the level of excess reserves. 8.24 The lack of liquidity and the segmented nature of the market are also revealed in the diversity of interest rates that coexist. In a liquid market there should be a single rate for each maturity of loan, which should lie somewhere between the cost to the borrower of obtaining funds from the central bank and the return to the lender on deposits at the central bank. The BOU's present discount policy is unclear to the banks. But since the previous discount rate was 14 percent, many banks still think in terms of this rate. There is no return on excess reserves at the BOU. The market now appears to be shifting into three brackets of interest rates, all between 0-14 percent, based on perceived credit worthiness. 8.25 The four foreign banks (Standard Chartered, Stanbic, Barclays, and Baroda) lend interbank to each other at a rate based on the ninety-one-day T-bill rate, although each bank calculates the rate a little differently so that it may vary by a percent or so among banks. The foreign banks lend to banks outside their group at a higher rate somewhat around 10 to 12 percent, although no recent data is available. Lending between other banks appears to be based on the earlier 14 percent BOU discount rate. Before December 1994 there was no standard method for setting rates, which could therefore vary by as much as 6 percent for the same borrower on the same day. 8.26 One reason for the range of interest rates among banks of the same credit standing is lack of information in the market. The market trades bilaterally: each bank rings the others to ask their price. There is no money broker or screen trading to bring all bid-and-offer information into one place and to assist in keeping pricing uniform. 8.27 Differences in rates among groups of banks are also caused by differences in credit risk. If T-bills were used as security this difference should be greatly reduced or even disappear. The price of funds is also affected by the term funds are borrowed for. The foreign banks tend to borrow overnight and to roll the funds if required. This short maturity reduces the liquidity risk to the lender. Local banks tend to borrow for terms of thirty to sixty days, reflecting their own weaker liquidity position and increasing the liquidity risk to the lender. 8.28 The establishment of a money broker to assist the creation of an informed market should therefore be considered once the level of bilateral interbank trading increases to a level that would make such intermediation profitable. Treasury bill market 8.29 In Uganda today, no true secondary market exists in T-bills. Stanbic is creating something of a market through its subsidiary Stanbic Bank International, but the greater volume of business is more akin to broking-buying at auction and on-selling to other banks or (mostly) nonbanks. 8.30 This lack of a secondary market is mainly due to problems in the primary market. Low demand for T-bills in the primary market means that few are sold and made available to trade. Most participants in T-bill auctions, moreover, are buying the bills with the intention of holding them to maturity rather than trading them. 8.31 The primary market in T-bills is not functioning effectively for several reasons: * T-bills are illiquid; there is no secondary market, and the BOU will not discount them-all of which limits their usefulness as a liquidity management tool. 96 * The illiquid interbank shilling market increases banks' need for highly liquid assets with which to manage clearing reversals. This increases the demand for excess reserve balances and lowers demand for T-bills. Banks have problems in deciding the volume of T-bills they can bid for because of the combined effect of volatility in the level of excess reserves and T-bill auction bidding requirements. As a consequence, banks participate in auctions only erratically and bid for fewer T-bills than their large excess reserve balances indicate they could bid for. * Banks' erratic participation in auctions and the limited choice of maturities offered, reduces banks' ability to acquire a portfolio of T-bills structured so that some bills are always within a week to maturity. Without a regular supply of T-bills approaching maturity in their portfolios, banks must keep higher excess reserve balances and buy more short-maturity T-bills at auction. The lack of a secondary T-bill market also increases the demand for short-maturity T-bills at auction. * The strong demand for short-dated T-bills (relative to demand for other maturities) results in a low yield, and this reduces demand for T-bills from the less liquid banks. Smaller banks in particular would prefer to keep excess reserves rather than to accept the 5-8 percent return on ninety-one day T-bills. The excess reserves can be used profitably to support foreign exchange business, and the risk of illiquidity in T-bills outweighs any advantage from a 5 to 8 percent return. 8.32 Banks buy T-bills for purposes of investment, managing liquidity, and the high weighting T-bills are given in judging capital adequacy. In Uganda today, T-bills are being purchased primarily as an investment as BOU's discount policy and the lack of a secondary market cause the banks to view T-bills as being of little use in liquidity management. However, the amount banks can invest in T-bills is reduced by the lack of an interbank market (which requires them to hold more excess reserves) and the difficulty they have in bidding for T-bills. 8.33 The factors reducing demand for T-bills have had a significant effect on T-bill auctions. Uganda's banks want to hold roughly U Sh 17 billion in excess reserves to manage clearing. There has, however, been U Sh 10 to 25 billion of excess reserves over and above the transactions demand of the banks. Although this full amount is in theory available to invest in T-bills, and holding these balances idle costs Uganda's banks something in the region of U Sh 1 to 3 billion a year in foregone interest revenue, T-bill auctions are chronically undersubscribed (Table 8.1). 8.34 Liquid shilling interbank market. A liquid interbank shilling market would reduce banks' clearing demand for excess reserves and so increases the amount of funds they were able to commit Table 8.1: Undersubscription in Uganda's T-billAuctions (U Sh million) T-bill maturity 1995 Auction Date date January 1 January 18 January 25 91-day (88)a (462) (986) 182-day (789) 65 (852) 273-day (1539) (74) (1100) 364-day (177) (103) (443) 'Numbers in parentheses indicate the shortfall between actual subscription values and the value of the T-bills offered in . to T-bills in the primary market. Potentially this is a significant source of demand. If interbank market liquidity improved to the extent that banks thought that they could rely on the market to fund half of their worst-expected adverse clearing, their demand for excess reserves would fall, 97 from an estimated U Sh 17 billion to U Sh 8.5 billion. The other U Sh 8.5 billion would then be available to invest in T-bills. Nor is this an unreasonable expectation. One Ugandan bank already has a policy of relying on the market for half its expected maximum adverse clearing. 8.35 A liquid interbank market would also increase the secondary market for T-bills by providing a source of funding for the trading portfolios of banks and brokers. 8.36 Use as a liquidity management tool. If more banks viewed T-bills as a liquidity management tool, a secondary T-bill market would develop. Banks would hold bills with the objective of trading if necessary, rather than holding them to maturity. 8.37 However, banks in Uganda today find T-bills too illiquid to be of use in managing their liquidity because of the BOU's discount policy. The earlier discount rate of 14 percent was far too punitive. With a ninety-one-day bill selling at around 6 percent, a discount rate of 14 percent translates into an effective cost of funds on an annual basis many times greater than 14 percent depending on the maturity of the T-bill discounted and on assumptions made about reinvestment. Banks would prefer to earn no interest at the BOU rather than risk this discount cost 8.38 A discount rate for T-bills set at a margin of 1 to 2 percent above the market rate would make T-bills more useful for managing liquidity and so increase demand. 8.39 Difficulty in committing funds to an auction. Banks hold much higher levels of excess reserves than would seem necessary because of the high degree of daily volatility in the level of excess reserves in conjunction with the T-bill auction rules, makes it difficult for banks to bid for T-bills. The daily change in total excess reserves from October 1994 through January 1995 is shown in Graph 8.5. Over the period the daily change ranged between extremes of +/- U Sh 10 billion. The peak to trough change was around U Sh 25 billion, about 7.5 percent of total deposits. For individual banks, clearing reversals of between 7 to 15 percent of deposits is not uncommon. Such volatility would be hard to manage in a developed market, let alone where banks must operate with an illiquid interbank market. 8.40 The rules of the T-bill auction Daily Escess Reserves require banks to have fundsequaltothefull 50 ----------------------------------------- 45 amount of their T-bill 40 bid in their account at 30 D25 the BOU on Tuesday, A 20 the day prior to the 15 auction. Banks must settle for successful bids on the day of the 94 95 auction or the October Peak to Trough A1: 25 billion Ush = 7.5% of total depo following day, Graph 8. 5 Thursday. A bank's ability to bid is thus very heavily influenced by the level of excess reserves it has on Tuesday and the level it can expect to have on Thursday, rather than the quantity of excess reserves it might expect to have on average over the week. Even if a bank knows its average excess reserves position over the week it will not be able to commit as much of this amount to the T-bill auction as it may want to because of the daily volatility in its excess reserves balance and the absence of a liquid interbank market which means that it has little ability to borrow interbank on Thursday, against the expectation of inflows over the rest of the week, to meet T-bill settlement requirements. 98 8.41 Altering the auction rules so that banks could base the quantity they bid on their expected spare reserve balances over the week following the auction would significantly improve participation in auctions. 8.42 The requirement that banks have funds in their settlement account equal to their full bid is intended to assure the BOU that bidders can meet their commitments. However, protecting BOU from default on a bid does not require that the full amount of the bid be available up-front. Banks and other large bidders with an established bidding record and adequate capital might be required to have 25-30 percent of the amount bid in their BOU account on the day of the auction rather than 100 percent. Competitive bidders without BOU accounts could be required to present cleared funds equal to 25 to 30 percent of the amount bid with their bid. 8.43 The period allowed for settlement of successful bids should be extended to the day of the auction (Wednesday) and the business days until the following Tuesday. Bidders could settle on any day inside this period at their discretion. The price paid for T-bills would be adjusted to reflect interest accumulated between the auction date and the settlement date. 8.44 To protect the BOU from failure to pay for bills, moreover, titles would not be registered until payment was cleared. A fine and exclusion from further auctions would be imposed on any party failing to settle. The penalty would be deducted automatically by the BOU from banks' accounts at BOU or in the case of non-banks, from the cleared funds provided with bids. The imposition of a loss-on-resale requirement on the bidder, automatically deducted from funds with the BOU, should deter bidders from trying to speculate against a decline in interest rates prior to settlement. Table 8.2. Bank Participation in 91 8.45 The impact of volatility in excess reserve day T-bill Auctions levels and the auction rules on bank participation Auction Date Number of Overlap in T-bill auctions is reflected in the lack of Banks of Bidders continuity in the number and names of banks Bidding participating in each auction. Each auction can 1994 attract a totally different group of bidders (Table November 9 6 some 8.2). There is also lack of uniformity in accepted November 23 4 some bids, illustrated in Graph 8.6. As the group December 7 3 none bidding in any one auction comprises only a part December 21 4 none of all banks and as the composition of the bidders 1995 differs between auctions, the bids in the primary January 4 3 none market reflect the differing perceptions ofa January 18 3 none different sub-set of banks at each auction. This contributes to the wide fluctuations in rates bid between each auction. If a larger group of banks consistently bid at each auction there would be more consistency in the bidding and a market consensus would develop over time, moderating extreme views of individual banks. 