MUNICIPAL FINANCE 21098 BACKGROUND SERIES Building Local Credit Systems George Peterson APRIL 2000 The World Bank CONTENTS I. INTRODUCTION 1 II. LOCAL BORROWING AND LOCAL CREDIT MARKETS: FRAMEWORK FOR ANALYSIS 5 The Credit Paradigm: When is Local Borrowing Desirable? 5 Foundations of a Local Credit System 7 Municipal Credit Systems and the Efficiency of Investment 13 What Constitutes "Excessive" Borrowing? 16 Formulating a Strategy for Municipal Credit Market Development 18 III. MODELS OF MUNICIPAL CREDIT SYSTEMS 22 The Theory of Intermediation: Bank Loans vs. Municipal Bonds 22 Municipal Banks as Specialized and Quasi-Monopolistic Lenders 23 Municipal Bonds 27 Choosing a Credit Model: The Advantages of a Mixed System 30 Colombia and the Czech Republic: Two Examples of the Transition to a Competitive Credit Market 31 IV. MUNICIPAL DEVELOPMENT FUNDS AND THE TRANSITION TO MUNICIPAL CREDIT SYSTEMS 36 What Are Municipal Development Funds? 36 The Transition to Self-Sustaining Credit System 36 Loan Repayments and Municipal Creditworthiness 38 MDF Lending: Development Bank Functions 44 Raising Capital: The Intermediation Role 48 Summary: The Transition to a Market-Based System of Municipal Credit 53 V. STRATEGIC CHOICES IN CREDIT SYSTEM DESIGN 54 A Set of (almost) Always Applicable Reforms 54 Introducing Parastatal Intermediaries into Municipal Lending 55 Building a Competitive Credit Market 58 ANNEX: INSTITUTIONAL CHARACTERISTICS OF MUNICIPAL DEVELOPMENT FUNDS 60 REFERENCES 62 BOXES Box 1. Financing India's Infrastructure Investment 7 Box 2. The Distorting Effects of Government Guarantees: An Historical Perspective 19 Box 3. The Distorting Effects of Government Guarantees: A Modern Example 20 Box 4. Municipal Bank of the Netherlands: The Western European Approach to Municipal Credit Intermediation 28 iii Box 5. Evolution of the U.S. Municipal Bond Market 33 Box 6. The Transition to a Competitive Local Credit Market Development: Colombia 37 Box 7. The Transition to a Competitive Local Credit Market: Czech Republic 39 TABLES Table 1. Loan Repayment Experience: Selected Municipal Development Funds 44 Table 2. Functions Performed by MDFs 50 Table 3. MDF Access to Capital Markets: Sources of Funds 53 Table 4. Sources of Funds: Selected MDFs 55 FIGURES Figure 1. Financing of Municipal Investment 13 Figure 2. Potential for Municipal Lending 24 FOREWORD Who Should Read This Report? Most nations of the developing world are trying to build stronger capital markets. A segment of the capital market that has acquired special importance is the financing of urban infrastructure. Explosive urban growth has multiplied demand for investment in basic infrastructure services, like water supply, waste removal, roads and mass transport. At the same time, decentralization strategies have shifted responsibility for much of this investment to local governments. Capital markets have had to adapt accordingly. No longer can a central government pay for local investment by raising national taxes or borrowing on international markets and using the funds simply to construct projects at the local level. Decentralized investment requires the development of decentralized capital financing. Over the past decade, both international agencies and national authorities have devoted a good deal of effort to the creation of local credit markets. This focus is now intensifying. The Asian Development Bank has announced that local credit markets will be key to meeting the massive investment needs of Asian cities, and that the ADB will focus more of its lending and technical assistance on their development. The Inter-American Development Bank has announced that an increasing proportion of its assistance will be channeled to local authorities through local lending and programs designed to strengthen local government access to credit. Yet, the record of accomplishment is frustrating. More than 50 countries in the developing world have set up special credit intermediaries to lend funds to municipalities. Almost always, it is asserted that these institutions are meant to pave the way to self-sustaining local credit markets, where funds can be gathered voluntarily from private savers and channeled through market intermediaries to cities and towns that need investment financing. Very few countries have made this transition, however. In fact, few have even laid the foundations for a true municipal credit market. Although Municipal Development Funds in some parts of the world now are celebrating their 25th anniversaries, they have largely remained captive instruments for on-lending funds provided by international institutions and central governments. If self-sustaining local credit markets are to emerge, a re-thinking of policy and institutions will have to take place. This report takes stock of international experience to shed light on a single question: How can efficient local credit markets be built? Not surprisingly, some of the lessons from past experience are negative. Attempts to design municipal development funds in particular often have been marred by perverse incentives, some of which recur so frequently that merely avoiding these errors in the future should substantially improve chances for constructing market-oriented local credit systems. The establishment of successful credit markets will be a joint enterprise. It will engage national government planners working on financial sector reform and fiscal decentralization. It will require the commitment of private sector bankers and other private financial institutions that supply credit. In decentralized government systems, states and v provinces will have a crucial role to play. Everywhere, cities, acting on their own and through collective instruments like Associations of Towns and Cities, will need to seize the initiative to help shape the credit systems that will be financing part of their investment budgets. Finally, international organizations will have to exert leadership in designing local credit mechanisms that can evolve into self-sustaining private credit markets as well as meet the agencies' short-run project implementation goals. To all these parties engaged in the task of building local credit systems, this report is dedicated. The report has been prepared under the Urban Management Program by George Peterson of the Urban Institute with Sonia Hammam, manager of this activity. I. INTRODUCTION Local Credit Systems Are Important... 1.1 Municipal credit systems have become important elements of development strategy because they stand at the crossroads of three of the most basic trends affecting the developing world. 1.2 Rapid rates of urbanization have magnified the need for infrastructure investment. The United Nations estimates that 2.4 billion people will be added to cities in developing countries between 1995 and 2025. This population growth will create unprecedented demand for capital facilities. In Asia alone, where three of the world's largest nations-China, India, and Indonesia-are urbanizing simultaneously, it is estimated that an annual average of US$280 billion will have to be invested in infrastructure projects over the next 25 years.' Part of this investment burden will consist of nationwide highway, power and telecommunications networks, but much of it will involve localized investment needed to provide residents of individual cities with water, waste removal, transportation, schooling and other fundamental services. 1.3 At the same time that urban investment needs are mounting, decentralization initiatives are transferring greater responsibility for planning and financing investment projects to local and state governments. A number of countries have amended their constitutions to formally assign more service functions to sub-national governments. In all, some 70 nations with populations over 5 million have launched decentralization initiatives. Foremost among the responsibilities being shifted to lower levels of govermnent is control over the local capital budget. On the expenditure side of the budget, local authorities now play a greater role in setting capital investment priorities and in project implementation; on the revenue side, they have greater responsibility for financing their own capital outlays. 1.4 Fiscal adjustments are impelling all levels of government to try to reduce their budget deficits. One of the favorite instruments of deficit reduction has been to cut back government subsidies in the financing of infrastructure projects. Studies in different parts of the world have found that capital investment support commonly is reduced more steeply than any other element of spending during periods of tight budget constraint.2 To sustain investment levels in the face of reductions in central-government capital grants local authorities have had to make greater use of borrowing as a way to finance capital projects. ' This estimate is from Asian Development Bank, Urban Infrastructure Finance, A New Vision: Devolution and Market-Based Mechanisms (April 1996). 2 For example, Hicks (1988) found that during the budget retrenchments of the 1980s, a sample of 25 developing nations cut central government real current spending by 7.8 percent but spending on capital projects and capital grants by 35.3 percent. 2 1.5 All of these factors have raised the stakes involved in finding effective ways for local and state authorities to access capital markets. The stakes involved in financing infrastructure investment are illustrated by the case of India (Box 1 below). Box 1. Financing India's Infrastructure Investment India exemplifies the challenge to capital markets inherent in raising infiastructure investment levels. The India Infrastructure Report, two years in preparation, estimates that urban infrastructure investment suitable for local financing constitutes about 20 percent of the country's priority infrastructure needs. Assuming that India is able to increase total infrastructure investment from 5.5 percent of GDP to the level of 8 percent found in Southeast Asia, and assuming that GDP growth rates can increase, the country still will be able to finance over the next decade only about half of the investment "needs" that experts have identified. Infrastructure investment, however, will climb rapidly. This growth will require "extraordinary measures" to transform India's capital market, particularly its illiquid debt market. The report estimates that by 2006 private investment may finance as much as 45 percent of India's total infrastructure spending (up from 25 percent at present). Nonetheless, capital markets will have to increase the amount of credit generated for public investment from $7 billion per year today to more than $20 billion by 2006. Much of this growth will occur in local and state credit markets-markets which, today, scarcely exist outside of specialized mechanisms for allocating credit from public financial institutions. A true local credit market is essential for India meeting its infrastructure investment objectives, within the framework of decentralization that the country has committed itself to through constitutional amendment. Source: The India Infrastructure Report, National Council of Applied Economic Research, New Delhi (1997). The magnitude of the investment challenge implies that financing cannot come only, or even primarily, from government or multi-lateral development banks. Private-sector savings must be tapped. Private equity investment in traditionally public services, like water supply systems or roads, can provide part of the needed capital. However, better-functioning credit markets will be critical. Local governments will need to increase the scale of direct public investment, which will require better local government access to credit financing. Even when infrastructure services are privatized, credit remains a large part of the financing package. In a typical private-sector infrastructure project, only 20 percent or less of investment funds are raised by equity. The rest of the financing consists of credit, often arranged in collaboration with local public authorities. 1.6 The credit systems that are established for local governments should do more than facilitate access to new sources of financing. They should help increase the efficiency of local investment and the efficiency of ongoing service delivery. By forcing local governments to acknowledge the true cost of capital, private market credit systems can help move local authorities toward more realistic project design and toward better cost recovery practices. The need to pay back private-sector borrowing often leads to institutional reforms-e.g., privatizing a water system or re-structuring it as a self-financing and institutionally independent public entity-that increase the efficiency of service operations. Finally, there are incentives for better financial management. Local authorities that must disclose their finances to the private credit market in order to demonstrate that they are creditworthy come under pressure to take financial management seriously. 3 ...But They Are Emerging Slowly 1.7 Developing countries have begun the process of building local credit systems. A handful of nations now have vigorous, if small, local credit markets. In most developing countries, however, local governments still rely on parastatal lenders for credit. More than 50 countries were identified during this study as operating municipal development funds or their equivalents in order to channel credit to local governments. Most of these funds were established with the assistance of multi-lateral development banks or bilateral donor agencies and are used largely to on-lend international program funding to local authorities. Municipal development funds (MDFs) typically have been envisioned as transitional institutions which will help prepare the way for self-sustaining municipal credit systems. Unfortunately, few developing-country MDFs have actually evolved into market-oriented suppliers of credit. Nor have they accelerated the development of competitive municipal credit markets. Not infrequently, MDFs have entrenched themselves as monopolistic institutions and resisted the entry of private-sector lenders into their domain. 1.8 Building efficient local credit markets will require reforms on a scale that goes beyond demonstration that municipal on-lending can be managed successfully by parastatal agencies. The track record of MDFs is mixed at best. Many have suffered high default rates which have only discouraged private financial institutions from making loans to local governments. Even when successful as lenders, MDFs typically enjoy special guarantee arrangements from the State that are not available to the private sector. To establish a broader municipal credit market, national governments will have to stabilize local government revenues, so that municipal officials and potential lenders alike can make realistic assessments about the future capacity of local governments to repay debt. Municipal financial statements need to be standardized. Local authorities need to acquire the capacity to prepare capital budgets and evaluate long-term financing plans. Practical arrangements for strengthening the security behind municipal loans need to be put in place. Once these conditions are met, private-sector lending to local governments can grow very rapidly, as several of the case studies presented in this paper demonstrate. 1.9 Perhaps most importantly, central authorities need to withdraw from the task of local investment decision making and local credit allocation. Allowing local governments to have direct access to private credit markets, constrained only by their ability to repay debt and accountability to local voters, is an important step in financial decentralization. The tradition of steering credit to municipalities outside the market, at below-market rates, has a long tradition that is difficult for governments to surrender. Moreover, central government has a strong and legitimate interest in regulating the local credit market to avoid excessive borrowing that can interfere with national economic adjustment priorities. However, this end is best achieved by straightforward regulation of the private market, not by setting up a parastatal system of municipal credit that substitutes for the market. Public institutions that lend to local authorities often are necessary first initiatives to introduce the principles of municipal credit to the economy, but they should be designed from the outset to make it easier to establish a private credit market and should either go out of business, or find specialized market niches that the private market 4 cannot serve, as soon as private lenders demonstrate that they are able to meet the credit needs of local authorities. II. LOCAL BORROWING AND LOCAL CREDIT MARKETS: FRAMEWORK FOR ANALYSIS 2.1 The idea of local government debt still evokes a good deal of ambivalence in the developing world. This ambivalence is reflected in the conflicting postures that international organizations appear to take on the subject. 2.2 The World Bank, as well as regional development banks like the Asian Development Bank, the Inter-American Development Bank, and the European Bank for Reconstruction and Development, have argued that greater borrowing by local authorities is necessary to finance the investment needs caused by urban growth, especially within a policy framework committed to greater decentralization of spending decisions. Consistent with this positive view of debt, the multi-lateral development banks and many of the larger bilateral donor agencies have assigned high priority to the development of credit systems that can raise the flow of lending to local governments. The share of donors' own funding which is disbursed as pass- through loans to local authorities has been rising. Several authors have stated that debt remains underutilized even in the financing of big cities' capital budgets.' 2.3 At the same time, however, the International Monetary Fund (IMF) has warned of excessive public sector debt.2 Debt levels in the developing world are said to endanger future repayment capacity, threatening the financial underpinnings of economic growth. Although much of this warning is aimed at central authorities, the IMF has raised concerns as well about mounting levels of sub-national debt. It specifically has identified local and state borrowing as important elements in the excessive public sector debt of such countries as Argentina, Brazil, Colombia, Hungary, and Russia-all of which are countries where international donor agencies have established or tried to establish credit systems for increasing local use of borrowing. 2.4 A similar ambivalence marks national policy toward local borrowing. Many countries have made local credit expansion a hallmark of decentralization policy, and have insisted that local authorities make greater use of debt and less use of state subsidies in financing future capital spending. Yet precautions against excessive levels of local debt can be found throughout the developing world. Most countries that permit local borrowing have established debt limits intended to keep local authorities from borrowing irresponsibly. Many prohibit local authorities from borrowing in foreign currencies. Some require central government approval of all sub-national debt. A few countries, like Chile, forbid local borrowing altogether. The Credit Paradigm: When is Local Borrowing Desirable? 2.5 These different perspectives suggest that it is useful to return to first principles in considering when it is desirable for local governments to borrow. We can start, as most texts on ' Bahl & Linn (1992) report that for a sample of 25 developing-country large cities, gross borrowing accounted for only 6 percent of revenue inflows. 2 International Monetary Fund, World Economic Outlook (April 1996). 6 local credit markets do, with an individual investment project of sizable scale. Local governments rarely maintain cash surpluses large enough to pay for the entire cost of big capital projects; it would be irrational for them to do so. Thus, they face a choice. They can accumulate savings in their current account budget until they are able to pay for the entire capital investment, or, if they have access to credit, they can borrow to finance all or part of the project. They then can pay for the project over time. 2.6 Borrowing allows a local entity to carry out a more ambitious capital program than otherwise would be possible. In principle, it also promotes intergenerational equity by having the generation of citizens that benefits from a capital facility's services pay for its construction. A standard rule of thumb holds that the period of debt repayment should approximate the useful life of the capital project, so as to match the time profile of costs and benefits. 2.7 Borrowing is not always an appropriate financing strategy, however. Borrowing to cover current account deficits has just the opposite intergenerational effects of paying for capital expenditures from current revenues. It shifts to the future the costs of services enjoyed by today's taxpayers. Many of the cases that have attracted the IMF's attention involve sub-national borrowing to pay for current expenditures.3 Central government regulations may prohibit local borrowing to finance operating expenses, but because resources are fungible and monitoring of capital projects often is inadequate, it is difficult to ensure that funds nominally borrowed for capital purposes end up financing investment rather than operating expenditures. 2.8 Within the sphere of capital financing, caution is warranted in generalizing from the way an individual project can reasonably be financed to a system of capital finance. Borrowing is a reasonable way to pay for a representative capital project. However, every community has an ongoing program of capital spending. The current generation of taxpayers was bequeathed a stock of capital facilities paid for in part from past operating budget surpluses. In a steady state-where investment constitutes an approximately constant share of total budget spending year after year-it will be equitable to continue financing the capital budget in approximately the same proportions between current-account savings and borrowing that has been used in the past. To suddenly switch to debt financing of all investment would shift the relative burden of paying for capital both backwards to previous generations and forward to new ones, while relieving current taxpayers of their share of costs. 2.9 A first guideline for responsible local borrowing is: increases in longer-term debt should be used to finance increases in capital investment. Analysts sometimes take this relationship for granted. It is assumed that, especially in the context of rapid urban growth and vast needs for infrastructure investment, new debt will be used to raise investment levels. In 3 In both Argentina and Brazil, for example, provincial and state governments have borrowed massively from provincial banks to pay for their operating expenses. As of June 1995, Brazilian states had $33.9 billion in short-term tradeable debt outstanding, as well as large amounts of longer-term debt and direct loans from state banks. The City of St. Petersburg, Russia, in March of 1995, authorized 2.1 trillion rubles (then about US $500 million) of very short-term borrowing, mostly in the formn of three-month bills, to help pay for operating expenses. Many other Russian cities followed a similar course. 7 practice, however, local governments that accelerate their borrowing often do so at the same time they are cutting back on their investment levels, or they may use a combination of heavy borrowing and one-time asset sales to barely sustain investment spending, while cutting back steeply on the contribution of current account savings. This is not a sustainable condition nor an indicator of healthy credit market development. It is a sign of fiscal stress.' Foundations of a Local Credit System 2.10 Like any other market, the local credit market can be visualized in terms of a "demand" side and a "supply" side. The demand side of the market is represented by local or state governments' desire to finance investment through borrowing and by their ability to repay debt. 2.11 On the supply side of the credit market, financial sector institutions have the role of gathering savings for on-lending to investors. Within the municipal sector, there is a special need to provide longer-term credits, since many of the infrastructure projects to be financed have long service lives and do not generate full economic returns until years into the future. Financial intermediaries serving the sector ideally should be capable of accessing long-term savings (like pension funds) or be able to absorb the risks of term intermediation, i.e., to collect short-term savings and use them to finance longer-term loans. 