SUBNATIONAL CAPITAL MARKETS IN DEVELOPING COUNTRIES FROM THEORY TO PRACTICE Editors | Mila Freire and John Petersen with Marcela Huertas and Miguel Valadez Subnational Capital Markets in Developing Countries From Theory to Practice Subnational Capital Markets in Developing Countries From Theory to Practice Editors Mila Freire and John Petersen with Marcela Huertas and Miguel Valadez A copublication of the World Bank and Oxford University Press © 2004 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, N.W. Washington, D.C. 20433 Telephone 202-473-1000 Internet www.worldbank.org E-mail feedback@worldbank.org All rights reserved. 1 2 3 4 06 05 04 A copublication of the World Bank and Oxford University Press. Oxford University Press 165 Madison Avenue New York, NY 10016 The findings, interpretations, and conclusions expressed here are those of the au- thor(s) and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the governments they represent. The World Bank cannot guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply on the part of the World Bank any judgment of the legal status of any territory or the endorsement or acceptance of such bound- aries. Rights and Permissions The material in this work is copyrighted. No part of this work may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or inclusion in any information storage and retrieval sys- tem, without the prior written permission of the World Bank. The World Bank en- courages dissemination of its work and will normally grant permission promptly. For permission to photocopy or reprint, please send a request with complete in- formation to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, World Bank, 1818 H Street N.W., Washing- ton, D.C. 20433, fax 202-522-2422, e-mail pubrights@worldbank.org. ISBN 0-8213-5464-7 Library of Congress Cataloging-in-Publication Data Subnational capital markets in developing countries: from theory to practice / edited by Mila Freire and John Petersen. p. cm. Includes bibliographical references and index. ISBN 0-8213-5464-7 Debts, Public--Developing countries. 2. Capital market--Developing countries. 3. Subnational governments--Developing countries. I. Freire, Mila. II. Petersen, John E. III. World Bank. HJ8899.S857 2003 332'.041'091724­dc22 2003061373 Contents Preface xvii Acknowledgments xxi Abbreviations and Acronyms xxiii Executive Summary xxix Chapter 1 Introduction 1 I Political, Legal, and Financial Framework 9 John Petersen and Mila Freire Chapter 2 Fiscal Devolution 11 Chapter 3 Market Setting and Legal Framework 29 II Borrowing Instruments and Restrictions on Their Use 47 John Petersen and Miguel Valadez Chapter 4 Subnational Governments as Borrowers 49 Chapter 5 The Nature and Design of Debt 63 Chapter 6 Debt Instruments and Methods of Sale 77 Chapter 7 Restrictions on the Issuance and Use of Subsovereign Debt 87 III Characteristics of Financial Market Regulation and Disclosure 111 John Petersen Chapter 8 Financial Market Structure, Regulation, and Operations 113 Chapter 9 Disclosure and Financial Reporting 129 v vi Contents IV Evaluating, Monitoring, and Assisting Subnational Governments 139 John Petersen and Marcela Huertas Chapter 10 Credit Analysis and Credit Ratings 141 Chapter 11 Monitoring and Intervening in Subnational Government Finances 155 Chapter 12 Designing and Implementing Credit Assistance to Subnational Governments 173 V Policy Guidelines 203 John Petersen and Mila Freire Chapter 13 Concluding Observations and Policy Guides 205 VI Country Case Studies 217 Latin America and the Caribbean Chapter 14 Argentina 219 Rodrigo Trelles Zabala Chapter 15 Brazil 261 Rodrigo Trelles Zabala and Giovanni Giovanelli Chapter 16 Colombia 279 Rodrigo Trelles Zabala Chapter 17 Mexico 299 Steven Hochman and Miguel Valadez Sub-Saharan Africa Chapter 18 South Africa 313 Matthew Glasser and Roland White Chapter 19 Zimbabwe 337 Roland White and Matthew Glaser Middle East and North Africa Chapter 20 Morocco 355 Samir El Daher Contents vii Chapter 21 Tunisia 365 Samir El Daher Asia Chapter 22 People's Republic of China 375 John Petersen Chapter 23 Republic of Korea 399 John Petersen Chapter 24 India 413 Pryianka Sood Chapter 25 Indonesia 443 Robert Kehew and John Petersen Chapter 26 The Philippines 461 John Petersen Eastern and Central Europe Chapter 27 Bulgaria 487 Peter D. Ellis and Kremena Ionkova Chapter 28 Czech Republic 503 João C. Oliveira and Jorge Martinez-Vazquez Chapter 29 Hungary 525 Pryianka Sood Chapter 30 Poland 545 Miguel Valadez and John Petersen Chapter 31 Russian Federation 571 Asad Alam, Stepan Titov, John Petersen Bibliography 593 Index 607 Boxes 2.1. Devolving Responsibility for Elementary School Teachers' Salaries in Romania 14 2.2. Rio and the International Marketplace 23 viii Contents 3.1. Banks and Securities Markets: Are Both Needed for Development? 32 3.2. Brazilian Banks' Excessive Concentration in Government Securities 33 4.1. Defining and Controlling Public Debt 56 4.2. China: Off-Budget Finance and the Transmuted Bond 59 4.3. Restructuring Subnational Government: From Few to Many (But How Many?) 60 5.1. Importance of the Rate-Setting Pledge 67 5.2. Intergovernmental Transfer Payments as Collateral 69 5.3. Importance of Feasibility Reports: The San Pedro Sula, Honduras, Sports Complex 72 6.1. Selecting an Underwriter through Competitive Negotiation 84 6.2. Rigging a City's Bond Sale 85 7.1. The Philippines: How Political Risks Can Inhibit Municipal Credit Markets 89 7.2. Examples of Language on the Binding Nature of Financial Obligations 90 7.3. The City of Cebu in the Philippines Considers a Deal 91 7.4. Johannesburg Comes Up Short 96 7.5. Example of Language Denying Central Government Responsibility for Municipal Debt 106 7.6. Example of Language on Securing Debt with Own Revenues 107 8.1. Commercial Banking in Transitioning Economies 116 8.2. The Bank for International Settlements' Reserve Requirements and Capital Rules 117 8.3. What Is a Security? 121 Contents ix 8.4. After 60 Years, Municipal Bonds Return to Romania 124 9.1. Disclosure over the Internet 131 9.2. Accounting for Accounting Differences 134 9.3 Why Did Czech Municipal Debt Grow So Fast? 135 10.1. Emerging Market Ratings and Bond Insurance 143 11.1. Example of Information Provided in the Debt Annex of French Subnational Government Budgets 163 11.2. In Argentina Trustees Make a Difference 166 11.3. Debt Adjustment and Subnational Insolvency in Hungary 168 11.4. Financial Stabilization to Address Subnational Bankruptcy in Latvia 169 12.1. The Subnational Government Retreat from the Private Credit Market in the Czech Republic 175 12.2. Moving from Soft to Hard Credit through Enforcement of Loan Collections: South Africa's Experience 178 12.3. The Philippine Local Government Unit Guarantee Corporation 183 12.4. The Tamil Nadu Urban Development Fund, India 185 12.5. Assisting Small Bond Issuers: The Bond Bank Option 189 12.6. A Brief Illustration of Grant-Loan Integration: An Example from Indonesia 194 18.1. A South African Parable 316 24.1. Recent Projects Financed by the Tamil Nadu Urban Development Fund 435 24.2. Basis for the AA+ Rating of the Madurai Municipal Corporation Bond Issue 438 x Contents Figures 3.1. Market Structures and Sources of Capital for Local Government Borrowing 37 3.2. Stages of Development in Credit Market Access 39 4.1. General Government Obligation 49 4.2. Government Limited Obligation 49 4.3. Public-Private Project Financing 50 4.4. Matrix of Subnational Government Financing Capacity 52 6.1. Debt Service Structures 79 12.1. Retail On-Lending by the Government Financing Institution 180 12.2. Wholesale On-Lending by the Government Finance Institution 181 12.3. Securitization of a Loan Pool 187 12.4. Mechanics of a Liquidity Facility 192 14.1. Distribution of Shareable Taxes under the Coparticipation Scheme, Argentina 223 14.2. Relative Fiscal and Debt Situations of Provinces, Argentina, 2001 226 14.3. Provincial Indebtedness by Type of Debt or Lender, Argentina, December 2001 227 14.4. Impact of the Devaluation on Provincial Debt, Argentina 230 14.5. Disbursement of Coparticipation Revenues, Argentina 231 14.6. Provincial Bond Debt Outstanding by Type, Argentina, End of 2001 233 14.7. Flow of Funds for the Salta Hydrocarbon Royalty Trust Bonds 241 14.8. Selected Debt Indicators, Salta and All Provinces, End-2001 243 14.9. Selected Debt Indicators, City of Buenos Aires and All Provinces, End-2001 250 Contents xi 14.10. Selected Debt Indicators, Buenos Aires and All Provinces, End-2001 256 15.1. Distribution of the Debt Stock in Bonds by State, Brazil, End-1996 272 15.2. Subnational Debt as a Share of GDP, Brazil, 1998­2002 274 16.1. Fiscal Balance as a Share of Total Revenue by Department, Colombia, 2000 284 16.2. Direct Subnational Debt, Colombia, 1996­2001 287 16.3 Allocation of Credit from Findeter 1989­99 289 16.4. Debt Stock, Capital District of Santa Fe de Bogotá, 1995­2001 296 17.1. Borrowing by Three State Governments, Mexico, 1994­98 304 18.1. Outstanding Municipal Debt, South Africa, 1997­2000 325 18.2. Outstanding Municipal Debt by Form, South Africa, 1997­2000 326 22.1. Typical Cooperative Joint Venture Arrangement for Expressway Development 385 24.1. Funding Approvals and Disbursements by the Tamil Nadu Urban Development Fund by Sector, as of 31 March 1999 434 24.2. Value of Capital Works Executed by Municipalities with Funding from the Tamil Nadu Urban Development Fund, 1993/94 to 1998/99 434 24.3. Simplified Flow of Funds in the Pooled Financing Scheme 441 27.1. Local Government Expenditure as a Share of GDP, Selected Countries, Various Years, 1998­2002 492 28.1. Municipal Debt Outstanding, Czech Republic, 1993­99 509 xii Contents 28.2. Local Fiscal Deficits, Czech Republic, 1993­2000 509 28.3. Composition of Municipal Debt Outstanding, Czech Republic, 1993­99 511 28.4. Average Composition of Municipal Debt Outstanding, Czech Republic, 1998­99 513 29.1. Sources of Local Government Revenue, Hungary, 1995­2000 530 30.1. Own-Source Revenue as a Share of Total Revenue, Szczecin, 1998­2002 552 30.2. Local Government Debt by Source, Poland, 1999­2001 555 30.3. Concepts of Surplus from Operating Revenues as a Source of Funds for Capital Spending 559 30.4. Structure of Operating Revenue, Wroclaw, 2001 561 30.5. Budget and Debt, Wroclaw, 1996­2002 562 30.6. Structure of Operating Expenditure, Lodz, 2001 564 30.7. Budget and Debt, Lodz, 1996­2002 564 30.8. Debt Burden as a Share of Operating Revenue, Lodz, 1996­2002 565 30.9. Budget Balance and Debt, Szczecin, 1998­2002 566 30.10. Debt and Debt Service as a Share of Operating Revenue, Szczecin, 1996­2002 567 30.11. Budget and Debt, Krakow, 1998­2002 568 31.1. Administrative Structure for Executing the Debt Strategy, St. Petersburg 589 Tables 7.1. Municipal Debt Limitations in Selected Eastern and Central European Countries 102 10.1. Credit Rating Volatility in Asia: Selected Standard and Poor's Long-Term Foreign Currency Sovereign Ratings 149 Contents xiii 14.1. Allocation of Responsibilities among Levels of Government, Argentina 224 14.2. Terms and Conditions of the Typical Consolidation Bond, Argentina 235 14.3. Provincial Bond Issues in Domestic and International Capital Markets, Argentina, 1994­2001 235 14.4. Features of the Bond Issue by the Salta Hydrocarbon Royalty Trust 241 14.5. Debt by Source, Salta, 1995­2001 243 14.6. Key Features of the Bond Program of the City of Buenos Aires 247 14.7. Main Characteristics of the Bond Issues by the City of Buenos Aires 248 14.8. Debt by Source, City of Buenos Aires, 1995­2001 250 14.9. Access to the Bond Market by the Province of Buenos Aires, 1994­2001 254 14.10. Debt by Source, Province of Buenos Aires, 1995­2001 256 15.1. Municipal Sources of Funds, Brazil, 1999 275 16.1. The "Traffic Light" System for Regulating Subnational Borrowing, Colombia 286 16.2. Potential Borrowers from Findeter 288 16.3. Terms and Conditions of Findeter Loans 289 16.4. Features of the Bond Issue by the Capital District of Santa Fe de Bogotá 293 16.5. Revenues and Expenditures, Capital District of Santa Fe de Bogotá, 1995­2001 294 16.5. Debt Service, Capital District of Santa Fe de Bogotá, 1995­2001 296 17.1. Spending and Own-Source Revenues as a Share of GDP by Level of Government, Mexico, Selected Years, 1991­97 301 17.2. Subnational Bond Issues, Mexico, 2002 310 19.1. Assets of Deposit-Taking Institutions, Zimbabwe, 1992­97 339 xiv Contents 19.2. Local Government Revenues by Source, Zimbabwe, 1995­98 343 19.3. Gross Public Debt of Local Authorities, Zimbabwe, 1994­97 345 19.4. Bond Issues by Local Governments, Zimbabwe, 1990­2001 346 22.1. Subnational Revenues and Expenditures, 1993 to 2001 379 22.3. Market Capitalization, Bonds, and Domestic Bank Credit as Percentages of GDP: 1995­2000 389 22.4. Outstanding Domestic Bonds and Issuance in 2001 390 23.1. Debt and Capital Spending of the Seoul Metropolitan Government, Republic of Korea, Fiscal Years 1996­2001 410 24.1. Third Tier of Government, India, 2000 418 24.2. Fiscal Decentralization, India, 1997/98 419 24.3. Finances of Local Bodies, India, 1990/91 and 1997/98 419 24.4. Terms of the Bond Issue by the Ahmedabad Municipal Corporation 425 24.5. Estimated Gap in Urban Infrastructure Financing, Tamil Nadu, 2002 428 24.6. Infrastructure Investment Requirements by Type of Urban Local Body and Sector, Tamil Nadu, 1996­2001 428 24.7. Lending Terms of the Tamil Nadu Urban Development Fund since 1998/99 433 24.8. Terms of the Bond Issue by the Madurai Municipal Corporation 437 24.9. Financial Indicators for the Tamil Nadu Urban Development Fund as of March 2002 439 24.10. Terms of Issue of the Water and Sanitation Pooled Fund 441 Contents xv 25.1. Central Government Lending to Local Governments, Indonesia, Selected Years, 1980­99 446 25.2. Banking Sector Assets by Type of Bank, Indonesia, End of March 2001 449 26.1. Local Government Loans and Deposits with Selected Financial Institutions, Philippines, 2000 467 27.1. Financial Performance Indicators for Sofia, 1996­2001 498 27.2. Credit Ratings of Selected Local Governments, 2001 499 28.1. Public Debt Outstanding, Czech Republic, 1993­99 507 28.2. Municipal Bonds Issued, Czech Republic, 1992­99 512 29.1. Equity and Debt Markets, Selected Countries in Central and Eastern Europe, End-1995 532 29.2. Terms of the Bond Issue by the Municipality of Budapest 539 29.3. Terms of the Bond Issue by the Municipality of Pecs 540 30.1. Current Structure of Subnational Government, Poland 551 31.1. Standard & Poor's Credit Ratings of Subnational Borrowers, Russian Federation, 1997­2002 578 31.2. Debt by Type, St. Petersburg, 1994­2001 581 31.3. Structure of Debt, St. Petersburg, 1994­2001 582 31.4. Debt Indicators, St. Petersburg, 1997­2001 583 31.5. Credit Ratings, St. Petersburg, 1998­2002 586 Preface This book examines institutional aspects of subnational capital mar- kets and presents case studies of subnational borrowing, showing what has worked, what has not, and why. As decentralization contin- ues and urbanization spreads, local authorities need to provide more services with fewer resources from the central government. Subna- tional borrowing, leveraging on reliable cash flows and prudent fiscal management, can be an alternative for funding some investments, especially when the useful life of the service is long (such as schools, roads, and public utilities). Worries that fiscal decentralization may contribute to structural deficits and fiscal imbalances are common, especially in countries where the main policy priority is to control aggregate public sector borrowing. When traditions of fiscal responsibility are weak, account- ability systems are immature, and administrative discipline is poorly developed, there is a risk that lower-level governments may abuse their borrowing authority, contributing to aggregate fiscal imbalance with adverse macroeconomic consequences. However, many analysts argue that, with an adequate legal frame- work and sound macroeconomic fundamentals in place, local gov- ernment access to capital markets is compatible with fiscal stability and promotes development of an important segment of the financial market. Necessary conditions include effective supervisory authori- ties; judicially enforceable contracts; tax decentralization; civic norms that promote fiscal prudence; availability of skilled staff; and adequate accounting, disclosure, and reporting standards. Increasing numbers of local governments are borrowing from banks and issuing bonds, although in these turbulent times the path has been neither steady nor smooth. During the 1990s 120 subnational governments in Latin America engaged in market bor- rowing, 150 in Eastern and Central Europe, 11 in East Asia, and nearly 500 in Africa. Domestic bond markets started to blossom in xvii xviii Preface Poland, Russia, South Africa, and Zimbabwe. However, by the late 1990s the international market for subnational debt entered a peri- od of sustained slump. Domestic bond markets, though not im- mune to the troubled economic conditions and ongoing structural changes, have continued to rise in importance. The participation of subnational governments in these domestic markets--where success will be critical to reenergizing international access--is the focus of this study. This book draws from the findings of the Global Program on Sub- national Capital Markets launched in 1998 by the World Bank with the sponsorship of the governments of Austria, Finland, Japan, Spain, Sweden, and Switzerland. Led by Augusto de La Torre, Mila Freire, and Marcela Huertas under the supervision of Danny Leipziger and Guillermo Perry, the program examined the experiences of subna- tional governments in accessing domestic and international capital markets. The objective was a mixture of fact finding and analysis of how countries deal with subnational borrowing and how they recon- cile it with budgetary and fiscal balance. The two-year program significantly advanced knowledge in this area. Among the outputs were regional studies for Latin America, Central Europe, East Asia, and Sub-Saharan Africa, undertaken in col- laboration with major universities, rating agencies, investment banks, and development banks. The findings were disseminated through seminars and workshops, international conferences, and training materials. The program brought to light the challenges of al- lowing subnational governments access to debt financing and the need for a regulatory framework that protects all interested parties: local governments, central governments, and bond issuers. This book also draws on a variety of other sources and studies, con- solidating that work and deriving lessons about how to improve ef- forts to promote credit market access. It pools information on the is- suing of subnational debt and its characteristics, analyzes the role of macroeconomic conditions and market development in the success or failure of such borrowing, and offers policy guidelines for ongoing efforts. The goal is to assist the World Bank and its clients to work as Preface xix strategic partners in developing and strengthening capital markets as sources of funds for local governments in emerging economies. The book provides a framework for analysis based on a systematic study of subnational governments as borrowers and the array of cred- it markets in which they may operate. Complementing the frame- work is a set of case studies that document the recent experience of 18 countries in developing markets for subnational borrowers. Mila Freire John Petersen Regional Advisor Professor Latin America and the School of Public Policy Caribbean Region George Mason University World Bank Acknowledgments The book was prepared by John Petersen and Mila Freire, lead editors, with the collaboration of Marcela Huertas and Miguel Valadez. John Petersen was lead author, with chapter contributions from Asad Alam, Michael D'Angelis, Colleen Butcher, Benjamin Darche, John Dunn, Samir El Daher, Peter D. Ellis, Giovanni Giovanelli, Mathew Glasser, Steven Hochman, Robert Kehew, Kremena Ionkova, João C. Oliveira, David Rosen, Pryianka Sood, Stepan Titov, Rodrigo Trelles Zabala, Miguel Valadez, and Roland White. Useful insights and mate- rials were obtained from many, including Barbara Boone (Moody's Investor Services), Jane Eddy (Standard & Poor's), and Gersan Zurita (Fitch Ratings). Also very valuable were inputs prepared for the earlier work on subnational borrowing by AB Asesores, with the assistance of Marcela Huertas, on the Latin American case studies; Benjamin Darche on the European and Central Asian case studies; SRIC Corporation, Research Triangle Institute on the East Asian studies, and Enrique Asturizaga; and Remy Prud'homme on the African case studies. The initial pro- ject was encouraged and approved by Frannie A. Leautier, Vice Presi- dent of the World Bank Institute, and funded by the President's Con- tingency Fund under the leadership of Mr. James D. Wolfensohn. The work was conducted under the supervision of Danny Leipziger, Direc- tor of the World Bank's Latin America and the Caribbean Region's Fi- nance, Private Sector, and Infrastructure Department. Eleoterio Coda- to, Anwar Shah, and Clemente del Valle served as internal reviewers. Useful comments were received from Anjali Kumar and Victor Vergara and from external reviewers. The book was edited by Meta de Coquereaumont of Communications Development Inc., and Mary Fisk, Office of the Publisher, served as production editor. xxi Abbreviations and Acronyms AMC Ahmedabad Municipal Corporation ARD/GFG Municipal bond consulting firm Arg$ Argentina currency unit AVK Securities and Finance Ltd./Corporation BANOBRAS Banco Nacional de Obras y Servicios Públicos BAPEPAM The Securities Regulatory Body in Indonesia BAPPENAS Indonesia's National Planning Agency BAPRO Banco de la Provincia de Buenos Aires BGN Bulgarian Currency Unit Leva BIS Bank of International Settlement BNA Banco de la Nación Argentina BNDES Brazil's Federal Development Bank BNP Bond Dealing Corporation [Argentina] BOCON Bonos de Consolidación BPP Banks Privatization Program BSF Bond Service Fund CAPMAC Capital Markets Assurance Company CARE Indian Credit Rating Agency CEF Brazil's Federal Housing and Savings Bank (Caixa Economica Federal) CIP Communal Investment Program COA Central Office of Audit CLF Consolidated Loan Fund CMC Chennai Municipal Corporation CMIP Consolidated Municipal Infrastructure Programme CPSCL Caisse de Prêts et Soutien aux Collectivités Locales CRISIL Credit Rating Information Service of India, Limited CRR Cash Reserve Ratio CSFB Credit Suisse First Boston CSN Brazilian company xxiii xxiv Abbreviations and Acronyms CZK Czech local currency DAK Indonesia government grant DAU Indonesia government equalization grant DBP Development Bank of the Philippines DBSA Development Bank of Southern Africa DFLA Development Fund for Local Authorities DS Debt Service DTC Depository of Notes in Argentina EBRD European Bank for Reconstruction and Development ECCI East Coast Constructions and Industries Private, Limited EMU European Monetary Union ESAP Economic Structural Adjustment Program ESR Environmental and Social Report EU European Union EUR Euro currency unit FCL Fiscal Coordination Law FDI Foreign direct investment FEC Communal Infrastructure Fund FI Financial Institution FINDETER Financiera de Desarrollo Territorial GAAP Generally Accepted Accounting Principles GDP Gross domestic product GFIs Government Financial Institutions GoH Government of Hungary GOI Government of India GoTN Government of Tamil Nadu GRP Gross regional product GSIS A pension fund HCMC Ho Chi Min City HDFC Housing Development Finance Corporation HIGC Philippines Government Guarantee HUDCO Housing and Urban Development Corporation HUF Hungarian currency unit IADB Inter-American Development Bank Abbreviations and Acronyms xxv ICICI Industrial Credit and Investment Corporation of India, Limited ICMS Value-added tax in Brazil ICRA Indian Credit Rating Agency IDP Integrated Development Plan IFI Type of foreign debt [Russia] IGT Intergovernmental Transfer IL&FS Infrastructure Leasing and Financial Services Limited IMF International Monetary Fund INCA Infrastructure Corporation of Africa IPTU Brazil property tax IRA Internal Revenue Allotment [Philippines] ISS Brazil services tax JBIC Japanese bank LALF South African Loan Portfolio LBP Land Bank of the Philippines LECOP Argentinian money bonds LGCF Local Government Credit Fund LGLA Local Government Loans Institution LGLB Local Government Loans Board LGU Local Government Unit LGUGC Local Government Unit Guaranty Corporation LIBID Interest rate LIBOR London Interbank Offered Rate Lpcd Litres Per Capita Daily LWUA Local Water Utilities Administration MBIA Bond Insurance Company MBS Mortgage Backed Securities MCMC Act Indian Act MDF Municipal Development Fund MFM Bill Municipal Finance Management Bill MLG Ministry of Local Government MoF Ministry of Finance MOHA Ministry of Home Affairs MUDF Municipal Urban Development Fund xxvi Abbreviations and Acronyms MUFIS Czech Local Government Fund NAFTA North American Free Trade Agreement NGO Nongovernmental organization NPA Non-Performing Assets NSE National Stock Exchange O&M Operate and Monitor OECD Organisation for Economic Co-operation and Development OMBs Open Market Bonds OTC Over the Counter OTP National Savings Bank [Hungary] PDAM Local water utility PDF Provincial Development Fund PNB Philippines National Bank POSB Post Office Savings Bank PPP Public Private Partnership PRI Panchayati Raj Institutions PSE Public Sector Enterprise RBI Reserve Bank of India RDA Regional Development Authority Funds [Indonesia] RDI Government Loan RSC levies Regional Services Council levies RTI/SCRI Consulting Firm S&P Standard & Poor's SEBI Securities and Exchange Board of India SLA Subsidiary Loan Agreement SLR Statutory Liquidity Ratio SOE State Owned Enterprise SSA Sub-Saharan Africa STP Sewerage Treatment Plant TE Total Expenditures TD Tunisian currency unit (Dinars) TNUDF Tamil Nadu Urban Development Fund TNUDP Tamil Nadu Urban Development Project Abbreviations and Acronyms xxvii TNUIFSL Tamil Nadu Urban Infrastructure Financial Services, Limited TR Total Revenues UBS United Bank of Switzerland UDI Unidades de Inversión [México] ULB Urban Local Body UNCHS United Nations Centre for Human Settlements UNDP United Nations Development Program UPAP Urban Policy Action Plan USAID United States Agency for International Development USD United States Currency Unit VAT Value-added tax ZAR Zimbabwe currency unit Executive Summary Raising capital for investment in infrastructure facilities is a universal concern in developing and transitioning economies. These long- lived facilities are crucial for building healthier, better-served popula- tions and for creating competitive economies. Properly planned, op- erated, and maintained, such investments provide benefits for many years. However, in these countries long-term capital is scarce and has many claimants. The challenge to raise funds comes at a time of transition and un- certainty. Devolutionary changes in the scheme of governance and dispersal of fiscal decisionmaking are pushing down responsibility for meeting capital needs and for subsequently operating facilities to a vast array of provinces, cities, and villages. Concurrently, there is increasing pressure to make government at all levels more account- able to citizens and more attuned to the demands of the market- place. This sensitivity to market behavior in the face of limited re- sources includes the drive to make more activities self-supporting, to curtail the provision of free service, and to shed services that the pri- vate sector can provide better. A critical issue in this transfer of responsibility and fiscal resources from the center to subnational governments is how to increase the access of subnational governments to financial markets, broadly de- fined as the banking system and the securities markets. The word markets implies a system with a variety of borrowers and lenders and with credit allocation based on pricing decisions that balance supply and demand. It also implies an array of alternatives for accessing cap- ital funds. Accordingly, the development of financial markets is an important objective for developing economies. As economies devel- op and financial markets mature, markets are expected to evolve to bring together subnational needs for investment capital and the sup- ply of funds. Will that happen? xxix xxx Executive Summary This book explores markets for subnational government debt and the successes and failures that emerging and transitioning economies have experienced in promoting the development of such markets over the past decade. It assesses these experiences and extracts lessons to inform future efforts to establish and strengthen local gov- ernment access to credit and improve the chances for success. The book has six parts. The first five parts establish an analytical framework for studying the way subnational government credit mar- kets function, the kinds of credit instruments and security arrange- ments available, and the main participants in the market. From this analysis, much of which deals with the technical design of transac- tions and the attributes of debt instruments, come a number of find- ings that form the basis for policy guidelines. Part six is a series of country-based studies that review the experiences of building mar- kets for local government debt. These studies show that a large vari- ety of economic and institutional settings influence the nature and extent of local government borrowing, from large-scale borrowing in international markets to building more effective state-sponsored in- termediaries. The Analytical Framework The analytical framework examines the components of the supply of and demand for subnational debt, noting several perspectives from which to assess the feasibility and desirability of introducing local government securities to credit markets. Part 1. Political, Legal, and Financial Framework Credit needs and market structures vary greatly and depend on the political, fiscal, financial, and legal settings in which they are embed- ded. Chapters 2 and 3 examine the tensions and difficulties caused by the devolution of fiscal systems and the risks associated with the decentralization of borrowing decisions. They discuss the needs for a hard budget constraint, managerial capacity, and transparency and ways to achieve them to promote effective markets. A description of Executive Summary xxxi the levels of financial market development and the frictions in creat- ing new financial markets and greater competition among credit providers sets the scene for later chapters. Also explored are legal sys- tems and their ability to support the operation of markets and to un- derstand the contracts they must enforce. Part 2. Borrowing Instruments and Restrictions on Their Use Borrowing is at heart an economic activity played out in financial markets. It is technical, with its own nomenclature and methods of analysis that revolve around balancing risk and rewards. Chapter 4 discusses the characteristics of potential subnational borrowers and what qualifies them as candidates. It walks through the wide variety of subnational government structures and conditions, noting that many localities are too poor and too small to have the need or the re- sources to borrow in markets. In many other cases both the need and resources are there, but inexperienced and immature governments require help to become creditworthy borrowers, a subject that reemerges later. The discussion concludes that, while needs and ca- pacities differ, it is best not to prejudge by overly restrictive classifica- tion systems. Where possible and with appropriate help, market ac- cess can help instill fiscal discipline and responsible behavior. Chapters 5, 6, and 7 lay the foundations for the kinds of borrow- ing that may be undertaken by subnational governments and ex- plore the nature of the resources that can be pledged to repay indebt- edness. They examine the variety of borrowing instruments, their technical design (maturity, interest payment schemes), and the methods by which they are offered to the market. Markets with suffi- cient competition and transparency should be allowed a good deal of flexibility in setting the parameters of individual transactions within a general framework of widely shared rules. The corollary is that the more developed a market is, the more it can be relied on to enforce the "rules" for subnational borrowing. Chapter 7 also turns to authorization and approval of subna- tional debt and the limitations that may be placed on its use and magnitude. The chapter examines the legal and market issues re- xxxii Executive Summary lated to the various types of security that may be pledged. A review of the wide variety of restrictions and possible pledges argues for avoiding ill-designed, overly restrictive cures that can cause more harm than good. To ensure that transparent procedures and overarching prudential restrictions are as useful as possible, there should be regularized rules that leave room for market-based determinations. Furthermore, no- tions of security must reflect the needs of financial markets, includ- ing sure and speedy remedies when transactions go awry. While comprehensive borrowing laws are a desirable end result, their con- struction should be incremental, occurring as borrowing needs arise and loan contracts are drafted. While not everything needs to be in place before markets can operate, participants should be mindful of elements that are still a work in progress. Part 3. Characteristics of Financial Market Regulation and Disclosure Chapters 8 and 9 consider the structure, operation, and regulation of the domestic financial markets in which subnational governments seek to borrow. Domestic credit markets are often small, have few participants, and are easily overwhelmed by the demands of the sov- ereign and the banking system. A recurring theme is the difficulty that markets have in matching governments' long-term borrowing needs with the limited investible funds and the short time horizons of investors. However, the types of investors that may be interested in purchasing subnational government debt are growing in rank and importance. Also examined are the links between domestic financial markets and international markets. The emerging bond markets, recently viewed as a promising alternative source of funds, languished after the world monetary crises of the late 1990s and subsequent econom- ic slowdown. Those markets, while sidelined, continue to have enor- mous potential capacity. However, until the international monetary system demonstrates greater stability and financial markets improve, the ability to tap that capacity is constrained by strong aversion to the risk found in emerging market credits. Executive Summary xxxiii Part 4. Evaluating, Monitoring, and Assisting Subnational Governments The operation of financial markets depends on assessments of the creditworthiness of market participants. Only in that way can risks be judged and offset by adequate rewards. As chapter 10 details, a major concern in emerging markets is assessing the operations and financial conditions of subnational government borrowers as a basis for assess- ing risk to investors. The archetype of such assessment, or credit analysis, is the opinions of rating agencies. Their practices and the in- fluence of their opinions are examined in chapter 8. Relying on credit ratings presents major issues for small and emerging financial markets where skills, resources, and markets for opinions are limited. Still, the value of the rating process in enforcing disciplined behavior can be great. The chapter also briefly examines the use of private credit en- hancements, in particular, the use of bond insurance. Because of the uncertainties in emerging markets, the dominant private sector firms have shown little interest; however, some important homegrown ap- plications have been found in emerging markets. Chapter 11 discusses how surveillance and analysis of credits by markets can complement and benefit from the monitoring of subna- tional governments by higher level governments. The information produced in a good monitoring scheme is useful to market and gov- ernment officials at all levels. In addition to illuminating how well governments are performing, such information is critical for enforc- ing prudential regulations. In practice, however, periodic financial reporting is a weak spot, often because it is not considered essential. Efforts to regularize and harmonize reporting and improve its con- tent are indispensable to market development. A related area is what the national government does when faced with local financial emergencies. Devising techniques that allow in- tervention to protect vital services, while seeing that weak budgeting and risky behavior by subnational governments are punished, is a difficult but necessary part of enforcing a hard budget constraint. A clear process for handling such circumstances and restoring local fi- nancial health reduces risk to markets but softens the budget con- xxxiv Executive Summary straint. Similarly, passing too much of the risk to the creditor can sti- fle market growth. Finding the right balance is the issue. Chapter 12 focuses on how to assist subnational governments in gaining access to credit markets. The process faces a number of pit- falls and limitations, not only among prospective borrowers but in the markets themselves. The design of credit assistance programs must deal with the fact that there are strong twin traditions of cur- tailing subnational government access to private credit markets and of meeting capital needs through national government grants and concessionary lending programs. The long-term loan capital avail- able for these programs typically is provided by multilateral and bi- lateral donor organizations through on-lending programs adminis- tered by central government agencies and guaranteed by the sovereign. While such competition makes private market develop- ment more costly and difficult, subnational governments will need to find a route if they are to meet large and growing needs for capital and to lessen obvious dependencies. Credit assistance to subnational governments covers a wide variety of possible aids, ranging from technical assistance and credit enhancements to specialized interme- diaries and direct lending programs. Depending on the creditworthi- ness and managerial capacity of the subnational government and the nature and depth of the financial markets, each technique has its ad- vantages and drawbacks. A promising technique is the use of a finan- cial intermediary that combines the borrowing of smaller govern- ments into larger issues that afford economies of scale while exposing the underlying borrowers to market-oriented requirements. Another is the use of credit enhancements to offer assurance to lenders while schooling borrowers in ways to improve their credit stature. A major challenge in programs that directly extend credit is the need to avoid making subnational borrowers captives of such pro- grams and to encourage borrowing in domestic markets where possi- ble. Part of the solution is to integrate grant and lending programs in ways that not only encourage but also reward governments for achieving credit market access. Chapter 11 reflects on ways to extend Executive Summary xxxv the maturity of debt in markets that are unwilling or unable to lend for the long term. Part 5. Policy Guidelines Chapter 13 provides policy guidance based on experience, both good and bad, with subnational credit markets: · While there may be no one "right way" of developing subna- tional credit markets (due to the enormous variety of circum- stances and structures), there are ways of achieving that end where it is both desired and possible. · Subnational borrowing appears to have clear positive effects on credit markets. While accessing credit markets imposes burdens and risks, exposure of subnational governments to the market's appraisal of transactions, demands for information, and re- quirements for budgetary discipline is beneficial and is an im- portant component of responsible self-governance. While not all political and economic systems are capable of supporting a market for subnational debt, emulating the required behavior and laying the foundations for a market flowering when condi- tions permit are worthwhile activities. · Weak, unstable, and corrupt central governments undermine the ability of subnational governments to achieve good credit ratings and the ability of financial markets to function fairly and efficiently. Instability in international financial markets renders them an unreliable, "fair-weather" source of funds, at least in the near term. Thus building viable domestic financial markets is an immediate task--one to which subnational gov- ernments may contribute if they are fiscally stable. · Subnational governments could be given broader powers, the more open and competitive the financial markets they are to enter and the more stable their fiscal circumstances. Regulatory schemes for subnational borrowing need to have prudential limits that are clearly stated, well-monitored, and enforceable. Good information systems are key components of success. xxxvi Executive Summary · In line with the development of markets, subnational govern- ments need a clearly stated range of types of security they can pledge. Intergovernmental transfers and shared taxes often pre- dominate as sources of funds. Especially in fiscal systems that want subnational self-determination as well as the economies of central collection of revenues, the ability to use and pledge these funds is vital to providing adequate security to investors. Where all subnational revenues are consumed by expenditures for vital services, subnational governments should not be borrowing. · Information and subnational accountability are key factors in the effective operation of markets. Assessment of risk, crucial for determining the cost of capital, requires reliable, complete, and timely information. Subnational governments need to de- velop the capacity to report and manage their affairs in plain view. Without the discipline of the hard budget constraint, markets have little reason to distinguish among credits, and the rationale for market allocation of resources is lost. · Credit assistance should be used surgically, with supporting ef- forts to build subnational skills to make financial decisions. Credit assistance should help subnational governments tap into private credit markets. The integration of grants and conces- sional lending, with access to conventional markets essential to avoid undermining conventional markets. For small borrowers and shallow markets, either bank-centered or specialized lend- ing intermediaries hold promise as ways to take advantage of economies of scale while preserving market-driven behavior and constraints. Part 6. Country Case Studies Part 6 consists of 18 country studies on subnational borrowing. The studies vary in approach, but the shared objective is to examine recent experience with subnational borrowing and to assess what has worked and what has not and the reasons for the successes and failures. The studies illustrate the wide range of government and market settings. Executive Summary xxxvii Latin America and the Caribbean Latin America has the highest volume of subnational borrowing in private markets of developing regions, but with borrowing highly concentrated among large provinces and major cities. Much of the credit has been used to fund accumulated operating deficits. Through the years, huge central government bailouts were engi- neered to avoid the collapse of the debtor governments. Accumulat- ed debt reached such great proportions in Argentina and Brazil, as it had in Mexico before them, that its continued financing has been a major source of economic destabilization. Argentina and Brazil both had some early success in the sale of subnational debt in the emerging international bond markets. How- ever, recent defaults have largely closed these markets to them. The excesses by a few large borrowers have precluded many smaller ones from enjoying market access, as national governments tightened the regulatory leash. A legacy of bad debt behavior continues to plague the weakened domestic markets. In contrast to Argentina and Brazil, Colombia essentially used a market-based mechanism to impose limits on local borrowing, al- lowing it to continue in a controlled environment. Colombia made a significant shift--though with restrictions--toward decentralization in the 1990s. Initially, borrowing restrictions were lax because of leg- islated mandates to increase central transfers, and subnational bor- rowing doubled relative to GDP. The transfers contributed to grow- ing fiscal deficits of the central government. In 2001 the central government implemented new laws to streamline the intergovern- mental transfer regime and free up extra revenues to address imbal- ances. The new laws require approval from the Ministry of Finance for additional debt and link borrowing controls to the fiscal health of local governments. Thus strong central control curbed an earlier ac- celeration in subnational borrowing. Surprisingly, continuing defi- ciencies in the regulatory framework have not led to widespread fis- cal difficulties, though decentralization and mandated spending have continued to strain fiscal balances. xxxviii Executive Summary Mexico, meanwhile, has taken a bold initiative to resurrect its previ- ously troubled subnational borrowing by requiring that bank loans and bond issues be rated by internationally recognized rating agencies. All these countries have large vertical imbalances, with subnational governments highly dependent on central government collection and transfer of revenues. However, even among the wreckage of the Argen- tine system, some positive lessons have emerged on the use of trustees and the structuring of loans to lessen the risk of devaluation. Sub-Saharan Africa Sub-Saharan Africa exhibits great contrasts in subnational govern- ment borrowing, with major challenges for the future. Both South Africa and Zimbabwe have had relatively sophisticated financial sys- tems and active municipal bond markets. In South Africa prior to 1994 municipal borrowing was largely limited to relatively well-off, historically "white" municipalities and was effectively underwritten by central government. From the late 1990s to the present the central government has undertaken a large number of structural, institutional, and policy reforms in subnation- al government. The ongoing change and uncertainty this introduced into the municipal sector, coupled with ongoing problems of bud- getary and financial management in many subnational govern- ments, have created a climate in which municipal lending has stag- nated. Notwithstanding the success of certain private sector initiatives targeted at improving access to the markets (in particular, the Infrastructure Finance Corporation of South Africa), private sec- tor lending to municipalities has declined while public sector lend- ing has expanded. As of mid-2003, with crucial aspects of the legal and regulatory framework governing municipal finances and munic- ipal borrowing still to be enacted, the future of the private municipal debt market--both direct bond financing and intermediated debt-- remains unclear. This lack of clarity exists despite the sophistication and liquidity of South Africa's capital markets and significant poten- tial demand from subnational governments that badly need funds for investment in local infrastructure. Executive Summary xxxix Zimbabwe illustrates a situation in which a small municipal credit market was created and sustained artificially for many years by cen- tral government policies that favored--even compelled--institution- al investment in subnational lending and, accordingly, entrenched moral hazard, with predictable results. The policy issues surrounding this market and attempts to reform it have been dwarfed by the im- pact of the larger political and economic crisis that has consumed Zimbabwe for the last few years. Until this nationwide crisis is re- solved, it is most unlikely that borrowing will become an effective, accessible form of financing for subnational authorities in Zimbabwe or that any initiative to modernize subnational borrowing practices will have a reasonable chance of success. Middle East and North Africa North Africa, as represented in the Tunisia and Morocco case studies, presents a different approach to the issues of subnational credit mar- kets. Both countries have continued in a highly centralized political system where subnational financial markets and subnational govern- ments are only slowly gathering new powers. With limited banking systems and nascent financial markets, the focus has been on on- lending activities, financed by donor loan programs. Meanwhile, most major infrastructure spending remains a central government responsibility, and subnational capital needs tend to be for small pro- jects. The institutional framework needed to make subnational credit market access a reality remains to be developed. The next stage of de- velopment has been to propose the use of financial intermediaries that would access markets on behalf of smaller subnational govern- ments. In this case, the specialized lending institution would play a twin development role: instilling more fiscal discipline and account- ability in subnational governments while providing a new invest- ment outlet for markets. Asia Asian nations that are emerging from highly centralized government structures reflect very different stages of economic development. xl Executive Summary China, Indonesia, the Republic of Korea, and the Philippines have di- verse experiences with subnational borrowing. The People's Republic of China presents something of an enigma: a highly centralized state that is loosely organized, with extreme variations in subnational fiscal capacity and high levels of invest- ment by companies owned by subnational governments that them- selves cannot borrow. Although China is a unitary state, it has de- volved a great deal of spending responsibility to its subnational units, which are both legion in number and, at the provincial level, as large in population as many countries. While the subnational gov- ernments are precluded from borrowing directly using their own credits, they effectively borrow through special-purpose vehicles, which are wholly owned companies that have their own revenues and often supply infrastructure needs on a quasi-commercial basis. Rationalizing the activities of these "off­balance sheet" borrowers, which often have to rely on borrowing from state-owned banks, is a major challenge the country faces as it carefully enters into a regime of financial markets--and the world's financial markets. Before the financial crisis of 1997 Indonesia had embarked on a plan to offer local water utility and housing bonds in its small but rel- atively active domestic bond market. Following the East Asia financial crisis, the government structure has undergone radical reform, and a rapid and pervasive devolution of government power is under way. Indonesia's difficult economic conditions, a beleaguered banking sys- tem, and political turbulence have delayed resumption of efforts to steer subnational governments toward credit markets. While access is being contemplated, many subnational governments have defaulted on outstanding loans from the national development fund, and all subnational borrowing has ceased. Resolution of defaults and restora- tion of domestic financial markets are needed before subnational bor- rowing resumes. How that borrowing should be accomplished within the new political framework is a topic of considerable debate. In the Republic of Korea there has been considerable borrowing by subnational governments, largely steered, if not controlled, by the central government. Subnational governments that conform to a na- Executive Summary xli tional planning framework are permitted to borrow on favorable terms from national entities. Korea also has employed a form of "forced lending" that requires participation in loans by private firms that undertake certain nationally franchised activities. Borrowing in private financial markets is growing, however, including bond sales. The Korean bond market was badly shaken by the Asian crisis of the late 1990s and has taken time to recover its footing. As it does, sub- national borrowers are likely to play an increasing role. In the Philippines a small but active municipal bond market oper- ates alongside a subnational credit system dominated by two govern- ment-owned banks and two municipal development funds. Recover- ing from a period of massive defaults that followed the Marcos regime, subnational governments have proved the most creditwor- thy borrowers in the Philippine economy, thanks to the large and steady transfers from the central government and the ability of state- owned institutions to intercept the transfers. A specialized bond in- surance company was formed to increase subnational government access to private capital markets, and all recent municipal bond is- sues have used its coverage to enhance creditworthiness. Meanwhile, government financial institutions, while still benefiting from man- dates that subnational governments use them as depositories, see their dominance challenged. South Asia India's federal system has undergone decentralization over the past decade, giving constitutional recognition to subnational govern- ments. The federal relationship between the center and the states has been under stress, with significant vertical imbalances, and subna- tional governments often find themselves without resources. Nonetheless, there has been progress toward a more market-oriented subnational government borrowing regime. A municipal bond mar- ket has emerged, and concessionary development fund loans have been revamped to reflect market conditions and discipline. Tamil Nadu has converted its development fund from a state-administered loan fund to a public-private fund. Acting as an intermediary, this xlii Executive Summary hybrid fund raises capital in domestic financial markets and comple- ments its lending activities with technical assistance. Europe and Central Asia The transitional economies of Europe and Central Asia have ad- dressed subnational government credit market access in a variety of ways, with several false starts and changes along the way. Bulgaria's experience is representative of the more slowly evolving situations of fiscal decentralization found in South-Central Europe (the Balkans). The case study focuses on borrowing undertaken by the city of Sophia, which successfully executed loans in the euro- market and managed to maintain a credible budget under trying cir- cumstances. Other transitional countries, with fewer resources and slower to set aside the old systems, are likely to follow models of bank lending and euro-market access by specialized intermediaries. The Czech Republic, Hungary, and Poland followed more conserv- ative paths in developing markets for subnational debt and avoided large-scale credit problems at the subnational government level. In each country a few major cities have successfully followed the sover- eign government into international bond markets, and a few have entered domestic debt markets. After initial flurries of activity, how- ever, most subnational government credit needs have been met by a combination of bank loans and development funds. Moreover, sub- national capital demands have been reduced through tightened legal limitations on borrowing and the availability of receipts from the sale of privatized assets. Growth in subnational government borrow- ing in credit markets also has been retarded by weak economies, the uncertainties created by the unsteady devolution of fiscal powers to subnational governments, and the reluctance of these governments to borrow. Recently, many of their capital needs have been met by concessional loans and grants. European Union structural funds will make borrowing easier for subnational governments as the inhibit- ing factors mentioned above lose potency. There have been several bond sales in these three countries' do- mestic markets, but the domestic bond markets have been slow to Executive Summary xliii grow. This is due in large part to high interest rates, the small num- ber of long-term investors, and the difficulties of attracting investors to small, soft-currency investments. The markets are generally illiquid, have few participants, and are preoccupied with financing sovereign debt. However, with these countries' pending accession to the European Union, domestic financial markets have begun to reflect the convergence and the broader European securities markets. The amply funded and far-reaching grant and loan activities of the European Bank for Reconstruction and Development now strongly influence the development of subnational government credit as several transitioning countries set their sights on meeting standards required to enter the European Union. While domestic bond markets will serve needs at the margin, the focus appears to be on bank lending and accessing the Eurobond market. As part of the process, it is likely that a role will emerge for specialized lending institutions ("bond banks") that can access euro credit markets to meet the needs of smaller borrowers. The Russian Federation has had the most dramatic (and chaotic) ex- perience with subnational borrowing, reflecting the fragmentation that occurred after the dissolution of the Soviet Union. In the mid- 1990s, Russian regions and major cities issued large amounts of debt in both domestic and international financial markets. Massive defaults of subnational debt followed the 1998 financial crisis, as payments of shared taxes and grants from the central government evaporated and the national government was forced into default and devaluation. However, the two major cities of Moscow and St. Petersburg managed to avoid defaulting on their debts and have continued to enjoy some limited access to much shrunken markets. The case study describes St. Petersburg's struggle to preserve its creditworthiness in the midst of the turbulent conditions at the turn of the twenty-first century. Chapter 1 Introduction The decentralization1 of governments throughout the world has brought new prerogatives and responsibilities to subnational governments as ser- vice providers to their local constituents. Part of a larger move toward greater democratization of government, reliance on markets, private provi- sion of many activities formerly carried out by governments,2 and global- ization of commerce and finance, decentralization also has encompassed a desire to use private capital markets as allocators of credit. In developing countries the twin tasks of building more dispersed and de- mocratic governments and opening economies to freer markets and greater private ownership have been attempted in tandem--and have proved a diffi- cult undertaking. A reduction in barriers to the movement of capital and goods has been a nearly universal objective.3 However, implementation of the required reforms has meant tough competition for domestic industries and in- creasing constraints on the fiscal and monetary policies of national govern- ments. In the face of economic slowdowns and unstable financial markets, many emerging and developing economies have found privatization and the opening up of their economies to be painful and unpopular. The steep price and uncertain benefits of joining global markets have their critics.4 Subnational governments, for their part, are being required to do more things, to do them more efficiently, and to be more self-reliant in raising re- sources.5 At the same time devolution and hard-pressed budgets have con- strained the ability of central governments to provide for the needs of sub- national governments. After years of neglect and with expectations rising, the needs for infrastructure are particularly daunting. The enormous fund- ing requirements cannot be met either practically or equitably without long-term investment. International lending and grant-giving institutions, another traditional source of funds, are also limited in their resources and restricted by rules and customary practice to dealing only through sover- eign governments. 1 2 Subnational Capital Markets in Developing Countries Nevertheless, the increased need of subnational governments to mobi- lize private capital to meet their infrastructure requirements is seen as a positive development in prompting the movement toward greater democ- ratization and decentralization. The day-to-day scrutiny of government op- erations by credit markets helps to reinforce transparency and encourage efficiency and prudence. However, the markets' rewards are not without risks. Efficient functioning of markets requires rules of procedure for both buyers and sellers and an overarching societal agreement on what is for sale and what is not. Realizing the Promise of Access to Financial Markets This book examines the experience of subnational governments in access- ing the credit marketplace, seeking lessons about how to realize the promis- es of credit market access while avoiding the pitfalls. The focus is on coun- tries that are either "developing," in the sense that they are attempting to attain more modern and productive economies, or "transitioning," in the sense that they are moving from a highly centralized and large government sector to a more decentralized and market-based one.6 Countries often have characteristics of both groups. The commonality is that they are rela- tively poor in the material sense, have a large central government sector, and have underdeveloped financial markets. Countries differ, however, in the role that subnational governments have in financial markets and in the nature of their financial markets. Whatever the goals of greater autonomy and capacity at the subnational level, subnational governments vary greatly in political power and deci- sionmaking authority, in part reflecting differences among unitary states, hierarchical federal states, and governance systems that recognize separate spheres for each level. These differences are embedded in constitutions and legal systems that condition the degree to which subnational governments are free to act and to control the resources with which to act. Some systems have been devolved at least on paper for many years (as in several Latin America countries), while others are starting from scratch (such as the transitioning countries of Eastern and Central Europe and sev- eral Asian counties). Institutional structure and history are important: many of the harsher lessons about balancing subnational debt finance7 with national macroeconomic stability have to do with the misadventures of poorly designed governmental systems and long-standing problems of fiscal mismatches, political corruption, and mismanagement. Introduction 3 The nature of credit markets also varies. Few developing countries have active securities markets. Where the markets do exist, subnational govern- ments have rarely been participants. Significant domestic markets for sub- national debt have begun to emerge in some countries, but the passage has been neither easy nor swift. In a few areas, such as South America, some subnational governments have been active borrowers, but results have been uneven. In most other parts of the world, such as South and South- east Asia, development at the subnational level has been slower. In the transitioning countries of Eastern and Central Europe, which have had sub- stantial government restructuring and credit market development, subna- tional borrowing from private sources has risen slowly.8 Several larger and better-known subnational governments have bor- rowed in foreign markets, an avenue that appeared promising until the twin crises of foreign currency collapse and global economic slowdown that occurred in the late 1990s. While restoration of world capital markets will help governments meet their financing needs, the more immediate is- sue for the vast majority of subnational governments in developing coun- tries is how to raise funds in domestic capital markets. Presenting Governments with Market-Based Alternatives Several vantage points are possible when surveying subnational govern- ment access to financial markets. One is that of the "macro-level" policy- maker determining how best to fit a municipal borrowing component into domestic and international credit markets, given a host of other pol- icy constraints and objectives. A second is that of the prospective subna- tional government borrower, intent on achieving as much flexibility as possible in financing decisions and on securing capital on the best possi- ble terms. A third is that of the lender or investor who needs information to assess relative rewards and risks (including remedies in case of trouble) and assurances that the rules of the game will not be violated or changed arbitrarily. While these three views are not always consistent, the ultimate objective is the same: to improve subnational governments' access to private credit markets in ways that are consistent with the overall fiscal health of govern- ment and the viability of the domestic financial markets. However attrac- tive the rhetoric, achieving the objective requires making choices and tak- ing risks.9 Without being prescriptive about a best approach in all cases, this book starts from the premise that subnational control and decisionmaking 4 Subnational Capital Markets in Developing Countries are desirable outcomes that should be cultivated and encouraged. While ac- knowledging the need for rules, the book also argues for maximizing com- petition among private sector financial options at the subnational govern- ment level, where possible and prudent. This reflects a belief that presenting governments with market-based alternatives is invariably better than any single "my way or no way" of doing things. The book further calls for a liberal view of the risks that subnational governments should be al- lowed to run; this liberal view entails a presumption that these govern- ments assume the risks at their own peril but that they do so in capital mar- kets that are fair and reasonably efficient.10 The terms credit market and capital market are used broadly. In many emerging economies the banking system is the leading provider of cred- it--and the banking system too is likely to be undergoing transforma- tion. The reliance on banks may be either a substitute for or a precursor to a functioning domestic securities market in subnational government obligations. Where possible, a securities market should be seen as a desirable means of obtaining long-term capital. Appealing to the rapidly growing numbers of nonbank institutional investors, securities markets were developing rapidly until repeated crises struck emerging bond and equities markets in the late 1990s. Activities in these markets slowed abruptly in reaction to unsettled conditions in global financial markets, a series of currency crises, and the general slowing of the world economy. The arguments in favor of securities markets do not deny the critical importance of the banking sys- tem as the bulwark of the financial system. To develop and thrive, markets for longer term debt require strong banking systems on which they can de- pend for a reliable system of payments. In many countries, the practical outcome may be to promote competition among institutions that lend to subnational governments or even to find ways to reproduce the benefits of competition. Immature capital markets should not deter efforts to create structures that can reproduce such benefits. Setting Out the Analytical Framework Thus the analytical framework for this study rests on the principle that a subnational government securities market is desirable and that subnational borrowing will be dictated largely by the operation of the market, working within a framework of rules necessary to keep it a free and efficient allocator among competing uses. Many conditions need to be met, but four are key: Introduction 5 · Subnational governments borrow of their own volition and rely on their own resources for security and repayment of debt. · Capital markets are free of excessive restrictions--with an arm's- length relationship between government and markets and banks-- and allocate resources on the basis of risk and reward. · The market has full access to the information required to assess the fi- nancial condition of borrowers and to determine risk and reward. · Subnational borrowing is subject to appropriate oversight by the cen- tral government before it is made available. The central government plays a supportive role, intervening only when well-established rules of borrowing are flouted or subnational government mismanage- ment threatens fiscal crisis. The ability--and desire--of governments and financial markets to achieve these conditions depend on several related policy and technical is- sues. A variety of government structures, schemes of devolution, and prob- lems of macroeconomic stability influence decisions about the nature and feasibility of subnational government borrowing (chapter 2), as do a coun- try's legal systems and financial market structures (chapter 3). Thus it is im- portant to understand the political, economic, and legal environments in which subnational government borrowing occurs. In analyzing options and possibilities for markets in subnational obliga- tions, some key questions need to be asked. On the borrowers' side, impor- tant issues are credit capacity, borrowing powers, and regulation within the government sector (chapters 4­7). What types of debt security are available? What debt instruments are to be used? What types of subsovereign govern- ments are good candidates to borrow? How is subnational debt to be autho- rized? What limitations should be placed on borrowing? What is the role of monitoring and oversight? What are the remedies in case of fiscal problems? On the investors' side, regulation, investor needs, and the operation of financial markets are important concerns (chapters 8 to 12). What is the fi- nancial market structure? Who are the potential investors, and what are their investment objectives and constraints? What is the regulatory frame- work of the marketplace? What is the role of disclosure, and how is it ac- complished? What is the role of credit analysis and credit ratings? How can the private sector mitigate risks? How should credit assistance be provided to comport with the market? These questions represent economists' familiar separation between the demand for loanable funds by the subnational government sector and the 6 Subnational Capital Markets in Developing Countries supply of funds by private suppliers of credit. This separation is conceptual- ly useful, but in practice matters are more complicated. The links between subnational governments and credit markets, even in countries with only a nascent financial sector, are diffuse and complex. Few actors are interested in only one financial relationship, such as borrowing, and the relationships are often more than financial. Nevertheless, the point of departure is that subnational governments are increasingly important economic actors, and the stage on which decisions are made is increasingly that of the market. All these issues are explored in the first part of the book. The second part is a series of case studies that discuss recent experiences in 18 developing and transitional countries. The case studies range from general reviews of subnational credit access on a countrywide basis to more detailed discussions of debt transactions and lending. The case studies pre- sent a rich variety of experiences, good and bad, with subnational govern- ment borrowing and offer lessons about which approaches have been suc- cessful and why. They also illustrate experiences that have been disappointing and attempt to explain why. Notes 1. The terms devolution, decentralization, and deconcentration are frequent- ly used synonymously to describe the process of giving more decisionmak- ing power to subnational governments. In practice, this process varies greatly, as does the degree to which fiscal powers are devolved. In many cases, decentralization has meant a dispersal of spending and taxing pow- ers that remain tightly controlled by the central government. In other cas- es, localities have been given a full range of taxing and spending powers, including the power to borrow. The terms are used interchangeably unless otherwise noted. 2. The conflict between government control and the freeing of markets was a common theme throughout the twentieth century and is treated on a global scale in Yergin and Stanislaw (1999). 3. This is the often-cited "Washington Consensus" for liberalizing trade and international financial flows. It has been actively promoted by the In- ternational Monetary Fund, the World Bank, the World Trade Organiza- tion, and the U.S. Treasury. 4. See, among others, Stiglitz (2002). 5. This book often uses the terms local, municipal, subnational, and sub- sovereign interchangeably, unless dealing in a specific context. The terms Introduction 7 can also encompass states, regions, provinces, and other subnational gov- ernments, depending on context. 6. The terms developing, emerging, and transitioning (transitioning is typi- cally applied to Eastern and Central European states that are changing from communist to democratic regimes) are used interchangeably. 7. Debt, loans, and bonds are used interchangeably to refer to subnation- al government debt finance, depending on context. 8. Noel (2000) notes that subnational debt markets have grown rapidly in Argentina and Brazil (representing as much as 5 percent of GDP), but they remain "embryonic" in most emerging and transitioning economies, including the relatively advanced states of Central Europe (p. 1). He fore- sees a clash between the rapidly rising needs for infrastructure finance and the limited development of the domestic markets. 9. Noel (2000) sees the movement as being from either a nonexistent or a monopoly market for subnational government debt to one of active com- petition among alternative sources of private capital (from a closed to an open system of financing investment needs). It is a movement filled with risks and tensions among key stakeholders: the national government, pri- vate investor institutions, and subnational governments. This is usually a simplification since there are competing interests within the sectors. The private sector has tensions among commercial banks, other financial insti- tutions, and securities markets. The central government may have compet- ing interests among agencies (the treasury and the central bank, for in- stance) and competition among the local governments themselves, which once were agents for the central government and now are striving to be- come more independent. 10. A source of continuing concern is that of the moral hazard that local governments present when they enter capital markets with an implied sov- ereign guarantee that the national government will be compelled to bail them out if things go wrong and they cannot pay their debts as promised. There are a large number of assumptions surrounding the implied exis- tence of such guarantees and quite a bit of history as well. The assumption in this book is that sovereigns as part of the move toward devolution are re- luctant to make such guarantees and are inclined to have their local gov- ernments face a "hard" budget constraint. Part I Political, Legal, and Financial Framework John Petersen and Mila Freire Chapter 2 Fiscal Devolution Devolution--the granting of greater political and fiscal responsibility and power to subnational units of government and the performance of more government functions at the subnational level--has been in full swing worldwide for the last decade. A 1994 World Bank report noted that of the 75 developing countries with populations greater than 5 million, all but 12 were in the process of transferring fiscal power from the center to subna- tional governments (Dillinger 1994). In the once highly centralized com- munist states with virtually no subnational autonomy, devolution has been a universal phenomenon. In some countries subnational govern- ments have long existed but frequently only as agents of the central or provincial government and with little real authority or financial autono- my. In other countries, a history of tension between competing "sover- eigns" at the center and in the regions has left a legacy of imperfect and damaged intergovernmental relationships. Principle of Subsidiarity There are well-rehearsed economic and political arguments in favor of de- volution that appeal to the efficiency and desirability of grassroots deci- sionmaking and accountability. To the economist the subnational govern- ment's greater knowledge of subnational needs strengthens the links between tax revenues and spending benefits that accrue to subnational tax- payers. Subnational authorities can respond more readily and effectively to local conditions, resulting in improved delivery of government services. Bringing expenditure assignments closer to revenue sources enhances ac- countability and transparency. Political arguments often adhere to the principle of subsidiarity, that is, in a democracy, the lowest level of govern- ment that can determine and effectively meet the needs of its constituency is the most appropriate structure of government. 11 12 Subnational Capital Markets in Developing Countries The details of the extent and effectiveness of devolution are specific to each country. The process can be complex and filled with uncertainty. Making the transition from a highly centralized system of governance to a more localized one is a serious task, subject to interruptions and miscalcu- lations along the way.1 In the end, the degree of devolution depends on the degree of de jure fiscal autonomy and de facto willingness and ability to tap resources. Countries vary greatly in both respects. Borrowing and Devolution from the Subnational Perspective Subnational access to credit markets usually derives from devolution. Bor- rowing becomes a critical issue of local initiative only when there is a move toward localized delivery of services requiring capital investments that will not be met by central government resources. Devolution is of great practi- cal consequence for credit markets and for how subnational governments access those markets. If effective, devolution places decisionmaking at the subnational level and erodes what has often been the de facto monopoly of the central government over subnational capital financing decisions, in- cluding the use of credit. With decentralization of finances and financial decisionmaking, investors and lenders care how well subnational governments are managed because they have money at risk, and their scrutiny drives greater transparency and ef- ficiency at the subnational level. However, subnational governments first must have the ability to raise and use resources and to make binding commit- ments that are politically and legally sustainable. For many countries, this constitutes a huge change in perspective and in the balance of political power. Measuring Fiscal Decentralization While the idea of devolving spending, revenue-raising, and borrowing deci- sions from central to local and regional governments seems conceptually clear, the process has proven cumbersome, contentious, complex, and confusing.2 The shifting down of spending responsibilities often has not been accompa- nied by a corresponding shifting down of resources, so that subnational gov- ernments have been faced with both mandated spending requirements over which they have little influence and weak and constrained revenue systems. Most devolutions have involved large shared-tax and fiscal transfer pro- grams that are not tied to specific spending programs. Furthermore, such fi- nancial management practices as public deposit management, investments, and borrowing procedures have been slow to adjust to the new devolution- Fiscal Devolution 13 ary regime, restricting the financial decisionmaking ability of subnational governments and their day to day management and planning (see box 2.1). In India, for example, State Finance Commissions are responsible for imple- menting the devolution of financial resources to subnational governments. They regularly review the finances of subnational bodies (panchayats and municipalities) and make recommendations on the sharing and assign- ment of state government revenues and grants in aid (see the India case study, chapter 24). Quantifying the amount of subnational autonomy in a fiscal system is dif- ficult. Internationally statistics on government finance leave large gaps in un- derstanding the nature of local revenue and expenditure systems and the de- gree of autonomy that subnational governments have to make "devolved" fiscal decisions (Ebel and Yilmaz 2001). For example, systems with substantial dictated expenditures or programs of large fiscal transfers and tax sharing that are subject to discretionary change at the center do not qualify as devolution, nor do categorical grants from the central government that are restricted to specific uses. Fiscal autonomy is also effectively lessened when subnational governments cannot control either the rate or base of local taxes. The upshot of devolution in many developing and transitional coun- tries is that subnational governments are undergoing structural change and typically have restricted power to borrow and limited own-source resources for securing debt. Understandably, would-be lenders, unfamiliar with the ways of subnational government and aware of the intergovernmental tu- mult, have been cautious in their lending. Budget Constraints and Local Control Apart from a subnational government's ability to raise taxes, levy charges, and commit resources as it sees fit, effective devolution requires a "hard budget" constraint at the subnational level. A hard budget constraint means that the subnational government must live within its resources and cannot depend on the central government to cover its deficits or repay its debts. A hard budget is possible, as long as certain basic services are provided and the risks are ac- knowledged and "paid for." Assumption of the risks by those who have decid- ed to take them is an important and often delicate point in governance. Free- dom to fail is one of the liberties and consequences that accompany greater subnational government freedom and responsibility in decisionmaking. Fiscal discipline is achieved only if those taking risks and failing are made to pay the price. Activities and borrowers deemed unsuitable for pay- ing a price for mistakes may be effectively precluded from the markets ei- 14 Subnational Capital Markets in Developing Countries Box 2.1. Devolving Responsibility for Elementary School Teachers' Salaries in Romania In 2000 the Romanian government passed to municipal govern- ments the responsibility for paying public elementary school teachers' salaries; however, the salaries were set uniformly at the national level. The central government transferred revenues to the municipalities to pay the salaries, dedicating a portion of the national value added tax (VAT) for that purpose. With the new spending requirements, local government bud- gets increased substantially and their composition changed. The discretionary portion of local budgets plummeted, while the portion going to employee wages rose. The new payments significantly increased overall transfers to localities, especially earmarked revenues. The change also affects the borrowing status of Romanian mu- nicipalities. By law, the increase in current operating revenues from permanent sources increased the amount that a local gov- ernment could borrow based on total current revenues.Howev- er, the new exposure of local government budgets to paying the salaries of a large and influential employee group and the un- certain reliability of future shared revenues from the VAT proba- bly reduced the amount investors are willing to invest. By im- pairing the future fiscal flexibility of localities and their ability to pledge funds for debt payment, the change may have made borrowing more difficult. The episode indicates the limitations of legally imposed ceilings and the importance of market per- ceptions in deciding what is prudent behavior. Source: Petersen 2002. ther by fiat or by the unwillingness of investors to invest. Both developed and emerging credit markets are full of examples where certain activities and facilities are held as essential to the public sector and cannot be used as collateral to secure borrowings. Markets in developed economies have found ways to achieve sufficient security. Fiscal Devolution 15 There has been considerable concern, particularly among central gov- ernments, about the destabilizing impact of fiscal decentralization, espe- cially of excessive subnational borrowing. They worry that decentraliza- tion will permit, if not encourage, subnational governments to spend too much, forcing central governments to run deficits of their own as they bail out the local excesses. This kind of destabilizing behavior arises pri- marily in one of two largely unrelated circumstances. One is the case of federal system countries with weak fiscal coordinating power by a central government that will not or cannot impose a hard budget constraint on the subnational governments.3 Another is the case of subnational govern- ments that are the putative borrowers from an entity such as a national development fund but the borrowing decisions are effectively made by the central government, with the localities merely "signing on the dotted line," or where the localities were placed in a position of substantial moral hazard because of the nature of the program design (see the In- donesia case study, chapter 25). Another part of the devolutionary equation is the need for local control of resources that can be used to secure debt. Two problems are common. First, subnational revenue systems are often inadequate, and meeting expen- ditures mandated by the central government exhausts the budgetary re- sources. As a practical matter, even if subnational governments have poten- tially viable revenue sources and can muster the political will, the inability to raise taxes and spend funds as they wish can be a severe constraint on the ability to borrow. Second, even where localities have substantial physical as- sets, they are legally precluded from using them to secure credit. This inabil- ity to pledge physical assets has been a constraint in many countries where bank lending, in particular, is secured by asset pledges. Impact of Devolution on Subnational Finances The impact of devolution on the ability to borrow has to do with how re- sources are assigned to governments and how resources are balanced against spending responsibilities. Several factors affect the resources avail- able to subnational governments for meeting both operating needs and debt commitments. Among them are the following: · The overall size of transfers and their size relative to a subnational government's overall operating revenues. · The extent of earmarking of transfers (as opposed to being generally usable or available for debt service). 