8.46 The erratic nature of the auction results is causing considerable uncertainty amongst the banks and Nonbank Financial Institutions (NBFIs) which is further reducing participation in auctions. Some banks are not bidding, or are placing their bids through banks whom they consider understand the market better, due to fear of having a bid accepted at the low end of accepted rates. In part this reflects the under-developed nature of treasury management in some banks; a successful low bid will be noticed unfavorably but the opportunity cost of idle funds will not attract such adverse comment. 99 Spread haten Maxim in & Miuimum Accepted Bids: SI Dap TBills 12 --- - - - - - - - - - - - - - - - - - - - - - - - - - - - 10- 2- 0 ""I AA numAI I n n>~ tiiati n limilan isSI2 Ml if 1.At ni x ni nR Graph 8. 6 8.47 In addition to changing the rules on bidding and settlement in auctions, the BOU could manage the level of excess reserves in a way that increases stability. The BOU could, for instance, select a target level of excess reserves based on banks' average demand for excess reserves for clearing. The target should be a margin above the banks' voluntary holdings to allow for downside error. 8.48 To achieve the target level of excess reserves, the BOU will need to: o Increase its ability to forecast factors affecting excess reserves (in order to set the size of T-bill auctions correctly). o Conduct open market operations to increase its ability to respond to short term fluctuations in the level of excess reserves. o Supply funds to the market through repurchase agreements or credit auctions. 8.49 Targeting the level of excess reserves is consistent with the current objective of controlling base money (Annex 4). This difficult task will require technical assistance and training. The BOU will also require an adequate balance sheet to enable it to conduct open market operations, for which it will require a portfolio of T-bills. 8.50 To control the level of excess reserves, the BOU must also have absolute discretion to sell T-bills to pursue monetary policy objectives irrespective of the funding position of the government. For as financial markets develop and an interest rate transmission mechanism strengthens, fluc*uations in excess reserves caused by the Treasury's reluctance to sell T-bills will lead to interest rate volatility and uncertainty. Having said this, coordination between fiscal and monetary policy is vital, and where a conflict exists, the BOU and MOF need a forum in which to reach a joint solution. 8.51 Maturity structure of T-bill portfolios. Uganda's banks presently have difficulty acquiring a T-bill portfolio with a well-structured maturity profile so that some T-bills would always be within a week of maturity. Should banks have to discount them to meet a short-term liquidity need, therefore, it would not be prohibitively costly. (Annex 5). The ability to maintain a supply of T-bills approaching maturity would improve banks ability to manage liquidity and reduce the demand for excess reserve balances. 8.52 Erratic participation in auctions due to auction rules inhibits Ugandan banks' ability to maintain a T-bill portfolio. The BOU's policy on T-bill maturities offered has changed twice in the last few months. Returning to offering a range of maturities each week would assist banks to structure their portfolios more advantageously. 8.53 Cutting off bids. The BOU has attempted to reduce fluctuations in the T-bill auction rate by cutting off bids it considers to be too high. The volatility is clearly undesirable, especially the 100 volatile ninety-one-day T-bill rate which has begun to affect bank behavior. Uncertainty about rates has caused at least two banks to shorten the maturity of deposits they are prepared to accept, which reduces their ability to extend longer term credit. 8.54 To reduce volatility in the rates T-bills go for at auction, Uganda must deepen the market rather than impose a cut-off rate on accepted bids. At present, the Ugandan T-bill market is too volatile for a cut-off rate (based on the maximum accepted yield in the previous auction) to be effective. Such a cut-off rule, biased toward lowering the rate, only confuses the market. For example, if the maximum yield bid drops from 8 percent one week to 5 percent the next, then at the next auction the cut-off would be set at 6 percent-much lower than the cut-off used two weeks earlier. While the BOU should continue to reject extreme bids, it will have to tolerate a high degree of volatility until changes in discount policy, auction rules, and the resulting reduced need for banks to keep massive excess reserves raise demand and deepen the market. Increased information 8.55 The broad dissemination of information on expected trends in market liquidity and prompt publication of auction results would also help banks to participate in the T-bill market. 8.56 Improved mutual understanding between the BOU and the banks will also assist market development. Regular contacts should be established at an operating level between staff of the BOU and the commercial banks. This contact need not be totally formal-meeting regularly at lunch proved very useful in New Zealand's banking sector. 8.57 Registry. Banks have responded positively to the suggestion that the BOU establish a secure, computerized T-bill registry operated on a commercial basis. A computer-based registry reduces the chance for fraud and theft and speeds processing time for transactions, especially where bills are to be divided into smaller parcels for retail investors. 8.58 At present T-bills in Uganda are bearer instruments that need to be filled in by hand by the Treasury prior to issue, a procedure that frequently delays the issue of scrip until a week after the auction. Hand-written scrip is also easier to forge. While pre-printed T-bill certificates are now available to the BOU and their use will both reduce issue time and the chance of fraud, computerized registry would further improve on speed of title transfer and security. 8.59 Large value transfer system. A secure system for transferring large sums is also needed in a liquid secondary market in T-bills to permit same-day delivery and payment for the bills. Such service would increase their value to banks as a means of liquidity management and therefore increase demand. Primary dealers 8.60 It has been proposed that Uganda establish a group of primary dealers whose job it would be to make sure that the T-bill offering was fully subscribed at a market-competitive cost to the Treasury, to broaden the market for T-bills by spreading ownership of T-bills beyond banks, and to create more liquidity in the secondary T-bill market. The proposal recommends that eight firms be licensed as primary dealers with exclusive access to the primary T-bill market. 8.61 The risk with this proposal is that giving a limited number of participants exclusive access to the primary market will reduce competition and raise transaction costs. This will be reflected in either a higher funding cost for government or lower returns to other purchasers of T-bills. The 101 benefits that primary dealers might be able to bring to the Ugandan market need to be examined carefully to determine that they are greater than the costs. 8.62 The first question to ask is whether primary dealers are likely to increase the primary market for T-bills beyond the size that could be achieved through implementing measures which encourage competition rather than limiting it. The market for T-bills consists of a wholesale market of large investors (banks, NBFIs, insurance companies, pension funds) and a retail market of small investors (individuals). In Uganda and elsewhere the wholesale market for T-bills is usually the largest market and also much more liquid than the retail market. Developing the wholesale market is thus a prime concern. 8.63 The main factors reducing the demand for T-bills of banks and NBFIs are the auction design, volatility of excess reserves and the lack of a liquid interbank market. Establishing primary dealers will not remove or reduce any of these constraints. The demand for T-bills from insurance companies and pension funds is currently reduced by T-bills' illiquidity and also the limited funds these organizations have at present. Unless primary dealers can assist in developing a secondary market to increase liquidity it does not appear that they will expand the demand for T-bills in the wholesale market. 8.64 Retail investors can already participate in the primary market through non-competitive bids. Expanding the retail market beyond the present group of participants within the constraints imposed by the level of Ugandan savings requires a mix of a distribution network which enables individuals to access T-bills more easily; a method of making small parcels of T-bills more liquid such as ability to discount to the BOU, brokers who would buy the T-bills or unit trusts which would be re-sold to the unit trust manager; and an alternative method of intermediation such as unit trusts which allow investment in T-bills through aggregating small savings. Primary dealers could perform some of these functions, but so could brokers and unit trusts who do not require exclusive access to the primary market. 8.65 Primary dealers' ability to stimulate a secondary market would be constrained by the same factors which presently constrain market development: lack of an interbank market; a relatively small number of participants; and problems in the primary market. Without special assistance from the BOU primary dealers will have no greater 20 - - - - - - - - - - - - - - - - - - - - inherent ability to make a market than a i..--. bank or a well capitalized broker. If primary dealers are given access to BOU credit there is a risk that they will use the 5-_1 funds to arbitrage the BOU rather than to 0 D 182 D 273 Day 3-4 Day make a secondary market. 8.66 Graph 8.7 shows the present Graph 8. 7 yield curve in the primary market. If a primary dealer obtained credit from BOU based on the ninety-one-day T-bill rate and then bought 182-day T-bills, it would make a low-risk 6.5 percent return simply by holding the T-bills-without taking the risk of market-making. 8.67 Primary dealers are sometimes used by the central bank to limit the number of bidders the central bank has to deal with. This is useful in a very large and geographically dispersed market such as the USA. However, in Uganda, all the banks and NBFIs have headquarters in Kampala and it is a simple matter for BOU to contract them and deal with them directly. There is no need for an intermediary to be imposed between BOU and this group. 102 8.68 Those advocating primary dealers have argued that-in their absence-fledgling brokers would go out of business for lack of income, wasting the training they had received. The equity market, they implicitly argue, should therefore receive a subsidy from the T-bill market. Such a cross-subsidy would have to be explicitly acknowledged and evaluated and the length of time specified. Authorities would therefore have to consider how to end the cross-subsidy after this time, especially if returns from other business activities had not materialized. It would be unfortunate if a desire to establish the equities market resulted in a higher cost structure in the government securities market in the medium or long term. 8.69 Finally, before considering the introduction of primary dealers, it is worth looking at the recent experience of other countries. Government securities markets in New Zealand and Australia have always worked effectively without primary dealers. South Africa eliminated the special position of discount houses in its government securities market in the early 1990s. The Federal Reserve Bank of New York diluted the position of primary dealers in fiscal 1992 to get more competition into the market in response to primary dealers' alleged abuse of its position, and because of the automation of auctions and open market operations that made it possible for the Federal reserve to deal directly with a greater number of participants. 