2.12 When no legal obstacles stand in the way, private suppliers of credit can then interact on market terms with municipal borrowers. Of course, in a competitive credit market municipalities must compete with all other borrowers for funds. Because local government lending involves a specialized kind of financial analysis and its development generally lags behind lending to central government and large private firms, specialized financial institutions often have been created to introduce municipal credit. 2.13 In a well-functioning credit market, the interest rate on municipal loans not only establishes the cost of borrowed capital but establishes the marginal cost of capital that municipalities should use in their general capital planning, even when projects are financed from a municipality's own resources. One of the important advantages of an active municipal credit market is to make clear the true "price" of investment. Local investment projects in general should produce a stream of benefits that outweigh the costs of financing them with borrowed funds. The Demand Side of the Market: Municipal Ability to Repay Debt 2.14 At the heart of any credit system is a reliable revenue stream that the borrower does not use for day-to-day operations. Borrowing for investment purposes is equivalent to 4 Hungary in the early 1990s illustrates this kind of strain on municipal budgets. In 1991, gross current account savings financed three-fourths of municipal capital spending. Thereafter, current account savings eroded steadily and sharply. By 1994, current account savings were negative. All of the capital budget was being financed by a combination of borrowing, capital grants, and asset sales. Borrowing was being used to cover operating deficits, as well. See Hegedus et al. (1996). 8 capitalizing an income stream. The borrower receives funds today to pay for project construction. In return, he signs away the right to an annual revenue flow in the future, which goes instead to the lender. The more certain and predictable is the revenue stream, and the greater the assurance that the borrower will use it to repay debt, the greater is the security behind a loan. 2.15 Three kinds of income streams or income promises typically stand behind municipal borrowing. These can be seen in Figure 1, which summarizes the sources of financing for municipal investments, as seen from the perspective of the municipal budget. Figure 1. Financing of Municipal Investment 1. MUNICIPAL BUDGET (GENERAL FUND): * LOCAL GOVERNMENT OPERATING BUDGET = Current Revenues * Own Source * Shared Taxes * Operating Transfers from Central Government - Operating Expenses = OPERATING SAVINGS _ CAPITAL BUDGET = Operating Savings + Capital Contributions from Users + Asset Sales + Investment Grants from Central Government + Proceeds from Borrowing and Bond Issues - Investment - Debt Service 2. FUND FINANCING (SPECIAL FUNDS) a Earmarked Streams of Revenue Set Aside to Finance Debt Service on Specific Capital Projects Source: Adapted from Pechon (1995), Gouame (1994), Standard & Poor (1996). 2.16 General Revenues. Every municipality has recurring expenditure obligations such as wages for municipal personnel and the operating costs of public service delivery. A municipality likewise has standard recurring revenue sources, such as receipts from fees and taxes. In decentralized fiscal systems, both sides of the local operating budget will be largely under the government's control, constrained primarily by the preferences of local voters. The municipality can raise local taxes, if necessary to balance its budget, or increase general fee levels. It has the option of cutting back on current expenditures by reducing municipal personnel, freezing wages, or, in extreme cases, trimming the kinds of local public services it provides. 2.17 The excess of local recurring revenues over local operating expenditures in the municipal budget is what American analysts tend to call "operating surplus" and the French 9 sometimes call "management savings." This surplus is labeled Operating Savings in Figure 1. The savings or surplus from the operating budget is carried forward to the capital budget, where it can be used to finance direct capital expenditures or to pay debt service on borrowing. The revenue side of the capital budget can be further augmented by strictly capital revenues, such as sales of municipal property, intergovernmental capital grants, and capital contributions from local developers or service users who connect to the municipal infrastructure network. These are one-time, non-recurring revenues. New long-term borrowing also adds to capital revenues. The overall budgetary relationships are illustrated in Figure 1, where revenues are shown with a (+) sign and expenditures with a (-) sign.5 2.18 A municipality's capacity to take on new "general obligation" debt-debt secured by the municipality's general budget-depends largely on its present and anticipated future operating savings. Extraordinary sources of capital income for a time may be used to repay debt, but over the long run operating savings must be sufficient to cover debt service with a margin of safety. Assessing prudent debt capacity in light of future revenues and future operating expenditures is the task of both the municipality and credit analysts working on behalf of lenders. 2.19 Figure 1 also shows that there is a tradeoff between using a municipality's current account savings to help pay for investment directly, on a pay-as-you-go-basis, and using it to pay annual debt service. Committing future operating savings to debt service increases today's borrowing capacity and can raise today's investment levels, but (unless the project itself generates additional revenue) it will reduce the net revenue available for investment in tomorrow's capital budget. A common complaint in the developing world is that a mayor will finance a flurry of public works projects by exhausting the municipality's borrowing capacity, then leave to his or her successor the burden of debt repayment without the possibility of making new investments. 2.20 General obligation debt is the most common form of local debt in developing countries. The borrower promises to take whatever measures are necessary, on either the revenue or expenditure side of the general budget, to ensure timely debt servicing. Of course, a promise of this kind is only as good as the borrower's ability and willingness to control its budget. In many countries, the central government fixes local tax rates, prescribes local tax bases, sets local fee schedules, and unilaterally determines tax sharing arrangements. The only discretion that local authorities have is on the expenditure side of the budget, and even there budget choices may be hedged by central government mandates regarding local service provision as well as nationally established wage scales for local government workers. In short, the general obligation debt pledge in a developing country may carry less weight than it does in a mature, decentralized fiscal system, simply because fewer budgetary options are open to the municipal borrower. 5 Figure 1 represents a stylized and simplified budgeting system. Accounting and reporting conventions vary greatly from country to country. Many countries do not have a separate, formal capital budget. Local accounting rues may require that expenditure items to be classified differently than shown in Figure 1. The principles of local budget analysis nonetheless are the same. 10 2.21 The willingness of the borrower to make budgetary sacrifices in order to repay debt may also be questionable. Defaults on municipal debt are not a rare phenomenon in the developing world-either today or in the nineteenth century among then-developing countries like the United States. Municipal "willingness to pay" is gradually shaped by the market itself, as it responds to defaults or financial emergencies. If municipalities perceive that defaults will shut them off from future credit access, or cause them to pay much higher interest rates for loans, they will tend to assign high priority to debt service. If defaults carry no significant penalty, either because government lenders forgive bad loans or because lenders continue to make new loans despite the fact that old loans remain unpaid, municipalities will become less inclined to make budget sacrifices in order to give priority to debt service. 2.22 The risks inherent in general obligation lending have caused lenders to seek various kinds of additional guarantees for their loans. In Eastern Europe, banks typically require municipalities to put up municipal property as collateral for municipal loans. Lenders everywhere commonly seek guarantees from central government that, if necessary, they will redeem local authorities' general obligation pledges. Failing that, they may seek legal protection that certain types of recurring revenues will be segregated from the General Fund and placed into special funds that can, by law, be used only for debt servicing. 2.23 Project Revenues. One forrn of specialized revenue financing, or Fund Financing as shown in Figure 1, involves dedicating the revenue stream from an investment project to debt repayment.6 The income stream in this case may consist of the revenue earned from water sales to households or the tolls collected from users of a newly built toll road. 2.24 Project finance of this kind can be used to fund either public or private investment projects. It is feasible only when it is administratively possible to charge customers for service usage. However, a broad range of municipal services falls into this category, including water, gas, and electricity distribution; solid waste collection; wastewater treatment (billed with water consumption); mass transit and toll roads; and even less conventional services like burial in municipal cemeteries. In all of these cases, at least partial cost recovery from users is not only feasible but economically desirable. Service pricing ensures that customers value service usage at least as highly as the price charged to them. It thereby restrains the "demand" for excess infrastructure capacity that exists when services are free. 2.25 Risk analysis of project financing, like that of general obligation debt, involves assessing the reliability of the income stream and the margin of safety that exists between projected income and debt service. Economic studies typically are used to forecast consumer demand for the final service and consumer willingness to pay. The lender gains protection from covenants written into the loan agreement that require the borrower to maintain a specified minimum margin of net revenues over debt service. If the margin falls below the designated 6 The range of special revenues that can be dedicated to municipal credit payment is virtually endless. In addition to project revenues, it includes such revenue sources as a gas tax, motor vehicle tax, local add-on to a national or state income tax, or a local sales tax. In each case, revenues are paid into a special fund which must be used exclusively or as first priority for debt servicing. For an analysis of the issues involved in this kind of fund financing, see Moody's Investor Service, "Analysis of Florida's Special Tax Securities," Perspectives on Municipal Issues (October 1996). 11 level, the operating authority is required to raise service prices or take other measures to restore the financial cushion. Project revenues are typically paid into a special Fund. They cannot be diverted to other budgetary uses unless the reserves available in the fund exceed a designated level. Ultimately, the lender has the right to foreclose on the income-generating property if debt service is not paid. 2.26 Worldwide, project financing has become one of the fastest growing components of international capital markets.7 In the United States, so-called "revenue bonds," municipal bonds backed only by project revenues, now account for the majority of lending to local and state governments. In developing countries, however, project finance within the public sector has proved more difficult to administer. The pricing of public services usually is subject to political constraints, whatever the wording of loan covenants may be. It has proved difficult for lenders to foreclose on publicly owned property, sometimes because the legal system does not recognize municipal bankruptcy or establish procedures to follow in the event of municipal default, sometimes because practical and political realities override legal formalities at a time of financial crisis. 2.27 For these reasons, project finance within the public sector has proved more risky than private-sector project finance. Most of the credit for developing-country project finance has flowed to private investors, even when the service is under public regulation and traditionally has been provided by local governments. International donor agencies have supported private-sector project financing as a cost-effective means of meeting many municipal investment needs and have begun to recognize the special demands it places on credit markets.8 2.28 When municipalities do use project financing to borrow, they typically reinforce the security of project revenues by additional pledges. Segregated project revenues still are used as the first line of debt service, but if these prove inadequate, the municipality is likely to pledge to make additional payments from its general budget. This combination of project financing and general obligation borrowing sometimes called "double-barrel" financing. 2.29 Central Government Transfers. A third kind of revenue stream that can serve as the basis for municipal lending consists of transfer payments from central or state authorities. In stable fiscal systems, transfers of this kind form part of a municipality's general revenues and are incorporated into the general obligation pledge. In developing countries, intergovernmental transfers more often are singled out as a distinctive source of credit security, and must be analyzed as a separate source of credit risk. 2.30 Local government budgets in most developing nations are highly dependent on shared taxes and intergovernmental transfers. However, transfer amounts can be highly unstable. ' Worldwide project financing reached $47.6 billion in 1996, up 83 percent from 1995. This follows a 50 percent growth rate the preceding year. See IFR Public Finance International (February 1997). 8 Some of the infrastructure financing credit funds now being set up in different countries will lend both to municipal governments and to private investors investing in local projects under municipal regulation. As one example of multi- lateral support, EBRD in 1995 signed an agreement with Lyonnaise des Eaux to help finance 12 to 15 private-sector municipal water supply, sewage treatment, and solid waste collection projects in Eastern Europe. 12 They may be reduced drastically because of a deterioration in the central government's own fiscal position, because of a change in party at the national level, or for other reasons. Local government borrowing capacity then becomes difficult to analyze, since it is a function of future central government transfer policy. In this environment, future intergovernmental financing policy often is the principal element of risk in local credit assessment. 2.31 Several nations have attempted to stabilize local revenue sharing arrangements and to make local resources more buoyant by guaranteeing in their constitutions that municipalities will automatically receive a prescribed or growing share of national tax collections. The first steps toward decentralization in Latin America, for example, typically have involved new revenue-sharing agreements that greatly increase municipal incomes and enshrine future municipal revenue-sharing entitlements in fundamental legislation.9 The security provided by guaranteed tax-sharing arrangements probably has done more than any other policy action to expand local credit markets and to accelerate private entry into municipal lending. 2.32 From the point of view of lenders, unimpeded access to central government transfers, before these go into the mix of a municipality's general budget, may constitute the best revenue stream against which to secure debt. Transfer payments in effect are diverted into a legally protected special Fund of the kind shown in Figure 1. The mechanisms for accomplishing this are various. Lenders may be given the right to "intercept" transfer payments before they reach the municipality, whenever a local government is past due on its debt servicing. That is, lenders have first claim on intergovernmental transfers as long as there is delinquent debt. In other cases, municipal borrowers may be required to establish special accounts at lending banks. Intergovernmental transfer payments flow into these accounts which are blocked for general municipal use until debt service has been discharged. In still other cases, transfer streams become the sole source of loan repayment. Amounts due to lenders are deducted at source by the central or state government and paid before transfer funds ever reach the municipality. Bookkeeping transfers of this kind are most common in servicing loans from public agencies. 2.33 Each of these arrangements represents an effort to ensure that adequate municipal revenues will be available to service a municipality's debts. Once a municipality has compiled a track record of timely debt payment, lenders are likely to be satisfied with a general budgetary pledge. At earlier stages of market development, or in lending to municipalities with an erratic repayment record, lenders are likely to require some form of additional guarantees that specific revenues will be set aside for debt payment and not commingled with the general budget. The Supply Side of the Local Credit Market: Financial Institutions 2.34 Even during periods of fast population growth and strong infrastructure investment, local government borrowing is likely to be a relatively modest portion-10 to 15 percent or less-of a country's total credit market. 9 Peterson, Decentralization in Latin America: Learning through Experience (World Bank, 1997). 13 2.35 The financial sector institutions lending to municipalities face a special challenge. Unless protected by special government regulation, they must compete on equal footing with all other intermediaries in domestic and international markets to assemble savings. At the same time, they must master the complexities of local government finances and local capital budgets, in order to make productive and creditworthy loans. 2.36 Financial sector competition is relatively new to municipal lenders in most parts of the world. Special financing circuits commonly have been used until recently to raise funds for municipal investments. Throughout Western Europe, for example, small savers' deposits often were assigned to municipal investment. Individual savings deposits were required to be made in the postal savings bank system or in special "municipal banks" at below-market rates of interest. These savings were aggregated by specialized financial intermediaries and lent to municipalities, also at below-market rates. Similar closed circuits supported credit financing in other sectors requiring long-term investments, such as small business development, agriculture, or housing mortgages. 2.37 In developed countries, segmented financing of this kind has largely disappeared. Although under pressure, closed-circuit institutions providing sector-specific credit still survive in many developing economies. Municipal credit in particular still is typically separated from the broader credit market and provided either through specialized public institutions or through publicly regulated private-sector monopolies. 2.38 In "repressed" financial markets of this kind, credit ends up being allocated largely by government. At below-market interest rates, demand for credit tends to greatly exceed the supply that can be financed from savings. Government institutions therefore decide which borrowers will receive funds within the municipal sub-market and allocate credit supply across sub-markets by establishing special lines of credit for specialized financing institutions. During the mid-1980s, two-thirds or more of the economy's entire credit flows were steered by government in Brazil, India, Pakistan, and Turkey, among a number of different developing countries.'0 Financial sector reforms have recently diminished government's role in steering credit and have opened investment banking to private sector competition. Municipal credit institutions, however, have remained relatively untouched by these reforms. Municipal Credit Systems and the Efficiency of Investment 2.39 Market-based municipal lending, in addition to tapping new sources of financing, should help increase the efficiency or productivity of municipal investment. The use of market- determined interest rates in lending, for example, has been found generally to be positively correlated with average investment returns. From the borrower's perspective, low-return investments that nonetheless are profitable when financed with subsidized loans become unprofitable at the market rate of interest, and are squeezed out when the investor must pay the full market cost of capital. Financial sector "deepening" also has been found to be positively correlated with investment returns. Financial intermediation by a bank or other lender adds the '° World Bank, Financial Systems and Development (1990). 14 lender's professional appraisal of a project to the investor's own assessment, potentially providing a double check on investment returns.1 2.40 The structure of municipal credit arrangements goes far toward determining how much of this potential for boosting municipal investment efficiency is realized in practice. Investment productivity is supported most effectively when participants on both sides of the credit market have a clear stake in a project's success. This kind of incentive is clearest in project financing, where the lender's repayment depends on the project's ability to generate adequate revenue. 2.41 General obligation loans create less incentive for project efficiency. Returns to the individual project may be largely irrelevant to the lender, if local debt is secured by the municipality's overall budget. General obligation loans instead create incentives for general financial discipline. In a well-functioning private credit market, municipalities whose budgets are out of control will be cut off from further credit or charged steep risk premiums. Local governments that want to use credit have a strong incentive to maintain budgetary discipline. A number of empirical studies have confirmed the practical importance of this incentive in restraining local budget behavior.'2 Indirectly, general budgetary discipline favors efficient investment by screening out risky projects that would jeopardize the budget. 2.42 Municipal loans backed solely by central government transfers or protected by unconditional government guarantees, however, can create perverse efficiency incentives. They encourage lenders to skip project review altogether, and to even ignore local financial conditions. As long as a loan is adequately covered by central government commitments, the lender has no reason to limit its lending to economically feasible projects. On the contrary, there is every incentive for profligate lending and for inefficient or corrupt local spending. For more than a century unwise government guarantees have encouraged inefficient infrastructure investment of the kind described in Boxes 2 and 3. ' Allocating capital through competitive credit markets, however, appears to have less decisive effects on returns to public investment then returns to private investment. A review of World Bank projects found that while private sector investment returns were substantially higher in countries that had de-regulated financial sectors and used market rates of interest for lending, investment returns on public projects were much less sensitive to the financial policy environment. Public authorities appeared to respond less to price signals than private firms in making investment decisions. See Kaufman, "The Forgotten Rationale for Policy Reform: The Productivity of Investment Projects." (World Bank Background Paper, 1991). 2 For review of this literature, see Goldstein and Woglom, "Market-Based Fiscal Discipline in Monetary Unions: Evidence from the U.S. Municipal Bond Market," International Monetary Fund (1991). 