16 Subnational Capital Markets in Developing Countries · The revenue sources legally available to subnational governments, the revenue potential of those sources, and the ability of subnational gov- ernments to use revenues for general rather than specific purposes. · The flexibility subnational governments have in setting rates or charges and defining tax and charge bases. · The overall political and institutional risk to which revenue and fiscal transfer systems are subject, that is, the potential for producing changes that can disrupt subnational finances. On the spending side, a related set of factors affects subnational govern- ment creditworthiness and credit access: · The degree of discretionary spending, the size and type of mandated spending, and the impacts of mandates on the future flexibility of subnational budgets. · How specific expenditure types are funded, such as those earmarked from specific revenue sources. · The degree of flexibility a subnational government has in adjusting its budget over the economic cycle or in response to changes in local conditions. · Demographic and economic factors that determine the demand for services and the ability of localities to control or plan for them. The more that subnational governments are expected to be self-reliant in financing their activities, the more these factors count. Conversely, to the extent that subnational government borrowing is formally guaranteed by the central government (or that credit markets expect a national gov- ernment bailout of subnational government debt in the event of difficulty), the less fiscal devolution has taken place; accordingly, the less important local fiscal affairs and demonstrated discipline are to private sector lenders. Devolution, Borrowing, and Macroeconomic Stability The subnational government's desire to pursue fiscal autonomy is one side of the devolution coin. The other is the central government's need to maintain macroeconomic and fiscal balance, which implies maintaining subnational debt under limits. Central government concerns over control of the macroeconomic bal- ance stem from its need to manage the national economy and currency and Fiscal Devolution 17 so the need to have centralized monetary and fiscal policies. Decentraliza- tion of a large share of public expenditures, even when subnational govern- ments are constrained by taxation and borrowing limits, can adversely affect aggregate demand and international competitiveness, undermining nation- al stabilization policy.4 Similarly, public debt at local levels that becomes ef- fectively "monetized" can interfere with monetary policy and, by extension, hamper the central bank's effectiveness in carrying out national policy. In theory, decentralization should establish a virtuous cycle of behavior by subnational governments that helps to maintain macroeconomic stabil- ity. Bringing expenditure assignments closer to revenue sources should en- hance accountability and transparency in government actions. Underpin- ning the downward shift in responsibility is greater reliance on the benefit principle; taxpayers should pay for the public services they receive and get the services they pay for, linking taxes to the benefits provided. Taxpayers are made aware of the cost of goods and services that they consume and, as consumers, they should be more concerned about efficiency and better able to do something about it. If poorly conceived and executed, however, decentralization can imperil macroeconomic stability. Given the greater difficulty in coordinating gov- ernment actions when subnational governments enjoy greater policymak- ing autonomy, the challenge is to design a system of multilevel public fi- nances that allows the efficient provision of local services while it maintains fiscal discipline nationally and subnationally (De Mello 2000). Much of the concern is rooted in the unwillingness of the central govern- ment to let go and in the web of political relationships between the central government and its subsovereign governments. Lack of discipline and transparency may induce subnational governments to spend beyond their means, leading to higher borrowing costs because of the risk premium asso- ciated with a higher probability of default. Avoiding these problems requires that subnational governments exer- cise fiscal discipline and that their fiscal position be effectively monitored. Thus, decentralization should include either firm rules or strong incentives for prudence in debt and expenditure management. While these notions are conceptually straightforward, decentralization in practice is the product of political decisionmaking, and the required changes create winners and losers. Not surprisingly, decentralization in many countries has been plagued by confusion and compromise that undermine both the trans- parency of fiscal relationships and the fiscal discipline of the newly em- powered subnational governments. 18 Subnational Capital Markets in Developing Countries Decentralization of Responsibilities and Revenues The literature on decentralization suggests rational guidelines for the allo- cation of responsibilities across government levels, assuming the manageri- al and technical capacity needed to carry them out. While the functions to be sorted out are many, as are the questions of local or regional points of service delivery and places of tax collection, a brief summation of princi- ples is possible. Looking first at expenditures: · The central level should retain expenditures that can strongly influ- ence aggregate demand, that involve income redistribution, and that have large economies of scale or public-good characteristics on a na- tional scale; examples include national defense, interstate communi- cations, foreign policy, and research and development. Subnational governments generally assume responsibility for local activities such as local infrastructure and services. · Sharing responsibilities should be considered in the case where the activity is national in scope but implementation is more effective at the subnational level, as in the case of education or health. On the revenue side of the ledger: · Base income taxation should be kept at the central government level to facilitate efficient collection and to preserve the government's macroeconomic stabilization and redistribution functions. · Overlapping tax bases between the center and subnational levels are common in partial assignments of income tax, where subnational governments can piggyback on the national income tax by applying surcharges. · To minimize unwanted tax-induced incentives, the central govern- ment should retain mobile tax bases such as the corporate income tax. A homogeneous tax system across all subnational governments is important to discourage enterprises from moving to areas with lower corporate taxes and eliminating tax competition among regions that could erode the tax base. · The central level should receive the most unevenly distributed and fortu- itous tax bases (such as natural resources) so that redistributive policies are possible and gross power differentials are not promoted inadvertently. · Single-stage and excise taxes, such as the property tax, utility fees, and betterment tax, can be effectively assigned to the subnational lev- Fiscal Devolution 19 el because the base is immobile and there is a close link between the tax and the benefiting user.5 However, these guiding principles can collide with reality. In many cases, levels of government have tradi- tional scopes of competency unrelated to their ability to raise rev- enues, requiring the transfer of funds from one level to another in an effort to balance resources and needs. Intergovernmental Transfers Transfers from the center reflect the disparity between decentralized rev- enues and the responsibilities associated with providing certain services at a subnational level.6 While many services can be decentralized, revenue sources at the subnational level are generally inadequate to fund the ser- vices. Intergovernmental transfers, whether as a proportion of a set of cen- tral collected taxes or as grants, help fill that gap. There is a vast body of literature on intergovernmental transfers (see, for example, Bird and Vaillancourt 1998). A critical issue is the impact of trans- fers during times of fiscal or other economic difficulty. The central govern- ment may need additional revenues, but the share of them appropriated to subnational governments is fixed. Grants from the central government tend to be more discretionary than shared revenues, a feature that may cre- ate revenue uncertainty for subnational governments in volatile economic times. Thus, an inherent tension exists between the predictability that is helpful to stabilizing subnational government finances and the rigidity that may destabilize the national fiscal balance. Transfers can be important to credit market development, since they constitute a large share of available revenue and may act as security on sub- national government loans and bonds. Transfers also can be limiting. In Hungary tight fiscal policies have constrained budgetary transfers from the central government, impairing the ability of subnational governments to meet the levels and standards of service required of them. Competing claims for scarce budgetary resources have led, in particular, to large fund- ing gaps for local infrastructure investments (see the Hungary case study, chapter 29). Subnational Borrowing as a Destabilizing Element Major financial crises in Latin America in the late 1990s and in 2002 were in part a product of excessive subnational borrowing and central govern- ment assumption of subnational debt. This experience highlights the nega- tive impact of subnational debt on the national aggregate debt exposure. It 20 Subnational Capital Markets in Developing Countries also underscores the difficulty that central governments can have in moni- toring the exposure of subnational governments. In theory, competitive capital markets establish interest rates for govern- ment debt according to differences in perceived risk and in the tax and reg- ulatory benefits that holding such debt may afford. Interest rates (risk pre- mia) reflect the borrower's creditworthiness when the risk is assumed by the subnational government and not absorbed by the central government through explicit or implicit promises of bailouts or guarantees. However, even where the central government backs subnational debt, market forces may induce greater fiscal discipline at the subnational level once the debt is traded on the open market. Greater fiscal discipline can improve resource allocation, eliminate waste, and benefit the local population directly by in- creasing resources. Two key assumptions are that capital markets are com- petitive and that bondholders or the governments themselves suffer the consequences. Without the threat of "pain," discipline fails on both sides of the market. Problems in Subnational Debt Markets Fiscal discipline by subnational governments depends in large measure on their relationship with the center. How much autonomy do they have, and will the central government step in--and, if so, when--to avoid financial calamity? Relationships between central and subnational governments can give rise to the problems of adverse selection and moral hazard. Adverse selection arises when asymmetric information or misaligned in- centives lead to decisions that would have been avoided with more informa- tion or a different set of incentives. Subnational governments have an in- centive to hide negative information about their finances from potential investors. Information asymmetries, common in all markets, must be miti- gated through legislation, regulation, and institutional development. Where well enforced, securities and tax fraud laws can be powerful antidotes. Moral hazard refers to the creation of incentives that distort behavior be- cause parties are not held accountable for the risks involved in their actions. A local jurisdiction with borrowing privileges needs to maintain fiscal disci- pline to retain an adequate credit rating and satisfy creditor scrutiny. Howev- er, where an explicit or implicit central government promise exists to bail out subnational government, the costs of default are transferred to the cen- ter and neither the borrower nor the lender faces the consequences of the borrower's failures. With the penalties removed, the costs of inadequate dis- cipline disappear, so that over-lending and over-spending are "rational be- Fiscal Devolution 21 haviors" for both borrower and lender. Perceived permissive behavior at the center inverts the incentive system, making it "profitable" for subnational governments not to live up to their obligations. Worries about moral hazard stem from ambiguities in the relationships among the sovereign government, subnational governments, and potential creditors. Often, national governments have relaxed the subnational budget constraint by permitting or even encouraging excessive spending. Creditors base their investment decisions on the financial viability of the subnational government to which they lend. Because of the unstable flow of revenues and less knowledge about the creditworthiness of the subnational govern- ment, creditors often seek a sovereign guarantee. Private sector lenders and multinational institutions and bilateral lenders alike often require that subna- tional loans carry sovereign guarantees. Private lenders are understandably circumspect about the moral hazard such behavior entails. They are unwill- ing to extend nonguaranteed loans in competition against the risk-free bor- rowing that subnational governments effectively enjoy (or risk-free lending that subnational government creditors enjoy) with a sovereign guarantee. Setting a Precedent If past interventions by the central government have set precedents for fu- ture interventions, the cycle is difficult to break. Moral hazard challenges confront countries from the start of decentralization. In the early stages, when subnational authorities are not fully in control of local expenditures, the central government is expected to fill expenditure holes, as has hap- pened frequently in transitioning countries. As a result, the subnational au- thority is not held fully accountable for its expenditures. This type of moral hazard should decline once the system of revenue sharing and grants is es- tablished and subnational governments are made accountable for the ser- vices assigned to them. However, the sequencing of assigning responsibili- ties and resources and applying appropriate restraints is often defective. A central government that has a history of bailing out subnational gov- ernments sends an implicit message that it will intervene in the future. Changing this perception can be difficult, since the causes are often deeply ingrained in the political and financial systems. In theory, intervention can be designed so that the subnational government bears the costs if it de- faults and needs help from the central government. However, convincing creditors that subnational governments need to be creditworthy to have access to credit markets requires a consistent and sustained policy of letting subnational governments default without bailing them out. In weak and 22 Subnational Capital Markets in Developing Countries unstable regimes the disruption caused by such failures may not be politi- cally sustainable. Why have central governments felt the need to intervene in subnation- al defaults? The answers are rooted in both politics and economics. De- faults where the creditor has effective remedies can lead to lost jobs and re- duced services as subnational governments are forced to pay up. Where creditors are not able to enforce claims, private lenders may simply stop lending to those they hold responsible, including the central government. Creditors may threaten a downgrading of sovereign debt if subsovereign debt does not receive central backing.7 Very large or systemic defaults may undercut the strength of financial institutions and cause them to close or rely on a state bailout of their own. This pressure to tie subsovereign obligations to the central authority re- inforces historical perceptions of the dependency of subnational authori- ties on the central government. The main prescription, besides disavowing any such implicit central guarantee, is to enforce local reliance on own- source and discretionary revenues. This, in conjunction with effective mar- ket regulation and stable central government policies on expenditure as- signments and transfers, should mitigate the moral hazard problems of subnational borrowing., However, while a competitive financial market structure should be used to enforce and help instill fiscal discipline, a myri- ad of other conditions needs to be met as well. Effectiveness of market dis- cipline depends on the extent of local accountability, which is in turn a function of transparency, available resources relative to expenditure assign- ments, fiscal management, and the political environment. Transparency and Financial Management Transparency--easy access to accurate and timely information about a gov- ernment's finances--is often the major obstacle to financial market devel- opment. The few subnational governments that have accessed internation- al financial markets have had to radically revise and upgrade their financial reporting practices (see box 2.2). Having little or no information on fiscal activity impedes reform. In the context of subnational government bor- rowing, transparency relates to budgeting, accounting, and auditing: · Budgeting. In many countries subnational government budgets do not distinguish between current and capital expenditures or between or- dinary revenues and loan receipts, or they provide inaccurate num- Fiscal Devolution 23 Box 2.2. Rio and the International Marketplace In the late 1990s Rio de Janeiro's municipal administration, in preparation for an international bond sale, gained a clear under- standing of the need for transparency and adequate information disclosure. It was the first and only municipality in Brazil to re- tain internationally recognized auditors to examine its books. Even though the city's financial reporting was among the more comprehensive for Latin American subnational authorities, it still suffered from serious gaps. The city did not produce a bal- ance sheet and, therefore, lacked a reliable view of its net asset position. In addition, it prepared financial statements in accor- dance with Brazilian legislation, which at times diverged signifi- cantly from international accounting standards. For the launch of its first issue of securities in the international market, Rio dra- matically improved its reporting system, even providing regular and updated information on the Internet. Source: Chapter 15, case study on Brazil. bers for capital expenditures. The design of the fiscal transfer system may create incentives that foster misreporting of the overall financial picture. As a practical matter, it may be impossible to determine whether borrowing is used for investment or for financing a subna- tional government's short-term deficit. This lack of information un- dermines borrowing rules and impedes the ability to monitor for problems and compliance. · Accounting. Deficient accounting rules and practices defeat trans- parency. Without a well-defined, uniformly applied set of accounting standards, it is impossible to judge a jurisdiction's financial health. Consequently, the absence of accurate reporting and clear applica- tions of definitions undermines the establishment of effective param- eters for borrowing, managing local assets and finances, and monitor- ing financial behavior. · Auditing. Independent, third-party auditing of accounts can help en- sure accuracy and legitimacy. Unfortunately, the list of possible audit- ing candidates in many developing countries is small, and consistent 24 Subnational Capital Markets in Developing Countries auditing standards may not have been developed. Furthermore, few incentives exist to promote professional discipline and minimal checks of auditing practices. Controlling Subnational Government Borrowing Macroeconomic stability requires reducing the moral hazard that allows subnational governments to borrow too much and investors to lend to them unwisely. Needed is a hard local budget constraint requiring that the future resources to pay the debt be prudently calculated and the door to the national treasury be resolutely closed to bailouts. To accomplish this, some key ideas need to prevail: · Rules for grants must be clear, and an effective monitoring system must be established for grants targeted to particular uses. Grants for capital purposes should be integrated with "market-based" loans to the extent feasible. · Central government lending to subnational governments should be curbed where possible and subject to the fiscal capacity of subnation- al governments. Subnational governments without access to private capital can be "taught" debt management through borrowing from the central government. However, the possibility of graduating to pri- vate capital markets must exist, and the lending programs should not undercut the operation of private credit markets. · Explicit limits on any sovereign guarantees should be set and the use of such guarantees should be avoided. Develop a prudent, rule-driven framework for subnational borrowing, setting forth clearly the appro- priate limitations and procedures to follow. · Any "implied" sovereign guarantee should be explicitly disavowed and procedures for dealing with defaulting or bankrupt subsovereign governments and with creditor rights and processes should be in place Regimes for Coordination and Control Controls on subnational fiscal relationships can be cooperative, rule-based, or direct central government regulation (Ter-Minassian and Craig 1997, chapter 7). The choice depends on the political heritage of the country and its form of government, the confidence that can be placed in the efficacy of market discipline, and success in imposing a hard budget constraint. Fiscal Devolution 25 The cooperative model involves negotiation between national and sub- national levels to establish limits on indebtedness that place subnational governments firmly in line with macroeconomic objectives and key fiscal parameters. This approach maintains overall deficit targets and growth guidelines for revenue streams at the central government level. The main drawback is that cooperation may not be politically possible or may be lop- sided, with the center unilaterally "forcing down" decisions or the subna- tional government refusing to cooperate without major concessions. In the absence of a strong center to enforce discipline, the approach requires shared fiscal discipline and a conservative borrowing mentality. Weak fiscal management or weak central government leadership will derail the stabili- ty of the system. A rules-based approach strengthens central control by embedding the framework for subnational borrowing within legislation. By establishing the rules upfront, it avoids the quarrelling between levels of government typical of the cooperative system. Uniform accounting standards are re- quired to eliminate subnational governments' circumvention of the rules. This entails the creation of a financial information system that provides data on the expenditures and financial operations of all levels of govern- ment. Enforcement can come through the market or administrative over- sight, with professionals attesting to observance of the rules and held cul- pable if they break them. Direct central government control may involve setting annual limits on borrowing, reviewing debt proposals, or formally sanctioning specific debt transactions. Central controls allow debt policy to be readily linked with overall macroeconomic policy, but the process has several shortcomings. When there are many subnational borrowers, approvals can be time con- suming. There are also issues of competency and corruption. Officials ap- proving the transactions may have little knowledge or interest, and layers of approval always open the door to political discretion and corruption. Central government approval, especially in the context of donor-funded on-lending programs, can be viewed as tantamount to a guarantee. Howev- er, in domestic private markets, central review and approval need not im- ply a guarantee, and borrowing governments can still face a hard budget constraint. Nonetheless, if the central government approves all subnation- al borrowing, it may be politically difficult not to bail out subnational gov- ernments that default.8 The regime of subnational borrowing controls needs to be examined in the context of the overall argument in favor of devolution of political deci- 26 Subnational Capital Markets in Developing Countries sionmaking. Direct administrative control may be the most comfortable and conservative approach from the center's point of view, but it means a diminution in the fiscal powers and prowess of the subnational govern- ments. It likely means less access to credit markets, a continuing subordina- tion of local self-sufficiency, and continuing, if not greater, reliance on cen- tral resources. Eventually, the rules on subnational borrowing will reflect the stage of market and political development of each country, the rigor in employing strict budgetary constraints at the subnational level, indepen- dence from political cycles, and the strength of accountability mecha- nisms. The next chapters review how these components of subnational market development fit together. Notes 1. A point not to be overlooked is that political and fiscal devolution calls for a substantial element of sacrifice on the part of national politicians that give away power and resources (and patronage) to lower levels of gov- ernment in the process. As a result, for devolutionary movements to be ef- fective, local political powers need to be persuasive and potent on the na- tional level. 2. Perhaps unappreciated is the difficulty national governments have in reorganizing themselves to operate on a more local basis. In many coun- tries, holding a central government job has been a reward for the brightest and best, and devolution has meant a step down in occupational status and a new constituency to serve. Familiar political and administrative power structures have been capsized in the process. 3. Argentina and Brazil have traditions of high levels of local and region- al autonomy. For largely political reasons, the imposition of hard budget constraints proved impossible, and the national governments accumulated large debt burdens to cover the operating deficits of local governments (see chapters 14 and 15 for Argentina and Brazil case studies). 4. Subnational governments, if left unconstrained in their fiscal con- duct, cause problems for the national government. For example, the impo- sition of local taxes on commerce and trade can adversely affect costs and revenues at the national level. Borrowing in foreign currency can lead to build-ups in foreign-denominated liabilities that create demand for foreign currency. Borrowing large amounts from banks can undercut creditworthi- ness if default looms. These problems are most evident in certain federa- Fiscal Devolution 27 tions, such as Argentina, where the central government has weak control over subnational fiscal behavior. 5. See table 14.1 in chapter 14 on Argentina for an example of distribu- tion of responsibilities that maps well to these guidelines. 6. Transfers can have other uses, including smoothing regional econom- ic shocks and providing targeted boosts to a regional economy. 7. This alleged "threat" is something of a curiosity in the absence of a specific pledge by the sovereign to make good on local debts. On the con- trary, among the major rating agencies, it is the act of bailing out failed borrowers in the absence of such a pledge that can lead to downgrading a rating. That is because once the bailout happens, the entire stock of local debt then becomes a potential contingent liability of the sovereign. The more likely causes of "implicit" sovereign guarantees are the weakness of the banking system if much of the debt defaults and the political power of other investors to force central actions to protect their investments. 8. Approval by the central government can mean different things, in- cluding not only implied "sponsorship" by the national authorities but also the kind of tax treatment a security will receive and its eligibility for investment by certain groups of investors. In Russia during the heyday of its "Wild West" municipal bond market, approval of local issues by the Fed- eration's Ministry of Finance was needed to obtain tax exemption and to permit investment by institutions, even though the status of a state guar- antee was unclear. In the absence of clear laws, would-be issuers would bar- gain with the central authorities for designation as an "approved" security, making it a political exercise (see Halligan 1996). Chapter 3 Market Setting and Legal Framework The financial operations of subnational governments are strongly affected by the financial market and the legal framework. The two are intimately linked. For markets to thrive, laws and regulations on their operation and structure should be in place and enforced. Financial markets, dealing in vast amounts of funds with numerous buyers and sellers, are by definition advanced marketplaces that are efficient and ultimately sustainable only to the degree that an equally vast variety of transactions is quickly and hon- estly handled. Borrowing rests on the premise that funds are lent with the expectation of their repayment and with compensation for their use. The debt instrument is a contract to that effect. The capacity of subnational governments to access credit markets, by bank loan or bond issue, depends on the perception of their debt contracts as a strong promise to pay so that funds can be secured on favorable terms. Financial Market Structures and Subnational Finance Credit market access for subnational governments is strongly tied to the character and stage of development of domestic financial markets.1 While financial markets vary greatly, some generalizations seem to apply. In grow- ing market systems, liberalization of capital markets and greater devolution have tended to go hand in hand, although not always at the same pace. These are difficult transformations, still under way. Stages of Development Domestic financial sectors are undeveloped in most emerging economies, with limited formal financial market activity and few institutions to supply credit and mobilize savings. The size and vigor of financial markets are typ- 29 30 Subnational Capital Markets in Developing Countries ically considered a leading index of economic development. Most studies have concluded that economic progress is closely allied to the appearance of private sector financial institutions that can marshal resources, accom- modate payments, pool risks, allocate credit and make equity investments, and monitor borrowers and ownership interests (World Bank 2002c). How- ever, many important provisos arise. For example, to operate well financial markets require strong supporting institutions. To meet these conditions, these supporting institutions must often undergo reform themselves. Most reformers agree on what the destination should look like, that is free markets working efficiently within an acceptable framework of social justice. Agreement on how to achieve the goal within the boundaries of ac- ceptable short-term costs is more elusive. Some would have all reforms in place before trusting the market to be an efficient allocator. Others believe that experience is the best teacher and that a better course is making incre- mental reforms to meet needs as they arise. Most emerging and transitioning economies are notable for the low level of bank deposits and lending relative to their gross domestic product (GDP). Bank lending to the private sector is limited as are investments by other fi- nancial institutions, so investment tends to be self-financed from firms' cur- rent savings. As an economy develops, there is greater use of external funds, first from banks and later from stock and bond markets; these, in turn, de- pend on the development of private nonbank financial institutions, includ- ing pension funds, insurance companies, mutual funds, and securities mar- kets. The operation of banks and nonbank financial intermediaries depends on the regulatory framework, while the potential size and scope of the fi- nancial system depend on the size and vigor of the economy. The Role of Banks and Securities Markets Generally, two different "visions" have evolved on the role of the banking sector and nonbank financial institutions in the provision of capital. Banks, as suppliers of credit, have dominated most financial systems throughout the world and continue to do so in most developing countries, as well as in Western Europe. In the United States and several other devel- oped countries, however, a broader model of a financial market system dominates, with strong securities markets for active trading of equities and debt among a variety of investors (see box 3.1) Sometimes, the banking system spurs development of capital markets. In the Philippines, private banks have begun to purchase municipal bonds rather than lend funds to municipalities. Lending to municipalities is dom- Market Setting and Legal Framework 31 inated by government-owned banks, with close political and financial ties to local politicians and captive deposits of subnational government funds (required by law). Nevertheless, financial and tax incentives and the rea- sonably developed underwriting infrastructure in the Philippines have cre- ated a nascent municipal bond market, actively assisted by international donors (see the Philippines case study, chapter 26). Debate continues on the advantages and disadvantages of the banking sys- tem and securities market approaches for raising capital, but in most coun- tries a mixture has been found to be useful, especially for long-term borrow- ing.2 The more important issue appears to be competition: can subnational governments seek capital in credit markets with effective competition among several private sector providers? For small countries with concentrated bank- ing systems and few nonbank institutional investors, the possibilities are re- mote for domestic financial markets to meet competitive norms. Another set of concerns is the regulation and optimal structure of finan- cial markets. In most developing and transitioning economies, banks dom- inate the financial sector, often growing out of a tradition of one or a few government-owned banks that monopolize credit provision. Financial lib- eralization and the emergence of securities markets as countries move to- ward greater private ownership and market-based economies tend to begin with the creation of a primitive "money market" (securities markets for short-term debt obligations), soon followed by an equity market and stock exchanges (Schuler, Sheets, and Weig 1998). The sequence is appropriate. Money markets, frequently narrow in the number of buyers and sellers and focusing on a few frequently traded short-term obligations, have often been the exclusive domain of banks trading overnight funds. A strong banking system is a necessary compo- nent of a market-based system. Securities exchanges are important because they offer a mechanism for making longer term debt obligations liquid, but without an efficient banking system and payments mechanism, stock and bond markets are unlikely to survive. The question is whether the money market and the stock market will form the launching platform for a longer- term debt market and if that market will be accessible to subnational gov- ernments. Another feature of many emerging and transitioning economies is the heavy reliance on the banking system to finance central government debt; this reliance can be an impediment to the development of markets for pri- vate debt and subsovereign debt (see box 3.2).3 Sovereign debt, commonly at generous yields, tends to squeeze out available capital for "riskier" pri- 32 Subnational Capital Markets in Developing Countries Box 3.1. Banks and Securities Markets: Are Both Needed for Development? Research indicates that financial structure does not explain cross-country differences in long-term GDP growth, industrial production, use of external funds by firms, or firm growth. However, there is a correlation between the level of a country's wealth on the one hand and development of nonbank financial institutions and the relative size of the securities market on the other. Financial structure tends to change with development, because banks and securities markets have different requirements for in- formation and contract enforcement. The information that banks collect is private and gathered in their direct relationship with clients, which limits the reliance on outside support ser- vices such as accountants or rating agencies. A bank can con- trol a borrower's conduct by threatening to withhold credit or to hold the borrower's deposits, but large financing needs may ex- ceed the resources of individual banks or violate their prudential restrictions. Securities markets depend on bondholder and eq- uity-owner protections, reliable and timely accounting and other information, and external analysis by credit rating agencies and investment funds, among others. Borrowers will seek access to security markets to increase com- petition for their investments and to enjoy more options in type of financing. Security markets require both interested investors and a strong support system to protect investor rights and the functioning of markets. Source: World Bank 2002c. vate sector loans. It also has lessened the desire of investors to finance sub- sovereign debt, often viewed as junior in status and, in effect, subordinated to the sovereign debt. Stock exchanges (often seen as the banner institution of capitalism) have been created in many emerging economies, often as by-products of Market Setting and Legal Framework 33 Box 3.2. Brazilian Banks' Excessive Concentration in Government Securities Banks can be exposed to financial turmoil because of banking capital adequacy regulations that foster reliance on the banks' domestic sovereign government obligations. For example, Standard and Poor's downgraded Brazilian Banks in mid-2002, not for weak private loan portfolios but for excessive concentra- tion in central government obligations. These are domestically favored under capital adequacy norms that view them as the most secure uses of capital. They are weighted as "zero risk" under the conventions of the prevailing capital adequacy sys- tem and do not require a capital allocation from a regulatory point of view. However, in the sterner world of international finance, a declin- ing currency and a shaky central government can lead to a downgrading of the banks for reasons that are diametrically op- posed to the capital adequacy norm. In Brazil's case there was an across the board downgrading of banks even though the do- mestic private loan portfolios were judged to be secure because of the weakness of the central government (which sets the up- per end of the scale in terms of the sovereign risk limit) and the heavy concentration of the banks in central government bonds, along with the rapid depreciation of the Brazilian real. Source: Standard & Poor's Research, Brazilian Banks in Time of Turbulence (July 24, 2002). converting government-owned industries into private concerns. An objec- tive has been to stimulate increased private capital formation and to create new sources of investment funds. Credit financing per se and the listing of debt obligations on exchanges have not usually been central features of the new financial systems. Furthermore, in many domestic markets, high infla- tion rates, unstable economies, bank domination of credit, and evolving le- gal systems have combined to discourage the use of credit market instru- ments by private entities. 34 Subnational Capital Markets in Developing Countries Both the bank-dominated and market-dominated models have their ad- vocates. Here, the presumption is that as other financial institutions grow in importance, there will be a natural push to develop debt markets as an outlet for their funds and to meet needs for asset-liability matching, diver- sification, and liquidity. Accordingly, the following chapters pay consider- able attention to the creation and operation of bond markets and the groups of "passive" investors that may supply credit by that means. Experience in Developed Economies Developed countries have used a variety of borrowing methods, institu- tions, and debt instruments to finance long-term subnational infrastruc- ture needs, but two forms predominate: government-sponsored financing institutions and bond markets. Specialized lending institutions have been the preferred approach in Western Europe, while direct access to bond mar- kets has been the preferred approach in the United States. Political history and financial circumstance have generally guided the choice of model. In Europe the unitary state structure, strong state-owned banking sys- tems, and the limited and subordinated role of subnational governments led to the late development of subnational capital financing responsibili- ties. State-owned banks specializing in extending credit to subnational gov- ernments emerged in the early twentieth century in a number of countries. In Austria, Belgium, Germany, Finland, France, Italy, the Netherlands, Spain, and Sweden, these institutions often were financed by special de- posits that provided low-cost funds for lending to subnational govern- ments.4 Sometimes localities formed cooperative financial institutions (as in Finland, the Netherlands, and Sweden) to finance their borrowing needs. For the most part, these institutions amounted to state-sponsored credit monopolies.5 They became increasingly market-oriented as deregula- tion enabled them to raise funds by borrowing in capital markets. In the late 1980s, with the move toward privatization, state-owned spe- cialized banks became candidates for private capitalization. An outstanding example is the transformation of the Credit Locale de France into a private stock company in 1987. No longer a depository institution, Credit Locale became a major bond issuer in the domestic and international markets. Its merger with the Belgian Credit Communal de Belgique led to the forma- tion of Dexia in 1996, a full-service, "relationship" banking institution that makes loans and underwrites subnational government bond issues through its affiliates. Dexia has bought shares in the specialized subnational lending institutions of Austria, Italy, and Spain. It has about a 40 percent share of Market Setting and Legal Framework 35 the French local government loan market and a 90 percent share of the Bel- gian market. Dexia seeks long-term relationships with its subnational clients and provides management and planning services in addition to banking services. This model is similar to that of the traditional European bank, which often has close operational relationships with its corporate clients. The U.S. experience is different. Its federal system has prevented heavy central involvement at the subnational level. State and local governments are responsible for most public works spending and have their own strong revenue systems. With a fragmented banking system and a highly devel- oped capital market, this structure of broad local responsibilities and sub- stantial fiscal capacity encouraged direct borrowing from bond markets, helped by the exemption of interest on state and local bonds from federal and state income taxes. The U.S. municipal bond market is by far the largest in the world, with some 14,000 new bond and note issues sold each year.6 As of late 2002 ap- proximately $2 trillion in U.S. municipal bonds was outstanding, an amount equal to all the corporate, financial institution, and government bonds outstanding in emerging markets (IMF 2002, p. 49). Size, wealth, a long tradition of federalism and subnational government autonomy, and strong institutional development combine to make the U.S. experience with subnational direct borrowing in the bond markets unique. Nonethe- less, other countries, including those in Europe, have permitted or actively promoted the direct use of bond markets by major subnational issuers. However, as noted, there is some specialized intermediation activity in U.S. municipal bond markets as well. In 12 states small governments can use bond banks, and many other states have special revolving funds for en- vironmental purposes that were created using special capital grants from the federal government in the early 1980s.7 In addition to tax exemption of interest income, the U. S. tax code contains provisions for the treatment of bank investments in subnational government obligations that favor bor- rowing by small governments.8 Emerging and transitioning economies are applying elements of both the European banking-based and U.S. capital market-based models. In addition, most have looked to donor-assisted loan funds as the principal sources of capital. Operated by national or state gov- ernment entities, municipal development funds are the sole source of long- term funds in many countries, with interest rates and terms much better than would be otherwise obtainable in domestic capital markets--where these markets exist. 36 Subnational Capital Markets in Developing Countries Municipal development funds usually have multiple objectives and may provide training, technical assistance, and grants in addition to loans. The funds have had mixed success. In Brazil state-level municipal funds have had impressive results in working with municipalities (on concessionary terms, however, effectively precluding competition from private market sources). A major issue is how to wean the funds from donor dependence and reshape them to promote more private sector involvement in subna- tional finance. One fund, the Tamil Nadu Municipal Development Fund in India, has been converted into public-private joint ownership with private sector management. It has sold bonds on the Indian capital market and in- creasingly acts as an intermediary (see the India case study, chapter 24). Multinational organizations, leery of creating a culture of dependency in credit programs, are seeking ways to leverage more private capital into fi- nancing infrastructure.9 An Array of Options Subnational governments have an extensive array of potential domestic sources of credit, from direct loans from agencies of the central govern- ment and various state-sponsored specialized loan funds to private capital markets (figure 3.1). What options are actually available to subnational governments depends on various limitations and controls, including laws and regulations, private sector capacity, and market-imposed limitations. For example, subnational governments may be precluded by law or prac- tice from borrowing in the private sector, or they may have nominal access but little interest in their securities. Most developing and transitioning countries are still emerging from a time when the central government was the exclusive provider of capital funds or took primary responsibility for capital projects. This was the case for the unitary states that emerged from the former communist states and for highly centralized countries elsewhere. Much of the story of improving access to credit markets for these countries is one of moving from a central government monopoly in the provision of credit or in responsibility for capital projects. The options among sources of funds are not mutually exclusive. Even in relatively undeveloped economies several sources may operate simultane- ously. In addition, financial structures are undergoing change. There has been a worldwide move toward privatization of bank ownership. A subna- tional government that has relied on a government-owned bank for capital might have to find a new source of funds, although in some cases the cred- Market Setting and Legal Framework 37 Local government borrower Limitations and controls Central Special Government Private Domestic International government development banks banks securities securities agency fund market market Public sector Private markets Figure 3.1. Market Structures and Sources of Capital for Local Government Borrowing it and depository activities of subnational governments are sequestered and retained by government-owned institutions. Access by subnational governments to international securities markets is still a rare occurrence. Few subnational governments have managed to at- tract foreign investors; national governments, concerned about the implica- tions for monetary and exchange rate policies, have banned access to inter- national securities markets or have supervised it carefully. For example, after the devaluation of the ruble in 1998 and widespread defaults, the Russian national government forbade new lending in the international markets by subnational authorities (see the Russian Federation case study, chapter 31). Barriers to the Market The financial operations of most subnational governments are still highly regulated by the center, effectively precluding or severely limiting private sector interest in their debt. A common restriction is that subnational gov- ernments must keep their deposits with the central government treasury or in a government-owned bank (see the South Africa case study, chapter 18). Moreover, the traditional insulation of subnational government from pri- vate sector financial institutions has led to mutual ignorance, if not distrust. Subnational governments may find that accessing private capital sources is uneconomic or politically difficult to justify. This can occur where the central government or its lending institutions provide direct funds on 38 Subnational Capital Markets in Developing Countries terms that are much better than those of the private credit markets. A ma- jor source of such concessionary loan funds has been donor-based on-lend- ing programs, which offer low rates of interest, generous grace periods, and longer term loans than the private sector can. Even where there is a long wait for the low-cost funds and waiting costs are high, local leaders are re- luctant to borrow from higher-cost private sources. Concessional lending has had mixed results, at best. There have been concerns about the long-term efficiency of channeling capital into low-re- turn public projects and about the incentives that are created for govern- ments in order to qualify for such borrowing. Hence, a special concern in meeting the capital financing needs of subnational governments is to wean them away from concessionary finance and to make them both better able and more willing to compete for funds in commercial markets. Of course, private sources of capital must be willing and able to make such loans. Bringing subnational governments and capital markets to the requisite lev- els of willingness and ability to receive and make loans and to issue and in- vest in securities is, as outlined in the following chapters, a great challenge. The Role of Subnational Government Borrowing in Developing Financial Markets Subnational obligations ought to be viewed in the context of credit market development. Their contribution to that development, while seldom ac- knowledged, can be substantial. Even with the tide of devolution, the no- tion that subnational governments would seek resources from private sources was novel and has taken some getting used to by private market participants. While the options for credit access differ by country, the transition from central government monopoly to private market competition has tended to move from exclusive reliance on the central government to increasing reliance on the banking system and then finally to access to securities mar- kets (figure 3.2). These sources are not mutually exclusive, and they may be tapped in tandem, with larger borrowers that are better able to compete moving toward the use of the securities markets as they develop. Business firms have followed much the same trajectory. As is the case with private firms, the evolution in credit markets often stops with the banking system, which dominates the financial landscape in many countries. However, in economies that are intent on developing securities markets, subnational governments will eventually either directly enter the securities market or be greatly influenced by it, accessing it indirectly through intermediaries. Market Setting and Legal Framework 39 Securities Commercial markets bank lending Local unit borrower Central government lending monopoly Figure 3.2. Stages of Development in Credit Market Access The move toward privatization in most developing and emerging economies has left subnational governments in a quandary. The presump- tion has been that financial markets were a private sector phenomenon and that governments had little to offer in the way of resources or security. Where government activities in the sector were profitable, they were viewed as belonging in the private sector and the solution was to privatize them. While that thinking continues in many countries, a more balanced ap- praisal sees the subnational government as a supplier of vital services and critical infrastructure for an improved society and thus as a strong stimulus for the development of capital markets. Good models are provided by Canada, the United States, and a growing number of Western European countries whose subnational governments and their enterprises are often the major providers of public services, using their own resources and mak- ing their own investment decisions. Since most such governments are well- run and benefit from their monopoly positions as potential borrowers, in- vestors operating in well-regulated financial systems see the obligations of these governments as safe outlets for their funds.10 In many emerging economies subnational government obligations can attain a level of credit- worthiness that makes them attractive investments for the private sector and for banks, although banks are limited in their ability to provide the 40 Subnational Capital Markets in Developing Countries long-term financing needed for infrastructure. Additionally, in some emerging economies subnational governments have already entered securi- ties markets, providing new investment outlets. The Legal Setting Underdeveloped legal systems and weak and corrupt judiciaries can make it difficult for subnational governments to convince wary private investors to invest in their obligations. Increasing subnational borrowings in domestic financial markets often requires adjusting regulations on borrowing au- thority and the issuance, registration, and servicing of debt, as in Morocco, for example. Market distortions that lead to a preference for one instru- ment over another (loans rather than bonds, for example) need to be re- moved. The regulatory and supervisory framework for subnational borrow- ing needs to be strengthened, especially prudential regulations and bankruptcy laws (see the Morocco case study, chapter 20). Institutions and large individual investors are interested in knowing that the obligations in which they invest are valid obligations that create enforceable claims against the obligors, based on underlying contracts. Contracts and their enforceability vary according to the legal system under which they are drafted and the transparency, competence, and honesty of the judicial system that enforces them. These attributes of legality and en- forceability of contracts, basic to the fair and efficient operation of capital markets, are often undeveloped and weak in developing and transitioning economies. Differences in Legal Systems Legal traditions have had a profound influence on fundamental concepts of ownership and creditor rights and consequently on the development of financial markets. There are wide differences in how quickly and how well laws pertaining to investor and creditor rights are enforced. These rights are not inherent in the securities but are determined by law. This can have a profound effect on the ability of countries to establish effective securities markets for debt and equity capital.11 Laws governing financial markets and securities strongly influence the accessibility and cost of capital for subnational governments. Security--what the lender can look to for assurance that a loan will be repaid or, if not, what asset it will be able to seize--is a fundamental con- cept in private capital markets. Numerous complications can arise. Some le- Market Setting and Legal Framework 41 gal systems are weak in describing creditors' legal rights: the priority of creditors' claims may be unclear, the legal presumption may run to the par- ty that has physical possession and use of the asset, no matter the liens against it. Laws governing secured transactions may be in conflict or may predate modern commerce and financial relationships and instruments. Thus laws may limit who can lend, what types of collateral may be pledged, and how the collateral is to be identified and physically kept. However, even where the ability to use collateral as security is clear, there may be no central registry to keep track of liens against collateral. The abil- ity to collateralize a right to revenues (and assign them, as in the case of tax collections) may be ambiguous or nonexistent (World Bank 2002c). The local financial obligations of some Czech municipalities are guaran- teed by national institutions, but most are not explicitly guaranteed and it is unclear how creditors would recover their money in case of default. The Czech Bankruptcy and Composition Act does not cover municipalities, an omission believed to contribute to the reluctance of the banking system to finance municipalities. Municipalities, for their part, have been granting loans and guarantees to businesses to support local development activities. Although these financial activities require the approval of municipal as- semblies, the procedures to be followed are not clear. The lack of debt mon- itoring and supervisory mechanisms softens local budget constraints and creates moral hazard incentives, contributing to higher fiscal risk (see Czech Republic case study, chapter 28). In addition to the complexities of pledging properties and gaining rights to revenues, a lender's willingness to accept pledges depends on enforce- ment and the ease with which claims can be settled. When the ability to enforce pledges surely and quickly is in doubt, loans are not forthcoming or are forthcoming only at higher rates. While enforcement problems are acute for loans to private parties in the case of moveable property, they are exacerbated in loans to the public sector by restrictions on pledges of pub- lic property and the difficulties in enforcing such pledges against the sover- eign or its subdivisions. A practical problem is the independence of the judiciary and its compe- tence to understand financial issues and adjudicate impartially. Modern se- curities markets depend on concepts that may exceed the mastery of tradi- tional judges. They also depend heavily on the integrity of a system of laws rather than on traditional or political relationships. An "absentee creditor" many miles removed from the scene who relies on a written contract is un- likely to obtain justice in settings with a corrupt or weak judiciary. 42 Subnational Capital Markets in Developing Countries Implications of the Legal System for Subnational Borrowing Subnational government borrowing is part of a larger legal fabric that sets the roles and responsibilities of public and private sector entities in the op- eration of a subnational government credit market. Several policy issues that have a substantial impact on subnational government borrowing can be covered only tangentially in this book. Examples are policies and prac- tices on ownership of public property and property rights associated with such ownership, the structure of intergovernmental revenue sharing, the adequacy and reliability of the accounting standards, regulation of the banking system and other financial sectors, and judicial enforcement.12 Nonetheless, an imperfect legal system and evolving fiscal situation are not necessarily impediments to the initial development of a subnational gov- ernment credit market. Many countries have had imperfect legal structures when they began pilot projects in market-based subnational government bor- rowing. Some now have developed substantial subnational government cred- it markets. Many of the risks can be diminished through provisions in the loan contracts, even before a fully developed legal framework is in place. In fact, the practical problems of developing "pioneer" transactions have ex- posed gaps in law and practices and prompted solutions. Both policy reforms and market practices are likely to be implemented incrementally. A successful subnational government credit market must be built both from the top down, by building a legal and policy framework to support efficient credit market operations, and from the bottom up, by ac- cumulating practical experience in banks and other lenders in making loans and in subnational governments in borrowing to finance high priori- ty investments and making timely debt service payments. Both tracks should move forward simultaneously. Legal Framework--Planning Ahead Several emerging and transitioning countries have confronted the reality of large-scale subnational government borrowing and been obliged to con- struct a legal framework after the fact to accommodate the interests of stakeholders on both sides of the market. Developing subnational markets can benefit from examining the difficulties in countries where subnational borrowing got out of hand. From Brazil to Russia excessive borrowing by some subnational governments in the absence of an adequate legal frame- work has exacerbated national economic crises. The promise of soundly based subnational borrowing is large, but the risks in badly prepared bor- rowing are also large. All parties (subnational governments, banks, and po- Market Setting and Legal Framework 43 tential investors) share an interest in fully understanding the policy issues surrounding credit market development and in having an appropriate legal framework in place before substantial borrowing occurs. Notes 1. The term financial market is used generically and encompasses both bank and non-bank institutions and other potential providers of capital, including various governmental entities that may lend to subnational gov- ernments. Although governments themselves are not participants in the equity markets, their companies and projects can be from time to time, and since stock exchanges frequently deal in both equities and debt, their exis- tence and activity are of interest here. The term credit is also used generical- ly to describe debt obligations and not just bank loans. 2. The crux of the matter is finding a financial structure that suits the needs of investors and potential borrowers. To finance infrastructure, local governments need access to long-term capital. On the investors' side, there are institutions and individuals that prefer long-term investments to offset their long-term liabilities. Banks, on the other hand, are constrained by the mismatch between their short-term liabilities (deposits) and the long-term assets that loans for infrastructure represent. Another desirable attribute of the securities market is the need for disclosure and the collateral benefits of wide scale information about the financial conduct of governments. 3. To the extent that a market for debt securities existed at all, it was dom- inated by central government or central bank obligations. Frequently, the debt market existed primarily if not exclusively to finance central govern- ment deficits. Regulation of investments and bank assets is tilted toward pro- viding a ready market for the national government debt. The extension of credit to private entities often has taken a back seat in banking operations. 4. Except for Finland, the Netherlands, and Sweden, all the specialized banks have been privatized. For a discussion of national experiences with specialized municipal lending institutions, see Peterson 1998. 5. Japan, as a unitary state with close fiscal linkages between the central and local governments, has followed a similar pattern. About 75 percent of its local government financing needs are met by specialized central govern- ment-owned entities. Low-cost funds are available from the Fiscal Invest- ment and Loan Program, financed indirectly by Post Office savings ac- counts. The other 25 percent is made up of direct borrowings from private sector institutions or bond issues. Subnational borrowings need the ap- 44 Subnational Capital Markets in Developing Countries proval of the central government. See Mihajek, p. 297. Lending to local governments is treated as part of the counter-cyclical policy in Japan and terms are made easier at times when the national government is seeking to stimulate the economy. 6. The term municipal bond is used generically to mean the obligations of subnational governments. 7. State revolving funds were set up as loan funds to replace the federal direct grant program for water pollution control. Some revolving funds em- ploy interest rate subsidies. Some also leverage borrowed funds on top of the capitalization supplied by the federal government and the 80 percent match required of the states. 8. Section 265 of the tax code permits commercial banks to partially write off the cost of capital for holdings on bonds that are sold by issuers that borrow less than 10 million dollars in a given year. The upshot is that interest rates on these bonds are usually 15 to 25 basis points lower than those on similar bonds sold by larger governmental units. 9. Chapter 12 addresses these issues of design in greater detail. The World Bank has recently emphasized the need to move municipal develop- ment funds toward "market" behavior; see, for example, World Bank 2001, p. 49. 10. These well-deserved reputations were not built overnight or without disappointments along the way. The United States, in particular, had a long and lurid history of defaults in the nineteenth century that gave rise to re- strictions (imposed by the states upon themselves) and hard-nosed market practices (imposed by lenders). These reforms ultimately laid the founda- tion for subnational bond markets viewed as being the safest next to that of sovereign bond markets. Bank regulation of investment requirements and required documentation, the large size and sophistication of the mar- ket, and the strength of the underlying revenue systems all contributed to the development of the quality of the market. 11. Several studies have examined the impact of the legal system on the development of the financial markets. The general conclusion is that of the four "families" of law, countries employing the French civil law tradition have the weakest protection for investors (and creditors) whereas those that follow the common law (British) tradition have the strongest. The oth- er two families of laws, the German and Scandinavian, fall in between. Fur- thermore, enforcement is often slow and subject to corruption. Regions (such as South America and parts of Africa and Europe) that follow the French civil law frequently provide an unattractive legal environment for Market Setting and Legal Framework 45 investors, often with high concentrations of financial markets and owner- ship of assets (see Burke and Perry 1998, chapter 4). 12. Additionally, laws relating to public procurement, tariff setting, and own-source revenues can have a substantial impact on the development of a subnational credit market. Stability and clarity in the generic laws and the ability to specify and monitor conduct in the loan contract are key in- gredients to limited obligation "project" financing. Part II Borrowing Instruments and Restrictions on Their Use John Petersen and Miguel Valadez 47 Chapter 4 Subnational Governments as Borrowers Subnational debt can be the obligation of a local, regional, provincial, or state government or of projects they sponsor through subsidies, partner- ships, or concessions with the private sector. Subnational governments en- ter into many types of legal and financial relationships, which can differ markedly among countries. In many places these relationships are evolv- ing, and even where they are established, they continue to be dynamic. Thus, policymakers and analysts must be prepared to examine a variety of factors and risk exposures when dealing with the debt transactions of sub- national governments. Subnational government borrowers have much in common with other borrowers such as public utilities and private firms. But there are also some special features relating to the powers, structure, and operation of subna- tional governments. For example, most subnational governments deal ex- clusively in domestic currency for revenues and expenditures. Thus, except for certain types of facilities (electric power, ports, airports, telecommunica- tions), they have little access to foreign currency payments. For some ser- vices, governments have powers approaching monopoly status that may be enforced by regulation. Additionally, governments are site-specific and un- able to change the geographic locus of business or the fundamental nature of the services they provide. They rarely go out of business.1 Debt Classification A fundamental distinction in classifying debt is whether the subnational government is the borrower, relying primarily on its taxing power and oth- er general governmental revenues to back the loan, or whether the govern- ment is just a party to the loan, as when the obligation is limited to a par- 49 50 Subnational Capital Markets in Developing Countries ticular revenue source of an enterprise to which the general governmental credit is not pledged (a limited or nonguaranteed obligation). This distinc- tion is reasonably clear in the United States, where a revenue-generating project or enterprise that is financed with a limited obligation is referred to as a revenue bond. Elsewhere, the distinctions between general and limited pledges can be blurry, as in the case of projects financed with a mixture of public and pri- vate funds, service and off-take contracts, profit-sharing arrangements, or concessions with guarantees of use. Confusion is especially likely in coun- tries where various government commercial and industrial activities are be- ing privatized. The credit structure may be especially complex, with a blend of risk factors involving both the public and private sectors, in "project fi- nance" cases, where the private sector is not only a direct investor in a proj- ect but also an equity provider and actively engaged in operation and man- agement. The following discussion and accompanying figures describe three pro- totypical financing and credit structures involving subnational govern- ments. For ease of exposition, the borrowing is assumed to involve a proj- ect, as is typically the case, although it could as well be used for other purposes, including relending, to meet emergency needs or to fund accu- mulated deficits. General Government Obligation With a general government obligation the government uses its general rev- enues to make debt service payments and owns and operates the project it- self (figure 4.1). In most countries this would be the likely structure for cap- ital expenditures for public safety, public education, health and welfare, and similar activities that are not revenue producing. The government is- sues the debt in its own name and pledges its general revenues. However, neither the financed project nor its earnings are specifically tied to repay- ment of the debt. In an important variant on this theme, the subnational government receives intergovernmental assistance, such as shared taxes or grants, that is pledged as part of the security. Government Limited Obligation (revenue obligation) In a government limited obligation, the debt is secured primarily or exclu- sively on the earnings of a project enterprise that produces revenues through charges and fees that are used to defray much or all of the costs of operation and debt service (figure 4.2). General revenues of the government are typi- Subnational Governments as Borrowers 51 Figure 4.1. General Government Obligation Figure 4.2. Government Limited Obligation cally not pledged directly, and there may even be a prohibition against their use. Common subnational enterprises are public utilities, such as water and sewer, electric distribution, local toll facilities, public markets, harvest pro- cessing facilities, and local ports and terminals. The debt is issued either by the project itself, which may be a limited-purpose special district, or on be- half of the project by the general government sponsor. Project Financings (public-private undertakings) In public-private undertakings, typically in utility-type projects, the gov- ernment contracts with the private sector, through concessions or partner- ship agreements, to build, own, or operate the project (figure 4.3). The gov- ernment may contribute in various ways to the financing, including equity 52 Subnational Capital Markets in Developing Countries Figure 4.3. Public-Private Project Financing interests, subsidies, and guarantees related to the demand for outputs or for supplying needed inputs. The private sector, or international lending enti- ties, also may contribute debt, equity, and various enhancements to the fi- nancial mix. The contract sets outs the obligations of the respective parties and the returns to each. The debt is typically issued in the name of the project and may be non-recourse, looking only to project earnings, owner- ship, or assets for security. Classifying Potential Subnational Borrowers Many subnational governments already have access to credit through gov- ernment-sponsored lending programs, bank lending, or sales of bonds in domestic or international capital markets. However, many more do not, and many factors influence whether and how they will gain access. Classification Based on Fiscal Capacity and Financial Acumen The fiscal capacity and financial acumen of subnational jurisdictions, which relate to the ability and willingness to pay, are fundamental consid- erations in determining which units are candidates for borrowing. Al- though these are not always correlated with size, private creditors generally prefer larger jurisdictions because of their greater sophistication, ability to draw on more resources, and ability to spread the fixed costs of debt trans- actions over larger volumes of borrowing. In most countries three groups of jurisdictions can be identified in terms of the likelihood for the issuance of subsovereign debt in private markets:2 Subnational Governments as Borrowers 53 · Those that already have access to capital markets because of their size, financial and managerial resources, and political clout. This group includes the largest and best known subnational governments with large economies and political muscle. · Those with limited or no access to capital markets but that can gener- ate adequate revenues to meet their responsibilities and otherwise are capable of borrowing private capital. This group consists of subna- tional governments that are large and capable of attracting private in- terest without direct central government help and those that are too small or that lack the managerial capability to attract private lending but that could gain access with assistance. One approach is to com- bine the needs and resources of individual governments and borrow as part of this larger group. · Those that cannot generate sufficient revenue to provide the current services they require or to build and operate the needed infrastruc- ture. Jurisdictions in this group, which for all practical purposes are "financial wards" of higher levels of government, do not have access to capital markets and most likely should not. Jurisdictions in the first two groups have the potential to use private credit resources under a regime in which central government assistance to municipal market development, if any, is accommodative and indirect, fo- cused on laws and regulations that create an enabling environment for sub- national government borrowing in credit markets. Subnational govern- ments in the third group, the very small and poor, neither can nor should borrow in private credit markets. As handy as the above triage of candidates for borrowing might appear to be, it is one that defies drawing strict lines of demarcation in practice. Advocates of light-handed intervention and believers in market solutions say that the market itself will define, better than government regulation, which jurisdictions fall into which category. Others argue that markets as- sume a symmetry of skill and information between buyer and seller that is not met in the case of subnational governments, especially those that are smaller and unsophisticated. Left untended, the unwary can wander into the credit market with unfortunate results. Like any classification scheme, this one is situational and dynamic. Some governments that are too small and too poor to gain access to credit markets using their general revenue funds may latch on to a project financing scheme that is creditworthy. Even subnational governments with otherwise 54 Subnational Capital Markets in Developing Countries insufficient own-source revenues might qualify for private credit if they can pledge a share of their intergovernmental transfers to secure the debt. Policies to Improve the Creditworthiness of Subnational Governments Government policies on intergovernmental finance and technical and credit assistance to small and unsophisticated jurisdictions affect how mar- kets assess creditworthiness. It is likely that countries with weakly financed and poorly managed subnational governments will have to forgo direct en- try into private credit markets or will need to devise policies to help subna- tional governments advance up the creditworthiness ladder. To promote subnational government access to private markets, the Philippines has used a four-quadrant strategy that considers two primary dimensions: a subnational government's wealth and the revenue-generat- ing potential of the proposed improvement (figure 4.4). For the smallest and poorest subnational governments that need to finance non-revenue- producing facilities, grants are the preferred means of assistance (lower-left quadrant). For subnational governments with adequate wealth and self- supporting projects, access to bond markets was the preferred financing mechanism for larger projects, with commercial bank lending at commer- cial rates with no grants or subsidies for smaller but commercially viable projects (upper-right quadrant). Because bank lending to subnational gov- ernments has been dominated by government financial institutions, an added dimension of the approach is to move from government financial Figure 4.4. Matrix of Subnational Government Financing Capacity Subnational Governments as Borrowers 55 institution lending (the loan and grant quadrant) to private credit sources (the loans and bonds quadrant).3 Government financial institutions were to facilitate the move to private capital as governments grew stronger and projects became self-financing (see the Philippines case study, chapter 26). Distinctions among Subnational Jurisdictions Approaches like these based on existing creditworthiness are useful for ana- lytic purposes, such as describing potential demand for credit and the likely size and viability of a subnational government securities market. However, should such distinctions be codified into law or regulation to identify which subnational governments can access credit markets? In developed economies credit markets effectively classify borrowers and reflect their cred- it assessments in the prices (interest rates) charged for borrowing, based on perceived differences in economic vitality, managerial efficiencies, financial condition, political sway, and the viability of individual projects. While detailed regulatory prescriptions are best avoided, senior levels of government have a legitimate interest in the financial market behavior of subnational governments, as chapter 2 describes. Even in mature markets, most national governments and some state governments employ regulatory classification systems to guard against imprudent behavior (see box 4.1). These classifications differentiate among jurisdictions in allowable maximum outstanding debt or, more typically, the maximum debt outstanding in rela- tion to some revenue source, such as a property tax. In the United States most state governments differentiate among subnational governments through le- gal classifications that can include differential borrowing authority. However, in subsovereign financial systems being put into place for the first time or be- ing radically redesigned, classifications may be overused, poorly designed, or unenforceable. The strongest argument against rigid regulatory classification is that upward mobility in classifications of financial strength and managerial maturity should be encouraged. Classifying a jurisdiction in a way that en- courages it to depend on external assistance and avoid responsible borrowing on its own is exactly the opposite effect that government intends to have. Ar- tificially limiting market access runs counter to the basic policy goal of pursu- ing greater private sector investment. Subnational Borrowers by Type of Entity Subnational debt also may be incurred by municipal enterprises and quasi- municipal entities created by agreement of existing municipalities or by 56 Subnational Capital Markets in Developing Countries Box 4.1. Defining and Controlling Public Debt How public debt is defined can determine the boundaries of subnational government borrowing. EU legislation, which limits public indebtedness under the deficit and debt limits of the Maastricht Treaty, defines public debt as the debt of the central, regional, and local governments, including social security funds but excluding the debt of public enterprises. The limitation thus is expressed in terms of the institutional units producing non- market services as their main activity rather than in terms of ownership of the facility. A concern has been how to coordinate debt at the subsovereign level with that at the sovereign level. Subnational governments have an incentive to place as much of their debt as possible on a self-supporting, commercial basis to avoid macro-level curbs on borrowing. Evidently, however, the EU definitions also include certain contingent obligations that subnational governments might enter into in support of commer- cial debt, such as obligations to purchase a commodity or ser- vice (an off-take guarantee) and pledges to make up project op- erating deficits or debt service deficiencies from general funds. The curbs on general obligation tax-supported debt embodied in the EU limits are akin to the individual state-based limitations of tax-supported debt that arose in the United States. In the United States the restrictions on general debt hastened the rise of the non-recourse revenue-bond obligation that is used for enterprise activities and other forms of non-recourse obliga- tions such as the moral obligation bond. These limited obliga- tions, many of which are de facto supported by taxes and fees raised by the general government, once represented only a small fraction of municipal borrowing. They now typically make up 60 to 70 percent of all bonds sold in the United States. One application of special districts is in the use of business im- provement districts. These special taxing units levy a tax in ad- dition to the normal taxes and have the powers and personnel to address the special needs of downtown areas, especially dis- tressed areas, including extraordinary sanitation and public safety needs. The concept has caught on in parts of Europe and may be spreading to developing economies as well. Source: Petersen and Crihfield 2000. Subnational Governments as Borrowers 57 national or regional legislation. These special-purpose arrangements are of four types: · Separate restricted funds, accounting arrangements, or special-pur- pose entities within a municipality, the revenues and expenditures of which are restricted to specific purposes and are separated from the general fund. These entities typically derive their power from the mu- nicipalities, although they may have considerable independence. · Entities created by agreement among municipalities to accomplish a special purpose, such as to provide fire protection across a broad area. Their revenues and expenditures can be separated from those of the organizing municipalities. Their powers can derive solely from the municipalities ("joint powers") or through state or national legisla- tion that limits or extends such combining powers. · Quasi-municipal entities created by state or national legislation to provide municipal services (such as water development, disease con- trol, or transport services) where needs do not necessarily relate to municipal boundaries. Their powers would be described in authoriz- ing legislation. · Public-private arrangements, such as project financing, where gov- ernments and private sector entities share in the ownership of proj- ects that usually are built and operated by the private sector partner. These arrangements have been heavily advocated by reformers as a way to re-capitalize projects and make enterprises, particularly public utilities, more efficient. In theory, such special-purpose subnational governmental entities might issue limited obligation debt based on their own revenue sources and the ability to borrow against them. In practice, however, issues are more complicated. As might be expected, there are two sides to the special-purpose, special- entity borrowing coin. Establishing such entities can allow services to be delivered by an appropriate entity with targeted taxes, fees, and charges. That characteristic is appealing to those that favor an application of the benefit principle and rational pricing of services. Moreover, since geograph- ic areas of traditional general governments have typically inherited a polit- ical and economic logic that may be long out of date, the case for promot- ing special service districts along the lines of economic service areas is often compelling. On the negative side is the possibility of a proliferation and fragmentation of local government and of diffusion of local revenue 58 Subnational Capital Markets in Developing Countries sources. There are also questions of the nature of the relationship between the governmental parents and their special-purpose children, a relation- ship that may rely on subsidies and guarantees, stated or implied, and the exposures that accompany them. These issues are illustrated in the case of the People's Republic of China. The People's Republic of China presents something of an enigma: a highly centralized state that is loosely organized, with extreme variations in sub- national fiscal capacity and high levels of investment by companies owned by subnational governments that themselves cannot borrow. Although China is a unitary state, it has devolved a great deal of spending responsi- bility to its subnational units, which are both legion in number and, at the provincial level, as large in population as many countries. While the subna- tional governments are precluded from borrowing directly using their own credits, they effectively borrow through special-purpose vehicles, which are wholly owned companies that have their own revenues and often supply infrastructure needs on a quasi-commercial basis. Rationalizing the activi- ties of these "off­balance sheet" borrowers, which frequently have to rely on borrowing from state-owned banks, is a major challenge the country faces as it carefully enters into a regime of financial markets--and the world's financial markets (see box 4.2). The practical implication of this discussion is that subnational govern- ment borrowing powers should remain flexible enough to address both common and special infrastructure problems. An example is the special taxing and fee district, which may permit a unit of government to gear its taxing and charging powers to the particular needs of subdivisions, as is the case in the United States. However, care needs to be taken to ensure that the legal and operational arrangements are clearly stated and that dealings are both correct and transparent. Cooperation among Subnational Governments For many projects, financing and operation are more efficient when the scale is larger than an individual subnational government. In many cases the desire to provide more local self-determination has led to the establish- ment of many small governments that are assigned service responsibilities that exceed their fiscal and managerial capabilities and encompass service areas that exceed their geographic boundaries (see box 4.3). Cooperation is imperative if services are to be prepared in a rational way and capable of be- ing financed by users on a local basis. Subnational Governments as Borrowers 59 Box 4.2. China: Off-Budget Finance and the Transmuted Bond Under the series of changes in the intergovernmental fiscal sys- tem that have occurred in China over the past two decades, Chi- nese localities found it increasingly attractive to hive off many activities into the off-budget category and have them carried on by government-owned entities. Given the austerity in many subnational governments and the changing mechanics of tax- sharing, the local government-owned companies had the ap- peal of raising their own revenues, being kept away from the formal budget calculations, and being able to pursue activities either not allowed or not financeable by the subnational govern- ment itself. While information is incomplete, it appears that such off-budget activity is about equal to that carried on by the regional and local governments on their formal budgets and may represent as much as 20 percent of Chinese GDP. One appeal of the off-budget financing is the ban against subna- tional borrowing from nongovernmental sources on the local government's own credit. However, the special-purpose entities that they create and own can borrow. This is especially impor- tant in financing infrastructure and has resulted in a phenome- non known as the "transmuted bond." To access credit, a Chi- nese subnational government will create an economic entity, which has a close, if legally murky, relationship to the parent, to accomplish the financing through the sale of "corporate" bonds. In some cases, such as Quinyang district of Chengdu City, bonds are sold locally to retail investors, although the usual pur- chasers are banks and investment funds. The debt of these special purpose vehicles, whose proceeds fre- quently are re-lent to the government and repaid by governmental funds, is widely understood to be a contingent obligation of the parent government. Because this transmuted debt is not subject to an orderly process of approval (and financial oversight and report- ing), and in view of the prohibition against government guaran- tees, this debt is seen as constituting a substantial risk for both the financial system and for the underlying government debtors. Source: China case study, chapter 22. 