8.70 Competition-which is needed for market development that would lower government funding costs-is encouraged by expanding rather than limiting the number of market participants. Any decision on whether or not to proceed with licensing primary dealers should be delayed until the changes suggested in discount policy, auction rules, etc. can be implemented and their impact on the market assessed. The decision should also be delayed until the desirability of a cross-subsidy from the government securities market to the equity market has been assessed. Developing a retail market for T-bills 8.71 Uganda is presently seeking to develop a market for T-bills outside of the banks, in particular among retail investors. Some of the incentive for this comes from having set policy targets on the basis of the net domestic assets of the banking system. Sales of T-bills to banks do not lower system net domestic assets whereas sales to nonbanks do. TBill Omefship 8.72 In developed markets 60000 - - - - - - - - - - - - - - - - - - - - - - - -, and in Uganda the holders of SOODO marketable government 10 goll 01e securities are typically 300oo o Ihs Cos C ce utns wholesale investors (banks, 200 H - NBFIs, pension funds) (Graph '00o 8.8). Banks favor short- maturity paper because of its liudiy whl.nuac 3.92 9.92 3.93 9.93 1.94 3.94 5.94 7.94 9.94 1.9 liquidity, while mnsuranceits companies, pension funds, and Graph 8. 8 mutual funds favor longer-maturity paper. Wealthy individuals will hold some marketable government paper but, in general, direct holdings of marketable paper by individuals are not substantial. Because retail investors primarily seek investment rather than liquidity management, they add little to secondary market liquidity. 8.73 The reason for banks and NBFIs dominating the market is that the capital market is a wholesale intermediation channel. Large sums of money are raised quickly in it with low overhead 103 costs. Participants must therefore have large sums of money and be well-informed and contacted easily-either directly by the organizer of the debt issue or at one remove through a broker. By contrast, retail investors have comparatively small sums of money and bid on an occasional basis. They are generally not well informed. Time must be taken to explain offerings and convince them to undertake transactions, and they are generally not easy to contact. Working with retail investors thus requires a more costly distribution network akin to a bank branch network. 8.74 Developing a retail investor base requires either that a marketable security (such as a T- bill) be sold through a retail distribution channel (such as a broker or mutual fund), or that a retail (nonmarketable) security be designed for purchase by individuals through a distribution channel (such as brokers or the Post Office). Such a retail bond would carry an administered interest rate based on deposit rates (which should increasingly reflect T-bill rates as the interest rate transmission mechanism develops) and could be sold on a commission basis by brokers or the Post Office. Banks would be unlikely to sell either a government retail debt instrument or T-bills to their clients as they are a direct source of competition for the deposits of the public. Banks are unlikely to use their branch networks to assist a competitor.. The amount raised through this method could be controlled by altering the interest rate offered. 8.75 A strategy for developing a retail funding base in Uganda would focus on developing brokers (and possibly mutual funds) to distribute T-bills, and on developing a retail savings bond to be sold though brokers or the Post Office. Brokers would be required to meet standards for capital, reputation, etc.. A retail savings bond could provide them with a product to sell that was more suited to small savers, making them more competitive with banks. A retail bond, sold on a commission basis would provide a source of income for brokers which would not involve granting them any special privileges in the T-bill market. 8.76 But developing a nonbank market for T-bills in Uganda is likely to be a slow process. If the BOU finds it difficult to meet the net domestic asset target with tools available (T-bill auction, discount rate), it may be useful to consider an alternative target. The target level of central bank net domestic assets could be set for a specific time and combined with a banking system net domestic assets target expressed as a trend. This would recognize the central banks' ability to control its own net domestic assets directly through T-bill sales, while the banking system's net domestic assets is subject to unknown and variable lag produced by the money multiplier effect and by the interest rate transmission mechanism. 8.77 Interest rate transmission mechanism. With indirect methods of monetary control, links between the level of excess reserves and short term interest rates generally arise. Lower levels of excess reserves lead to higher interbank and T-bill interest rates. Higher interbank and T-bill rates lead in turn to higher bank lending and deposit rates, which encourage saving and discourage expenditure and investment, slowing growth and dampening inflation. This series of connections is called the interest rate policy transmission mechanism. 8.78 In Uganda the interest rate policy transmission mechanism is developing but is weak at present. Graph 8.9 shows the weak connection between the level of excess reserves and the ninety- one-day T-bill rate-the first link in the chain. As the interbank rate is increasingly based on the ninety-one-day bill rate, the connection will strengthen in the interbank market 104 1 Day TBill Yield & Excess Reserves 45 ----------- ------------ -- 25 4 20 i 30 \ 25 1 A 201 Eess 10- all a 5 0 0 9.93 19 1.94 3.94 5.94 7.94 9.94 1 1.95 Graph 8. 9 8.79 The link between T-bill rates and bank lending and deposit rates (Figure 8.10) is also very weak. On the deposit side, the general excess of liquidity since mid-1994 has led to a collapse of the interest rate paid on demand deposits. But for large time deposits, banks are now linking the rate to the T-bill rate. This picture is clouded, however, by weak banks who are paying higher rates to attract deposits and remain liquid. 8.80 Minimum lending rates appear to be related to interbank 91 Day TBill Yield & Average Maximum and business and to the ninety-one- Minimum Deposit Rates day T-bill rate (Graph 8.11). Some banks are now also 25 ---------- ------- beginning to link their base (or 20 reference) rate to the ninety-one- 15 day T-bill rate, but the margin , 10 - I &x charged above the base rate, and 5 lending rates in general, are related more to what the market 8.93 10.931293294 4.94 6.94 8.94 10.9412.94 will bear. Uganda's weak banks, moreover, need a wide Graph 8. 10 net profit margin to cover bad debt, and high costs. Because of the impact of the weaker banks on the interest rate structure, stronger banks can take a wider profit margin. The large size of noninterest costs compared to interest expense also reduces lending rates' sensitivity to T-bill rates. 8.81 Strengthening T-bill rates' affect on lending rates St Day IBill Yield & Average Maximum and requires lowering banking Minimum Landng Rates intermediation costs, in part by 30 improving banks' ability to 25 enforce contracts and realize 20H ld collateral. Strengthening the link i 15 A Mn would also require that weak 10 A Max banks either recapitalize and s improve the quality of their credit control or exit the market. 8g 4 4.94 6.94 8.94 I O 8.82 Coordinating fiscal, Graph 8. 11 monetary, and exchange rate policy. As the link between excess reserve levels and money market rates strengthens it will become increasingly necessary for the central bank to manage the level of 105 excess reserves more exactly, for volatility in the level of excess reserves will increasingly affect interest rates. 8.83 Yet, in Uganda today, the Treasury is funding the budget deficit with off-shore funds. From the perspective of national funding, therefore, T-bill sales are unnecessary and even add an unwanted debt-service cost to the deficit. Offshore deficit financing and BOU interventions in the foreign exchange market to purchase dollars, on the other hand, add to the money base and has to be sterilized to some degree. The extent to which sterilization through issuing T-bills is necessary will depend on the expected inflationary impact of an increase in base money and-when the interest rate transmission mechanism is more developed-on the need to control excess reserves. 8.84 An understanding is therefore required between the Treasury and the BOU on the need to issue T-bills for monetary control on how the conflict between funding and monetary policy needs can be resolved. Such an understanding should be based on the following: * Using a base money process, the T-bill issue program, which corresponds to the chosen base-money path, should be discussed with the Treasury. * The Treasury should seek to keep inflation low and approve T-bill issue programs which are geared to do so. * The same process should be followed if an excess reserve target is used. 8.85 Mutual cooperation in the establishment of T-bill issues should help to familiarize the BOU and the Treasury with each others' views and create a constructive working relationship for solving issues of common interest, such as: * How to match government needs for domestic financing with T-bill issues. government and the BOU might investigate, for example, how to raise more domestic revenue and reduce foreign finance. * How to improve the efficiency of the domestic market for T-bills so that the domestic financing of the government was achieved on the best possible terms consistent with monetary control? * How the foreign exchange market could be deepened so that the market could handle large volumes of transactions without the need for BOU intervention. * How to coordinate fiscal and monetary policy in the even of inflation. Foreign exchange market 8.86 A distinction needs to be drawn between trends that cause appreciation or depreciation over time and factors that cause volatility immediately. The following discussion is confined to volatility, a problem that can be addressed through analysis of the structure of the market. 8.87 To reduce volatility in the exchange rate of the Uganda shillings to the US dollar, the BOU needed to intervene over a prolonged period in 1994. The BOU repeatedly purchased dollars, which increased the money base, creating an additional complication for monetary management. Since mid- 1994 BOU foreign exchange interventions have corresponded closely with the rising level of excess reserves (Graph 8.12). 106 8.88 Volatility in the exchange rate derives primarily from the Excess Reserves & CanulativeBOU FX market's small size relative to the Purchases size of some transactions. The Ugandan foreign exchange market 5oooo - - - - - - - - - - - - - - - - - is subject to erratic and lumpy 4oooooo 350D0000Ees esre dollar inflows, which it then has sec.oooo-* Emess Reserws trouble absorbing, although in the 2500000c-amuWn past year that problem has self- O n' corrected somewhat: 5oo o o 0. * Aid agencies have " Aid agenies have 1.t94 3.94 5.94 7.94 9.941.9 changed their habit of ts ris1 exchanging large Graph 8. 12 amounts of dollars for shillings on an irregular basis. When the deregulated market could not cope with the amounts involved, its stability was threatened. Aid agencies are now exchanging smaller amounts of dollars on a more frequent basis, although banks still feel there is room for improvement. (Most of Uganda's dollar inflows are related to aid). * Exporters are now bringing in foreign funds for pre-financing of crop purchases and are spreading their dollar flows so as to avoid disrupting the market. The second- largest source of dollar inflows is coffee exports. 