15 Box 2. The Distorting Effects of Government Guarantees: An Historical Perspective During the 19th century, major infrastructure development projects like railways were financed largely by international lending. Foreign lenders were unfamiliar with economic conditions in the countries where they were providing project finance, and could not devote the time needed to master the budgetary arrangements or regulatory rules of each country. Instead, they insisted on obtaining central government guarantees for loans. One consequence was a rash of wasteful investment. In India, railway companies were guaranteed a minimum five percent rate of return on all investment, with government obligated to make up the difference in the event of a revenue shortfall. The interest clause in railway bond covenants was backed by a general obligation pledge from the central government. In the words of one historian, "All the money came from the English capitalist, and so long as he was guaranteed 5 percent on the revenues of India, it was immaterial to him whether the funds that he lent were thrown into the Huoghly River or converted into brick and mortar." The list of uneconomic or partially built railways was impressively lengthy. In the recent past, Mexico's rush to build a national highway system through private construction of toll roads was financed primarily by project lending. State banks, acting on political direction, provided much of the financing and supplied guarantees both to private lenders and contractors. One consequence was a vastly over-designed and expensive toll road system. A second consequence was a hemorrhaging of the state banks that provided guarantees. When the economic earnings of the toll roads proved inadequate to service private lenders' debt, state guarantee institutions ended up with much of the losses. Sources: Eichengreen, "Financing Infrastructure in Developing Countries: Lessons from the Railway Age," in Mody (ed.) Infrastructure Delivery: Private Initiative and the Public Good (1996); "A Wrong Turn on Mexico's Tollways," Los Angeles Times (June 25, 1996). Box 3. The Distorting Effects of Government Guarantees: A Modern Example Municipalities in Turkey began disowning debt on a large scale in 1996. Both Istanbul, the largest city, and Ankara, the capital, informed the national Treasury that they would not be repaying foreign debts. As of June 1996, approximately $400 million in debt payments due to Japanese and German lenders had been abandoned by cities. Conflict with the central government during the first half of 1996 over grant-in-aid payments exacerbated the fiscal situation. In the words of one municipal spokesman, "They don't pay us (grants); we don't pay them (banks and bondholders)." The international credit ratings of Turkish municipalities on stand-alone debt plummeted. Despite this turmoil, the credit market remained largely untouched by events. Some municipal debts were explicitly guaranteed by the national government. All other debts contained cross-default clauses, so that default on any one credit triggered default on the others, including those guaranteed by the Treasury. Under these conditions and to preserve national credit market access, the central government assumed responsibility for all payments not made by municipal borrowers. In total, between $8 billion and $10 billion in municipal foreign currency debt is outstanding. An international banker was quoted as stating: "As long as I am paid, I don't care where the money comes from." In the midst of the crisis and following Istanbul's and Ankara's disavowal of repayment obligations, a new Dm200 million loan was extended to the municipality of Izmir through an international syndicate of commercial banks, with a central government guarantee. Lenders, in effect, ignored the quality of the local government general obligation pledge and looked only to the sovereign guarantee. 16 Source: Euromoney, Emerging Markets Report (July 1996). 2.43 Private-sector lenders often acknowledge that, when municipal loans are fully protected by government guarantees or intercept provisions, they spend no time at all on municipal project review and do not attempt to master the complexities of a local government's budgeting.'3 Government guarantees have this perverse effect whether offered in support of general obligation or project loans. 2.44 An inescapable dilemma underlies the use of government guarantees. They have been responsible for a good deal of inefficient investment, but at many times they also have been essential to stimulating municipal credit flows. Without loan guarantees, much of 19th century urban infrastructure investment would not have taken place when it did and much of today's municipal lending would not occur. This accounts for the continued use of guarantees. The designers of credit systems need to recognize, however, that transferring credit risk to the central government eliminates most of the incentives private lenders otherwise would have for monitoring local investment projects and investigating local finances. At the earliest possible stage of market development, private lenders should be required at least to share in the financial risks associated with municipal lending. What Constitutes "Excessive" Borrowing? 2.45 Contentious discussions have erupted on occasion between local officials, national officials, and representatives of international organizations over "excessive" local borrowing. Participants in these discussions often talk at cross purposes because they implicitly adopt very different interpretations of what "excessive" borrowing means. The framework presented in this chapter can help clarify these debates. 2.46 For those involved directly in municipal lending, the critical measure of creditworthiness is a municipality's ability and willingness to repay its debts. From this perspective, it is prudent for local authorities to borrow as long as they can repay their debts with a margin of safety. "Excessive debt" is debt the locality cannot repay, or is at risk of not paying. A municipal finance officer or bank credit analyst normally would have this perspective. 2.47 From the perspective of economic efficiency, "excessive" local debt implies that local authorities are borrowing to finance projects that have lower rates of return than projects in the rest of the economy or rates of return that are lower than the true cost of capital. Credit could then be more productively re-allocated to other sectors. A national investment planner or analyst for a multi-lateral development bank might maintain this perspective on "excessive" borrowing " See Peterson, "Colombia Municipal Development Financing: Current Performance and Future Role of FINDETER" (Inter-American Development Bank, 1996), for interviews with commercial bank lenders who acknowledge that they do not examine a project's economics or the general financial condition of the local borrower when lending to local authorities. They examine instead the adequacy of the municipality's pledge of intergovemmental tax-sharing to cover debt servicing. 17 used to finance "excessive" local investment.'4 This condition is most likely to arise in segmented markets, where the amount of municipal lending and investment are determnined by government, or municipal loans are subsidized or guaranteed by government. 2.48 In practice, a third, macroeconomic criterion often has been used in defining "excessive" sub-national debt. It is debt that contributes significantly to total public sector borrowing in an environment of aggregate public sector budget deficits. Under these conditions, the public sector in effect borrows to finance current account deficits and crowds out credit that would be used to finance private investment. 2.49 By this last standard, sub-national debt has sometimes been held to be "excessive" even when it is within local governments' capacity to repay and the state-local sector is using debt to finance productive investment. Colombia provides a case in point.5 Local government borrowing surged when the 1991 constitution formally promised a rapid increase in tax sharing with local authorities, as part of the nation's commitment to decentralization. The promised growth in intergovernmental transfers was quickly used by local governments to collateralize a surge in bank loans, which in turn were used primarily to finance investment, especially for expansion of local water and wastewater systems. So far so good. Increased investment in the water and wastewater sector was an explicit national planning priority, and by law the growth in tax sharing was supposed to be used by municipalities exclusively to finance social investment. The use of government tax sharing as the primary security for municipal borrowing may have led to some inefficiency in investment, but in its broad outlines the local investment response was just what national authorities had hoped for. 2.50 Unfortunately, the rapid increase in the local share of national tax collections exacerbated the aggregate public sector deficit. The central government did not transfer to local authorities sufficient new service responsibilities to offset the revenues being transferred. Nor did central government cut back its own spending commensurately with its revenue losses. The result was a rapidly mounting central governrnent deficit and a substantial increase in total public sector borrowing. In partial response, Colombia's national government introduced a number of emergency measures to limit sub-national borrowing. 2.51 Excessive revenue-sharing commitments by central government lay at the heart of this impasse. Fiscally unbalanced decentralization of this kind, in which the central government transfers revenue to local authorities faster than it transfers expenditure responsibilities and faster 14 Note that if all project benefits can be captured in financial returns, if loans are made at competitive market rates, and if project revenues are the sole source of debt security, the first two perspectives on debt will coincide. Under these limited conditions, the economic return to a project is also its financial rate of return, and the financial rate of return provides the lender with the only revenue stream to repay debt service. A loan that is financially prudent for the lender and borrower then is also economically desirable. '5 International Monetary Fund, Colombia: Endeudamiento Territorial Gesti6n y Consideraciones Macroeconomicas (July 1995). 18 than it can cut back central spending has triggered fiscal imbalances in a number of Latin American nations, including Brazil and Argentina.6 2.52 There is indeed an "excessive debt" problem in these circumstances. However, it is paradoxical to assign blame to the local credit market. In Colombia's case, the explosive growth in public sector deficits occurred at the level of national government and national public enterprises. Local governments continued to accumulate budget surpluses. Imposing credit ceilings on local authorities still may be an appropriate short-term response. It prevents local governments from permanently "locking in" the fiscal imbalance by claiming that the future growth in revenue transfers is needed to repay local debts which already have been collateralized by the promised transfers. However, an emergency measure of this type can only buy the central government time to restore fiscal neutrality in its decentralization policy. Central authorities must either transfer additional expenditure obligations to the local level (and cut back national-level spending), or retract part of the constitutional promise to increase tax sharing. Central government deficits cannot permanently be cured by restrictions on local borrowing. 2.53 As Colombia's example illustrates, local credit markets in developing countries rarely can be considered in isolation of national budget conditions. As long as a large part of local revenue consists of intergovernmental transfers or tax-sharing payments, municipal finances will be inextricably bound up with central government finances, even if the credit market itself is operated by private institutions. Formulating a Strategy for Municipal Credit Market Development 2.54 Figure 2 provides an overview of a country's potential for municipal credit market development at a point in time. Along the vertical or y-axis is arrayed the stability and adequacy of the municipal finance system. This measures potential, creditworthy demand for municipal borrowing. Along the horizontal or x-axis is shown the degree of financial sector development. This measures the economy's ability to assemble financial savings and to use financial intermediaries to supply investment financing in the form of credits. A few countries have been located on the Figure for illustrative purposes, using only qualitative judgments as to their conditions in 1996. 2.55 Broad strategies for developing municipal credit markets should begin with an assessment of where a particular country stands in Figure 2. In the lower left quadrant of the Figure are countries where any kind of municipal credit market development is premature. Kenya, for example, is plagued by a municipal finance system that leaves municipalities highly dependent upon central government transfers. Transfer policy, however, is not spelled out in law. Central authorities determine transfer amounts by non-transparent criteria, and often end up transferring less than has been promised. The different levels of government are also trapped in interlocking debts that remove the incentives for any one party to clear its debts. Municipal lending in particular has been crippled by a parastatal lending system that has awarded financing 16 Vito Tanzi, Fiscal Federalism and Decentralization: A Review of Some Efficiency and Macroeconomic Aspects (Annual Bank Conference on Development Economics: May 1995). 19 capriciously, then failed to collect debt service. More than 90 percent of the loans made by the Local Government Loans Authority are estimated to be in default or seriously past due. In these conditions, it is pointless to introduce a new or more "serious" form of municipal lending. The only reasonable priorities are to stabilize the municipal financing system and to find a solution to the accumulation of interlocking government debts.'" 2.56 Bulgaria in 1996 is another country where municipal lending of any kind would be premature. With the banking system in collapse, inflation careening upward, and municipalities uniformly running deficits they cannot finance, any municipal loan program would only accumulate destructive experience that would set back eventual development of a municipal credit market. The country's overwhelming priorities are to stabilize inflation and establish a comprehensible and realistic system of municipal finance. Figure 2. Potential for Municipal Lending . Czech Republic U. South Africa (Former White Townships) 0. U_ Chile Poland Colombia o~~~~~~~~Pln 0 co Philippines South Africa . | . Romania Brazil (Former Black Townships) Co 2:% Bulgaria Kenya Financial Sector Development 2.57 For countries in the middle range of Figure 2, there are more options. Traditionally, countries in this stage of development have used Municipal Development Funds to meet local governments' credit needs. Most of the states of Brazil, for example, utilize municipal development funds supported by the World Bank or Inter-American Development Bank. Some of these are in their fourth generation of large-scale external funding. Other countries, like " See Paul Schmoke, Local Government Finance in Developing Countries, pp. 83-87, 131-133, and 158-160. 20 Romania, are taking the first tentative steps toward establishing a municipal development fimd, with external support. Countries have found other possibilities, however. Poland, for example, for several years weighed various options, including setting up a specialized municipal development fund in the form of a municipal bank, owned by municipalities and a Municipal Development Fund to lend directly for local on-lending. While it deliberated, another kind of credit market emerged. The government created Environmental Funds and an Environmental Bank to finance local environmental projects at subsidized interest rates. Meanwhile, a private- sector municipal bond market and bank loans emerged as a means to finance projects not eligible for environmental subsidies. Municipal credit market became possible once the municipal financing picture stabilized and inflation was brought under control. 2.58 Countries in the upper-right quadrant of Figure 2 face a transitional decision. They are ready to integrate municipal lending into the general, private-sector credit market, but must decide whether and how they will make the transition. Will municipal credit supply be turned over to the competitive private sector? Or will publicly controlled institutions gradually be modified so that, while formally remaining monopolies, they are forced to act more like competitive intermediaries, raising capital at market rates and becoming exposed to the full risks of lending without state guarantees? 2.59 Some of the countries shown in the Figure represent exceptional circumstances. South Africa, for example, possesses a sophisticated financial sector fully capable of raising and allocating capital to projects involving known risks. At one time, it had an active municipal bond market. The country's municipal finance system, however, is undergoing fundamental reform as former black townships are consolidated fully with former white townships. This process has left the budget picture and borrowing capacity of local authorities in doubt. The increasing use of subsidized government loans as a kind of fiscal transfer has undermined private lending. In preparation for South Africa's new constitution, the Financial and Fiscal Commission was able to establish the principle that municipal lending should take place at market rates of interest within a competitive environment. However, development of a new municipal credit system will require greater clarification of the expenditure responsibilities and permranent revenue sources of local governments.'8 2.60 Chile illustrates another kind of special case. The country has developed very advanced capital markets. Its local government financing system is one of the most stable in the developing world. However, the country has chosen not to develop a municipal credit market. In the late 1960s. one of the destabilizing factors in Chilean society was the local government sector, which for a time was running budget deficits of 50 percent and greater. Subsequent constitutions in Chile have forbidden local governments from incurring deficits or borrowing. Local infrastructure investment is either financed by the national government, through capital grants to municipalities or direct investment by decentralized regional administrations, or financed by private sector owners of infrastructure projects operating under local government regulation. If municipal governments finance investment projects, they do so out of current account savings. 18 Office of the President, Municipal Infrastructure Investment Framework (Republic of South Africa: October 1995). 21 2.61 In sum, at each stage of credit market "potential," as shown in Figure 2, a different set of policy issues confronts those wanting to push ahead with local credit market development. Some of the key choices to be made are addressed in the next two chapters. III. MODELS OF MUNICIPAL CREDIT SYSTEMS 3.1 Regardless of its current stage of credit market development, any country that plans to introduce new institutions into the municipal credit market, or to reform municipal lending procedures, should be clear about the kind of credit system it eventually wants to establish. Municipal "credit" comes in many forms. 3.2 Countries in the upper right quadrant of the Figure shown in the previous chapter (Figure 2) face an almost immediate choice of credit models. Countries toward the middle of the Figure are more likely to be establishing the basic rules that will guide the municipal credit market and the institutions that will participate in its growth. The precedent setting effect of these choices should not be underestimated. If a public sector bank with monopoly powers and subsidized interest rates is given the responsibility for municipal lending, it will be difficult later to dismantle it or to convert into an unsubsidized competitor. A country that begins by legally restricting municipal lending to banks may find that it has inadvertently suppressed a budding municipal bond market, or vice versa. 3.3 Very few countries, when taking the first steps in municipal lending, have articulated how they hope to build on their initial experiences to create a self-sustaining municipal credit market. Of the many project loans provided by multi-lateral donors and bilateral aid agencies to strengthen municipal credit institutions, only a rare few provide even a glimpse of how the counterpart institution is expected to grow beyond reliance on donor funding. It is perhaps not surprising, then, that few have done so. The Theory of Intermediation: Bank Loans vs. Municipal Bonds 3.4 One fundamental choice that market designers face is the nature of the credit instrument. In Western Europe, specialized municipal banks supply most municipal credit in the form of long-term bank loans. In the United States and Canada, municipalities borrow primarily by selling municipal bonds in the capital market. There are some signs of convergence between these two credit systems. Municipal bonds have gained a foothold in Europe, and have become an important part of the municipal credit market in some places, like the Nordic countries. In project finance, European banks (as well as Japanese financial institutions) and European service firms have become major financiers of some types of U.S. infrastructure projects, like toll roads or water privatization. 3.5 Still, the choice between a municipal credit system based on bank lending and a system based on bond issuance highlights many of the key issues to be decided in credit system design. 3.6 The differences between bonds and bank loans are rooted in financial sector intermediation. In modern intermediation theory, banks perform what is called "delegated 23 monitoring."' Municipalities and other borrowers could, in principle, deal directly with individual lenders, by borrowing investment funds from large financial institutions like pension funds or insurance companies. They could even borrow from individual savers. However, unless the loan at stake is large, it is inefficient for each saver to try to monitor financial condition and all the other factors affecting loan payment. A bank performs this intermediation and monitoring function. It gathers savings from numerous sources, assembles specialized professionals capable of loan appraisal and loan oversight, allocates capital, and then monitors both its loans and the financial condition of borrowers. 3.7 Of course, there are costs associated with this intermediation. Big lenders like insurance companies or pension fumds may find it more efficient to make large loans to low risk clients on their own, without going through banks as intermediaries. When borrowers have strong finances and transparent budgets, monitoring can be relatively straightforward. The bond market represents another kind of disintermediation. Local governnent bonds are sold directly to household savers as well as to financial institutions. The system still requires monitoring of municipal financial conditions. However, the monitoring function has come to be performed by specialized credit rating agencies who make the results available to the market as a whole. Whereas banks typically seek to build loan departments that possess proprietary information and proprietary methods of analyzing creditworthiness, in a municipal bond market, informnation on local financial condition is provided by issuers to the market at large. Rating agencies widely disseminate their assessments of this information. Municipal Banks as Specialized and Quasi-Monopolistic Lenders 3.8 Bank lending can occur either in a competitive environment populated by many commercial banks or in a regulated environment where a single banking institution has a monopoly on servicing the municipal sector. In Western Europe, municipal lending in most countries began as the responsibility of a single specialized bank, within a regulated banking system in which each sector of the economy tended to have its own development bank to meet its long-term financing needs. The development banks were either owned or tightly regulated by government. Box 4 summarizes the development of one of these institutions, the Municipal Bank of the Netherlands, and compares it to similar municipal lending institutions that have evolved elsewhere in Western Europe. Box 4. Municipal Bank of the Netherlands: The Western European Approach to Municipal Credit Intermediation Brief History The Municipal Bank of the Netherlands (BNG) was founded in 1914 at the initiative of the Dutch Association of Municipalities, with the motive of improving individual municipalities' access to the credit market. However, only a few cities subscribed to the new institution. In 1922, the bank was extensively reorganized and the Dutch government became the major shareholder. The bank has grown to the point where it now has more than $60 billion of assets, about 90 percent of which consist of loans to municipalities and other kinds Douglas W. Diamond, "Financial Intermediation and Delegated Monitoring," Review of Economic Studies (1984), Li, 393-414. 