60 Subnational Capital Markets in Developing Countries Many countries have achieved this goal through associations of subna- tional governments. Subnational governments often have legal authority to "collaborate or associate to perform public works"4 through contractual relationships among participating subnational governments. Even with such legal authority, cooperative projects still need a legal contractual framework to permit subnational governments to work together in a way that enables the jointly created entity to access financing and avoid the in- efficiency of separate financing of each government's share of the cost of a joint project.5 Box 4.3. Restructuring Subnational Government: From Few to Many (But How Many?) The path to the democratization of the formerly communist Eastern and Central Europe states has not been easy. Restruc- turing unitary systems of government to foster more self-gover- nance has led to a proliferation of subnational governments. In Hungary the number of subnational governments doubled after the 1990 reorganization, and the same pattern was seen in the Czech Republic, the Slovak Republic, and Ukraine. Meanwhile, the new units were given extensive service responsibilities un- matched by expanded local revenue-raising powers. In large part, the difficulty has been in deciding which govern- ing model to follow. After the fall of communism, territorial frag- mentation was greatest in the state systems that followed the Napoleonic (or Southern European) system. The central govern- ment maintained a strong local presence through the prefecture system of administration and an array of national services reaching down to the local level. A key responsibility of subna- tional governments was to represent local interests to the cen- tral government, which retained the major sources of revenues, doled out grants, and imposed national standards. An associated difficulty in devolution schemes has been the dis- regard of the optimal size of government needed to deliver local Subnational Governments as Borrowers 61 services efficiently. Traditional concepts of "community" often led to high levels of government fragmentation. The idea, again, was representing the locality to the center, as opposed to exer- cising true self-sufficiency. In the countries following the Northern European model (and Western federated systems), there was greater effort to achieve the optimal subnational government size needed to match as- signed service responsibilities and revenues. The central gov- ernment does not have a presence at the local level, and locali- ties have more responsibility for delivering local services and for deciding what those services should be and how much to spend on them. Reconciling the two conflicting views of the proper role of sub- national governments has been a big source of tension. Efforts by central authorities to promote regional cooperation have of- ten been resisted by new subnational governments that jealous- ly guard their new autonomy and local resources. Source: Davey and Gabor 1998. Notes 1. This is not to say it cannot happen. In Poland, the old state's adminis- trative districts (the Voidvoidships) were replaced by a new structure of counties, the Poviat. In other countries, there have been massive reorgani- zations and amalgamations. However, a new name on the government building is not the same thing as its being abandoned: somebody else picks up the duties and the liabilities. 2. Later, there will be a discussion of concessionary finance and techni- cal assistance. At this stage, the concern is with identifying the likely "po- tential market" for private sector capital access and under what conditions. 3. It should be noted that sale of bonds by local governments in the Philippines is restricted by law to finance "self-supporting" projects. 4. Fed. LLSG, Article 16, Bosnia and Herzegovina. 62 Subnational Capital Markets in Developing Countries 5. In Latvia, the Law on Self-Government determines the right of local governments to "cooperate." However, the legislation does not state that institutions commonly established by self-governments can be juridical persons with their own budget. Thus, there is a question whether the "joint entity" can borrow, which means that each participant has to borrow on its own. This results in an inefficient structure for jointly financed projects. Chapter 5 The Nature and Design of Debt Types of debt are defined by the kind of security given by the borrower. Creditors want to know not only where the money is expected to come from but also what their remedies and security are in the event of default. Knowing the security on debt is important: uncertainty about remedy and security can create risks that build inefficiencies into the market. If the remedies and security are not deemed adequate, markets may set risk pre- miums so high that credit is unaffordable to many jurisdictions. What remedies should be available to creditors by law? This question is critical. Any framework for subsovereign borrowing needs to spell out what powers a jurisdiction has to pledge assets and revenue streams and to exer- cise its powers to set taxes, tariffs, and other levies. It is also desirable to spell out how such security can be affected by default or other financial emergency. General Obligations, Special Pledges, and Limited Obligations In most emerging market economies general purpose subsovereign debt has had some form of sovereign government backing. In many cases the subsovereign governments were merely administrative units of the central government under a unitary government concept. In other cases the only long-term funds available were supplied by international lending entities that typically required a sovereign guarantee. However, this is changing. Devolution has meant that national governments are encouraging subna- tional governments to borrow on their own credit. Precisely what constitutes the credit of a subnational government, with- out an explicit or implicit guarantee by the national government, is often 63 64 Subnational Capital Markets in Developing Countries unclear. Expressions such as "general obligation" or "balance sheet" debt often mask an unresolved question of ultimate security: what remedies are available to an investor if a subnational government fails to pay on time and in full? Aside from the subnational government's good faith and the prospect of national government assistance if things get difficult, subsovereign general obligations have often been backed by the ability of creditors to seize finan- cial and physical assets. For a number of reasons, this physical collateral system is not a sound approach to securing credits. Subnational govern- ments with physical assets that are unrelated to their municipal service re- sponsibilities, such as a commercial enterprise, might be better off to divest themselves of the asset to avoid diverting scarce city management capacity to manage a potentially private activity. In addition to the problems in enforcing a claim on pledged public property are problems with title--does the locality actually own the prop- erty? The legal nature of the public domain continues to be unclear in many emerging and transitioning countries. Municipal assets that are used directly or indirectly to provide vital services should not be (and more of- ten are not permitted to be) risked as collateral. The pledging of physical collateral can divert the government's attention from making sure its gen- eral revenues are sound enough to support borrowing. Despite these draw- backs, there can be times and places where a subnational government owns non-vital property that is "alienable" and useful in bolstering its creditwor- thiness. The meaning of the general obligation pledge is also subject to varia- tion. The term full faith and credit originated in the United States and is generally understood to mean more than a general, unsecured promise.1 Debt not backed by specific revenue flows should be backed by a pledge of all general revenues as a source of debt service payment. The subnational government could be specifically obligated to use any and all of its general resources, including an increase in taxes and fees, to meet debt service obligations. Stronger and more specific remedies for creditors are likely to improve investor confidence in subsovereign debt in emerging market economies. Several kinds of limited security can be pledged to secure subsovereign debt: · Physical or monetary assets. · The right to operate a facility or provide a service. The Nature and Design of Debt 65 · Selected revenues, such as those from tariffs, fares, or rentals; particu- lar taxes or special levies; and grants or shared taxes (intergovern- mental transfers). · Power to set specific tax rates, utility tariffs, and other levies. · Executive agreement to budget for and recommend payment of fu- ture debt service, without an explicit binding pledge that those ap- propriations will be made.2 · Assignment of the payment of future intergovernmental transfers. · Pledge by a higher level government to exact certain penalties against defaulting lower level government borrowers.3 Pledging Assets, Operating Rights, and Revenue Streams A common pledge backing debt in developed markets is one that is restrict- ed to a particular revenue stream or enumerated subnational government assets or one giving the creditor the right to step in and perform the activi- ty and receive the revenues in the event of default. However, carving out specific revenues and giving the creditors rights to assets and operational powers both raise a host of operational and policy issues.4 Local officials may be hesitant to pledge assets because their loss in the event of a default would be dramatic. There may be assets or revenue streams that are so vital to maintaining basic governance that they should be protected from a debt service pledge. Examples include: · Intergovernmental transfers or local dedicated taxes that are intend- ed to provide services to selected segments of the population. · Transfers or taxes that are earmarked for mandated purposes. · Physical facilities deemed essential to the public health, safety, and welfare, such as water supply, fire equipment, and hospitals. Subnational governments have a limited number of pledgeable assets, especially non-vital properties. To the extent that governments are forced to pledge these assets to support debt, their future ability to secure loans is diminished and their financial flexibility is reduced. As a practical matter, private sector lenders are of two minds about asset pledges. Lenders are anxious to have as much collateral as possible to apply leverage to reluctant debtors. Lenders may be willing to take marketable as- sets, such as vacant land, office buildings, parking lots, or sports facilities, but they are not likely to foreclose on indispensable physical assets or on es- 66 Subnational Capital Markets in Developing Countries sential service facilities, such as water or wastewater treatment plants, town halls, city streets, or fire stations. If the legal system provides adequate assur- ance, lenders are more likely to secure local debts by pledges of actual or po- tential revenue streams that are sufficient to cover the debt service. Because of the practical and political problems of tying up essential fa- cilities, a prohibition on pledging properties considered essential to public health and safety could be included in authorizing legislation with little impact on a market's development of other useful security devices. Mini- mum essential services can be defined by law, with the borrowers deciding what fits the definition. These might include services necessary for human health and safety, such as water, sewer, and refuse collection. Use of the as- sets relating to these services could be pledged, but the law should require minimum essential services to be continued at all times. These restrictions would not be barriers to borrowing, since lenders are not interested in re- possessing pipes in the ground. They want revenue streams. A pledge of revenues from public utilities is appropriate for financing re- lated to the same utilities but not if the pledge is used to secure unrelated financing. Part of the concern is simple economics. When a jurisdiction subsidizes general expenditures at the expense of utility charges, resources are misallocated. The service that does the subsidizing tends to be under-al- located, and the service that gets the subsidies tends to be under-priced and thus over-allocated. Nevertheless, in cases where the demand for or the supply of the burdened service is relatively price inelastic, using revenues from that service is tempting since the costs of collection are low and the certainty of collections is high.5 Pledging to Set Tax Rates, Tariffs, and Other Levies Where tariffs, rates, or charges can be increased or decreased at the discre- tion of the subnational authorities, a rate covenant to set and maintain the charges at adequate levels to meet operating costs and pay debt service is a useful financing tool. However, subnational governments in emerging and transitioning economies usually have quite limited revenue-raising powers, a legacy of unitary states with a center monopoly on decisions and revenue raising.6 Even subnational governments with considerable power to set rates and establish levies can experience ambiguity about their ability to pledge to set tariffs, tax rates, or other charges at a level sufficient to service a debt be- cause of questions about whether such covenants unlawfully bind future The Nature and Design of Debt 67 administrations (see box 5.1). Without such a forward-looking and binding contract ability, a pledge is probably worthless. In many emerging market economies the primary cause of debt service default and payment arrears is failure to increase the rates and charges that were to be the source of revenues for debt payment. Subnational jurisdic- tions would benefit from clear legal authority to covenant future tariff or tax Box 5.1. Importance of the Rate-Setting Pledge South Africa offers an interesting example of the importance of the rate-setting pledge in a revenue bond and of the potential problems when its application is uncertain. Several South African cities are attempting to implement privatization plans that involve nonrecourse revenue bonds. Debt service payments on the bonds would rely exclusively on the water tariffs of the privately operated water treatment plants. The tariffs are likely to need to be increased over time to meet rising operating costs and offset unforeseen expenditures. However, national legislation gives a national minister the discretion to set water tariff rates, in effect overriding local control and contracts. Should that happen, the private concessionaires want the local communities to make up any shortfall in revenues by raising property taxes. Having such a clause in the contract would improve the credit- worthiness of the bonds, but it raises other problems. First, if wa- ter rates are not raised, then taxes must be, eroding the general tax base of the municipality. Second, the water ratepayers and the property taxpayers are not the same people. The great ma- jority of water users own little if any taxable property, and they greatly outnumber the better-off property taxpayers. This opens up the potential that local elected officials might even welcome a reneging on water rate increases by national officials, since it would shift the burden of the debt to the wealthier--and less nu- merous--constituents that pay the property taxes. Source: Petersen and Crihfield 2000. 68 Subnational Capital Markets in Developing Countries increases to secure debt. Jurisdictions may choose whether to use this mech- anism, but they should have the legal authority to make such a covenant. Intergovernmental Revenue Intercepts In many countries, subnational governments can assign to creditors their in- terest in specific revenue streams, such as shared taxes and grants, received from higher-level governments (box 5.2). Called revenue intercepts, these as- signments are attractive to creditors because of the promise of predictable revenue streams for paying debt service. Intercepts can be designed to ensure that adequate funds are available to meet debt service payments before they come due (an ex ante intercept) or to be tapped only in the event of a de- fault (an ex post intercept). Another variant is to have a bank "stand-by" credit facility to advance money should funds not be on hand to meet debt service payments, with that loan then repaid out of future intercept receipts. Some have argued that the pre-assignment of revenues to pay debt ser- vice tempts subnational governments not to budget for or pay debt service and induces intercept-protected creditors not to adequately assess the un- derlying worth of the investment being financed or the subnational gov- ernment's financial performance. If these problems are thought to be com- pelling, charging borrowers that routinely use the intercept to pay debt service bills a large penalty or an administrative fee would ensure that the subnational government is always better off collecting its own revenues and paying its own bills. Enterprise or "Self-Supporting" Limited Obligation Financing A common use for special pledges of revenues and assets is for a self-sup- porting enterprise that generates its own means of repayment, without re- lying on recourse to general revenues.7 This limited obligation involves the pledge only of revenues from a specified system or project for repayment. This implies the creation of a special fund to receive the revenues that will be expended to meet costs as- sociated with the enterprise, including debt payment. This concept focuses credit concerns on the viability of a particular project or system, rather than on the viability of the subnational government. It legally isolates cer- tain self-sustaining activities and projects from the general affairs and fi- nancial backing of the sponsoring government. Even poor or unsound gen- eral purpose jurisdictions can have viable enterprises.8 The Nature and Design of Debt 69 Box 5.2. Intergovernmental Transfer Payments as Collateral In many emerging market economies subnational governments are highly dependent on transfers from the central government for a major portion of revenues. While these transfers can be volatile, transfer intercepts are attractive for covering debt service payments. As a general rule, if intergovernmental payments are used for pledging, the historic or expected level of transfers should cover the debt service payments by a fraction greater than one. In the Philippines cities receive about half of all revenues and the provinces about three-quarters through intergovernmental trans- fers from the national government. The smaller and more rural the subnational government, the higher the proportion of transfers to total revenues. In the Philippines, government-owned banks (the de facto required depositories for subnational governments) have got- ten deeds of assignment of transfer payments to cover bank loans. As aid is received, the banks have a right of offset against any loan amounts owed the banks prior to dispersal for other purposes. Mexico recently enacted legislation that permits states and cities to sell debt secured by a master trust that holds federal tax partici- pation payments. Payments are made to the trust, which in turn pays out principal and interest to bondholders. Aguascalientes was the first Mexican city to issue bonds under the trust in De- cember 2001. The bonds were sold in the domestic peso market. Intercepts can have a powerful impact on subnational borrow- ers, especially small and remote governments. The assignment to bondholders of state payments to local school districts (which typically make up over 50 percent of revenues for the districts) is common in the United States. It is the basis for the high credit ratings enjoyed by local school districts covered by such programs. As a result of this widespread appreciation of the impact of state assistance and other small-borrower prefer- ences, local schools are among the lowest cost borrowers in the U.S. municipal bond market. Source: Authors. 70 Subnational Capital Markets in Developing Countries Enterprise financing has several advantages: · It establishes a relationship between the cost and the price of services, promoting more efficient operations. The cost-price relationship need not be absolute and can be modified, but it has the advantage of making any subsidy transparent. · If the utility has run a surplus to subsidize other governmental func- tions, then the added "tax" burden on utility users becomes evident. · Replacing general revenues that have subsidized enterprise opera- tions with dedicated revenue structures will free up general revenues for other purposes.9 · Management and operation of revenue-producing facilities tend to be more efficient and the facilities better maintained since they need to be in shape to produce revenues. This can be encouraged by con- tractual provisions protecting income and value, paired with credi- tors' active interest in assets and their operation. · When there are legitimate reasons to use general revenues as well as specific revenues, it may be better to use general revenues to reduce the amount of debt incurred. For example, this can be done by mak- ing a municipal "equity" investment in the asset up front, and bor- rowing to build or acquire the rest of the asset, pledging only rev- enues produced by the asset, or even a part of the asset's operations, to meet debt service requirements. This practice is common for many municipal utility operations in Western Europe. Limited obligation, self-supporting financing also has several disadvan- tages: · The expressions asset stripping or security dilution convey the concern of existing creditors of subnational governments that have relied on a utility to generate subsidies for the general fund when those rev- enues are instead peeled off and pledged to a utility-specific pur- pose.10 Where prior lenders have looked to the overall revenues as a source of repayments, a subsequent sequestering or stripping away of revenue streams weakens the credits and creditworthiness of the ju- risdiction. · Limited obligations may impede redistribution of infrastructure and services among population groups (for example, from better-off groups to poor ones) by keeping potentially redistributable revenues The Nature and Design of Debt 71 for the benefit of an already privileged area. Preferential and redistrib- utive policies typically require financing from general funds. · Enterprise financing is a contract between the public sector acting on behalf of the enterprise and the investor, who typically requires re- strictions that reduce the financing options of borrowers in the fu- ture.11 For example, borrowers must meet certain conditions before issuing more debt secured on the enterprise earnings (additional bonds test), must conform to certain requirements about reserves and insurance, and must abide by a rate covenant. To improve creditworthiness and expand revenue sources, some subna- tional governments have used utility surpluses to subsidize the general bud- get.12 However, transparency implies that utility surpluses used for cross-sub- sidization should be identified in a specific tax or surcharge that reflects the added cost of cross-subsidies. Without this transparency, it is not possible to see whether the utility is operating at an economic optimum in getting the most delivered service per unit of input. Having reached that optimum, the redistribution becomes a clear added cost for some and a benefit for others. In addition to the traditional "natural monopolies," such as public utilities provided by subnational governments, other candidates for complete or par- tial financing through revenue bonds are more commercially oriented rev- enue-producing activities, such as transportation terminals, public markets, farm processing plants, industrial estates, tourism facilities (including hotels), and toll roads and bridges. Critical to their suitability to revenue bond financ- ing are the reliability and growth of revenues, the technology used, the facili- ties' adequacy, and construction costs and future operating costs. Determin- ing risks in these technical and economic factors requires engineering studies and market demand studies to obtain objective estimates of the net revenues available to pay debt service (see the section on grant and loan integration in chapter 12). Especially for new, free-standing projects with no operations ex- perience, failing to do engineering and feasibility studies or not having them performed objectively by skilled professionals can lead to severe problems for the sponsoring subnational government, especially if that government pledges its own credit as part of the security (box 5.3). Special District Financing Special district financing is a variant of enterprise financing. A special dis- trict is created to provide infrastructure and services to a subset of the pop- 72 Subnational Capital Markets in Developing Countries Box 5.3. Importance of Feasibility Reports: The San Pedro Sula, Honduras, Sports Complex Projects that are intended to be self-supporting should generate sufficient revenues to pay for their operation and to meet capi- tal costs. In many cases projects are monopolies--because they are essential in a technical sense (water and electricity), exclu- sive in location (toll roads and bridges), or subject to a high de- gree of market control through government regulation (solid waste and parking facilities). Other projects, such as sport or cultural venues, are not essential, are subject to competition, and face greater market demand risks. In all cases facilities may be subject to construction and technological risk, such as cost overruns, startup delays, or failure to produce output of the ex- pected amount or quality. Assessing these factors and associated risks is the role of the engineering and marketing studies conducted to establish a project's feasibility. Emerging market economies often lack the technical skills needed for engineering and market demand studies and the independence needed for objective analysis. The difficulties of separating project promotion from technical analysis--and the unfortunate consequences of not doing so-- are reflected in the fate of the sports complex built in the munic- ipality of San Pedro Sula, Honduras. Expected to largely pay for itself, the complex was built to host the Central American Games. To partially meet expected costs of $25 million, some $15 million in bonds were sold, which were expected to be off- set by a variety of revenues. The project ended up costing $36 million, and net project revenues fell far below expectations. The city made a general obligation pledge in addition to the project revenues, and it is now in serious financial difficulty. The issuance of the bonds was not the problem. It was the lack of analysis that permitted the city to take on large and unknown risks. Among other problems, the feasibility study failed to do the following: The Nature and Design of Debt 73 · Analyze the market for "special-seat" sales, although these were expected to generate nearly 80 percent of the oper- ating revenues. · Identify the assumptions used in the construction esti- mates. · Examine alternatives, such as upgrading an existing stadi- um. · Consider using the private sector and more equity in the construction project. · Identify the various risks or contingencies if the complex's revenues were not realized. In short, there was never a credible assessment of the project's economic prospects nor of the impact on the guaranteeing mu- nicipality's finances. Source: Kehew 2002. ulation or geographic area that demands special types or levels of service.13 Special districts have been used to provide urban services (such as water, sewer, and roads), to areas that are developing rapidly or that have special needs (such as downtown areas). They are common in Western Europe and the United States and are beginning to appear in developing countries. As noted in chapter 4, the proliferation of subnational governments makes the need for cooperative ventures in project financing especially important among small governments. If there are special benefits that can be ascribed to a particular area, the special district provides a mechanism for recovering the costs associated with the benefits. Special districts can transcend political boundaries or unite jurisdictions into a single financing unit to provide a regional service. Some types of new developments such as public utilities and transporta- tion, storm drainage, and parks increase the attractiveness of an area and enhance property values, as well as other indices of economic activity and worth. If the costs of the capital improvements are borne by the public sec- tor, the public sector should have some way to capture its investment. A 74 Subnational Capital Markets in Developing Countries special district can do that by adjusting its taxes or charges to pay for capi- tal improvements that benefit specific properties.14 A successful special tax relies on good and timely measurement of values and an efficient collection system. In the United States, for example, some special taxing districts are administered by private for-profit organizations that undertake the calculations and do the tax billings as agents for the governments. The entire revenue-raising mechanism is meant to support obtaining credit and is specified in the loan or bond agreement. This im- proves the administration of taxes, and the integrity and efficiency of the system becomes, in effect, a matter of contract with bondholders. In West- ern Europe the special district or special authority covering all or parts of more than one political jurisdiction has facilitated the subsequent privati- zation of services, such as water utilities in France and the United King- dom. Notes 1. In the United States the term full faith and power means the applica- tion of the general taxing power to the repayment of the debt. That power in its traditional and strongest formulation has meant the imposing of tax- es "unlimited as to rate or amount" sufficient to repay the debt. As a practi- cal matter general obligation defaults are almost unheard of and are promptly cured by a mandamus from a court to levy taxes and the inter- vention of state governments to make sure that happens. States in the United States are usually very sensitive to the risk of getting a bad reputa- tion because of the failure of a local government to pay its debts. Default by a general government means the loss of local governing powers, sometimes with the appointment of a control board or a receiver to take over opera- tions until the debt is resolved. 2. This type of security is known in the United States as appropriation or moral obligation debt. It recognizes that the debt is not a full faith and credit binding obligation but rather is subject to the will of successive legislatures. Its origins are lease rental debt that holds that the obligation runs only from fiscal year to fiscal year and is subject to legislative reconsideration each year. 3. This is a seldom applied but potentially useful approach. The senior level does not commit to pay the creditor directly; rather, it agrees that it will withhold payments from the locality. It in effect avoids a contingent claim by the private party on the funds. The Nature and Design of Debt 75 4. The Argentina case study, chapter 14, discusses two pledges that were used in tandem: intergovernmental transfers from a specific asset (hydro- carbon). 5. There are other cases to be made for taxing utility consumption. In many cases, the consumption can be used as a proxy for income. In others, the ability to piggyback on the billing process lowers collection costs. Last, the ability to shut off utility services provides a powerful means of enforce- ment. In many countries, however, shutting off utilities to non-payers is ei- ther illegal or extremely unpopular. This is particularly the case where utili- ties have been provided for free or heavily subsidized. 6. Problems of local revenue raising are particularly acute in transition- ing economies. In the communist system taxes were buried within the state-owned corporate system and were frequently negotiated and changed by administrative fiat. Since the taxes were at the corporate level, citizens were unaware of the burden and have often resented the adoption of visi- ble, explicit taxes (see Estirn 2002). 7. This is the traditional notion of the enterprise revenue bond. Many variations have been designed by states in the United States to circumvent restrictions on tax-supported debt. This approach encourages the alloca- tion of the full costs of services to the beneficiaries, which is desirable eco- nomically because it leads to efficient allocation of scarce resources. 8. During the Great Depression of the 1930s in the United States, some states defaulted on their general obligation securities but continued to pay on revenue bonds supported by the motor fuel tax. People and businesses would forgo paying property taxes (on which states relied heavily) while continuing to use automobiles and purchase fuel. States subsequently shift- ed their tax systems to rely more on the sales tax. 9. In some places, such as in South Africa, the subsidy runs from the util- ity to the general fund, rather than the other way. 10. However, most economists would applaud this elimination of the cross-subsidy on efficiency grounds. 11. For example, there may be requirements that the borrower not pledge the same asset to another lender, except under stated conditions, that the revenues provide certain coverage of the debt service (rate covenant), and that revenues be retained for use on the facility and to ben- efit bondholders (closed loop). Negotiation of these restrictions and the as- sociated tests is an integral part of the borrowing transaction. 12. Subsidies can be hard to detect. Where the utility is part of the gov- ernment, the allocation of costs can be highly judgmental. A government 76 Subnational Capital Markets in Developing Countries may allocate many of its administrative and other costs to the utility or it may receive utility services below cost or for free. 13. A district may be "dependent" and overseen by the governing body of the municipality or "independent'' with an autonomous elected or ap- pointed board. In many areas storeowners or homeowners form associa- tions to manage the district and levy charges. The key is the ability to levy taxes and charges and to seize properties that do not pay. 14. In its most common form in the United States and a few places in Europe, the tax district uses property taxes (percentage of taxable property value) or assessments (fixed dollar levies). However, the district can use other bases to charge for the benefit or service, including square footage (or meters) or front footage (or meters), number of vehicular trips (for roads), impervious surface (for drainage), lumens of light (for lighting), residential bedrooms (for educational facilities), square footage of space (for parking), and so forth. The key is that there be a logical connection between the im- provement and the form of charge that is used. Chapter 6 Debt Instruments and Methods of Sale Debt instruments are the legal embodiment of a credit transaction, setting out the terms and conditions of the loan, including how the principal is to be repaid, how long a debt will be outstanding, and how interest is figured and paid. Method of sale considerations involve the procedures by which debt is offered to the final investors and the debt obligations exchanged for the bond proceeds. The general parameters of what instruments should look like and how sales are conducted are often covered in a nation's securities laws. As a rule, the pre- cise details of these matters are determined by the market. Financial markets are fluid, and what might be attractive one day can be unattractive the next. Inflexibility is costly. However, in new markets both the borrowers and lenders are often unaccustomed to the process and perhaps unwary of the risks. A major concern at the national level is to avoid creating regulations that interfere with the flexibility of lenders and borrowers in structuring debt in ways that best suit both parties. This chapter examines several of the alternatives that may be used in the design (often referred to as structur- ing) of subnational government debt transactions. It describes debt struc- ture and illustrates the range of instruments available to suit the profiles of issuers and investors. Maturity or Term of Debt The maturity of a debt instrument refers to the period from the time the funds are borrowed to the time the principal is due to be repaid. The maturity should be matched to the economic life of the asset that the debt is financ- ing. Ideally, the amortization of the liability on one side of the balance sheet is matched by the depreciation of the asset financed on the other side. Thus 77 78 Subnational Capital Markets in Developing Countries infrastructure assets, such as water systems, roads, or municipal buildings, which typically have lives of 15 to 30 years, should be financed with long- term bonds of similar duration. Matching asset life to debt term is also sound public policy because then facilities can be paid for by those who use them. In many emerging market economies, however, private investors are un- able or unwilling to extend loans beyond a few years. Even if longer term capital is available, the upward sloping yield curve--the longer the term of the debt, the higher the interest rate payable--may cause borrowers to pre- fer shorter-term debt. Investors want extra compensation for the lack of liq- uidity of long-term lending and the increasing uncertainty about economic conditions, price levels, and interest rates far into the future. However, this is not always the case. Short-term interest rates may be temporarily driven up by liquidity shortages and efforts to defend the currency. If expectations are that the prevailing level of interest rates is unsustainably high, and if rates are expected to fall, then the yield curve may be inverted, with short- term rates higher than long-term rates. In such cases, some borrowers may borrow on a short-term basis, if they believe long-term rates will fall. Oth- ers may choose to lock in the relatively lower long-term rates. There is also a tradeoff between the lower rates typical of short-term debt and refinancing risk. If the debt is shorter in maturity than the life of the asset, the borrower is exposed to refinancing risk--new debt may have to be raised during the life of the asset at a higher rate than the original loan. If the borrower's credit risk has worsened, it may not be possible to re- finance. Refinancing can, of course, work in favor of the borrower, if, for example, interest rates fall or the borrower's credit improves. In the case of general obligation bonds, this could happen as a result of the improved general creditworthiness of the subnational government. In the case of project finance, the construction and initial phases of operation are riskier than the later phases of a mature project, when it may be possible to refi- nance at lower rates. However, financiers are aware of this and rely on the later phases to provide some compensation for the additional risk taken at the outset. Thus they would probably reserve for themselves the right to re- finance. All in all, maintaining an unhedged position is risky and usually not advisable with public funds. Debt Service or Repayment Structures There are several common cash flow profiles of debt, which describe the ways a borrower pays interest and principal over the life of the liability (fig- Debt Instruments and Methods of Sale 79 ure 6.1). In addition, interest rates may be fixed or floating and bonds may pay interest on a variety of "coupon" dates.1 Loan Structure and Cash Flow Profiles The debt service (that is, combined principal and interest) may be paid in approximately equal installments over the life of the debt, which is called Level debt service Debt service Interest principal Time Level principal Debt service Interest principal Time Term loan (or "bullet") debt service principal Interest Time Figure 6.1. Debt Service Structures 80 Subnational Capital Markets in Developing Countries level debt service. Another, more conservative approach is the level principal structure, in which the principal is repaid in equal increments and interest in declining increments, leading to a more rapid repayment of debt. This front-end-loaded structure frees up future borrowing capacity quickly and leads to progressively smaller debt service payments. Alternatively, the debt service schedule may be structured to increase over the life of the debt. A term bond structure typically has periodic interest payments but the princi- pal falls due at the end. This back-end-loaded structure, sometimes called a bullet loan, is common in short-term securities and bank loans. The variations on loan structures are practically limitless. Their shapes can be influenced by grace periods, deferrals of payment of the principal or interest or both for periods of time. Such structures are used when loans are to be paid from project earnings and there is a construction or start-up pe- riod before receipts start to flow. Original discount bonds, called zeros when they fully discount future interest payments, pay no or reduced interest. The investor realizes a return by buying the bond substantially below its principal value. Such bonds can be issued at discount or created syntheti- cally by investment banks by stripping the coupon off a standard term or serial bond. Zeros are attractive to parties who want to secure a fixed amount of capital in the future without being exposed to reinvestment risk. Zero bonds are created synthetically when the coupon stream is stripped from a bond and sold to an investor who is interested primarily in an annuity flow. Cash flow profiles can be engineered to match the cash flows generated by the activity being financed. Liabilities can be index linked, where rev- enue flows are expected to vary with an index, such as inflation or an input cost. Interest payments can go up or down, depending on the movement of the index. As noted, amortizing payments can be structured with an es- calating profile, with lower debt service in the early years. This is common in commercial property finance, for example, where there is a "ramp-up" period when rentals are expected to escalate, and can be appropriate for certain municipal assets. Similarly, interest payments for initial periods can be deferred by using bond proceeds to pay interest costs in early periods (capitalized interest). Fixed or Floating Interest Rates Bank loans or municipal bonds may be made at fixed or floating rates of in- terest. In emerging market economies, the variable rate may be the only in- terest payment structure available for obligations beyond a short maturity. Debt Instruments and Methods of Sale 81 Both have advantages and disadvantages. Floating rate debt implies contin- uous uncertainty about the cost of debt, but it can be appropriate where the matching revenues are expected to vary with changes in interest rates. However, this is not usually the case for municipalities. Financial flexibility and access to liquidity are important considerations for floating rate bor- rowers. If there is limited ability to change taxes or rates to respond to ris- ing interest rates, then over-reliance on variable rate debt is worrisome. The rating company Standard and Poor's generally recommends that the com- bined short-term debt and variable rate debt not exceed 20 percent of total debt, but the share depends on the circumstances and degree of flexibility and matching of revenues with debt service.2 Cash Flow Concerns There are several considerations in deciding on the cash flow of municipal bonds. Bonds may pay interest on a variety of "coupon" dates. Although semiannual payments are the most common in developed markets, struc- tured loans can have varying coupon profiles (semiannual, quarterly, even monthly) to suit the cash flow requirements of the borrower and the capac- ities of the issuer. Most municipal bonds in emerging markets have had short maturities and many have had term bond or bullet maturity struc- tures, meaning that most loans to subnational governments have been for construction and start-up costs. Implicit in the repayment structure has been the requirement that the borrower roll over the loan into a new one at maturity or come up with alternative means of long-term financing. This approach subjects issuers and lenders to great uncertainty about future debt service requirements and effectively holds borrowers hostage to future changes that may be forced on them when they come back to the market to renew the loan. Legal Restrictions A final area of policy regarding the structure of instruments concerns re- strictions that may be placed on interest rates or on the maximum maturi- ty of bonds. Interest rates may be capped by "usury rates" that set an ab- solute ceiling on rates. While this was once common practice in the United States, the restriction has disappeared for all practical purposes over the last 20 years. Limiting interest rates has the effect of rationing capital away from governments during periods of high interest rates. Such restrictions continue as a matter of contract in variable rate instruments, however, where a cap is specified or a borrower may purchase a rate cap contract 82 Subnational Capital Markets in Developing Countries from a commercial bank that will agree to pay the excess interest for a fee. The other common restriction is on maximum maturity of bonds, which is often specified in conjunction with the expected useful life of the improve- ment being financed. Again, these restrictions are seldom effective and the market itself provides the limitation on how far it will extend debt, espe- cially at fixed interest rates. Methods of Sale for Securities Municipal securities can be sold to investors in a number of ways. Bonds can be auctioned competitively to the highest bidder or placed with the fi- nal investor, much as a direct loan is made from a bank. In most emerging markets the offering is made through negotiation, with the borrower sell- ing its bonds through a financial services firm (such as an investment bank- ing firm or, for larger issues, a combination of firms, called a syndicate). The firm underwrites the issue, agreeing either to buy all the bonds offered at a certain price or to act as an agent and make a "best effort" to sell the bonds, receiving a commission on the bonds sold. Characteristics of Markets: Setting Interest Rates and Other Terms A competitive sale environment requires an active market with a large number of issuers offering fairly standardized securities and a large number of investors interested in owning them. The large volume of activity results in a number of bankers following the market, making bids, and placing bonds to investors. It also means that there are other professionals who help to design the issues, prepare documents, and run the auctions.3 The competitive auction, with several underwriters bidding on a bond issue, is common in the U.S. municipal bond market but a rarity in other markets. It may, however, become more prevalent as markets thicken in activity and experience develops.4 A strong point in its favor is the transparency of the transaction, since barring collusion among bidders, the public auction clearly identifies how the bonds are priced. Where bond markets are less homogeneous and sales are irregular, is- suers typically rely on negotiations, hiring an underwriter to help prepare the issue and seek out possible investors. The negotiations can be made competitive by injecting elements of competition among firms into the un- derwriting selection process and subsequently by holding underwriters to the projected terms of the issue. To help achieve competition, the issuer may employ the services of a financial adviser knowledgeable about the de- Debt Instruments and Methods of Sale 83 sign of transactions and the marketing of securities. The adviser usually helps the issuer select an underwriter and, among other tasks, helps to en- sure that the issuer is being dealt with fairly by the underwriter (box 6.1). The underwriting process has the advantage over the use of a best effort marketing arrangement of guaranteeing that sufficient funds will be bor- rowed. However, the investment banker undertakes the risk of reselling the issue and demands more remuneration when acting as an underwriter than when acting as a placement agent. To make a profit and to cover risk and expenses, the underwriter buys the bonds at a discount--for less than the value at which they are reoffered to the final investors. This price difference is known as the spread. The mechanics of selling bonds and setting interest rates and other terms differ for various domestic securities markets. In countries with rela- tively small and inactive markets, the terms of the bond offering may be set well in advance of the sale date. The bonds then may be sold on a given day with a discount or premium to make returns competitive with then- prevailing conditions. Fixing terms before the sale date puts the under- writer at greater risk, so issuers pay an interest rate premium. Another ap- proach is to commit to having the bonds underwritten at a certain mark-up or in relationship to some regularly published interest-rate index, usually that on government bonds. Finally, the terms can be determined by offer- ing the bonds at a proposed structure and then changing the terms to meet the effective demand from investors in what amounts to an "informal" auction. The terms and their acceptability to the issuer remain open until the sales contract (the bond purchase agreement) with the underwriter is signed. The bond instruments or other evidence of ownership then are deliv- ered physically or electronically and money is exchanged for them (settle- ment). Depending on market conventions and the nature of the security, the issuer or the underwriter may have selected a paying agent or a trustee to receive funds from the issuer and to pay interest and principal. The trustee also oversees the bond contract between the issuer and ultimate buyers of the bonds, the investors, and looks after the interests of the in- vestors, making sure that the terms of the bond contract are observed. Importance of Impartiality and Transparency Beyond the procedures set down by national enabling laws, the specifics of the bond offering are a matter of contract among the underwriter, the is- suer, and the ultimate investor. Thus the issuer needs to obtain sound legal 84 Subnational Capital Markets in Developing Countries Box 6.1. Selecting an Underwriter through Competitive Negotiation The city of Krakow proposed a 15 million zloty bond issue in 1996. With the assistance of a financial adviser, the city sent a solicitation to a large number of investment banking and com- mercial banking firms, describing the project and needed funds, providing information about the city, and asking for proposals. The solicitation and selection process contained several ele- ments designed to make the choice of firms transparent and competitive. The solicitation contained a tentative maturity structure for the issue and asked respondents to price the bonds (provide interest rates) and indicate their gross profit, as- suming that the bonds had been sold on a given day. In addi- tion, the respondents were asked to estimate an itemized list of costs and to indicate which costs would be met from their prof- its and which would be paid by the city. Firms were asked to cri- tique the structure and suggest alternatives and to describe their experience and financial capacity. A combination of factors was used in selecting the finalists, but the cost of borrowing was the most important. All costs, includ- ing future interest payments and fees paid by the city, were made comparable by using an all-in-cost internal rate of return calculation. Responses were analyzed by a committee, and indi- vidual firms were contacted to clear up any questions. Of the eight firms and syndicates that responded, the top three were invited to make presentations and to make their best and final offer. A syndicate was selected. The final offer committed the underwriting syndicate to price the proposed bonds on a par with Polish Treasury bonds of the same maturity, a highly ag- gressive bid. Subsequently, Krakow received an investment grade credit rat- ing from Standard & Poor's and sold bonds (Deutsche mark de- nominated) in the Euro market in late 1997. It was the first Pol- ish city to do so. Source: Petersen and Crihfield 2000. Debt Instruments and Methods of Sale 85 and financial advice that is independent of that given by the underwriter and the final investor. The transparency of the method of sale matters. The large amounts in- volved in bond sales and the ability of financial firms to make large profits on bond issues can be a temptation for corruption (see box 6.2). Investments Related to Borrowing An important element of subnational government borrowing is the types of investment that are permitted for the following. · Proceeds of a borrowing that are awaiting application to the intended purpose. · Funds held for paying debt service, including intercepted funds and reserve funds. Box 6.2. Rigging a City's Bond Sale Saõ Paulo was a heavy borrower in the Brazilian bond markets. As of January 2000 its outstanding debt was over 10 billion reais or nearly 1,600 reais per capita (equivalent then to about $800), some 20 times the average debt of Brazilian municipali- ties. The city's large appetite for borrowing was driven by more than its fiscal needs. In early 2000, when the national government was negotiating an arrangement to allow the city to refinance its debt, a scandal broke out involving corruption in previous city bond sales. The mayor, who was formerly the city's finance officer, was accused of having rigged past bond sales. He sold bonds at a steep dis- count to a select group of underwriters and then participated with them in the profits when the marked-up bonds were reof- fered on the open market. The mayor was removed from office by court order in March 2000. Source: World Bank 2001. 86 Subnational Capital Markets in Developing Countries In the case of bonds, such funds should be held in a custodial arrange- ment, segregated from other funds of the subnational governments, and invested with minimal credit risk exposure.5 The funds must be available when needed for their intended purposes, and there should be no market risk associated with liquidation of the investments. The legal framework for subnational governments is often silent on the parameters for investing bond-related funds, sometimes with unfortunate results.6 Too much rigidity, as, for example, requiring that bond proceeds be held by the national treasury in non-interest-bearing accounts, can make bond issuance less attractive and more awkward to structure efficiently.7 Reg- ulation of allowable investments may be desirable, balancing flexibility and the need for prudent investment instruments in a changing environment. Notes 1. A useful guide to concepts and terminologies used in designing and mar- keting subnational debt, along with many illustrations, is World Bank 2002c. 2. In the United States, where there is an upward sloping yield curve from short-term to long-term maturities, there has been a reward of 100 to 200 basis points for using variable-rate instead of fixed-rate debt. Debt typically can be called at the reset date, allowing flexibility to restructure debt. 3. Other professionals include financial advisers, legal counsel, auditing firms, and printers to produce the documents. There also may be banks to handle the investment of proceeds and to oversee payments under the debt contract (trustees and paying agents). 4. In Romania, for example, some cities are beginning to solicit bank loans on the basis of "bid sheets" that set forth the structure and terms that the city seeks and then asks for the respondent to fill in the interest rate. A motivating factor is the law on procurement, which generally forbids ac- quiring goods and services without a competitive bidding process. 5. In South Africa, municipalities may invest in a relatively short list of invest- ments, including bank deposits and government securities (LGTA, Section 9). 6. In the 1998 Odessa bond issue in Ukraine, the proceeds of the bor- rowing were invested in the Ukrainian interbank market at negative arbi- trage, many of the proceeds were unaccounted for, the projects were not completed, and the city defaulted on payment on the bonds. 7. In Romania local general governments must deposit funds in non- interest-bearing national treasury accounts, while enterprise funds can use private bank accounts. Chapter 7 Restrictions on the Issuance and Use of Subsovereign Debt Most national governments place restrictions on the use of debt by subna- tional governments, thereby substituting national policy for local flexibili- ty and the regulating effect of markets on municipal borrowing. As chapter 3 shows, there are several regimes for regulating debt, but in most cases there is an overarching set of rules that governs subnational debt issuance. National regulations typically cover the authority of subnational govern- ments to borrow and restrict the purpose of borrowing, the maturity, the amount of borrowing or debt outstanding, the use of proceeds, and the type of security given or recourse available to the lender. While the differ- ences can be arbitrary, short-term debt is generally that due within one year or less, and long-term debt is anything due more than a year after it is incurred. Subnational government guarantees, which ought to be treated like any other debt, also may be subject to regulation. Subnational Government Authority to Borrow Subnational government autonomy is generally based on principles set forth in the national constitution,1 although the laws on subnational gov- ernment borrowing are often scattered across the legal landscape, reflecting the fact that defining such activities is frequently an afterthought. Deter- mining a subnational government's legal authority to borrow and the asso- ciated legal parameters can require reconciling conflicting laws, regula- tions, and decrees. Explicit authorization and procedural requirements are essential, especially where these governments have had no experience with issuing financial obligations that are valid, binding, and enforceable. 87 88 Subnational Capital Markets in Developing Countries Some people argue for a minimalist approach to subnational debt legis- lation that gives authority to the minister of finance or other central gov- ernment authority to issue regulations, to be approved by the government, so that regulations can be adapted readily to experience and circumstance (Glasser 1998). This approach may provide some flexibility in an emerging market and in a changing environment. However, over the long term, all legal criteria and conditions for borrowing should be expressly contained within the legal framework, whether as law or regulation. In Indonesia the implementation instructions for subnational govern- ment borrowing under Law 25 mandate that donors and external govern- ment lenders conclude direct agreements with subnational governments, but the agreements must be cosigned by the Ministry of Finance. External creditors do not have an explicit right to secure their debt with the general allocation grant intercept mechanism or the right to a sovereign guarantee, but these security structures can be negotiated with the Ministry of Fi- nance--an opaque requirement that allows for considerable political inter- ference in the approval process (see Indonesia case study, chapter 25). Binding Nature of Debt A frequent issue in developing credit markets is the concern that a commit- ment by a subnational government may not bind a subsequent govern- ment. Even if the problem is one of market perception rather than law, the lack of clarity about who exactly is bound and for how long can create un- certainty about the political commitment of succeeding governments to re- pay the debt (see box 7.1). For long-term finance to be available for subna- tional investment, capital markets must be confident that a financial obligation is binding on succeeding governments. In Bulgaria the law prohibits a municipal council from contracting debt or extending short-term interest-free loans within six months of the expira- tion of its term of office (Municipal Budgets Act, Article 40 [4]). The intent is to prevent the issuance of debt for politically popular projects that may win over the electorate while encumbering the municipality with excessive debt that will be binding on the succeeding municipal council. While this requirement has little practical effect now, given the limited municipal bor- rowing activity, as the municipal credit market matures it could prevent a subnational government from taking advantage of commercial financing during a period of low market rates. If long-term debt financing for subnational investment is to become widely available, capital markets must have confidence that succeeding leg- Restrictions on the Issuance and Use of Subsovereign Debt 89 Box 7.1. The Philippines: How Political Risks Can Inhibit Municipal Credit Markets In the Philippine city of Cebu a newly elected mayor publicly questioned whether his administration would be bound to hon- or a debt incurred by the prior council. The mayor eventually withdrew his comments, and the city paid the debt on time and in full. However, the financial community lost confidence in the city, and as a result lenders have been inclined to limit loans and bond maturities for subnational governments in the Philip- pines to the current administration's term of office. To counter this maturity limitation, some subnational govern- ments have held voluntary referendums to demonstrate popu- lar support for specific project debt financing and thereby over- come financial institutions' fears of the political risks associated with long-term lending. Source: DeAngelis and Dunn 2002. islative bodies will honor the financial obligations of their predecessors. This principle should be explicitly affirmed in any subnational debt legisla- tion. The governing law should specify the binding nature of subnational obligations to repay duly authorized debt (see examples of France and Ro- mania in box 7.2). Authorizing and Approving Subnational Debt Subnational government borrowing can be approved by the subnational government executive or governing body; the community at large through a referendum; or state, provincial, or national authorities. Each approval mechanism can be conditioned by a variety of considerations, including the financial capacity of the borrower to repay debt, the purpose of the bor- rowing, the form of borrowing, and its consistency with national econom- ic policy. Most mechanisms also act as a curb on local officials' prerogatives by enforcing certain disciplines and limiting their authority to borrow. 