8.89 Other factors affecting the markets ability to absorb large flows but that could be moderated include: * The lack of an interbank market in shillings, which limits individual banks' ability to respond to large dollar inflows, since a bank might not be able to obtain sufficient shillings to exchange for dollars. A more liquid interbank shilling market would assist liquidity in the foreign exchange market. * The lack of-a liquid money market, which affects the pricing of forward foreign exchange in a way that reduces demand for forward cover (Annex 6). The lack of a forward market limits market participants' ability to organize the timing of currency flows and places more pressure on the spot market. A liquid interbank shilling market would assist the development of a forward market. * Ugandan banks' unfamiliarity with forward foreign exchange and its integration into overall treasury management. An educational course on treasury management above should include a component on foreign exchange. * BOU prudential limits on the open foreign exchange positions banks are allowed to carry overnight. One of the biggest problems banks face in managing foreign exchange transactions is difficulty in laying-off dollar purchases. Limiting open positions sometimes forces banks to sell dollars into a thin market, leading to price fluctuations. While limiting open positions is a necessary prudential safeguard, and it would be unwise to institute a blanket relaxation of controls, some banks have requested permission to hold larger open positions. The BOU should consider permitting selected banks- based on capital adequacy, the skill and experience of staff in the bank's foreign exchange operations and the quality of the bank's internal risk control procedures-to have larger open positions. 107 Action Plan Actions for the BOU to complete within the next few months: o Adopt a dual discount window with T-bills under ninety-one-days to maturity to be discountable on demand at 1 percent above market and banks without T-bills to be allowed overnight funds (secured against their statutory reserves) at a 2 percent margin over the average maximum lending rate. o Survey all banks to establish their transactions demand for excess reserves. o Begin to develop the information framework needed to forecast the level of excess reserves in the system. o Change T-bill auction rules to reduce the amount that banks with an established bidding record have to hold in their BOU account prior to the auction and to extend the settlement period for auctions to four days. o Resume offering a range of T-bill maturities at each auction. o Develop a forum for meeting with counterpart staff in other banks. Actions to be completed by September 1995: o BOU to have completed documentation and procedures for repurchase agreements. o Technical assistance to have begun how best to establish a large-value transfer system in Kampala. o Contract to be signed for the investigation of a computer registry for T-bills. o Bankers Institute to have let a contract for a course on treasury management. o Government to be reassessing the market's need for primary dealers. o BOU should review its policy on the size of banks' permitted open foreign exchange positions, with a view to increasing positions for well-capitalized banks which have adequately trained staff and adequate risk management systems. Action to be completed by March 1996: o BOU and MOF to have developed a forum for the coordination of monetary, fiscal and exchange rate policies. Action to be completed by June 1996 o BOU to be fully capable of managing banks' level of excess reserves. 108 9. Long-Term Finance 9.1 The long-term financing activity in Uganda is limited in variety and magnitude. There is no functioning capital market for equities and no long-term debt instruments. Medium and short- term tradable debt certificates are limited to Treasury Bills. Lease financing and mortgage lending are currently provided on a very small scale. Contractual saving institutions do not play a significant role in the economy. This chapter describes the existing institutions and highlights potential reforms that can enhance long-term finance as an important potential contributor to economic growth by encouraging savings and facilitating the financing of new investments. Development Banks 9.2 Development banks are the most important providers of long-term funds for productive investments in Uganda. The three active institutions in this category are: * East Africa Development Bank (EADB)-which is prnmarily owned by three East African governments (Kenya, Tanzania and Uganda) in equal shares, * Uganda Development Bank (UDB)-which is fully owned by the government of Uganda (GOU), and * Development Finance Company of Uganda (DFCU)-which is jointly owned by Commonwealth Development Corporation (CDC), International Finance Corporation (IFC), Uganda Development Corporation (UDC), and Deutsche Investitions und Entwicklungsgesellschaft (DEG). EADB 9.3 EADB operates in the three East African countries, intermediating funds provided by international lenders and donors. Its sources of funds include the World Bank, African Development Bank (AfDB), European Investment Bank (EIB), SIDA (Sweden), FMO (Dutch), JEXIM (Japan) and the Swiss Government. Grant fund are also provided by NORAD (Norway) and DANIDA (Denmark). 9.4 EADB went through a major restructuring effort in the last two years. Special emphasis was given to collecting on the existing portfolio (with legal action and liquidation procedures against all non-performing borrowers), calling upon the three governments to pay arrears on loans made by the bank against their explicit guaranties, and introducing a new organization structure. Quantitative targets set for the two years restructuring period were fully met and the bank expanded its activities. For 1994 the bank recorded a profit of 1.7 million SDR (the currency in which the bank keeps its books). This profit amounts to 7.4 percent of EADB's capital and 2.1 percent of its total assets. With the 1994 profits, the bank's accumulated losses fell to less than 2.6 percent of its paid-up share capital. 9.5 During 1994, EADB provided new loans in the total amount of SDR 17.1 million (equal to US$ 25 million). Ugandan businesses received SDR 7 million (equal to US$ 10.3 million), making EADB the largest source of long-term loans in Uganda. The bank's liquidity position (with SDR 28 million in deposits and bank balances), and the willingness of lenders and donors to provide EADB with resources (including some additional investment in the bank's equity base), 109 will make it possible for the bank to expand its operations if viable projects are submitted for financing. An additional area for EADB activity is in the field of equity financing, where EADB is currently establishing a venture capital fund as a source for equity finance. The venture capital fund is to be managed by a venture capital specialist and funded by various participants and special donors funds. DFCU 9.6 DFCU was created for the provision of long-term finance. Resources were made available by the CDC, EIB and from the funds available through BOU apex department (discussed below). DFCU portfolio for December 31, 1994 amounted to U Sh 5.7 billion with some U Sh 0.9 billion allocated to loan loss reserves. Although much effort is put into collections, almost 24 percent of the December 31, 1994 portfolio was not performing. DFCU is currently putting much emphasis on collections, and new loans are approved after a full project evaluation is performed. For 1993 and 1994, DFCU recorded a substantial loss. For 1994, the loss amounted to U Sh 0.7 billion (on total equity of U Sh 3.8 billion). Even before making provisions for bad debt, DFCU recorded a loss of half a billion U Sh for 1994. It seems that only a much larger volume of business will enable DFCU to record a profit, as the resulting increase in revenue and gross profit will make it possible to cover the administration cost which amounted to U Sh 0.7 billion. Of this U Sh 0.7 billion, U Sh 0.3 billion was financed from a grant, while the remaining U Sh 0.4 billion charged to the profit and loss account. 9.7 DFCU has currently enough resources to fund viable investment proposals submitted. Total disbursements increased from U Sh 1.5 billion in 1993 to more than U Sh 4 billion in 1994, and the plan for 1995 is to disburse close to U Sh 7 billion. With these disbursements, DFCU is the second largest provider of long-term funds in Uganda. 9.8 DFCU is also pioneering venture capital activity. DFCU is currently managing a venture capital fund, financed with USAID grant funds, which has already invested U Sh 2 billion in eight projects. UDB 9.9 UDB is a fully owned parastatal. Its portfolio of U Sh 54 billion is heavily concentrated in industry, both private and public. The bank was never properly capitalized and has accumulated substantial losses, thus showing a negative net worth of more than U Sh 5 billion. 9.10 UDB's loan portfolio is presented as a net figure of U Sh 24 billion, after U Sh 15 billion are provisioned for bad debt, and overdue interest of more than U Sh 16 billion is suspended. The performing segment of the loan portfolio is very small, and the rest of the net portfolio represents non-performing loans which UDB management is convinced can be collected and/or the collateral can be used to cover the net debt. 9.11 In recent years UDB had very limited activity and total disbursements for 1994 amounted to only U Sh 2.3 billion. For the same year, administration costs alone amounted to U Sh 3.1 billion. UDB employs some 150 staff. 9.12 With its current performance UDB is not a viable intermediary, and cannot be considered for privatization or any participation from investors outside the government. For most 110 international lenders, UDB will not be considered as an eligible intermediary, as it does not meet minimum requirements for solvency and capital ratios set by most lenders. BOU 9.13 BOU has a development finance department that acts as an apex organization for several international lenders. The department is currently managing several Government funds as well as two World Bank lines of credit, and one each from IFAD and EIB. These foreign funds are available via BOU to several financial intermediaries, including commercial banks and development banks. UDB, with its negative net worth is not eligible to borrow from the World Bank or EIB funds. The amounts available under the above mentioned four lines of credit are U Sh 55 billion, and most of these funds are not yet disbursed. DFCU has applied for funding, as well as several commercial banks, with UCB and Baroda taking the lead. However, several of the larger commercial banks are not presently active in intermediating these funds. Reform Strategy For Developmental Banks 9.14 The current economic stabilization and growth in the Ugandan economy are expected to create a stronger demand for long-term loans in the next few years. Such demand will be a result of increased investor confidence, the need to modernize local manufacturing plants, and the opening up of export opportunities. 9.15 The existing sources of long-term financing, EADB, DFCU, and those commercial banks intermediating BOU's apex fund, seem to be able to provide enough resources for the next few years. One area for improvement will be to get more commercial banks involved in intermediating BOU's apex fund. It is important to find out what is keeping banks like Barclays, Standard Chartered and Stanbic from lending long-term to their customers from the long-term funds made available to them by BOU, and to create an environment for stronger commercial banks to fully participate in these programs. 