24 of local government units, such as housing authorities. The bank presently is owned 50 percent by the national government and 50 percent by municipalities. Municipal Creditworthiness In recent decades, the bank has experienced negligible problems with municipal loan repayments. In addition to the bank's appraisal of individual loans, municipal credits are backed by the Municipal Fund which disburses State transfers to municipalities and is adrninistered on behalf of the government by BNG. Special provisions allow Municipal Fund monies to be used to support municipalities that are in financial difficulty and to guarantee their loan repayments. Municipalities drawing on the Fund in this way are subject to limitations on further borrowing and must follow designated steps in restoring fiscal health and prudent financial management. As a result of the excellent loan repayment record, municipal loans in the Netherlands carry no additional risk weighting for reserve purposes (i.e., they are classified as carrying the same risk as central government debt). BNG itself enjoys an AAA rating from all of the international rating agencies, allowing it to raise financing worldwide at low cost. Development Bank Functions The Dutch system relies on decentralized project preparation. BNG does not participate in local project development. It appraises loan applications primarily from a financial point of view. Part of the Bank's mission, however, is to help implement national policy toward local authorities (e.g., in financing the recapitalization of the municipally owned housing stock to meet European Union subsidy guidelines). The Bank handles a wide array of financial transactions for local governments. It is responsible for 90 percent of financial transactions between central government and local government units, including the monthly payment of grants and shared taxes. The majority of municipalities maintain their accounts with BNG. Some of the other Western European municipal credit institutions play an even more active "relationship" role with local governments through local technical assistance. Credit Local de France, for example, seeks to establish a permanent financial advisory relationship with its municipal clients; this includes advising municipalities on budget preparation and capital planning. In the German system, a municipality owns shares in the local municipal bank, which by law provides the municipality with long-term credit as part of a permanent partnership. Raising Loan Funds For a long period BNG, like other municipal credit intermediaries in Western Europe, enjoyed regulated access to personal saving deposits and also received substantial government capital contributions. However, in recent years financial sector liberalization has forced it to move into the competitive capital market to raise funds. Through the 1980s and into the early 1990s, competitive capital was raised primarily in the form of direct, privately tailored loans from large domestic financial institutions, principally pension funds and insurance companies. These instruments allowed BNG to match long-term liabilities with the long-term assets it held in the form of municipal loans, while also serving the financial institutions' needs for safe, long- term fixed income investments. In recent years, however, pension funds and insurance companies have invested increasingly in the equity market and real estate, reducing the volume of funds available for private lending. BNG has turned to international markets for most of its incremental funds. About two-thirds of new funds are now raised through international bond issues. Because BNG is exempt from income taxation and earns stable fees on the financial transactions it manages on behalf of government, it is able to on-lend funds to municipalities at modest margins over its own low cost of capital. The other municipal financial intermediaries of Western Europe have gone through similar evolutions. The predecessor of Credit Local de France (CLF) historically enjoyed monopoly access to small-saver deposits collected through the French postal savings systern. These deposits paid low, below-market interest rates, which the institution passed on to municipal borrowers. Financial sector liberalization changed this relationship. Today, CLF raises more than 80 percent of its funds through international bond issues. Credit Communal Belgique (CCB), in contrast, retains monopoly access to individual savings accounts and 25 individual savings bonds in Belgium, although this monopoly position is expected to be eliminated in the future. Exclusive access to regulated household deposits has both lowered CCB's cost of funds and given CCB far more savings than it needs to meet municipal credit demands. CCB therefore has acted as a financial intermediary in supplying funds to the rest of the Belgian financial system. As part of the accelerating deregulation and consolidation of European financial markets, Credit Local de France and Credit Communal Belgique announced in July 1996 that the two institutions would merge, seeking to establish a pan-European institution specializing in sub-national financing. Government Ownership and Government Guarantees Government historically has owned a controlling share of municipal banks and provided them with de facto guarantees against municipal defaults. Both of these "special relationships" are now disappearing. Credit Local de France, the largest of the municipal banks, is now in the final stages of a privatization begun in 1993. As of 1996, the government's share of equity ownership had been reduced to 12 percent, with the final block of government shares scheduled for sale shortly. Management of the Banco de Credito Local of Spain, formerly owned by the State, likewise has been transferred to a financial group (Argentaria) which is now completing its privatization. National governments also are eliminating their guarantees of municipal loans, in line with both decentralization policies and financial sector liberalization. Previously, for example, municipal budgets in France were reviewed and approved by the State, including approval of local bprrowing and the local budget's provision for debt servicing. Most local revenues also were provided by the State. It was believed that this relationship carried with it an implicit obligation on the part of the State to provide additional revenues, should its approved budget prove inadequate to cover debt service. For decades, the State did indeed provide these implicit guarantees. With decentralization of the French local government system in the 1980s, however, the State expressly abandoned its guarantee position, forcing the re-negotiation of many outstanding high-interest municipal loans as local finances came under pressure. Spain's Banco de Credito Local (BLF) has made one of the most recent transitions from state-protected status. It therefore may be of special interest to developing countries. Until the 1980s, BLF was funded by direct contributions from the State budget and by State loans made on favorable terms. Its funds were replenished by the State when municipalities failed to make full debt payments. As a first step toward liberalization, BLF was required to begin raising part of its funds on the competitive bond market, but its bond issues were supported by Govemment guarantee. The State guarantee later was withdrawn, and BLF was folded into a larger financial organization being prepared for privatization. Most recently, local and regional borrowing has surged, both as a result of decentralization of service functions and in anticipation of the doubling of unrestricted revenue sharing to 20 percent of national tax collections, which took place in 1997. The enlarged transfer commitment has provided sub-national govemments with the wherewithal to support much larger debt service. Overall Competitive Position and Market Supply of Credit Between 1965 and 1975, BNG, like several other municipal banks, enjoyed a formal monopoly in the Dutch municipal credit market. Government policy then removed the monopoly. BNG's low cost of funds, tax exemption, and historical relationship with borrowers, however, makes it difficult for commercial banks to compete with it. The most substantial competition occurs in the form of direct loans by pension funds and insurance companies to larger municipalities, which do not pass through the intermediation of BNG. Competition has reduced the dominance of some of the traditional municipal banks in Europe. Credit Local de France, for example, in 1995 accounted for 42 percent of local authority lending in the French market, where competition is intensifying, both from bank loans and an emerging municipal bond market. Opponents of de-regulation feared that subjecting municipal banks to competition would drive away credit from the municipal sector. In all of these countries, however, long-term credit flows to local governments have increased substantially since the introduction of competition. 26 3.9 Long-term ties between a certain class of borrowers and a bank specializing in meeting their development needs sometimes is called "relationship banking." Relationship banking is most valuable during a borrower's start-up stage or when an institution is first entering the credit market. A relationship bank that is protected from competition can afford to take a longer-run view of its partnership with clients. Empirical studies have found, for example, that monopolistic development banks tend to subsidize clients' borrowing in their early stages of development, then later recover these costs by charging more than market rates of interest once the clients have become better established. 3.10 In the municipal sector, one argument for creating a specialized municipal bank, protected by the state from competition, is that only a relationship bank of this kind can afford to support localities in their earlier stages of learning about the credit market. The banking partner is likely to help the municipality prepare its loan applications and structure its investment projects. It may handle the municipality's payment system and keep its surplus cash on account. It will help the community in budget management, and even lend to a new borrower at the same interest rate charged larger and more experienced municipalities. A bank can afford to nurture small communities in this way only if it is subsidized by the State, or if it knows that, as monopoly lender to the sector, it is creating a long-term client with which it will conduct more profitable business in the future. 3.11 As borrowers grow in economic strength, the advantages to them of relationship banking tend to diminish.2 They no longer require the same kinds of financial hand-holding from their bank. In competitive markets, experienced borrowers that have established a "reputation" for good financial management and have compiled a record of prompt debt servicing will find other institutions willing to lend to them, often at lower costs. Their loans now require less intensive monitoring and involve less credit risk. This maturation of borrowers can place strain on a closed credit system. 3.12 Municipal credit systems of this kind share the spirit of a credit union. In fact, a number of the municipal banks in Western Europe began as communal banks, created by associations of towns and cities out of a spirit of mutual assistance. Members were motivated by collegial concern for creating credit access for local governments in general as well as by self- interest. Like all credit unions, municipal credit systems of this type must confront internal pressures. Financially strong members may be tempted to withdraw from the system, especially if they are forced to accept artificially low rates for funds on deposit at the municipal bank as well as higher interest rates on loans. 3.13 Municipal financing institutions of the Western European type traditionally have been used by governments to help implement national policy toward municipalities. They continue to play this role, though to a lesser degree than formerly. The Municipal Bank of the 2 Douglas W. Diamond, "Monitoring and Reputation: The Choice Between Bank Loans and Directly Placed Debt," Journal of Political Economy (1991), vol. 99, no. 41, 689-721; Charles P. Himmelberg and Daniel P. Morgan, "Is Bank Lending Special?" in Peek and Rosengren (eds.), Is Bank Lending Important for the Transmission of Monetary Policy? [Boston: Federal Reserve Bank of Boston, 1995]. 27 Netherlands (BNG) administers the government's special credit assistance program for distressed cities. When the government, following EU guidelines, reduced recurring subsidies for municipally owned public housing, it set up a special credit program to recapitalize the municipal housing stock. This program, too, was administered through BNG. 3.14 In Great Britain, the government-run Public Works Loan Board (PWLB) serves as an instrument for macroeconomic control of sub-national borrowing. The central government establishes loan ceilings for local authorities. The fact that upwards of 85 percent of municipal borrowing occurs through the PWLB makes compliance with the loan ceilings easier to monitor and enforce. The potential for using a single credit institution owned or controlled by government to implement financial policy toward sub-national government helps explain the appeal that municipal banks have had in countries with historically centralized governance systems. Municipal Bonds 3.15 Municipal bonds are a form of direct access to the capital market. They represent a less personal and less permanent connection between borrower and lender than is incorporated into relationship banking. Bond underwriters try to cultivate continuing ties with municipal authorities, but the essence of a public bond market is that every deal is competed anew in an open market place. 3.16 A municipal credit system based on bonds will possess certain fundamental characteristics that distinguish it from a credit system based on a single municipal bank. A bond market is inherently more decentralized. Municipalities issuing general obligation bonds receive all of their funds up front. They are not subject to partial payments based on a bank's monitoring of their project construction progress, as is typical of bank loans to local governments. 3.1-7 The aggregate supply of credit to the local sector is similarly more difficult for central authorities to monitor or control when each municipality sells bonds directly to independent investors than when they borrow from a single authority whose credit ceilings are set by government. Central governments may find it more difficult, too, to influence local governments' investment choices or to satisfy themselves that local government projects have been well planned, when municipalities obtain credit directly from the private capital market rather than from a government-authorized municipal bank. A bond market is an awkward instrument for exerting central government control over local borrowing, but a natural instrument for financing decentralized investment choices. 3.18 Bond markets rely on public disclosure of municipal financial information to function effectively. Most financial systems utilizing bond issues have extensive public disclosure requirements that issuers must comply with. A municipality increasingly must perform this self-monitoring function as long as it has bonds outstanding, and must report publicly to the market on any significant changes in financial condition. Credit-rating firms have developed a presence in every municipal bond market of significant size. They also report publicly and exert substantial influence over the market. Many financial institutions are 28 precluded from buying bonds that are rated below a certain threshold of credit risk by the principal rating agencies. 3.19 A municipal bond market unbundles the various support functions that a municipality can receive from a municipal bank. Whereas a relationship bank seeks to build an all-encompassing partnership with a municipality, a bond market only supplies capital. Municipalities make separate decisions about where to maintain their liquid deposits, where to obtain financial advisory services or technical assistance on project design, and how to manage their internal payment systems. Bidding for each of these support activities can be competed separately in a sub-market of its own. Unbundling services and subjecting them to competition is likely to lower total costs. However, unbundling may deprive municipalities of the benefits of having a comprehensive partner familiar with how a municipality's financial activities fit together. 3.20 Unbundling the various financial services provided to municipalities also requires that several competent firms already exist in each of these fields, so that there can be meaningful competition in each sub-market. At a minimum, the skills must exist in the economy to form such firms once demand for them becomes apparent. For this reason, bonds are likely to emerge as a significant financing option at a later stage of development than comprehensive municipal banks, in countries where both systems operate side by side. 3.21 Reliance on bond financing can be especially problematic for small communities. The fixed costs of bond issuance (legal costs, underwriter fees, printing, etc.), when spread over a low volume of bonds, can make the total cost of bonds uncompetitive with other sources of financing. Nations that rely heavily on the bond market for municipal credit supply have solved this problem through "bond banks" which pool the issues of numerous small communities. Bond banks, as their name implies, are a type of financial intermediary. They perform some of the functions of a specialized municipal bank-such as helping to strengthen a municipality's credit quality, and assisting small municipalities in the paperwork necessary to borrow-while using bonds as the way to raise funds and to on-lend to individual local governments. Box 5. Evolution of the U.S. Municipal Bond Market State and local governments in the United States have approximately $1.3 trillion of municipal bonds outstanding. That represents roughly 8 percent of the domestic credit market. Brief History The U.S. municipal bond market is now a mature and relatively stable market. During the 19th and early 20th centuries, however, it was very much a developing-country market, facing many of the same problems that today's developing countries face in creating and testing different kinds of market institutions. At the time of most rapid urban growth in the United States, cities themselves were responsible for financing the bulk of their capital improvements. This decentralized form of public investment was consistent with the use of locally-issued bonds as a decentralized financing tool. Between 1840 and 1880, when urban populations exploded (Chicago, for example, grew from about 30,000 to 2,000,000), local government debt also exploded, rising from $25 million to $821 million. At the turn of the century local government debt (excluding the debt of states) exceeded that of the federal government by 50 percent. Between 1902 and 1932, total state and local debt again grew by a factor of almost 10 times. 29 Especially in the earlier period, municipal bonds issued by rapidly growing midwestern and western cities were sold largely to investors in the eastern United States and to Europeans. The growth of the municipal bond market revealed serious cracks in the credit system. Periodic recessions precipitated bond defaults. Foreign investors in particular sought protection from both the federal government and the states, arguing that they were implicit guarantors of the debt of local jurisdictions within their boundaries. The courts rejected this argument. State governments did try to limit their bond-market exposure. Most sta4es established debt ceilings for local governments. They prohibited municipal guarantees of private loans as well as municipal borrowing to finance joint ventures with private entrepreneurs. Voter approval of general obligation bond issues was commonly required. To further limit their risks, many states formally amended their constitutions to prohibit state assumption of municipal indebtedness. Limiting Credit Risk Over time, three important mechanisms have emerged to reduce credit risk in the U.S. bond market. Credit rating agencies now rate virtually all publicly issued bonds. Credit rating has become a large and sophisticated business that reliably distinguishes risk. However, in its early years credit rating was much more haphazard. Of all the municipal bonds that defaulted during the Great Depression of the 1930s, 48 percent were rated AAA, the highest safety classification, from Moody's, the principal bond-rating agency of the time. Almost 80 percent of defaulting bonds were rated AAA or AA, the two highest safety categories. One author has concluded, "During this period municipal rating was characterized by superficiality and inexperience. The staff at Moody's did not exceed four people." The current sophistication of the U.S. credit rating system does not imply that start-up institutions in developing countries will be able to discriminate municipal credit risks successfully from the beginning. Public disclosure of municipal financial conditions has greatly improved the information from which risk assessments are made. In the wake of New York City's financial crisis of the mid-1970s, voluntary disclosure guidelines were adopted by the Municipal Finance Officers Association and the Public Securities Association. These guidelines comprehensively identified the kind of municipal financial information that should be made available to the public before a bond issue could be sold, and recommended standardized formats for presentation. Later, the disclosure guidelines became mandatory and regulated by government. Municipalities also are now required to report any significant changes in their financial or legal condition, as long as bonds are outstanding. Private bond insurance has further reduced purchasers' risks. Almost half of all municipal bonds issued today are insured by private insurance companies for timely payment of interest and principal. Unlike free government guarantees, private bond insurance does not create perverse efficiency incentives. Municipalities must pay a premium for insurance coverage. The insurance companies have specialized staffs that assess the risks involved in a municipality's finances or in a project financed by revenue bonds. The greater the credit risk, the greater premium a municipality must pay to obtain insurance. Often, the insurance company will advise a municipal borrower on how it can restructure a project to reduce economic risk. Raising Capital and Tax Exemption In the United States, upwards of 75 percent of municipal bonds are now sold to individuals, either directly or through mutual funds. This reflects the historical exemption of municipal bond interest from personal income taxes, which makes the bonds attractive holdings for well-to-do households. Tax exemption has sometimes been viewed as critical to the success of a municipal bond market. Tax preferences undoubtedly have lowered the cost of capital for U.S. municipalities. However, a strong municipal bond market existed in the United States long before the country even introduced the income tax. Moreover, tax exemption brings some important disadvantages for a bond market. It effectively segments capital supply to the municipal sector by making municipal lending unattractive to many long-term investors for whom municipal bonds would otherwise be logical holdings. Pension funds in the United States are themselves exempt from income taxes; therefore the tax-exempt status of municipal bonds carries no special value. Pension funds are the fastest growing source of capital in the United States and in many other parts of the world, but in the U.S. they do not supply capital to the municipal sector, because of the tax situation. 