90 Subnational Capital Markets in Developing Countries Box 7.2. Examples of Language on the Binding Na- ture of Financial Obligations France, Code Général des Collectivités Territoriales (Article L.1612-15 ) · The only obligatory expenditures of subnational authorities are expenditures necessary to pay debts that come due and expenditures that have been expressly determined by law. Romania: Local Public Finance Law, 1998 (Article 48) · Local and judet councils and the General Council of the Municipality of Bucharest can approve the contracting of internal or external loans, for a long or a medium term, for public investments of local interest, as well as for refi- nancing the public debt, under the provisions of this chap- ter. · Local and judet councils and the General Council of the Municipality of Bucharest may decide upon contracting loans by the vote of at least two thirds of their members. · The local public debt incurred under the provisions of para- graph (1) represents a general obligation which needs to be reimbursed, according to the agreements concluded, from the sources available to the territorial administrative unit, with the exception of special purpose transfers from the state budget. Source: DeAngelis and Dunn 2002. Approval by the executive. Authorization often depends on the maturity and size of the borrowing. Authorization by the chief executive seems ap- propriate for relatively small amounts and for the short term, where the fu- ture financial health of the local jurisdiction is not at risk. Large amounts or long-term borrowing should require authorization by legislative act of the governing body. Longer term borrowing involves trading off future fi- nancial flexibility in exchange for investment capital today. Restrictions on the Issuance and Use of Subsovereign Debt 91 Approval by the legislature or the community. For large amounts of debt, a local governing body should determine the key borrowing issues: for what purpose and for how much? Experience suggests that without local govern- ing body approval, the probability rises dramatically that debt will be repu- diated or that the tax or tariff changes needed to meet debt service obliga- tions will not be enacted. Furthermore, public debate on debt policies and plans helps keep the process open and visible (see box 7.3). Legislative approval can range from simply approving the borrowing as part of the budget process to voting and authorizing a particular transac- Box 7.3. The City of Cebu in the Philippines Considers a Deal Subnational governments may lack the knowledge and proce- dures needed to fend for themselves as borrowers with ex- panded opportunities. The experience of the city of Cebu, Philippines, illustrates what can happen when politics and me- dia influence local financing decisions. The city and its mayor were actively seeking financing for a ring road, a core element of a development plan. In mid-1998, two firms, one based in Hong Kong and one in Austria, proposed that the city enter into a $500 million loan and $75 million letter of credit to fund the ring road. The mayor liked the idea, and the firms, eager to fa- cilitate the process, paid several councilmen from the mayor's party to travel to Europe to see, secretly it turned out, examples of what the firms had financed. The councilmen signed letters of intent to enter into an agreement in three months. Subsequently, things began to get sticky. Political opponents asked whether this was a scam. The local representative of the investment firms was arrested on 32 counts of passing bad checks and after posting bail skipped town for Hong Kong. Lo- cal bank officials indicated that three signatures had been forged on documents shown in the transaction, evidently guar- anteeing $150 million in city funds. The mayor established a committee to meet with the financiers, but the financiers re- fused to meet. By this time the Cebu scam saga was receiving daily press coverage in Cebu and Manila. (Box continues on the following page.) 92 Subnational Capital Markets in Developing Countries Box 7.3. (continued) The mayor announced in late November that the financing con- sortium was embarrassed by the episode and would give the city a grant instead of a loan or at least contribute a few million pesos to the project. By the next day, however, the consortium changed its mind about any "grants instead of loans" and it was clear that the deal would collapse. There are two perspectives on the ill-fated deal. One is that sub- national governments lack the skills needed for such major de- cisions, so such decisions should be kept out of their reach. However, blandishments by project proponents who stand to gain (or by con artists who wish to defraud) are facts of life at all levels of government. A more positive perspective is that local party politics and news-hungry journalists are ready to shine bright lights on shady deals. A possible solution? Having local banks and bond dealers compete openly for the deal or help with the due diligence promotes disclosure and limits politically influenced decisionmaking. Source: Petersen and Crihfield 2000. tion and approving its terms. Special voting requirements may be em- ployed. A "supermajority" vote by a subnational governing body has sometimes been used to demonstrate political support (see box 7.2). Such a supermajority vote can be required for certain types of subnational bor- rowing or for debt that is authorized just before an election, to avoid the perception of a politically motivated project. Borrowing approval also can come from the community voting at large through a referendum. A popular vote may encourage citizen participation in decisionmaking and win community backing for the long haul, whatev- er the changes in the elected council. Voter approval for borrowing is not without problems, however. It adds time and expense and can turn finan- cial decisions into political battles that may have little to do with the mer- its of the proposed financing. As with legislative approval, referendums may be limited to certain types of debt or special circumstances. Restrictions on the Issuance and Use of Subsovereign Debt 93 Approval and review by higher government level. National (or state) review and approval of subnational borrowing plans is not uncommon in emerg- ing market economies and may be predicated on specific conditions. These could include the financial capacity to repay debt, as measured by credit analysis or a formula specified by law or regulation. Other relevant consid- erations include consistency of subnational borrowing with national eco- nomic policy (such as the timing of the borrowing) and the purpose and form of the borrowing. Such oversight can be used to prevent irresponsible borrowing at the local level, but it raises a number of issues. Reviews intro- duce delays, require oversight capacity at the national or state level, and provide an entry point for political rather than economic considerations. In general, advocates of market discipline argue that the marketplace, aided by appropriate disclosure rules, borrowing rules, and investor analysis, will do a better job of assessing financial capacity to repay debt. There may be circumstances in which higher level review is appropriate. A state or national authority, while leaving the decisions to incur debt at the subnational level, might certify the procedures used in the borrowing process.2 Certification can help build investor confidence and relieve indi- vidual investors of some of the due diligence that otherwise would be re- quired. However, in most emerging market economies there may be little capacity at the higher levels of government to undertake this task for secu- rities offerings. Imposing a procedural review that cannot be promptly exe- cuted slows the development of the market and opens a forum for political second-guessing and bickering. Even where higher government approval is not routine, it might be de- sirable when a subnational government wishes to exceed its debt limit to have a central government entity authorize such an exception in certain narrow circumstances: · The subnational government has a high degree of creditworthiness. · The projects to be financed will clearly increase subnational revenues, and will be self-financing or will reduce subnational expenditures in future years and be effectively self-financing. · The money is needed to respond to a natural disaster or civil calamity. Additionally, some countries (for example, Ukraine) have tried to pre- vent "pyramid" schemes by prohibiting any refinancing of outstanding debt. Such prohibitions become problematic when a borrower experiences financial difficulties and a legitimate restructuring of debt would benefit all 94 Subnational Capital Markets in Developing Countries parties. Provisions could be made for exceptions in such cases, perhaps with special approval procedures. Equal Treatment of all Forms of Debt The legal framework for subnational debt should not differentiate based on the legal form of the debt. The authorization process, debt limitation, and allowable purposes for issuing debt should be uniform for loans and bonds. The decision of a subnational government about whether to use loans or bonds should be based on market factors rather than legal factors.3 The Ro- manian Ministry of Finance adopted a regulation requiring its approval of subnational government bond issues (even though the Law on Local Public Finance did not require central government approval) but not of bank loans, thereby creating a legal environment favoring loans over bonds.4 Ukraine has substantially different authorization procedures, amount limitations, and allowable purposes for bonds, loans, and guarantees.5 In the Philippines loans may be taken for any purpose, but bonds may be sold only for "rev- enue-producing" facilities (see Philippine case study, chapter 26). Details relating to the terms of subnational debt, such as maturity and in- terest rate limitations, are often not expressly set forth in the legal frame- work and so are open to interpretation. An area under intense scrutiny in many emerging market economies is the currency composition of the debt (see the South Africa case study, chapter 18). Except in unusual circum- stances, subnational governments have limited ability to raise foreign cur- rency funds themselves and are poorly positioned to hedge or speculate against currency fluctuations. Sofia, Bulgaria, used a U.S. dollar-denominat- ed loan to fund the purchase of buses in 1994. During the term of the loan, the exchange rate rose from less than 30 lev to the dollar to more than 3,000 lev to the dollar. Subnational governments are often exposed to such risk through the on-lending programs of multilateral and bilateral lending pro- grams. Subnational governments have sometimes been charged loan premi- ums for currency risk protection and in other cases have simply borne the risk directly, though often the risk has been mitigated by sovereign guaran- tees. Central government approval is often required for subnational govern- ment assumption of currency risk, which seems to be a sound policy.6 Restrictions on Short-Term Debt Short-term financing can be a useful part of a subnational government's regular operations. It can be used to cover operations in anticipation of an- Restrictions on the Issuance and Use of Subsovereign Debt 95 nual tax revenues or of nonrecurring revenue, such as from the sale of as- sets, receipt of a grant, or issuance of long-term debt. Operating expenses can be financed from borrowed funds or from a municipality's working capital, which usually is less expensive and more reliable, but not all subna- tional governments maintain adequate working capital funds. In that case subnational governments generally either match outflows with inflows or attempt to get advances from the national treasury, as in Romania. Events can slow the receipt of revenues or cause unexpected surges in spend- ing that lead to cash shortages. Ideally, governments would carry reserves to smooth fluctuations in working capital flows, but liquid reserves can be a source of political bickering and a temptation to politicians with other priori- ties. In some countries surplus funds are returned to the central government for redistribution or are held in non-interest-bearing national treasury accounts, yielding no benefits to subnational governments from investing surpluses. Borrowing to meet short-term financing needs can provide opportuni- ties for banks and subnational governments to develop working relation- ships and allow bankers to become familiar with the governments' finan- cial affairs. Provided that the financing is repaid within the budget year and that carrying debt beyond the budget year is prohibited, there is no a prioi reason to limit such financing to capital spending. Dangers of Misuse A major concern is that short-term debt will be used to bridge an ever-grow- ing gap between recurring revenues and recurring expenditures, reaching levels that compromise a subnational government's ability to deliver basic services. The "snowballing" of short-term debt as governments run chronic operating deficits has been a leading cause of financial emergencies, causing banks and other investors to lose confidence in a government's ability to run surpluses and repay its short-term debt. Allowed to accumulate too long, short-term debt can reach unsustainable levels, requiring a high proportion of revenues to be devoted to debt service at the expense of public services. Eventually, creditors may deny further credit extensions when they perceive that the floating debt has reached excessive levels (see box 7.4). This hap- pened to New York City in the 1970s and more recently to subnational gov- ernments in Argentina, Brazil, and Mexico, contributing to financial crises. Nature of Restrictions Short-term borrowing should be restricted to financing intra-year cash flow budget deficits. The debt should be repaid within the budget year, with no 96 Subnational Capital Markets in Developing Countries Box 7.4. Johannesburg Comes Up Short South African municipalities may legally borrow to finance both routine and unusual short-term needs, but they are required by the national constitution to settle their short-term debts by the end of the fiscal year. The usual form of borrowing has been a bank overdraft, which creates an unsecured debt. In some cases, the overdrafts were not being settled as required by law at the end of the fiscal year. For Johannesburg, curing the snowballing short-term debt problem led to other problems for the nation's financial sector. In 1997 Johannesburg found itself in a difficult position. It had accumulated a large amount of outstanding short-term debt to finance the start of a capital spending program that was to have been funded by the sale of long-term debt. Domestic markets closed to the city in late 1997. Ultimately, the Development Bank of Southern Africa (DBSA) stepped in and made a loan secured on a specific tax source. However, the rescue had its repercussions. The South African commercial banks, which had refused to roll over the short- term loan, were disturbed that the DBSA had "peeled off" part of the general revenue base and had done the deal on terms that were only marginally better than they were offering. Source: Authors. refinancing beyond the end of the budget year. The volume of short-term borrowing also can be limited, with a ceiling set at some percentage of total budget revenues (see box 7.5 for some common formulas). Lithuania and Romania limit short-term debt to 5 percent of revenues. A further protection against excessive debt accumulation is a require- ment that short-term borrowing be paid off in full at least once a year, with appropriate safeguards against immediate re-borrowing. Because natural disasters or financial emergencies may make this difficult to enforce, some provision for national-level approval of exceptions might be needed. Restrictions on the Issuance and Use of Subsovereign Debt 97 Restrictions on Long-Term Debt Long-term debt allows subnational governments to acquire or build capital improvements more quickly than they could on a pay-as-you-go basis. It al- lows more equitable payment schemes, since users can be made to pay for the capital cost of facilities as they are used over time. However, there are also costs and risks. Long-term debt limits a subnational government's fu- ture budget flexibility. Unwisely used, it can burden citizens with high tax- es or service charges. Many countries permit long-term debt only for capi- tal spending and not for operating deficits (sometimes called the "Golden Rule"). Some countries are even more restrictive, limiting bonds to "self- supporting" revenue-generating activities, as in the Philippines. Underly- ing these regulations is the conviction that governments should only bor- row long term when the proceeds of the debt will contribute to some future capacity to repay. Competing Prescriptions for Long-Term Debt Use Long-term debt is clearly appropriate for capital investment when the term of the borrowing is related to the useful life of the capital asset being built or acquired. Less clear is whether multiyear debt should be allowed for oth- er purposes in many developing countries, such as work-outs as part of a fiscal recovery package or extraordinary expenses related to the transition and restructuring of governments. Typically, regulation of the purpose of long-term debt either allows sub- national governments to borrow for any public purpose authorized by law-- leaving it to the local jurisdiction to decide what is wise and appropriate and to the markets to decide whether the stated purpose is worth financing--or limits borrowing to specific public purposes. These might include: · Building or acquiring a capital asset whose anticipated useful life will equal or exceed the term of the borrowing. · Funding self-supporting revenue-generating projects. · Funding accumulated operating deficits as part of a legal or adminis- trative restructuring. · Funding extraordinary needs, such as recovery from natural or hu- man-caused disasters. Each alternative has its advantages. Two factors favor a more liberal autho- rization policy. First, subnational government finance is an evolving art, and 98 Subnational Capital Markets in Developing Countries there should be room to adopt new forms and techniques of local finance. Second, if national policy favors decentralization (the operating premise of this book), then local managers should have decisionmaking flexibility. How- ever, some restrictions may be appropriate for nascent subnational govern- ment debt markets. Specific limitations can provide clarity about what is per- missible, which may reassure young capital markets, particularly where there is a perception that the public needs protection from politicians or managers who might try to use long-term debt for their own short-term gain. Public Purpose Debt: Distinguishing between Public and Private Benefit A frequent issue is the evolving standard for "public purpose" and the cre- ation of legally defined boundaries distinguishing private and public bene- fit. The limited debt capacity of many subnational governments in emerg- ing markets might best be devoted to projects that clearly serve a direct public purpose. Yet many subnational governments have inherited activi- ties and facilities of a commercial nature, including ownership and opera- tion of entrepreneurial businesses. Moreover, many reformers call for sub- national governments to involve the private sector in the delivery of goods and services and the ownership of facilities. Such public-private engage- ments can rapidly turn to questions of how the subnational government can provide guarantees or even loans to make the facilities more attractive for private ownership or operation. In other words, both customary and new ways of "doing business" compound the difficulties of making bright- line distinctions between public purpose and private benefit. The governing law should distinguish between debt issued for a public purpose and debt issued for a publicly owned, but inherently private, en- trepreneurial activity. In defining "public purpose," debt legislation should prohibit general obligation debt or subnational guarantees for the benefit of such private entrepreneurial activities. Sometimes a public purpose clause in the subnational debt legislation also explicitly prohibits the use of subnational borrowing authority to incur an obligation solely or primarily to benefit a private property owner or a business.7 This can preclude issuing debt for the types of public-private projects being considered in many emerging market economies in which assets are to be transferred to the pri- vate sector or jointly operated by the public and private sectors. Thus the standard for what constitutes a public purpose may be difficult to define in many contexts. Financing certain private entrepreneurial activi- ties can be argued to have an indirect public benefit, such as increased em- Restrictions on the Issuance and Use of Subsovereign Debt 99 ployment, economic activity, or housing. This standard may be an appropri- ate issue for regulation, which should provide the necessary flexibility. The standard could initially be defined very conservatively and later expanded. Additionally, an argument could be made that, although the nature of the project is not directly related to a subnational unit's ability to pay the debt, creditors would rather be associated with a public purpose project that en- joys general political and popular support and enhances a subnational gov- ernment's willingness to pay the debt associated with such a project. Regulations concerning the public purpose need to be carefully phrased with provisions such as the following: · The public purpose is paramount in the expenditure or loan, any pri- vate gain must be incidental to achieving that purpose, and such gain must be of a customary and appropriate degree. · Financing that extends beyond the current budget year may be issued solely for investment or refinancing of debt issued for investment that serves a public purpose authorized in the municipal budget. · The proceeds of a borrowing may be spent only on the investment for which the debt has been authorized, unless both the subnational governing body and the debt holders agree otherwise. Restrictions on Amount of Debt There is little agreement in practice on the amount of debt that a subna- tional government should be allowed to carry. A review of the case studies (chapters 14­31) shows that approaches occupy a spectrum: at one end is Hungary (chapter 29), with a Law on Local Self-Government that gives sub- national governments unlimited borrowing authority; at the other end is the Republic of Korea (chapter 23), with very detailed borrowing criteria in- volving multiple measures of debt service and requiring higher level ap- proval of individual issues. An International Monetary Fund publication (Ter-Minassian 1997) argues for rules-based control of subnational borrow- ing, with limits on the debt of individual jurisdictions that "mimic market discipline."8 Such control could be framed in terms of a ratio between max- imum annual debt service and a conservative projection of the revenues that would be available to pay debt service.9 The argument against such control is that there can be occasions when it is desirable for a local jurisdiction to temporarily exceed a given ratio as it 100 Subnational Capital Markets in Developing Countries invests for the future or spends down accumulated reserves. The argument favoring such control is that it protects the public from reckless borrowing by officials or elected representatives who may not be sensitive to the long- term risks. Distinguishing General Fund and Limited Obligation Debt For general fund debt an allowable ratio of outstanding debt or debt service to available resources for repayment can be set to provide reasonable pro- tection without interfering with sound management. An escape clause could permit the debt limit to be exceeded for exceptional cases under emergency legislation or with special permission from higher level authori- ties (or local referendum, as in the United States). For self-financing or en- terprise projects, where the pledge is clearly limited to project revenues, the debt limit need not apply. However, few emerging market economies make this legal distinction for limited obligation debt. Some Practical Problems in Designing Limits Limitations on subnational debt are widespread. Common debt limits are on the following: · The amount of indebtedness issued, usually expressed as a ratio of ac- tual or potential source of revenues, such as taxable property values. · Annual debt service as a percentage of uncommitted annual rev- enue,most commonly 15 percent of recurring revenue. · Short-term indebtedness, generally to mature within one fiscal period but often violated in practice. · Long-term borrowing is restricted to capital investments and borrow- ing in foreign currency is prohibited. Provisions are frequently open to differing interpretations, and enforce- ment can be uneven and fractious. Poland provides an example. In Poland, Regional Audit Agencies are charged with ensuring that cities comply with borrowing restrictions, but interpretations have been inconsistent. Al- though the borrowing law is silent on the agencies having oversight for project selection, cities and regions often have argued over the desirability of specific projects. The main reason for ambiguity is that legal limitations on debt are not adequately detailed in regulations, leaving several questions about their ap- plication: Restrictions on the Issuance and Use of Subsovereign Debt 101 · Is it the debt service installments (principal and interest) payable in any single year that may not exceed the designated percentage or the total principal amount of the debt at the time of borrowing? While the intention would appear to be to limit debt service install- ments in any single year, the language often expressly refers to the "borrowing." · If the test is based on annual debt service, exactly how is the formula calculated for future years? What assumptions are to be used for debt service when the interest rate is variable? What assumptions are used to predict future revenues?10 · If compliance in subsequent years is not tested at the time of is- suance, what is the effect of violating the debt limitation in a future year? Is there any impact on the validity of the debt? What would prevent debt structures that defer a substantial portion of the princi- pal repayment to later years? If not clarified, these issues can cause substantial confusion, permit po- litical skirmishing, and create barriers to the development of a credit mar- ket for subnational borrowing. The annual debt service limitation should be tested at the time of is- suance; if the debt service is within the limitation, it should not be sub- ject to claims of violations of the limit in subsequent years. Each annual installment can be calculated as a percentage of the total current revenues of the budget in the year in which the debt is issued (or the prior year, if the data are more verifiable), assuming the interest rate at the time of is- suance (providing it is based on an independent index, to prevent use of an artificially low rate to achieve compliance with a debt limitation.) This interpretation would be an incentive for "substantially equal annual debt service." Many transition economies have restricted debt service to a percentage of budgeted revenues (table 7.1). Poland holds annual debt service to no more than 15 percent of budgeted revenues. Debt carried beyond the cur- rent year may not be greater than 60 percent of budgeted revenues.11 The limit on annual debt service falls to 12 percent when total public debt (sov- ereign and subsovereign) hits 55 percent of GDP, and further borrowing is prohibited when the total hits 60 percent of GDP, the EU standard. Roma- nia limits the annual debt service of subnational governments to 20 percent of total current recurring revenues, including the shared wage tax.12 Lithua- nia holds borrowing to 10 percent of current revenues. 102 Subnational Capital Markets in Developing Countries 60 to . 100 loan a than state, by exceed clause a annually the enue. more rev of from individual cannot loans. approved include reported revenues. total for commission. be year of must and total expenses transfers liquidity ceiling must of and ilnius)V budget . and . register audits. general for next year year percent authorization stated. loans to documents debt 5 short-term borrowing to loan a assets, to fiscal fiscal debt percent investment public by lower all independent (30 explicitly for limited apply within within on registration the outstanding performance. in primary are restrictions not only Finance. impose unless debt approved with percent of investment. external with repaid repaid tax. does 20 budget used principal loans be Other be can be for reported have to on be revenues. guarantees. Ministry only be must secured income must can must guarantee, is unpaid guarantee; Finance balance must the be governments restriction limited based of state of must is loans of no credit loans state credit budgeted state cash no no borrowing forints cannot personal of Countries stock are is is Ministry effect. incurred Subnational million Loans shared Debt-service Debt Short-term There The municipalities Long-term commission Short-term There Long-term Carry-forward percent There this Debt Short-term External · · · · · · · · · · · · · · · · European a is Central for and exceed "revenue" (excluding there total percent limit cannot of budget grants); 5 borrowing. Eastern of ratio percent approved Borrowing-to-revenue None. Borrowing 10 in earmarked sub-limit short-term None. None. Selected in fees, limit current environ- revenue under under taxes, Limitations own ratio includes total total includes current of (local commitments. of grants. of commitments). of 2002. liability revenues, excluding liability Debt fines). Dunn service service service percent percent percent percent and 70 revenues interest mental Debt potential guarantee 10 revenue, earmarked 15 (debt potential guarantee 20 revenues. Debt Municipal DeAngelis 7.1. Source: ableT Country Hungary Lithuania Poland Romania Restrictions on the Issuance and Use of Subsovereign Debt 103 Central Government Review and Exceptions Laws governing subnational debt sometimes provide for central govern- ment authority to review and approve requests to exceed the debt limit when the subnational government can demonstrate that its local revenue base would support a greater amount of debt. Exceptions might include: · Additional financing for more creditworthy subnational governments. · Financing of investments that have a positive net impact on cash flow either by generating increased revenues or by reducing operat- ing expenses; examples are utility and energy conservation projects. · Natural or civil calamity. Considerations of Security and Collateral Authorizing the use of various forms of security and collateral to secure sub- national government obligations is an important part of the legal underpin- nings of a subnational credit market. In view of the imperfect security pro- vided by a general obligation pledge, subnational borrowing is often reinforced by additional, specific pledges of revenue, property, or a third-par- ty guarantee. Some projects that have the potential to be "self-supporting" may not require the pledging of general credit. Revenues A subnational unit should be legally authorized to pledge to a creditor specified revenues over which it has spending discretion. The revenues should be identifiable and held apart from other funds. A creditor should have a first-priority secured position to such revenues, a critical element in the structure of a revenue-secured debt. Lack of experience with such pledges and with judicial recognition and enforcement creates uncertainty. It is clearly advantageous to have an express provision in the law for subna- tional governments to secure their loan repayments with identifiable fu- ture revenues and to ensure a creditor that it has a first-priority secured po- sition to such revenues. While a legally protected pledge is important, its absence has not pre- cluded subnational borrowing when alternative arrangements have been available. Banks have loaned to subnational governments that deposit their funds (or some portion of them) with the bank. The possession of these de- posits, backed up by a right of offset, reduces the risk that arises from not having a legally protected pledge.13 104 Subnational Capital Markets in Developing Countries Physical Property The ability to sell property is essential to the ability to pledge or mortgage property to secure a loan. However, the authority for subnational govern- ments to use property as collateral is often clouded in legal uncertainties. Some countries have clear distinctions between subnational property in the "public domain," which is used to carry out mandated government functions, and property considered to be in the "private domain," which is unrelated to such essential government functions. Private domain property may be encumbered or otherwise disposed of, but public domain property is "inalienable." The value of property as collateral also depends on the legal procedures to be followed by a creditor in case of a default. If it takes several years to foreclose on property, its value as collateral is substantially diminished. Of- ten, procedures relating to subnational property have not been established or have not been used enough to create a reasonable expectation based on precedent. Whatever the legal status of these issues, in many countries there ap- pears to be a consensus that a subnational government may sell or other- wise encumber property that is not used in carrying out its mandated ser- vices.14 Such practical understanding of the parameters of this authority seems to be based more on historical practice than on legal provisions. So even if the authority to use physical property as collateral is unclear, banks often lend on the basis of physical property as collateral since banks are fa- miliar with this type of collateral. Additionally, bank regulatory require- ments often establish preferred capital reserve requirements for loans se- cured with physical property. Property in the private domain that is owned by subnational governments may be sold or otherwise encumbered to se- cure debt.15 Intercepts One of the most used and effective forms of security for subnational debt in emerging markets is an intercept provision. Intercepts give a lender a first claim on intergovernmental transfers due to the subnational govern- ment in the event of nonpayment. A number of countries specifically use legislatively authorized intercepts of intergovernmental transfers to en- hance the ability of subnational governments to offer reliable security for their borrowing. Depending on the size and continuity of the transfers, in- tercepts can provide strong encouragement of credit market development without any implied central government guarantee or other cost to the na- Restrictions on the Issuance and Use of Subsovereign Debt 105 tional treasury. Thus intercepts merit particular consideration in the devel- opment of subnational borrowing policy and law. Intergovernmental transfers are of several types. Some give subnational governments a specified portion of national tax revenues. Others are distrib- uted not by formula but through annual appropriations by the national leg- islature or as a percentage of national revenue raised in each region or locali- ty. Still others are provided as subsidies for specific projects (Bahl and Linn 1992). In the initial stages of credit market development, the share of the revenues derived from the central government pursuant to an established and reliable formula is often a preferred source of security for lenders. This form of security has opened the credit market to subnational governments that otherwise would not have access to it and lowered their interest costs. Contractual intercept provisions must be carefully drafted to prevent abuse and overuse. Subnational governments can come to rely on the in- tercept rather than on the discipline of making timely debt payments. If the intercept law is too permissive, an ambitious mayor and council can tie up a disproportionate portion of a subnational government's main revenue sources for years to come, jeopardizing mandated service delivery. In some countries problems arise because the central government and private lenders cooperate too closely in the administration of intercepts. The central government may make automatic payments to the commercial lender from a subnational government's allocation of a shared tax and then transfer only the residual funds to the subnational government, with- out any clear accounting for the intercept. To prevent this, the following assurances are important as part of applicable rules and regulations: · That the subnational government not only enter into the intercept arrangement voluntarily but be in control of negotiating the specific terms and conditions. · That there be clear conditions for when the intercept would be acti- vated; the intercept should operate only in the case of defaults, not as a substitute for regular payments. · That, at a minimum, the transferring government provide a clear ac- counting for any intercept funds diverted to a lender. Alternatively, intercepts can be administered through a special fiduciary arrange- ment established at the local level. To discourage subnational governments from over-reliance on the inter- cept to cover delinquent debt payments, consideration can be given to lim- 106 Subnational Capital Markets in Developing Countries iting or imposing a financial penalty each time the intercept is used.16 Also, to preclude any future question of an implied central government guaran- tee, the legislation might explicitly state that no central government guar- antee is to be inferred for such credit without explicit central government authorization. The law could require that each subnational debt instru- ment contain a statement on its face that there is no express or implied central government guarantee and that the instrument does not represent any obligation of the central government (see boxes 7.5 and 7.6 for sample language). Box 7. 5. Example of Language Denying Central Government Responsibility for Municipal Debt Romania Local Public Finance Law 1998, Article 50 (1) The local public debt does not represent a debt or respon- sibility of the government, and it shall be reimbursed ex- clusively from the revenues though which the respective loan was guaranteed by the authorities of the local public administration. (2) The documents registering the local public debt shall in- clude a clause through which the respective territorial ad- ministrative unit places itself under the obligation to reim- burse the debt, and to pay the interest and the commissions associated with that debt exclusively from the revenues of the respective local public authority; the government has no payment obligation whatsoever, and the credibility or taxation capacity of the government must not be used for guaranteeing the reimbursement of the debt contracted by the territorial administrative unit or of the payment of interest or commissions associated with that debt. (3) The documents registering the local public debt which do not comply with the provisions under paragraph (2) shall not be considered as valid. Source: As quoted in DeAngelis and Dunn 2002. Restrictions on the Issuance and Use of Subsovereign Debt 107 Box 7.6. Example of Language on Securing Debt with Own Revenues Lithuania, Decree on Usage of Bank Credits by Local Authori- ties, 1998, Article 14 · When taking a loan, the municipality must guarantee its re- payment only by the means of the municipality budget, and the municipality enterprise, only by the assets, which could serve as a source to recover the loan. Romania, Local Public Finance Law, 1998, Article 49 (1) The due installments deriving from the contracted loans, the interest and commissions due by territorial adminis- trative units, shall be provided in the local budget. (2) The loans contracted by territorial administrative units can be guaranteed by the local public authority, from any rev- enue source, with the exception provided under article 48, paragraph 3. Any guarantee by revenues is valid and shall apply from the moment the guarantee is offered; the rev- enues representing the guarantee and which are collected by the local budget shall be subject to the respective guarantee agreement, which shall apply with priority against any other request of third parties addressed to the respective local public authority, irrespective of whether these third parties are aware of the guarantee agreement or not. The document through which the agreement of guaranteeing through revenues is concluded must be reg- istered with the city hall or with the respective judet [county] council, and with the debtor. (3) All loan agreements concluded according to the provi- sions of this law shall be considered as fully authorized and shall constitute obligations to be enforced on the re- spective local budgets. Source: As quoted in DeAngelis and Dunn 2002. 108 Subnational Capital Markets in Developing Countries Reserve Funds A reserve fund, segregated from other funds of the subnational government and available only for debt payments should the government run into pay- ment difficulty, enhances debt security. Governments should consider cre- ating such a fund for securing debt. How that fund may be invested and by whom it shall be kept are important considerations. Subnational Government Guarantees In many emerging market economies subnational guarantees of municipal- ly owned utility enterprises are a common financing device. Because these and other contingent obligations can present problems for controlling sub- national debt, the guarantee should not be a mechanism for incurring debt indirectly that could not be incurred directly. Guarantees of third-party debt by subnational governments should be as follows: · Authorized in the same manner as subnational debt. · Restricted to projects in the public interest for which subnational debt could be issued. · Limited to third parties created or controlled by the local govern- ment.17 · Counted toward the debt limitation in the same manner as direct debt or as a percentage of the amount until a payment is made on the guarantee, when the full amount would be allocated. The initial per- centage could be based on some determination of the creditworthi- ness of the guaranteed party, although creating such credit distinc- tions may be too sophisticated a process in a new subnational market. Notes 1. In Poland, for example, decentralization and self-governance are key constitutional principles (Constitution, Article 16.2). The Constitution of South Africa provides that the "executive and legal authority of a munici- pality is vested in its municipal council" (Section 151 [2]). Chapter 9 of the European Charter on Local Autonomy calls for local authorities to have ac- cess to capital markets to borrow the funds needed for capital investment. 2. In the United States, the bond counsel (a law firm specializing in mu- nicipal bond transactions) often drafts the needed resolutions and con- Restrictions on the Issuance and Use of Subsovereign Debt 109 tracts and provides opinions on whether the transaction conforms to ap- plicable laws and regulations. Some states in the United States also have procedural checks. Texas local government issuers must obtain approval on procedures but not on use of proceeds or other substantive matters from the Office of the Attorney General before issuing debt. The office reviews the proposed bond issue's supporting documentation, certifies validity, and issues an opinion. The opinion is needed for the bonds to be legally bind- ing and relieves individual purchasers of the need to inquire into the process by which the bonds were issued. Oregon requires that localities pre- pare bond documents following recommended guidelines. Many states re- quire that prospective sales be reported to a central office and placed on an official calendar. North Carolina requires a filing and approval before sale, largely a procedural matter. 3. Certain forms of debt may nonetheless have additional legal require- ments relating specifically to their form. For example, a publicly offered bond issue should be required to conform to standards of appropriate dis- closure to investors. 4. This requirement has subsequently been repealed by the Ministry of Finance. 5. Law on Securities and the Stock Exchange, Law on Local Self-Govern- ment and the Budget Code. 6. In Romania a Government Debt Commission has been created to "ap- prove" any local government debt issued in a foreign currency (Law on Lo- cal Public Finance, 1998). In the Philippines a Monetary Board "render[s] an opinion of the probable effects of the proposed operation on monetary aggregates, the price level, and the balance of payments" (Central Bank Act, Republic Act No. 7653, Sec. 123). 7. The Romanian Law on Local Public Finance allows only projects in the "public domain" to be financed with debt (Article 48 [1]). The Viet- namese Law on the State Budget authorizes a province or city to finance only infrastructure investments. 8. Teresa Ter-Minassian, editor, Fiscal Federalism in Theory and Practice, Washington DC, International Monetary Fund, 1997, pp. 171­172. 9. Although many government's debt ceilings are expressed in terms of the debt principal outstanding in relationship to either a measure of tax base or revenue flows, the best prescriptions are likely to use annual debt service in relationship to available revenues. Actually crafting such a re- striction calls for considerable care in definitions. See Charles Smith, "Mea- suring and Forecasting Debt Capacity: The State of Oregon Experience," 110 Subnational Capital Markets in Developing Countries Government Finance Review (December 1998), pp. 52­54. As Smith points out, the legal limit is usually much higher than the effective market limit. 10. A debt limitation is most effective when it is an "issuance" test rather than a "continuing compliance" test that may be violated in subse- quent years. Unfortunately, many such debt limitation provisions are writ- ten as requiring compliance in each year. 11. Law on Public Finances 1998, Articles 113 and 114. 12. Law on Local Public Finance 1998, Article 51). 13. In certain transitioning economies the formerly centrally owned banks have retained powers of offset on deposits that put them at the head of the line of creditors. In other countries the ability of the banks to exer- cise offset powers is limited, and depositors may elect to sever their rela- tionships and withdraw funds. Where banks possess considerable powers to enforce security, they may stifle competition from the bond markets. They may assist the development of bond markets by acting as trustees on behalf of bondholders. 14. In Latvia local governments are expressly prohibited from guaran- teeing a loan "with property that is necessary for the performance of gov- ernmental functions." 15. The requirement of a "public purpose" and the limitation on collat- eral to be in the "private domain" may effectively prevent the use of a fi- nanced project as collateral for the debt issued to finance the project. 16. In the Philippines local governments may pledge no more than 20 percent of "internal revenue allocations" (Local Government Code, Sec- tions 287 and 324 [b]). 17. In Latvia, a local government cannot guarantee the debt unless it owns at least 50% of the borrower.(or an association that is at least 65% subnational-unit owned). Regulations of the Cabinet of Ministers on Self- Government Borrowings and Guarantees, 4/2/97. Part III Characteristics of Financial Market Regulation and Disclosure John Petersen Chapter 8 Financial Market Structure, Regulation, and Operations Any examination of options for subsovereign borrowing must consider the supply side of the equation. To what extent does a market for subsovereign obligations exist, and how should would-be borrowers access it? Perhaps more relevant in most emerging market economies is the question of where subsovereign securities fit into an overall strategy to develop domestic finan- cial markets. Promoting private capital markets has been a primary objective of financial market regulators and international donor and lending institu- tions that wish to encourage private ownership and functioning markets. A financial market along the lines outlined in preceding chapters would have some level of effective competition in rates and terms and would in- volve private capital, even though government entities also might supply capital. The financial market would be primarily domestic, with borrowers and lenders (or issuers and investors) subject to domestic rules and dealing in local currency. A key objective of many governments in recent years has been to create a municipal bond market for subnational securities. Most of the liberaliza- tion and subsequent growth of the domestic securities markets has focused on privatization and the desire to promote private sector equity ownership. It is in this setting of recasting the roles of the private and public sectors and capital markets that subnational borrowers must navigate. Financial Market Structure The topic of financial market structure and development far exceeds the scope of this book, but it is vitally important for judging the various link- 113 114 Subnational Capital Markets in Developing Countries ages that subnational governments may forge with the capital markets. Most national debt markets are dominated by banks and by central govern- ment and state-owned enterprise debt. Early securities market growth in emerging and transitioning economies has focused on equity markets, and the few bond markets that exist are dominated by national governments and the commercial banking system, with private capital debt markets coming later and hesitantly. Corporate borrowing has traditionally been through the banking system, and there are few corporate bond issues. Nearly all bank lending to corporations is short term; long-term bank fi- nancing is almost nonexistent. Companies have relied on retained earn- ings or direct foreign investments to meet their long-term financing needs. The ratio of the volume of listed securities of exchanges or transactions on the exchanges to the overall GDP is a rough indicator of the relative role of financial markets in the economy. A more precise measure of credit mar- kets would look at listed securities in the debt market (including any ex- change listings, as well as bonds in the over-the-counter market) in relation to GDP. The relative size of the banking sector can be measured by the three ratios of bank loans and investments to GDP, the size of securities markets to listed securities, and domestically held debt to GDP. Similar measures of other financial institutions and intermediaries provide indices of the development of domestic financial markets.1 Government's Role in Credit Market Development Where subnational credit markets end up on their journey toward more openness and competition depends on policies, luck, and how a variety of competing interests are balanced. One commentator on the development of municipal credit argues that the development of subnational govern- ment borrowing should be tied to the methodical and sequential develop- ment of financial markets as a way of minimizing several risks inherent in the process (Noel 2000). In this view, there is a progression from state-con- trolled monopolies on lending, to oligopolies (often holdovers from the state-run system), to open, competitive markets, based on the following prerequisites: · Reduced moral hazard. · Greater market transparency. · Strong financial market governance. · A level playing field among investor groups. · Subnational government capacity to manage and budget. Financial Market Structure, Regulation, and Operations 115 While basic laws need to be in place, in the end markets are developed by champions and risk-takers. The best lessons on market building are those that are taught by mistakes in an environment of accountability and discipline, where public resources and private fortunes are won or lost. In many developing and transitioning economies, the private sector is a recent arrival on a scene that has been dominated by the state (box 8.1). The extent of central government involvement in the allocation of credit is not always immediately apparent. Government ownership of the banking system and other financial institutions (such as retirement funds and insurance com- panies) can be very influential in deciding which borrowers' needs are served and on what terms (box 8.2). For example, in laying out prudential rules and reserve requirements for financial institutions, governments can mandate or build in large incentives to invest in certain classes of obligations. A common market support approach has been to require that reserves contain government bonds (both sovereign and subsovereign) or to set capital adequacy rules that favor these investments. These market develop- ment measures are often relaxed over time as a domestic market begins to emerge (Noel 2000).2 Monetary policy requires that authorities have lever- age over bank portfolios. In theory, open market operations can be carried on in any security. However, for subnational borrowers there is a continu- ing problem of adverse selection, as securities whose markets are directly manipulated by the monetary authorities are either supported or subverted for reasons unrelated to the subnational issuer. Relationships Affecting Markets The recurring turmoil in world financial markets has focused attention on the relationship between the finance industry, especially banking, and oth- er industries. The extent of interlocking ownership, control of boards, and self-dealing between financial institutions and their nonbank affiliates has been at issue. Although subnational governments have been a relatively minor player in such concerns in Asia, in South America the relationship between municipal and provincial governments and the banking system has come under considerable scrutiny. Large cities and states in that region may own banks that serve as in-house providers of credit. While efforts have been made to privatize the banks or place them on an equal footing with private competitors, they still can come under political pressure to fi- nance their governmental parent units. Financial institutions that may be called "banks" do not necessarily follow prudential practices, just as regula- tors do not necessarily regulate nor are laws enforced. 116 Subnational Capital Markets in Developing Countries Box 8.1. Commercial Banking in Transitioning Economies All banking systems in the transitioning economies of Central and Eastern Europe evolved from a single state-controlled bank that was responsible for both monetary policy and commercial banking. These monobanks routinely extended a high volume of credit to state-owned companies to direct production along the lines determined by central planners. The bank did not screen credit or base funding decisions on creditworthiness since credit allocation was a political decision, nor was loan payment enforced. The goal was to get capital funds out ac- cording to the plan. Planned economies hid inflation and guar- anteed jobs for all, so the standard countercyclical activities of banks were not relevant. Loans to subnational governments were seen as just another production (local services) and em- ployment policy of the state. In the move from central planning to market planning, the pro- tean monobanks were split into commercial bank and central bank activities, with commercial banking often set up along sec- toral lines. New banks were allowed to form, and limited entry of foreign banks was allowed. At the outset, regulation by the central bank was often weak and subservient to political inter- ests. Newly created commercial banks were also weak, with small depositors, unknown portfolios, and flaccid regulation. They also remained under state ownership and susceptible to political influence. Nonperforming loans were simply rolled over, and lax lending policies were used to keep state industries going. Inflationary pressure was created as the central bank printed more and more money, encouraging rapid disintermedi- ation and abandonment of the currency. Many small banks were established, but they were unsuper- vised and often closely tied to new private enterprises. Bank scandals erupted in Albania, Romania, and Russia, including Ponzi schemes that drew in thousands of gullible small deposi- tors. Growth of bank loans has not kept pace with growth in the Financial Market Structure, Regulation, and Operations 117 real sector. Private firms, while borrowing for working capital, rely more on retained earnings and direct foreign investment than in more mature systems. Commercial banks, attracted by the high yields and low risk, have tended to lend to each other and invest heavily in the national government's obligations. Source: Berghof and Bolton 2002. The ratio of the volume of listed securities of exchanges or transactions on the exchanges to the overall GDP is a rough indicator of the relative role of financial markets in the economy. A more precise measure of credit mar- kets would look at listed securities in the debt market (including any ex- change listings, as well as bonds in the over-the-counter market) in relation to GDP. The relative size of the banking sector can be measured by the ratio of bank loans and investments to GDP and the size of securities markets to listed securities and domestically held debt to GDP. Similar measures of other financial institutions and intermediaries provide indices to the devel- opment of the domestic financial markets.2 Box 8.2. The Bank for International Settlements' Reserve Requirements and Capital Rules Prudential regulations can have a major impact on the market for various types of obligations. Many sovereign governments have effectively built in markets for their securities by requiring that financial institutions hold a certain amount of sovereign di- rect or guaranteed obligations as part of their reserves. Banks may be required to put up government debt as collateral if they wish to hold government accounts. For example, prior to the availability of deposit insurance for large denomination (Box continues on the following page.) 118 Subnational Capital Markets in Developing Countries Box 8.2. (continued) accounts, the collateral requirement on public deposits was a powerful incentive for banks to hold U.S. municipal bonds. One source of information on potential demand for subsover- eign obligations (as well as an overall measure of perceived subsovereign risk) is seen in the weights that banks must apply to their assets to calculate their capital adequacy. Although these have varied internationally, they are increasingly coming into conformance with the Bank for International Settlements' (BIS) capital adequacy ratios (ratio of bank capital to performing loans; nonperforming loans carry special provisions). The BIS minimum is currently 8 percent. Virtually all countries have sys- tems that meet or exceed the BIS standards. Under the BIS regime, loans to the sovereign government of the same country as the bank are assigned a 0.0 sectoral risk weight (they are assumed to be domestically risk free) and those of private sector firms are assigned a 1.0. Recognizing that the relationship between the central government and sub- national governments varies from country to country, the BIS al- lows the central bank to assign the appropriate risk weight. Thus the weightings provide the central bank's opinion of the risk of loans to the subnational governmental sector relative to loans to the sovereign and the private sectors. In the United States the BIS credit factors range from 0.1 for general obligations to 1.0 for private activity (corporate) bonds. In foreign countries subnational government obligations with explicit central government guarantees have BIS ratios of 0.0 (which makes them tantamount to direct sovereign obligations), and those without such guarantees have ratios of up to 1.0 or even higher. Ratios can be changed to recognize overall changes in sectoral credit strength. This happened in South Africa, where the ratio was increased from 0.1 to 1.0 for subna- tional government securities when the national government an- nounced that it would no longer guarantee municipal and provincial debt. Financial Market Structure, Regulation, and Operations 119 Prudential rules for other financial institutions such as insurance companies and pension systems have similar impacts on vari- ous types of security. To the degree that subnational securities have been lumped together with sovereign securities, they have often benefited from favorable treatment. However, to the ex- tent they are seen as tantamount to corporate debt and loans, they can be disadvantaged. The capital rules are being revised (Basel II), with an emphasis on the underlying creditworthiness of the obligor as well as on the character of the securities. This development may enhance the role of credit ratings in the determination of capital adequa- cy and give a boost to both information systems and credit analysis in domestic markets. Market Development and Regulation Since the 1980s, there has been a worldwide move to lessen direct regu- lation of financial markets and to open markets to greater domestic and in- ternational competition. This has involved all aspects of financial markets, from privatizing banking systems to creating stock exchanges to support the privatization of formerly state-owned enterprises. One result has been a greater number of domestic firms in the securities business and more openness to foreign firms doing business in domestic markets. The en- trance of foreign firms has been important because they bring not only capital and competition but also experience in financing subsovereign obligations. There also has been a move toward greater self-regulation by industry participants and away from regulation by administrative fiat and direct gov- ernment involvement in investment decisions. Less regulation by ministeri- al fiat and less official involvement in individual transactions have made way for more general rules of fair dealing and capital adequacy and rules of 120 Subnational Capital Markets in Developing Countries the road for functioning markets. Thus the presence of more firms and a greater variety and number of financial instruments in the market means a need for more regulation and more sophisticated regulation. This changed regulatory mode depends on the operation of self-regulatory bodies rather than on central government agencies and is not without costs and risks. Where do subnational governmental borrowers fit into the emerging se- curities market regulatory scheme? Since subnational government securi- ties are still a rarity, the question is just beginning to be asked in most places. Emerging markets have seen a variety of regulatory schemes, includ- ing requirements designed to encourage sound business operations. In Chile publicly offered issues of corporate securities must be rated by a li- censed rating agency. Indonesia's securities regulatory body, BAPEPAM, has similar requirements, which have been instrumental in creating the na- tional rating agency, Perfindo. In Mexico the requirement that states and municipalities be rated by at least two credit rating agencies in order to bor- row in commercial markets has created strong demand for ratings and helped build an active bond market. The fundamental concept of regulation is to define the financial system and its rules of operation. That is easier said than done. Countries have dif- ferent legal traditions that can influence the nature of a market's operation (box 8.3). Countries also have different traditions in regulating the banking system and other financial institutions, with the biggest debate between advocates of the "universal" banking systems and advocates of the separa- tion of the banking system and the securities markets (as in the United States until recently.)3 Most emerging and transitioning economies come out of a bank-oriented financial system, often with government-owned or favored universal banks that have seen virtually every phase of domestic financial commerce as fair game. As financial markets broaden and mature, the regulatory boundaries between financial institutions need to be defined. For example, in addition to prudential regulation of traditional financial institutions such as banks, insurance companies, and pensions, there are new entities to regulate such as mutual funds, clearing and settlement operations, securities depositories, and markets in derivatives and asset-backed securities. Subnational govern- ments that enter these markets are exposed to both the opportunities and risks that attend a dynamic marketplace and the ways it is regulated. Financial market regulation has a variety of roles to play in emerging markets, and the end results may not always be in harmony. Among the competing objectives are the following: Financial Market Structure, Regulation, and Operations 121 Box 8.3. What Is a Security? Defining a security is important from a legal perspective for es- tablishing what an investor can look to in support of the obliga- tion and from a securities regulation perspective for characteriz- ing the nature of the transaction and the instrument involved. Efforts to regulate securities and to harmonize laws across countries have been hampered by different concepts of what constitutes a security. For example, in the Spanish-speaking world and in the civil sys- tems of Eastern and Central Europe the concept of a security has differed from that which evolved under English common law and exists in many English-speaking countries today. A se- curity in Spanish-speaking countries is embodied in the concept of a titulo valor, which encompasses only a limited number of specific physical documents that have the right of ownership embodied in the document. Thus the only evidence of owner- ship for the security investor is the existence and possession of the document itself. The titulo valor instrument is like money, since it can be transferred physically without re-registration or even endorsement and is payable on presentation. The titulo valor proved woefully inadequate as a concept for ev- idencing ownership given the nature of modern transactions. Not only does it pose physical safekeeping and transfer prob- lems, but it does not fit the needs of new financial instrument constructs. New instruments necessarily rely on book entry and dematerialization, such as variable rate securities, derivatives, and investment contracts. New definitions of security now be- ing enacted into law in Latin America rely on the economic ba- sis of what constitutes the security, rather than on the strict defi- nitions of what physical instruments qualify as titulo valor. · Market development: Some regulations are intended to provide incen- tives to market development, especially as part of the effort to priva- tize government-owned institutions. Opening up markets, particular- ly to international capital flows and competition, is not without 122 Subnational Capital Markets in Developing Countries controversy. With proper regulation, however, the objective of en- couraging competition and efficiency appears to be sound even if the means of achieving it are not entirely clear. · Market integrity: Regulators want to foster lively, creative markets yet protect the integrity of the payments system and avoid excessive risk taking. This requires prudential measures to minimize systemic risk and protect the solvency of individual firms. · Fairness: Regulators are keen to prevent fraud and manipulation and to protect investors and prevent monopoly power. Asymmetry of dis- closure information (the issuers control it, the investors need it) can be an invitation to manipulation and fraud. · Efficiency: Markets are allocators of capital resources. Realizing the benefits requires competition among players and fair price discovery mechanisms, but the participants must be limited to those that have adequate capital and experience and meet standards of behavior. In most emerging market economies, subnational governments come to markets as largely untested small borrowers. Where they have adequate rev- enue bases, they can be viewed as potentially strong "credits" notwithstand- ing their small size. Even where banks have dominated direct lending, they benefit from the development of a securities market. The market provides banks with more liquidity as investors, even while promoting more compe- tition among them by providing an alternative source of funds to direct bank loans. Furthermore, a more developed credit system allows banks other ways to earn fee incomes, such as acting as trustee and credit enhancer. The Securities Marketplace Domestic securities markets are in various stages of development, with differ- ent intensities of competition, technological development, and philosophies on regulation. Several measures of development are possible. One is the amount of trading in formal markets (exchanges) versus over-the-counter transactions. Over-the-counter transactions take place in electronic markets of dealer-to-dealer trades in securities that are not listed on the exchange or that can be traded off the exchange as well as on. Typically, registration and listing requirements are softer and less expensive than on formal exchanges. Over-the-counter transactions give rise to several questions. How much off-market trading is reported, and how are such trades cleared, that is, how is the ownership of securities exchanged against payments of cash? How Financial Market Structure, Regulation, and Operations 123 integrated are the markets? Is there a single market or are there segmented markets by types of instruments? The great advantages are ease of access for dealers and low cost for issuers. However, with few investors, over-the- counter markets in developing economies often languish from a lack of volume. Most investors buy to hold, and the liquidity provided by a deep and active market remains a goal rather than a reality. In many transitioning economies, stock markets are a new development, often a by-product of the privatization of formerly state-owned enterprises. Many exchanges are small, with little activity and maybe a short lifespan (World Bank 2002c). To accelerate development of the exchange, regulators in emerging market economies have often required that all securities (equity and private debt) trade on the stock exchange. However, normal exchange listing requirements, typically modeled on those in developed countries, and the related registration fees can be burdensome, especially for new com- panies and small companies. One answer has been to create a separate brack- et for smaller, higher risk companies, as in Japan. Another has been to allow the development of an over-the-counter dealer-to-dealer market or to restrict certain classes of offerings to sophisticated institutions and individuals. This approach provides trading liquidity to otherwise less liquid shares without exposing the general public to undue risk. New credits can be allowed to sea- son before graduating to an exchange listing. However, having a number of separate markets can lead to an undesir- able diffusion of resources. Recently, the move has been toward fewer orga- nized exchanges and screen-based, fully reported trading as opposed to the open-cry, single place market.4 According to proponents of integrated mar- kets, this leads to more self-policing. The more integrated and transparent a market's operation, the better defined are the market's participants and scope and the more likely that market competition, especially foreign com- petition, on the basis of price and quality of service, will discipline behav- ior without direct regulatory involvement. Even advanced markets such as the United States must continue to work to achieve the correct mix of governmental oversight and market freedom that balance the goal of reasonable access to the markets by would-be is- suers of debt with that of protecting the investing public. The boundaries of regulated activity can shift depending on evolving circumstances and events. Achieving an appropriate balance is even more difficult in develop- ing countries, where an infrastructure of experience and legal mechanisms is not yet in place. Efforts to register and list new subnational government bond issues in a national stock exchange are discussed in box 8.4. 124 Subnational Capital Markets in Developing Countries Box 8.4. After 60 Years, Municipal Bonds Return to Romania In November 2001 two small Romanian cities, Predeal and Man- galia, issued municipal bonds, the first subnational government bond issuances in the country since 1941. The local currency- denominated issues were small (5 billion and 10 billion lei, or about $175,000 and $350,000) and short-term (maturities of two years). However, they were viewed by both the communities and the underwriting firms as a first step to opening up a fledg- ling capital market as an alternative to commercial bank lending. The proceeds of the bond issues were used for modest capital improvements, including a new sea wall for Mangalia and site improvements for Predeal's ski slope, the major business in the resort town. Much effort went into designing the transactions and documen- tation, which all parties concerned saw as a pioneering effort. Eager for the exposure, both cities decided to list the stocks on the Bucharest Stock Exchange. This required Security Commis- sion approval, but the commission's registration forms and dis- closure requirements were designed for private companies and ill-suited for the municipalities. The commission staff and the applicants set about devising new standards and prospectus. The stock exchange, which acts as registrar and depository for securities issues, entered into contracts with the cities to re- ceive regular reporting information, also a first. By late November the new bonds were listed for trade, repre- senting only the second and third listing of debt securities by the exchange. (The listing ceremony made the evening televi- sion news.) The national government is planning to list its own small-denomination note offerings on the exchange in hopes of cultivating more individual investments. Meanwhile, the Securi- ties Commission is drafting new regulations to govern future bond municipal issuances. Source: Petersen 2002. Financial Market Structure, Regulation, and Operations 125 Nature of Investors In assessing the market environment, the nature of investors is an impor- tant consideration. Emerging market economies typically have few in- vestors, and their appetite for long-term securities in locally denominated debt is often limited. Both institutional and individual investors look at the tradeoffs between risk and return. The tradeoffs can be very steep in the case of a domestic currency securities market in a developing economy. Many investors are reluctant to make long-term bets in currencies subject to large fluctuations. National government securities, with large demand for funds and offer- ing high returns, often sop up most of the supply of investible funds. Fur- thermore, capital requirements for banks and other financial institutions often reinforce the desirability of holding sovereign securities.5 However, in many cases, these capital requirements, whether through oversight or in- tent, can give preferential treatment to subnational debt, which may be considered "governmental" for purposes of the calculations.6 The supply of long-term investible funds is especially limited. Because of the need to match assets against liabilities, banks are typically a poor source of such funding. When banks lend for even intermediate periods, the structure of the obligation is typically a variable interest rate, and the loan is often callable, should the need arise. Ideally, longer term funds would be forthcoming from institutions with long-term liabilities (pension funds and insurance companies) and individual investors with long-term savings. While several transitioning and developing countries have em- barked on programs to promote these long-term investing institutions in the private sector, progress has been slow and the barriers daunting. There are several reasons. First, any pool of domestic long-term capital is avidly sought by the na- tional government and the banking system. Second, the institutional in- vestors may be circumspect about making long-term investments in the lo- cal currency. Local currencies in small countries can be extremely volatile, with major uncertainties about future value, making them unappealing to investors with other options. Third, raising long-term funds is especially difficult if countries lack considerable liquidity (an active secondary mar- ket), as is often the case in emerging market economies. Despite the difficulty, countries should try to promote longer term sav- ings and to mobilize those savings for infrastructure investments. Banking laws and regulation of institutional investors should not discriminate 126 Subnational Capital Markets in Developing Countries against subnational issuers. In many emerging market economies long- term institutional investors with less need for liquidity, such as private pen- sion systems and insurance companies, find themselves in a position simi- lar to that of banks. The high yields on government securities, reserve requirements, and prudential requirements are impediments to investing in nonsovereign bonds, including those of subnational governments (Rosen 2002). Tax Laws Tax laws have a powerful effect on the development of credit markets and the motivation to participate in such markets. The distorting effects of tax laws on financial markets are well-known. Financial institutions and trans- actions, as highly visible components of the payments system, are relative- ly easy game for oft-frustrated tax collectors. A common measure is a flat rate withholding tax on interest income that is preemptive of any further payment due. Another is a turnover tax on transactions. Yet another mea- sure is to grant tax exemptions on the interest received on long-term bank savings deposits and on foreign currency deposits held domestically. De- pending on design and enforcement, all these tax strategies can stifle the development of bond markets. Tax laws can favor the securities of subnational governments over other financial instruments. The exemption from federal income taxes is the ma- jor reason for the low interest rates on municipal bonds in the United States. Both Poland and the Philippines have extended limited tax exemp- tion to subnational government bonds (see case studies, chapters 30 and 26). Most economists argue against tax exemptions for subnational debt on the grounds that exemptions distort the allocation of capital between the public and private sectors (Leigland 1998). Notwithstanding the shortcom- ings, such tax exemptions are common and can kick start market develop- ment. They have the advantage of being implemented through market ac- tivities--to enjoy the benefit a government has to initiate the borrowing and be prepared to repay the debt--and, if kept relatively simple, adminis- trative costs are low. International Markets Large subnational borrowers and intermediaries that cater to them may have the option of borrowing in the "emerging-markets" tier of interna- Financial Market Structure, Regulation, and Operations 127 tional financial markets. This source of funds had been growing rapidly un- til the series of financial crises and setbacks in the late 1990s. Since then in- ternational borrowing activity has receded rapidly. The hope remains, how- ever, that when the broader international market recovers, subnational borrowers will again find it an attractive source of funds. International financial markets are segmented between prime quality sovereign borrowers (public and private) that are able to borrow at the low- est rates and higher risk borrowers that have to pay higher rates of interest. The emerging market sector is part of this high-yield segment of the market (Rosen 2002).7 Although the high-yield segment performed very well in the mid-1990s, it subsequently fell on hard times, making it difficult for emerg- ing market borrowers of less than prime quality to sell bonds in the inter- national markets. As a consequence subsovereign borrowers have been ex- cluded from international debt transactions. Domestic bond markets in emerging market economies appear to be ex- panding at least in part because international markets have become too dif- ficult to access. However fleeting, the exposure that some subnational bor- rowers had in the international market provided useful lessons for both domestic and international markets. Chief among them were those related to market expectations for disclosure documentation and the importance of internationally accepted credit ratings. These two subjects are dealt with in chapter 9. Notes 1. None of these measures is flawless. For example, there may be a large number of securities listings, but turnover may be low. Bank loans may be highly concentrated, with a high rate of nonperforming loans. Nonethe- less, analysis indicates a positive correlation between the growth in finan- cial markets and the pace of economic development (World Bank 2002c, chapter 5). 2. Noel (2000) sees the preferential treatment of government securities as leading to moral hazard in the financial system. 3. Until 1999, commercial banks in the United States were prohibited under the Glass Steagal Act from underwriting or dealing in corporate secu- rities for resale. This prohibition was a product of reforms enacted in the 1930s after notorious abuses in the stock market. While it is too early to speculate, similar abuses in the late 1990s and early 2000s may lead to a re- visiting of the recent reforms. 128 Subnational Capital Markets in Developing Countries 4. See, for example, "Survey: Financial Centers," The Economist, 9­15 May, 1998. The number of exchanges may be fewer but their physical loca- tion will be less important as trading occurs wherever a computer can be plugged in. 5. As noted, the capital adequacy requirements discussed above that are used internationally favor the investment of reserves in sovereign securi- ties. Countries are not above using this, as well as domestic laws, to build in a captive market for their own securities. These often may be sold at be- low-market interest rates so that the market value is much less than the par value. However, the par value is what counts in meeting the legal require- ments. 6. In the Philippines, the central bank reduced the risk weight assigned to local government obligations that are backed by an intercept of their in- ternal revenue allotment and guaranteed by the Local Government Unit Guarantee Corporation from 100 percent to 50 percent, in effect making them more attractive investments for banks (see Tirona 2003 and the Philippine case study (chapter 26). 7. Interest rates are benchmarked to U.S. Treasury bond yields (usually the 10-year or 20-year maturity). Thus a prime borrower will enjoy a small "spread," that is, will trade at a hundred basis points above the U.S. Trea- sury security (seen as the safest and most liquid security). High-yield bor- rowers will trade at large spreads, which may amount to several hundred basis points above the Treasury bill rate. Chapter 9 Disclosure and Financial Reporting Information disclosure about issuers is a necessary condition for the effec- tive operation of a securities market. Information--consistent, complete, timely, and comparable--is essential for judging the risks and rewards of investments. While information does not always answer all the questions (and bad information can give the wrong answers), an absence of informa- tion makes it difficult even to know what questions to ask. Emerging and transitioning economies face particular difficulty with disclosure. Many countries are undergoing dramatic changes in their fiscal structure just as the structure and regulation of financial markets are changing as well. Direct guarantees by the sovereign are being replaced by newly minted local own-source revenue and transfer systems, as well as more specific pledges of assets and revenues. Some countries, such as South Africa, rely heavily on revenues pledged on commercial public utility oper- ations. Other countries, for a variety of reasons, may choose to restrict long-term debt to self-supporting commercial operations.1 The ability of subnational governments to generate resources to support themselves or to generate surpluses for general revenue purposes depends on efficient technical and managerial operations. Even where governments rely primarily on transfer payments, information on trends in transfer pay- ments compared with local expenses becomes vital to determining relative credit quality. Without uniform, regular, and reliable reporting, comparing and tracking the performance of subnational governments become impos- sible tasks, and market decisions are based more on faith than fact. Regulating Disclosure Disclosure can be required by the central government, by securities market regulation, or as a byproduct of market operations, through contracts and 129 130 Subnational Capital Markets in Developing Countries market practice and convention. Disclosures to securities markets originate with the borrowers themselves, the subnational governments. Borrowers may be assisted by the central or provincial government authorities in ac- cumulating information, but the borrowing government is responsible for disclosures as the party financially responsible for timely and full payment of debt service. A closely related concept is that the party that controls de- cisions to honor obligations and thus has the relevant information is the one responsible for providing the information.2 In securities markets, disclosure is aimed at helping investors make in- formed investment decisions. An often overlooked but practical by-product of securities disclosure is that the performance, condition, and prospects of borrowers become publicly available information. These economic and fi- nancial factors are of material interest to many others in the market besides investors. Also, the concept of disclosure reaches beyond investor "protec- tion" (that is, avoidance of fraudulent behavior) to encompass support for the rational allocation of resources on the ability to evaluate rewards versus risks, whatever their levels.3 Generally, formal disclosure requirements are met when the issuer sends published reports to the marketplace. In the bond markets, there are usual- ly two phases in the process. First, the would-be borrower issues a docu- ment in conjunction with the initial sale that describes the transaction and provides pertinent information about itself, the security pledged, and the use of the bond proceeds, which is variously called an official statement or a prospectus. Second, after the sale, the borrower provides a stream of contin- uing information with respect to itself and the obligation, a process called continuing disclosure. The timing and scope of reporting information are im- portant, and technology is changing the reporting process (see box 9.1). Another class of recipients of the information analyze it and convey their opinion to investors. The most important of these are the rating agencies, which, as is discussed in the next chapter, often act as a surrogate for dis- closure to individual investors. Disclosure documents can be available from a central depository, using information received on a recurring or event- driven basis.4 The broad policy objective of developing a thriving securities market ar- gues for balancing the need to protect investors with the need to ease ac- cess for certain classes of borrowers. Often, standards are lower for smaller issuers or for lower risk securities.5 The content of disclosure statements can be dictated by the regulator's detailed list of required documents and schedules or by a flexible standard that relies on the issuer and its agents to Disclosure and Financial Reporting 131 provide information that investors need in reaching an investment deci- sion. In practice, the two approaches are usually combined. Regulators pro- vide a list of generic types of required information, leaving the particulars to the issuers. Since the scope and detail of meaningful disclosure can vary markedly, the trend has been to rely on market forces and self-regulatory bodies to specify the details of disclosure. Box 9.1. Disclosure over the Internet Electronic transmission of information over the Internet is changing the processes of bond sales and information disclo- sure in the securities markets. Although electronic transmission of data has been possible for many years, it was not until the use of the Internet became widespread that issuers were willing to move their bond sales to the Internet, taking bids in real time. An early experimenter was the city of Pittsburgh, which held its first Internet competitive bond auction in early 1997. That year, several large municipal bond issuers permitted bidders to file bids conventionally, in sealed envelopes, or over the Internet just prior to the close of auction. In 1998 municipal issuers began to publish preliminary official statements over the Internet. Again Pittsburgh led the way. In- vestors could contact the city for a printed copy if they wished. While the city had previously printed 750 copies of the official statement at a cost of $15,000, once it began posting the state- ments on the Internet, it received only four requests for hard copies. Many issuers have started to post their budgets and fi- nancial statements on the Web. The economies of posting bond disclosure over the Web are considerable for both bond sales and information disclosure. The access to a large number of investors and underwriters at low cost promotes improved disclosure. Just as exchanges in many emerging market economies are leap-frogging the stages of securities market development in many developed (Box continues on the following page.) 132 Subnational Capital Markets in Developing Countries Box 9.1. (continued) countries, so too are new information technologies swiftly changing the flow of information in the markets. In a related use of the Internet, subnational governments in Ro- mania can access a Web site that provides a self-diagnostic pro- gram that allows them to compare their financial ratios with those of other governments. Administered by the Romanian Bankers Institute and funded by the U.S. Agency for Internation- al Development, the Web site also contains model loan and bond documents, a collection of state laws relating to local bor- rowing, and a listing of consultants and financial institutions in- terested in municipal finance. Source: Authors. Accounting Standards and Financial Disclosure Uniform accounting standards for subnational government financial state- ments are critical to disclosure. In many countries accounting systems are under review with an eye toward improving their timeliness, transparency, and conceptual consistency.6 International bodies are also working toward cross-country comparability. Strong accounting practices are central to im- proved financial management. The adoption of accounting standards has been expedited where the standards have been required for borrowers wishing to sell bonds or take out loans.7 Accounting standards vary greatly among countries and between the pri- vate and public sectors (see box 9.2). Most governments come from an orien- tation of controlling expenditures and revenues, stressing the legality of their actions and reporting on their conformance with legislation. This has led to the use of cash accounting techniques and has obscured the economic pur- pose or life of the expenditures. The biggest concerns with cash accounting techniques are their focus on short-term financial assets and liabilities and the ability to alter the results by accelerating receipts or delaying payments. It Disclosure and Financial Reporting 133 is not unheard of for governments to simply put the "bill in the drawer" or to delay making a payroll until the next fiscal year.8 It is the case, however, that much credit analysis focuses on cash flows, particularly those flows related to the availability of cash to pay debt service in full and on time.9 When the government borrower is involved in an enterprise activity, it often uses accrual accounting techniques that conform to those used in the private sector. This has a sound economic rationale for determining the worth and period income performance of an activity. However, credit analy- sis typically requires conversion to a cash basis to ensure that adequate cash will be available when needed to meet debt service requirements.10 No accounting system is foolproof, and all are susceptible to misunder- standing and manipulation.11 What matters most is whether the principles are being observed (that is, the accounts correctly kept in accordance with the chart of accounts and their definitions) and whether someone is sys- tematically checking the books. An example of the importance of the con- sistent application of accounting principles in understanding what is going on is shown in box 9.3. Another important issue is the frequency and independence of audits. Most subnational governments rely on audits performed by auditors from higher levels of government. The auditors typically check for compliance with program requirements rather than assess financial condition or assign costs to activities. Independent audits, which are sometimes required, may be difficult to implement because of a dearth of audit skills in the private sector or prohibitive costs for small borrowers (see Hungary case study, chapter 29). In some countries government financial records are not pub- licly available, and bank secrecy laws impede public disclosure of some por- tions of the financial statement. The very unavailability of such financial data is a warning flag that financial risk cannot be assessed and that politi- cal and legal risks are particularly important. In addition to financial statements, appropriate disclosure may require information about the operations and characteristics of the service provid- ed and the market served.12 For example, investors in an enterprise-based security that looks to cash available after operating expenditures to repay debt want to know about the operating characteristics of the enterprise and the market it serves in order to judge how efficiently it is being operated and whether there are any concerns about such issues as the strength of de- mand, supplies, labor relations, environmental matters, and lawsuits. The list of items worthy of disclosure can be long, and the particulars will be dictated by the nature of the operation and the security pledged.13 Thus, an 134 Subnational Capital Markets in Developing Countries Box 9.2. Accounting for Accounting Differences Differences in accounting and financial recordkeeping can make it hard to analyze the performance of governments and their en- terprises. In some countries uniformity in these practices in the private sector arose from the tax systems and securities laws requirements. Because most subnational governments do not pay taxes on their activities and do not list their securities on stock exchanges, the pressure for prompt reporting and uni- form accounting has been lacking. Disclosure of information is meant to support analysis of the risk and reward relationship. Appraising "economic" risk--the risk that the borrower will be able to pay interest and principal as promised--depends on knowledge of its financial perfor- mance (operating statement) and condition (balance sheet). Since most problems involving "willingness to pay" are pro- voked by fiscal stress, strong financial reporting practices sup- port assessment of this risk as well. From a disclosure standpoint, the immediate objectives are get- ting financial data on a comparable basis; measuring the avail- ability of dependable, recurring revenue streams to make debt service payments; and measuring liquid reserves available to continue meeting debt service requirements should the recur- ring revenues be interrupted. With proper reporting, other items, such as the strength and stability of the underlying econ- omy, other indebtedness, and the mix and costs of inputs used by the borrower, also are disclosed or can be calculated from the financial statements and their footnotes. Source: Authors. important initial disclosure will be the intention or contractual commit- ment of the issuer to provide information on a recurring basis in the future. Disclosure requirements do not mandate that every investor be able to read every document and understand every nuance of every deal. When disclosure requirements are particularly stringent, securities regulators may Disclosure and Financial Reporting 135 Box 9.3 Why Did Czech Municipal Debt Grow So Fast? It is not clear why the outstanding debt of municipalities in the Czech Republic grew so rapidly during the 1990s. There was no evident correspondence between the reported accounting flows of the revenues and expenditures of municipalities and their accumulation of debt. While the accounting reports indi- cated that the municipalities' fiscal balances were reasonably stable during the 1990s (that is, with rather small deficits alter- nating with small surpluses), the aggregate amount of munici- pal outstanding debt continued to rise rapidly. This apparent discrepancy is thought to reflect a lack of unifor- mity in accounting practices that led to an inability to know what was actually going on. The evident inconsistency could have resulted from the following causes: differing interpreta- tions by municipalities of accounting procedures and terminolo- gies, including treating loan receipts as revenues; off-budget fi- nancial operations, including the treatment of grants from the state budget; and extrabudgetary funds that were inappropri- ately recorded as revenue by some municipalities. The Czech problem was not unique, as the accounting and financial report- ing systems used by subnational governments during the tran- sition often have been artifacts of the old unitary state system, which were not designed to measure their fiscal performance or condition. Source: Czech Republic case study, chapter 28. decide to promote reliance on private advisory and information services to examine disclosures and make informed judgments for which they are paid by investors. These opinions are published and become a "baseline" of the assessment process. An example of such services is provided by the credit rating agencies, which post ratings on issuers and issues and keep them un- der surveillance while the debt is outstanding. However, even if regulators are not doing the substantive reviews and forming opinions about the ade- 136 Subnational Capital Markets in Developing Countries quacy of disclosure, they need to institute meaningful safeguards to ensure that those who do (such as financial advisers, rating agencies, and other in- formation providers) are professionally qualified, behave ethically, are not manipulating the market, and are free of conflicts of interest. Chapter 10 turns to the subject of credit analysis and credit ratings. Notes 1. In the Philippines, bond issues by local governments are restricted to self-supporting projects. However, absent any definition of the term self- supporting, the restriction is not very effective. 