9.16 UDB is in a different category. If not restructured and reformed, it cannot play a role in the provision of long-term funds. UDB is currently assisted by Commonwealth experts in an effort to improve its performance. Two areas which need specific emphasis are: (i) the collection of loans, and (ii) streamlining the organization while reducing the number of staff. If successful (even partially), collection of non-performing loans can result in income against arrears of principal and interest previously provisioned or suspended. As UDB has a negative net worth of U Sh 5 billion and total provisions and suspended interest above U Sh 30 billion, collecting only one third of this amount will turn net worth to a positive U Sh 5 billion, creating the basis for a revitalized UDB. Based on the above analysis, it is recommended that GOU should support, with technical assistance, the collection efforts of UDB on its existing portfolio and support efforts at organizational restructuring to reduce staff and administrative costs. During this stage, new funds from foreign sources (guaranteed by the GOU) should not be made available. After 12 to 18 months, UDB's success in collecting on its non-performing portfolio and in streamlining its operation will be reviewed. On the basis of UDB's performance, the demand for long-term credit and the activities of other long-term credit providers, GOU's strategy towards UDB's future role in the economy can be formulated. 111 Other Long-Term Financial Institutions ULC 9.17 Leasing has been introduced in Uganda only recently. Uganda Leasing Company (ULC), owned by DFCU, CDC, IFC and two local financial companies (one bank and one insurance company), just started operations and by the end of August, 1995 contracted its first lease. The shareholders of ULC have committed to provide US$ 10 millions in equity investments and long term loans to ULC. In many developing countries, leasing is a popular financing option for medium-sized companies in need of medium-term funds for equipment, commercial vehicles and computers. It seems that two issues will limit ULC's growth-funds available to ULC are mainly in foreign exchange while most lessees prefer local currency leases. With low interest rates in local currency, local businesses are reluctant to accept foreign exchange risks in investments intended to support activity in the local market. Another more technical issue which affects the leasing of commercial vehicles, is the problem of differentiating between the vehicle owner (the lessor) and the person or entity that has the legal responsibility for third party liability and for traffic violations (the lessee-operator). Without such a distinction, a lessor of vehicles is responsible for all damages to third parties and fines for traffic violations. HFCU 9.18 Mortgage lending is very limited in scope in Uganda. The company active in this subsector is Housing Finance Company of Uganda (HFCU), jointly owned by the National Housing & Construction Corporation (NH&CC) and by DFCU. HFCU is financing itself by receiving mainly short-term deposits from the public, and providing residential mortgage loans for 10 years. The total amount of public deposits with HFCU reached U Sh 3 billion for December 31, 1994. Outstanding mortgages for the same date were U Sh 2.7 billion, containing some 320 individual loans (averaging U Sh 8.5 million). Recognizing the liquidity problem that may result from using short-term deposits for long-term mortgage loans, HFCU is beginning to mobilize long-term resources. Recently a US$ 1 million 7 years loan was arranged from DFCU. The amounts involved are small, and the number of mortgage loans insignificant for solving Uganda's hou,ing problem. 9.19 Mortgage loans currently carry an interest rate of 18 percent interest (which is charged monthly, thus actually costing the borrower 20 percent per year), whereas depositors are paid only 4 percent interest. The resulting spread of 16 percent is large by any international standard. However, HFCU needs this spread to cover administrative expenditures (staff, office and depreciation) of U Sh 0.5 billion (which is more than 6 times the interest expense to depositors), to suspend uncollected interest of almost U Sh 100 million, and to show a modest profit for the year of U Sh 44 million (on net worth of U Sh 1.4 billion). HFCU is also acting as an agent for GOU in collecting GOU's low-income housing support program loans and managing the collection of long- term credit provided by GOU to buyers of government owned housing units. 9.20 HFCU estimates the value of a modest house which it is willing to finance at U Sh 25 million. HFCU will be financing 75 percent of the house, or U Sh 18.75 million. At 20 percent interest, such 10 years loan will require a monthly payment of U Sh 340,000. By international standards, such mortgage payments can only be afforded by families which earn at least U Sh 1.1 to 1.3 million a month, or U Sh 13 to 15 million annually. Such income levels are not common in 112 Uganda, making it impossible to consider HFCU type loans as a potential solution to mass housing finance. Alternative solutions must be explored, including the potential for financing cheaper housing units, more in line with the income level of the population. One potential solution, financing apartments in high-rise buildings (where the cost of land and infrastructure per housing unit is reduced), requires the recognition and registration of ownership on individual apartments, which then makes it possible to register individual mortgages to specific apartments. Such registered ownership of apartments in multi-unit buildings does not currently exist in Uganda. Insurance 9.21 The insurance industry in Uganda consists of 25 to 30 companies, most of which are very small. Total insurance premiums paid during 1994 were less than U Sh 20 billion, and out of this total the two largest companies accounted for half. Four to five medium size companies reached a level of U Sh 1 billion in premium revenue each, whereas the rest shared the remaining 20 percent of industry revenue (more than 20 companies sharing total annual premiums of about U Sh 4 billion). NIC 9.22 National Insurance Company (NIC) is fully owned by GOU. Unlike some other African countries, Uganda never prohibited private companies from entering the insurance sector, as a result of which NIC never had a monopolistic position nor was it entitled to a captive market in the parastatal sector. NIC went through a major reorganization in the last two years, decreasing staff from 400 to 160. NIC currently operates 12 offices and provides services in all the provinces of Uganda. It suffered a loss of U Sh 150 million in 1994, out of total revenues (premiums) of U Sh 5 billion. These losses were largely incurred due to high claims on motor insurance. The company has a capital base (net of accumulated losses) of U Sh 1.5 billion, and the total funds employed amount to U Sh 8 billion. The Government's decision to privatize NIC should prove to be beneficial to both NIC and the insurance sector as a whole, if an international insurance group can be encouraged to become a strategic investor in NIC. Such foreign participation may help in upgrading the insurance services provided in Uganda to international standards: PWICO 9.23 The other large company, Pan World Insurance Company (PWICO) had similar volume of premiums as NIC. However, PWICO is even less active than NIC in life insurance, thus making it larger than NIC in premium revenues on general insurance. PWICO is operating 8 offices (out of which three are agencies) with a staff of 85. The company grew very fast in recent years, with gross premium revenue more than doubling in two years. For 1994, the company earned a before tax profit of U Sh 270 million and its net worth reached almost U Sh 0.5 billion, whereas total funds employed reached U Sh 5 billion. PWICO pioneered the introduction of medical and travel insurance. Medium and small size companies 9.24 There are four to five additional companies which earn annual premium revenues of U Sh I billion each, and a much larger number of very small companies. This proliferation of small companies is a result of very low capital requirements to operate as an insurance company. This is 113 an unhealthy situation since it may precipitate systemic failure to meet justified claims, by allowing the under-writing of policies by companies which lack both the required know-how and financial resources. Furthermore, this environment may lead to unfair and irresponsible competition. Such a large number of insurance companies will also impose a very heavy regulatory and supervisory burden on the insurance commission, which is to be created in the BOU according to the new insurance law to be enacted shortly. Insurance regulation and supervision 9.25 Before the enactment of the new insurance law, the Insurance Commissioner in the MOF (and his staff) responsible for monitoring the sector according to the old law were dismissed, while a new commissioner in the BOU is not yet in place. This is a cause for concern. There is a need for immediate action to enact the new insurance law, operationalize the new commission in BOU, and start to formulate new regulations and enforce compliance amongst the insurance companies. Reform Strategy For Insurance 9.26 The insurance industry needs immediate consolidation into a workable number of companies. These companies must have an adequate capital base and skilled staff, as well as more international exposure. To reach the above goal, various actions are required: 9.27 The insurance law has to be enacted as soon as possible and implemented by: (i) nominating (immediately, not even waiting for the law to be enacted) a commissioner and several senior professional staff, and (ii) by drafting regulations and introducing a reporting and supervision system. 9.28 The Insurance Bill 1995, should be corrected to address the following three issues: o The treatment of foreign companies is not clear, and one has to recognize the fact that by discriminating against them, Uganda may lose valuable services to local businesses. One has also to recognize that if foreign insurance companies are artificially kept out of Uganda, nothing prevents local large businesses from going abroad for their insurance needs. The best solution will be to allow foreign companies to operate in Uganda as locally registered subsidiaries or as branches, as long as they provide enough capital locally. (Note: the current law proposed which requires foreign companies to hold five times the required capital in comparison to local insurance companies, should be brought in-line with the banking law requirement for foreign banks to hold twice as much capital as locally owned banks. With time, even these double standards should be removed completely). o The identification of reinsurance as a separate subsector by law, will discourage the reinsurance activity of insurance companies. A policy of allowing insurance companies to also provide reinsurance will help in keeping a larger share of the reinsurance business in the country (with insurance companies joining treaties with each other and smaller companies reinsuring with the bigger ones). o The need to prepare and submit every second or third year, an actuarial evaluation of the life insurance business is missing from the draft. 9.29 NIC should be prepared for privatization, with a serious effort to encourage a large internationally known insurance group in joining NIC as a strategic investor. 114 National Social Security Fund 9.30 The National Social Security Fund (NSSF) operates as a provident fund. Participation in the fund is compulsory for all formal sector employees, with the exemption of GOU employees, the military and the police, as well as some other exempted groups (such as teachers). The total number of contributors is around I million, and the number of actively contributing members around 800 thousands. The size of the NSSF contributing group is therefore a small fraction of the labor force, due to the low level of participation in formal sector employment and the large number of exempted groups. The relatively small number of contributors affects NSSF's financial significance (amounts collected and available for investments), its financial performance (in terms of the ratio of administrative costs to income from assets), and its contribution to solving old-age support problems (by serving only a small fraction of those in need). 