30 Commercial banks are another large source of capital. However, tax regulations now forbid them from deducting the interest costs of borrowed funds if they also purchase tax-exempt securities. For a long time, commercial banks were the biggest purchasers of U.S. municipal bonds. For the last 15 years, since revision of the tax laws, commercial banks' holdings of municipal securities have steadily declined. Tax exemption, in short, has isolated the U.S. municipal credit market from some of its natural sources of long-term credit supply. One consequence of this separation from the rest of the capital market has been that the net savings to municipalities from tax exemption are less than half the net cost to the federal treasury due to lost tax payments.3 Tax exemption is an economically inefficient form of subsidy. In the U.S. context, however, it provides municipalities with a stable, decentralized subsidy that is not subject to annual appropriations by the federal government nor allocated by federal agencies. This form of decentralized support for local governments fits the U.S. model of fiscal federalism. Tax exemption, however, is not critical to the functioning of a local bond market. In most circumstances, it represents an undesirable preference for the municipal portion of the credit market relative to other sectoral sub-markets. Sources: Petersen (1974); Jacobson (1996); Hillhouse (1936). Choosing a Credit Model: The Advantages of a Mixed System 3.22 Largely by historical accident, most developed countries' municipal credit systems have evolved so that they emphasize either bank lending organized around one or more municipal banks or a municipal bond market. Credit Local de France, the world's largest municipal bank, has been able to rapidly expand its local authority lending into other European countries, which have a history of similar institutional arrangements, but has encountered great difficulty in entering the U.S. credit market, where long-term bank loans to municipalities are considered an oddity. Conversely, until the last decade, few municipalities in Europe issued local bonds. 3.23 Partisans of bank lending and municipal bond markets have debated which system better satisfies municipal financing needs at lower cost. Disagreements have carried over to developing countries, where technical advisors familiar with their own countries' municipal credit structures tend to recommend these structures for adoption in the developing world. 3.24 Developing nations, however, have no reason to opt a priori for one of these end points over the other. Bank lending to municipalities can operate side by side with a municipal bond market. The same municipal financing system and basic legal structure can support both markets, as long as neither system is subsidized preferentially. Competition between the different forms of credit supply is likely to lower the cost of capital to local governments and increase the flow of information about credit quality. 3.25 In most developing countries the banking system now provides the vast majority of credit to the economy. In Asia and Latin America, for example, it is typical for at least 80 percent of all investment financing to be handled by the banking system, despite the rapid growth 3 In recent years, municipal bonds in the United States have had interest rates roughly 15 percent lower than taxable bonds of similar risk. The mean marginal federal income tax rate of municipal bondholders is 31 percent. The difference between these two rates is a deadweight efficiency loss. 31 of stock and bond markets.4 The banking sector's share of finances is even higher in most of Africa and Eastern Europe. Under these financial sector conditions, it is normal for municipal credit supply also to begin in the banking sector-often with a single municipal bank or a municipal development fund that can evolve into a bank. 3.26 A specialized municipal bank, protected by the state from competition, can lay the groundwork for a broader municipal credit market if it screens borrowers' credit quality and collects loan payments as vigorously as any private lender. The European municipal banks helped establish municipalities as highly credit worthy borrowers by imposing exacting loan standards. The history of municipal credit worthiness made it relatively easy to integrate municipal lending into a competitive credit market, once financial sector liberalization took place. As illustrated by the case histories of Colombia and the Czech Republic in the next section, the period of "protected" lending to the municipal sector can be quite brief, as long as national authorities want to encourage market competition in municipal lending and the underlying conditions for a municipal credit market (a stable municipal finance system and efficient financial internediaries) are favorable. 3.27 Once financial markets are de-regulated, bond markets will develop at their own speed through competition. In countries where the banking sector does not do a very good job of screening credit risk, however, it may be advisable to introduce some of the institutions usually associated with bond markets at an earlier stage than the market would do. In particular, independent credit rating agencies can help clarify credit risk. Although they were originally developed to provide credit information to buyers of bonds on the capital market, independent credit ratings can help inform all market participants about credit risk. Within the municipal sector, credit ratings have been used by banks to help assess the risk attached to municipal loans, and even by municipalities that want to negotiate a more favorable lending rate from banks. Many countries-among them Korea, India, Thailand, South Africa, and Chile-have helped create independent rating agencies in order to disseminate objective information about credit risks more efficiently. Most agencies of this kind are linked to one of the international credit- rating firms.5 Colombia and the Czech Republic: Two Examples of the Transition to a Competitive Credit Market 3.28 Colombia and the Czech Republic are two countries where competitive local credit systems have evolved from municipal credit monopolies. Boxes 6 and 7 summarize the history of their transitions. The two countries share fundamental characteristics. Each initially operated under a regulated banking system that essentially restricted municipal loans to a single 4 International Monetary Fund, World Economic Outlook (October 1996). Within the capital market, the market for debt instruments typically has lagged well behind the equity side of the market. In Thailand, for example, the bond market's value at the end of 1995 was 6.8 percent of the stock market's capitalization vs. a comparable U.S. figure of 150 percent. See Euromoney, 1996 Guide to Thailand and OECD, Emerging Bond Markets in the Dynamic Asian Economies (Paris: 1993). S see Credit Ratings in Emerging Markets (London: Financial Times Press, 1996) for a comprehensive discussion of the credit rating agencies now functioning in different countries and their rating procedures. 32 source. With de-regulation, competitive commercial bank lending to municipalities first emerged, specifically encouraged by public policy. Somewhat later, a municipal bond market emerged. 3.29 In both countries competition has produced a segmentation of the municipal credit market. The smallest municipalities borrow primarily from parastatal institutions. Mid-sized cities borrow principally from commercial banks. Large cities finance most of their capital investment needs from municipal bonds. Segmenting the municipal credit market in this way was not planned. It has resulted from a competitive sorting out of efficient credit access. Small municipalities value most highly the technical support that the parastatal lender provides in helping develop investment projects and planning for loan repayment. Mid-sized cities prefer to avoid the bureaucratic delays associated with parastatal lending and do their own planning. Large cities, once they have established a reputation for debt repayment, find it cheaper to go directly to the capital market. 3.30 Both Colombia and the Czech Republic have used their internationally supported Municipal Development Funds strategically to accelerate the development of a self-sustaining private credit market. In Colombia, the sub-national credit institution, FINDETER, from its inception has supplied credit to the municipal sector through commercial banks, in an attempt to engage commercial banks in municipal lending. It operates as a second-level financing institution which re-discounts commercial bank loans to municipalities. Commercial banks accept all of the credit risk of their municipal lending. The banks' successful experience with municipal lending through FINDETER prompted them to enter the municipal credit market with their own resources, and subsequently led other suppliers of capital to enter the market as underwriters and purchasers of municipal bonds. 3.31 In the Czech Republic, as in most of Eastern Europe, a single state savings bank formerly supplied virtually all credit to local authorities. Shortly after the market was opened to competition, MUFIS, or the Municipal Infrastructure Financing Company, was introduced to strengthen competition among commercial banks. MUFIS provides long-term loans to commercial banks for on-lending to municipalities. It can also finance the purchase of municipal bonds. The commercial banks in this case perform all credit analysis and assume all credit risk. They must repay their loans to MUFIS, in full and on schedule, whether or not they are paid by municipal borrowers. Partly because of MUFIS' presence in the market, municipal credit market competition has intensified. Eight commercial banks now are actively engaged in municipal lending. A vigorous domestic and international municipal bond market emerged a few years after bank lending began. Box 6. The Transition to a Competitive Local Credit Market Development: Colombia Colombia now has a three-tier sub-national credit financing system. The Territorial Financing Institution (FINDETER), now an independent parastatal institution owned by government and local authorities, was the original source of credit for municipal borrowing. FINDETER began as an infrastructure financing window within the National Mortgage Bank. It has evolved into a development bank for the municipal sector, working through the commercial banking system. FINDETER reviews local investment projects from a technical, economic, and developmental standpoint. For 33 projects it approves, FINDETER agrees to refinance up to 85 percent of a commercial bank's loan to the municipality to finance the project. The commercial bank performs its own financial analysis and assumes all credit risk, including risk on the portion of the loan refinanced by FINDETER. (The bank remains responsible for servicing its debt to FINDETER regardless of its own repayment experience.) Banks participating in the FINDETER program have enjoyed a good repayment record. Less than 2 percent of loans are classified as non-performing. A voluntary intercept provision strengthens loan security. This requires a municipality to set up a special account into which intergovernmental revenue-sharing payments flow. The lender has first claim on the resources in the account, as long as municipal loan payments are due. All of the financial safeguards built into the FINDETER program have been carried over to direct commercial bank lending. In particular, the voluntary intercept guarantees are used by private commercial lenders and have been extended to other kinds of local revenues that can be sequestered, such as property tax receipts or project specific revenue streams. Initially, commercial banks made municipal loans only with a prior agreement that FINDETER would re-finance a loan. The good credit experience of the banks, however, led them to commit their own resources. Competition within the sector has reduced bank loan margins and lengthened the average lending period for commercial financing of infrastructure projects. Municipal demand for direct bank loans has grown rapidly because, although the lending periods of direct bank loans typically are shorter than FINDETER's long-term loans, the approval process is far faster. Intermediate-sized cities and departments now borrow primarily through commercial bank loans from own resources. Small cities and towns continue to rely heavily on FINDETER. Bond issues are the most recent development in the municipal credit picture. The largest cities now finance their credit requirements primarily through bonds. Bonds are typically issued in serial form, with differing maturities, ranging up to eight years. By the end of 1995, Bogota had issued its eighth series of municipal bonds. Bonds are preferred because they enable the largest cities with sound finances to raise capital less expensively than through commercial borrowing, and because they give city authorities more flexibility in managing the resources received (i.e., the covenants regarding funds' usage are less rigid). The table below summarizes the overall growth in local and departmental credit between 1991 and 1994. FINDETER's shrinking share of the market has been a concern to that institution. However, from a broader market perspective, the reduction in FINDETER's role is an encouraging measure of the integration of municipal lending into the commercial credit market. (Note: the jump in aggregate sub-national credit in 1993 and 1994 reflects the surge in borrowing that resulted from the new intergovernmental revenue-sharing commitments written into Colombia's 1991 constitution. The guaranteed growth in central-government transfers was used as collateral for local borrowing to finance social investment (see discussion in Chapter 2). Municipal and Departmental Debt (Billions of Pesos) Source of Credit 91 992 993 994 FINDETER 15 42 80 78 Commercial Bank Own Resources 939 31 59 24 Municipal Bonds 6 4 66 47 Note: US$ 1.00 = approximately 1,000 Colombian pesos Source: Peterson, Municipal Credit Institutions in Latin America. FINDETER: Its Current Role and Future Mission (1996). 34 Box 7. The Transition to a Competitive Local Credit Market: Czech Republic Municipal borrowing to finance local infrastructure projects is relatively new in the Czech Republic. Under the former regime, almost all local capital investment was financed either by state subsidies or direct state construction. When autonomous local governments were established after the collapse of Communism, a single institution, the former State Savings Bank, became responsibie for virtually all municipal lending. The Czech Parliament debated two models of municipal credit supply. Legislation was introduced to formally establish the Savings Bank as the sole supplier of municipal credit. In the end, however, the municipal credit market was fully de-regulated. Beginning January 1, 1993 a new tax and local government financing system was adopted. It provided Czech municipalities with a variety of stable, shared taxes, and solidified local governments' budgets. A diversified municipal credit system has developed. Small municipalities borrow principally from the State Environmental Fund, at zero percent interest, for targeted environmental projects. Mid-sized cities borrow largely from commercial banks. More than 1,300 commercial bank loans were made between 1992 and 1995. All of the large commercial banks actively lend to the municipal sector. All Czech cities over 100,000 have now issued municipal bonds. The municipal bond market developed quickly on the heels of commercial bank lending. The same banks that make municipal loans typically serve as municipal bond underwriters. Both Prague and Ostrava (the third largest city) have obtained international credit ratings from Standard & Poor. Both cities are now rated 'A' for foreign-currency bonds. Prague was the first city in Central and Eastern Europe to issue a foreign currency bond in the European market (for US $250 million). As can be seen from the table below, all segments of the Czech credit market have grown rapidly. The competitive environment has been supported by the activity of a municipal development fund (the Municipal Infrastructure Finance Company or MUFIS). It was established with the express purpose of enhancing competition among commercial banks in the municipal sector. MUFIS provides commercial banks with long- term loanable funds, at market rates of interest, for on-lending to local governments. The banks perform all municipal credit analysis, and accept all credit risk for their loans. Municipal Credit Market Growth in the Czech Republic Structure of Debt Outstanding (Millions of Czech Crowns) Type of Debt 1993 1994 1995 1996 Commercial Bank Credits 2,500 4,900 8,700 11,600 Domestic Bonds 30 610 1,230 3,300 Foreign Currency Bonds 0 7,290 7,290 8,610 Zero-Interest Loans* 100 1,100 2,800 4,400 Note: Figures are not inflation-adjusted. Inflation rates over the period ranged between 9% and 13% per annum. Approximate exchange rate US$1.00 = 28.0 Czech crowns. With competition, real interest rates in the municipal sector have declined and average lending terms have been extended. Commercial banks now make infrastructure loans of up to 8 years with their own funds and up to 15 years when using intermediary funds. Municipal bonds are issued for five- to ten-year terms. Growth in the credit market has been supported by an exemplary local government repayment record. There have been no defaults on commercial-sector lending; problem loans in the commercial sector represent less than 0.8 percent of all municipal loans. The Czech National Bank classifies municipal loans as the second safest loan category after national government debt. For reserve purposes, municipal loans carry a risk weighting of only 0.2 over national government debt. The table below shows that with development of the municipal credit market, the structure of bank lending to municipalities has shifted dramatically to long-term lending. The growth in overall long-term credit has 35 allowed Czech municipalities to increase the share of investment in municipal budgets, despite large cutbacks in government capital subsidies. The investment share of aggregate municipal budgets in 1995 stood at 38.4 percent. Term Structure of Commercial Bank Credit Type of Credit 1993 1994 1995 1996 Short-Terma 32% 18% 25% 6% Medium-Termb 41% 39% 26% 25% Long-Term' 27% 42% 49% 69% a. Less than one year b. 1-3 years c. 4+ years Source: Peterson, Matouskova, and Tajcman, Monitoring Report: Municipal Infrastructure Financing Program, Czech Republic (1997). IV. MUNICIPAL DEVELOPMENT FUNDS AND THE TRANSITION TO MUNICIPAL CREDIT SYSTEMS What Are Municipal Development Funds? 4.1 Municipal development funds (MDFs) are parastatal intermediaries that provide credit to local governments and to other institutions investing in local infrastructure.' Around the world, more than 50 developing countries have established intermediaries of this kind. In larger, federal nations, such as Brazil and India, individual states often have established their own municipal development funds. The funds have been particularly favored by multi-lateral development banks, which find them a practical way of channeling loans to local governments, without themselves taking on the administrative burden or financial risk of direct lending to local authorities. 4.2 MDFs come in a wide variety of institutional forms. The annex summarizes the institutional arrangements for a sample of MDFs in different parts of the world. All of the MDFs have a substantial degree of governmental control. At one extreme, an MDF may operate as a fund or "loan window" within a government agency. Other Funds have legally independent status. However, in these cases, the institutions generally have their management appointed by the government, have the majority of their shares owned by government, or receive funds for on- lending from government subject to government conditions. Three of the MDFs in the sample have closer connections with the private sector. As noted in Chapter III, MUFIS in the Czech Republic and FINDETER in Colombia conduct their on-lending through private commercial banks which bear the full credit risk of loans they make. A third fund, IDI in India, is a joint venture of government and private capital. The Transition to Self-Sustaining Credit System 4.3 MDFs have dual objectives that can be difficult to reconcile. In the short run, they are charged with disbursing publicly provided credit to local governments and seeing that local investment is efficiently implemented. In the intermediate and longer run, they are supposed to pave the way for a self-sustaining credit system that can access private-market capital. 4.4 In the latter role, MDFs can be viewed as a kind of demonstration project. Their goal is to demonstrate, by practical example, that lending to local governments for infrastructure investment can be good business. In countries that have a sound municipal finance system and active banks or other financial intermediaries (i.e., those in the upper right portion of Figure 2), operating a successful municipal loan business, using donor capital but operating as close as possible to private market terms, may be sufficient to attract private credit suppliers into the municipal credit market. ' A variety of abbreviations have been used to describe these institutions. They are sometimes called Municipal Credit Institutions (MCls) or Municipal Credit Facilities (MCFs). For an early summary, see Davey (1988). 37 4.5 In the majority of cases there will be deeper impediments to private sector lending. An MDF can, through its operations, lay bare these obstacles so that public authorities can address them. Until a system has acquired practical lending experience, it is difficult to know whether the laws regarding collateral foreclosure have to be rewritten to accommodate municipal lending, or whether the kind of financial reporting municipalities make to government also provides an adequate basis for credit judgments. Government is unlikely to fully appreciate the difficulties that sudden or unpredictable cutbacks in revenue sharing can cause for local debt servicing. It is better to learn this lesson during a period of parastatal lending to municipalities than in a period of private sector lending, when a wave of municipal defaults could turn away private credit suppliers form the municipal market for a decade or more. 4.6 Municipal Graduation to Private Sector Borrowing. A municipal development fund seeking to develop the private credit market will see itself as part of a continuum in municipal lending. Its success can be measured by the number of municipalities that "graduate" to private sector borrowing, by having demonstrated their creditworthiness under MDF tutelage. In a dynamically developing credit system, the "best" municipal borrowers will regularly be promoted out of the parastatal support system into the private market, while the MDF continues to reach out to smaller and poorer communities, preparing them for credit market participation. 4.7 Unfortunately, few MDFs see their mission in this light. Once established, MDFs have fought to retain monopoly lending powers in the municipal sector. They have resisted losing good clients to private lenders. They have been slow to surrender control over local project preparation. They have resisted reductions in municipal loan subsidies. On the other side of the market, MDFs have displayed little interest in becoming true financial intermediaries which raise capital on competitive terms in the private capital market. 4.8 In fairness, MDFs have received highly mixed signals from international donors and governments as to what their priorities should be. National governments often want to maintain control over municipal lending and municipal investment through MDF operations. International donors want MDFs to disburse loan funds quickly and to compile a record of both successful investment projects and good loan repayment. These objectives are difficult to meet if the highest quality borrowers continually are outgrowing the MDF. It becomes tempting to keep the entire municipal sector within the parastatal lending framework, and to continue to ensure high demand for MDF loans by subsidizing interest rates or matching market-rate loans with government capital grants. 4.9 This chapter examines the track record of MDFs. It starts from the perspective that MDFs should encourage the development of self-sustaining private-sector municipal credit markets, but that this is not their only objective, and that for many MDFs the transition to private market lending is not achievable in the near term. 4.10 Throughout the chapter, we draw on a special sample of MDFs that were examined in institutional detail for this report (see Annex). To simplify tables, we have limited text comparisons to a representative sub-set of MDFs. These were chosen to reflect the variety of institutional structure and financing methods found in the full sample. 38 Loan Repayments and Municipal Creditworthiness 4.11 Loan repayment lies at the heart of all credit systems. If MDF loans are not repaid, continuing injections of government capital will be required merely to sustain MDF lending and private lenders will see the sector as involving high risks that discourage their entry. The foremost obligation of an MDF in helping build a local credit market is to establish a track record of regular loan repayment. 4.12 It would not be surprising if, in the start-up phase of municipal lending, problem loan rates ran fairly high. In fact, one of the primary reasons for having a state-sponsored development bank lead the way in sectoral lending is to clarify the credit risks that exist and to help bring these down to acceptable levels. Depending upon what is discovered, improved creditworthiness may require changes in local financial management, changes in the intergovernmental financing structure, or changes in the way project loans are appraised both by borrowers and the MDF lender. 