2. A guarantee by a third party (such as the national government) has sometimes been seen as a reason to require less disclosure on the part of the actual borrower. That concept has been rejected in U.S. practice, where a guarantee (or insurance) does not obviate the need for full disclosure by the borrower. In South Africa and elsewhere the custom has been to relax re- quirements when the national government is the guarantor. 3. This is not just an academic distinction but goes to the heart of mar- ket regulation. If the primary purpose is to avoid fraud and investor loss, the emphasis should be on screening out high-risk securities that regulators feel might cause loss to the investors. This substitutes a bureaucratic deci- sion for that of the marketplace. The other approach, and the one stressed in the U.S. philosophy, is to require full disclosure, and then to let the mar- ket decide on the appropriate rate of return to offset the level of risk, no matter what its magnitude. 4. In the United States this role is played in the municipal market by a limited number of officially sanctioned (but privately owned) repositories as well as a central repository operated by the Municipal Securities Rule- making Board. 5. Traditionally, government securities have belonged to this lower risk disclosure class, although that tradition has been eroded in the United States and elsewhere and the exceptions are less likely. 6. One team of investigators reviewing the Latin American markets stress the problem of financial information: "The first problem is the quali- ty of municipal or subnational management and accounting, which is of- ten poor and incomplete " (Freire, Huertas, and Darche 1998). 7. International Federation of Accounting (IFAC) Guideline for Govern- mental Financial Reporting. The IFAC is attempting to develop widespread adoption of generally accepted accounting standards. Disclosure and Financial Reporting 137 8. Credit analysts are concerned about cash flows, and not all cash ac- counting is considered bad. For example, state and local governments in Mexico are on a conservative system that accrues expenditures but treats revenues on a cash basis. This treatment is viewed favorably by rating agen- cies since it understates revenues while fully accounting for costs as they are incurred (interview with Jane Eddy, Standard & Poor's, March 25, 2002). 9. It is customary for credit analysts to restate accounting reports on a cash basis to assess the availability of cash to meet debt service payments. Revenue bond contract indentures are expressed in terms of minimums of available current revenues after meeting expenses (cash outlays) in rela- tionship to debt service needs. 10. Asset valuation techniques differ among countries. Those that use a historical basis can greatly understate the replacement value of plant and equipment in periods of high inflation. For example, water utilities with much of their investment in underground piping and reservoirs may have major assets that have expected useful lives of 40 to 100 years. Utilities that use current market values for assets will appear to be much less leveraged (ratio of debt to total assets) than those that do not. However, their current depreciation charges are likely to be higher, which makes them appear less profitable. 11. A recent study of earnings management by local governments in Sweden and elsewhere found that use of the accrual system let govern- ments manipulate reported earnings (deficits and surpluses) by altering de- preciation rates, asset write-downs, and pension costs. The statistical analy- sis found that governments that were exposed to high levels of scrutiny by public groups and capital markets were less inclined to manage earnings (Stalebrink 2002). 12. The word appropriate is used because once beyond a simple general government balance sheet pledge (and likely even in that case), the infor- mation needed to assess risk will be specific to the local government. For example, a government that relies heavily on utility revenues will find its ability to pay debt heavily influenced by the operations of those utilities. If the raw material or labor costs are rising rapidly or users are not paying their bills, timely debt service payments may be endangered. 13. The list of items to consider can be found in various trade and pro- fessional publications. A good starting point for generic items is the Gov- ernment Finance Officers Association's Disclosure Guidelines for State and Lo- cal Government Securities (GFOA 1991). Part IV Evaluating, Monitoring, and Assisting Subnational Governments John Petersen and Marcela Huertas Chapter 10 Credit Analysis and Credit Ratings Credit analysis is a demand-side activity. Investors and their advisers exam- ine information on issuers and their obligations and make judgments on the rewards and risks of investments. Credit risk, typically taken to mean the economic, legal, and political risk inherent in a particular obligation, ultimately boils down to default risk.1 Information used in credit analysis can be garnered from a variety of sources, such as government statistical data or the local newspapers, as well as issuers and borrowers. Credit analysis demands resources and analytical skills that many in- vestors, especially individuals and smaller institutions, lack. Thus most in- vestors rely on the opinions of experts (box 10.1). An independent, objec- tive system of credit ratings of high quality is an essential component of the development of a vibrant capital market. It is especially important for security markets, with numerous investors that must rely on information provided by issuers and others. If the ratings are respected and used, the rating companies have the clout to demand full disclosure by issuers. To the degree that these companies are successful in obtaining data and that their ratings reflect legitimate risk indices, the entire market is aided by the categorization of debt and the monitoring of performance. The role of credit ratings is not without controversy. For emerging mar- ket economies, with their chronic shortage of trained analytical staff, rat- ing agencies offer a pool of skilled analysts who can assess credit quality on behalf of all investors, using a standard methodology (at least standard to each agency). On the negative side this concentration of opinion, using methods that are proprietary and not fully disclosed, can lead to a danger- ous dependence on a handful of experts who can influence the market without an effective check.2 141 142 Subnational Capital Markets in Developing Countries International rating agencies are sensitive about their impact on the markets. They have had considerable difficulty with "regulatory rating" (a requirement that bonds be rated before they can be listed on the exchanges or sold to the investing public), which can lead to "shopping" for the high- est rating or an acceptable rating at the lowest cost. Requirements for mandatory ratings can lead to the creation of national agencies that are not technically competent and can be politically influenced. The major agen- cies prefer a free market for their opinions, with investors deciding which agencies' opinions are worthwhile;the agencies themselves are leery of be- ing regulated by anyone other than the market. The development of credit ratings in emerging markets has followed two often overlapping tracks. Along one track are various market partici- pants who create a domestic rating agency, sometimes in alliance with an established international rating agency. The focus of these homegrown agencies has been on meeting domestic regulatory requirements. General- ly, the opinions of these domestic agencies have carried little weight inter- nationally. Along the second, more common, track are the major interna- tional rating agencies that have opened national offices or acquired local rating firms. Subsovereign Ratings The appeal of credit ratings is clear: they provide a third-party opinion by experts that informs investors without the skills or resources to carry out their own investigations of the relative creditworthiness of competing in- vestment opportunities. Their appeal is especially strong to investors that have a diverse portfolio of securities, where each represents only a small part of the total holdings. Furthermore, credit ratings have positive effects on the working of subnational governments. Preparing the data for ratings and undergoing review help instill discipline in subnational government officials and staff. The rating agencies' demands for continual updating (with the threat of a down-grading if a government's performance is subpar or the required information is not provided) can strongly encourage good behavior. The rating agencies, for good reason, place considerable emphasis on governments keeping them well informed as a measure of good finan- cial management. The concept of creditworthiness is important. It measures the compara- tive risk of "payments difficulties." Rating agencies do not rate the compar- ative market values of securities or general market risks per se. Each agency Credit Analysis and Credit Ratings 143 Box 10.1. Emerging Market Ratings and Bond Insurance International credit ratings began in the 1980s with Western Eu- ropean countries and corporations that were active in the Euro- market. There were very few subsovereign credits to rate, since most subnational governments relied exclusively on bank lend- ing and sovereign guarantees. The rating agencies later entered the emerging markets by first rating sovereign borrowing in hard currencies and then the pub- lic or private corporations that seemed likely to generate hard currency to pay back international bondholders. The next instru- ment to be developed was the asset-backed security (ABS), which is secured by pools of underlying loans aggregated by the issuer. The ABSs started off with car loans, credit card accounts, and mortgages. These markets soon were flooded at the higher end, mainly by U.S. and Western European issuers, and margins were very thin. Attention again turned to emerging markets. The assets that back ABSs are typically dollar-denominated se- curities consisting of export receivables, credit cards, and tele- phone receivables. The ABS approach allowed issuers to bor- row at much lower rates than in the domestic markets. However, access to these markets requires having a credit rat- ing, and getting a rating has usually required obtaining credit enhancements from third parties. The need for enhancements in turn stimulated the growth of bond insurance. This was accomplished by structuring the debt through an off- shore origination and securing the debt by receivables gathered through a trust. The future receivables are held by an offshore trust, and obligors are required to make payments to the trust. Payments never enter the country of the issuer, thereby avoiding problems of convertibility and mitigating sovereign risk. These obligations thus are not constrained by the sovereign rating of the borrower's country. The device has been used successfully by Argentine provinces that were able to pledge offshore oil rev- enues to repay bonds sold internationally. The funds were re- ceived offshore and so escaped the convertibility restrictions im- posed by the Argentine national government in late 2001. Source: Authors. 144 Subnational Capital Markets in Developing Countries has its own formula for weighing various factors, but the agencies typically look at the same factors in rating subsovereign credit risk. Except in the United States and a few other developed countries, the rat- ing of subsovereign government risk is very much in its infancy. For emerg- ing and transitioning countries, the number of subnational bond ratings by recognized international rating agencies, while growing, is still low. Nonetheless, the rating agencies have been staking out the subsovereign government area, and many observers believe that progress in the develop- ment of subsovereign securities markets will depend on establishing a cul- ture of ratings to guide the market. According to the rating agencies, quite a few subnational governments also are seeking ratings to bolster their overall visibility and credibility.3 Each rating agency has its own rating formula. Reflecting prospects for ultimate or partial repayment, ratings range from AAA for the highest cate- gory, which is usually conferred only on sovereign credits, down to C or D categories, which are assigned to bonds that are in default. While the major agencies have different ways of weighting each factor, they agree on the major analytical underpinnings for judging the creditworthiness of subsov- ereign credits:4 · Sovereign rating ceiling: The rating of the national government usually sets the top limit on the rating that a subsovereign government can enjoy. National governments set monetary and fiscal policy and usual- ly have first claim on foreign exchange. They also can change the rules of the game for subnational governments. Exceptions can arise if the debt is secured by offshore assets or hard currency revenue streams. · Economy: Fiscal health is usually closely linked to the health of the subnational economy. Diversification in activity, which often comes with size, helps balance the economy's performance. Demographics are important. A high dependency population (the very young and very old) and a population growing too rapidly for a country's capac- ity are both negatives. Higher income and more educated popula- tions are a plus, as are an acceptable distribution and rate of growth in income. · Structure and management: An assignment of functional spending re- sponsibilities consistent with revenue resources is a positive. Inter- governmental transfers are examined for their size and predictability. The willingness and ability of the national government to detect and stem financial emergencies is a positive. The rigor and timeliness of Credit Analysis and Credit Ratings 145 budgetary and financial laws are examined and can be either a posi- tive or negative, depending on the flexibility they provide to locali- ties. Past performance in achieving budgetary balance is important. The timeliness and comprehensiveness of financial reporting and the application of consistent standards are all positives. · Fiscal performance: Revenue composition and trends are considered. The ability to set rates at the local level is a positive. Tax burdens should be in balance with those in neighboring regions. Effective use of charges and fees is viewed favorably, but large transfers of general funds to local enterprises are not. Composition and trends in expen- diture are reviewed for consistency and pace: high and rising pro- gram costs are worrisome; steady shares among programs and slow growth are reassuring. Capital spending and maintenance spending are positives; a large wage bill is a negative. The ability to budget and to meet budgets is a positive. Surpluses in current operating budgets are a strong positive, as are capital budget planning and making many expenditures from current revenues. · Financial position: Liquid assets and marketable real assets are favor- able factors, as are healthy reserves in relation to annual expendi- tures. Outstanding debt is considered. Short-term debt is a concern if not periodically retired. Long-term debt and contingent debt (guar- antees) is generally a negative unless used in support of productive (self-supporting) activities. Short maturity debt with principal due at term, called bullet maturity, is a negative because of continuing pres- sure to refinance and the potential burden on current revenues. Over- lapping debt of other governments that relies on the same economic base is considered. · Legal framework: The lack of clear laws, legal precedent, or an effective judicial system is a major impediment, especially where there are re- stricted revenue or enterprise-based pledges. A history of repudiations or insolvencies is a large negative. Approval of borrowings by higher level governments and other restrictions on local borrowing may be positive factors if carried out in an efficient and nonpolitical fashion, but these can be negatives if the process is complex and political. · Accounting and financial reporting: The basis and quality of financial records are examined, and prompt, consistent reports are a positive. So are timely and independent audits. Cash flow information or cash basis accounting that provides reliable information on cash available to pay debt service is a positive. Evaluation of liquid assets and ac- 146 Subnational Capital Markets in Developing Countries counts receivable can influence credit assessments because required investments in government bonds can be risky and accounts may be in arrears. Opinions on credit quality are not static, and the relative importance of factors can change over time. A range of national policies not directly relat- ed to local debt can alter the mix and weighting of credit factors. For exam- ple, laws governing purchasing policies, public employee retirement bene- fits or wages, or the reassignment of functions and revenue sources all can shift the focus of analysts. Credit rating analysts are especially sensitive to the changing missions and roles of subnational governments, especially as part of fiscal adjust- ment. For example, the responsibilities of subnational governments for in- frastructure provision have increased greatly in many transitioning coun- tries. Meeting these needs has led to changing balance sheets and operating statements, as subnational governments assume more debt to meet capital spending requirements. The increasing levels of indebtedness and debt ser- vice at the subnational government level are seen as a natural development and not necessarily as indicators of deteriorating credit quality. The impor- tant issues are the purposes for which the debt is used and how surely and quickly the revenues to pay debt service are growing. Expanding the Market for Ratings International rating agencies have been establishing beachheads in subna- tional markets, both to cover the changing circumstances of subnational borrowers and in anticipation of new markets. This process is illustrated in South Africa. CA Ratings (now affiliated with Standard and Poor's), Fitch Rat- ings, and Duff & Phelps (absorbed by Fitch Ratings) actively promoted their products, even though the South African municipal bond market was mori- bund. Despite the market's small size and cloudy prospects, the agencies continued to show substantial commitment to following municipal debt. One role for the rating agencies in South Africa was to monitor out- standing debt for banks, insurance companies, and other institutional in- vestors that had neither the analytical capacity nor the desire to invest in any. Before 1994 South African municipal bonds carried an implicit sover- eign guarantee. When that was revoked, investors suddenly had to distin- guish among municipal credits that, for all intents and purposes, had been homogeneous in the presumption of carrying no default risk. The transfor- Credit Analysis and Credit Ratings 147 mation into a new government structure presented new elements of risk. The rating agencies pooled the credit research for their subscribers, who had little interest in following individual credits on their own. During the late 1990s more insurance companies were formed to handle nontraditional business, including emerging markets. These insurers han- dle non-investment-grade paper (rated in the fourth tier of ratings, BBB or Baa, or higher), and no longer price under the assumption of zero loss.5 Non-investment-grade paper requires higher reserves and may have less than the highest bond rating. Insurers make money where the perception of risk exceeds the actual risk and can alter the actual risk through close monitoring and direct involvement. Risk perceptions may be institutional- ized in various prudential restrictions placed on lending institutions and investors. These perceptions and restrictions cause credit spreads, that is, the differentials in interest rates, among classes of debt. The insurer, by su- perior access to information, deeper analysis, and ability to diversify risk, can effectively narrow these spreads by "renting out" the use of its credit rating. It charges premiums for this service, thereby enjoying a return on the capital it commits (in addition to its interest earnings). The difficulties of the Asian financial crisis in 1997 and the Russian de- fault of 1998 sent the international financial markets, particularly the emerging markets, into a prolonged decline, with severe effects on subna- tional government borrowing. Nevertheless, interest in new debt issues re- mains in some corners. Fitch Ratings provides ratings for the privately fi- nanced South African bond bank, Infrastructure Corporation of Africa (INCA), on its municipal investments and holdings. With the rapid change in the South African subnational government structure, investors want to stay current under the assumption that once the government structure set- tles down there will be a flood of new issues. Borrowers, too, are anxious to position themselves favorably and are keeping ratings up to date. Each agency has compiled data for more municipalities than it has been called on to rate, and each makes an effort to recast data reported in standard formats. However, not all segments of the investor community are familiar with or convinced by rating resources and opinions. Some investors express reservations about the value of credit ratings in general. Once the ratings are published, all investors must be aware of them and calculate the effects into their pricing decisions. It seldom pays to bet against the rating of a re- spected agency. Rating agencies suffer from inherent difficulties that go with being both financially viable and having a powerful effect on market behavior. First, 148 Subnational Capital Markets in Developing Countries their methodologies are necessarily proprietary. If everyone could apply the rating formula, no one would pay for a rating. Second, important factors used in ratings can be largely subjective. What is the risk of political insta- bility, including debt repudiation? (Even when the "right people" win con- trol in a country, bondholders and creditors can lose if the terms of out- standing debt are unilaterally changed.) Third, in publishing opinions the rating organizations generally assume that certain conditions and relation- ships will prevail. In a rapidly changing world, the assumptions may not hold. These problems are compounded for small agencies in developing countries, where there are few users of ratings and few issues to rate. The economics do not justify retaining skilled employees, and there is too little business to sustain competition among opinions. Credibility of Ratings The problem of credibility arises from cases where rating agencies have failed to foresee financial disruptions or have lagged behind rapidly mov- ing events, calling the rating process into question. Recent events in the U.S. market have shown that the rating agencies are not infallible and that investors and regulators are a goad to better performance.6 Another exam- ple of the fallibility of ratings is the precipitous downgrading of several sov- ereign credits in Asia during the ongoing financial turmoil. In December 1996 all of the countries were listed as having either stable (nothing on the horizon to suggest a downgrading) or positive (indications that the rating may be upgraded) credit outlooks. Not only were the ratings reduced over the next two years, but the countries also went through a continuing peri- od of negative outlook (indications that the rating may be reduced) on Standard & Poor's Creditwatch, which exacerbated the uncertainty about how far they would fall. The precipitous declines in the ratings of Indone- sia and the Republic of Korea and the serious slides of Malaysia and Thai- land caused havoc for them in the markets (table 10.1). Several other emerging market sovereign ratings have been downgraded in recent years. The drops were especially sharp following the Russian de- valuation and default in the summer of 1998, which sent all the emerging markets into a tailspin. Prior to its currency and credit crash, Russia had in- vestment grade sovereign ratings from both Standard & Poor's and Moody's on some of its Euromarket obligations. Governments, trying to protect cur- rencies, depleted foreign reserves. Depletions were followed by devalua- tions, flights of capital, and widespread concerns over domestic firms and Credit Analysis and Credit Ratings 149 Table 10.1. Credit Rating Volatility in Asia: Selected Standard and Poor's Long-Term Foreign Currency Sovereign Ratings December 1996 September 1997 December 1997 September 1998 India BB+ BB+ BB+ BB+ Indonesia BBB BBB BB+ CCC Korea, Rep. of AA- AA- B+ BB+ Malaysia A+ A+ A- BBB- Philippines BB+ BB+ BB+ BB+ Thailand A A- BBB BBB- Note: The dividing line between "investment grade" and "noninvestment grade" is drawn between the BBB and BB categories, using the Standard and Poor's nomenclature. The equivalent dividing line for Moody's is between Baa and Ba. Duff & Phelps and Fitch Ratings use the same symbols and demarcation points as does Standard and Poor's. Source: Standard and Poor's. banks making payments in foreign currencies and, ultimately, domestic currency. Subsovereign government credit ratings were also lowered, but selective- ly. Typically because of the lowered sovereign rating, the effective estimate of "macro" creditworthiness and the cap on the subsovereign ratings both fell. Between October 1997 and October 1998 Standard & Poor's lowered seven of the 18 ratings on subsovereign governments (two in Korea and five in Russia). Subnational government ratings in Central Europe and South American were not affected. Whether changes in credit ratings anticipated, coincided with, or stimu- lated turmoil in the financial markets is an important question, and it is be- ing asked with increasing frequency. Once rated, issuers run the risk that the agencies may change their minds as economic and political conditions change. Relatively well-rated Malaysia was shocked to have its rating dropped from A to BBB­ just days before a large international bond offering, a move that was sure to cost the country higher interest rates. The Malaysian prime minister called for controls over the market power exerted by the rat- ing companies. The ratings for some lower-rated Asian borrowers were not changed amid the market tumult: evidently the rating agencies got it right for India and the Philippines in the first place. Both of these on-the-fringe- of-creditworthiness countries had lagged behind the formerly high-rated "tigers" in economic growth and the pace of capital market development. Unfortunately, neither the financial markets nor the rating agencies have enjoyed any respite from the turbulent market conditions and recur- 150 Subnational Capital Markets in Developing Countries ring crises of the last five years. After a few years in the mid-1990s of what can best be described as euphoria in the emerging markets, growth has failed to occur.7 The South American credits have been especially hard hit, and several subnational borrowers have defaulted. Nonetheless, there are some bright spots, with Mexico a leading recent example. Furthermore, the difficulties in the international markets have underscored the need to de- velop domestic markets. Without stronger domestic markets, a resumption of access to the international markets is unlikely. Private Bond Insurance Allied with the development of international credit ratings has been the development of commercial bond insurance. Bond insurance acts as a third-party guarantee that debt service will be paid on time. The attraction is that the insurer carries a high credit rating from the internationally rec- ognized rating agencies. This third-party guarantee of debt with a high credit rating lowers the cost of borrowing by more than the cost of the in- surance premium. Growth of Bond Insurance Bond insurance originated in the United States and has been tremendously successful in the municipal securities market. Insurance covers half of the dollar volume of municipal bonds. For bond insurance to catch on, in- vestors must find value in the promise of insurers to meet the debt service payments, and investors must perceive differences in credit quality among issuers, usually expressed in different rates of interest demanded to offset the perceived differences in risk. The commercial insurer has a high rating from the recognized rating agencies that carries with it the promise of a lower interest rate for the insured borrower. While these are accepted no- tions in the highly developed subsovereign markets in the United States, they are still novel ideas in emerging markets. Not surprisingly, the idea of bond insurance has been most successfully applied to sales in international currency markets. In the 1990s bond insurers underwent a transformation and began to take a much broader approach. Commercial bond insurance became an in- ternational commodity as the U.S. bond market became saturated and inter- national markets became larger and more complicated. While all major in- surers had an AAA rating and stringent reserve requirements, some of the smaller insurance firms that emerged had less than prime grade and covered Credit Analysis and Credit Ratings 151 credit risks of less than investment grade. The international bond insurance market appeared promising until 1997 and the Asian financial crisis. In 1996, Standard & Poor's asked chief executives of the international insurance industry for their view of future international expansion. At the time, international business made up about 2 percent of the bond insur- ance companies' "book" (Smith 1998, p. 5). The executives estimated rapid growth to 9 percent of outstanding business in 2000 and 17 percent by 2005. The rapid expansion was expected to come in Asian markets. In 1996 a consortium of firms started up ASIA Ltd., which was to be a nonprime grade competitor for Asia business. Also, the relatively small insurer Capital Markets Assurance Company (CapMAC) reached heavily into the interna- tional markets in hopes of opening up new frontiers of profits. The Asia turmoil laid both ASIA Ltd. and CapMAC low, and CapMAC was subse- quently absorbed by the bond insurance giant MBIA. Problems in Emerging Markets The international financial turmoil of 1997 sent a strong warning that the risks of the new emerging market frontier may not have been adequately understood. On the other hand, the slow entry of the major companies was well rewarded since they avoided large capital charges and the down- grading that crippled ASIA Ltd. The insurance industry had a bad experi- ence once before, when it entered the real estate market. While the growth of private insurance can be expected to continue, it is likely to be much slower in the emerging market area than had originally been thought (Veno and Smith 1998). The primary bond insurers were not too seriously affected by the 1997 and 1998 plunges. The primary companies had only 3 percent of their par exposures in foreign-based insurance policies. Municipal-type international business is about two and half times as profitable as domestic work and has been largely restricted to superior, investment-grade issuers. With some- what less competition in the field, the possibility of higher premiums ap- peared to improve. The crises in the Asian bond markets in 1997 was followed by the broad- scale emerging markets crisis of the summer of 1998, precipitated by the Russian government's devaluation and default. The major insurers were spared the fallout because they had been slow to add Asian credits to their risk portfolios, but ASIA Ltd. was caught in the downdraft because of its re- gional concentration. Although given a respectable A rating by Standard & Poor's on its creation in 1996, ASIA's rating was lowered to BA the next year 152 Subnational Capital Markets in Developing Countries as rating downgrades of the policies in its portfolio caused a major erosion of its capital position. Short of widespread defaults, a massive systemic downgrading of credits is the worst thing that can happen to an emerging market insurer.8 As with international financial markets generally, there was a sharp con- traction in international private market insurance at the turn of the twen- ty-first century. The major insurance companies are not risk takers. They are really "rating upgraders" and "credit endorsers" rather than insurers in the classic sense. If they can avoid risk, they will. Underwriting policies and supplying enhancements on an international scale to government borrow- ers with less than investment grade issues is extremely costly since the rat- ing agencies make much heavier exactions in terms of reserves that are re- quired to be set aside to offset the higher risks. As a result, the use of insurance is likely to develop in emerging market economies as part of do- mestic schemes to encourage market access. Notes 1. Credit risk is distinct from market risk or interest rate risk, which usually pertains to how the entire debt market (interest rates and exchange rates, in the case of foreign currency denominated debt) will perform. 2. The rating agencies have come under close examination and criticism regarding both their methods and influence on markets (see While 2001 and International Monetary Fund 1999). Liu and Ferri (2002) question the dominant influence of sovereign ratings (country ceiling effect) on the rat- ings of firms. 3. In addition to bond-specific purposes, governments may use credit ratings to promote general investor confidence achieve name recognition, improve communications, and strengthen their ability to negotiate lines of credit or bolster the credit capacity of enterprises they own (see Eddy 2000). 4. The rating agencies publish articles and reports that outline their rat- ing criteria for various markets and instruments (see, for example, Moody's 1998). 5. By convention the value of this paper can be carried on the books at purchase price by financial institutions. With the emphasis on marking all securities "to market" (current prices), that practice has fallen out of favor. 6. The rating agencies missed badly on Enron, keeping its debt at invest- ment grade until just days before its bankruptcy. In congressional hearings Credit Analysis and Credit Ratings 153 the agencies maintained that they were duped along with others by the fraudulent financial information put out by the company. Nonetheless, the Securities and Exchange Commission is undertaking a study of the rating agencies and the need for more federal oversight of their activities. 7. Net long-term private sector resource (liability transactions of one- year or more original maturity) flows from capital markets to developing countries declined from approximately $160 billion in 1996 to zero in 2001. In other words, new long-term lending was completely offset by re- payments of outstanding debt (see World Bank 2002a). 8. The involvement of the Asian Development Bank and other owners of ASIA Ltd. was hoped to provide a certain degree of insulation because of the "management insights" and one would suppose the political clutch that the owners represented. The tumble in Asian ratings had terrible con- sequences for ASIA's insured portfolio. Chapter 11 Monitoring and Intervening in Subnational Government Finances A national government has a justifiable interest in subsovereign finances in general and in subsovereign indebtedness in particular.. The kinds of infor- mation required to understand the financial condition of subnational gov- ernments and subsovereign debt are much the same for governments and investors. As a result of this common interest, an active securities market is an important way to stimulate continuing interest in local financial condi- tion. Subjecting governments to continuing scrutiny and applying pressure for greater transparency are viewed as advantages of a securities market sys- tem that relies on private capital. Furthermore, what the central govern- ment is willing and able to do to avoid and cure the financial problems of subnational governments is of fundamental concern to investors. Financial monitoring may focus only on borrowing localities or on fi- nancial reporting by all localities, including annual budget and expendi- ture reviews. Much of the information needed for local debt monitoring can be generated by an active municipal securities market that demands continuing disclosure and by the availability of audited, standardized fi- nancial statements. The evolution of the credit market may be the major factor in the evolution of the relationship between the central government and its subnational partners. Once market-dictated transparency and regu- lar reporting are achieved, there should be less need for ongoing direct su- pervision or regulation of subnational jurisdictions. Central government leadership in prescribing reporting practices and making reports available to the public can advance the development of private markets. The political and financial relationships between sovereign and subsov- ereign governments are rich and varied. They are evolving along new lines, 155 156 Subnational Capital Markets in Developing Countries many of them unique to a country's tradition and position along the devo- lutionary scale. National government oversight and intervention in subna- tional government financial affairs vary fundamentally in federal systems, which leave important prerogatives to the states and their subnational gov- ernments, and in unitary governments, which have a strong sovereign cen- ter. The United States, Canada, India, and several Latin American countries, for example, have a federal system of government with specific powers and prerogatives reserved to each level. Local governments are typically subor- dinate to state or provincial governments, although often possessing some degree of independence. In unitary systems all powers of the state are de- rived from the central government, which has oversight over subnational governments. Rather than prescribe a single approach to monitoring and oversight of subsovereign conditions, therefore, this chapter first reviews international experience in developed and emerging market economies and then draws some guidelines. Examples from the United States Oversight and intervention by the states in the affairs of local governments vary greatly in the United States. As a general rule the older states in the Eastern part of the country (the original colonies) have tighter controls and oversight over local governments. In these so-called "Dillon Rule" states lo- cal governments are the progeny of the parent states and have only the powers expressly given to them in the state constitutions and by the legis- latures.1 Since the local governments are seen as accountable to the state, they often have strict reporting requirements to the states. If a local gov- ernment gets into trouble, the state is typically in a position, if it chooses, to step in and take over government operations, including removing local- ly elected and appointed officials. Direct Intervention Because the administration and finances of local governments in the Unit- ed States have been at a high level since the Great Depression of the 1930s, there are only a few examples of direct intervention. However, it can be very sweeping when used. The state appropriates functions and monitors. In the mid-1970s the State of New York stepped in to help resolve the financial crisis in New York City, establishing a control board for the city with approval power over all finan- cial decisions. The control board remained until the city had enjoyed two Monitoring and Intervening in Subnational Government Finances 157 years of budgetary balance, a total of five years. The state took back the city sales tax and used it to secure the city's debts. A new financing vehicle, the Municipal Assistance Corporation, was created to sell bonds backed by the special sales tax and to refund outstanding city notes as they came due. Debt service payments on the refunding bonds had first call on the sales tax revenues; the city had access only to what remained. The federal gov- ernment initially refused to provide special assistance, though it did ac- commodate the workout of the financial problem by providing a liquidity facility to the city. It also sponsored federal legislation that permitted the city's pension system to invest in city and Municipal Assistance Corpora- tion securities without violating federal prudential standards. The pension systems financed most of the recovery and bought some $4 billion in Mu- nicipal Assistance Corporation bonds. When the city of Philadelphia faced a financial emergency in the 1980s, it too came under a New York City­style state control board with oversight of all spending decisions. Washington, D.C. also had a financial control board that had to approve budgets and expenditures and that took over day to day control of key city services. Elected officials effectively lost con- trol over spending decisions. The state takes over. When the city of Chelsea, Massachusetts, was on the brink of insolvency in 1991 (it had little debt outstanding but was default- ing on payroll and vendor payments and there was widespread corrup- tion), the state governor removed all elected officials and appointed a re- ceiver. The receiver reported only to the governor and ran all aspects of the city, approving all contracts, tax levies, and the like. The state also created a special guarantee program to back the city's bonds, which were sold to fund several improvements. After three years a new city charter was written and approved by the state legislature, elections were held, and the city was turned back to elected officials. The city had to meet certain tests, includ- ing tests of financial operations, to stay out of receivership. Similar strong approaches have been used in the small cities of Ecorse, Michigan, and East Saint Louis, Illinois. In both cities a receiver was ap- pointed either by a state court (in Michigan) or by the governor (Illinois) to direct the financial affairs of the local government. The state creates an oversight institution and strengthens it. When the city of Bridgeport, Connecticut, ran into financial difficulty in 1991, the state of Connecticut first tried to use a limited control board approach. The board had budget approval but no power to oversee or enforce implementation of the budget. The city overspent its budget and, at odds with the state, at- 158 Subnational Capital Markets in Developing Countries tempted to go into bankruptcy under Chapter 9 of the federal bankruptcy code, which has provisions for defaults by local governments.2 The state of Connecticut opposed the city's bankruptcy petition, and the bankruptcy court ruled that the city was not technically insolvent.3 The state subse- quently stiffened the powers of the control board and provided transitional aid, and the city did not default on its debt. The Nonintervention Tradition Alongside this tradition of municipal intervention in the eastern United States is another tradition of much less oversight and nonintervention, in which local governments have much more autonomy. This appears to be especially prevalent in states west of the Mississippi River. When Orange County, in the state of California, had insufficient funds to pay its debt on time in December 1995, the county entered into bankruptcy (providing immediate protection from its creditors) and defaulted on $200 million in short-term debt. The state of California refused to become involved, and the county entered into extensive litigation and subsequent settlements on its own without state intervention or oversight. In a similar case in the 1980s the Washington State Power Supply Sys- tem, a large regional utility owned by several local governments (a combi- nation of special districts and municipalities) in three states, defaulted on revenue bonds. The bonds had been sold to finance the construction of five nuclear power plants. The Supreme Court of the state of Washington ruled that the basic contract on which the borrowing had been secured was in- valid and the borrowing itself was thus invalid (ultra vires).4 Because of the limited obligation nature of the pledge, the bondholders were simply out of luck, having no recourse to the underlying municipal governments that were clearly not guarantors of the projects. Construction of the plants ceased, and no liability was incurred by the underlying jurisdictions.5 None of the state governments tried to bail out the bondholders. Examples of Monitoring and Oversight in Other Countries Several other examples give a sense of the wide range of monitoring and in- tervention by higher levels of government. Canada In the Canadian federal system the provinces have parental powers over lo- cal governments and effectively control their finances. This is in contrast to Monitoring and Intervening in Subnational Government Finances 159 relationships between the Canadian national government and the provinces, which are highly decentralized. Localities rely on the property tax (although legislation is at provincial level, local governments can set their own rates) and transfers from the provinces. Local government capital spending and borrowing are generally subject to provincial approval, and most borrowing is done through provincial intermediaries (bond banks) that provide additional security through provincial pledges. South Africa South Africa illustrates the pressure of a changing governmental structure on intergovernmental fiscal relationships in an emerging market economy. The country has moved from a highly centralized system of government to one whose constitution recognizes three spheres of government (national, provincial, and local). The rapid amalgamation of white municipalities with the less affluent black townships has led to a variety of problems, in- cluding nonpayment of property taxes and utility bills by the newly ab- sorbed areas. Since South Africa's public sector financial structure places much of the fiscal responsibility on local governments, the nonpayment of taxes and charges has caused widespread fiscal stress. Insolvent local gov- ernments are under the control of the provinces, whose position is even more tenuous. Responding to the fiscal problems at the local level, the national govern- ment has instituted "project viability," requiring quarterly reports from municipalities on their financial position. Distressed governments are sub- ject to supervision. While the supervision provisions have not yet been tested, the quarterly financial monitoring is probably the most regular and frequent anywhere in the world. Argentina Argentina has a historically highly decentralized system of government, with significant powers given to the provincial governments. Much like the U.S. and Canadian systems, the provincial governments are the parents of the local governments. The provinces vary greatly in income and level of development. The central government raises taxes, most of which it then transfers to the provincial level to provide services. This financial structure obviously places great importance on intergovernmental transfer mecha- nisms. All three levels of government are permitted relatively free rein to borrow, which they have done primarily to cover operating deficits. Most of the financing has been through province-owned banks whose invest- 160 Subnational Capital Markets in Developing Countries ment decisions were strongly influenced by the needs of state and provin- cial governments. The result has been large and increasing amounts of un- sustainable debt, especially during the 1980s. In the 1990s the central government stepped in to bail out the provinces and cities by replacing subnational debt with national debt. The national government essentially closed the window on provincial bank lending to provincial governments. However, the provinces have continued to borrow from private banks and to pledge future intergovernmental transfers. A re- curring problem has been a lack of discipline in borrowing to cover current deficits. Since the provinces and municipalities have a high degree of inde- pendence, the central government's ability to control their behavior is lim- ited. In a new approach, the federal government and the provinces have entered into numerous agreements intended to control provincial spend- ing and borrowing. Brazil Like South Africa's, Brazil's constitution provides nominally equal status to all three levels of government. The country has had a long-standing if un- steady tradition of federalism. As in Argentina the lack of effective control by the central government led to the running up of high levels of indebt- edness by the states and the two largest cities, followed by widespread de- faults in the 1980s. The debts were rescheduled by the central government to convert short-term debt to long-term debt. A major problem was that the national government had no effective control over the amount of debt incurred by subnational governments. In the final analysis Brazil was un- willing to allow massive defaults. As in Argentina negotiations between the states and the central government are ongoing. Since 1998 and the passage of the Fiscal Responsibility Act, the central government has curbed impru- dent fiscal behavior and set tight conditions for subnational government borrowing. Transitioning Economies in Europe The transitioning economies of Eastern and Central Europe emerged from highly centralized unitary systems where the subnational government sub- divisions were service delivery points for the center and highly dependent on the central government for fiscal transfers. In addition, subnational governments owned various enterprises that generated revenues but that often operated at a loss. Financial reporting systems were designed for mea- suring levels of inputs or for tax purposes only and so provided little infor- Monitoring and Intervening in Subnational Government Finances 161 mation on the financial condition of the government. Auditing was done by state offices and was notable for both low quality and frequency.6 Economies tended to operate on a cash basis with a small and highly cen- tralized banking sector and no functioning capital markets. Major capital spending was financed by grants or soft loans and was directed by the central government or financed on a pay-as-you go basis by the locality in the case of smaller routine projects. Since subnational governments had no existence be- yond the central government, monitoring and interventions consisted mainly in the removal of officials who failed to perform as instructed. Little considera- tion was given to coping with financial emergencies of subnational govern- ments, although Hungary enacted legislation on municipal bankruptcies. More recently, governments in these transitioning economies have been moving to greater local autonomy. Financial reporting systems have been put in place to provide more useful information about local condi- tions. These systems tend to follow the European model of full accrual ac- counting, and the balance sheets are often spotty and inaccurate because of unresolved questions of ownership, value of real assets, and accounts re- ceivable. Capital financing has relied primarily on specialized loan funds or commercial banks (themselves often undergoing privatization and car- rying suspect balance sheets) that have traditional relationships with the subnational governments. Recently, loans from the European Bank for Re- construction and Development (EBRD) and grants related to accession to the European Union have become the dominant sources of long-term cap- ital for subnational governments. Establishing a Central Government System of Monitoring and Intervention Establishing a framework for monitoring subnational performance--deter- mining the appropriate institutional roles and authority to intervene and identifying under what circumstances and with what limited powers--can raise major issues of intergovernmental relationships and accountability. As a practical matter the financial information routinely provided to the central government by subnational governments may be the primary source of centralized information about the current status of subnational debt. However, the information forms need to be carefully designed, cor- rectly filled out, and promptly returned. Because debt issues have special information needs, careful consideration should be given to requiring sub- national governments to report clearly specified information about the 162 Subnational Capital Markets in Developing Countries debt. If sufficiently detailed and frequent, periodic reporting by subnation- al governments can allow central government monitoring of their financial compliance with their debt obligations. Information Needs A system of reporting that provides complete and detailed information on outstanding subnational debt issues is basic to understanding the issuer's fi- nancial condition. Such systems can be structured in various ways. France requires that the annual municipal budget include a detailed annex on out- standing debt (see box 11.1). Romania plans to establish a public debt reg- istry system. For greatest effect, such a system should be integrated into a more com- prehensive system that collects data on subnational finances in a form use- ful for analysis of financial condition. Reports on indebtedness might be re- quired to include basic descriptions of the nature, terms, and other key characteristics of the debt; certification of compliance with the debt limita- tion; and information about the collateral pledged. Notification by both lender and borrower should be required in case of a payment default, and the information should be available to the public. Having such a repository of information allows the central government to maintain a current inventory of outstanding subnational debt and makes it possible to enforce the debt service limit and monitor aggregate subna- tional borrowing as part of overall public debt management. The inventory, which could be updated annually through improved subnational debt re- porting practices, should be open to the public and prospective lenders. Ideally, financial oversight would come through market forces that de- mand the timely provision of information, which in turn determines ac- cess to the market, thereby exerting pressure for financial discipline. Where the institutions and market players are in the formative and untested stage, however, a "seed-planting" role for government is likely to be required. It is important not to discourage market initiatives or to weaken market incen- tives. Legal requirements that bond market participants disclose and send information to a central point help markets work more efficiently and prod subnational governments into assuming reporting responsibilities. Formulating and Enforcing Intervention While financial monitoring may identify problems, monitoring alone is unlikely to eliminate or cure all problems. Intervention may be needed Monitoring and Intervening in Subnational Government Finances 163 Box 11.1. Example of Information Provided in the Debt Annex of French Subnational Government Budgets Every budget presented to the local, county, or regional coun- cils in France (as well as to councils of local government associ- ations) must include a debt annex on the status of all outstand- ing loans as of January 1 of the fiscal year that includes information on the following: · Year the loan was contracted or bond was issued. · Bank or financial institution that provided the loan. · Amount of principal borrowed / debt issued. · Purpose of the loan / bond. · Maturity of loan / bond. · Currency and rate if loan / bond is in foreign currency. · Interest rate (fixed or floating). · Index used to determine the rate, if floating. · Payment schedule (annual, semi-annual, quarterly, or monthly payments). · Grace period (number of months, years). · Principal outstanding on January 1 of the fiscal year. · Interest payment for the fiscal year. · Principal payment for the fiscal year. · Principal outstanding on December 31 of the fiscal year. An annual total is calculated for the last four items above. These data also must be provided for loans guaranteed by the local government to a third party, with the name of the beneficiary of the guarantee. Source: DeAngelis and Dunn 2002. when a subnational government is in fiscal distress. What steps can higher- level governments (or others) take to protect citizens and creditors and to correct whatever is causing the financial malaise? While the remedies may be of most immediate interest to lenders and investors, their form and en- forcement are questions of national policy interest since they affect issues 164 Subnational Capital Markets in Developing Countries of self-governance, the delivery of essential services, and the health of fi- nancial markets. A viable municipal borrowing market need not have a detailed statutory intervention process. Rather, the parties can define the intervention and re- ceivership processes contractually, and these processes can be customized for a particular deal. However, there may be constitutional restrictions on the ability of a subnational government to contract for intervention and further practical problems of having courts enforce the contract. So while subnational governments should be free to negotiate monitoring and inter- vention provisions with creditors, a codified national approach helps to de- marcate the relationship between subsovereigns and the financial markets. To ensure greater certainty about creditors' and debtors' rights and to avoid the fallout that an individual default might have on other subnational ju- risdictions, it is usually better if national policymakers develop an interven- tion process through law or regulation that provides a clear framework for dealing with subnational financial emergencies. Claims after default: Who gets priority and how to collect? A legislated de- fault cure process should include a ranking of creditors and remedies. Vari- ous options are available for establishing the priority of claims. In some countries subnational governments are able to put owners of bonded debt at the head of the line. In others, the depository bank or the higher level of government gets that position. In some countries domestic creditors come before foreign creditors, a position that is likely to discourage foreign lend- ing. In the case of security, the first to take physical possession may have the advantage. Options for remedies are numerous. Creditors could be given the right to intercept funds that are due to a jurisdiction from other levels of govern- ment (see chapters 5 and 7). They could have a right to trigger imposition of an additional tax within the defaulting jurisdiction or the appointment of a receiver to control expenditures or the operations of a jurisdiction. Cit- izens also need protection to preserve minimum essential services, such as public safety and water and sanitation. Creditors should have the right to apply to courts for execution on their security interests and for judicial in- tervention. Courts should be empowered to deal with insolvency and the priority of claims among creditors and to discharge debt where the local ju- risdiction could not otherwise be made solvent. Enforcing remedies in the event of default. Predictable and timely enforce- ment of remedies for nonpayment is essential to transform a psychology of nonpayment to a hard credit culture. That requires a legal framework that Monitoring and Intervening in Subnational Government Finances 165 clearly lays out the negative consequences of a default. Failure needs to in- volve pain for the erring parties. Also important is the judicial system's effec- tiveness in enforcing financial and other commercial contracts and property rights. A lender will find comfort in a well-defined legal and political process that clarifies what happens in the event of a default and the conditions for which a lender can force a claim for payment or foreclose on collateral. In emerging market economies there is often little or no experience in judicially enforcing financial obligations against defaulting subnational government debtors. Only a record of precedent will determine how the ju- diciary will enforce such claims. Until a system has acquired practical lend- ing experience, including experience with defaults and remedies, it is diffi- cult to know whether the laws on collateral foreclosure are adequate. Substantive and procedural defects in the legal framework for a remedial enforcement system may become apparent only after there has been practi- cal experience with enforcement. Providing for a bondholder representative. In the event of a default in the payment of a subnational bond issue, the legal framework should give bondholders the right to designate a representative to act on their behalf and to pursue remedies in concert. Otherwise, each bondholder would have to pursue remedies individually, at great cost to all parties involved. That could constrain the type of collateral pledged since it might suggest that collateral must be in a highly liquid form that would allow each bond- holder to readily take possession of its share. The way around this is to designate a representative bank or trustee to look out for the bondholders' interest and act as their surrogate. Not all trustees are alike, but as markets mature, investors will find that the role is increasingly valuable in protecting their interests. Having dependable and skilled trustees also will improve the market's perception of credit quality and lower the costs of borrowing for issuers (see box 11.2). Recovery from insolvency. The insolvency of a subnational government raises concerns that do not apply to the typical corporate insolvency. Gov- ernments do not "go out of business," so procedures are needed for manag- ing the affairs of an insolvent subnational government and its relation- ships with creditors and for helping it regain financial stability. Such procedures could be initiated by the central government, by the subnation- al government, or eventually by its creditors. The procedures should clearly define what constitutes subnational insolvency. Regulations need to cover setting deadlines and defining minimum service requirements, order of payments, and limitations on the competencies of elected officials. 166 Subnational Capital Markets in Developing Countries Box 11.2. In Argentina Trustees Make a Difference Having the right trustees can make a difference in protecting bondholders' interests. During its latest financial crisis, Argenti- na has gained important experience with how the selection of trustees influences the strength of a debt transaction. In Argentina subnational governments use an intergovernmen- tal payment, the co-participation payment, for securing loans and bonds. There are two ways to intercept this payment if used as security on a loan or bond. In the more common way, the intercept occurs at the source of disbursement--at the Ban- co de la Nación Argentina (BNA), the commercial bank of the federal government. In the second way the intercept occurs when the provincial bank or the financial agent of the province receives the revenues from the BNA. Recent Argentine devaluations and widespread defaults have tested these trustee mechanisms. At the first intercept level at the BNA, every bond with a trustee has been honored. At the second intercept level hazards have arisen when province- owned banks were involved, but not when the banks had been privatized. The Province of Chaco issued three bonds for which the provin- cial-owned bank (Banco del Chaco) was a trustee. When hard times arrived in 2001, the province unilaterally deferred amorti- zation of the bonds and ordered the bank to return the funds collected in the trust escrow accounts. Bondholders brought suit against the province and Banco del Chaco. The province was sued because it had unilaterally deferred capital payments, and the bank was targeted because it broke the Argentine Trust Law by accepting and implementing the province's order. There was a very different outcome when the Province of Rio Negro deferred amortization of all of its bonds in January 2002. The province had established a trust in its financial agent (Ban- co Patagonia), a former provincial bank that had been priva- tized. Banco Patagonia continued to honor payments to bond- Monitoring and Intervening in Subnational Government Finances 167 holders and to enforce the intercept provision. In other words, private banks have been resolute trustees, and creditors are aware of this. During 2001 some commercial bank lenders proposed that the federal government permit interception of co-participation rev- enues at the Central Bank, before the funds ever got to the BNA. The private banks made the request because they believed that the Central Bank had greater independence than the BNA. The federal government rejected the proposal. Source: Argentina case study, chapter 14. A subnational government that defaults on its debt and other payments is likely to have poor financial management, overestimating its financial capacity and allowing expenditures to increase faster than revenues. It may require assistance in building a stronger financial base and in establishing good financial management policies and practices. Procedures for addressing subnational government insolvency can vary considerably. Practices in Hungary and Latvia are informative and illustrate two very different approaches. Hungary relies on the court system, with al- most no actions needed by the Ministry of Finance or the Ministry of the Interior (box 11.3). Latvia relies on the Ministry of Finance (box 11.4). In both cases a supervisor or trustee is appointed to assist the subnational gov- ernment to prepare a financial remediation program and to supervise im- plementation of the program. Latvia offers the possibility of low- or no-in- terest financial facilities to aid in implementing the financial stabilization program. In France the Crédit Local de France often requires a financial protocol to stabilize subnational finances, including raising local taxes and reducing expenditures, as a condition for additional guaranteed loan fi- nancing for subnational governments in difficult financial positions. 168 Subnational Capital Markets in Developing Countries Box 11.3. Debt Adjustment and Subnational Insolvency in Hungary Under the provisions of the 1996 Municipal Debt Adjustment Act, debt adjustment may be initiated by the municipality or by its creditor through court petition. The conditions for meeting a default situation are defined from the point when an invoice or call for payments or an acknowledged debt has not been paid within 60 days, an obligation required by court decree is not met, or an obligation resulting from a previous bankruptcy de- cree is not paid. Once a series of notification conditions have been met by the city and the creditor and the court has deter- mined that default conditions do exist, the court appoints a fi- nancial trustee. The trustee monitors the business operations of the local government and ensures the provision of mandated public services. The financial trustee must sign all obligations and payments, and the local government's bank cannot enforce any liens or make payments without the countersignature of the trustee. For creditors the debt adjustment process means that all debts become due, and all claims continue to accrue interest and penalties. Debts must be reported to the financial trustee within 60 days. Deadlines are not extended, and a creditor who fails to report on time must wait until two years after completion of the adjustment process for enforcement of the debt. The municipality's actions are severely limited once the debt adjustment procedure has been initiated. In particular, the mu- nicipality may not assume additional debt, create new enter- prises, or purchase ownership interests in enterprises. A debt adjustment committee (composed of the financial trustee, the mayor, the notary, the head of the council finance committee, and an additional council member) prepares a draft emergency budget, including a detailed listing of mandatory public functions and their financing. However, there are severe limitations. The emergency budget will not fund public health, Monitoring and Intervening in Subnational Government Finances 169 social, and educational facilities with a usage rate of less than 50 percent or facilities whose costs are more than 30 percent higher than the national average. Compromise negotiations are initiated to define the reorganiza- tion program and the debtor-creditor agreement, and the com- promise agreement is submitted in writing to the court. If the agreement meets the requirements of the law, the debt adjust- ment procedure is complete and the compromise is published in the Enterprise Registry. The financial trustee may supervise implementation of the compromise. A compromise agreement may include liquidation of some assets of the local government. Source: Hungary case study, chapter 29. Box 11.4. Financial Stabilization to Address Sub- national Bankruptcy in Latvia The Local Government Financial Stabilization Act of 1988 lists three conditions as a basis for financial stabilization action: the inability of the local government to meet its debt commitments, a value of debts greater than the market value of local assets, and a debt service ratio greater than 20 percent. The troubled local government, on recommendation of the chairman of the municipal council, the Minister of Finance, the Minister of Special Assignment, or the state auditor may initiate a financial stabilization process. The municipal council must vote on the proposed application for a stabilization plan. If the council rejects the plan, the Cabinet of Ministers may determine that the local government nevertheless should enter a stabiliza- tion program. (Box continues on the following page.) 170 Subnational Capital Markets in Developing Countries The Stabilization Act sets out options that local governments should review while carrying out their stabilization program: im- proving tax collection capacity, promoting regional develop- ment, advancing amalgamation, privatizing municipal assets, and identifying cost efficiencies to reduce local expenditures. A supervisor is appointed to assist the local government in de- veloping and implementing the stabilization program. The su- pervisor makes proposals to improve the budget (which should include finding cost efficiencies to reduce local expenditures) and to monitor budget implementation for compliance with the stabilization program. At the request of the Minister of Finance, the supervisor also can control all municipal expenditures and sign the municipality's payment orders. Source: DeAngelis and Dunn 2002. Notes 1. Dillon was a state of Kansas judge who in the late nineteenth century laid out the theory of expressed and implied powers for local governments under the constitution of the states. 2. The federal municipal bankruptcy chapter is permissive in that a state can forbid a subdivision from filing under the chapter. The State of Con- necticut, however, did not legislate such a provision until after Bridgeport had filed for protection. Since Bridgeport was found not to be technically bankrupt, the issue of whether a state could prohibit filing after the filing had been made was not decided. Most states have opted out of Chapter Nine. 3. In expert testimony, it was pointed out that the city had $400,000 in cash balances and had not demonstrated that it could not get more by sim- ply raising taxes or cutting expenditures. 4. The court reasoned that the utility only had the ability to charge for electricity actually produced and distributed. It did not have the legal abili- ty to levy charges and pay for electricity not produced or received. This Monitoring and Intervening in Subnational Government Finances 171 pledge of payment even in the event electricity is not produced or received (a "hell or high water" provision of payment) was necessary to meet debt service in case of delays in completing construction, as happened here be- cause of massive engineering and construction problems and environmen- tal concerns. 5. There was, however, securities fraud litigation. This was ultimately dismissed, since the standard for proving securities fraud is a difficult hur- dle for plaintiffs when it involves government officials. 