9.31 NSSF is collecting 15 percent of all gross earnings (until recently the 15 percent applied to a small portion of gross income and contributions were insignificant). The employer carries 10 percent and employees are charged 5 percent from their gross income. Collections amounted to U Sh 8 billion in 1994, and are estimated to reach U Sh 9 billion in 1995. Payments to retirees are very small. They amounted to less than half a billion in 1994, and are estimated to be below U Sh 0.6 billion in 1995. Contributors (or their families) are entitled to their savings (contributions plus declared "interest" on accumulated contributions) either upon reaching retirement at the age of 55, in case of death, or in case of becoming an invalid unable to work. In the year 1993/94, NSSF announced "interest" on accumulated contributions of 14 percent, and in the year 1994/95 "interest" of 10 percent was announced. However these NSSF liabilities were not covered by income. NSSF's income on its portfolio was actually smaller, in both years, than its administrative costs, thus resulting in a net loss to the fund (and reducing each contributor's accumulated savings). The declaration of "interest" was against the creation of a deficit, which reached U Sh 2.9 billion in June 1994 (after the 14 percent "interest" was added to the liabilities), or 22 percent of NSSF liabilities. 9.32 NSSF is the largest holder of long-term funds in Uganda, and current collections are expected to be substantially larger that payments in the coming years. However, the main investment venue of NSSF are deposits with commercial banks and Treasury Bills, namely short- term investments. NSSF is currently expanding its investment portfolio by investing in a large office building in a prime location in Kampala. This building was started many years ago and left unfinished until very recently when construction has resumed. In effect, no meaningful long-term investment strategy was formulated beyond the above mentioned office building, and this inadequacy needs to be addressed. Thus NSSF's contribution to the financing of economic growth is very limited. Its long-term funds are invested short-term, while all long-term financing to local businesses is provided from abroad. 9.33 An ILO report advocating the conversion of NSSF into a pension fund was prepared in August 1994, and approved by GOU in principle. However, this report only provides a general framework for a pension scheme without any details. NSSF is currently working on a detailed proposal for the conversion of NSSF from a provident fund into a pension scheme. Reform Strategy For NSSF 9.34 Two existing policies of NSSF should be reviewed and reformed: 115 o The granting of "interest" to contributors at the expense of creating a deficit should be reviewed in light of the contingent liability imposed on the GOU. In a way, current contributors are subsidizing those members who retire during the early years. Alternatively, GOU will have to cover the deficit from its budget. Both solutions have positive and negative aspects. However, the fact is that the current NSSF policy was adopted without a thorough evaluation of this practice and its possible consequences. o The investment policy for NSSF funds should be evaluated, in order to support the long-term investment needs of the economy. The current practice of investing long- term funds in short-term deposits and T-bills does not take advantage of the only available source of long-term funds. 9.35 The move to a pension scheme should be carefully evaluated in terms of its contribution to solving Uganda social problems (as only a very small segment of the population, and not the poorest, are in formal employment and therefore eligible for pension payments) and its budgetary implications (consider the need to commit budget resources and their alternative uses) before a final decision is reached. 116 10. Macroeconomic Implications of Reform 10.1 This chapter offers a description and some approximate quantification of the macroeconomic effects of financial sector reforms that are being proposed in this report. The effects are fiscal, because public expenditure outlays will be required, and monetary because reforms will affect the fiscal deficit, non-banks' monetary behavior and the ways in which monetary policy is conducted. 10.2 Of these effects the most tangible and readily quantifiable are the fiscal effects. These are dealt with in the section A. An important underlying assumption is that the fiscal stance of the government will remain unchanged, and that domestic and external resources will be mobilized sufficient to prevent additional monitization. The resources required are assessed at the end of section A. The monetary effects go wider and are more speculative. They are more ambiguous, but also more significant for the real economy. They are discussed in section B. The direction and timing of effects on the real economy are briefly considered in section C. A. Fiscal Impact 10.3 The reforms in the financial sector rehabilitation program with potential fiscal implications are: * Compensation paid to depositors in insolvent banks which are closed. * Recapitalization of the UCB in association with its sale to private investors. * Restructuring and recapitalization of the balance sheet of the Bank of Uganda and other measures to staunch its current operating losses. * Investment to improve the payments system and financial support needed to maintain unprofitable UCB branches in rural areas. 10.4 These are reviewed in the following paragraphs. The magnitude and timing of the fiscal impact depends on the magnitude of the compensation paid to depositors, the balance sheet losses to be made good both at UCB and BOU, and on the choice of instrument to be used for recapitalization at both banks. 10.5 There are in Uganda no significant government or parastatal deposits in banks which are likely to be closed. However, government guarantees on letters of credit issued by UCB (and other off-balance sheet items) would have to be honored, as and when they arise. These off-balance sheet items will have to be incorporated in the bank's contingent liabilities. Such items are not covered by the indemnification arrangements which are likely to be put in force. 10.6 Deposits in "Problem Banks". As mentioned in Chapter 2, eight problem banks' have been identified, some of which may either have to be closed or kept open if the government so decides, through injection of new funds as part of a deal with new owners. Of these eight, Barclay's has recently been upgraded to "satisfactory", the insolvency of Cooperative is being resolved with the assistance of USAID, 1 Camel ratings issued by BOU state that Tropical, Sembule, Nile, Centenary, Kigezi, Barclays, Co-op and UCB, are showing unsatisfactory performance. In the mission's view, the first five represent banks for which closure may be a serious option. 117 and UCB is being dealt with separately. This leaves five banks which need resolution through liquidation or recapitalization. Deposits in these banks amounted to U Sh 228.5 billion in December 1994. The five banks for which closure may be a serious option had deposits totaling U Sh 36 billion. Of these, the three banks most seriously affected had deposits of U Sh 28.2 billion at the end of 19942. 10.7 Under the provisions of the deposit insurance fund, depositors in failed banks are only entitled to compensation of up to U Sh 3 million per depositor. The sums insured, based on Mission estimates, are U Sh 18.7 billion for all five banks, and U Sh 15.7 billion for the three worst affected. The current resources of the deposit insurance fund (U Sh 568 million) are inadequate to met charges of this size. Remaining depositors would have to be compensated at a cost to the fiscal budget after accounting for the sales revenues realized from the liquidation of the banks. 10.8 If depositors are only compensated up to the limit required under deposit insurance arrangements, there will be a budgetary saving, but there may be some consequential loss of depositor confidence in the banking system. The money supply could, at least temporarily, contract, probably bringing some, albeit small, contraction of GDP. On the other hand there would be moral hazard consequences if fuller compensation were paid. Senior bank executives, if they are confident of full coverage of deposits and are also confident that they will not be replaced in the event of insolvency, will not be concerned with the "down-side" risk of non-performing loans. Their lending patterns will therefore be riskier than otherwise, which is a classic example of moral hazard in bank lending. Furthermore, depositors would see no reason for prudence in their choice of banker because they expect full compensation in the event of bank failure. 10.9 If the reforms to the financial sector proceed according to schedule, indemnities will fall due in the course of the fiscal year 1995/96. The immediate cost to the budget would depend on how the Government decides to resolve the problem bank issues. If, as discussed earlier, the Government opts for liquidation, then the immediate cost to the budget for compensating depositors (either fully or up to the limit required under deposit insurance) would be offset by the value realized through the sale of the bank's assets at the time of liquidation. If the depositors are paid only up to U Sh 3 million (the current deposit insurance limit), then the costs to the Government are likely to be relatively small, probably negligible. If, on the other hand, depositors are fully compensated, the budgetary cost will be at least the amount of negative net worth-the "hole" in the balance sheet. If the bank is acquired by a new owner, again the Government would have to inject funds to and at least bring net worth to zero. 10.10 The fiscal burden of resolving the problems facing the five banks is shown in Table 10.1 It is assumed firstly that the Government's action to intervene in the operations of the banks leads to the liquidation of the two smallest banks with negligible fiscal cost as only deposits covered by deposit insurance are compensated. In the case of Nile, Sembule, and Tropical Africa, it is assumed that the Government injects funds spread over two years, to cover their negative net worth . Deposits in these banks on 31 December 1994 were (in U Sh billion): - Nile Bank 16.5 - Sembule Bank 7.1 - Tropical Bank 4.6 The recent portfolio audits of Nile and Sembule banks by Ernst and Young, suggest that their net worth is approximately negative U Sh 6 bil and 2 billion, respectively. For Tropical Africa bank, the December 1994 position, as per BOU's records, is negative U Sh 2.2 billion. This gives a combined negative net worth of U Sh 10.2 billion. 118 Table 10.1.Paments to De ositors in Failed Banks S1995/91996/9. 1997/9 5.1 5.1 nil 10.11 Recapitalization of the Uganda Commercial Bank. UCB is estimated to have had, on 31 December 1994, a negative net worth of U Sh 82.4 billion on total liabilities of U Sh 265.6 billion. Among the liabilities were deposits of U Sh 130.4 billion and administered funds of U Sh 22.5 billion. The administered funds are donor funds for directed credit programs which UCB has managed on behalf of the government, e.g. for the Rural Farmers' Scheme. A high proportion of these credits are in default, are uncollectible and have been provided against. Administered fund liabilities could [and should] be excised from the liabilities to be transferred to the privatized UCB. If the balance sheet were shorn of these liabilities UCB's net worth would improve from negative U Sh 82.4 billion to around U Sh negative 60 billion. 10.12 The buyer of UCB might not be willing to take over the deposit and non-deposit liabilities (minus the administered funds) unless there were a positive margin on shareholders' funds. This is because the book valuation of assets may be unreliable and subject to fluctuation. A purchaser might reasonably expect, as part of the sale agreement, an injection of assets from the Government into the bank of the order of U Sh 80 billion. 10.13 The recapitalization of UCB can be by way of an up-front cash payment, or by means of the issue of government bonds, or by some combination of the two. The government's interest is to recapitalize as far as possible by bond so as to reduce up-front payments from the budget. A bond gives the investor a performing asset to help restore the bank to solvency and profitability, with a positive net worth. It also mitigates the adverse monetary implications and the resulting macroeconomic consequences for the government from an up-front cash payment which will increase the budget deficit. However, there is no market for medium term GOU bonds, so the asset would be illiquid. The UCB is now quite liquid, with cash in vault and net balances with the BOU of U Sh 28 billion. Nevertheless prospective buyers are thought Zo be very unlikely to accept a high proportion of bonds in the recapitalization package. The more liquid the asset structure the better the recapitalized bank can fulfill its intermediation function, and the better its new owner will be able to cope with eventualities. 