4.13 Table 1 summarizes the loan repayment experience of our representative group of developing-country MDFs. Repayment rates can be seen to vary greatly across funds-ranging from zero problem loans in some systems to systems that have had as high as 80 or 90 percent non-performing loans.2 On balance, Table 1 shows that unsecured municipal lending involves a good deal of financial risk. Most of the municipal development funds enjoying high repayment rates benefit from special types of loan guarantees. Loans backed only by municipalities' "general obligation" pledge to repay debt from local revenues have proved risky. Non- performance rates run well above levels that would be commercially viable, unless lenders were compensated by steep risk premiums. More alarmingly, there is little indication within individual funds that the frequency of problem loans declines over time. The usual pattern has been for problem loan rates to increase as a system matures. 2 The Calcutta Municipal Development Fund appears to hoid the non-payment record. More than 90 percent of local loans eventually defaulted. The State of West Bengal which operated the fund not only failed to press for collections but objected to the principle of municipal lending for capital investment. The program had been initiated by a predecessor government of a different party. 39 Table 1. Loan Repayment Experience: Selected Municipal Development Funds Loan Non- Automatic Country Performance Intercept Remarks and MDF Rate Guarantee Brazil, State of 2% Yes Loans are collateralized by VAT tax-sharing transfers from Parana state to local government. The state bank deducts debt service PrAM/PIMES payments directly from municipalities' accounts. Colombia 2% Voluntary in Municipalities open special account to deposit inter- FINDETER loan contract govermnental tax-sharing transfers. Bank has right to access account. Municipal loans are made and payments collected by private commercial banks. Czech Republic 0 No Municipal property commonly used as collateral. Loans are MUFIS (only in oper- made and payments collected by private commercial banks. ation 2 years) Ecuador 5% Yes Local government loans are collateralized by local BEDE government accounts at central bank. These receive all government transfers as well as all local govemment own- source revenues. Repayment of BEDE loans has first claim. Honduras 50%+ No No collateralization. BANMA does have the right to BANMA intervene in local tax collections if loans are delinquent. It can establish a local office and use incremental tax receipts to repay its loans. In past, BANMA continued to make loans to local authorities in arrears. Performance now improving. Indonesia 30% No No collateralization or other guarantee structure. RDA Jordan 30% Yes Collateralized by government transfers. However, bad loans CVDB joined with a falling economy that reduced transfers produced a situation where delinquent loan payments for many municipalities exceeded central govemment transfers and the guarantee system broke down. Performance now improving. Kenya 80%+ No LGLA continues to make new loans to municipalities in LGLA arrears. Forms part of an interlocking bad-debt situation, where state also fails to pay amounts legally due to municipalities. Morocco 0% commune No Local budgets must provide for loan repayments before they FEC will be approved by central govemment. Central government 20%+ regies implicitly guarantees commune loans through official autonomes approval of municipal budgets. Debt payments are deducted at source by the Ministry of Finance. Municipal enterprises (regies) have not participated in central govemment budget review and implicit guarantee. Rising arrears led to requirements that regie loans be guaranteed by local municipality. Philippines 20% No MDF has the legal right to impose intercepts, but in past did MDF not exercise this right. System is now being reformed. 40 4.14 Two aspects of MDF lending appear to be especially problematic. One is the custom of providing initial grace periods on payment of principal and occasionally on both principal and interest. These characteristics of MDF loans tend to reflect parallel conditions in the international loans that the MDFs use for financing. Whatever their advantages and disadvantages in sovereign international loans, extrapolation of the same characteristics to municipal lending seems inappropriate. A case can be made that a development bank's loans to finance revenue-generating projects should not require repayment until the project is in operation and revenue starts to flow. The construction period, however, should be far shorter than the normal loan grace period of five years. Grace periods for loans supported by general municipal revenues are still less persuasive. They lure municipalities into underestimating the costs of debt servicing. When the need to start repaying principal kicks in, debt service ratios often rise precipitously. Moreover, a lengthy grace period creates the unhappy possibility that the political administration borrowing funds never has to face up to the full costs of borrowing. These are bequeathed to its successor. 4.15 MDFs also can introduce new, political risks into the municipal credit system, making it more difficult for market-based lending to begin. In some of the MDF systems, non- payment of debt service carries no significant penalty. Municipalities in default can continue to receive grants from central government and frequently continue to receive new loans from the MDF, even though outstanding loans are not being repaid. Municipalities faced with tight budgets may fall into the habit of negotiating with central authorities over deb-service waivers or debt re-scheduling. A new local administration may argue that a previous administration, controlled by a different party, undertook the original borrowing, and that the current administration should not be obliged to repay debt incurred for a project it did not favor. 4.16 The fact that MDFs typically are owned and operated by governmental institutions makes it almost impossible for them to deal with municipalities in a commercial manner. Most MDFs have control over discretionary subsidy funds that they can use to match municipalities' market-rate borrowing or to write down interest rates. Their combined control over subsidies and credit invites borrowers to try to obtain a larger subsidy mix, either before the fact by negotiating more favorable financing terms or after the fact by pleading poverty and not making full loan repayment. In short, MDF lending practices can exacerbate municipal credit risk by creating political risk that manifests itself as unwillingness to pay. 4.17 Two of the MDFs with the best repayment records-MUFIS of the Czech Republic and FINDETER of Colombia-are second-tier financing institutions which lend to municipalities through private commercial banks. The private banks assess municipal creditworthiness, assume credit risk, and collect loan payments using standard commercial methods. These practices have produced very good loan repayment records. However, the use of 41 commercial banks as final lenders is not feasible until the municipal sector has demonstrated that credit risks are controllable.3 Attempts to Reduce Credit Risk: Loan Ceilings 4.18 The erratic repayment record of municipal development funds has led governments and external donors to look for structural features that can enhance the creditworthiness of municipal lending. 4.19 One common approach has been to establish numerical guidelines that would automatically identify prudent borrowing ceilings, based on a municipality's capacity for debt repayment. The guidelines are often transferred from municipal credit analysis standards applied in developed countries. For example, several municipal development funds in Latin America, including the Urban Development Fund of the State of Rio Grande do Sul, Brazil, and the Municipal Development Program administered by the Ecuadorian Development Bank, limit total debt service (including debt service on MDF loans) to 15 percent of a municipality's revenues, defined to exclude one-time or discretionary grants from the State. A similar standard has been written into law for all municipal borrowing in Poland. The 15 percent threshold for debt service traditionally has been a basic guideline for determining prudent levels of municipal indebtedness in the U.S. municipal bond market. 4.20 The Cities and Villages Development Bank (CVDB) of Jordan and the Communal Infrastructure Fund of Morocco (FEC) have adopted the basic credit assessment procedures used in France. These project a municipality's budget surplus from existing operations, or its "~management surplus," then limit new municipal debt to levels that can be serviced within this surplus.4 4.21 Numerical ceilings on borrowing can help MDFs resist municipal entreaties for unrealistically large loans. However, ceilings of this kind have been only moderately effective in improving overall creditworthiness. Empirical studies in the United States have found that state debt ceilings exert a modest positive effect on credit ratings and result in modestly lower 3 During the late 1980s, the World Bank attempted to implement a municipal on-lending program through commercial banks in Nigeria. Even though the banks would have been at risk for only 20 percent of financing, and would have benefited from a substantial float of scarce foreign exchange, they declined to participate in the program. They felt there was no way for them to control for political risk in the intergovemmental financing system. In 1994, Credit Local de France similarly concluded that municipal financial conditions in Hungary were too unstable to justify its committing capital to development of a municipal bank. 4For example, Jordan's Cities and Villages Development Bank calculates the capacity to service additional debt by projecting a municipality's current revenues from transfers, taxes and fees, then subtracting projected expenditures on recurrent operating costs, service on existing debt and capital investment financed from recurrent revenues. The difference is the management surplus. Local authorities are permitted to borrow up to the point that would absorb 75 percent of this management surplus. 42 municipal default rates.5 However, the rules of thumb that make sense in assessing municipal borrowing capacity in developed countries often have less relevance in assessing credit risk in developing countries. The largest component of local authorities' budgets in the developing world is intergovernmental revenue sharing or central government transfers. When these are subject to drastic year-to-year changes because of annual budgetary calculations, projected debt service ratios which are based on today's level of intergovernmental receipts may be meaningless. 4.22 Eagerness to disburse loan funds, especially in projects financed by international donor agencies, can further erode the utility of credit limits or advance credit analysis. Jordan's CVDB, for example, adopted credit limits based on future, projected income, after a World Bank report found that the previous system of credit ceilings resulted in slow disbursement of donor funds. However, CVDB's projections of municipalities' future income, especially the revenue to be generated from new municipal investment projects, proved highly optimistic, leading to overstatements of future municipal income and hence municipal capacity to service debt. The CVDB eventually had to waive debt service payments for two years for all municipalities and restructure most of its outstanding loans.6 Guarantees 4.23 The inability to fully protect debt repayment through borrowing ceilings has prompted lenders, including municipal development funds, to look for other guarantees that would protect their loans. 4.24 The first preference of lenders usually is a central government guarantee. Occasionally, an explicit, open-ended governmental guarantee has been provided. The first municipal bonds issued in Hungary after the collapse of the Communist government, for example, were supported by an explicit national government guarantee. More often, the government guarantee is implicit. Under the French municipal finance system in effect until the mid-1980s, the central government reviewed and approved municipal budgets, including plans for municipal borrowing. It was thought that this system carried with it a government commitment to guarantee municipal debt repayments, through special funding assistance and management intervention in the event of threatened default, as had been the practice for many years. However, with the fiscal decentralization of the 1980s, the central government guarantee disappeared. A large volume of outstanding municipal loans had to be restructured when municipalities encountered difficulty in repayment. This experience has raised questions in some North African countries, whose MDFs are modeled after the original version of the French municipal credit system, as to whether the implicit state guarantee for municipal borrowing, Capeci, "Local Fiscal Policies, Default Risk and Municipal Borrowing Costs," Brandeis University, Department of Economics, No. 259 (1990); Goldstein and Woglum, "Market-Based Fiscal Discipline in Monetary Unions; "Evidence from the V.]. Municipal Bond Market" (IMF, 1991). 6 World Bank, the Hashemite Kingdom of Jordan Intergovernmental Fiscal Relations and Municipal Financial Management Sector Study (May 1995); William Dillinger, "Jordan Cities and Villages Development Bank," (World Bank, 1990). 43 based on similar state review and approval of local budgets, will survive decentralization initiatives. 4.25 Perhaps the most effective form of municipal loan guarantee has been the intercept provision. These give the lender first claim on intergovernmental transfers otherwise due to the municipality. Intercepts of this kind can be found in a number of Western countries. The majority of states in the United States now have intercept laws, which allow distressed central city governments or distressed school districts to assign lenders the first claim on state aid whenever a municipality or school district is not current in its debt payments. This form of guarantee has opened the credit market to jurisdictions that would not otherwise have access to it, and has lowered their interest costs. In the Netherlands, borrowing through the Municipal Bank of the Netherlands is secured by a common Municipal Fund, through which central government aid is distributed to local authorities. The Municipal Fund in effect guarantees each municipality's debt repayment, reducing the amount of funding available for distribution as grant assistance in the event that a loan guarantee has to be utilized. 4.26 In developing countries, intercept provisions have been generally successful in reducing municipal arrears. Table 1 shows that most of the countries with intercept arrangements have arrearage rates of 2 percent or less. Loan repayment rates remain high even when lenders, as in Colombia, cannot access central grant funds directly, but must enter into voluntary contractual arrangements with municipal borrowers to have grant transfers deposited into a special local account which can be drawn upon by the lender in the event that debt service payments are not made in full and on schedule. The importance of intercepts in securing loan repayment has led several MDFs to incorporate coverage requirements into their loan authorization procedures. The various state MDFs in Brazil, for example, where municipal loans are secured by intercepts of state-collected VAT revenues, contain provisions requiring that a municipality must have sufficient VAT transfer entitlements to collateralize all debt service. 4.27 Although intercepts have been effective in reducing municipal loan arrears, they are not without problems. An intercept arrangement does not overcome the risk that a state or central government will drastically reduce the intergovernmental transfer that serves as loan guarantee. The loans made by Jordan's CVDB were secured by intercept authority. However, when the economy turned down and intergovernmental transfers were cut, these proved insufficient to fully cover debt service obligations. The risk that central authorities will reduce local tax-sharing receipts or grant transfers can be protected against only by other types of measures, such as a constitutional provision stipulating local transfer entitlements,' or a track record of stable transfer policy. 4.28 A more fundamental shortcoming of intercept provisions is that they frequently fly in the face of decentralization initiatives intended to place more responsibility for budget choices on local governments. Intercept arrangements can become mere book transfers at the central level. This is especially true when, as often is the case, the same Municipal Development In Brazil and Colombia, as well as several other Latin American countries, local entitlements to specific shares of national tax collections are now spelled out in the Constitution. 44 Fund that makes loans to local authorities is responsible for allocating grants and handling local authorities' financial accounts, or is linked to these agencies. In these cases, the MDF may merely credit itself with the loan payment due, subtract the amount from the revenue sharing or grant payment to a municipality, and send the municipality a check for the difference. Local authorities may not even be aware that this process has taken place. Bookkeeping arrangements of this kind, though effective in preserving the MDF's financial position, defeat the purpose of making local governments aware of the costs of investment and having them take into account these costs in deciding the magnitude and type of investment they will finance. 4.29 Other kinds of guarantees have been offered to protect municipal loans. In Central and Eastern Europe, it is customary to collateralize municipal loans with real property owned by the municipal government. This practice reflects both the extent of commercial property holdings by local governments and the reliance on real property as collateral throughout the banking system. Lenders in the Central and Eastern Europe region as well as elsewhere recently have moved toward more liquid forms of collateral. Lenders may require that municipal borrowers maintain their ordinary business accounts with the lending authority and grant the agency the right to automatically debit any loan payments due. These practices most commonly are found among commercial banks or municipal development funds that operate through commercial banks. Liquid collateral of this kind has proved extremely effective in reducing loan defaults and is consistent with the principle of fiscal decentralization, since it requires a deliberate choice on the part of a municipality to assign to a lender the right to debit its current account. The "price" of loan repayments is immediately visible in the municipal balance. MDF Lending: Development Bank Functions 4.30 Most MDFs are either literally development banks or are institutions charged with similar functions. Their goals extend beyond credit supply to increasing the efficiency of local investment, promoting the development of service pricing and cost recovery in local investment projects, upgrading municipal financial management capacity, and ensuring that critical investment projects get built. MDFs for the most part have addressed this range of goals by trying to bring a host of different developmental functions under their own roofs. Table 2 shows that the typical MDF, in addition to lending funds to local authorities, performs project appraisal from both a financial and economic perspective, blends loan funds with capital grants or subsidies for high-priority projects, oversees local project preparation and construction, provides technical assistance in financial management, and collects loan repayments, among many other responsibilities. 4.31 The rationale for bundling together these different functions is straightforward. In principle, it allows an MDF to coordinate all of the activities critical to the success of an investment project. Without coordination, local authorities may never make use of loan funds because they are hoping to become eligible for a capital grant instead, or a project may get financed but never be completed because of inadequate local implementation capacity or poor technical design. As the pre-eminent institution dealing with local government investment, an MDF can assemble the critical mass of technical expertise necessary both to improve local 45 investment practices and serve as a comprehensive counterpart for international technical assistance. 4.32 Nonetheless, there are dangers in internalizing all of the functions associated with local government investment financing in a single institution. One danger is bureaucratic delay. In performing the various roles assigned to it, an MDF may become a cumbersome bottleneck that slows down the investment process. In the early 1990s, for example, the average elapsed time between initial project identification and municipal loan approval at the Ecuadorian municipal development fund was more than 32 months. FINDETER in Colombia has been able to reduce the period between initial loan application and loan approval to a little more than eight months, but municipalities nonetheless have been shifting loan demand to commercial banks, citing the delay in FINDETER project review as the principal shortcoming of its lending process. 4.33 A second danger is that a development bank empowered to perform the full range of functions illustrated in Table 2 will act as a monopolist, squeezing out competition and retarding development both of a private credit market and of local governments' capacity to plan and execute projects on their own. If the municipal development bank, for example, has the ability to blend subsidy finds with its loans, it will tend to underprice pure commercial lenders, forcing them out of the market. If municipal authorities must utilize MDF loans in order to gain access to technical assistance, other institutions will find it difficult to compete for loan business. A strong MDF can likewise stand in the way of decentralization initiatives. If the MDF must approve project design and project service pricing arrangements, for example, in order for a project to obtain financing, it will be tempting for local authorities simply to look to the MDF for technical guidance in the first place. Many MDFs maintain a roster of approved technical consultants, who view their true client as the MDF, not local governments. Table 2. Functions Performed by MDFs Economic & Credit Financial Financial Capital Assessment Country and Appraisal of Construction Technical Subsidy & Payment Post-Project MDF Projects Oversight Assistance Allocation Collections Monitoring Brazil No, for small Little Yes Yes Yes No PrAM/PIMES projects Colombia Yes Yes Yes No currently. No No FFNDETER Yes, in past. Czech No No No No No No Republic MUFIS Ecuador Yes Yes Yes Yes Yes No BEDE Honduras Yes Yes Some Loan Yes No BANMA Forgiveness Indonesia Yes No Yes Some Yes No 46 RDA Jordan Yes Yes Yes Yes Yes No CVDB Kenya Yes Yes Yes Yes Yes No LGLA Morocco Yes Yes Yes Yes Yes No FEC Philippines Yes Yes Yes Yes Yes No MDF Unbundling Options 4.34 The development functions that a multi-purpose MDF performs are critical to improving the quality and cost-efficiency of local investment. Whether these functions should be centralized in a single institution, however, is an open question. An alternative path is to unbundle the functions as much as possible, so that some of them can be assumed by the private sector, once it becomes equipped to do so, and others can be performed by municipalities themselves. This strategy introduces competition into local investment design and financing, while keeping the MDF in a pivotal coordinating role. 4.35 Assigning Responsibility for Credit Analysis, Credit Risk, and Payment Collections to Commercial Banks or Other Specialized Private-Sector Firms. Commercial banks and other commercial lending institutions routinely conduct credit analyses and perform loan operations. In countries where private-sector banks already are functioning, such responsibilities may be more efficiently carried out by commercial banks than by the MDF directly. Not only is commercial bank participation likely to boost municipal loan repayment rates and reduce MDF risk exposure, but commercial banks' participation in municipal lending as an MDF partner can hasten their independent entry into the market and clear the way for other private-sector institutions to lend to the municipal sector. 