6. Noel (2000, p. 15) sees auditing as possibly the weakest link in the lo- cal government budgetary framework, with the central audit office as the culprit. A difficulty in many countries is the shortage of private sector tal- ent and the high cost of outside auditors. The costs and difficulties of fi- nancial administration at the local level often are seen as a practical argu- ment against having direct credit market access. Chapter 12 Designing and Implementing Credit Assistance to Subnational Governments Chapter 3 identifies three groups of subnational governments based on their readiness to access private financial markets as indicated by their fi- nancial condition, managerial skills, and (to a certain extent) size. The first group includes jurisdictions that already have access but could enjoy more and better options given a more supportive regulatory and policy environ- ment. The second group could achieve access with help, including credit assistance that complements the operation of credit markets. The third group cannot access financial markets, even through market-oriented inter- mediaries, because of inadequate revenue sources. Borrowing programs should not be created for these subnational governments because borrow- ing will not solve this problem and could even exacerbate it. The question then is how to assist subnational governments that do not now have the resources to be self-financing, possibly because they do not have an adequate tax base. If the central government chooses to assist these jurisdictions by establishing a predictable and stable system of inter- governmental transfers, even smaller governments can have adequate local revenues. Revenue streams from both local sources and intergovernmental transfers can be used for capital investment, with or without borrowing. Once a subnational government has reliable revenue streams, it has the potential to support debt. Access to borrowed capital should be available to the extent that the amount of borrowing represents an acceptable level of risk. Private markets still may not serve these jurisdictions because of the small size of their financing needs, their inability to conduct analysis and planning, or their inability to deal with capital markets concepts and prac- 173 174 Subnational Capital Markets in Developing Countries tices. For this group, market intermediaries and technical assistance could be made available to bridge these gaps. This middle group of potential bor- rowers is the major focus of this chapter. National governments can provide an environment that promotes the marketability of local debt by implementing good macroeconomic and regu- latory policies. Beyond that, several questions arise when considering assis- tance for subnational governments that do not have access to financial mar- kets. Should assistance be given to help subnational governments gain access to credit? If so, what form should the assistance take in order to en- courage private capital market participation and to minimize the crowding out of private capital providers? While designing national credit assistance programs that concentrate on the most needy governments seems a worthy policy objective, making cheap credit available from the central government is not without hazards. Whether as loans or grants, assistance programs have the potential to undermine private credit markets (see box 12.1). Assistance can take several forms, ranging from technical assistance and financial assistance to direct lending and interest rate subsidies to encour- age private market participants to join a transaction. At least three basic questions should be asked to determine whether to use a given technique: · Does the assistance technique leverage private sector investment? · How likely is it that the assistance will crowd out private sector capital? · Does the technique increase the risk of moral hazard? How likely is it that it will be misinterpreted as a central government guarantee? These questions are explored in the context of several forms of assistance that might be provided to promote private capital market development. Technical Assistance Technical assistance to help subnational governments become familiar with credit market practices and to become more creditworthy is the most likely form of assistance to attract private sector interest and the least likely to crowd out private capital. It is also the least likely to raise the risk of moral hazard. Technical assistance and training in accounting and budget- ing, identifying and analyzing capital investment projects, and operating and managing facilities expand managerial skills and encourage more effi- cient financial practices. Designing and Implementing Credit Assistance to Subnational Governments 175 Box 12.1. The Subnational Government Retreat from the Private Credit Market in the Czech Re- public Sometimes progress to more open markets for subnational gov- ernments can be reversed by national government policies, as happened in the Czech Republic. Immediately following liberal- ization commercial bank loans to subnational governments be- gan to grow in the early 1990s. This growth was soon cut short by competition from state-based loan funds and capital grants. Commercial loans to subnational governments had been en- couraged through the formation of the Municipal Investment Fund, a USAID-supported project that provided a discount facili- ty to commercial banks. Czech cities tend to have heavy urban infrastructure responsibilities, and capital spending accounts for a large share of their budgets, typically about 30 percent. How- ever, the national government then chose to follow a less trans- parent capital grants policy, which together with low-cost loan programs that ignored creditworthiness undermined the emerg- ing bank lending market. The soft loans from the state created moral hazard, and as conditions deteriorated, cities began to default. An estimated 73 percent of the State Environmental Fund loans to subnational governments were nonperforming, for example. Finally, to meet the EU pre-accession Maastricht convergence requirements on government debt limitations and to conserve credit access for national government use, the central govern- ment ruled that it must approve all subnational government loans. This effectively stifled subnational government borrow- ing from private sources. Source: Czech Republic case study, chapter 28. Technical assistance works much better with practical applications than with abstract principles and when focused on creating local institutional and technical capacity. Technical assistance in capital planning, cash flow projections, and project management are particularly supportive of in- 176 Subnational Capital Markets in Developing Countries creased capital market access. These skills allow the subnational govern- ment to work within budget constraints, to match revenues and expendi- tures, to figure out how much to borrow and for what purposes, and to de- termine how quickly it can and should repay loans. Either public or private lending entities can help provide access to markets, especially if standard- ized documentation and processes are developed. Standardization helps to resolve questions of security and keep costs down.1 Financial Assistance Financial assistance to help subnational governments gain access to pri- vate credit can take several forms. However, direct financial assistance that is insulated from market testing has significant drawbacks and risks, be- cause the risks of adverse selection and moral hazard (see chapter 2) are al- ways involved. To the degree that assistance from the center is institution- alized, it can foster a culture of long-term dependency and impede market development. The lure of cheap credit provides an incentive for subnational govern- ments to be or appear to be needy rather than self-sufficient. Direct assis- tance also creates hidden subsidies in the form of contingent guarantees and enhancements. It can crowd out the private sector, which typically sets higher credit standards and charges more for lending. Direct assistance usu- ally is less efficient at leveraging private sector resources than is technical assistance. Concessionary financing (with terms and conditions more favorable than those available in the commercial market) can also distort choices. Financial assistance reduces only capital costs to the borrower, not future facility oper- ating costs, which will increase with the new investment. A borrower whose only source of credit is through preferential assistance, rather than capital from hard credit sources, may not have been required to fully investigate op- erating costs or to build them into budget planning. The governmental bor- rower may have little or no capacity to properly operate, maintain, and ulti- mately replace the facility, which then rapidly slips into decline. However, concessional finance for subnational governments continues to have a role in most economies, either to encourage desirable activities or to surmount barriers. Furthermore, careful design can reduce the drawbacks and risks, even if it cannot eliminate them. Fundamentally, direct financial assistance should always have an exit strategy and a plan for shifting obliga- tions to commercial credit markets. The assisting government can thereby Designing and Implementing Credit Assistance to Subnational Governments 177 absorb some of the risks that are unacceptable to the private credit market. This might mean finding a way to eliminate a narrow risk (such as environ- mental risk) by providing risk insurance. Alternatively, it might mean taking a junior lien in order to comfort potential private lenders or providing a guarantee on the "long end" of a debt structure if commercial lenders are able to provide short- and medium-term principal maturities. Direct Lending Direct lending can be an inefficient form of financial assistance and is like- ly to crowd out private lenders and invite moral hazard. Many direct lend- ing programs aimed at subnational governments have been directed from the center. These loans are often made to unwilling and inattentive subna- tional governments, which end up treating them as grants. However, there can be constructive direct lending roles. The International Finance Corpo- ration's A/B loan syndication and certification structures have demonstrat- ed that leverage efficiencies can be achieved in the private sector with such instruments, if they are well designed. To increase leverage and reduce crowding out and moral hazard, direct lending should be designed to induce cofinancing by commercial lenders. The smallest possible direct lending role required to achieve this objective will minimize the risk of crowding out and maximize the efficiency of the assistance rendered. Thus, for example, if a 5 percent junior lien position will induce the private sector to join in cofinancing a loan, the provider of this form of assistance should be prepared to forgo a larger loan program. Although direct lending programs have had a poor record of loan repay- ment, the tide appears to be turning in some countries (see box 12.2). Cred- it discipline, if it is instilled into direct lending programs, can help prepare borrowers for the realities of the private market as long as sufficient eco- nomic inducements can be designed to enable borrowers to graduate to pri- vate market access. Debt Service Subsidies and Public-Private Cofinancing Debt service subsidies resemble direct lending in that they constitute ongo- ing payment streams to support subnational borrowing and so can be inef- ficient. They are more likely to lead to moral hazard than are more indirect or softer forms of financial assistance, such as insurance, partial guarantees, or technical assistance. Nonetheless, they can be useful tools if they are 178 Subnational Capital Markets in Developing Countries Box 12.2. Moving from Soft to Hard Credit through Enforcement of Loan Collections: South Africa's Experience Development banks have had a very poor loan repayment record, which has made many observers skeptical of the ability of subnational governments to make the transition into private markets with hard credit demands. However, some countries are seeking to correct the situation by holding delinquent borrowers responsible. South Africa demonstrates one way of doing this. In January 1996 the Development Bank of Southern Africa (DBSA) inherited the Ministry of Finance's development loan portfolio for subnational governments. The portfolio consisted of some 390 loans representing about 900 million rand ($50 mil- lion) that had been made to subnational governments primarily under the pre-1994 regime. At the time of the transfer most bor- rowers were on time with their payments. Amid the turmoil of the transition to the new governmental structure, many of the subnational government obligors began to go into default. The DBSA, which saw itself as a bank with commercial incentives and a capital position to protect, recoiled at the growing delinquency rate. While the original terms of the loans might have been concessionary, the DBSA's role was to keep the payments on schedule and to instill discipline into bor- rowers. DBSA was not expected to lose money and erode its capital base; its goal was to make reasonable returns to capital, while promoting longer term, socially useful development. Accordingly, DBSA moved to deal with the subnational authori- ties to bring loan payments back on schedule. Loan officers were assigned to each region and given procedures for going after overdue loans. In three provinces, 32 of the 40 loans that had defaulted were put back on a timely basis using technical assistance and the threat of closing off future credit. In South Africa both government and private lenders have the power to seize assets of borrowers. Source: Petersen and Crihfield 2000. Designing and Implementing Credit Assistance to Subnational Governments 179 well designed to provide the smallest subsidy necessary to induce private capital market participation and if they are used solely when this is the only tool that will make the borrower creditworthy. Linked deposits and co-lending programs are two devices used to subsi- dize interest costs through the private credit system. With linked deposits, a commercial bank might receive a deposit from a government intermedi- ary that agrees to a reduced rate of interest if the bank agrees to use the de- posited funds to make a loan for a particular purpose to a subnational gov- ernment. The private institution still takes the credit risk, does the credit analysis, and administers the loan. With co-lending, the government inter- mediary makes a loan for a portion of the principal amount at a reduced rate of interest, while the private lender makes its share of the loan at the conventional rates. The borrower gets the advantage of the blended rate on its loan. The private lender, however, still has its principal at risk and ad- ministers the loan, with the intermediary as a partner in the transaction. Major public-private infrastructure projects often have capital needs that exceed the financing capacity of a developing country's nascent credit market or banking system. A cornerstone of a credit assistance program should be a lending facility designed to attract rather then supplant private capital in financing subnational government infrastructure projects. Thus, in another creative use of cooperative devices, donors could require recipi- ents of their credit to design loans to attract private sector participation in infrastructure projects. This participation might be by private financial in- stitutions or by project proponents that may bring their own equity and debt financing, such as in a public-private project-financing scheme. Such loans could be coursed through a government financial entity (GFI), which could retail the loan directly to a qualifying project (figure 12.1) or wholesale the loan proceeds to a private financial institution (PFI), which would then on-lend to a project (figure 12.2). The government finance insti- tution, as a condition for receiving the loan from the donor, could be re- quired to construct deals that attract private sector participation in infra- structure projects. The cofinancing approach could entail bank loans or bond issues where there are different tranches with different lien positions, maturity structures, and loan repayment mechanisms.2 The idea is for the government finance institution to leverage private sector funds by taking various cofinancing positions in the transactions that provide comfort to the private participants or by taking positions with greater risk or less liquidity. The government finance institution is able to better absorb the added expo- sure because the donor credit line has been constructed for that purpose.3 180 Subnational Capital Markets in Developing Countries For these "market-friendly" co-participation variants to succeed, the re- turn on investment to private sector participants needs to be competitive with that obtainable elsewhere. Thus, the government finance institution, in taking the long view and acting as a catalyst for financial market devel- opment, would need to act as companion and facilitation lender and de- sign issues that would stimulate private participation. For many such insti- tutions, accustomed to market monopolies when lending to subnational governments, this would be a difficult role. The use of an on-lending facility is customary practice for donor-based loans, but the active engagement of private sector banks and financial in- stitutions is not. Getting the government finance institution to behave in this market-building way, perhaps compensating it for its catalyst role, would need to be wired into the donor's loan conditions.4 A combination of inducements and requirements might be built into the loan, to encour- age use of the facility while ensuring that the government finance institu- tion does not gain all the advantages of long-term money and drive out in- vestment through commercial banks and the bond market. Donor credit facility PFI GFI Local unit Project Private proponent Note: In this example, a subnational government and a private participant join in a project, financed by funds from a private finance institution and from a government financing institution. The government financing institution looks to a donor credit facility for loan funds or enhancements. Figure 12.1. Retail On-Lending by the Government Financing Institution Designing and Implementing Credit Assistance to Subnational Governments 181 Donor credit facility GFI PFI Local unit Project Private proponent Note: In this example, a subnational government and a private participant join in a project, financed by a private fi- nance institution from funds that have been on-lent in part from a government financing institution. The government financ- ing institution looks to a donor credit facility for loan funds or enhancements. Figure 12.2. Wholesale On-Lending by the Government Finance Institution Guarantees, Insurance, and Intercepts Guarantees are a traditional and important form of financial assistance (see chapters 5 and 10). Their contingent nature makes their cost difficult to measure at the time the guarantee is given. While guarantees can lead to lax lending practices and impede the development of effective private mar- kets, guaranteeing specific risks or specific maturities may be worth consid- eration. The ability of a credit assistance provider to reduce or eliminate specific risks in a transaction (such as certain environmental hazards or the repudiation of certain contractual obligations) or to back maturities that the domestic private sector is unwilling to provide can leverage private cap- ital investment. Properly designed and implemented such use of guaran- 182 Subnational Capital Markets in Developing Countries tees can reduce the risks of crowding out and moral hazard. Some of the World Bank's guarantee operations have begun to demonstrate the utility of guaranteeing specific risks or maturities as a means of inducing private capital providers to participate. One way to reduce the risk that such credit enhancements will crowd out commercial lenders is to price them according to the degree of risk pre- sented by each borrower. In this way local borrowers and commercial lenders see the costs involved in securing guarantees and so are more likely to treat the guarantees as having a cost. When this is done according to commercial standards, with costs and expected losses reflected in the fees charged, the guarantee is transformed into a form of insurance. While some might oppose the idea of charging needier borrowers more than those that are better off, there must be incentives to improve financial op- erations if subnational governments are ever to stand on their own in cred- it markets. Buying down part of the costs with grants but making the is- suers borrow at risk-adjusted rates on the margin may be one way to force governments to pay attention to market interest rates and to scale projects accordingly. Another option is to price enhancements with "seasoning" premiums that can be partially rebated as borrowers live up to their obliga- tions and see their circumstances improve. As discussed in chapter 5, intercepts of national payments to subnational governments are a form of financial assistance that need not have any sig- nificant cost to the national government.5 Intercepts can be a powerful cred- it enhancement--and an almost essential one, given the highly centralized system of tax collection in many emerging market economies.6 A stream of stable, predictable intergovernmental transfers can be made pledgeable and interceptable, which can enhance creditworthiness so long as the use of the transfer is not overly restricted. Significant penalties or administrative fees when the intercept is exercised could encourage subnational governments to manage the debt payments in a businesslike fashion and not to misuse the intercept mechanism to cover lax practices. An intercept mechanism can leverage private sector funding rather than crowd it out. In the Philippines intercepts are being combined with guarantees, in the form of bond insur- ance, to enhance bonds sold by subnational governments (see box 12.3). To qualify for insurance, borrowers must achieve a minimum credit rating and pledge a portion of their future intercept payments to debt service. In case of default, the insurance company continues to pay the debt service to in- vestors and assumes their rights to receive the intercept. Designing and Implementing Credit Assistance to Subnational Governments 183 Box 12.3. The Philippine Local Government Unit Guarantee Corporation Under the sponsorship of the Philippine Bankers Association, a banking consortium of 22 domestic and foreign banks has created the Local Government Unit Guarantee Corporation to provide guarantees on loans made by participating financial institutions to local governments. Some 230 million pesos were raised by sub- scription from the participating banks, deposited into a special ac- count, and made available for backstopping the guarantees. The guarantee is expected to stimulate private commercial bank interest in local government credits. For institutional and regula- tory reasons, local governments have been borrowing only from government financial institutions. The government and government financial institutions are under pressure to open up the local government debt market to greater competition and to develop a municipal bond market. The government financial in- stitutions are being privatized. The guarantee program gives comfort to the private banks as they start lending to local governments. The program is expect- ed to serve as an enhancement for bond issues. The guarantee depends in large part on the pre-assignment of the local govern- ment's intergovernmental transfers to the corporation, so that the transfers can be tapped in the event of default. The initial pro- gram was geared to the 120 largest local governments;once the guarantee system was in place, however, the program soon reached down to smaller government units. As of early 2003 the corporation had insured 11 bond issues amounting to over 1.6 billion pesos (about US$35 million). Bond issue activity slowed during the political turmoil and economic slowdown of the early 2000s, but the insured bonds are paying debt service on time and in full. As a result, the use of the pro- gram in the case of a default remains to be tested. Source: Philippines case study, chapter 26. 184 Subnational Capital Markets in Developing Countries Intermediaries for Small Borrowers Should a special intermediary be created for jurisdictions that cannot ac- cess credit markets through existing market mechanisms? Special interme- diaries should complement rather than replace existing commercial lend- ing and underwriting institutions. In some countries the private sector may be able to provide such intermediation, without the need to create a new government agency or function. While this may be desirable in principle, a small issue may not attract the market's attention because it would not be economical to finance in the formal securities markets. Many intermediary models are available, including bond banks (see be- low), bond pools, revolving loan funds, and municipal lending institu- tions. Such an institution might borrow in its own name and use the pro- ceeds to purchase debt instruments of subnational borrowers (bond banks), or it might assemble and repackage municipal debt instruments and make them available to the market (bond pools). A major attraction of such structures is that they can provide economies of scale in issuance and, be- cause of the larger size of issuance, improve the chances of attracting inter- est from secondary markets. Any intermediary function has costs, which may include administrative costs, subsidized re-lending rates, or credit enhancement costs. However, with a properly designed and efficiently run intermediary, the costs will likely be less than those involved in outright capital grants. Intermediaries have the additional virtue of helping local officials understand the trade- offs involved in debt finance (Noel 2000). Intermediaries can be designed to provide several services to subnation- al governments (see box 12.4), including access to capital markets for gov- ernments that otherwise would not have access, savings on the fixed costs of debt issuance, streamlined and standardized borrowing procedures and documentation, assistance with capital planning and cash-flow projec- tions, and pre-structuring of loan packages. The higher-level government also may decide to offer direct financial assistance, such as credit enhance- ment (see chapter 11) or the re-lending of intermediaries' funds at subsi- dized interest rates. The more passive the financial assistance and the more it is used in tandem with normal credit channels, the better, to avoid the moral hazard risks associated with direct financial assistance that is insu- lated from market forces. Overall, it is better to expose the novice borrow- er to the actual costs of capital and the discipline of the market, at least on the margin. Designing and Implementing Credit Assistance to Subnational Governments 185 If the objective is to promote local self-sufficiency, it is generally advis- able to avoid enhancement methods that are nontransparent, reward dys- functional governments, or crowd out private investment. If there is a stream of stable, predictable intergovernmental transfers for jurisdictions lacking the resources to be self-sufficient, these transfers could be made pledgeable and interceptable. This would enhance creditworthiness and leverage private sector funding at little or no cost to the national govern- ment. However, the extent to which otherwise impecunious governments should be encouraged to borrow remains a judgment call. For subnational governments with slim prospects for financial self-sufficiency, it may sim- ply be a way for the higher level providing the transfers to pass the buck of indebtedness. Box 12.4. The Tamil Nadu Urban Development Fund, India The Tamil Nadu Urban Development Fund evolved from a mu- nicipal trust fund to a fund financed and managed by the public and private sectors. The initial fund was financed entirely by the public sector, and while it was financially viable, it was too small to meet the demand for urban infrastructure investment. To increase the impact of the fund, it was converted into an au- tonomous financial intermediary. The new fund has 30 percent participation by the private sector and is managed by Tamil Nadu Urban Infrastructure Financial Service Ltd., a private man- agement company. Operations have been widened to include urban infrastructure projects sponsored by private investors. To further pursue the project's objective of poverty alleviation, a new grant fund was established to finance poverty alleviation projects for specific low-income populations. In addition, the participating financial institutions have committed to contribute an amount equal to 44 percent of the Tamil Nadu government's initial contribution. The ultimate objective of the fund is to pro- vide self-sustainable financing while mobilizing private savings for urban infrastructure investment. (Box continues on the following page.) 186 Subnational Capital Markets in Developing Countries The fund is administered by a board of trustees nominated by the government of Tamil Nadu and the participating financial in- stitutions. The participating financial institutions include Indus- trial Credit and Investment Corporation of India, Ltd., the lead- ing managing partner of the Tamil Nadu Urban Infrastructure Financial Service Ltd.; Infrastructure Leasing and Financial Ser- vices, a leader in the development and financing of private in- frastructure projects in India on a limited recourse basis; and the Housing Development Finance Corporation, a leading fi- nance corporation in housing and regional development. The strong reputation of these institutions in India's business and fi- nancial community should help the fund raise additional re- sources from other private investors. Source: India case study, chapter 24. Securitized Loan Pool Another mechanism for credit assistance to subnational governments is the securitized loan pool. Securitization means the sale of a bundle of future cash flows arising from a specified underlying pool of loans. Proceeds from the loan payments are passed through to the investor in the form of inter- est and principal payments. Several variations are possible: the debt service payments may or may not be secured by the underlying loans themselves (and the underlying security that they individually provide) and may or may not have recourse to the issuer. Certain restrictions are placed on the loans admitted to the pool, either for the benefit of the investor (nonrecourse) or at the insistence of the pool sponsor or enhancer (where the pool is enhanced). The pool can be accessed either directly by individual subnational governments borrowing or indi- rectly by borrowing from a government finance institution (GFI) or private financial institution (PFI) that holds the pooled loan portfolios (figure 12.3). Several configurations of securitization are possible, from pooled issues carrying an "umbrella" guarantee or access to a liquidity facility or bond in- surance to strictly nonrecourse pooled securities that provide an "over- Designing and Implementing Credit Assistance to Subnational Governments 187 Donor credit facility Local GFI unit Loan Bonds pool Local Investors unit PFI Figure 12.3. Securitization of a Loan Pool pledge" of revenues to the underlying subnational government securities. An overpledge means that the flow of payments on the underlying loans is fractionally higher than that on the securitized debt. Thus qualifying loans would generate more debt service than the bonds sold by the pool. The ex- cess earnings over the debt service could be used to pay the debt off faster or could be retained as income by the government financial institution that originated the pool. The pool of loan obligations can be open or closed, with an open pool permitting replacement of debt that matures or defaults with comparable loans. As depicted in figure 12.3, the pool could be backed up by a donor-based enhancement to increase the marketability of the bonds. Securitization makes possible relatively large bond issues that create the potential for large trading volumes. Bond pools help investors become fa- miliar with subnational government credits and provide comfort for enter- ing into future transactions. For example, the prospect of future pool fi- nancing would permit banks to extend the maturities on the new subnational government underlying loans. The pooling and securitization technique could be especially useful in devising standard form documenta- tion and in providing better market access to small borrowers. The pool ap- proach would provide economical access to the credit markets for smaller 188 Subnational Capital Markets in Developing Countries localities that are creditworthy, extend maturities on loans, and ultimately work to reduce their loan costs.7 A bond pool has potential drawbacks, however. First, sale of the subna- tional government loan assets might encounter resistance from existing lenders, in some cases eroding balance sheets by reducing the stock of per- forming loans among their assets.8 Thus, successful bank lending experience can work as an impediment to expanding the market options for subnation- al government borrowers. A second concern is the need to establish the legal status of a bond pool with respect to securities and banking regulations.9 This proved to be an impediment in efforts to create a pool of loans to sub- national governments in Poland (DeAngelis and Putnam 1999). Bond Banks Another mechanism for assistance to subnational government is the bond bank, which borrows in its own name and uses the proceeds to purchase debt instruments of subnational borrowers. Bond banks originated in the United States to improve access to the financial markets for small local governments. The operation and scope of bond banks have varied, depending on relative fi- nancial priorities and their legal and political environments. As the bond banks gained experience, they frequently took on specialized areas of activity, such as financing environmental activities, local schools, short-term borrow- ing, and equipment leasing. They also moved into limited obligations, struc- tured transactions, and credit enhancements (see box 12.5). A survey of state bond banks in the United States shows a variety of ad- ministrative and program structures and financing experience that can be useful to municipal credit markets in emerging market economies (Petersen 1998). Because bond banks compete with private lenders and dealers and often can finance at lower costs or on better terms, bond banks have been resisted by commercial banks and securities dealers in many states.10 After early adoptions in several states, the bond bank movement in the United States slowed in the face of opposition from competing interests and con- cerns about stretching state credit enhancements too thin. More recently, interest at the national level in replacing recurring capital grants from the central government with revolving funds has reactivated interest in state- based financial intermediaries, including the traditional bond banks. The bond bank concept has been slow to catch on in developing and emerging market economies, for a variety of reasons. First, in transitioning economies subnational government credit needs have been an orphan. The Designing and Implementing Credit Assistance to Subnational Governments 189 Box 12.5. Assisting Small Bond Issuers: The Bond Bank Option The United States is often thought to have a highly sophisticat- ed financial market, with knowledgeable and skilled investors and issuers. However, that is not necessarily the case for the es- timated 40,000 subnational government issuers in the U.S. bond market, many of them small and unsophisticated. Their access to markets has improved as a result of well-established legal and regulatory processes, the availability of skilled advisers, and competition among potential lenders. A combination of state- backed financial intermediaries such as bond banks, private bond insurance, and preferential federal tax policy keeps com- petitive pressure on dealers and banks to provide services to small issuers. As a result, the typical U.S. small local govern- ment credit has become very competitive in the markets. To encourage this largely market-driven process, a good deal of attention has been given to upgrading local government finan- cial management practices and reporting. States have long had an oversight function for local governments in their jurisdictions but have worked at it with increasing vigor in recent decades. Bond banks, bond insurance, and other organized lending and credit enhancing programs have required local governments to report their financial condition regularly following generally ac- cepted accounting principles. These developments, along with the widespread use of credit ratings and recently adopted secu- rities-related reporting requirements of the U.S. Securities and Exchange Commission, have also worked to standardize and regularize financial reports. Source: Petersen 1998. initial thrust was to create equity markets to handle the new private inter- ests, and the need to finance the central government has taken precedence in debt markets. Second, disruption in the political and fiscal structures in transitioning economies has made subnational governments appear to be poor credit 190 Subnational Capital Markets in Developing Countries risks. There was no clear sense from one year to the next just what respon- sibilities and powers subnational governments would have in the emerging regime. Such issues as ownership of property, for example, placed a cloud over lending practices that traditionally had been tied to the provision of physical collateral. In the absence of information and experience, the "name" and size of a subnational government have had disproportionate importance. Third, the economics of transactions--it is more efficient to do the due diligence and promotion of one large loan or bond deal than to round up several smaller transactions--has resulted in a strong tendency to leave the smaller subsovereign loans to the commercial banking system or municipal development funds and to use the bond markets only for larger, more profitable transactions. Fourth, in many transitioning countries domestically derived or donor- induced development funds offering loans on concessional terms and asso- ciated grant programs have effectively undercut competition from the pri- vate sector. Long-term loans, much larger and on more favorable terms than commercial markets can offer, are frequently tied into grants.11 Fifth, a legacy of protection afforded by "special" municipal borrowing windows, such as development funds, have shielded local governments from the temptations and tribulations of private sector financial markets. The difficulty is that the new fiscal order calls for local governments to be more self-reliant and market-oriented. The bond bank approach offers an opportunity for smaller subnational governments to enter the market to- gether, enjoying the benefits of a broader market appeal while minimizing the risks of a single mistake or misfortune. With experience, stronger gov- ernments may find it better to borrow on their own in the markets. These and other inducements to prudential behavior by governments can be built into the mechanics of the bond bank operations. Liquidity Facilities Various credit enhancements can be used to help financial markets mature and better meet the needs of subnational borrowers. Emerging markets are chronically short of long-term investible funds, as both institutional and individual investors are leery of making long-term commitments of their cash. Thus, one approach is to enlist their short-term investments into long-term capital for borrowers. A useful tool is the "put" or "tender" op- tion, which allows investors to cash in their holdings of bonds at set dates prior to the debt's maturity date. Put options are usually found in variable- Designing and Implementing Credit Assistance to Subnational Governments 191 rate interest markets, where bonds are repriced on a recurring basis accord- ing to interest rate fluctuations.12 Bonds carrying the option, no matter what their final maturity, trade like comparable obligations that are due at the next put/repricing date.13 Put options allow investors an "early out" and so, by definition, provide market "liquidity." The presence of the liquidity facility, which is typically a bank stand-by loan agreement or letter of credit, ensures the current bondholder, if it elects to put the bonds, that a purchaser for the bonds will always be there and the money will be returned quickly on demand. Unless the issuer pays off the bond, the bond that is put back can either be resold by a repricing agent (usually a securities firm hired for that purpose) at the prevailing rates of interest or converted into a bank loan from the liquidity facility to "warehouse" the security until a buyer is found.14 There are many mechanical details in designing and operating a liq- uidity facility, but where markets are thin, the put option can provide in- vestors with liquidity where a secondary market has yet to take root. It provides some intriguing possibilities among the arsenal of forms of donor credit assistance. For example, a credit facility provided by a do- mestic bank might be backed up by a donor-assisted loan facility. Under the terms of a "put" option that would be incorporated into bond issues, the liquidity facility could be availed by "qualifying" subnational govern- ment bonds.15 The basic mechanics of a liquidity facility are depicted in figure 12.4. The figure shows the alternative pathways that a bond might take, de- pending on borrower needs and market conditions. The issuer first sells bonds with a put option to investor 1 (pathway A). Under normal circum- stances (pathway B), the repricing agent reprices the bonds to maintain a market acceptance for them, and if investor 1 puts its bonds, they are resold to investor 2. However, if the repricing agent is unsuccessful in im- mediately reselling the bonds, the liquidity-providing bank provides a loan to pay off investor 1 (pathway C). Were the liquidity facility itself to be incapacitated for some reason, the stand-by loan agreement, in this case provided by a donor providing a stand-by commitment, would be activat- ed (pathway D). Put options and liquidity facilities involve fees, and the economics of such a devices are improved when there is a relatively large volume of se- curities involved, such as with a pool or a bond bank. The repricing of bond issues at the time of the put date means that the debt service pay- ments will change for the underlying borrower after each put and repric- 192 Subnational Capital Markets in Developing Countries Donor standby loan facility Local government Liquidity issuer facility Repricing Use put agent option Investor #1 Investor #2 Explanation: A. Issuer sells bonds to Investor #1 that have a put option. B. If investor elects to exercize put, bonds are placed with repricing agent and resold to Investor #2. C. If repricing agent is unable to place bonds with an investor, the bonds are placed with the liquidity provider, which in turn activates a loan to the issuer. The loan is repaid when the bond is resold or paid off. D. If liquidity provider has insufficient assets to carry the bond until it can be resold, then it can borrow from donor standby loan facility. Figure 12.4. Mechanics of a Liquidity Facility ing. However, repricing reflects only changing interest charges, and the changes in debt service are easier to absorb if the principal component is much smaller in the first place; that can happen if the maturity of debt is stretched out. The great advantage to the put option is its ability to extend the maturi- ty of debt. A liquidity facility allows the bond, if it is otherwise creditwor- thy (current in its debt service payments) to be priced as short-term debt, while it allows for a longer maturity, that is, the date when the issuer must by contract repay the principal.16 Having a bond's principal payable in a se- ries of installments over 10 to 15 years, as opposed to 2 to 3 years, dramati- cally lowers the annual debt service. Designing and Implementing Credit Assistance to Subnational Governments 193 Grant and Loan Integration Of great importance in inculcating credit market discipline is to ensure that the availability of grants does not undercut the use of credit at market rates by subnational governments and projects that can afford it. Grants should be tailored to meet the needs of projects that are not creditworthy or that have capital costs that, once sufficiently reduced through grant assistance, can be partially financed at market rates. Capital subsidies come in a number of forms, from paying explicit subsi- dies to offset interest rates, to extending credit on very favorable terms, such as below-market rates and with extended grace periods, to providing longer maturities than are available in capital markets. The form of the sub- sidies has operational implications. Gains from Integrating Loans and Grants Encouraging efficient use of resources and access to credit markets argues for integrating grants and loans.17 A conscious regimen of exposing subna- tional governments to private market demand and credit expectations will benefit the development of both private lenders and government borrow- ers. To foster that process, grants generally should take the form of an ini- tial capital grant that lowers capital costs to levels that can be financed by a loan. The conditions on the loan should be similar to those in private mar- kets, with the exception that the loan will likely be of longer maturity. While the concept is generally applicable to both revenue- and non- revenue-producing projects, the initial application will likely be limited to non-revenue-producing projects. The grant also might take the form of loan "forgiveness," with a proportion of the original debt principal written down if the borrower meets its debt obligations on time. The idea is to create posi- tive incentives for the borrower to be faithful in meeting its obligations. While the advantages of integrating grants and loans are easy to see, ac- tually accomplishing the integration requires technical guidance and data on project costs, benefits, and the resources of customer groups (see box 12.6).18 Such data are likely to be sketchy, but formulating general parame- ters will help in decisionmaking. Perhaps one of the more straightforward applications of integration would be in such utilities as water supply and waste disposal, typically large users of capital. However, the integration concept should be applicable not only to most enterprise activities that generate some revenues but also to social or non-revenue-producing facili- ties that are supported by general revenues and transfers. 194 Subnational Capital Markets in Developing Countries Box 12.6. A Brief Illustration of Grant-Loan Integration: An Example from Indonesia It is useful to illustrate some of the concepts and terminologies of the grant-loan integration concept. The example is based on Indonesia, although the technique is generic in concept. Determining "market-proxy" costs. Suppose that constructing and equipping a facility will cost 20 billion rupiahs (Rps) if it is built in an efficient manner to meet projected demands. Sup- pose also that its annual operating (O&M) costs (labor, materi- als, energy, routine maintenance) will be Rps 1 billion, again with efficient operation and adequate maintenance. These con- struction and operating figures are derived from a feasibility study conducted for the project. Note that the cost figures are calculated irrespective of how the facility is financed and are based solely on technical and economic efficiency grounds. In addition to the annual operation and maintenance costs, there is a potential annual debt service (DS) component. Thus, were the project to be totally financed at "market rates" and on a long-term basis (the economic life of the asset), then the full "market-proxy" annual debt service would be DS', applying a standard level debt service schedule. Leaving aside any equity contribution to be made by the com- munity, the capital cost is first estimated at 100 percent debt-fi- nancing over the useful life of the improvement. K = D' where K is the capital cost of the facility and D' is the debt amount. An annual cost to repay the debt can now be derived, depending on the interest rate and the maturity. For comparison purposes using the market proxy approach, an annual level debt service factor can be applied that is designed to pay inter- est and principal at approximately constant amounts over the life of the loan (see chapter 6). A 20-year loan at a 15 percent rate of interest would require an annual payment equal to 0.1598 of the principal each year. At Designing and Implementing Credit Assistance to Subnational Governments 195 level debt service the "market proxy" annual debt service cost is about Rps 320 million a year on a 20-year Rps 20 billion loan. Thus, with the market-proxy financing calculation, the annual total combined operating cost (Rps 1 billion) and capital cost (Rps 0.32 billion) of the facility (E') would be E' = O&M + DS' or total annual expenditures of Rps 1.32 billion. This annual market proxy cost is a standardized starting point of the analysis. It is intended to replicate what the full costs would be annually were the project financed at market interest rates. Therefore, it represents a proxy for the cost of capital in the economy, making allowance for the fact that long-term financ- ing (that is, financing for the useful economic life of the facility) is unlikely to be possible in the immediate future. The next step is to compare this full cost concept with what is affordable. Determining affordability. The project's full annual cost needs to be compared with what the community and users can afford. This can be the most difficult part of the process. Assume that the facility will charge tariffs based on charges to residential and commercial sectors and their expected volumes of usage. The first step is to determine whether annual revenue, R, would cov- er the full market-proxy costs, E', which includes the market proxy debt service, DS'. While a certain portion of users will be able to afford the costs at the stipulated rates and volumes, there may be a large num- ber of users who will not, and cross-subsidization may not be practical or may be too burdensome. Adjustment of the rev- enue for the means-tested revenue constraint shows that it is feasible to raise annual revenues only to equal R*. The amount of debt service that is affordable under the project- ed performance of the project and the affordability revenue constraints is calculated as affordable annual revenue minus (Box continues on the following page.) 196 Subnational Capital Markets in Developing Countries Box 12.6. (continued) operating costs (R* ­ O&M = DS*). The constrained value can then be used to determine the amount of grant that is needed to make the overall or "blended" annual cost of the project af- fordable. Amount of the capital grant. Assuming that the affordable mar- gin of debt is borrowed on market-proxy terms, the ratio of af- fordable debt service to market-based debt service (DS*/DS') also yields the ratio of affordable debt service in the project to debt services that would be required at full market rates. Ac- cordingly, the ratio of D*/D' (where D' = K, assuming that the project were to be fully debt financed) represents the ratio of af- fordable debt to the entire project cost, which is equal to the ra- tio of affordable debt service to the market-based debt service. Thus, if D* = 0.5 of D', the capital grant will need to equal half of the initial project costs. The capital grant is calculated leaving the interest rate, maturity, and debt service structure un- changed. In other words, the debt borrowed at the margin is on market terms (with the notable exception that the maturity is longer than normally obtainable). The affordable annual debt service (and hence overall annual revenue) is achieved by buying down the capital cost of the project through a grant. The capital grant to fill the gap equals the full capital cost minus the amount of debt that can be bor- rowed under the affordability criterion: G = K ­ D*. If the community can afford only Rps 116 million a year in total revenue, the project will need a capital grant that will reduce the debt service to half the market-proxy level, or a capital grant of Rps 10 billion. The facility's operating costs would be the same (Rps 100 million), but the required debt service would be Rps 16 million instead of the full market proxy amount of Rps 32 million. Designing and Implementing Credit Assistance to Subnational Governments 197 Technical and Market Analysis Subnational government projects need to be subjected to an affordability analysis. This "means testing" of project costs against reasonably available local resources helps ensure that the availability of grants does not discour- age creditworthy governments from borrowing and does not create a cul- ture of subsidy dependence that retards the development of capital mar- kets. In a variant of adverse selection (see chapter 2), projects that could be financed at least partially at market rates in commercial markets elect not to borrow because they believe they can get grants. This not only reduces the overall grant funds available to needy subnational government proj- ects, but it delays realization of projects. Grant funds should be reserved for needy projects and for projects that would become affordable to subnation- al government with a partial grant subsidy. Grants can also assist subna- tional governments to fund projects that may not be affordable in their early years but become so as they mature and as financial markets develop. Technical and affordability analysis has two phases. First, technical pa- rameters, based on engineering best practices, are needed to determine the most efficient operation at various scales and alternative processes and the associated reasonable costs for constructing the facilities. This analysis is the stuff of standard feasibility studies and yields a standardized annual cost function and the required capital stock investment and its cost.19 The second phase is the most critical in establishing the needed amount of the grant. The required capital investment is translated into a standardized annual debt service cost by applying a factor that reflects a commercial cost of capital on the assumption that the debt could be borrowed for a period of time that corresponds to the useful life of the project. Thus in addition to the facility's operating and capital cost figures, studies are needed of the likely us- age and applicable rate structure in order to project operating revenues. Facil- ities that have a high proportion of low-income users are the most likely can- didates for grants.20 The subsidies to facilities will be "means tested," and the subsidy will come from lowering (or in some cases removing completely) fu- ture debt service through a capital grant that reduces the amount to be bor- rowed. This up-front grant is suggested rather than subsidized interest rates or operating subsidies, which require ongoing administration and surveillance and tend to conceal the amount of subsidy (Varley 2001).21 To decide how large the capital grant should be, an objective measure is needed of an "acceptable burden" of user charges that may be paid annual- ly by the poorest users (residential and commercial). These constraints on the affordable charges then are converted into a constraint on the overall 198 Subnational Capital Markets in Developing Countries annual revenues that will be available to pay the operating and debt service costs. This constrained sum is then compared with annual operating costs (assuming efficient technical and economic operation) and the prototype "market-proxy" annual debt service, for a combined annual revenue re- quirement. The excess of annual revenue requirements using market proxy values over the needs-constrained revenue projections is the proportion by which the annual debt service must be reduced to qualify the project for debt financing. Where the acceptable level of annual charges is equal to or less than anticipated operating costs (excluding any debt service), it is un- likely that any part of the project should be considered for debt financing. Planning projects on a self-sustaining basis depends heavily on the abil- ity to develop skills in engineering and financial consultancies. Often the subnational government is unable to fund the study from the project's own resources. However, if it did, this might present a moral hazard problem, since any subnational government will prefer grant to loan funding. The long-term efficiency of an integrated grant-loan program might best be served by having the central government commission and pay for an ob- jective third-party analysis of need and affordability. Sometimes standard- ized "prefeasibility" analysis provides an acceptable level of analysis. Such studies are routinely done by registered engineering firms that use cost curves to estimate facility costs under varying conditions and sizes. Costs are often adjusted for local factor costs and specific items such as land. The analysis should be required for all capital grant programs seeking project financing. For determining the amount of the grant, project costs should be calculated at the annual amount of revenue that would be need- ed in the absence of the grant at some "indicative interest rate." Only after the affordability test has been applied, taking into consideration the likely amount that could be charged in tariffs or taxes, should the amount of the grant be calculated. If the independent feasibility or prefeasibility study finds that the subnational government and its enterprise can pay for a por- tion of the facility through a loan, then receipt of the grant should be con- tingent on also taking out a loan (or finding another way to pay for its share of the project). Designing and Implementing Credit Assistance to Subnational Governments 199 Notes 1. The costs of developing pioneer bond issues are considerable since they represent for public and private parties alike a heavy investment in learning skills, developing documentation, and charting new procedures. In the Philippines, these costs for four relatively small bond issues ranged from 4 to 5 percent of the total issue proceeds. Bond issues are very much subject to economies of scale since the novelty and complexity of a deal may have little to do with the size of the issue (see Financial Executives In- stitute of the Philippines). 2. For example, the Asian Development Bank advocates the use of pub- lic-private financing vehicles. See Asian Development Bank Commercial Co- financing and Guarantees (1999) and ADB, Office of Co-financing IED Semi- nar on Commercial Co-financing and Guarantees (12 May 1999). Similar structures are used by other international agencies, including the World Bank, USAID in its Development Credit Assistance program, and by various state governments in the United States. 3. It should be recognized that the credit line in this case amounts to a letter of credit or stand-by loan facility. The donor then looks primarily to fee income from the credit line, not the actual exercise of a loan. Any loan would be at commercial rates set high enough to discourage use of the facili- ty except in emergency. By its very presence, the facility is intended to lend confidence to the market and obviate its use. Furthermore, having the im- primatur of a highly rated bank and its surveillance of the arrangements cre- ates a halo effect in ensuring the markets of the facility's prudent operation. 4. It is useful to note that some roles can be accomplished under existing domestic market conditions, but others may be more realistic with the Bank's employing credit line assistance. 5. If the aid is to be provided anyway, making it pledgeable and inter- ceptable does not add to the cost. Any administrative costs could be borne by the borrower. 6. Traditionally, rating agency analysis gave intercepts of intergovern- mental transfers only modest credit-enhancing power in the United States. However, the power of an intercept is substantially increased if the flow goes through a trustee-administered "lock box" arrangement in which debt holders have first access to the revenue. This provision, coupled with the historical record of intergovernmental payments, led to the intercept gain- ing greatly in stature as an enhancement device. It is almost universally used for local school financing in the United States. 200 Subnational Capital Markets in Developing Countries 7. This has been the experience with pools and bond banks in the Unit- ed States, which are usually run by state entities. Some private banks and investment firms have formed pools (mutual funds) as well. 8. In the Philippines, one of the major government finance institutions, the Land Bank of the Philippines, for example, had high nonperforming loan rates for commercial loans (17 percent) and agricultural sector loans (34 percent) as of 2000. The nonperforming loan rate for local govern- ments, by contrast, was virtually zero. 9. The Philippine securities and exchange authority declared that local government securities are "exempt" entities for purposes of registration but that securitization of private sector loans is subject to special registration procedures that can make securitization a cumbersome and lengthy process, with tax implications as well. 10. Because they aggregate small issues into one large issue, bond banks can provide economies of scale, but that process reduces the amount of business available to regional dealers and banks. On the other hand, large money center dealers may support the creation of bond banks if they think they will get the underwriting business. The money market dealers have lit- tle political influence, however, compared with local investment firms. 11. For example, loans from the European Bank for Reconstruction and Development are frequently tied to grants that reduce the effective interest costs to very low levels. This is advantageous to the few subnational gov- ernments that get the financing but not to the governments that press ahead for loans that do not fit into the donor's particular game plan (Noel 2000). 12. The pricing can be based on a formula relating to the reference rate, such as short-term government securities. A problem with that approach is that the rate may go out of touch with the market if the government refus- es to accept bids for its notes. Another approach is having a repricing agent set the rate at whatever level it takes to sell the bonds. If a buyer cannot be found, the repricing agent puts the bonds to the liquidity facility, which lends the money to pay off the investor that is cashing in the security. 13. Puts may be at any prestated value and a put at par or a slight dis- count is commonly used. For example, a put at a discount is one way to discourage puts from being exercised too often. 14. The loan rate from the liquidity facility is usually set at a market in- dex plus several points. There is also a fee for making the facility available. 15. Qualifying obligation might be defined to be bonds sold for infra- structure purposes that are timely in payments and that meet certain dis- Designing and Implementing Credit Assistance to Subnational Governments 201 closure and credit criteria. An important by-product of the liquidity facility is that it can help generate demand of disclosure and for credit ratings. 16. For example, there might be substantial investor interest in fixed- rate investments of a medium maturity (five years). A 15-year bond with a one-time put at year five would be attractive, and the annual debt service much lower. To work, the liquidity facility must be backed by a very high- grade credit so that there is no doubt that the facility will be there to oper- ate. It is for this backing that a donor stand-by loan could be very effective. It is difficult to see how the systemic risk would be any greater than with a direct loan made by the donor. Also, the private sector would be stakehold- ers, unlike the direct loan scenario. 17. Analysts looking at Indonesia have argued that the availability of grants can be a significant disadvantage in starting a credit market culture (see Smoke 1999, pp. 1561­85). On grants undercutting loans as a problem in credit market development (see Weitz 2001, p. 5). Lewis (2002) encour- ages the use of market-proxy loan rates on on-lent donor funds in order to help develop private market access by local governments and discusses the need to blend loans and grants, with the size of the grants conditioned by national priorities, benefit spillovers, and the fiscal capacity of the local governments. A recent World Bank (2002a) Project Concept Document states approvingly that it appears that capital grants for the Specific Pur- pose Grant Fund (Dana Alokasi Khusus, or DAK) would depend on the in- come of governments and the nature of the project, with wealthier govern- ments eligible for only limited grants since they qualify for commercial borrowing on most investments. The report also looks at improved integra- tion of municipal credit with the capital market, "including closer to mar- ket determined rates" (p. 22). 18. To the extent that certain projects might be considered national public goods, they might be candidates for a national subsidy irrespective of local resources. These are points of judgment and national policy, but the initial assumption is that most projects will have large components of local benefit and that these benefits should be weighed against local re- sources to pay for them. 19. These technical studies often result in "cost curve" studies that pro- vide a baseline for the costing of services and facilities under specified con- ditions. Deviations in individual projects are obviously to be expected, but there is a baseline from which to start. 20. Note that there may be a good deal of cross-subsidy at the local level as richer users subsidize poorer ones. The idea is that there are limits on 202 Subnational Capital Markets in Developing Countries how much cross-subsidy can occur in a locality without driving out the richer ratepayers and that in some localities there will be too few rich users to offset the costs of serving the poor. 21. Varley (2001, p. 5) argues that subsidized rates and other soft terms lead to buildups of hidden liabilities and crowd out private sector suppliers of credit. Part V Policy Guidelines John Petersen and Mila Freire Chapter 13 Concluding Observations and Policy Guides Subnational government borrowing is not an end in itself. Ideally, it should be used to obtain long-term capital for expenditures that provide benefits that stretch into the future. Repaying debts represents the fulfilling of an intergenerational contract obligating those who benefit from improve- ments to pay their share of costs over time. Subnational governments are the legitimate parties to effectuate the obligation and the agents to see that its terms are fulfilled. Successfully incurring and paying off debt--raising funds in capital mar- kets, employing the funds in useful improvements, and repaying the debt according to the contract--is an affirmation that the subnational govern- ment is capable of planning for the future and fulfilling its obligations. Suc- cessful debt transactions are both products of financial prudence and fore- sight and installments toward financial independence (DeAngelis and Dunn 2002). That said, the gap between the ideal and the real in subnational govern- ment borrowing in private financial markets is great. Subnational govern- ments, as junior and often freshly minted government units, must find ways to enter financial markets that are themselves young and troubled in legal and economic environments that are often in transition. Credit market access has been approached from various angles: the needs of potential borrowers, the organization and regulation of the securi- ties market, likely investor groups and their regulation, the need for infor- mation to analyze credit, and the rating and private insuring of securities. This book examines, in particular, the tools that senior governments and donors might choose in developing markets, looking at forms of credit as- sistance and methods by which higher level governments monitor and, when necessary, intervene in the affairs of subnational governments. 205 206 Subnational Capital Markets in Developing Countries While it is clear that different structures of government and levels of credit market development affect the particular circumstances of each country, the following observations on policies and practices can serve as a point of departure in appraising a country's willingness and readiness to promote markets for subnational government securities. They also can stimulate debate on existing markets and on how access to them might be improved and new markets for subnational securities might best be engen- dered. Security Pledges, Instruments, and Methods of Sale: From What Sources Should Subnational Debt Be Paid, What Forms Should It Take, and How Should It Be Sold? Determining the appropriate scope and pace of subnational government borrowing and the forms it should take has presented problems for nation- al governments, financial markets, and subnational government borrowers. Over-regulation and encrustations of out-of-date, ill-defined, and conflict- ing laws have caused problems. Subnational credit access has often been an afterthought, in terms of both fiscal powers and financial market develop- ment. Overall, generic laws with broad formulations of policies and simple parameters based on easily obtainable and objective criteria are better than specific procedures that must be followed with respect to borrowing. Governing laws should make clear the legal status and remedies available to investors in subnational government obligations. This is frequently not the case, especially where subnational government obligations once carried explicit or implicit sovereign guarantees. The ultimate security and the en- forcement process for creditors should be explicit and easy to call on. Flexi- bility is important in setting the boundaries of prudential behavior. Parties to debt transactions should be able to design security provisions to meet specific needs and circumstances, as well as general requirements. Essential services, for example, can be defined and minimum service levels protected in the case of assets and intergovernmental transfers used for pledges. In addition to general obligation debt (supported by general revenue), lo- cal jurisdictions should be able to offer limited obligation security arrange- ments (revenue bonds) that do not involve a pledge of general revenues. Subnational governments should be able to enter into tariff setting and oth- er covenants for limited obligation debt. If higher level governments retain ratemaking approval, provisions should be made for prior approval of rate adjustments or for some indemnification against default where the subna- Concluding Observations and Policy Guides 207 tional government lacks the ability to adjust rates. The issues involved in rate setting and minimum service levels are thorny in low-income countries, but full costs need to be identified and any subsidies made explicit. Subnational governments should generally be able to assign revenue. This includes having the ability to pledge intercepts of intergovernmental revenue transfers. However, certain limitations on such pledges make sense. An ex- ception can be made for revenues that are necessary to provide minimum essential services. This could be achieved through a regulation specifying a maximum share of transfers that may be pledged to debt service pay- ments.1 So long as intergovernmental transfers constitute a large propor- tion of local revenues, as is usually the case, any prohibition against pledg- ing funds from these sources effectively foreshortens the fiscal planning horizons of subnational governments. Subnational governments should have the ability to create or to join with others in creating special service districts to address service needs relat- ed to specific areas or activities. Where government jurisdictions do not correspond with the rational service area, subnational governments should be strongly encouraged to cooperate. Applying the benefit principle, they should have revenue powers that allow them to capture a share of the val- ue created by their activities and investments made within or on behalf of those districts. In both developing and developed countries electorates fre- quently are more supportive of taxes and charges that are directly related to specific physical improvements and service betterments. The financial marketplace should be free to decide on the types of in- struments and associated payment mechanisms to employ. Unless there are compelling reasons to place restrictions on all borrowers, there is no basis for treating subnational governments differently than other borrowers so long as there is full disclosure and competitive norms are met. However, if there is no effective competition in financial markets (including a reason- able basis of shared knowledge by borrowers and lenders), then more over- sight is likely to be needed. At a minimum, strict rules are required on pub- lic notice and disclosure of proposed transactions. Wherever possible, it is best to introduce competition into financial markets. As a first step, competition can be promoted by requiring subna- tional governments to use formal solicitation and bidding procedures for bank- ing services, underwriters, advisers, investment services, and other profes- sional specialties. Clearly, public and timely reporting on the terms and conditions of loans and bond offerings is a necessary complement to sup- porting a competitive regime. Even where financial markets are not fully 208 Subnational Capital Markets in Developing Countries developed and effective competition is limited, the bidding process and full disclosure of transactions should help prevent monopolistic behavior and encourage entry by private lenders. Borrowing Power: How Much Can Be Borrowed, and Who Must Approve? Restrictions on borrowing powers are appropriate in many emerging mar- ket economies, where the objective is to balance local self-determination with limited experience and shallow capacity in financial markets. At the same time, the object of political devolution is to link self-determination with fiscal self-sufficiency and local accountability. If financial markets are developing in the right direction, this goal may best be accomplished by using incentives that operate through the market. The issues are ones of se- quence and scope: within what boundaries should the market decide on lending, and how should those boundaries change as the market matures? While the focus of much of this book is on long-term borrowing to meet infrastructure needs, many subnational governments now rely on short- term loans to meet cash cycle needs or to finance budget shortfalls. Such short-term borrowing, while a useful tool when responsibly limited to a sin- gle fiscal year, has often been the Achilles heel in budgetary discipline. Short-term debt should be used only to meet cash flow shortfalls in antici- pation of realistic income streams within the fiscal period. That means that under most non-emergency circumstances, short-term debt should be paid off by the close of the fiscal year. The corollary, often built into governing law, is that long-term debt should be limited to capital investment in property, plant, and equipment. It should not be used to finance operating deficits except as part of a financial emergency re- covery plan as defined by statute and regulation. Effective enforcement of such provisions on the appropriate use of debt requires regular reporting of borrowing and the purposes for which it is undertaken. The reporting should be based on a chart of accounts that is clear and analytically meaningful. Limitations on outstanding debt constitute a basic form of restriction for debt that is secured by general revenues. The restriction should be related to the tax base (where the subnational government is largely self-reliant and has control over its revenue resources) or some measure of recurring revenues (which is more typically the case where localities rely on intergov- ernmental payments). However, limitations on total debt are only a rough gauge of permissible debt burden. Where possible, the limitation should be Concluding Observations and Policy Guides 209 expressed in terms of annual or maximum debt service. For example, total debt service (principal and interest) on general obligation long-term debt should be limited to a maximum percentage of projected recurring annual revenues. For self-supporting debt issued to finance revenue-generating projects, however, the market should determine acceptable ratios of debt service coverage. Such projects will vary according to the technical and economic aspects of the improvement being financed and the security be- ing pledged. Definitions matter, and the terms used in limitations need to be precisely defined. Guarantees constitute a problem in the application of debt limitations, since the extent of guarantees is usually a missing link in debt limitation calculations. There are few quantitative restrictions on the use of guaran- tees. The common solution is to value against the debt limit that portion of guarantees that appears likely to be called on within a given fiscal period and to treat the remainder as contingent debt that is in effect self-support- ing and not counted against the debt limit. Again, the problem is less the guarantees themselves than the reporting of the guarantees. Approval of borrowing by a jurisdiction's legislative body is sufficient in most cases to obligate the unit, so long as the debt outstanding after the proposed borrowing falls within pre-stated legal parameters. Some coun- tries have provisions for citizen referendum, although this is not customary in most countries and can be expensive and disruptive. Some countries re- quire that the local budget be approved by a central government agency and that anticipated borrowings be included in the budget. Such routine budgetary review by national authorities, so long as it observes broad and general parameters, need not be overly intrusive and can help in the timely reporting of information and in the formation of macroeconomic policy. However, requiring specific prior approval of transactions by senior levels of government diminishes local flexibility and responsibility and opens the door to delay and political manipulation. Waivers of limitations in unusual cases by cognizant state or national authorities may help flexibility. Financial Market Regulation and Disclosure: What Should the Market Look Like, and How Will It Perform? In most developing and transitioning countries banks dominate the finan- cial system, but with a nascent securities market beginning to broaden the financing landscape. For the most part, financial markets still do not meet the long-term credit needs of local governments. That gap is filled by na- 210 Subnational Capital Markets in Developing Countries tional or regional government-administered on-lending programs funded by multinational donors. The observations here assume a desire to develop the securities markets as an effective alternative to a near-exclusive reliance on the banking sector or a specialized lending institution. The regulatory framework for banking and securities markets should apply to subnational government borrowers just as it does to other borrowers. That framework seeks to foster the competitive norms of market efficiency and development while preserving the integrity of the payment system and protecting investors. Generally, subnational governments should enter financial markets on an equal footing with private firms, while recognizing the distinctions that flow from their taxing and governing powers. Where banking and financial regulations favor the national government, consid- eration could be given to according the same benefits to subnational gov- ernments, along with appropriate limitations. A secondary market for securities is important to investor liquidity, but es- tablishing such markets is inherently difficult where financial markets are small. Formal listing of subnational securities on exchanges should be re- quired only where the potential size of secondary activity justifies the time and expense involved. Furthermore, the practical limitations on attracting long-term investments in local currency need to be considered. Alterna- tives might include intermediaries capable of borrowing on behalf of sub- national governments on domestic and international markets and a liquid- ity facility to back up instruments that provide built-in liquidity, such as put-option bonds. In some cases a secondary market for subnational debt can be developed as part of the over-the-counter markets that operate among banks and se- curity dealers. These are likely to be more efficient for smaller issuers, whose bonds are traded infrequently. Investor protection needs to be bal- anced with economical access for smaller issuers, which should be a funda- mental tenet of both registration and disclosure requirements. A key concern in securities market regulation is proper disclosure. Like other securities, subsovereign securities should be subject to disclosure stan- dards that require both information at the time of the initial offering and regular reporting to investors subsequently. The standards should focus on the process and generic needs. Some of the information required of govern- ments is different from that required of private firms, and disclosure re- quirements should reflect that. The actual data needs for meeting such standards may best be left to self-regulatory bodies in the market and to participants in individual transactions. Concluding Observations and Policy Guides 211 Subnational government financial information needs to be reported in clear, consistent formats and promptly after the close of the fiscal period. For debt-monitoring purposes reporting on a cash or modified accrual basis is especially useful, as are cash-flow statements. Charts of accounts should re- flect the needs of debt analysis, terms should be clearly defined, and users should be trained in their application. Audits should be independent, re- curring, and punctual. Where governments are too small to afford inde- pendent audits, borrowing is most likely to be successful thorough a mar- ket intermediary or trustee relationships that allow for funds to be sequestered to ensure payment. A central repository of financial information on government borrowers is a useful tool in promoting efficient disclosure. The repository should have current data on debt outstanding and information on security pledges and liens against real and personal property if that information is not recorded elsewhere. Markets will not thrive without information, and making infor- mation broadly available is good public policy. Credit Analysis, Credit Ratings, and Bond Insurance: How Can Risk be Measured and Mitigated? Credit analysis is a product of the credit market's need to assess risk. A fi- nancial market becomes viable only when there is a variety of competing investors and, similarly, investments with different risk and reward charac- teristics. Where there are large numbers of "passive" investors in securities that are widely held and transactions are diverse and numerous, these in- vestors generally rely on the opinions of specialists. This need is often rein- forced by various prudential requirements that are framed to ensure the in- vestment quality of institutional portfolios. Credit ratings, typically shorthand expressions of relative ranking among credits, are the leading form of institutionalized credit analysis. They assist in developing an active securities market by pooling skills to develop opin- ions. Credit ratings play an important role. They focus on credit risk (risk of payment delay or default), which then is used to help judge overall risk and reward. The use of ratings has grown steadily as markets have expanded, and they promise to play an increasing role in the regulation of banks and institutional investors. Credit ratings have the positive side benefit of ranking governments on their perceived ability and willingness to pay their debts and avoid finan- cial difficulties. The ratings are easily understood--hence their popular ap- 212 Subnational Capital Markets in Developing Countries peal--and contribute to the essential task of improving finances by provid- ing an incentive to upgrade one's rating. However, credit ratings tend to centralize and dominate credit analysis, and the precise basis for the ratings is not always clear since their calculation is based on proprietary criteria. Credibility requires accuracy, a reputation for objectivity, and freedom from influence. The demand for ratings should derive from the market it- self (even if part of that demand is a function of regulatory requirements on institutional investors), with competition among ratings companies. In domestic financial markets, it is not a good idea to have "official rating agencies" or to have the government set standards for ratings. Foreign mar- kets are likely to require internationally accepted ratings. Private sector bond insurance and other forms of credit enhancement are im- portant in developed financial markets and may have application to subsov- ereign credits. The major bond insurance companies were seeking opportu- nities in emerging market economies during the 1980s and 1990s until the financial crises of the late 1990s dulled their appetites. With the possibility for diversification of holdings within countries limited and confidence in many currencies eroded, commercial bond insurance will