10.14 Three modalities of capital injection by the Government are therefore examined: * 100 percent cash. * 50 percent in cash and 50 percent with an index-linked five year bond yielding 4.5 percent interest in real terms. * The same as in 'V but with an up-front payment into an escrow account of the debt service due in the first three years of the bond. 119 10.15 The rationale for using an index linked bond - whose face value is adjusted every year by an index of general inflation - is that the outlook for inflation is uncertain, and a risk premium is attached even to short term instruments such as one year Treasury bills. The purchaser of UCB could therefore be looking for an asset which is protected from the vagaries of domestic economic fluctuations. This asset would yield a premium rate of interest in recognition of its illiquidity. 10.16 The fiscal costs of the three modalities defined above are shown in Table 10.2: Table 10.2. Costs of recapitalizing UCB under the different options (as part of its sale). Options for recapitalizing 1995/96 1996/97 1997/98 1998/99 1999/00 UCB a. Cash 80 0 0 0 0 Interest 0 0 0 0 0 Redemption 0 0 0 0 0 Total % of gross domestic revenues (12.4 %) (0 %) (0 %) (0 %) (0 %) b. Cash 40 0 0 0 0 Interest 1.8 1.98 2.18 2.4 2.64 Redemption 0 0 0 0 58.6 Total % of gross domestic revenues (6.5 %) (.26 %) (.25 %) (.24 %) (5.6 %) c. Cash 40 0 0 0 0 Interest 5.96 0 0 2.4 2.64 Redemption 0 0 0 0 58.6 Total % of gross domestic revenues (7.14 %) (0 %) (0 %) (.24 %) (5.6 %) Notes: Modality "a" refers to full cash recapitalization. Modality "b" refers to 50 percent cash and 50 percent bonds. Face value of domestic bonds are indexed to inflation rate of 10%. Yearly interest payments are 4.5%, and first year's debt service payment is made at the same time as the issue of bonds. Modality "c" is identical with "b" but with an up-front payment into an escrow account of the first three years debt service payments. 10.17 Recapitalization of the Bank of Uganda. As already discussed in Chapter 5, the BOU has started to operate at a loss. Its audited accounts show an operating loss of U Sh 3.3 billion in 1993/94, due largely to falling receipts of interest on advances and to falling foreign exchange commissions . This fall in net revenues is a consequence of stabilization (a decline in seigniorage) and of foreign exchange liberalization. The BOU intends to offset the fall in its earnings by further cost-cutting initiatives (beyond the initial staff reductions discussed in Chapter 5) and to ask the Government to recapitalize the institution. 10.18 The BOU's balance sheet contains a number of claims considered "doubtful", including debt service payments made on behalf of government to the IMF and other creditors, and advances to parastatals. These, together with small realized foreign exchange losses on external credits administered by the BOU and losses on barter trade arrangements amount to U Sh 66.3 billion.5 The BOU's audited accounts for the year ending 30 June 1994 include, in addition to the operating loss of U Sh 3.3 billion. a large, unrealized foreign exchange loss of U Sh 37 billion which the BOU believes may subsequently become reversed. They also show a large increase in gross loans and advances to the Uganda government (though a fall in advances net of deposits), without any corresponding increase in interest receipts. The losses and doubtful claims underlying this figure are (in U Sh billion): 120 10.19 Efforts to restructure the asset base of BOU, would necessitate that doubtful assets be realized for what they are. The resulting fall in the volume of asset would be reflected in a corresponding fall in BOU's net worth (share capital and reserves). The net worth is already negative due to accumulated past losses. To rectify this situation, the government needs to recapitalize BOU either by issuing it bonds or by transferring funds from its deposits at the central bank into BOU's share capital. 10.20 There is a case for an interim injection of U Sh 61 billion which will eliminate its current negative reserves (which stand at negative U Sh 26 billion), establish positive reserves of U Sh 20 billion, and increase its share capital from U Sh 15 to U Sh 30 billion. The positive reserves would be a useful buffer against future capital losses in the foreign exchange account and would facilitate monetary policy interventions. It is likely that further recapitalization will be required when the BOU's loan book is thoroughly audited to take into account government and parastatal borrowing from the BOU. 10.21 The macroeconomic consequences of BOU's recapitalization should be viewed in light of how BOU had financed its past losses (operating losses) and what options of deficit finance are available to the Government. If past BOU losses were covered by resorting to monetary finance, there are four feasible options concerning BOU's recapitalization: (1) to do nothing, which implies that BOU continues to monetize its operating losses and the government ignores its negative net worth, (2) the Government recapitalization BOU using zero interest bonds, which rectifies its negative net worth but is not able to resolve BOU's operating losses, (3) to recapitalize BOU with interest paying bonds, which addresses both its negative net worth and provides BOU with an additional source of income, and (4) a direct transfer of funds from the Government's account (in BOU) to BOU which allows it to increase its paid-up capital and purchase additional assets (government bonds or foreign exchange) from the domestic market. 10.22 These four options should be evaluated in terms of their impacit on BOU's balance sheet position (rectifying its stock position) and its liquidity/income flows. If strengthening BOU's balance sheet is deemed important, the option of not recapitalizing BOU (option 1) should be discarded. If improving BOU's income flows is considered important (to stem its operating losses), then the option of racapitalizing with zero interest bonds (option 2) can also be discarded. Between the two remaining alternatives, the least inflationary option is to recapitalize BOU by issuing interest bearing government bonds (option 3). If option 4 were implemented, BOU's efforts to build its asset base by purchasing these assets from the domestic market would result in a significant injection of liquidity into the Ugandan economy. 10.23 Although the option of issuing interest bearing bonds raises the Government's domestic debt burden, it provides enough leeway for the GOU to finance its future fiscal deficits using non-inflationary sources of finance.6 There is, however, the inevitable opportunity cost of using such forms of non- inflationary finance (eg. external or domestic bond finance) which should be considered in deciding which option to use. The Mission recommends that option 3 be used in the recapitalization of BOU. - debt service payments on behalf of govemment 38.5 - advances to parastatals 22.4 - foreign exchange losses on FSU and EDF credits 3.3 - barter accounts 2.1 6 It is important to note that if the Government's fiscal deficit were totally monetized (by printing currency), there is little difference between options 2 and 3 in terms of their inflationary impact. However, since this is not the case, there is a macroeconomic distinction to be made between interest- free and interest bearing bonds. 121 10.24 On the basis of this recommendation, the fiscal burden facing the Government-if the nominal interest rate on BOU's recapitalization bonds was 12 percent per annum, would change Table 19.2 to read: 10.25 Payments system improvement and subsidies for rural bank branches. At the time of writing this report no estimates were available of the costs of keeping open the rural branches of UCB which may not to be desired by UCB's eventual purchaser or which cannot be merged with the branches of other Table 10.3. Fiscal costs of(i) recapitalizing UCB, (ii) intervention in remaining five problem banks, and (iW) recapitalizing BOU Options for recapitalizing 1995/96 1996/97 1997/98 1998/99 1999/00 UCB a. UCB - Cash 80 0 0 0 0 Interest 0 0 0 0 0 Bond Redemption 0 0 0 0 0 Five problem banks 5.1 5.1 0 0 0 BOU - Interest 7.32 7.32 7.32 7.32 7.32 Total 92.42 12.42 7.32 7.32 7.32 Total (as % of gross domestic revenues 14.3% 1.6% .8% .7% .6% b. UCB - Cash 40 0 0 0 0 Interest 1.8 1.98 2.18 2.4 2.64 Bond Redemption 0 0 0 0 58.6 Five problem banks 5.1 5.1 0 0 0 BOU - Interest 7.32 7.32 7.32 7.32 7.32 Total 54.22 14.4 9.5 9.72 68.56 Total (as % of gross domestic revenues 8.4% 1.9% 1.1% 1% 6.3% c. UCB - Cash 40 0 0 0 0 Interest 5.96 0 0 2.4 2.64 Bond Redemption 0 0 0 0 58.6 Five problem banks 5.1 5.1 0 0 0 BOU - Interest 7.32 7.32 7.32 7.32 7.32 Totat 58.38 12.42 7.32 9.72 68.56 Total (as % of gross domestic revenues 9.1% 1.6% .8% 1% 6.3% Notes: BOU's recapitalization is strictly in terms of bonds valued at U Sh 61 billion, with a nominal interest rate of 12 percent (option 3) Modality "a" refers to full cash recapitalization of UCB. Modality "b" refers to 50 percent cash and 50 percent bonds. Face value of domestic bonds are indexed to inflation rate of 10 percent. Yearly interest payments are 4.5 percent, and first year's debt service payment is made at the same time as the issue of bonds. Modality "c" is identical with "b" but with an up-front payment into an escrow account of the first three years debt service payments. institutions such as the Cooperative Bank. These costs are difficult to estimate. In terms of the costs of improving the payment system, it is expected that infrastructural costs will be about US$ 5 million, and technical assistance should amount to US$ 10 million. 122 Fiscal costs of refonn 10.26 The main costs of rehabilitating the financial sector are thus to be sought in terms of the fiscal impact of compensating depositors in banks to be closed, and in servicing the recapitalization bonds to be issued to UCB and BOU. Government payment of an additional U Sh 2 billion into the deposit insurance fund has no immediate macroeconomic consequence because the funds will be invested in Treasury bills. The insurance fund represents a contingent liability for the government with consequences for macroeconomic balances only when insurance claims are made. 10.27 As shown in Table 10.3, the sums involved range from a crippling 14.3 percent of domestic revenues in 1995/96 if UCB's recapitalization is all in the form of cash, to a still unmanageable 8-9 percent of revenues if one of the options of part payment in the form of bonds can be agreed. External assistance would be required to enable Uganda to bear this burden. In summary, the foreseeable fiscal costs of financial sector restructuring should be bearable without serious risk of destabilization, so long as there is external assistance in the initial period of reform. In the subsequent years the fiscal costs of reform could be covered by a small, temporary, increase in fiscal receipts. 10.28 A donor-supported external financing package designed to cover the costs of depositor indemnification of the interest payments for UCB and BOU in 1995/96 would amount to US$ 63 million if the procedure outlined in Modality c were adopted.8 B. Monetary Impact 10.29 Reforms may have a variety of monetary effects. The main potential effects arise from: * Any loss of confidence in the banking system arising from bank closure. * Banks' endeavors to strengthen their balance sheets, under the impetus of tighter prudential supervision and to conform with capitalization requirements. * Higher bank lending made possible by the recapitalization of UCB. * Development of the market for Treasury Bills and commercial paper. 