4.36 Commercial lenders will be willing to assume municipal credit risk only if the risks can be well defined and limited. The same array of guarantees that have been effective in reducing MDF payment problems should be available to the commercial sector. At present, intercept provisions often are available only to state lending institutions. In Brazil, the Philippines and other countries, the law historically did not permit private lenders to intercept state tax-sharing payments, even if a borrowing municipality voluntarily sought to enter into such an agreement. Only the public MDF could use intercepts. Limitations of this kind impede the spread of municipal lending beyond a state-sponsored facility. One of the principal reasons for the success of FINDETER in spreading its municipal lending model into the commercial sector has been the ability of banks and municipalities to enter into exactly the same type of voluntary intercept agreements that have been included in FINDETER loan contracts. 4.37 In a number of countries with active capital markets, including India, Thailand, Chile, Colombia, and Eastern Europe, independent credit rating firms now exist. These firns can 47 be introduced to the municipal credit market by MDFs. Experience reveals that credit assessment is a function that MDFs generally perform poorly. They do not have highly trained technical staffs and often are restricted by political factors in awarding loans.' MDF reliance on independent assessment of municipal loan risk can both improve repayment experience and introduce to the municipal credit market firms that will play a crucial role in allocating credit in a fully developed, private market. 4.38 Separating Subsidies from Loans. Linking state capital subsidies exclusively to MDF loans is one of the surest ways to deter expansion of the private credit market. Unbundling the subsidy function need not reduce the magnitude of subsidies. Rather, the subsidy decision can be made by a different government agency or agencies, based on the characteristics of the project (e.g., its external benefits or importance to national priorities) or the characteristics of the local government seeking financial assistance (e.g., the borrower's financial capacity). In an unbundled system, interest-rate subsidies and matching grants are awarded according to transparent criteria, independently of the source of loan financing. They are not tied exclusively to MDF loans. 4.39 The evolution of the private credit market in Poland illustrates the potential advantages of unbundling and targeting subsidies. In Poland, most municipal environmental projects have been financed in their entirety through the Environmental Bank and regional Environmental Funds. Initially, the criteria for defining an "environmental" project, one eligible for subsidized lending from the Enviromnental Bank, were wholly vague. This vagueness delayed the introduction of a commercial credit market. Municipalities preferred to queue for the possibility of receiving subsidized loans, rather than borrow at market rates from private lenders. Gradually, however, the Environmental Bank and Environmental Funds clarified and limited their project eligibility criteria. As soon as they did so, a private credit market sprang up. The first projects to be ruled ineligible for subsidized environmental lending were municipal investments in new rolling stock and in road construction and road repaving. (These projects had initially been described as potentially eligible for environmental loans on the grounds that they would reduce fuel consumption and vehicle emissions, or reduce dust pollution.) Municipalities immediately shifted their credit demand for financing these projects to commercial banks and municipal bonds. A vigorous commercial market in both kinds of municipal lending emerged in less than two years. 4.40 In most countries, the subsidy share of overall MDF financing has fallen in recent years. Where capital subsidies can be matched only with MDF loans, however, subsidy allocations continue to deter market entry by private-sector lenders. 4.41 Technical Assistance for Project Preparation. One of the tightest controls that MDFs typically exert is over local government project preparation. The MDF not only must approve local project design, in order for the project to receive MDF credit financing, but 8 In the Czech Republic in 1996, for example, the State Environmental Fund made almost as many municipal loans for only a moderately less total loan volume than the Savings Bank, the largest private-sector lender. The Environmental Fund had six municipal credit analysts; the Savings Bank had 52 municipal credit analysts. 48 typically the MDF oversees all aspects of project design and development. The MDF is likely to maintain a list of qualified technical consultants; local authorities must hire their consultants from this list. International support for MDFs often involves setting aside small grants or subsidized loans for project development and pre-feasibility studies. Municipalities that make use of this project development assistance normally must agree to borrow from the MDF to finance the approved project. All of these mechanisms are intended to stimulate the flow of eligible project financing. However, they also have the effect of squeezing out commercial loans, which do not have access to the same sources of subsidized project development finance. 4.42 In the case of small municipalities, the argument can be advanced that MDF supervision is necessary to ensure that local projects are efficiently designed. Even in this case, however, a strategy for gradual unbundling is appropriate. Small-scale projects can be a proving ground for decentralized project preparation by local governments. The technical benefits gained by having MDF staff review and approve small-scale individual projects, like neighborhood water distribution projects or road surfacing, are likely to be outweighed by the time savings gained from allowing local authorities to move ahead on their own. The experience of taking full responsibility for project preparation is an important ingredient of local capacity building. 4.43 For municipalities of any size, separating project preparation from credit financing furthers the conditions for a competitive credit market. Municipalities that have prepared their own project specifications can "shop around" for the best financing terms. If the project design must be prepared, or reviewed and approved, by a professional team associated with the MDF, it becomes difficult to seek financing from any other source. The bundling of credit and technical assistance for project preparation deters the competitive development of both markets, that in technical consulting for local governments and that in credit financing. Raising Capital: The Intermediation Role 4.44 One of the functions of an intermediary financial institution is to raise "loanable funds" from savers or from the capital market. Tables 3 and 4 summarize how developing- country MDFs have raised loan funds. As can be seen, MDF experience in tapping the overall capital market is quite limited. This contrasts with the experience of municipal banks in Western Europe (see Chapter II) and of Environmental Revolving Loan Funds in the United States, both of which have become efficient vehicles for mobilizing funds from the private capital market for urban and local environmental investment. Table 3. MDF Access to Capital Markets: Sources of Funds Public or Government Regulated Competitive Supports Parallel and/or Donor Municipal Institutional Capital Market Private Country and Funds Deposits Savings Municipal Credit MDF Market Brazil Yes No No No No PrAM/PIMES 49 Colombia Yes No Yes Yes (small) Yes FINDETER Czech Republic Yes No No No Yes MUFIS Ecuador Yes Yes No No BEDE Honduras Yes Yes Yes No No BANMA Indonesia Yes No No No No RDA Jordan Yes Yes Yes Partial No CVDB (see text) Kenya Yes Yes No No No LGLA Morocco Yes Yes Yes Partial No FEC (see text) Philippines Yes Yes No No No MDF 4.45 MDFs typically draw on a variety of sheltered sources for their loanable funds (see Table 4). Almost all benefit from significant capital contributions from government. These may take the form of equity share purchases by Government ministries which become the legal owners of the MDF. Alternatively, Government may contribute zero-cost capital to the MDF's loan fund. Often, Government will transfer the proceeds from multi-lateral or bilateral loans to the MDF at no cost to the MDF. Under these arrangements, the Government borrows from an international institution like the Inter-American Development Bank or World Bank, then repays the external loan from general revenues. The Government in effect contributes to loan financing to the MDF at zero cost. The MDF builds up its permanent capital base by collecting on the initial round of loans it makes to municipalities. 4.46 International agency loans constitute a large share of MDFs' on-lending pool. In more market-oriented systems, the MDF is responsible for servicing the external loan from its earnings on municipal lending. The margin between the MDF's cost of capital and its municipal lending rate then must cover the institution's administrative costs as well as municipal default risk and currency devaluation. None of the MDFs we examined operate on a pure market model of this kind. However, several approximate this arrangement, especially for recent increments in international financing. 4.47 Aside from their two major sources of financing, MDFs tap numerous types of below-market funds. These may be considered initial efforts at financial sector intermediation, although the MDFs typically function within markets sheltered by government regulation. Some MDFs serve as a kind of municipal credit union. Local authorities are required by law to 50 maintain their liquid deposits in the MDF; these are used by the MDF to finance loans to other municipalities. 4.48 MDFs' efforts to access the broader capital market almost always have involved public savings or savings subject to public regulation. Experience with this kind of financial mobilization has been mixed, but includes some prominent failures. The National Mortgage Bank of Brazil, which at one point financed a large part of the local infrastructure investment associated with housing construction in Brazil, borrowed heavily from the national social security system, then went bankrupt when it was unable to collect on its loans or maintain loan values in the face of extreme rates of inflation. At the time, this was the largest failure of a public financial institution in the developing world. BANMA of Honduras sold bonds to the national teachers' public pension fund, which purchased them at government instruction. It became delinquent in repayment. BANOBRAS, the local public works financing institution of Mexico, sold its bonds to state banks, which purchased them at govermment instruction. In Colombia, the MDF initially sold a special class of bonds to the banks (largely public banks) that participated in its municipal on-lending program. The bonds were preferentially treated by the Central Bank, which allowed them to count against reserve requirements. In India, the national Government has steered the savings of public-sector financial institutions into municipal loan funds. 4.49 Only a few institutions have sought to raise funds in the competitive capital market, and then only under special conditions. FINDETER attempted to sell $50 million of bonds in the competitive market-a very small share of its financing base. However, it succeeded in placing only $10 million, or 20 percent, of the bond issue. The commercial banks to which the bonds were marketed saw FINDETER as a competitor in the municipal credit market; they preferred to lend directly to local governments rather than finance a parastatal intermediary. The Country and Villages Development Bank of Jordan competes in the open market with other financial institutions for savings deposits from liquid governmental institutions, like the Social Security Institute. FEC of Morocco follows the pre-reform model of local government financing in France. It belongs to the government-owned Caisse de Dep6ts et Gesti6n, which now raises a substantial part of its funds through market-rate bond issues. The proceeds from these bonds are shared with FEC at the cost of capital. 51 Table 4. Sources of Funds: Selected MDFs Country and MDF Brazil: World Bank Loan: 48% PRAM (1991) State of Parana Capital Contribution: 52% Initially, State funds were used for complementary matching grants, not loan financing. Later, the State contribution was used to help capitalize the loan fund. Colombia: FINDETER (1994) Donor Loans and Grants: 37% Domestic Loans and Bonds: 8% Government Capital Contribution: 49% Retained Earnings: 5% Ecuador: World Bank & Inter-American Development Bank loans: 73% BEDE (projected 1993) Govemment of Ecuador: 27% Municipalities: 0.2% Government funds used for complementary grants Honduras: Municipal equity: 14% BANMA (1991) Municipal deposits: 17% Donor loans: 55% Government capital contribution: 12% Other: 2% Jordan: Local government deposits: 20% CVDB (1991) Central bank advances: 21% Retained earnings: 15% Government capital contribution: 13% Donor loans: 12% Time deposits, govt. institutions: 11% Other: 8% Kenya Donor grants and loans: 78% LGLA (1990) Retained earnings: 17% Other, including local deposits: 5% Morocco: Up to 1988, capital came primarily from concessional lending from the state-owned FEC Caisse de Depots et de Gesti6n (CDG) and Central Bank rediscounting of FEC loans. Later, CDG issued market-rate bonds and on-lent funds to FEC at its cost of capital. Obstacles to Capital Raising 4.50 Why has it proved difficult for MDFs to mobilize private sector funds? Where municipal repayment histories are poor, the underlying income stream is too weak to allow MDFs to be seen as creditworthy, based on their municipal lending operations. A government guarantee is then needed to make MDF borrowing creditworthy. Governments willing to provide such guarantees on behalf of an MDF have found it more convenient to borrow from the multi- lateral development banks or through donor agency programs than to encourage the MDF to go directly to the competitive capital market with their support. 52 4.51 In Western Europe, the most common transitional strategy for introducing parastatal municipal banks to the competitive capital market has been to require the banks to begin raising funds through domestic bond issues, but with a state guarantee. Over time, the extent of the state guarantee is reduced and eventually withdrawn altogether. Of course, the bank's underlying municipal loan repayment stream must be built up at the same time, so that the bank becomes capable of repaying debts from its own resources. 4.52 In the United States, the State Revolving Funds used for subsidized environmental lending have been designed from the outset to leverage financing in the private capital market. Government equity contributions, for example, are commonly used to establish Debt Service Reserve Funds. The institutions typically maintain debt service reserves equal to two years' debt service obligations. This structure reassures the marketplace that external debt obligations can be met, even if the municipal loan repayment stream should falter. The Reserve Fund can be further strengthened by a State commitment to replenish the fund whenever it needs to be topped for debt servicing. The credit ratings of State Revolving Funds of State Bond Banks, and thus their cost of capital, closely reflect the reserve structure that is set up, as well as the moral obligation pledge of the State to replenish the reserve fund form its own revenues when necessary.' 4.53 Municipal funds and municipal banks structured in this way can have excellent access to domestic and international credit markets. Credit Local de France and the Municipal Bank of the Netherlands, for example, enjoy AAA bond ratings and are among the most active issuers of bonds on worldwide capital markets. Most states in the United States sponsor bond banks or Environmental Revolving Loan Funds, or both, which operate on similar principles and are very active and successful borrowers in the U.S. capital market. 4.54 None of the developing-country MDFs examined in this study has been designed with access to the private capital market in mind. Many of the MDFs enjoy substantial levels of government capital contributions, but these tend to be used to pay for the Fund's administrative expenses, or to subsidize municipal lending rates. They are not set aside in a form that reduces the risk to bond purchasers or institutional lenders to the MDF. Moreover, the moral obligation of the State usually is vague. Whereas a national Government usually provides multi-lateral and bilateral lenders with explicit and highly structured guarantees, we found no cases where Government provided comparably well-defined guarantees or even well-defined reserve funds to lessen the risk of competitive market financing by MDFs. 4.55 Surprisingly, multi-lateral lenders appear not to have placed a high priority on having MDFs develop into self-sustaining financial institutions. Multi-lateral banks have not required that MDFs raise capital in the private market as part of their development as financial intermediaries, nor have they urged structural changes that would facilitate capital market access. 9 Standard & Poor's, "Criteria for Rating State Revolving Funds," Credit Week (January 1992). 53 Summary: The Transition to a Market-Based System of Municipal Credit 4.56 Two types of models can be pursued in expanding municipal development funds into self-sustaining, market-based systems of municipal credit. In one model, the MDF gradually sheds its state-protected status, and becomes an instrument for gathering capital competitively on the open market while continuing to on-lend to municipalities. This is the model that has been followed in most of Western Europe through the development of specialized municipal banks. 4.57 In the second model, most of the growth in municipal credit supply occurs outside the MDF, through the entry of private sector institutions into the municipal credit market. The MDF may be designed to fade away as the private sector gains capacity to service municipal credit needs. Or the MDF may continue to play an important but limited role, by meeting the credit needs of certain types of local governments (e.g., smaller municipalities) and by serving as a benchmark for private-sector competition. 4.58 Whichever model is followed, an explicit strategy for MDF support of broader credit market development is needed. Without such a strategy, MDFs tend to persist in their original form, retaining monopolistic powers and retarding development of a competitive municipal credit market. V. STRATEGIC CHOICES IN CREDIT SYSTEM DESIGN 5.1 A review of experience in developing municipal credit markets should have usefulness for future policy design. What lessons are there to be learned about how national governments or international donors can more effectively build these markets? Some conclusions do seem possible, even if the strategy for implementing them has to be adjusted to fit the history and institutions of each country. A Set of (Almost) Always Applicable Reforms 5.2 A basic set of easily understood principles underlies a sound municipal credit market. Reforms to put them in place will bring greater clarity to municipal finances, increase the number of creditworthy investment projects, and make it easier for would-be suppliers of credit to enter the municipal market, by allowing them to distinguish more quickly and more reliably between good risks and poor risks. Most of these basic reforns can be initiated by countries still in the lower left quadrant of Figure 2 (page 23), even if municipal lending is but a future prospect. At the same time, few countries in the developing world have satisfactorily established all of the pre-conditions for an effective municipal credit market. The items therefore can serve as a check list for countries with more substantial municipal lending experience. Are there parts of the municipal credit market's permanent foundation that are still not in place, even though a superstructure has been erected and is functioning? 5.3 The list of straightforward but critical reforms begins with businesslike collection of local taxes and user charges. Local authorities almost always have more power on the revenue side of their budgets than they choose to exercise. The city of Ahmedabad, India, for example, moved from serious deficit to surplus and the ability to enter the long-term credit market for the first time, within two years, based on a new city manager's determination to tighten octroi collection. Other municipalities that wish to start borrowing to finance investment usually have unrealized revenue potential. A serious attitude toward collecting local fees and taxes probably is the first prerequisite both of true decentralization and municipal creditworthiness. 5.4 A second class of reforms involves accounting, auditing, and disclosure of municipal financial condition. National policy should establish a clear, comprehensible set of municipal accounts applied uniformly across the country. These local government accounts should be independently audited. The results should be reported to the financial community and to the public at large. This is a very different approach from that now found in many countries, where municipal accounts are audited, if at all, by internal auditors and, though reported to central government, are not released to the public. Without reliable financial data, meaningful credit analysis cannot take place, and municipalities cannot hope to access private capital, except with government guarantees. 5.5 A next step is for a municipality to prepare a several-year rolling capital investment and infrastructure maintenance plan, together with a separate capital budget showing how its priority projects will be financed. This plan does not have to be complicated-a 55 few pages describing key projects and how they will be paid for may suffice for many towns. Only projects that are ready to go should be placed in the current year's capital budget. For each capital project, there is a financing plan, which indicates how the project will be paid for, including the use to be made of credit. Each year the capital plan is "rolled over" by adding a new year to the planning horizon. Each year's capital budget should include a comparison of last year's actual spending with planned spending, gradually building pressure for more realistic capital budgeting based on the municipality's own capacity to produce operating savings and borrow, not merely to capture grant funds from the central government. 5.6 Central government can help establish a more positive environment for municipal credit by stabilizing the system of intergovernmental transfers and shared taxes. In most countries that are just establishing municipal credit markets, these flows form the largest part of local governments' capacity to repay debt. "Stability" of intergovernmental transfers does not mean that levels must remain fixed, or that central authorities cannot adjust flows to take account of macroeconomic conditions. Rather, it means that the process for establishing transfer levels and allocating the total amount among municipalities is clearly spelled out in grant or tax-sharing formulas; that adequate advance notice must be given to local governments of changes; and that local authorities have the opportunity to participate in discussions with government when the principles of intergovernmental revenue-sharing are modified. 5.7 Finally, central government should draw a clear line as to what types of local projects it will finance (or co-finance) through grants and subsidized credit. A policy statement should identify the limited range of projects that will receive subsidies. Government should restrict the list of grant-eligible local investment projects to a range it can actually finance over a reasonable period of time. Its commitments to intergovernmental capital financing should appear as part of central government's own multi-year rolling capital plan. Governnent should make clear that all other local investment projects will have to be financed from local savings or local borrowing. Clearly drawing the line on subsidy policy is a fundamental step in clearing the way for private financing of local investments. 