be slow to take hold. Furthermore, the private bond insurance industry is highly dependent on credit ratings of their portfolios. The volatility and generally lower-rung ratings given to emerging market credits create heavy capital requirements and make it difficult for companies to price their products competitively. However, domestic credit enhancement programs that have sufficient capi- tal, are market-oriented, and use insurance principles in determining appro- priate charges may hasten development. Domestic bond insurance in the Philippines is a promising "home-grown" alternative that can assist local borrowers (see Philippines case study, chapter 26). Financial Oversight, Monitoring, and Intervention: How Should the Central Authorities Monitor Subnational Financial Conditions and React to Financial Emergencies? Even in countries where the credit operations of subnational governments are largely autonomous and subject to general rules, positive action by high- er-level government has a place. This is especially so in requiring the collec- tion of timely, complete, and pertinent financial information. Without com- parable and consistent information on borrowers, financial markets operate in a cloud of uncertainty, with personal and political relationships dominat- ing decisions rather than objectively measured conditions and results. Concluding Observations and Policy Guides 213 A regular and universal reporting system for subnational governments, found- ed on an accounting system relevant to the information needs of investors and prepared by properly trained officials, is a prerequisite for market de- velopment. Most important is the ability to report direct and contingent debts outstanding, current debt service requirements, and cash funds avail- able to meet those demands as well as baseline operating expenses. While the ability to support other measures of performance and conditions is highly desirable, reliable basic data on meeting pending debt obligations and regular operating needs are indispensable to market development. Gathered data should be made public. In countries with active financial markets for subnational obligations, disclosures may suffice since self-inter- ested participants conduct the reviews and analysis. However, in most cases these data should also be subject to review by the appropriate national-level agencies to ensure that the numbers are right and to monitor the condi- tions of governments. Such monitoring need not be intrusive, but it can provide a warning if subnational governments are violating the rules or showing signs of financial weakness. Intervention by higher levels of government in a subnational government fi- nancial emergency should be comprehensive and thought out in advance. Interests need to be balanced to avoid moral hazard. The responsible par- ties should bear the risk, sharing the pain of mistakes and bad fortune with those that would have enjoyed the fruits of investments. Intervention mea- sures should provide for creditor rights, remedies, and workouts, as well as for the financial recovery or dissolution of the debtor unit. There should be added flexibility in terms of making specific pledges of security and reme- dies a matter of contract. However, it is best to have in place a statutory framework to define rights, essential services, and the procedure for re- composition of debt. Interventions should be rare, and not be used as a backdoor means for the higher-level government to bail out subnational governments and their creditors. Credit Assistance and Financial Interventions: How Can Credit Assistance Encourage the Development of Private Capital Markets? For most emerging market economies subnational government access to private financial markets is an achievable goal. However, it is not achiev- able overnight and may not be achievable for all subnational governments. Meanwhile, many emerging and transitioning countries will continue to 214 Subnational Capital Markets in Developing Countries depend on various forms of assistance to help satisfy the capital financing needs of subsovereign governments. For the most part, such aid will either be sanctioned or administered by central government agencies and, in all likelihood, will be funded by multilateral and bilateral assistance agencies. These sources of funds are not adequate to meet all needs, but the prospect of grants and loans on concessionary terms makes them attractive. The longer-term policy objective, of course, is to make subnational government borrowers self-reliant and the markets in which they borrow adequate sup- pliers of long-term capital. How best to move in that direction? Credit assistance should be provided only to the level needed to permit a subnational government to access private credit markets. This requires inte- gration of grants that might be given with the loans. Subnational govern- ments, to the extent possible, should face the costs and demands of private credit markets at the margin in meeting their financing needs. Borrowing is not appropriate in many settings. Subnational governments that are too poor and too small to borrow in the private market should not be encour- aged to borrow until their underlying financial situation makes that feasible. Direct lending, interest subsidies, guarantees, insurance, and other fi- nancial assistance should be designed to provide subnational governments with incentives to access the market on their own. Such assistance should rec- ognize differences in creditworthiness and reflect those differences in the in- terest rate and loan amount. Debt contracts, even if given on preferential terms, need to be written to commercial standards and enforced. The goal of exposing subnational government borrowers to the discipline of private markets needs to be encouraged at every step. Financial intermediaries that pool smaller loans into larger offerings, as in bond banks, can provide economies of scale and give investors opportu- nities for greater depth and liquidity in the secondary market. Bond banks and loan pools can be sponsored by either the public sector or the private sector but should function as financial institutions and be subject to credit market discipline. Other devices work through private financial markets and encourage their development. These include government-sponsored co-lending programs, credit enhancements, and liquidity support facilities. Their success will depend on private sector investors gaining confidence in the domestic market as a place to put long-term capital and in subnational government issues and loans as prudent and profitable investments. Donor lending programs need to promote innovations in assistance that advance the enlistment of pri- vate capital market sources. Skillful use of enhancements that leverage the Concluding Observations and Policy Guides 215 amount of donor aid to encourage private market participation needs to be encouraged further. Note 1. This is akin to a minimum coverage requirement often found in limit- ed obligation bonds. Part VI Country Case Studies Chapter 14 Latin America and the Caribbean Argentina A weak central government, declining economy, and uncontrolled deficits undermine the role of the credit markets in subnational finance. Rodrigo Trelles Zabala Lessons Long plagued by macroeconomic instability and political up- heavals, Argentina appeared to have set the right course with the Convertibility Plan in 1991, undertaking a number of reforms and pegging the Argentine peso to the U.S. dollar. However, the solutions to the country's many related structural problems ei- ther never took hold or proved to be the wrong ones. A declin- ing economy and growing government deficits undermined the national administration, and the ensuing devaluation and de- fault not only closed the international financial markets to Ar- gentine borrowers but also crippled the domestic markets. Subnational borrowers have played a major part in the nation's recurring financial crises. The reasons for this include the loose federal structure of government (in which the provinces, not the federal government, form the core of the system), the appetite for deficit spending, and the extensive government ownership of 219 220 Subnational Capital Markets in Developing Countries assets, including commercial banks. Argentine subnational gov- ernments have borrowed heavily from the banks, in the domestic bond market, and abroad, and by 2001 debt service was absorb- ing 25 percent of provincial spending. Devaluation and high do- mestic rates of interest (due to variable rate bank loans) pushed debt service beyond sustainable levels, precipitating widespread defaults, and in November 2002 the national government took the ultimate step of defaulting on multilateral loans. In Argentina the weak federal government has no effective con- trol over provincial borrowing, which it monitors but cannot reg- ulate. Provinces historically have owned captive banks, which they have tapped for funds. While bank loans make up only part of subnational borrowing, the high interest rates and the shutting down of the domestic market were a disaster for subnational governments. The crisis has led provinces to resume the practice of issuing interest-bearing notes that serve as a substitute curren- cy, undermining the central government's monetary policy. Argentina provides a vivid reminder that fixes at the top or in one sector cannot cure systemic problems--and that the subnational credit market may be not only a victim of a financial crisis but also a contributor to it. An elastic revenue system; heavy reliance on negotiated transfers from the central government; and a large, expensive, and protected public workforce have reduced the ability of subnational governments to manage their finances responsibly. In addition, the bottom-up political structure has not provided the public will to make the changes needed. Case studies look at the experience of three Argentine subna- tional borrowers--the province of Salta, the city of Buenos Aires, and the province of Buenos Aires--that were able to ac- cess international capital markets during the interval between crises in the mid- to late 1990s. The cases show what happened when the reforms that investors were betting on at both nation- al and subnational levels failed to materialize. The national ex- perience shows that extensive rebuilding is needed to solve the problems endemic to the Argentine political system if it is to cope with the challenges of a modern open economy. Country Case Studies: Argentina 221 Until the early 1990s Argentina experienced recurring periods of slow eco- nomic growth and high inflation, a cycle that led to the devaluation of the Argentine peso and the imposition of exchange controls. The Convertibili- ty Plan, introduced in 1991, marked a sharp change. It appeared to finally get macroeconomic management right. Based on tighter monetary policy, tax system reforms, privatization, and liberalization of the economy, the plan reduced inflation rates from more than 1,300 percent in 1990 to 0 per- cent in 1996 and pushed GDP growth from 0.1 percent in 1990 to more than 7 percent a year in 1991­94. Foreign direct investment increased five- fold, reaching US$6 billion in 1993. The Convertibility Plan fixed the ex- change rate to the U.S. dollar, established the independence of the Central Bank, and made the monetary base equal to the external reserves. In the mid-1990s it also appeared that the new government could han- dle shocks when they arose. The Mexican crisis in 1994­95 led to a sharp recession in Argentina marked by capital outflows, declining bank deposits, rising interest rates, reduced liquidity, and increased market volatility. GDP fell by 4 percent in 1995, and the unemployment rate reached a record 18.4 percent. The government responded quickly and effectively, restoring financial equilibrium with cuts in government spending, tax increases, and proactive measures to promote fiscal discipline at the provincial level. After a temporary recovery, conditions deteriorated sharply in the rest of the 1990s. Exports fell, the trade deficit grew, GDP growth plunged to 0.5 percent in 2000 (in part because of the Brazilian devaluation), and the fiscal deficit reached 3 percent of GDP. The central government's total outstand- ing debt, not including provincial debt, reached US$132 billion in June 2001, with interest payments absorbing 22 percent of the annual budget. A combination of political factors and economic mismanagement deepened the economic crisis, leading to general unrest among Argentines and to the fall of the de la Rua government at the end of 2001. Political turmoil--involving the establishment of two interim govern- ments--and the persisting economic crisis led to the devaluation of the peso, which soared to an exchange rate of more than 3 to 1 with the U.S. dollar from the initial parity it had held for 10 years. In late 2002 the gov- ernment was holding ongoing negotiations with the international finan- cial community, led by the International Monetary Fund (IMF), on how to correct the huge fiscal imbalances and put the economy back on track. In November 2002 the country defaulted on loans from the World Bank and the Inter-American Development Bank. In the midst of this crisis the feder- al government had to reduce its budget deficit, a difficult task given the 222 Subnational Capital Markets in Developing Countries constitutionally defined independence of provinces. Payments to provinces were under close scrutiny, and the government felt the need to reduce discretionary federal transfers (those that do not depend on consti- tutional provisions). This put further pressure on provincial budgets and make timely servicing of provincial bonds difficult. Intergovernmental Relations Argentina's government comprises three levels: the federal government; 24 provinces, including the city of Buenos Aires, which has the rank of province; and 1,911 local governments (municipalities) with borrowing powers. Provincial governments form the core of the country's political or- ganization. Provinces have their own constitutions and executive and leg- islative branches of government. Provincial governors and legislators are elected directly to four-year terms. The municipalities are dependent on the provinces, which dictate their organization and taxing powers and, in some cases (such as the province of Chubut), have delegated taxing powers to municipalities. According to the Constitution, the federal government can intervene, with the approval of the Congress, "in the territory of a province in order to guarantee the republican form of Government." As a result, the federal government is able to assume control of a province at any time and replace an elected governor with a federal appointee. Subnational Revenues: The Coparticipation Scheme Argentine provinces have three major sources of direct own-source rev- enues: the sales tax, the property tax, and the vehicle registration tax. The sales tax is the most important, accounting for about 60 percent of total di- rect revenue. Indirect revenues come in the form of federal transfers: two unconditional and 10 conditional transfers. Unconditional transfers, repre- senting 70 percent of the total, include general treasury support and the co- participation revenue, which is the cornerstone of subnational finance in Argentina. The gross coparticipation1 transfer accounts for 90 percent of federal transfers to provinces and 52 percent of provincial revenues. The copartici- pation law mandates that 89 percent of revenues from the federal value added tax, 64 percent of income tax revenues (after a fixed reduction), and 50 percent of a variety of other revenues go into the gross coparticipation fund. Of this, some 15 percent is retained at the federal level to finance the social security system, and another 546 million Argentine pesos (Arg$) a Country Case Studies: Argentina 223 year is allocated to a fiscal imbalance compensation fund distributed to the provinces under a separate formula. The balance, the net coparticipation fund, is shared among the federal government (42.3 percent), the provinces (54.7 percent), and special emergency and equalization funds (3 percent) (figure 14.1). The distribution of the coparticipation revenues among provinces was set in 1988 based on fixed percentages reflecting each province's share of total spending at the time: 43.7 percent for the more developed provinces (such as Buenos Aires, Santa Fe, Mendoza, Córdoba, and the city of Buenos Aires); 19.1 percent for the intermediate provinces; 27.3 percent for the low-density provinces; and 9.9 percent for the less developed provinces. As a result of subsequent adjustments to this formula, the current distribution of resources among provinces is based on arbitrary criteria emerging from bilateral negotiations between each province and the federal government. Provincial governments also have coparticipation schemes, for transfer- ring revenues to their municipalities (three provinces, Jujuy, La Rioja, and San Juan, have no coparticipation system). Unlike the national coparticipa- tion scheme, the provincial systems allocate payments to municipalities on Special purpose funds Federal 3% government 42% Provinces 55% Source: World Bank staff estimates based on Argentine Ministry of Economy data. Figure 14.1. Distribution of Shareable Taxes under the Coparticipation Scheme, Argentina 224 Subnational Capital Markets in Developing Countries the basis of objective indicators such as area, population, other municipal revenues, and similar factors. Expenditure Responsibilities Under the Constitution the provinces have jurisdiction over education, municipal institutions, provincial police, provincial courts, and other mat- ters of purely provincial or local concern (table 14.1). The federal govern- ment has jurisdiction only over the areas explicitly assigned by the Consti- tution: customs, national defense, foreign relations, issuance of currency, federal public debt and property, regulation of shipping and ports, regula- tion of banks and banking activity, and regulation of international and in- terprovincial trade and commerce. Responsibility for the remaining public services is shared among the three levels of government. There has been extensive devolution of public services to the provincial level in Argentina. The share of the federal government in public sector spending fell from more than 70 percent in 1986 to less than 55 percent in the 1990s. Provinces and municipalities are responsible for the other 45 percent. Since the provinces' direct revenues account for only 18 percent of their total revenue, they depend heavily on federal transfers. Recently provinces have used proceeds from privatizations to reduce deficits and cover capital spending. Table 14.1. Allocation of Responsibilities among Levels of Government, Argentina Federal and Provincial and Federal provincial municipal Municipal government governments governments governments Defense Higher education Basic education Markets Foreign affairs Preventive health Health care Cemeteries Currency and banking Economic Water and sewerage Solid waste collection regulations development Regional and local and disposal Public debt Justice and roads Local streets and Interprovincial security Land use drainage transport Housing Fire control Parks Trade regulation Passenger and Mail and telex cargo terminals Gas and electricity Source: World Bank. Country Case Studies: Argentina 225 Regulatory Framework for Subnational Debt There are no national regulations on the ability of subsovereign entities in Ar- gentina to raise debt. Under the 1991 Convertibility Plan, however, the provinces were prevented from rolling over existing borrowings from local banks and had limited access to provincial banks, their traditional source of financing. The Constitution allows each province jurisdiction over its borrowing. Approval procedures vary among provinces, but most provinces need a fa- vorable opinion from their controller institution (general accountant office or general prosecutor office, for example). These procedures establish that debt should not finance current expenditures and that the debt stock can- not exceed a certain share of annual revenue, a limit that usually ranges from 20 percent to 25 percent.2 For municipal borrowing, authorization is required from the municipal council and, in some cases, from provincial fi- nancial authorities. Foreign currency debt requires the approval of the Ministry of Economy under Resolution 1075/93 of the ministry. Banks are prohibited by the Central Bank from lending to subnational governments in either foreign or local currency and from underwriting provincial bonds (Central Bank Rule A 3054) unless the Ministry of Economy authorizes the transaction or the bond issue on an exceptional basis. Another important rule is Resolution 571/95 of the Ministry of Econo- my, which sets the criteria for lenders to subnational borrowers. Among these, the most notable are experience in local or international subnational debt markets, a sound financial position, and "good" loan terms (interest rate, maturity, amount, interest payments, amortization payments, and up- front fees). In cooperation with the largest Argentine bond custody compa- ny and the major stock exchanges, the Ministry of Economy has developed ways to better monitor provincial bonds. The bottom line is that the Min- istry of Economy can monitor, but not control, subnational borrowers. Subnational Indebtedness Argentina's provincial debt reached US$29.4 billion (100 percent of consoli- dated provincial revenue) at the end of 2001, while the consolidated provin- cial fiscal deficit rose to US$6.5 billion (2.4 percent of GDP).3 Few provinces have made an effort to cut spending, with 60 percent of expenditures in 2001 going to salaries and interest payments. All provinces ran a fiscal deficit in 2001. Figure 14.2 shows the relative position of provinces based on their operating deficit and accumulated debt as a share of their total revenue. 226 Subnational Capital Markets in Developing Countries The provinces have pursued different debt strategies. The province of Buenos Aires accessed the bond market in 2001, issuing four bonds for a to- tal of US$737 million. In the second half of the year the bond market was closed, and the province had to issue compulsory money bonds (Pata- cones) to pay salaries, contractors, and suppliers (see section below on compulsory money bonds). Córdoba tried to privatize its provincial bank and its electricity company to pay short-term commercial bank loans with bullet amortizations. Because of the high country risk, the privatizations never took place. Some provinces--such as La Pampa, San Luis, and the city of Buenos Aires, which had run fiscal surpluses in previous years-- faced a sharp fall in revenues and had to fund their fiscal deficits in an ad- verse financial environment. The situation is complicated, especially since the peso devaluation in January 2002. Still, the situation of provinces has improved as a result of a federal rescue through a debt swap. Formosa faces the worst situation, and Debt as percentage of total revenues 250 Formosa 200 Chaco Río Negro Corrientes Jujuy Buenos Aires 150 Tierra del Fuego Córdoba 100 Average Entre Ríos Chubut 50 Santa Fe City of Buenos La Pampa Aires San Luis 0 0 5 10 15 20 25 30 35 40 45 Fiscal deficit as percentage of total revenues Note: The size of the bubbles represents the relative size of provincial revenue. Source: World Bank staff estimates based on Argentine Ministry of Economy data. Figure 14.2. Relative Fiscal and Debt Situations of Provinces, Argentina, 2001 Country Case Studies: Argentina 227 the provinces of Chaco, Córdoba, Buenos Aires, and Jujuy also confront large problems. During the first half of 2001 many provinces tapped the bond and banking markets, but during the second half financial conditions tightened and most provinces returned to issuing money bonds. Bonds account for the largest share of provincial indebtedness, which totaled US$29.4 billion at the end of 2001 (figure 14.3). However, the debt with banks is the most expensive because it is linked to the average rate for certificates of deposit (as published by the Argentine Central Bank) plus a rate spread or adjusted by a rate multiplier. These floating rates are recalcu- lated every month. During the second half of 2001, when Argentine sover- eign risk increased dramatically, some provinces faced real annual interest rates of 45 percent. More than 30 banks have made loans to provinces, but four banks--Banco de la Nación Argentina, Banco de Galicia, Banco Frances, and Banco Rio--clearly dominate the market, with almost 60 per- cent of total bank lending to provincial governments. In 2000 the federal government implemented a voluntary refinancing program for provinces through the Provincial Development Fund. The Bonds 38% Banks 32% Provincial Development Fund (refinancing program) 10% Provincial Development Multilateral Fund (bank organizations Consolidated privatization 11% debt program) Others 3% 5% 1% Note: The figure excludes short-term debt with suppliers and employees. Source: World Bank staff estimates based on Argentine Ministry of Economy data. Figure 14.3. Provincial Indebtedness by Type of Debt or Lender, Argentina, December 2001 228 Subnational Capital Markets in Developing Countries nine participating provinces (Catamarca, Chaco, Chubut, Formosa, Jujuy, Neuquén, Río Negro, Tierra del Fuego, and Tucumán) made commitments to reduce their 1999 deficits by 20 percent, to borrow no money other than that provided by the Provincial Development Fund, to implement certain structural reforms, and to keep increases in their debt stock to no more than their 2000 fiscal deficit. In return, the fund committed to finance the provinces' 2000 fiscal deficits and the rollover of debt principal payments. The program was repeated in 2001, and two more provinces (Misiones and San Juan) joined the scheme. Both programs were unsuccessful, for two main reasons: the federal government lacked the enforcement capacity to ensure that the provinces met their targets, and politically negotiated waivers were given. Although a few provinces met their fiscal targets, most did not because there was no system to reward those that did and punish those that did not. Multilateral lenders are also important sources of credit, accounting for 11 percent of provincial indebtedness. Until the recent devaluation of the Argentine peso, this type of debt had the lowest cost and longest maturity. However, devaluation greatly increased the cost of servicing this debt in do- mestic currency. Under an agreement between the federal government and the provinces signed on 27 February 2002, the federal government will pro- vide some kind of hedging to help provinces meet the cost of this debt. Al- though some municipalities have borrowed indirectly from multilateral or- ganizations, no province has indicated whether it will help its municipalities with such debt. The debt under the Provincial Development Fund's bank privatization program4 has a long term and low cost (7.6 percent annual fixed interest rate) because the funds were provided by multilateral organizations (the World Bank and the Inter-American Development Bank). Some 17 of the 24 provinces have been involved in programs to privatize provincial banks. The largest provincial banks (Banco de la Provincia de Buenos Aires, Banco Ciu- dad de Buenos Aires, and Banco de Córdoba) were not privatized, and other banks returned to provincial ownership because of the poor performance of some private managers and the nontransparent process of privatization. Debt Service In 2001, before the most recent debt swap, debt service payments absorbed more than 25 percent of the operating revenue of provinces. More than 45 percent of the debt service payments went to commercial banks, often for short-term loans subject to refinancing risk. There was no possibility of refi- Country Case Studies: Argentina 229 nancing such loans because of the run on deposits during the second half of 2001. In addition, banks that were active players in the provincial lending market were also active players in the sovereign bond market, a situation that complicated the refinancing of provincial debt. The average life of provincial debts at the end of 2001 was six years, but some provinces had to face impor- tant due dates with no possibility of refinancing those payments. Thus at the end of 2001, as the federal government worked out a swap for sovereign debt, most of the provinces followed suit in a provincial debt swap. Provincial Debt Swap In establishing the criteria for the debt swap, the federal government ex- tended eligibility only to bank loans, provincial bonds denominated in Ar- gentine pesos and U.S. dollars, and provincial debt with the Provincial De- velopment Fund. Bonds and bank loans would be exchanged for loans issued by the Provincial Development Fund with a federal government guarantee. In exchange for this better guarantee, creditors agreed to extend the maturities of their loans and bonds by three years, established a three- year grace period for principal payments, and lowered interest rates (70% of the original interest rate with a maximum of 7% for fixed interest rates and a maximum of London interbank offered rate, or LIBOR, plus 3 percent for floating rates). The transaction involved 18 provinces and US$18 billion. By 14 Decem- ber 2001, the last day on which creditors could enter the provincial swap, more than 450 bank loans and 70 provincial bond had entered the swap. International credit rating agencies (Fitch Ratings, Standard & Poor's, and Moody's) considered the debt swap a "selective default" because it involved a reduction in net present value for creditors. As a result, some borrowers (such as the city of Buenos Aires) did not enter the swap. On 19 November 2001 Fitch Ratings published a press release explaining its concerns about the debt swap stemming from the change in terms and conditions of bonds and loans and the reduction in net present value of the debt exchanged. Impact of the Devaluation Among the first economic actions by the new government was to devalue the Argentine peso and establish a new parity for exchanging dollar-de- nominated debt for peso-denominated debt. Parity was set at Arg$1.4 to US$1. Some provincial debts were excluded from the exchange parity be- cause they were incurred under foreign laws (including all multilateral loans and some provincial bonds). 230 Subnational Capital Markets in Developing Countries After the devaluation, provincial debt instantly increased (figure 14.4). The debt stock, which had been Arg$29.4 billion at the end of December 2001, rose to Arg$62 billion by June 2002. Most provinces could no longer afford to service their debt. For international debt, ultimate responsibility rests with the federal government, which acts as guarantor. Indeed, before the present financial crisis the central government had implicitly bailed out some provinces by making their payments to multilateral lenders and then intercepting coparticipation revenues to cover the debt service costs. Collateral for Subnational Borrowing Two main types of collateral back provincial loans and bonds: coparticipa- tion revenues and hydrocarbon royalties. Most subnational borrowing is backed by pledged coparticipation revenues. Coparticipation Revenues. There are two basic mechanisms for collateraliz- ing a borrowing with coparticipation payments (figure 14.5). The first, and the more common and safer of the two, is the intercept at the source of the Millions of Argentine pesos 28,000 24,000 Dec 2001 (before devaluation) June 2002 (after devaluation) 20,000 16,000 12,000 8,000 4,000 0 Consolidated Provincial Provincial Banks Bonds Multilateral debt Development Development organizations Fund (bank Fund privatization (refinancing program) program) Source: World Bank staff estimates based on Argentine Ministry of Economy data. Figure 14.4. Impact of the Devaluation on Provincial Debt, Argentina Country Case Studies: Argentina 231 disbursements: Banco de la Nación Argentina, the commercial bank of the federal government. The second is the intercept at the provincial bank (or the financial agent of the province) that receives the coparticipation rev- enues from Banco de la Nación Argentina. These mechanisms have been tested as a result of the recent Argentine default, and some interesting differences have appeared. At the first level of interception every bond with a trustee has been honored. At the second level, however, behavior has differed depending on whether the provincial bank had been privatized, and moral hazard problems have arisen. Priva- tized provincial banks did not follow provincial instructions to default on bonds for which those banks served as trustee. In contrast, provincial banks that had not been privatized followed provincial instructions to default on bonds for which they were the trustee. The province of Chaco issued three bonds for which the provincially owned bank (Banco del Chaco) acted as trustee. When hard times came in 2001, the provincial government issued two decrees (1845/01 and 1869/01) unilaterally deferring payments on the bonds and ordering its bank to return to the province the amount collected in the trust escrow ac- counts.5 Contrast the experience of the province of Río Negro, which had established a trust with its financial agent, Banco Patagonia (its former provincial bank). In January 2002 the province postponed principal and in- Taxpayers Provincial Province (1) bank (1) Banks that receive Second tax payments level of Provincial Province (2) interceptions First bank (2) (at provincial Central level of level) Bank interceptions (at source) Provincial Province (n) bank (n) Banco de la Nación Argentina Source: World Bank staff estimates based on Argentine Ministry of Economy data. Figure 14.5. Disbursement of Coparticipation Revenues, Argentina 232 Subnational Capital Markets in Developing Countries terest payments on all its bonds except those for which Banco Patagonia acted as trustee, because the bank would not follow provincial instructions to default on the bonds covered by the trust. In 2001 some commercial banks proposed that the federal government intercept coparticipation revenues at the level of the Central Bank. Clearly, they perceived the Central Bank as having greater independence than Ban- co de la Nación Argentina. The federal government rejected the proposal. Municipal governments receive their share of coparticipation revenues through the financial agent of their province. Two interesting cases are the provinces of Buenos Aires and Mendoza. Because these provinces allow their municipalities to pledge their share of coparticipation revenues at the first level of disbursement (Banco de la Nación Argentina), lenders see these mu- nicipalities as more secure, enabling them to reduce their borrowing costs. Hydrocarbon Royalties. Four provinces have issued bonds backed by hy- drocarbon royalties (oil and gas) as collateral (Mendoza, Neuquén, Salta, and Tierra del Fuego). Although transactions backed by hydrocarbon royal- ties are much more complicated to structure than those backed by copartic- ipation revenues, all the bond issues were successful because investors per- ceive this type of collateral as the safest.6 Hydrocarbon royalties back the most successful Argentine provincial bond issue, the Salta Hydrocarbon Royalty Trust. One of the main advantages of hydrocarbon royalties is that concession- aires pay the royalties to the provinces through private local banks (includ- ing offshore banks), avoiding federal and provincial government interfer- ence. During the financial crisis affecting provinces in the second half of 2001 and the first half of 2002, there were no defaults on loans and bonds backed by hydrocarbon royalties. Still, the use of hydrocarbon royalties as collateral is rare, mainly be- cause only 10 provinces receive such royalties. Neuquén receives the largest amount, more than US$400 million in 2001. Experience with Subnational Bonds Argentine provincial bonds are of two types: those known as compulsory bonds, for which the investor must accept the terms and conditions of- fered, and those issued by conventionally accessing capital markets. At the end of 2001 more than 135 provincial bond issues were outstanding, with a total value of US$11.4 billion, and bond issues in international and domestic capital markets accounted for 40 percent of the bond debt (figure 14.6). Country Case Studies: Argentina 233 Money bonds 15% International capital markets Other 31% compulsory 45% Local capital market 9% Source: World Bank staff estimates based on Argentine Ministry of Economy data. Figure 14.6. Provincial Bond Debt Outstanding by Type, Argentina, End of 2001 Few local governments have floated bond issues in the capital market; most of the municipal bonds that have been issued have been compulsory bonds. At the end of 2001 the municipal bond debt outstanding reached US$110.5 million, and no municipality had accessed the international market. Compulsory Money Bonds. In 2001, facing revenue shortages, many provinces began to issue bearer bonds, reviving an old scheme of using bonds to pay salaries and other expenses.7 Money bonds are printed at the same size as Argentine pesos--indeed, they look like currency. Money bonds can be used to pay provincial taxes and are commonly accepted as money at their face value. The most well-known provincial money bond is the Patacon, issued by the Buenos Aires provincial treasury. Patacones are the only money bonds that can be used to pay federal taxes, and federal transfers to the province of Buenos Aires are made in Patacones. Most mon- ey bonds are short-term notes that pay a fixed interest rate and are backed by the full faith of provincial treasuries. This backing means little, however, because most provincial treasuries have defaulted on their bonds. 234 Subnational Capital Markets in Developing Countries Provinces such as Tucumán never gave up using money bonds, while provinces that have done better in managing debt and accessing capital markets in the past decade, such as Buenos Aires, only recently became ac- tive issuers of money bonds. Buenos Aires has been the largest issuer: at the end of 2001 its debt outstanding in money bonds reached almost Arg$1 billion. By comparison, the debt outstanding of all provinces in this type of bonds on the same date was Arg$1.8 billion. The Provincial Development Fund has also issued money bonds (Letra de Cancelación de Obligaciones Provinciales) to pay coparticipation revenues. In all, some Arg$4.3 billion in money bonds were outstanding at the end of 2001. Local revenues continued to decline in the first half of 2002, and at the end of June the stock of money bonds (excluding those issued by the Provincial Development Fund) reached Arg$4.1 billion and Arg$ 7.4 billion including those issued by the Provincial Development Fund. Some of these bonds, such as those issued by Buenos Aires, Córdoba, and the Provincial Development Fund, have wide acceptance and a liquid secondary market with low volatility. Nonetheless, these bonds adversely affect monetary pol- icy, not only because they prevent open market policies but also because they can be used to buy U.S. dollars. For this reason, under an IMF financial rescue package for Argentina, provinces would stop issuing money bonds. The IMF and the Ministry of Economy have discussed options such as buy- back programs for retiring these money bonds from the market. Other Compulsory Bonds. Other compulsory bonds typically are the result of debt consolidation related to judicial decisions, debt restructurings, and old provincial debts. Because of their compulsory nature, most of these bonds replicate the terms and conditions of consolidation bonds issued by the federal government (table 14.2). Although the compulsory bonds were not issued in the capital market, many were listed on the Buenos Aires Stock Exchange and in the Buenos Aires over-the-counter market (Mercado Abierto Electrónico). These listings helped provinces gain knowledge about bond issuance. Bonds Issued in the Capital Market. Capital markets have proved to be a good avenue for lowering the cost of funds and extending debt maturities for Argentine provinces: in the past several years 12 provinces have ac- cessed the bond market, and seven of them have reached international capital markets (table 14.3). The earliest and most active issuer is the province of Buenos Aires, which launched its first issue in 1994; the province is the second largest issuer, after the city of Buenos Aires. Because of the size of its fiscal deficits, Buenos Aires cannot finance them through Country Case Studies: Argentina 235 Table 14.2. Terms and Conditions of the Typical Consolidation Bond, Argentina Most common name BOCON (bonos de consolidación) Currency U.S. dollar or Argentine peso Interest rate 30-day U.S. dollar London interbank offered rate (LIBOR) or average interest rate for savings accounts in Ar- gentine pesos Maturity 16 years Grace period for principal 6 years Grace period for interest 6 years Interest payments Monthly, beginning in month 73; 119 payments of 0.84 percent of the principal and a final payment of 0.04 percent of principal Period of capitalization During the first 72 months Collateral None Status General obligation Source: Argentine Ministry of Economy. Table 14.3. Provincial Bond Issues in Domestic and International Capital Markets, Argentina, 1994­2001 Average Average Total amount issue interest Average (millions of (millions of rate life Year Issues Issuers U.S. dollars) U.S. dollars) (percent) (years) 1994 1 Buenos Aires 100 100 9.83 3.00 1995 4 Buenos Aires, Neuquén 283 71 10.15 3.17 1996 3 Buenos Aires, Mendoza 479 160 9.25 5.08 1997 8 City of Buenos Aires, Mendoza, Tierra del Fuego, Tucumán 1,156 144 10.64 7.43 1998 2 Buenos Aires 164 82 7.83 4.34 1999 10 Buenos Aires, Formosa, Misiones, San Juan, Santiago del Estero 946 95 12.88 5.09 2000 9 Buenos Aires, City of Buenos Aires, Chaco 1,306 145 12.05 5.96 2001 4 Buenos Aires, Salta 859 215 11.84 6.12 Note: The table excludes treasury bills. Source: World Bank based on Argentine Ministry of Economy data. the banking system. Thus it is a regular issuer in the bond market, while other provinces are opportunistic issuers. The bond issued by the province of Buenos Aires in 1994, a U.S.-dollar fixed-rate bullet bond, was the only provincial one issued that year. Four bond issues were launched in 1995: three by Buenos Aires and one by 236 Subnational Capital Markets in Developing Countries Neuquén. The Neuquén bond was the first (and still the only one) backed by coparticipation revenues and hydrocarbon royalties. It was also the first with a trust structure under the Argentine Trust Law (Law 24441). Neuquén was the first subnational issuer to use the largest Argentine custody house, Caja de Valores, as trustee. In 1996 Mendoza province issued its first bond, collat- eralized by hydrocarbon royalties. The bonds, due July 2002, were reported- ly fully repaid despite the Argentine default and devaluation, demonstrating the strength of security arrangements through offshore trusts. The province of Buenos Aires has issued bonds in the market every year except 1997, when the province achieved a fiscal surplus as a result of its privatization program. In that year, the most successful in the 1990s for bond issues, four provinces accessed the market. The market confidence prompted Mendoza to issue a bond with no collateral. (Unfortunately, the confidence proved to be misplaced.) The city of Buenos Aires launched four bonds, one denominated in Argentine pesos, two in Italian lire, and the last in U.S. dollars. All general obligation bonds, they were sold to fund an accumulated deficit. Tierra del Fuego completed a successful transaction in October 1997, offering a fixed rate bond backed by hydrocarbon royalties. Tucumán became the first province to tap international capital markets, backing its bonds with coparticipation revenues. Its program included two bonds (US$200 million each), one issued with a fixed interest rate and the other with a floating rate. In 1998 financial conditions tightened because of the Asian, Russian, and hedge fund crises. Most of the provinces refinanced their loans in the banking market, with the province of Buenos Aires the only subnational government accessing the market. The next year, 1999, was a complicated one for Argentine provinces not only because of the Brazilian devaluation but also because of presidential and gubernatorial elections. Also a very difficult year was 2001. The Argentine financial problem was the eye of the hurricane. Despite the turmoil, Salta launched the first Ar- gentine subsovereign bond rated better than the sovereign. Indeed, the bond received investment-grade ratings from the three major rating agen- cies. The bond was denominated in U.S. dollars and backed by hydrocar- bon royalties. Among these provincial bonds, 85 percent were issued with fixed interest rates and the most common currency used was the U.S. dollar. The debt in dollar-denominated bonds increased significantly after the currency devalu- ation. In addition to bonds in Argentine pesos and U.S. dollars, the province of Buenos Aires and the city of Buenos Aires issued bonds in deutsche marks, Country Case Studies: Argentina 237 yen, Swiss francs, Italian lire, and euros. Coparticipation revenues, hydrocar- bon royalties, or both were usually used as collateral. The offering process of- ten was protracted and challenging. Except for the province of Buenos Aires, subnational governments that accessed capital markets lacked experienced and permanent debt offices and reliable information. The lack of well- trained and properly prepared subnational debt offices has hampered the ex- pansion of the subnational bond market in Argentina. Every local bond issue used a trust scheme, which has proved to be a safe measure, especially during a financial crisis. While many provinces default- ed on their bonds after the sovereign default, there was no default on bonds with a trust scheme. There has been no common approach to dealing with the defaults. A few provinces have taken actions to reschedule their pay- ments, others have done nothing, and still others have deferred payments. Most bond issuing activity has been at the provincial level, with just three municipalities--Guaymallén, Bariloche, and Bahía Blanca--accessing local bond markets. Bahía Blanca is the only one with debt still outstand- ing as of June 2002. Recent Developments in Subnational Finance (Up to End-2002) Although all provinces face a deep liquidity crisis, their fiscal situation varies. The province of Buenos Aires, with the largest economy (35 percent of GDP), has accounted for two-thirds of the total provincial deficit in the past three years (1999 through 2001) on average. Even so, the relative per- formance of other provinces is much worse. While the central government has made efforts to rationalize spending, mainly by cutting salaries for civil servants, only a few provinces have followed this example. Many others have financed their imbalances by issuing money bonds. The federal government's dwindling resources and its declining ability to assist subnational governments are major concerns in the protracted crisis. After the second half of 2001, because of the dramatic decline in federal revenues, the federal government failed to transfer the minimum amounts required under the federal compromise, an agreement fixing the monthly transfers owed to the provinces.8 At the end of 2001 most of the provinces (including the city of Buenos Aires) signed a new agreement with the feder- al government allowing it to use notes to pay all past-due amounts to provinces and up to 40 percent of the amounts due after November 2001. The agreement also allowed the federal government to reduce transfers due 238 Subnational Capital Markets in Developing Countries after January 2002 by up to 13 percent of the total (as long as it makes cor- responding reductions in the federal budget) and allowed provinces to use these reductions as a credit on debt service payments to the Provincial De- velopment Fund. The persisting economic crisis was causing problems in the coparticipa- tion scheme. To address these problems, the federal government signed an agreement with provinces on 27 February 2002 aimed at sharing the costs of the crisis. The agreement includes restructuring provincial debt through the issuance of a central government bond that will be used to assume provincial debt in a new debt swap. A precondition for the debt restructur- ing is a reduction in the provincial fiscal deficit of about 60 percent. A deadline of the end of 2002 was set for passing a new coparticipation law. The law was not passed, and the system remains unchanged. To guarantee their debts, the provinces pledged 15 percent of their gross coparticipation revenues. However, the amounts to be collected from provinces will be inadequate to pay the bond launched by the federal gov- ernment, and the national Treasury will face imbalances beginning in the fourth year after the issue. The agreement also calls for the central government to provide currency hedging for provincial multilateral debt. Provincial debt issued in foreign markets will receive the same treatment as central government debt. At the end of August 2002 the federal government issued a decree (1579/02) establishing the terms and conditions for the debt-restructuring program: · The bond issuer is the Provincial Development Fund. · The interest rate is 2 percent annually. · The grace period for interest payments is 7 months, and for principal payments, 36 months. · The maturity is 16 years. · The amortization is monthly in 13 years. · The principal is to be adjusted by an index based on the consumer price index. · The issue date is 4 February 2002. An important difference between the 2001 and 2002 debt swaps is that the latter extended eligibility to debts denominated in any currency. To reach an agreement with the IMF, the federal government was required to sign an agreement with each province establishing an ordered financial Country Case Studies: Argentina 239 package. The aim was to avoid the issuance of provincial money bonds and to reduce fiscal deficits by up to 60 percent of the 2001 fiscal deficit. The provinces appear to be better situated than in previous years to achieve these goals: Some of them receive oil royalties that are settled in U.S. dollars. Be- cause of inflation, tax revenues are again growing. The interest payments for the 2002 debt swap are much lower (almost 40 percent less) than those for the 2001 debt swap; salaries, which account for 50 percent of spending, have not been increased or indexed to inflation. Nonetheless, a sound monetary policy and a real commitment from the federal government to cut spending are required to avoid hyperinflation pressures and to change the situation dramatically for both the federal government and the provinces. Key Issues for the Viability of Subnational Bonds During the 1990s Argentina addressed several reforms that had positive ef- fects on capital markets. Nevertheless, other major reforms are still pending. Among the structural problems facing Argentina, building a workable fiscal relationship between the federal government and the provinces is one of the most pressing. Structural change is needed to make the revenue- sharing mechanism simpler and more equitable. Despite a constitutional requirement for change in the structure of the coparticipation scheme, there has been a political stalemate: such changes require approval by the federal government and all the provinces, something very difficult to achieve politically. Unless additional revenues are allocated to the copartic- ipation funds, which can happen only after economic growth resumes, the issue will remain unresolved because no province is likely to agree to re- duce its share of revenues to benefit another. The federal government's need to balance its fiscal accounts in the face of falling tax revenues and lack of external financing requires cost-cutting efforts that also involve the provinces. These efforts are complicated by the institutional inconsistency in Argentina, where provinces depend on the central government for most of their revenues but have constitutionally granted economic and financial independence and thus are not subject to central government interference. An agreement with the provinces on the structure of the intergovern- mental fiscal relationship will have to wait until institutional changes are supported by a consensus on the urgency of reform. That consensus has not yet emerged. Meanwhile, provinces need to increase local tax revenue but lack a sound structure for collecting taxes and appear unable to curb 240 Subnational Capital Markets in Developing Countries tax evasion. There is continual renegotiation of the compensation that the federal government assigns to the provinces for service responsibilities that have been decentralized. In some cases this issue is holding up further de- centralization, such as for police and judicial services in the city of Buenos Aires. No agreement has been reached on how much additional funding the city should be awarded to provide these services. The default and devaluation raised important questions relating to sub- national bonds. Some provinces (such as Salta and Tierra del Fuego) are do- ing their best to avoid defaulting on their bonds, but will this effort be rec- ognized by rating agencies and investors in the future? With just a few provinces having debt management offices, what would have happened if provinces had had well-trained debt managers? With independent trusts playing a key role in avoiding provincial defaults, will these structures lead to the reconstruction of the Argentine subnational debt market? Can the provinces manage the currency risks of multilateral loans, or will the feder- al government bail out provinces again? Regardless of the answers, it is clear that Argentina needs to rebuild its bond market. The challenge is to learn from the past and improve on it. Province of Salta: A Bond Issue Backed by Hydrocarbon Royalties The province of Salta had its first public debt issue in February 2001. The bond was issued by the Salta Hydrocarbon Royalty Trust with a targeted maturity of 12 years but an actual maturity of 15 (table 14.4). It is the first asset-backed structure for an Argentine subnational issuer rated higher than the federal government. The structure includes a strong security pack- age enabling the bond to just reach international investment grade (figure 14.7). The transaction was considered very successful not only for its long maturity but also for its relatively low cost for an Argentine province at the time. However, the marketing period was long because of the financial problems Argentina experienced at the end of 2000. The Province Salta is one of Argentina's major provinces, with an area of 155,488 square kilometers and 3 percent of the country's population. At the time of the is- sue the province was managed by a strong administration, elected in 1999 for a second four-year term and with a positive record in financial manage- ment and administrative reform. Country Case Studies: Argentina 241 Table 14.4. Features of the Bond Issue by the Salta Hydrocarbon Royalty Trust Feature Details Issuer Salta Hydrocarbon Royalty Trust, a trust established in the U.S. state of Delaware Amount US$234 million Market Qualified investors in Europe and the United States Issue date 28 December 2000; offered and closed in February 2001 Issue price 100 percent Interest rate Fixed at 11.55 percent a year Interest payment period Quarterly on 28 December, March, June, and September Maturity date 28 December 2015 Expected maturity date 28 December 2012 Amortization Bullet Targeted amortization Starting 2.25 years after the issue date, with the first targeted amortization on 28 March 2003 Ranking Direct and unsubordinated Credit ratings · Moody's: Baa3 (global) · Standard and Poor's: BBB­ (global) · Fitch Ratings: BBB­ (global) Sources: Salta Hydrocarbon Royalty Trust offering circular, Moody's, Standard & Poor's, and Fitch IBCA. Argentina Concessionaires Province of Salta Argentine collection account Salta United States Hydrocarbon U.S. collection Royalty Trust account Political risk insurance policy (offshore) Liquidity reserve Investors account Source: Salta Hydrocarbon Royalty Trust offering circular. Figure 14.7. Flow of Funds for the Salta Hydrocarbon Royalty Trust Bonds 242 Subnational Capital Markets in Developing Countries Economic Performance. Following several years of robust economic growth, Salta suffered an economic downturn in 1995, paralleling the na- tional recession triggered by the Mexican crisis. Like all Argentine provinces, Salta was deeply affected by the capital flight from the banking system and the reduced availability of international liquidity. Salta has a di- versified economy by Argentine standards, but its per capita income and education levels are below the national average. Manufacturing is its main activity, accounting for almost a fourth of its production. Its chief exports are agricultural products, industrial products, and fuel and energy. Brazil has historically been Salta's most important export market, accounting for 30 percent of total exports, followed by the United States at 10.6 percent and Bolivia at 9.2 percent. Hydrocarbon production and exploration activi- ties in the province increased sharply with the deregulation of the 1990s. As a result, hydrocarbon royalties rose from US$16 million in 1991 to US$20.7 million in 1995 and US$37.5 million in 1999. Financial Performance. At the time of the bond issue the province derived current revenues from three main sources: gross coparticipation transfers (71 percent, with net coparticipation transfers accounting for 44 percent), provincial taxes (15 percent), and provincial nontax revenues (5 percent). Current spending goes primarily to personnel costs (54 percent in 2000) and transfers to municipalities (11 percent). Salta had strong revenue growth in the early 1990s, reflecting national economic trends following the implementation of the Convertibility Plan in 1991. The province has maintained a relatively small fiscal deficit com- pared with other provinces, and it achieved a fiscal surplus in 1996 as a re- sult of its privatization program and the transfer of its pension fund to the federal government. The privatization program, considered very successful, included two banks, a water supply company, and an electricity utility. Debt Profile. Salta's debt stock increased by 85 percent in 1995­2000 as a result of fiscal deficits in those years (table 14.5). The province has been re- ducing its debt with the national government, its bond debt, and its con- solidated debt while increasing its bank and multilateral debt. The growth in its commercial bank debt implies a higher cost of funding and shorter maturities. Salta's ratio of debt to economic production is worse than the average for Argentine provinces (figure 14.8). However, thanks to the tight admin- istration by its government, its ratio of debt service to operating revenue matches the national average. Moreover, the province has a slightly lower cost of funding than the average. Country Case Studies: Argentina 243 Table 14.5. Debt by Source, Salta, 1995­2001 (millions of Argentine pesos9) Source 1995 1996 1997 1998 1999 2000 2001 Banks 34.5 146.1 247.5 268.9 361.6 435.7 350.6 Multilateral lenders 8.8 31.6 21.4 74.1 105.3 105.4 147.4 Provincial Development Fund (bank privatization program) 16.7 50.0 50.9 50.9 50.9 50.9 49.3 Bonds 99.3 100.3 123.5 93.7 76.2 82.0 324.0 Other debt 375.6 234.2 104.6 83.7 92.8 90.0 74.9 Total 534.8 562.0 547.9 571.3 686.8 764.0 946.2 Note: Data are as of the end of December of each year. Source: World Bank staff estimates based on Argentine Ministry of Economy data. Average interest 30 rate 24 Provinces 20 Salta 10 Debt per capita/ income per capita Debt service/ operating revenue Debt/economic production Source: World Bank staff estimates based on Argentine Ministry of Economy data. Figure 14.8. Selected Debt Indicators, Salta and All Provinces, End-2001 (percent) 244 Subnational Capital Markets in Developing Countries Issue Development The Salta bond issue represented the first time that a subnational Argentine government was involved in a true sale of its future royalties. This made le- gal due diligence particularly important. The transaction was analyzed by Argentine and U.S. counsel, with both focusing on key aspects of the hy- drocarbon concessions (terms and conditions, validity of permits, hydro- carbon royalties), the relevant hydrocarbon laws, the province's rights to the hydrocarbon royalties and other revenues, and the validity of the col- lateral documents and arrangements. Also important was the regulatory and constitutional framework governing the province's revenue-raising powers and expenditure responsibilities. The structure of the notes was the key to their receiving the first global investment grade for an Argentine subnational bond. The notes were struc- tured as a U.S. dollar issue to tap a deep and mature market, important for a first international issue. The structure included four innovative features that had never been used before in Argentina: · The province sold its hydrocarbon royalties to a trust in a true sale under Argentine law. · The trust, established in the U.S. state of Delaware, issued the notes. · Target amortizations were scheduled to be due in 2015, but failure to make a targeted principal payment does not constitute an event of default. · Salta used a political risk insurance policy for its bond, the first Ar- gentine subnational issuer to do so. Reasons for the Issue. Like other subnational governments, the province was facing an increasingly short-term and high-cost debt structure because of the large share of its debt contracted from commercial banks. This made an international bond issue with a longer maturity and a fixed cost of fund- ing an attractive alternative. The province also viewed the issue as a good opportunity to gain credibility in international markets. The issue was structured as a single transaction, and all the funds raised were used to pre- pay commercial bank loans. Credit Rating. The Salta bond issue was the first Argentine transaction si- multaneously rated by the three major rating agencies--Fitch Ratings, Moody's, and Standard & Poor's. All ratings just reached international in- vestment grade. The main factors supporting the ratings were these: Country Case Studies: Argentina 245 · The true sale of the royalties, which mitigated the risk of provincial interference with the transaction. · The convertibility and transferability insurance policy covering 31 months of interest payments. · The reserve liquidity fund covering six months of interest payments. · The irrevocable and unconditional payment instructions delivered by the province to all the concessionaires with concessions dedicated to bond repayment. · A flexible amortization schedule, which means that failure to pay tar- geted amortizations is not an event of default. · Levels of collateralization that can sustain significant drops in oil and gas prices and lack of growth in oil and gas demand. · Proven reserves representing 29 years of gas production and 33 years of oil production at 1999 levels. Underwriting and Marketing. The syndicate acting as lead manager charged a gross fee of 2.75 percent, considered high by industry standards. The issuance took almost a year. The issue was marketed to qualified insti- tutional investors in the United States under Rule 144A and outside the United States under Regulation S. In line with common practice, road shows were the main presales marketing technique used. Interestingly, in- vestors formally requested a meeting about the issue with the Ministry of Economy, and it was the first time such a request had been made for an Ar- gentine provincial bond issue. Seven large institutional investors sub- scribed to the offer, and all attended the meeting held in Buenos Aires at the Ministry of Economy. The lead manager made presentations to the in- vestors and took them to the province to build knowledge and confidence. Key Factors Affecting the Issue Salta had little tax authority or revenue flexibility, and its expenditures were rigid. These fiscal constraints, together with the province's lack of vis- ibility in the international markets, prompted the decision to use hydrocar- bon royalties as collateral and issue the bond through a trust. The regulatory framework also played a key role in the bond issue for several reasons. Two sets of regulations--those for oil and gas--had to be taken into account. The transaction was the first involving a true sale of royalties by a subnational government, and possible changes in hydrocar- bon royalties and currency exchange transfers were being contemplated. 246 Subnational Capital Markets in Developing Countries The perception in international capital markets of Argentina's financial condition had a large influence on the offer price and coupon rate of the is- sue. In addition, after the 1997 Asian crisis and the ensuing crises in other emerging markets (Russia in 1998, Brazil in 1999), investors had become cautious about emerging market bonds and demanded larger spreads over U.S. treasury bonds. However, while the issue was adversely affected by the weak sovereign position at the time of the launch, the province was con- sidered highly competent and gained broad credibility during the road shows. There were several reasons for this. The province, which was well managed by a strong team, had implemented a series of reforms aimed at improving tax collection and controlling spending. The province had pri- vatized its bank and other provincial companies and transferred its pension fund to the national government. Despite competent management, the province still had a lot of work to do in financial reporting and disclosure. Financial statements were not pre- pared or audited in accordance with international standards, and the bud- geting process was still elementary. The insufficient and inconsistent statis- tical and financial information on the province complicated financial and economic due diligence. Given these inadequacies, the security structure of the issue was the key to its success. Recent Developments In 1999 and 2000 the effects of the Brazilian devaluation and the slow- down of the Argentine economy led to a worsening of Salta's fiscal position and increased its fiscal imbalances. Debt (excluding short-term arrears to suppliers and employees) was equal to 76 percent of provincial revenue, and debt service absorbed 20 percent of revenue. In December 2001 the province's total debt stock reached Arg$946 million. To improve its debt profile, the province decided to participate in the debt swap program promoted by the Argentine government in December 2001. However, it could include only commercial bank debt and consoli- dation bonds in this swap. On 27 February 2002 Moody's downgraded the Salta Hydrocarbon Royalty Trust from Baa3 to Caa1 because of the rede- nomination of dollar-denominated contracts between private parties at the exchange rate of 1 to 1. While most private companies found it impossible to obtain authoriza- tion from the Central Bank to transfer money overseas to pay their debts, most provinces received authorization from the Ministry of Economy to pay their international bonds. This shows that at times of deep crisis such Country Case Studies: Argentina 247 as that experienced in Argentina, the strength of the credit arrangements backing a bond and its overall security structure is as important as the po- litical will to make good on payments. City of Buenos Aires: A Debut in the International Bond Markets The city of Buenos Aires made its debut in the international bond market with a euro medium-term note program in March 1997 equivalent to US$500 million (table 14.6). The city launched four issues from April to June 1997 and a fifth and final one in July 2000. The notes could be issued in a variety of currencies, including the Argentine peso, U.S. dollar, Italian lira, and euro. The first series was issued in U.S. dollars and targeted primar- ily to the U.S. market (table 14.7). The purpose of the program was to refi- nance the city's debt stock and restructure its bank, Banco Ciudad de Buenos Aires. It was also aimed at gaining credibility and a sound reputa- tion for the city among global investors. Table 14.6. Key Features of the Bond Program of the City of Buenos Aires Feature Details Issuer City of Buenos Aires Arranger Chase Manhattan International/Chase Bank AG Dealer Chase Manhattan International Currency Various hard currencies, including the U.S. dollar, Argentine peso, Italian lira, pound sterling, Swiss franc, yen, and euro Amount Up to US$500 million equivalent in series Maturity Variable by series (up to 30 years) Issue price At par, discount, or premium over par by series Method of issue Continuous basis with syndication if needed and minimum offerings of US$10 million equivalent Interest rate Fixed, variable, or zero coupon, depending on the series Fixed rate notes Payable in arrears on agreed dates Variable rate notes Interest borne separately in each series by reference to such benchmarks as the LIBOR and London interbank bid rate (LIBID) Interest periods As agreed between issuer and dealers Zero coupon notes Bear no interest and normally issued at a discount Status Direct, unconditional, unsecured, unsubordinated ranking pari passu with all obligations of issuer Source: City of Buenos Aires offering circular. 248 Subnational Capital Markets in Developing Countries Table 14.7. Main Characteristics of the Bond Issues by the City of Buenos Aires First issue Second issue Third issue Fourth issue Fifth issue Currency U.S. dollar Italian lira Argentine peso Italian lira Euro Hedging Swap Swap Swap Amount in original currency 250 million 100 billion 150 million 69 billion 100 million Issue date 11 April 1997 23 May 1997 28 May 1997 10 June 1997 7 July 2000 Maturity date 11 April 2007 23 May 2004 28 May 2004 10 June 2005 7 July 2003 Interest Semiannual Annual Annual Annual Annual 11.25 percent 10 percent 10.5 percent 9.5 percent 9.5 percent Amortization Bullet Listing Luxembourg Stock Exchange/PORTAL Arranger Chase Manhattan International Rating Moody's: B1; Standard & Poor's: BB­ Source: City of Buenos Aires offering circulars. Each of the issues sold well, thanks to the city's good international repu- tation and low indebtedness. Despite the city's growing fiscal deficit, its debt at the time of the issue was equal to only 1.4 percent of its annual eco- nomic production. This, coupled with a targeted reform program, helped achieve reasonable ratings, which strengthened market perceptions. The notes were placed at a fairly large spread over the benchmark U.S. treasury bonds, but the city was more interested in achieving a placement well di- versified by region and investor than in minimizing costs. The Issuer The city of Buenos Aires, located at the mouth of Rio de la Plata, was founded in 1580 and has been the capital district of Argentina since 1880. Its population of 3 million represents 8.6 percent of the country's total. The city is administratively independent from the province of Buenos Aires and has no fiscal or political relationship with it. The city was granted its autonomous status (similar to that of a province) following constitutional reforms in 1994. Before these adminis- trative changes the president of Argentina appointed its mayor, and the federal government made most key decisions. The city's constitution, ap- proved in October 1996, provides for executive, legislative, and judicial branches. The city has a decentralized administration consisting of "com- munities" managed by an elected seven-member administrative board. These communities are responsible for secondary services, such as main- taining streets and parks, but have no independent revenue-raising powers. Country Case Studies: Argentina 249 Economic Performance. Besides being the federal capital and key financial center of the country, the city is a major driver of the economy, contribut- ing more than a quarter of GDP. Thanks to a strong concentration of ser- vices and industry, the city's per capita income grew by 90 percent over the past decade to reach Arg$22,400 in 2001, about three times the national average. The city was affected by the Mexican crisis of 1995 but less so than other parts of the country; its production fell by 1 percent, while national GDP declined by 4.4 percent. Production growth in the city averaged a strong 5.7 percent in 1992­98. Argentina's most recent economic crisis, which led to a contraction in GDP of 3.4 percent in 1999 and 0.5 percent in 2000, started to affect the city's finances only in 2001. Financial Performance. At the time the bond program was launched in 1997, the city of Buenos Aires derived more than 90 percent of its revenue from local taxes, mainly turnover taxes (57 percent), property taxes (16 per- cent), and motor vehicle licensing fees (9 percent). Federal transfers con- tributed only 6 percent of revenue, far less than the 50 percent typical for most other provinces. The city is a net contributor to Argentina's subna- tional system: while the federal government collects about a third of its to- tal tax revenue in the city, it gives back to the city only 1 percent of its total transfers to provinces. The city maintained a solid financial position from 1996 when it re- ceived autonomy to 2001 when it was affected by the Argentine crisis, with operational surpluses each year. The situation was sharply different before 1996. The city had large structural deficits amounting to US$1 billion over the period 1991­96. It generally funded the deficits through late payments to suppliers and short-term loans. Growing spending coupled with shrink- ing revenues led to a surge in the fiscal deficit--from US$9 million in 1995 to US$349 million (13 percent of revenue) in 1996. Debt Profile. When the bond program was launched, the city had a mod- erate level of direct debt by national and international standards, with a debt stock of US$1.16 billion, about 1.4 percent of annual economic pro- duction (table 14.8). The moderate level of debt was possible because of the substantial transfer of outstanding debts to the federal government that oc- curred when the city's new constitution was adopted. This debt was later refinanced as part of the Brady bond program, in exchange for offsetting claims against the federal authorities. At the end of 2001, before the debt swap and devaluation, the city of Buenos Aires had a strong debt position relative to the average for Ar- gentine provinces (figure 14.9). The only debt indicator on which the 250 Subnational Capital Markets in Developing Countries Table 14.8. Debt by Source, City of Buenos Aires, 1995­2001 (millions of Argentine pesos) Source 1995 1996 1997 1998 1999 2000 2001 Banks 448.9 591.1 226.1 119.0 35.5 13.8 12.8 Multilateral lenders 31.8 23.2 19.2 17.6 22.9 45.2 55.0 Bonds 0.0 0.0 450.7 450.7 450.7 545.8 545.8 Other debts 264.5 481.3 420.6 340.3 498.9 451.6 780.7 Total 745.2 1,095.6 1,116.6 927.6 1,008.0 1,056.4 1,394.3 Note: Data are as of the end of December of each year. Source: World Bank staff estimates based on Argentine Ministry of Economy data. Average 30 interest rate 24 Provinces 20 City of Buenos Aires Debt per capita / 10 income per capita Debt service/ operating revenue Debt/ economic production Source: World Bank staff estimates based on Argentine Ministry of Economy data. Figure 14.9. Selected Debt Indicators, City of Buenos Aires and All Provinces, End-2001 (percent) city's performance was close to the provincial average was the cost of funding. At the time of the bond sale the city's financial management and re- porting systems were reasonably effective by Argentine standards. Howev- Country Case Studies: Argentina 251 er, there were problems stemming from the different accounting treatment of revenues and expenditures, the incompleteness and inconsistency of some of the information, and the lack of audited financial statements. Issue Development The bond issues, used to capitalize Banco Ciudad de Buenos Aires (US$100 million) and restructure short-term obligations, reduced the city's exposure to short-term volatility in interest rates and market appetite, but they in- creased the city's exposure to currency risk.10 Moreover, the longer-term obligations require that the city accelerate reform in order to meet its liabil- ities. The city's de facto assumption of the obligations of its bank was a concern, because it could create a precedent for future bailouts. The bank had a large share of nonperforming loans, a small capital base, and narrow profitability. Moody's and Standard & Poor's, generally optimistic about the prospects of Argentina and the city, rated the bonds B1 and BB­. These ratings were a major factor in the eventual placement of the issues. Nonetheless, both rat- ing agencies expressed concerns about the city's ability to tackle fundamen- tal structural problems in revenues and expenditures and about the ineffi- ciencies of Banco Ciudad de Buenos Aires. The first issue under the program (US$250 million) sold extremely well in the market despite a rapid weakening of the benchmark U.S. treasuries that increased the spread from 330 basis points to 370. Even at the larger spread and in a tightening market, however, the issue was a resounding success. The issue was twice oversubscribed, and more than two-thirds was sold to U.S. investors. An important feature was that it attracted new mon- ey rather than investors selling out of existing portfolios. The second issue, a peso issue equivalent to US$150 million, was struck in record time, with marketing starting on a Friday and price-fixing taking place on the following Monday. The second issue had a narrower spread over the benchmark 2006 Argentine treasury bonds (95 basis points, com- pared with 140 basis points for the first issue). In line with the strategy of market diversification, the peso transaction was followed by lira issues, which also performed well. The transaction as a whole was considered highly successful. Book de- mand was high in all cases, with issues oversubscribed about twice, and all series of notes were sold out. Interest in the bonds came mainly from institutional investors, which purchased about 90 percent of the issues on average. 252 Subnational Capital Markets in Developing Countries Key Factors Affecting the Issues Macroeconomic conditions in the country and the city played a major part in the success of the bond issues. Especially significant were the economy's resilience to the Mexican crisis and the importance of the city to national GDP, employment, and income. Another key factor was the city's strong revenue raising powers, a sign that repayment did not depend on central government transfers. Conditions in the bond market affected placement dates and price-fix- ing arrangements and determined the underwriting and marketing process. Because the bond sale was relatively small and a debut for the issuer, a full underwriting commitment could not be obtained from the arrangers. In- stead, the bonds were sold on a best-efforts basis. The city's reputation and its plans for reform also contributed to the suc- cess of the issues, despite the city's less-than-optimal financial perfor- mance. Among the greatest concerns for the rating agencies was the fiscal deficit, considered a sign of structural problems and a constraint on reform. The city's financial reporting system, while needing improvement, did not adversely affect the issues, though it slowed the due diligence and rating process. Recent Developments In 1999 and 2000 the city of Buenos Aires was able to maintain a strong fis- cal position despite the economic crisis in Argentina. By cutting capital spending and reducing the budget for noncore activities, it achieved sur- pluses of 2.3 percent of total revenue. In 2001, however, a decline in own- source revenue led to a deficit of almost Arg$250 million, equivalent to 8.4 percent of total revenue. While revenues remained relatively stable in the first half of 2001, they started to decline in August 2001, when they aver- aged 9 percent less than in August 2000, and fell sharply for the rest of the year. In December 2001 revenues were 46 percent less than in the same month in 2000. With the city of Buenos Aires deriving 90 percent of its rev- enues from own sources, this dramatic decline in own-source revenues had a big impact on the city's solvency. City authorities decided not to participate in the provincial debt swap promoted by the federal government in November 2001, which the rating agencies considered a partial default. The city's debt stock was sustainable. Annual debt service reached US$196 million at the end of 2000, equivalent to 6.1 percent of current revenue. Almost all debt had been issued at a fixed rate, and about 46 percent was denominated in Argentine pesos. Country Case Studies: Argentina 253 At the end of December 2001 the city council approved the Economic and Social Emergency Law, which allows the executive body to issue bonds to pay employees and suppliers and to contract additional debt of up to US$218 million. In February 2002, in response to the deterioration in eco- nomic activity in Argentina, Standard & Poor's reduced its rating of the Buenos Aires foreign currency bonds to CCC+ on the global scale, and in June 2002 Moody's rated the city Ca. In late 2002 the city was conducting negotiations with investors to restructure its bonds. Province of Buenos Aires: An Extensive International Bond Program The province of Buenos Aires launched a euro medium-term note program in 1994 totaling US$3.2 billion. The intention was to finance provincial needs but also to gain credibility and a sound reputation in global markets. The notes could be structured with maturities ranging from 30 days to 30 years and issued in currencies including the Argentine peso, U.S. dollar, euro, yen, deutsche mark, Swiss franc, and Italian lira. All issues under the note program (except for the 30th) were sold at fixed rates, and all bonds had bullet maturities. In addition, the province engaged in a wide variety of debt swaps, all against the U.S. dollar. Under this note program Buenos Aires had frequent recourse to the in- ternational bond market in recent years (table 14.9). Its record as an issuer in those years shows that it was a relatively regular issuer, it had strong debt management capacity, and its exposure to currency risk was very high, which led it to declare a default after the Argentine devaluation. The Issuer Buenos Aires is the largest province in Argentina, with a population of 13.8 million. The province is a net contributor to Argentina's subnational sys- tem, receiving only 23 percent of federal transfers, well below its share of the national population (38 percent) and GDP (35 percent). Economic Performance. Buenos Aires is the main driver of the Argentine economy, contributing more than a third of GDP. Per capita income in the province reached Arg$6,980 in 2001, a little less than the national average. The service sector accounted for almost 50 percent of production in the province in 2001, with finance, real estate, and insurance alone contribut- ing almost 20 percent. Manufacturing is the main economic activity, repre- senting more than 31 percent of economic production. 