10.30 The effects of reform on monetary conditions will vary through time. These effects are not quantifiable with the available information but a qualitative assessment can be made. 10.31 Systemic consequences ofbankfailure. If banks fail and their depositors forfeit their deposits, a contraction of the money supply will result unless depositors are fully compensated for the loss of their financial claims.9 If bank failure leads to a loss of confidence in the banking system or in particular institutions, the voluntary withdrawal of deposits brings a secondary reduction in the money supply. Any serious loss of confidence would reverse the slow remonetization of the Ugandan economy of recent years. 7If we add on the costs of improving the payment system (U Sh 13.9 billion), Modality "a" could cost the Government up to 16.5 percent of gross domestic revenues in 1995/96, whereas the range with the other two options is 10.6 - 11.2 percent of gross domestic revenues. 8 Assuming an exchange rate ofU Sh 925 = USS 1. 9 Money supply (M2) equals the sum of bank deposits and of notes and coins in circulation. 123 The monetary effects of a loss of depositor confidence in parts of the financial sector will need to be countered by an emergency provision of liquidity by the BOU. 10.32 There has so far been no sign of a general loss of confidence in the Ugandan banking system. One small bank, (Teefe) has already been closed successfully (although the extent to which its depositors were compensated is unclear). With careful handling of reform it should be possible to avoid the potentially serious consequences just outlined, and any impact on the real economy. 10.33 Balance sheet strengthening. As part of the reform process, banks will be expected to strengthen their balance sheets and to achieve a minimum target for core capital of 4 percent of risk adjusted assets. They can manage this either by operating profitably, and so increasing their net worth, or by attracting additional outside capital. Either way they will have to adopt conservative lending policies and to pay greater attention to the quality of their lending. This may well, for a time, continue to deprive some potential borrowers of access to bank credit. 10.34 Lending by a recapitalized and privatized UCB. UCB's capacity to lend has been circumscribed in recent years by bad debt and by its management's endeavors to restore its liquidity and profitability. If UCB is successfully privatized and recapitalized, with the burden of non-performing assets removed from its balance sheet and replaced by government bonds, the bank's new owners and management will be well placed to seek new lending opportunities. New lending by UCB which is not offset by recoveries by NPART on non-performing assets, can represent a potentially inflationary addition to the domestic assets of the banking system if past lending practices were followed by UCB. In practice, however, the new owner of UCB is likely to lend cautiously in view of Uganda's history of poor credit discipline. A large proportion of the cash element of the recapitalization package is likely to be used initially to acquire domestic or foreign securities. 10.35 Treasury bill market. An important objective of financial sector reform is to establish a liquid financial market permitting indirect monetary control. Interest rate signals in the market for Treasury bills will eventually directly affect the inter-bank market rate and the base rate used by banks for determining lending rates. 10.36 When this structure is fully established, it will permit a more accurate fine-tuning of monetary policy. Currently, the principal instrument of monetary control is the relatively blunt and insensitive one of fiscal management. Increases in officially held net foreign assets have to be offset by changes in the government's financial position vis-a-vis the BOU in order to adhere to the base money growth path agreed with the IMF. This is Uganda's main target variable for conducting monetary policy. 10.37 Indirect monetary control is most effective as a device for restraining credit expansion to the private sector times when banks are eager to lend into markets which are keen to borrow. At present, Uganda faces a very different situation from this. Banks have excess liquidity and are reluctant to lend. Under new banking and clearing house regulations, they have high liquidity needs and are distrustful of the inter-bank and Treasury bill markets as liquid placements. There is no effective interest rate transmission mechanism. Changes in the supply of Treasury bills do not have the expected impact on T-bill rates which are very volatile. Furthermore, bill rates have little influence on the banks' base rates. 124 10.38 The reform program comprises measures to change the BOU's discount policies so as to make the Treasury bill market and interbank market more liquid. These measures are unlikely to have an early impact on the conduct of macroeconomic policy because of the small volume of Treasury bills relative to the money supply. If open market operations are to become a powerful tool for economic stabilization, there will first have to be a serious effort to tackle the problems of bad debt and credit discipline which currently restrains banks from lending up to their prudential limits. In addition to this, efforts to increase the supply of Treasury bills and to make the secondary market more liquid, are important pre-requisites to implement contractionary monetary policy at times of excessive credit expansion. C. Impact on the Real Economy 10.39 Initial Impact. The impact of the reforms is likely in the first instance to be contractionary. The government is most unlikely to depart from its tight financial program formulated in the context of its ESAF arrangement with the IMF. Any fiscal costs of restructuring over and above those provided by new external assistance, are thus likely to be met by (temporarily) higher taxation. This will depress domestic incomes and activity, albeit only to a limited extent. 10.40 Systemic effects (if any) from bank closures and the effect of the strengthening of banks' balance sheets will also temporarily depress domestic activity and limit the financing of new business ventures. Banks will keep their high intermediation margins while they struggle to restore their profitability, so the cost of credit will remain high for a while longer. The licensing of new banks will be more cautious than it has been in the recent past. 10.41 Medium to long-term impact. The benefits of financial consolidation in the banking sector should, however, become apparent within the medium term as stronger financial institutions emerge which are better able to compete and to innovate. The slow but steady past increase in indicators of financial depth - such as the M2/GDP ratio and the ratio of bank deposits to GDP - will then resume and accelerate. These benefits for the real economy will be reinforced by the development of a secondary financial market which will enable financial institutions to achieve greater efficiency in the deployment of their assets. Financial institutions will require smaller liquid balances and will be able to reduce their intermediation margins without loss of profitability, so reducing the cost of credit to non-bank borrowers. 125 Annex 1 Performance Contract with Merchant Bank 1. The performance based compensation contract with the merchant bank has the following characteristics. On the basis of due diligence (a detailed audit of the firm to be sold) undertaken by the merchant bank and other independent audits, it is possible to calculate the volume of recapitalization needed to resolve UCB's negative net worth and - on the basis of its existing loans - place enough capital in the bank to meet prudential standards (currently 4 percent of risk adjusted assets). This is the total cost (say U Sh X) which would have to be incurred by GOU to bring UCB back to solvency. Assuming UCB has intrinsic value (on the basis of its distribution network, its name, and its ability to mobilize low cost deposits), the merchant bank should be able to find a potential buyer who - in negotiation with GOU - would agree to take over UCB if the government injected U Sh Y. The intrinsic value of UCB would ensure that X is greater than Y. In the contract with Morgan Grenfell to sell UCB, GOU has agreed to compensate the merchant bank according to the following formula: Compensation = Fixed Sum (F) + (D (X - Y - F), where (D is the fraction of the difference between the actual cost of recapitalizing UCB (X) and the total cost to the Government of selling UCB (this includes the recapitalization package negotiated with the buyer (Y) and the fixed fee to the merchant bank (F)). Since F and cx are determined by the outcome of negotiations between the merchant bank and GOU, and X is calculated after the process of due diligence, it is clear that if the merchant bank were interested in maximizing its revenues, it can only do so by minimizing Y. This proposed market solution to resolve the issue of UCB will not only be more efficient in using market information (Y), but will also minimize the recapitalization cost which the government would have to bear., However, the terms of the contract would have to be designed to address the moral hazard on the part of the merchant bank to delay the sale of UCB on account of a low market valuation of UCB (or a high value of Y). On the other hand, since the contract only addresses one part of the principal's goal (the actual sale of UCB) but does not include the larger concern - ensuring a sound and efficient privately owned UCB - the contract would also have to address the possibility that a buyer who is willing to accept a low Y, may not necessarily represent a high quality investor. a-I Annex 2 Detailed Analysis of BOU Inspection Reports 1. Data analysis. To analyze the data from these reports, the thirteen items listed under "Other Supervisory Matters" should be computerized in a spreadsheet. Ratings could then be readily compared with previous inspection reports for the same bank or for other banks. This would allow a gross ranking based on quantitative comparisons, which would reflect general patterns of weakness within the banking system. Because this table could be changed at will in future inspection reports, analysis of the sector would become a dynamic process. 2. Following verification and modifications, the most recent quarterly reports relating to insider lending, excess concentration of credit, and large nonperforming assets should also be attached to the BOU inspection report to allow for ready reference during subsequent review. In the future, these attached reports could be expanded by a column showing the status of an entry in the last report, and the degree of change during the previous period. Columns should be added to get totals, which would facilitate the analysis of trends as well as levels in the sector. 3. BOU off-site supervision reports. The format of off-site-reports would benefit by: * Substituting a twelve month, rolling information requirement for present system of year-to-date information provided each quarter. This is necessary because the wide variation in fiscal year-end dates for individual banks make year-to-date peer group comparisons impossible. * All banks reporting lending to other Ugandan banks of about U Sh 5 billion more than they report borrowing from Ugandan banks. This suggests that either the banks have inconsistent definitions of these accounts (no standardized accounting), or that they are accounting creatively (with respect to float, etc.) in order to understate their liabilities. in either case, the discrepancy should be eliminated. * Two banks report no nonperforming loans-an anomaly that needs to be examined. 4. The balance sheets in the present reporting system do not balance because they report beginning-of-the-year data on capital rather than on-going information. An additional line item showing year-to-date changes in capital position should therefore be added, so that banks will also report their current net worth position. a-2 -3~ rrU 0 n 7J Q Q O Z tti c-n c22 c_- C
World Bank Group · Pre-2003 Economic or Sector Report
Uganda - Financial Sector Strategy Update (Vol. 2 of 2) : The Main Report
View original document
The full text is hosted by the publishing organisation. lawenc.com indexes the metadata and links to the official source.
Full text
Key facts
Organisation
World Bank Group
Document type
Pre-2003 Economic or Sector Report
Country
Uganda
Source
World Bank