5.8 None of these "pre-conditions for a well-functioning municipal credit market" concerns the lending instrument itself. In fact, each of the reforms is a simple matter of strengthening municipalities' capacity for local financial management or making it easier for outside institutions to get a true reading of local financial condition. The items on the list reflect the fact that the greatest constraint on municipal credit market expansion, almost everywhere but certainly in the first stages of market development, is finding clearly creditworthy municipal borrowers that have well-defined investment projects they want to pay for by borrowing. Introducing Parastatal Intermediaries into Municipal Lending 5.9 Parastatal intermediaries can "jump start" the municipal lending process. They make most sense where there is a wide difference between the cost at which central government can obtain credit in the market and the costs at which local governments can obtain credit on their own. In these circumstances, central authorities can lower the cost of capital for local governmental units by borrowing on the bond market or obtaining funds from multi-lateral 56 lenders, then on-lending the funds through a government-supported financial intermediary. Government borrowing on behalf of municipalities makes much less sense in macroeconomic terms if the State budget already runs a large deficit and the State risks jeopardizing its own future access to credit markets should it borrow large sums on behalf of local governments. 5.10 A parastatal intended to help build a self-sustaining municipal credit market should be designed from the start either to go out of business as a lender as swiftly as possible, by graduating its clients to the private credit market, or to evolve into a competitive supplier of municipal credit, which raises financing without governrment help. The process of transition may require decades to complete, but the goal of clearing the way for an efficient private credit market provides a framework for measuring the institutional progress of municipal development funds. 5.11 A Municipal Development Fund of this kind will have an ever-changing client base. At the lower end of the credit-risk distribution, it will constantly be seeking to prepare new municipalities for lending, by strengthening their creditworthiness and weaning them from dependence on subsidized capital. At the higher-quality end of the risk distribution, the MDF will continually be graduating its best clients into the private credit market. A basic set of policy rules can help guide the operations of intermediaries taking on this difficult mission. 5.12 First, where a country has many different parastatal intermediaries providing credit to municipalities, these should be consolidated or at least managed according to a consistent set of lending principles. Often, multiple intermediaries will have sprung up over time, supplying credit on inconsistent terms. In the Philippines in the early 1990s, in addition to a World Bank-supported Municipal Development Fund, the Philippines National Bank, the Land Bank, the Development Bank of the Philippines, and several other government finance institutions all provided subsidized credit to local governments. In the words of one report, "a laissez-faire approach to local governmental unit lending [was] adopted; whoever ha[d] the funds and the motivation to lend to LGUs could just lend."' This proliferation of governmental lenders not only sent contradictory signals to local governments about central authorities' willingness to subsidize credit and the priorities the government had for project development but bred inattentiveness to loan oversight.2 When the Philippines, with external assistance, rationalized its municipal lending program around a common set of principles for credit market development, one of its first steps was to scrap the multiplicity of on-lenders in favor of a strengthened MDF charged with preparing the way for private credit market operations. 5.13 At each stage of its operations, a parastatal lender should strive to put in place mechanisms that, if successful, can be quickly and simply spread to the private sector. For 'Philippine Institute for Development Studies, Local Government Access to Private Sector Financing (October 1994). 2 In 1989 and 1990, for example, all of the local government loans outstanding at the Development Bank of the Philippines were in arrears. In 1987, 98.2 percent of GSIS loans to local governments were in arrears; by 1990 the arrearage rate had risen to 100 percent. The local government sector was so crippled by bad loans that in 1990 the Government approved a local debt forgiveness program. See Peterson, Local Government Access to Credit: a Strategy for Long-Term Development of the Philippine MDF (World Bank, August 1993). 57 example, intercept arrangements that are introduced into parastatal lending and succeed in reducing credit risk should be available to private lenders in the same form, not, as often now is the case, prohibited for lenders outside the public sector. Matching capital grants from government should be available for eligible projects regardless of the source of credit financing, not tied to the use of the parastatal's loans. 5.14 The period of sheltered lending should be used to build the legal basis for broader credit market development. The legal foundation should cover such subjects as the types of collateral that municipalities can provide for loans and how this collateral can be accessed by lenders when loans go into default; the legal procedures that will be followed in the case of municipal insolvency and where debt ranks in the priority of claims on insolvent borrowers; under what conditions, if any, and through what procedures government guarantees will be provided to municipal debt; as well as the procedures for central government oversight of municipal borrowing. The legal framework should be drafted in anticipation that it will apply to a competitive market utilizing many different forms of lending to local governments, and establish common rules that apply across lending instruments. 5.15 The parastatal intermediary should look for opportunities to engage the private sector in municipal lending as a partner in co-financed credits, on the principle that the more quickly private lenders become introduced to municipal lending the more quickly they can assume principal responsibility for servicing at least the upper stratum of the municipal credit market. In some contexts, the parastatal lender can best assist private-sector entry by providing long-term capital for municipal lending to commercial banks. In other contexts, it may join with private lenders to co-finance local privatization initiatives. The public financial institution may agree to take a more risky, subordinated debt position in order to encourage new kinds of private- sector lending, such as lending for local-government project finance. At times, the MDF may find that it can support development of the private municipal credit market more effectively by insuring private lenders against certain kinds of political risk than by providing financing. 5.16 The parastatal should move its own lending rates toward market rates of interest. It does not have to lend at the same risk-adjusted rate as private lenders; part of its function is to seek out riskier borrowers that the private market will not service. However, if MDFs charge the same nominal rate of interest as private lenders, the eventual shift to private sources of financing will be facilitated. Municipalities will be exposed to the market cost of capital in their budget planning. Municipal borrowers will not have a perverse incentive to perpetuate their budget problems so as to remain eligible for lower-rate borrowing. 5.17 The mission outlined above for Municipal Development Funds is quite different from the one MDFs usually now are assigned. It requires a different set of indicators of project achievement. MDF success would be measured by such indicators as: growth of the overall private-sector municipal credit market, not the rate of loan disbursements of the MDF 58 * the rate at which municipalities "graduate" from MDF borrowing to the private market * MDFs' ability to raise capital on their own, as true financial interrnediaries, rather than to remain dependent on international institutions' lines of credit Measures such as these almost never now appear in donor project documents setting forth MDF objectives. Their absence indicates that the mission of building a permanent municipal credit market is not taken very seriously relative to other functions, such as disbursing project funds or exerting quality control over the design of individual infrastructure works. 5.18 MDFs that are asked to support credit market development in the form outlined above cannot achieve some of the other objectives that are commonly demanded of them. They cannot, for example, become fully self-sustaining financially. They will require either continuing injections of governmental operating support, or an equity endowment that allows them to perform their developmental role within the spread they can earn on their credit operations. Building a Competitive Credit Market 5.19 The final stage of municipal credit market development involves building competition into the market. Competition provides the greatest surety of self-sustaining market orientation. 5.20 Perhaps the most important step that countries can take in supporting competition is also the simplest: to refrain from passing laws that establish legal monopolies and approving policies that establish de facto monopolies. A common legal framework can support the development of competitive bank lending and municipal bonds. Market competition then can sort out the financing roles of these instruments, given the attractiveness of each to different classes of lenders and different classes of borrowers. 5.21 Attempts to force feed a particular model of credit market development generally have failed. Often, these initiatives have tried to replicate a particular nation's municipal credit market structure. Attempts by United States aid agencies to introduce the US municipal bond market model as the first form of commercial lending to the municipal sector have floundered in countries like the Philippines and Indonesia, where bank lending is a more natural first model, as it is the predominant form of credit financing throughout the economy. In contrast, municipal bonds have grown swiftly in the Czech Republic, Colombia, and Brazil, virtually without foreign donor support. In Eastern Europe, European aid agencies often have sought to replicate their own national histories by creating municipal banks possessed of legal monopolies on municipal lending. This kind of nearsightedness does not fit the needs of developing countries trying to attract private capital into municipal infrastructure investment. The legal and institutional framework for municipal lending should be broad enough to accommodate any credit arrangement that is mutually advantageous, without attempting to force municipal credit down a particular institutional path. 59 5.22 A more difficult issue to resolve is the unbundling of all the functions that traditionally have been packaged together in Municipal Development Funds. These functions include: * technical assistance in preparation of municipal investment projects * technical assistance in assessing household capacity to pay and tariff structures * assessment of municipal creditworthiness and capacity to borrow * construction oversight * advice on privatization alternatives * making loans to municipalities 5.23 Where local governments have control over significant fiscal resources, it will be more efficient eventually to separate these functions and allow specialized providers to compete for municipal business. 5.24 In the meantime, MDFs face a choice of which of the above functions they should relinquish and which they should retain, as unbundling gets underway. In practice, MDFs almost always have tried to retain the financing function as long as possible. Financing seems to carry with it more institutional income and hence a more robust base for institutional survival. It may be more useful developmentally, however, for MDFs to retain their technical assistance functions and shed their financing responsibilities. As noted throughout this Report, once the basic policy and legal framework is in place, private suppliers of credit are likely to be able to provide financing for creditworthy projects. Technical assistance-and particularly technical assistance in helping municipalities understand and choose between alternative financing and privatization options, or technical assistance in financial management that lessens credit risk for all lenders- may in the end prove the scarcest resource in the municipal sector. Where this is true, government may instruct an MDF to sell off its municipal loan portfolio, and specialize in technical assistance that makes municipalities more effective users of private-sector financing.3 3 A policy of this kind is currently under discussion in South Africa. ANNEX: INSTITUTIONAL CHARACTERISTICS OF MUNICIPAL DEVELOPMENT FUNDS Municipal Development Fund Legal Organization Brazil: State of Rio Grande do Sul. PIMES was established as a municipal Integrated Program of State development fund administered by Improvements (PIMES) BADESUL, the development bank of the Rio Grande do Sul, which owns and controls it. Lends directly to municipalities. Brazil: State of Parana. Municipal Action BANESTADO, the Parana state bank, Program (PrAM) and Urban Development originally served as financial agent for Fund (FDU) PrAM. The program then was institutionalized as a revolving fund, FDU, administered by BANESTADO with technical assistance provided by a legally autonomous organization that in practice functions as a department of the State of Parana Secretariat of Planning. Acts as direct lender. Colombia: FINDETER Operates as a discounting facility for commercial banks lending to municipalities. It is now administered by legally independent institution with an independent Board. Majority owned by the Government. Czech Republic: MUFIS Legally independent entity that onlends to private commercial banks for municipal credits. Can participate in municipal bond purchases. Owned 49% by Ministry of Finance, 49% by a state development bank, and 2% by Union of Towns and Communities. Ecuador: Municipal Development BEDE is a development bank owned 99% Program (PDM) & Ecuadorian by the government, but with policy of Development Bank (BEDE) increasing municipal share of ownership. PDM operates as a special program within BEDE. 61 Honduras: Municipal Bank (BANMA) Legally, BANMA is a bank owned by municipalities, which must use it for all of their financial transactions. In practice, it is tightly controlled by Government. India: Tamilnadu Development Fund Owned by state of Tamilnadu, and administered by the Ministry of Finance. Indonesia. Regional Development RDA is an "account" or lending window Account (RDA) within the Ministry of Finance for municipal lending. Jordan: Cities and Villages Development Began as a lending window within the Bank Ministry of Municipal and Rural Affairs and Environment, then became legally independent. Its Board is still chaired by the Minister. Kenya. Local Government Loans Legally independent, but controlled by the Authority Ministry of Local Government and chaired by its Minister. Morocco. Communal Infrastructure Fund Began as a department inside the Caisse (FEC) de Depots et de Gestion which is the global public financing institutions for the state, then became a legally independent state institution. FEC is largely controlled in practice by the Ministry of the Interior. South Africa. Development Bank of Owned by the Government. Now directed South Africa to concentrate on financing of sub- national infrastructure investment. Zimbabwe. General Development Loan Dependency of the Ministry of Local Fund Government, Rural, and Urban Development. 62 REFERENCES Credit Ratings Cheyre, Hernan V., "Credit Ratings for Infrastructure Projects," processed, paper presented to World Bank Conference on Urban Finance (August 1995), [Santiago: Duff & Phelps]. Financial Times, Credit Ratings in Emerging Markets [London: FT Financial Publishing, 1996]. Moody's Investor Service, Perspectives on Municipal Finance, various issues [New York: Moody's Investor Service]. Standard & Poor's, Municipal Finance Criteria [New York: Standard & Poor's, 1996]. Intermediation Theory Berlin, Mitchell and Jan Loeys, "Bond Covenants and Delegated Monitoring." Journal of Finance, vol. XLIII, no. 2 (June 1988), pp 397-412. Diamond, Douglas W., "Monitoring and Reputation: The Choice between Bank Loans and Directly Placed Debt." Journal of Political Economy, 1991, vol. 99, no. 4, pp 689-721. Diamond, Douglas W., "Financial Intermediation and Delegated Monitoring." Review of Economic Studies (1984) LI, pp 393-414. Dittus, Peter, Corporate Governance in Central Europe: The Role of Banks [Basle: Bank for International Settlements, August 1994]. Himmelberg, Charles P. and Donald P. Morgan, "Is Bank Lending Special?" in Joe Peek and Eric S. Rosengren (eds.), Is Bank Lending Important for the Transmission of Monetary Policy? [Boston: Federal Reserve Bank, 1995]. Is Bank Lending Important for the Transmission of Monetary Policy? [Boston: Federal Reserve Bank, 1995]. Petersen, Mitchell A. and Raghuram G. Rajan, "The Benefits of Lending Relationships: Evidence from Small Business Data." Journal of Finance, vol. XLIX, no. I (March 1994), pp 3-37. Petersen, Mitchell A. and Raghuram G. Rajan, "The Effect of Credit Market Competition on Lending Relationships." Quarterly Journal of Economics, May 1995, pp 407443. Preece, Dianna C. and Donald J. Mullineaux, "Monitoring by Financial Intermediaries: Banks vs. Nonbanks." Journal of Financial Services Research, 8 193-202 (1994). Rajan, Raghuram, "Insiders and Outsiders: The Choice between Informed and Arm's-Length Debt." Journal of Finance, vol. XLVII, no. 4 (September 1992), pp 1367-1400. 63 Municipal Development Funds in Developing Countries Davey, Kenneth, Municipal Development Funds and Intermediaries, World Bank Working Paper WPS32 [Washington, D.C.: World Bank, July 1988]. Gall, Pirie, Municipal Development Programs in Latin America [New York: Praeger, 1996]. Peterson, George E. and Bruce Ferguson, Municipal Credit Initiatives, research paper prepared for World Bank, Urban Development Division [Urban Institute: August 1991]. Project Finance Euromoney, Project and Infrastructure Finance in Asia [London: Euromoney Books, 1995]. Fabozzi, Frank and Peter Nevitt, Project Financing [London: Euromoney Books, 1995]. Public Works Financing, various issues. Colombia FINDETER Hemrnndez, Rafael, "Second Municipal Development Program." FINDETER II (CO-0138). Junguito, Roberto B., Ligia B. Melo, and Martha A. Misas, "Descentralizaci6n y Finanzas Puiblicas: Colombia 1967-1994." "Observaciones al Proyecto de Ley Sobre Disposiciones en Materia de Endeudamiento Territorial" (June 7, 1995). "Plan Para Poner en Marcha el Sistema Nacional de Cofinanciaci6n" (July 22, 1993). "Programa de Preinversi6n y Asistencia Tecnica: 'El Monitor Local."' (April 1995). "Proyecto de Ley por la Cual Dictan Disposiciones en Materia de Endeudamiento: Exposici6n de Motivos" (FINDETER). Repulblica de Colombia Departamento Nacional de Planeaci6n, "Ajustes al Sistema Nacional de Confinanciaci6n." Superintendencia Bancaria, "Endeudamiento Publico" (1995). Torres, Fernando, "El Credito como Fuente de Financiamiento de las Entidades Territoriales" (FINDETER). Turbay, Gabriel, "Credit Financing for Municipal Infrastructure in Colombia: Some Key Lessons from FINDETER's experience." 64 Valenzuela, Luis Carlos, "Seminario Endeudamiento Territorial e Impacto en el Modelo de Descentralizaci6n." Realidad del Endeudamiento Territorial en Colombia. Czech Republic Frishkopf, John J., "Development of the Bond Market in the Czech Republic." The Financier, vol. 2, no. 1 (February 1995). Peterson, George E., [Washington, D.C.: The Urban Institute, 1997]. Eastern Europe (Comparative) Peterson, George E., East European Municipal Credit Seminar: Conference Proceedings and Analysis [Washington: Urban Institute, July 1994]. Pigey, Juliana and George E. Peterson, "Municipal Credit in Eastern Europe: Background Paper," prepared for USAID, East European Municipal Credit Seminar (June 1994). Hungary Credit Local International, "Market Study for a Communal Bank in Hungary." [Paris: August 1994]. Hegedus, Jozsef, Judit Kalman, Rita Nayyar-Stone, George Peterson, Juliana Pigey, Steven Swanson, and Andrea Tonko, "Subsidy System of Municipal Infrastructure Investments in Hungary." [Budapest: Metropolitan Research Institute and the Urban Institute: March 1996]. Peterson, George E., "Demand for a Municipal Credit Facility in Hungary," processed. [Washington: The Urban Institute, April 1995]. India Allaudin, K., and P. V. Rajaraman. 1996. "Urban Infrastructure Financing-India-Tamilnadu New Financial Intermediary: The Evolution." Invited Paper under IDF Technical Assistance Programme on Urban Sector Reforms in India. The Times Research Foundation. Bhatnagar, K. K. 1996. "HUDCO's Experience in Urban Infrastructure Financing." Paper prepared for Government of India, Ministry of Urban Affairs and Employment, under IDF Technical Assistance Programme on Urban Sector Reforms in India. World Bank, "Confronting Rapid Urbanization: Lessons from Five Countries." The Experiences of Brazil, Canada, China, Korea, and Malaysia and their Relevance to India. Dasgupta, Anupam. 1995. "Debt Funding of Capital Assets." Paper prepared for National Workshop on Municipal Bond Experiences, December 4-5, Bangalore. 65 "Debt Market Component for Urban Infrastructure Financing: An Introduction." FIRE(D) paper prepared by National Institute of Urban Affairs, New Delhi. Govermment of India, Ministry of Urban Affairs and Employment. 1995. "Centrally Sponsored Scheme of Infrastructural Development in Mega Cities: Guidelines for Implementation. August. Mehta, Dev. 1995. "Urban Infrastructure Development: Need for Financial Intermediary." Paper prepared for National Workshop on Municipal Bond Experiences, December 4-5, Bangalore. Mehta, Meera. 1995. "Municipal Bonds for Urban Infrastructure: Potential and Policy Issues." For USAID, New Delhi. Mohan, Rakesh. 1995. "Urban Financing in the 1990s and Beyond: An Overview." Paper prepared for Conference on Urban Finance, Plenary Session II. September. "Municipal Bond Market for Urban Infrastructure." 1995. FIRE(D) Working Paper Series No. 2, National Institute of Urban Affairs, New Delhi. Namabiar, M. Madhavan. 1995. "Urban Infrastructure Financing-Indian Experience Tamilnadu. Paper prepared for National Workshop on Municipal Bond Experiences, December 4-5, Bangalore. Ramachandran, C. 1996. Key Note Address to the International Workshop on Urban Infrastructure Finance in the context of Municipal Decentralization in India. March 8-9, The Krishna Oberoi Hotel, Hyderabad. Technical Support Services, Inc. 1995. "The Financial Institutional Reform and Expansion (F.I.R.E.) Project." Reports by Indian Consultants for a National Seminar on Municipal Bond Experiences. Bangalore, India, December 4-5. Times Research Foundation. 1996. "Democratic Decentralization in the Context of the 74th Constitution Amendment Act, 1992. Paper prepared for Government of India, Ministry of Urban Affairs and Employment, under IDF Technical Assistance Programme on Urban Sector Reforms in India. Romania Bank Nederlandse Gementeen, "Policy Recommendations for a Municipal Credit Facility," processed (September 1996). Government of Romania, European Bank for Reconstruction and Development and Bank Nederlandse Gementeen, "Report on Outcome of Municipal Credit Workshop." (May 1996). 66 Russia Progressor Research and Innovation Center, St. Petersburg Administration Municipal Loan, 1995- 96 [Moscow: Progressor Research, 1995]. _ City of Cherepovetz Municipal Bonds Issue, 1994-2000 [Moscow: Progressor Research, 1995]. Research Triangle Institute, Capital Finance Workshop [Washington: Research Triangle Institute, 1996]. Sliva, A. Ya. and G.B. Polyak, "Comments on the Draft Law, 'On the Fundamentals of Local Self- Government,"' processed. Ter-Minassian, T.J., J. Diamond, P. Martinez-Mendez, and P. Parente, "Russian Federation: Selected Issues in Public Debt Management," International Monetary Fund, Fiscal Affairs Department, Washington, D.C. [October 1992]. World Bank, Fiscal Management in Russia [Washington: World Bank, 1996]. . Russian Federation. South Africa Miller, John, et al. 1995. "South Africa: An Assessment of Municipal Environmental Services and Infrastructure." Prepared for USAID. April. Minis, Henry P., Jr. and Thomas Cochran. 1996. "Financing Municipal Infrastructure in South Africa." Recommendations for Policy Implementation. Paper prepared for USAID. March. Riordan, Rory. 1995. "The Changing Face of Intergovernmental Fiscal Relations in South Africa." Paper prepared for the Habitat Urban Finance Conference, World Bank, Washington, DC, September 10-13. Southeast Asia Organization for Economic Cooperation and Development, Emerging Bond Markets in the Dynamic Asian Economies [Paris: OECD, 1993]. The World Bank Urban Development 1818 H Street, NW Washington, DC 20433 USA Internet: www.worldbank.org/urban