254 Subnational Capital Markets in Developing Countries Table 14.9. Access to the Bond Market by the Province of Buenos Aires, 1994­2001 Amount in Amount original in U.S. Interest currency dollars Issue Due rate Issue numbera Currency (millions) (millions) date date (percent) 1 U.S. dollar 100 100.0 14/07/94 14/07/97 9.50 2 U.S. dollar 15 15.0 16/08/95 16/08/98 11.50 3 U.S. dollar 100 100.0 19/10/95 20/10/98 11.50 4 (swap) Deutsche mark 150 104.5 07/12/95 07/12/98 10.00 5 (swap) Deutsche mark 250 170.2 05/03/96 05/03/01 10.00 6 Swiss franc 200 159.0 23/10/96 23/10/03 7.75 7 Euro 100 108.8 13/07/98 12/07/02 7.88 8 ­ Reopening 6 Swiss franc 75 55.7 23/12/98 23/10/03 7.75 9 U.S. dollar 150 150.0 19/03/99 15/03/02 12.50 10 (swap) Euro 175 185.0 06/05/99 06/05/04 9.75 11 (swap) Euro 150 151.9 12/07/99 12/07/06 10.63 13 (swap) Euro 300 289.7 03/03/00 03/03/05 10.75 14 U.S. dollar 350 350.0 29/03/00 29/03/10 13.25 15 ­ Reopening 13 (swap) Euro 50 48.3 14/04/00 03/03/05 10.75 16 Yen 3,000 27.9 24/05/00 27/05/03 4.25 18 (swap) Euro 100 96.5 05/07/00 05/07/04 10.00 21 U.S. dollar 100 100.0 27/09/00 01/08/03 12.75 22 U.S. dollar 160 160.0 31/08/00 05/09/07 13.75 23 Euro 100 89.4 06/09/00 06/09/02 9.00 27 (swap) Euro 300 276.3 30/01/01 30/01/03 10.25 28 (swap) Euro 300 274.4 23/02/01 23/02/04 10.38 30 U.S. dollar 74 74.0 28/09/01 28/09/06 24.17 a. Numbers missing from the sequence in the column correspond to the number of a treasury bill issued by the province. Source: Province of Buenos Aires Public Credit Office. Financial Performance. The province derives more than 55 percent of its rev- enue from provincial taxes, mainly turnover taxes (23 percent), property taxes (7 percent), and motor vehicle licensing fees (5 percent). Federal transfers pro- vide the other 45 percent, a share similar to that for most other provinces. Buenos Aires was strongly affected by the fall in its own revenues since 1999. From 1998 to 2001 operating revenues fell by more than 15 percent, while op- erating expenditures rose by 11.7 percent. Even so, total expenditures increased by only 2.3 percent, reflecting cuts in capital spending and investments. During this period the accumulated fiscal deficit totaled almost Arg$8 billion. The province privatized its electricity utility (for more than US$1.2 bil- lion) and its water company (US$440 million). However, the province still owns several companies (railroads, a bank, a hotel, and a shipyard). More- Country Case Studies: Argentina 255 over, it retains ownership of Banco de la Provincia de Buenos Aires, the sec- ond largest Argentine bank. From time to time the bank generates signifi- cant costs to the province because of nonperforming loans resulting from unsound credit management practices. In the second half of 2002 the province bought the bank's nonperforming loan portfolio by issuing a provincial bond for US$1.3 billion. Because Buenos Aires predates the Argentine republic and joined the Ar- gentine confederation only after the national Constitution was adopted, it has certain prerogatives. One of them is that its provincial bank is not gov- erned by the Argentine Central Bank. Because of the provincial bank's im- portance, however, the two banks maintain close coordination. Debt Profile. Buenos Aires has a stable and well-trained debt management team that has gained much experience in debt markets since 1994 as the province has pursued a debt strategy focusing on bonds. The province's finan- cial management and reporting system are reasonably effective by Argentine standards, though it has problems resulting from the different accounting treatment of revenues and expenditures, the incompleteness and lack of con- sistency of some information, and the absence of audited financial statements. Huge provincial deficits have led to substantial growth in the debt stock of Buenos Aires. In 2001 the province's indebtedness increased sharply be- cause of its enormous deficit and the capitalization of the provincial bank (table 14.10). Almost US$3.7 billion of the province's debt at the end of 2001 was issued under foreign laws; accordingly, this part of the debt increases as the Argentine peso is devalued. Even so, at the end of 2001, before the debt swap and the Argentine devaluation, Buenos Aires had debt indicators simi- lar to the average for provinces. The exception was debt service as a share of operating revenue, where Buenos Aires exceeded the average (figure 14.10). Buenos Aires was the first province to sign the agreement with the feder- al government required as a condition of the negotiations with the IMF on a financial assistance program. Like most of the provinces that later signed such agreements, Buenos Aires committed to reduce its fiscal deficit by up to 60 percent of the 2001 deficit. It achieved a substantial part of the deficit reduction by defaulting on bond and loan payments. Without structural reform the reduction is unsustainable, because the main problems that led to those deficits remain unsolved. Key Factors Affecting the Issues Moody's and Standard & Poor's are the credit rating agencies that rate Buenos Aires. Since the province's first launch under the program, provin- 256 Subnational Capital Markets in Developing Countries Table 14.10. Debt by Source, Province of Buenos Aires, 1995­2001 (millions of Argentine pesos) Source 1995 1996 1997 1998 1999 2000 2001 Banks 2,108.8 2,053.5 2,024.4 2,046.7 2,030.1 2,341.9 2,631.2 Multilateral lenders 169.2 173.8 330.3 450.2 727.2 907.2 968.7 Provincial Development Fund (refinancing program) 0.0 0.0 0.0 0.0 0.0 0.0 421.3 Bonds 319.4 694.4 735.3 725.7 1,385.1 3,340.6 6,412.8 Other debts 587.3 688.7 692.5 770.9 844.1 886.9 1,087.2 Total 3,184.7 3,610.4 3,782.4 3,993.5 4,986.5 7,476.5 11,521.3 Note: Data are as of the end of December of each year. Source: World Bank staff estimates based on Argentine Ministry of Economy and Province of Buenos Aires Public Credit Office data. 40 Average interest rate 30 24 Provinces Buenos Aires 20 Debt per capita/ income per 10 capita Debt service/ operating revenue Debt/ economic production Source: World Bank staff estimates based on Argentine Ministry of Economy and Province of Buenos Aires data. Figure 14.10. Selected Debt Indicators, Buenos Aires and All Provinces, End-2001 (percent) Country Case Studies: Argentina 257 cial ratings have generally changed with the sovereign Argentine rating. The rating agencies have expressed concerns about the province's ability to tackle fundamental structural problems in revenues and expenditures and about the inefficiencies of Banco de la Provincia de Buenos Aires. Dur- ing the first issues the ratings were an important factor in accessing the market. The key factors have been the significant revenue raising powers of the province, its large tax base, and its strong negotiating position with the federal government. Buenos Aires usually leads every negotiation between the provinces and the federal government. Moreover, Buenos Aires is by far the best-known Argentine subnational debt issuer. The provincial administration's commitment to reform has proved to be weak. While the province privatized some of the companies it owned, it continues to own Banco de la Provincia de Buenos Aires, its largest source of quasi-fiscal deficits, as well as other corporations that are not a core part of provincial activity. Provincial authorities have been unable to cut fiscal deficits or implement serious reform since 1999. They have tried to reduce fiscal deficits by cutting capital spending, but year after year the decline in revenues has exceeded the spending cuts. Like the other provinces, Buenos Aires has much work to do in improv- ing financial reporting and disclosure. Its financial statements still are not prepared or audited in accordance with international standards. Recent Developments In recent years the province has maintained stable expenditures by cutting capital spending, but the fall in revenues forced it to finance substantial fis- cal deficits. Provincial authorities decided to participate in the provincial debt swap promoted by the federal government in November 2001. As usu- al Buenos Aires was the largest player, entering the debt swap with a target- ed amount of more than US$6.4 billion. On 29 January 2002 the province declared a default on some bond pay- ments, initiating the largest Argentine provincial default in history. During the first half of 2002 the province continued issuing money bonds (Pata- cones) to finance its fiscal deficit. At the end of June 2002 the outstanding debt in Patacones reached Arg$2.4 billion (2.5 percent of annual economic production in the province). On 25 July 2002 the first tranche of Patacones was due, for an estimated amount of Arg$500 million, but because of its fi- nancial situation the province had to exchange these bonds for a second tranche (Patacones B). 258 Subnational Capital Markets in Developing Countries The province has been greatly affected by the Argentine devaluation. According to the province, its outstanding debt reached Arg$21.3 billion at the end of March 2002. Unlike other bond issuers that were already in- volved in debt restructuring, such as Santiago del Estero and the city of Buenos Aires, the province of Buenos Aires declared that it would wait for the sovereign debt restructuring before renegotiating its debt. In February 2002 Standard & Poor's reduced its rating of the Buenos Aires long-term foreign currency bonds to CCC+ on the global scale, reflecting the deterio- ration of economic activity in Argentina. In June 2002 Moody's downgrad- ed the province's foreign currency debt rating to Ca. Notes 1. Gross coparticipation includes different programs such as the Nation- al Fund for Housing (FONAVI) that were historically earmarked revenues, but since 2000 these revenues have been converted to nonearmarked rev- enues. 2. Most provinces have their own interpretations of the terms revenues and debt service. 3. The exchange rate at the end of 2001 was US$1 to Arg$1. 4. The privatization program was created in the mid-1990s to encourage provinces to privatize their financial institutions. Originally funded by the World Bank, the Provincial Development Fund later was capitalized by the national treasury. Thus in the late 1990s, the fund supported provincial bank privatization using its own assets. 5. Some bondholders brought suit against the province and its bank-- against the province for unilaterally deferring payments and against the bank for breaking the Argentine Trust Law (Law 24.441) by carrying out the province's order. 6. Bond issuance in Argentine provinces typically takes an average of four to six months if the bonds are backed by coparticipation revenues, and six to nine months if backed by oil revenues. 7. The money bond is not unique to Argentina. They have been used in unusual circumstances in the United States. The states of Michigan (in the 1980s) and California (in the early 1990s) issued "warrants" to pay employ- ees and suppliers during cash crises. The warrants were very short term and were issued at a discount. Banks accepted the warrants from the payees and then cashed them in at maturity. Country Case Studies: Argentina 259 8. In December 1999 the provinces and the federal government signed a federal compromise fixing the total monthly transfers owed to the provinces until the end of 2000. By the end of the period fiscal difficulties in Argentina and the pending negotiations with the IMF on a financial res- cue package brought provincial transfers under tough scrutiny. In Novem- ber 2000 a second federal compromise was signed that fixed total transfers to provinces for 2001, obligated provinces to pursue fiscal discipline, and required the federal government to increase funds for unemployment and social programs and to allow provinces to administer part of these funds. 9. From April 1991 to January 2002 the Argentine peso and the U.S. dol- lar were at parity. 10. The city was able to hedge euro and Italian lira debt against the U.S. dollar, but it could not hedge its U.S. dollar debts. Thus after the devalua- tion its indebtedness increased dramatically. Chapter 15 Latin America and the Caribbean Brazil A past of excessive borrowing by a few large states makes the future difficult for all subnational entities. Rodrigo Trelles Zabala and Giovanni Giovanelli Lessons Brazil's experience with subnational borrowing serves as a cau- tionary tale of the deep and lasting effects that weak central control, macroeconomic instability, fiscal indiscipline, and insuf- ficient regulation can have on a country's public finances. This story in large part reflects the legacy left by imprudent lending by state banks and failure to subject the states to the discipline of the capital market. It also reflects the gyrations of Brazil's po- litical system as it alternated between decentralization and re- centralization. The latest phase of democratization has led to advanced devo- lution of political and fiscal authority to the states, giving them substantial power to generate revenue and a large degree of au- tonomy. Subnational borrowing powers have traditionally been extensive and flexible. There was abundant borrowing in the 1960s and 1970s, with both domestic and foreign bond issues permitted as well as financing from state-owned banks, which 261 262 Subnational Capital Markets in Developing Countries often amounted to "lending to oneself." Financing by state- owned banks proved to be a key source of fiscal indiscipline, exacerbating already weak central controls and the ambiguous intergovernmental framework, where the assignment of expen- diture responsibilities is particularly opaque. In addition, pres- sures on state budgets, such as generous pension plans for re- tired public servants, made balancing the budgets difficult. Brazil has suffered multiple bouts of macroeconomic instability, starting with debt defaults by the central government in the 1980s and hyperinflation in the mid-1990s. This instability has pushed local finances over the edge, leading to a need for three rounds of bailouts in recent years. The moral hazard that central guarantees and recurring bailouts have introduced in local fiscal behavior has been difficult to erase. Credit enhancements-- such as the Central Bank's appropriation of intergovernmental transfers to guarantee repayment--have removed incentives for creditors to factor local fiscal health into their financing deci- sions. As a result of the most recent default, however, the cen- tral government prohibited any additional borrowing (with the exception of refinancing existing debt) until 2010. It also insti- tuted stricter controls for managing outstanding local debt and placed a cap on state spending. Several characteristics of the crises serve as useful lessons. Al- though 30 percent of local debt took the form of bonds, the bond debt problem was concentrated in a handful of states ac- counting for 90 percent of this debt. A large share of debt was incurred with state banks that lacked incentives to perform competent analyses of local financial conditions and, in many cases, resulted in the obvious conflict of having a governing body lend to itself. Additionally, the absence of the private sec- tor from subnational lending eliminated a potential source of evaluation and control. This last characteristic is a curious one, since Brazil's financial markets are relatively developed by Latin American standards. The central government's current stranglehold on local debt and financial operations has not addressed the underlying prob- Country Case Studies: Brazil 263 lems of Brazilian states, particularly the inability to cure persis- tent fiscal deficits and the continued rollover of highly subsi- dized debt. Legislation has focused on administrative controls and restrictions, and little has been done to correct the deficien- cies in market mechanisms. Regulatory reforms have been pro- posed--including laws relating to bankruptcy, contracts, and disclosure--to foster a prudent, market-based institutional framework. Some headway has been made in these areas, and there is hope that further reform, together with improvements in local fiscal health and retirement of the existing debt burden, will open the door to a sustainable capital market for local obli- gations in the medium term. Sovereign Context Brazil is politically structured as a federation. While the revenue sources of the different tiers of government are reasonably well laid out by the Con- stitution, there is much overlap in the provision of services. The country has a large municipal sector with around 5,500 units--ranging from small rural enclaves to the massive urban centers of São Paulo and Rio de Janeiro. These two megacities have tended to overshadow much of the rest of the country economically and politically and, not coincidentally, ac- count for more than two-thirds of the municipal debt. Much of the politi- cal history of the country has been marked by a tug-of-war over resources and influence between the wealthier regions in the Southeast and the poor regions in the Northeast. Brazil has been plagued by a history of fiscal and financial instability. The large debts accumulated in the past by some of the provinces and the two largest cities have imposed a big burden on the country's finances, and their refinancing through a series of federal bailouts has led to major macroeconomic problems. The highly decentralized public sector and heavy personnel expenditures have contributed to persistent public sector deficits. Transfers from the central government dominate local revenues, accounting for about two-thirds on average, and the many very small mu- nicipalities depend heavily on them. 264 Subnational Capital Markets in Developing Countries As a result of recent financial reform efforts, direct borrowing from the financial markets is now tightly regulated and municipal borrowing is cur- tailed. Borrowing is limited to subsidized loans from two state-controlled banks--in effect, the government lending to itself. Private lending to mu- nicipalities is thereby effectively precluded. Efforts to introduce private lending require changes in the concessionary loan practices as well as other reforms to improve creditworthiness. The moral hazard resulting from a tradition of interference and bailouts of troubled loans presents a major ob- stacle to creating an efficient market for subsovereign credits. Macroeconomic Conditions The Brazilian economy is the largest in Latin America and the tenth largest in the world by GDP, with a strong export-oriented private sector. Before the introduction of the Real Plan in 1994 Brazil's economic performance had been characterized by macroeconomic instability. The events of the 1970s and 1980s--the oil shock, the debt crisis, the rise in real interest rates, and the decline in foreign direct investment and credit--caused a drastic contraction of the economy. State intervention, poor fiscal manage- ment, exchange rate management, and general indexation of wages con- tributed to hyperinflation and state and federal fiscal deficits. In 1980­88 annual inflation averaged 200 percent, and in 1989­94 it soared to an aver- age 1,260.3 percent. After the Real Plan was introduced in 1994, however, inflation decelerated, falling to a manageable 9 percent in 1996, the year that Rio de Janeiro floated a municipal bond issue in the international bond markets. Aimed at curbing inflation and building a foundation for sustained eco- nomic growth, the Real Plan was designed to address persistent deficits in the federal government's accounts, expansive credit policies, and wide- spread backward-looking indexation. The plan was implemented in three phases. The first, addressing the fiscal deficits, had as its centerpiece the cre- ation of the Emergency Social Fund by constitutional amendment in Feb- ruary 1994. The second phase, initiated in March 1994, began a process of monetary reform by introducing a new index, the real unit of value, aimed at eliminating the distortions in relative prices in the economy. In July 1994 the federal government initiated the third phase of the Real Plan by adopting a new currency, the real, with an initial ceiling of parity with the U.S. dollar, and removed the real unit of value. By promoting deindexation of most prices and adopting a floating exchange rate subject to a parity cap, the federal government was able to orchestrate an abrupt deceleration of Country Case Studies: Brazil 265 inflation, a convergence in the growth rates of tradable and nontradable goods, and greater competition in all sectors. Large imbalances remained in public finance. Brazil's current account, which ran an average deficit of 0.02 percent of GDP between 1990 and 1994, deteriorated to a deficit of 2.5 percent in 1995 and 3.2 percent in 1996. In addition, Brazil's external debt ratios remained relatively high. At the end of 1996 total external debt stood at $178.1 billion, equivalent to 322.7 percent of exports, up from 296 percent the year before. Annual debt service obligations were also heavy, reaching 49.3 percent of exports in 1996 and 57.3 percent in 1997. In the fall of 1997 the Brazilian currency came under attack as a result of the general anxiety about emerging markets that grew out of the East Asian crisis. Unlike Argentina, Brazil had not tied its currency to the dollar but al- lowed its targeted exchange value to crawl downward, allowing some room for inflation. It raised interest rates to defend the currency and appeared to be faring well until the Russian crisis in the summer of 1998 brought on another crisis in confidence, intensified by the threat of Minas Gerais to de- fault on its debt to the federal government. In January 1999 Brazil devalued its currency. Assisted by a loan from the International Monetary Fund, it immediately implemented a targeted infla- tion monetary policy that contained inflation: in 1999 the consumer price index rose by 4.9 percent, in 2000 by 6.2 percent, and in 2001 by 9.4 per- cent. Brazil's debt management strategy focused on extending the maturi- ties of federal debt by indexing government securities to the U.S. dollar and the inflation rate. That debt structure, combined with the Argentine de- fault at the end of 2001, led to a new Brazilian debt crisis. In August 2002 the federal government received a package of financial assistance from the International Monetary Fund: a $30 billion loan that was to be disbursed in two installments, the first ($6 billion) before the presidential election and the second ($24 billion) when the newly elected president took office. Meanwhile, anticipating political change, the international financial mar- kets reacted nervously to the election campaign and the real faced contin- ued downward pressure in world markets. Structural Reforms In the 1990s Brazil undertook myriad reforms as it attempted to liberalize its economy and contain the size of its government sector. The Cardoso ad- ministration, entering office in October 1994 with a clear agenda of reform, made great progress in privatizing state-owned enterprises and improving 266 Subnational Capital Markets in Developing Countries the climate for foreign investment. However, other initiatives critical for consolidating the public sector were not implemented, including cutting public sector payrolls, reforming the tax structure, overhauling the social security system, and reforming the civil service. Brazil's privatization program is among the largest and most compre- hensive in the developing world. The government has eliminated several distortions in the program, most notably the distinction between resident and nonresident ownership of companies, which had prevented foreign participation in such sectors as mining, transport, petroleum, electricity, and telecommunications. It also improved the regulatory regime and intro- duced tax exemptions and incentives for investments in less developed re- gions and export-oriented zones. Between 1991 and 1995 Brazil privatized 41 companies, for total revenues of $9.2 billion; privatizations in 1996 raised another $6 billion. The privati- zations also transferred $8.1 billion in debt to the private sector. Foreign di- rect investment, which rose from $2.2 billion in 1994 to $17 billion in 1997, accounted for a third of the privatization proceeds. However, difficulties in the public sector persisted, proving to be largely impervious to reform. Intergovernmental Relations The Brazilian federal structure, established by the 1988 Constitution, con- sists of the federal government, 26 states, one federal district, and an unde- fined number of municipalities (roughly 5,500 today). The 1988 Constitu- tion set the powers of the federal government, which include national defense, social security, monetary policy, control of public debt, interstate and foreign trade, and the establishment of general norms for civil ser- vants. It granted states all powers not otherwise reserved for the federal government. The Constitution also delineated some concurrent responsi- bilities of the federal government and states, including education, tax legis- lation, and social assistance, and it specified that federal law, while limited to general norms, prevails in case of conflict with state legislation. Unlike other federal constitutions, which typically subject municipali- ties to the control of their state, the 1988 Constitution recognized munici- palities as a third tier of government with the same constitutional status as states. Accordingly, states cannot impose on or prohibit the actions of the municipalities within their jurisdiction. The Constitution left the division of functions and responsibilities between states and municipalities ambigu- ous, merely reserving for municipalities the power to legislate on subjects Country Case Studies: Brazil 267 of local interest and provide for local services. The loose controls on the lo- cal sector led to the emergence of a large number of small municipalities, an outcome fostered by the intergovernmental transfer system.1 Revenue Raising Capabilities The 1988 Constitution explicitly defined the division of tax responsibilities between the levels of government. In addition to assigning a specific tax base to each level of government, the Constitution created a system of rev- enue sharing that redistributes resources among levels of government and geographic regions. Direct Revenues. The Constitution assigned states receipts from the value added tax and authorized them to tax automobiles and real estate. Since the value-added tax is the highest yielding tax in Brazil, this assignment gave states much independence, particularly in the wealthy Southeast. States retained some flexibility to set the rates on interstate sales, subject to the minimum and maximum limits established by the Senate. Municipalities were assigned a tax on services, an urban property tax, and a real estate transaction tax. These are all locally assessed and collect- ed, although the tax on services is subject to a maximum established by federal law. Revenue Sharing System. The 1988 Constitution substantially increased the amount of taxes shared by the federal government. Brazil's revenue- sharing system has two main parts: the participation funds and the state value added tax. The participation funds consist of fixed shares of the federal govern- ment's two principal taxes: the income tax and the industrial product tax. Under the 1988 Constitution the federal government is required to transfer 21.5 percent of the participation funds to the states. Within each group of states, 95 percent of the funds are distributed among states on the basis of population and per capita income, with poorer states receiving a larger share. The other 5 percent is distributed in proportion to the area of states, to cover the relatively higher expenditures associated with a dispersed popu- lation. The federal government distributes another 22.5 percent of the par- ticipation funds to municipalities, transferring 10 percent of this amount to state capitals and distributing the other 90 percent among all other munici- palities on the basis of population and the state's per capita income. The participation funds represent a substantial redistribution of rev- enues among regions. On average, the less wealthy states of the North, Northeast, and West-Central regions receive twice as much as the states of 268 Subnational Capital Markets in Developing Countries the South and Southeast. The participation funds doubled in size between 1967 and 1992 and have been a predictable and reliable source of income over the past 10 years. The state value-added tax is the second major tax-sharing arrangement. Under the Constitution states are required to transfer 25 percent of their proceeds from the value-added tax to the municipalities within their territo- ry. Of this amount, 75 percent must be distributed on the basis of the origin of tax collections. The other 25 percent is distributed according to formulas established by each state legislature. The Constitution expanded the base of the state value-added tax by abolishing federal taxes on fuel, mining, trans- port, and electricity and incorporating these into the state value-added tax. Expenditure Responsibilities In contrast to the explicit provisions on revenue sharing, the Constitution leaves unclear how expenditure responsibilities are to be divided between federal and subnational governments and between states and municipali- ties. This ambiguity has led to friction over their roles. To match the in- crease in revenue sharing mandated by the 1988 Constitution, the federal government proposed a program of decentralizing expenditures. When this proposal was rejected by the Congress, the federal government trans- ferred some expenditure responsibilities to states and municipalities on an ad-hoc basis. These included suburban railways and highways in São Paulo and Rio de Janeiro, transferred to their state governments, and federal hos- pitals in Rio de Janeiro, transferred to the state and municipality. The feder- al government also unloaded some health care costs onto subnational gov- ernments by reducing federal compensation payments. Despite the federal government's decentralization efforts, the 1988 Con- stitution extended central control over two main areas: personnel and state debt. Under the Constitution state and local governments cannot dismiss redundant civil servants or reduce nominal salaries. Public employees have the right to retire after 35 years of employment (30 years for women and teachers) and to receive a pension equal to their final salary plus any subse- quent constitutionally mandated increases. This mandate has proved to be onerous, substantially reducing the fiscal flexibility of states and munici- palities. Pension benefits are particularly troublesome: constitutionally pro- tected, very liberal, and unfunded, they represent an ongoing drain on cur- rent revenues.2 Reforms have been undertaken, but they are forward looking, and civil servants employed at the time the 1988 Constitution was adopted continue to be protected by its provisions (World Bank 2001). To Country Case Studies: Brazil 269 restrict growth in the protected classes of civil servants, local governments reportedly are attempting to privatize services and hire workers on a tem- porary basis. In response to the profligate borrowing of the past, the 1988 Constitu- tion also provided that any state or municipal government wishing to bor- row, domestically or internationally, must obtain approval from the Senate. Subsequent tightening of statutes and regulations has sought to rein in subnational borrowing and reduce the need for further bailouts by the na- tional government. Regulatory Framework for Subnational Borrowing Brazilian states and municipalities traditionally have had access to a wide variety of debt funding sources: · Domestic bond issues. · Domestic private commercial banks. · Federal intermediaries, such as the Federal Housing and Savings Bank and the Federal Development Bank. · State-owned commercial banks. · Foreign institutions, including multilateral development banks and private commercial banks. · Informal sources, such as arrears on salaries and on payments to sup- pliers. Under the 1988 Constitution the Senate retained the authority to regu- late state borrowing. It adopted a resolution regulating such borrowing on the basis of a state's existing debt stock, its revenues, and its capacity to ser- vice debt. However, the Senate reserved the right to grant exceptions, and it often did so. In 1998 the Senate adopted several new measures to control subnational debt. One of these, Senate Resolution 78, prohibits the issuance of new subnational bonds until the end of 2010 except to finance the rollover of previously issued bonds. In addition, Resolution 78 contains the following: · Prohibits borrowing from own enterprises or suppliers. · Limits new debt to no more than 18 percent of real net revenues.3 · Limits annual debt service to no more than 13 percent of real net rev- enues. 270 Subnational Capital Markets in Developing Countries · Limits debt outstanding to no more than 2 times real net revenues.4 · Prohibits governments in default from accessing new borrowing. · Requires governments to have a primary surplus before obtaining new loans.5 · Prohibits governments from contracting new debt during the last six months of their term. The Law of Fiscal Responsibility, adopted in 2000 by the Senate, takes a more comprehensive approach, extending beyond subnational govern- ments to the federal government as well. The law contains the following: · Limits all personnel costs--including pensions and permanent and temporary personnel--to 60 percent of current revenues. · Limits the net stock of debt to no more than 2 times net current rev- enues for states and 1.2 times for municipalities. · Allows states and municipalities that exceed the debt stock limit 15 years to adjust to the requirements. · Authorizes new debt only when debt service does not exceed 11.5 percent of current revenues. · Forbids borrowing between levels of government, except for federal institutions. External borrowing by states is largely exempt from federal regulation unless it requires a federal guarantee, in which case the Ministry of Finance has the authority to grant or deny federal backing. Still, the National Mon- etary Council of Brazil, in its Resolution 2280, established conditions for the external credit operations of states and municipalities. The two most important provisions of this resolution are the following: · The proceeds of the external credit must be used to refinance the is- suer's outstanding domestic financial obligations, with preference given to the obligations with a higher cost of funding or shorter ma- turity than the external debt. · In cases where the issuer has no credit rating, the issuer must estab- lish a sinking fund escrow account with a balance equivalent to the monthly debt service obligation (principal and interest). The federal government and the Central Bank have attempted to tight- en regulations on the supply side. Central Bank Resolution 2461, adopted Country Case Studies: Brazil 271 in 1998, prohibits private banks from increasing their holdings of state debt other than bonds. However, it does allow them to adjust the composi- tion of their state debt portfolios as existing debt matures. Central Bank regulations also prohibit states from borrowing from their own commercial banks, although this rule has not been strictly enforced. In addition, the Central Bank prohibits public sector banks and financial institutions from having more than 45 percent of their equity in the form of loans to or investments in public sector entities. The Federal Housing and Savings Bank and the Federal Development Bank are both subject to this limitation. Municipal development funds are not subject, though they are limited by the Fiscal Responsibility Law. The Central Bank also controls borrowing in its capacity as adviser to the Senate: every borrowing request must be directed to the Central Bank, which analyzes each case and makes a recommendation to the Senate. Interestingly, all limitations on subnational borrowing are based on ad- ministrative controls, with no market-oriented mechanisms in place. Intro- ducing a market-based system of credit allocation remains a dream as the country continues to try to dig itself out of a legacy of fiscal profligacy. Recurring Subnational Debt Crises The regulatory framework to control subnational debt emerged as a conse- quence of three bailouts by the federal government during the 1980s and 1990s. The first followed the debt crisis in 1989, caused by the heavy do- mestic and international borrowing in the 1970s and the shocks to the economy in the early 1980s. When the federal government defaulted on its external debt in the 1980s, subnational governments did the same; when the federal government reached an agreement with foreign creditors, it had to assume the subnational foreign debt of $19 billion. The outstanding debt plus arrears were rescheduled for up to 30 years. This initial bailout in- cluded only the foreign debt of states and municipalities.6 After this first bailout subnational governments started to pressure the federal government to reschedule their debt held by federal institutions. In 1991 a second round of negotiations began, concluding in 1993 with an- other bailout, this time covering only debt with federal institutions ($28 billion). As in the previous bailout, the debt was rescheduled for up to 30 years and interest rates were subsidized. As part of this second bailout the federal government took steps aimed at reducing the need for future bailouts: it prohibited itself from lending to 272 Subnational Capital Markets in Developing Countries states and municipalities in default, and it adopted a constitutional provi- sion allowing itself to intercept intergovernmental transfers to pay debt ser- vice. The limits that the Senate established relating to debt service permit- ted the capitalization of debt service obligations that could not be met. During the negotiations that began in 1991 the states made several at- tempts to include their bonds. These attempts failed, and, not surprisingly, these bonds led to another subnational debt crisis. By the mid-1990s the high interest rates that states faced and the capitalization clause had led to a dramatic increase in their stock of debt. Bonds accounted for 30 percent of the debt not yet refinanced, and the domestic bonded debt of states rose from 2.3 percent of GDP in 1991 to 5.4 percent by mid-1996. However, the debt in bonds was not a widespread problem: four states accounted for more than 90 percent of the almost $30 billion in debt stock in bonds (fig- ure 15.1). This time the solution was a conditional bailout that included a fiscal and financial restructuring program, privatization of public companies, and the sale of state-owned banks, and the negotiations were held on a state-by-state basis. Another important difference was the requirement Sao Paulo 36% Rio Grande do Sul 17% All others 9% Rio de Janeiro 16% Minas Gerais 22% Source: Central Bank of Brazil. Figure 15.1. Distribution of the Debt Stock in Bonds by State, Brazil, End of 1996 Country Case Studies: Brazil 273 that states entering the program make a down payment equal to 20 percent of the debt to be rescheduled. This requirement led to the privatization of state-owned companies and banks. Again the debt was rescheduled for up to 30 years, with a fixed real interest rate equal to 6 percent. This interest rate was heavily subsidized, since the debt of the federal government car- ried much higher rates. Some 25 states and 180 municipalities participated in the refinancing program. All states and municipalities offered their own revenues and revenue transfers as guarantees, but only up to a maximum of 15 percent of their revenues. At the end of 2001 the debt restructured under this program had amounted to more than $100 billion, and a series of new rules had been imposed to control subnational debt (see section on regulatory framework). The Fiscal Responsibility Law represents a landmark in the control of sub- national debt. Even so, rules cannot be seen as a solution to the underlying fiscal problem of persistent operating deficits. At best, rules can restore confi- dence and encourage better fiscal and financial management practices. Sub- national governments' inability to achieve surpluses and their continued rolling over of debt, coupled with the large federal subsidy on outstanding debt, are fundamental problems that Brazil has not yet addressed. Subnational Credit Market While a few states and municipalities have tapped international credit and bond markets, subnational governments have financed their needs mostly through public financial institutions or loans provided by the federal gov- ernment. The debt of states has steadily increased as a share of GDP since 1998, while that of municipalities has remained a fairly constant share (fig- ure 15.2). During the past decade states have issued bonds underwritten by their own banks and then sold to investors and other market participants. Mu- nicipalities have relied mainly on funds provided by the Federal Housing and Savings Bank and Federal Development Bank and by municipal devel- opment funds established with grants from the World Bank and the Inter- American Development Bank. Private banks have played almost no role-- surprising, given the Brazilian financial sector's size and level of development. Why have commercial private banks stayed away from the subnational credit market? There are several plausible explanations: 274 Subnational Capital Markets in Developing Countries Percent 18 State debt 16 Municipal debt 14 12 10 8 6 4 2 0 Dec-98 Dec-99 Dec-00 Dec-01 Jun-02 Note: Figure excludes the debt of state-owned companies. Sources: World Bank and Central Bank of Brazil. Figure 15.2 Subnational Debt as a Share of GDP, Brazil, 1998­2002 (percent) · States and municipalities borrow funds at subsidized, below-market interest rates, making it impossible for private banks to compete for their business. · Private banks have been "burned" in the past by several subnational defaults. · The private sector offers loan maturities that tend to be much shorter than those offered by public financial institutions. · Private financial institutions generally do not offer grace periods for repaying loan principal. Meanwhile, public banks are under close scrutiny by the federal govern- ment, which is trying to prevent the public sector from lending to itself. The government's strategy for doing so is to have public banks lend to pri- vatized infrastructure companies at subsidized interest rates. However, these below-market interest rates imply that some projects and capital in- vestments being financed are not economically efficient. A credit policy re- quiring market interest rates ensures that projects are economically effi- cient and that capital investments are carefully selected and analyzed. The Country Case Studies: Brazil 275 high real interest rates seen in Brazil in 2002, however, make most capital investment projects unviable. Brazil, then, presents a paradox. It has large financial markets, but those markets are not tapped by municipal governments and access to them is very restricted for states. Subnational borrowing is dominated by state- owned and federal banks and an assortment of specialized funds that lend at subsidized rates (table 15.1). Still, the subnational credit market is clearly a big market in Brazil and the biggest in Latin America. To further expand that market, the federal government should promote a market-oriented funding policy to help break away from the old tradition of borrowing from public institutions. Ending that tradition will be difficult without numerous public sector reforms, but surely could be part of a package of such reforms. Table 15.1. Municipal Sources of Funds, Brazil, 1999 How the rate Program Source of funds Lending rate was determined Caixa Economica Mandatory workers' 8­12 percent Margin over the cost of Federal (Federal Housing contributions to FGTS; funds and Savings Bank) credit from Inter-Ameri- can Development Bank (IADB) BNDES (Federal PIS-PASEP and FAR 5­8 percent for 2.5 percent margin. FAT Development Bank) employer social subsidized regions funds carry rate of TJLP insurance contributions and activities; plus 2.5 percent, although up to 16 percent only part has to be paid in for standard loans cash. Remainder is capi- talized indefinitely. Federal Treasury Federal budget 6­9 percent Political negotiation bailout Paraná municipal IADB (formerly World 10.14 percent 3.5 percent over IADB development fund Bank) reference loan rate Minas Gerais municipal World Bank 9.04 percent 3 percent over World Bank development fund reference loan rate Ceará Development World Bank and 9 percent Spread over base rate Bank and municipal Federal Development development fund Bank Private sector Market 34.5 percent for Market commercial loans two-year commer- cial loans to prime borrowers Source: World Bank 2000. 276 Subnational Capital Markets in Developing Countries In addition, subnational governments need to develop better fiscal and financial management practices to generate confidence among private lenders, which are both skeptical of government credits and conditioned to expecting bailouts. As a result of this lack of confidence, municipalities did not have access to medium- and long-term private funds to finance their capital investments. Moreover, in contrast with many other countries, where large cities have been encouraged to borrow from private banks, in Brazil large cities borrow proportionally more from public banks than smaller municipalities do. Interestingly, the Central Bank's ability to intercept intergovernmental transfers to service subnational debt provides investors with a much bet- ter safeguard than those available in other Latin American countries with subnational credit markets at similar levels of development (such as Ar- gentina). This type of credit enhancement, however, also has costs. It eliminates the incentives for lenders to analyze potential subnational creditors, because they think that their loans will be repaid no matter how the loan proceeds are invested. Further, it eliminates the incentives for state and local governments to analyze their projects, because they know they can gain access to the credit market by pledging their revenues to the Central Bank. All this makes clear that Brazil's subnational credit market has a low lev- el of financial intermediation and efficiency--and that measures are need- ed to reduce the cost of funds and increase efficiency. A recent World Bank study (2001) proposed the following initiatives: · Strengthening contract enforcement. · Reforming the bankruptcy law. · Extending the maturities of commercial bank loans. · Increasing the efficiency of the judicial sector. · Strengthening the rights of secured and unsecured creditors. · Improving the quality of information provided to the market. · Introducing better accounting standards and practices. · Developing a stronger framework for sharing creditor information among financial institutions. · Adopting a new, more comprehensive securities law. These recommendations point to the importance of the legal and regu- latory framework in developing local credit markets. A clear priority is re- form of creditors' rights to rank secured creditors first. Another is reform of Country Case Studies: Brazil 277 the bankruptcy law, to move away from the tradition in Brazilian legisla- tion of favoring debtors, and there is a clear need for a comprehensive se- curities law. Today legislation relating to securities is dispersed among the civil code, commercial laws, financial sector rules, and special laws apply- ing to particular financial instruments. The quality and availability of information need to be improved not only to reduce uncertainty but also to add greater transparency to the cred- it system. Some initiatives already have been taken in this area. For exam- ple, the Central Bank has created the Credit Risk Data Center, a system that provides monthly information on credit operations of 20,000 real (equiva- lent to roughly $6,500 today) and above. Finally, extending the maturities of debt will help achieve a more stable macroeconomic framework. Clearly, much work needs to be done to develop a private credit market for subnational borrowers. In an important step, the authorities appear to recognize the need to move away from captive sources of funding in the medium term. A market-oriented funding policy will lead to a better alloca- tion of funds and a better assessment of investment projects by subnation- al governments as well as lenders. Limits need to be imposed on debt not to reduce or discourage municipal borrowing but to ensure that loans are used to fund capital investments and that the investments financed are economically efficient. Notes 1. The mechanism for distributing federal aid is an unintended but ef- fective inducement to form small municipalities. This mechanism favors small municipalities, which derive up to 90 percent of their revenues from transfers. Lenient requirements for incorporation allow the federal trans- fers to become a revenue source and thus a means of employment for would-be government officials and workers. 2. In the municipality of Rio de Janeiro retiree payments, fixed at the level of the retirees' final salary and indexed to salary increases for their last position, grew from 26 percent of payroll in 1993 to 35 percent in 1997. Since the city bureaucracy is growing slowly, the number of retirees will one day surpass the number of employed workers. See World Bank (2001, p. 23). 3. Real net revenues are total revenues less receipts from credit opera- tions less property sales less transfers for specific purposes less specific grants for specific projects. 278 Subnational Capital Markets in Developing Countries 4. The limit declines by 0.1 annually until 2008, when it reaches 1.0. 5. The primary deficit or surplus is equal to total revenues less total ex- penditures less interest payments. 6. The rules of the bailout were set by Law 7978 (27 December 1989). Those of the second and third bailouts were set by Law 8727 (5 November 1993) and Law 9496 (11 November 1997). Chapter 16 Latin America and the Caribbean Colombia Despite fiscal difficulties, the country has succeeded in using the private market mechanism to raise funds while limiting local borrowing. Rodrigo Trelles Zabala Lessons Colombia has made a significant shift--though with restric- tions--toward decentralization, but the consequent shift to sub- stantial transfers has caused fiscal imbalances for the central government. Strong central control curbed an earlier accelera- tion in subnational borrowing. Continuing deficiencies in the regulatory framework surprisingly have not led to widespread fiscal difficulties, though decentralization and mandated spend- ing have continued to strain fiscal balances. Use of credit by subnational governments grew sharply in the 1990s because of inflexibility in local expenditures. Borrowing restrictions were lax because of legislated mandates to increase central transfers to finance required expenditures, and private banks, provided with an intercept mechanism, were content to lend. As a result, subnational borrowing doubled relative to GDP. The substantial increase in debt led to enactment of a law 279 280 Subnational Capital Markets in Developing Countries requiring approval from the Ministry of Finance for additional debt and tying borrowing controls to the fiscal health of local governments. In addition, new rules required banks to increase capital reserves for riskier subnational loans, increasing the cost of borrowing from commercial banks. The new law slowed the growth of subnational debt substantial- ly, and the central government has been able to avoid large bailouts of subnational governments. In contrast to Argentina and Brazil, Colombia essentially used a market-based mecha- nism to impose limits on local borrowing, allowing borrowing to continue in a controlled environment. Findeter, an innovative government financial intermediary, has played an important role in facilitating local borrowing by pro- viding loans to subnational governments that cannot access the private market and by discounting loans made by private banks. However, it has experienced some difficulties in recent years. For the central government, chronic fiscal imbalances have been a continued concern. In 2001 it implemented new laws to streamline the intergovernmental transfer regime and free up extra revenues to address its imbalances. Bogotá is the star of Colombian subnational borrowing. It has successfully issued domestic debt and in 2001 became the first and only Colombian city to issue an international bond. The city is in the rare position of having its credit rating constrained by the sovereign rating (which is dampened by political instability). More recently, macroeconomic instability has put a strain on Bogotá's financial position, but this does not detract from its history of fiscal prudence and several years of operating sur- pluses. Its record of competent management has secured its position as one of the strongest municipal borrowers in Latin America. Indeed, even in the face of refinancing pressures stemming from the short-term maturity of its obligations, the city has successfully managed the currency and interest rate risks on its outstanding debt. Country Case Studies: Colombia 281 Colombia is a unitary country with 43 million inhabitants, 75 percent of whom live in urban areas. Although beset by domes- tic turmoil, and despite the flagging world economy, the coun- try has generally turned in a good economic performance in the past few years. Inflation has been declining, dropping from nearly 17 percent in 1999 to 8 percent in 2001, but economic growth also has been slowing, from 3.4 percent in 1999 to 1.6 percent in 2001. In 2002, thanks to the tight monetary policy of the Central Bank, inflation remained relatively low (7 percent). However, the country had another year of slow economic growth (an estimated 1.6 percent), with adverse effects on public revenues and expenditures. Recent forecasts put the fiscal deficit for 2002 at 4 percent of GDP, well above the target of 2.6 percent. Two of the main sources of pressure on the national budget are growing military spending and a rise in pensions. President Alvaro Uribe Velez, elected in May 2002, focused his presidential campaign on fighting drugs, guerrillas, and paramilitaries, which explains the increase in military spending. The government plans to solve the pension problem with a national referen- dum to reform the pension system in 2003. In Colombia all levels of government, including subnational units, rely heavily on domestic debt markets to fund their deficits. In addition, the feder- al government has been working toward a domestic government bond market since 1995--with much success compared with other governments in the re- gion. However, Colombia's economic and political situation undermines in- vestor confidence, a situation exacerbated by the crises in other Latin Ameri- can countries, such as Argentina, Brazil, and República Bolivariana de Venezuela. As a result, the voluntary government bond market has remained closed since August 2002. The national government needs to restore public confidence, particularly investor confidence, because it is almost impossible for the government to forgo borrowing from the domestic debt market. Intergovernmental Relations Colombia's 1991 Constitution defines three types of subnational territories: departments (states), districts (municipalities with the status of depart- 282 Subnational Capital Markets in Developing Countries ments), and municipalities.1 The country has four districts: Bogotá (capital district), Barranquilla, Cartagena, and Santa Marta. The Constitution com- mits the central government to providing compensating resources when it imposes spending or service requirements on subnational governments. The central government has honored this commitment so far, but continued fis- cal imbalances could lead to exceptions or limits. Colombia's recently launched and still incomplete process of decentralization has led to prob- lems in maintaining fiscal balances at the national level because of resource transfers to subnational governments as well as problems in avoiding unsus- tainable deficits at the subnational level (Dillinger and Webb 1999). The decentralization was begun in 1983 several years after the military lost control of the government. An early landmark in the process was Law 78 of 1986, which required that mayors be elected by the people rather than appointed by the governors of departments. Similarly, the 1991 Con- stitution mandated that governors of departments be elected rather than appointed by the president. In addition, the Constitution committed the central government to expanding the revenue sharing system (situado fis- cal, or "situado") to ensure adequate provision of the services it is intended to support. Revenues Since 1983 departments have collected taxes on liquor, cigarettes, vehicles, and lottery sales; these taxes form the core of departments' own revenues. In addition, departments receive transfers from the central government through the revenue sharing system, established in 1971 by Law 46 to transfer 13 percent of the central government's ordinary revenues. The 1991 Constitution and Law 60 of 1993 expanded the revenue sharing sys- tem by adding the income, customs, and value-added taxes, increasing the system's share of the central government's revenues to 22.1 percent in 1993 and to 24.5 percent in 1996. Law 60 required that 15 percent of the shared revenues be distributed equally among the departments and the other 85 percent according to specific social indicators.2 The Congress is required to review this sharing formula every five years. In addition, under Colombian law all hydrocarbon royalties must be dis- tributed to subnational governments according to a formula directing 47.5 percent of royalties to producer departments, 12.5 percent to producer mu- nicipalities, and 8 percent to municipalities that are ports, with the other 32 percent redistributed across the country. The discovery of oil has in- creased the importance of this revenue source. Country Case Studies: Colombia 283 For municipalities, locally raised taxes cover about a third of expenditures in the aggregate. Wealthier municipalities raise more, poorer ones less. The primary sources of local tax revenue are the property tax and the business tax (on gross turnover). The many other local taxes tend to be unproductive. An interesting exception is the contribucion de valorization, a local betterment fee based on the user-pays or benefit principle. Some observers believe this tax could be used more extensively (Ahmed and Baer 1997). In the mid-1990s municipal spending was equal to about 6 percent of GDP, but municipal tax revenues were only about 2 percent of GDP. For most municipalities, then, transfers from the central government are criti- cal. Under Law 60, 60 percent of the shared revenues (participaciones munic- ipales) transferred by the central government to municipalities are to be dis- tributed according to the number of inhabitants with unsatisfied basic needs, and the other 40 percent according to such indicators as population size, administrative efficiency, and improvements in the quality of life. Transfers to municipalities were expected to grow until 2002. Thus as a result of the new arrangements introduced by the 1991 Con- stitution and Law 60, the central government has been transferring almost 47 percent of its total revenues to subnational governments. In addition, the rules mandate that any increase in its tax base must be shared with sub- national units. These heavy demands have caused continuing fiscal prob- lems for the central government. Expenditures Subnational governments have little autonomy in managing their expendi- tures. Consider this example in education: subnational governments are re- sponsible for paying teachers' salaries, but the size of the salaries is deter- mined through negotiations between the central government and the national teachers' union. As part of the ongoing decentralization process, the central government transferred responsibility for education, health care, and investments in water and sewerage to subnational governments in the 1990s. However, a lack of capacity to handle these services led to re- consideration of the transfers to some municipalities. After a review, the services were transferred only to departments and to some larger munici- palities with proven management capacities. The revenue-sharing system stipulates how resources transferred by the central government are to be spent. Departments are required to spend 60 percent of the revenues for education, 20 percent for health care, and the remaining 20 percent for other purposes. Municipalities must apply the 284 Subnational Capital Markets in Developing Countries transfers to basic education (30 percent); health (25 percent); water supply (20 percent); physical education (5 percent); and housing, welfare, debt ser- vice, and other uses (20 percent). The earmarking and tight rules have made it difficult for departments to balance their budgets. Part of the diffi- culty stems from the fact that the central government continues to set workers' wages and the terms of employment. For subnational governments, the budgetary inflexibility resulting from the earmarking of most of their revenues and the mandated spending linked to transfers can lead to unsustainable fiscal deficits, reflected in rising levels of debt. Only 12 departments achieved a fiscal surplus in 2000. Among the 20 that had fiscal imbalances, 9 had deficits exceeding 15 per- cent of their total revenue, and for Vichada the fiscal deficit was almost 150 percent of revenue (figure 16.1). Subnational governments also suffered se- rious effects from the country's slow economic growth in 1999 and 2000. Vichada Choco Caldas Nariño Tolima Caqueta Santander Norte Quindio Guainia Vaupes Risaralda Huila Sucre Guaviare Boyaca Cauca Putumayo Cundinamarca Atlantico Meta Guajira Amazonas Cordoba Cesar Magdalena Antioquia Casanare Bolivar Santander Valle deCauca SanAndres Arauca ­160 ­120 ­80 ­40 0 40 80 Percent Source: World Bank based on Colombian Ministry of Finance. Figure 16.1. Fiscal Balance as a Share of Total Revenue by Department, Colombia, 2000 Country Case Studies: Colombia 285 Regulatory Framework for Subnational Debt The earmarking of revenues and the centrally determined use of transfers encouraged growing use of credit in the early 1990s. Banks expanded their lending on the strength of the constitutional mandates to increase transfers to local governments. The weak reporting by and control over local governments and the ability to use intergovernmental revenue transfers to secure debt also loosened constraints on borrowing (Ahmad and Baer 1997). Subnational borrowing had been rare in the past, but during the 1990s subnational bank debt rose as a share of GDP--from 2.6 percent in 1991 to 4.6 percent in 1997, including indirect debt and the debt of subnational government-owned companies; direct debt in 1997 was 3 percent of GDP (Dillinger, Perry, and Webb 2001). Until 1997 the central government re- quired prior approval from the Ministry of Finance for any subnational borrowing. In 1997 a new law, Law 358, was enacted to curb the excessive use of credit by subnational governments. Under this law, called the "traffic light" law, the Ministry of Finance analyzes two indicators of indebtedness before approving subnational borrowing: · Capacity to pay, measured by the ratio of interest payments to oper- ating surplus (the operating surplus is defined as current revenues less fixed current expenses). · Sustainability of debt, measured by the ratio of debt outstanding to current revenues. Based on these indicators, a subnational government might be free to borrow or might face restrictions (table 16.1). On the supply side the Central Bank implemented various policies relat- ing to subnational borrowing in the past decade. Since 1999, however, it has tightened regulations, requiring that banks maintain capital reserves for the full amount of any loans to subnational governments with a "red light." This regulation has made the loans costly to lenders and thus to bor- rowers, supporting the effectiveness of the traffic-light system. In Colom- bia, unlike in Argentina and Brazil, the Central Bank has always been pro- hibited from lending to subnational governments. 286 Subnational Capital Markets in Developing Countries Table 16.1. The "Traffic Light" System for Regulating Subnational Borrowing, Colombia Rating Indicator Result Green Interest as % of operational savings less than 40% No restrictions on and debt stock/current revenues equal or less than 80% lending Yellow Interest as % of operational savings equal or greater Lending only with than 40% but less than 60% and debt stock/current Ministry of Finance's revenues equal or less than 80% authorization Red Interest as % of operational savings greater than 60% No lending, unless the or debt stock as % of current revenues greater than 80% subnational agrees to adjustment plan Source: Law 358 of 1997. Subnational Debt The growth in subnational debt was a direct consequence of the decentral- ization. The inflexibility in local expenditures made it difficult to adjust spending, resulting in fiscal imbalances. Also contributing to the growth in debt was the regulatory framework. The framework had been poorly de- fined until 1997 and the passage of Law 358, which introduced a stricter approach to regulating subnational borrowing. On the positive side, in Colombia, unlike in Argentina and Brazil, subnational governments did not own banks, so nontransparent lending practices were avoided. According to the Ministry of Finance, commercial banks account for more than 50 percent of total lending to subnational governments, and fi- nancial corporations account for more than 15 percent. Public and private banks lend to subnational governments at variable interest rates and re- quire that they pledge specific revenue sources to repay loans. Some gov- ernments have pledged shared revenues even though they had only limit- ed ability to use these earmarked resources for debt service. Subnational indebtedness grew during the second half of the 1990s, but the "traffic-light" controls under Law 358 appear to have put on the brakes (figure 16.2). The increase in indebtedness between 1996 and 1997 was the largest of the period in both relative terms (74 percent) and absolute terms (1.6 billion Colombian pesos [Co$]).3 Subnational borrowing then tight- ened, and despite the recession of 1999, subnational indebtedness as a share of Colombia's GDP remained stable from 1999 to 2001. Because of Country Case Studies: Colombia 287 Billions of Colombian pesos Percentage of GDP 8,000 4.0 7,000 3.5 6,760 6,000 3.0 6,047 5,000 2.5 5,151 4,000 4,287 2.0 3,832 3,000 1.5 2,000 2,196 1.0 1,000 0.5 0 0.0 1996 1997 1998 1999 2000 2001 Source: World Bank based on Colombian Ministry of Finance. Figure 16.2. Direct Subnational Debt, Colombia, 1996­2001 the limits established by the traffic light law, debt service requirements for most departments are less than 10 percent of their total revenue. In 2000 only two departments devoted more than 10 percent of their total spend- ing to interest payments (for Valle de Cauca the share was 42.10 percent, and for Arauca, 54.87 percent). Despite the pressures facing subnational governments, the central gov- ernment has not had to conduct comprehensive bailouts. During 1998 the departments of Valle de Cauca and Santander Norte were not servicing their debt. Valle de Cauca renegotiated the terms of its debt with the banks and at the end of that year reached a restructuring agreement with them. This mar- ket solution to a subnational debt problem differed from the approaches adopted in such Latin American countries as Argentina and Brazil. Findeter: A Financial Intermediary for Subnational Credit In 1989, under Law 57, the Colombian government created Findeter (Fi- nanciera de Desarrollo Territorial) as a second-tier financial institution to fi- 288 Subnational Capital Markets in Developing Countries nance or rediscount commercial bank loans made for municipal capital projects. The central government owns 92.53 percent of the company, and the departments own the remaining shares. Findeter has evolved from a municipal development fund that disbursed credit at subsidized rates to a bank that provides credit at market rates as well as technical advisory services. It also has improved its efficiency. In the past gaining access to funds took an average of 18 months, but in recent years Findeter has reduced the wait to approximately 6 to 8 months. While the in- stitution's original purpose was to lend to local governments, today it can serve a broader range of borrowers, including the private sector (table 16.2). Since its inception Findeter has provided credit, directly or indirectly, amounting to almost Co$3.7 trillion, allocated across a variety of uses (fig- ure 16.3). Subnational governments that cannot access the private credit market finance most of their projects through Findeter. The terms and con- ditions of the loans it provides differ substantially from those provided by commercial banks, because most bank loans are short to medium term while Findeter's are medium to long term (table 16.3). As a second-tier lender, Findeter rediscounts commercial bank loans for subnational governments for up to 100 percent of the loan. However, the commercial banks perform the financial and risk analysis and bear the full credit risk (a key factor in reducing moral hazard). Findeter thereby pro- vides a ready market for the loan but does not assume the credit risk of the counterpart commercial bank. Subnational governments have seen the relative interest rates on their loans decline. Two factors have contributed to this. Because most subna- Table 16.2. Potential Borrowers from Findeter Private sector Public sector Others Companies and individuals Departments NGOs dedicated to such activities as involved in education Districts local cultural activities Private companies that Municipalities provide public services Municipal associations Nongovernmental Metropolitan areas organizations (NGOs) Decentralized organisms involved in public services (not included in the federal budget, such as housing agencies) Source: World Bank based on Findeter. Country Case Studies: Colombia 289 Pipelines 30% Roads 23% Others Institutional 9% development 5% Telephone networks 5% Education Sewers 13% 15% Note: Underlying values are in 1999 Colombian pesos. Source: World Bank, based on Findeter data. Figure 16.3 Allocation of Credit from Findeter 1989­99 Table 16.3. Terms and Conditions of Findeter Loans Credit amount Up to 100 percent of the project cost Amortization Up to 12 years, including 3 years' grace for principal payments Up to 6 years, including 1 year's grace for principal payments for preinvestment projects Amortization system Quarterly Interest payments Quarterly Rediscounts Up to 100 percent of the credit Annual interest rate (rediscounts) Average fixed interest rate for bank certificates of deposit plus 2.5 percent Fees Surveillance: flat fee of 1 percent of the credit amount Commitment: annual fee of 0.75 percent of the undisbursed amount Source: Findeter. tional governments have pledged shared revenues for loan repayment and banks can intercept these revenues, subnational governments are seen as strong credits. Findeter has an outstanding track record in refinancing mu- nicipal loans, with only 2 percent of its loans nonperforming in 1996. Be- 290 Subnational Capital Markets in Developing Countries cause of its low rate of nonperforming loans during the past decade and its medium- and long-term investment perspective, Findeter was able to en- courage commercial banks to extend maturities and lend directly to local governments. Although the scheme in which commercial banks performed the finan- cial and risk analysis and retained the credit risk worked initially, recently Findeter has found it difficult to have commercial banks as intermediaries. Because of excess liquidity, the commercial banks' portfolio in Colombia declined by 8 percent between November 2000 and December 2001. Finde- ter's credit line now has to compete with commercial banks in the market. In the current economic cycle, with local governments already carrying ex- cessive debt and having no additional revenue streams to pledge, commer- cial banks are not finding adequate guarantees to act as intermediaries.4 Recent Developments In 2000 the Colombian Congress passed Law 617 to establish a regulatory framework for making fiscal adjustments at the subnational level. The main goal is to provide a long-term solution for subnational fiscal imbal- ances. The law was designed to free current revenues to fund operating ex- penditures fully and capital investments partially. The law sets specific lim- its, such as restricting personnel expenditures to no more than 50 percent of nonearmarked current revenues by 2004. In addition, it establishes a prohibition on funding current expenditures with debt and restricts short- term treasury borrowings. To encourage fiscal discipline, the central gov- ernment restructured more than Co$849 billion of subnational debt through the fiscal adjustment program. According to the law, the central government can provide guarantees for subnational governments if they agree to the following requirements: · Implement a fiscal adjustment program. · Reduce operating expenditures. · Adjust legislative expenditures to enable a reduction in the expendi- tures of legislators. · Reschedule debt to improve payment capacity. · Obtain new credits from banks to finance the fiscal adjustment program. In June 2001 a legislative act was approved requiring that the three types of intergovernmental transfers be combined in a new general partici- Country Case Studies: Colombia 291 pation system. The system became effective in January 2002. The funds will grow annually by the average annual percentage change in national current revenues in the previous four years. During a transition period (2002­08) transfers will grow by the rate of inflation plus 2 percent in 2002­05 and by the rate of inflation plus 2.5 percent in 2006­08. Since the new general participation system is still being tested, its results are uncertain. Nonetheless, the new scheme clearly allows the central gov- ernment to reduce its fiscal imbalances, since the system does not include the additional revenues that have resulted from the national tax reform in- troduced by Law 633. Thus subnational governments will benefit less than the central government. Capital District of Santa Fe de Bogotá: First Subnational Issuer in the International Bond Market Unlike many other municipalities, Bogotá sought to take full advantage of the greater opportunities offered by the decentralization process that began in Colombia in 1983.5 Decentralization gave residents the chance to choose local representatives through elections every three years, and it granted municipalities more independence to address the needs of their residents along with full responsibility for financial management. (A new law aimed at extending the term of elected municipal representatives and allowing greater flexibility in seeking reelection is to be implemented in 2004.) The Law of Urban Reform of 1989 was designed to help municipali- ties improve their operations through such mechanisms as expropriation, land banks, land readjustment, land improvement taxes, designation of priority areas for urban expansion, and transfer of construction and devel- opment rights. For both political and technical reasons, however, many Colombian municipalities never made full use of their ability to improve financial self- sufficiency through their own tax base, cost recovery policies, and other initiatives. Instead, they preferred to continue their dependence on manda- tory revenue sharing by the central government. For many of them the consequence was a precarious financial situation--a result of fluctuating transfers and large municipal debts, most of which are guaranteed by fu- ture national transfers. Bogotá is one of the few municipalities that undertook tax reform, sought out new revenue sources, reorganized and streamlined sectoral in- stitutions, and found ways to improve its operations to the point where it 292 Subnational Capital Markets in Developing Countries was able to successfully float local and international bond issues to cover some of its funding needs. Despite these positive actions, Bogotá's financial health deteriorated recently as a result of macroeconomic problems in the country that affected direct transfers as well as property values and business activity, two key factors in determining own revenue for Bogotá. Moreover, the city's revenue structure is incompatible with its growing financial needs. Recent reluctance by the city council to approve several new tax and cost-cutting initiatives has added to the problem. In addition, the capital investments completed in the past few years will demand greater current spending for operation and maintenance. Despite these dampening factors, the city's experience with bond issues illustrates its relative strength in municipal financing. In 2001 Bogotá sold US$100 million in bonds in the international market, becoming the first and so far only Colombian city to access that market. Despite the 2001 de- valuation of the Colombian peso, the city saw a good opportunity in the international market. It undertook the bond issue not only because of the sound financial condition it had achieved, but also as a marketing strategy to show itself to the world. While this was Bogotá's first time borrowing abroad, the city has a strong record in bond issues, having earlier launched 11 bond issues in the local bond market. Features of the Bond Issue The launch was very successful and obtained a low interest rate of 9.5 per- cent (table 16.4). The issue did not carry a sovereign guarantee. Given the uncertain situation in Colombia and the difficult straits of the emerging economies of Latin America, the market reception was gratifying. The city was able to issue the bond at a fixed interest rate, a very important feature because almost all of its debt has variable interest rates. Nonetheless, this bond clearly implies more currency risk exposure for the city. Bogotá's au- thorities are working to reduce the risk exposure. The Issuer Bogotá, the capital district of Colombia, had an estimated population of 6.6 million in 2001, 16 percent of the country's total, and occupies an area of 1,732 square kilometers. Administratively, the city is not part of the de- partment of Cundinamarca but has direct fiscal and political relationships with the central government. The city was granted its autonomous status, similar to that of a department or municipality, following constitutional re- forms in 1991. The city is a net contributor to Colombia's subnational sys- Country Case Studies: Colombia 293 Table 16.4. Features of the Bond Issue by the Capital District of Santa Fe de Bogotá Feature Details Date of issue 2001 Issuer Capital District of Santa Fe de Bogotá Currency U.S. dollar Amount US$100 million Maturity 2006 Amortization Bullet Interest rate 9.5 percent annually Interest periods Semiannual Market International bond market. The notes were issued under the U.S. Securities and Exchange Commission Rule 144A and Regulation S. Purpose Funding infrastructure projects. Status Direct, unconditional, unsecured, unsubordinated ranking pari passu with all obligations of the issuer. No sovereign guarantee. Covenants The district will not allow liens on its assets or revenues to secure any of its external indebtedness in the form of securities unless the notes are secured equally. Other covenants exist. Cross-default Failure to pay any public external indebtedness or external debt constituting guarantees of the district for amounts greater than US$20 million. Governing law State of New York Rating Fitch Ratings: BB+ (global) Standard & Poor's: BB (global) Source: World Bank based on Moody's and Standard & Poor's. tem: it accounts for more than 20 percent of GDP but for only a small per- centage of central government transfers. Bogotá owns eight independent companies that provide a wide range of services, including water, housing, energy, telephone, television, and mass transit. All are controlled by an independent board of directors but subject to budgetary oversight by the district. The mayor is elected for a three-year term and cannot be reelected to con- secutive terms. Council members are also elected for a three-year period. Economic Performance. As the country's capital and main financial center, the city is a major economic engine, contributing more than 20 percent of the country's GDP with less than 17 percent of its population. During the first half of the 1990s Bogotá's economy grew faster than that of the na- tion. During the second half of the decade, however, the gap between the national and city growth rates narrowed, and Bogotá could not avoid the economic recession that began in 1997. The city's per capita income is 50 percent higher than the national average. 294 Subnational Capital Markets in Developing Countries Manufacturing accounts for 16 percent of the city's economic activity, and finance and real estate account for almost 30 percent. Other services represent 27 percent (transport and communications 10 percent, construc- tion 7 percent, and trade 10 percent). Although exports do not play a key role in Bogotá's economic base, the city administration is committed to in- creasing exports by promoting agreements with the Cundinamarca depart- ment to improve transport, communications, and infrastructure. Bogotá is by far the largest urban center in the country. Since 1990 its pop- ulation has grown by almost a third. Immigration into the city is a concern for authorities because of the demands it imposes on infrastructure. Immigration also has an impact on labor indicators. In June 2001 the unemployment rate in Bogotá reached 18 percent, compared with a national rate of 15 percent. Financial Performance. The city has a positive track record of sound finan- cial and fiscal management, reflected in the string of operating surpluses it has achieved since the early 1990s. In addition, Bogotá has an aggressive investment plan to meet the needs of its growing population. However, the fall in revenues since 1997 has hampered implementation of the in- vestment plan and caused rescheduling and deferral of some projects. Current revenues reached their peak in 1998 at US$1.3 billion, while to- tal revenues reached their highest level in 1999 at US$1.9 billion (table 16.5). The economic recession that began in 1997 affected revenues, but the administration was able to cut some expenditures to offset the decline. National transfers to the city reached their peak in 2000, accounting for 35 percent of current revenues that year. In 2001, in response to the economic Table 16.5. Revenues and Expenditures, Capital District of Santa Fe de Bogotá, 1995­2001 (millions of U.S. dollars) Item 1995 1996 1997 1998 1999 2000 2001 Current revenues 770 1,058 1,271 1,274 1,108 995 1,033 Current expenditures 586 773 859 882 925 782 715 Operating balance 94 186 330 297 112 155 231 Capital revenues 85 254 223 234 820 554 613 Capital expenditures 336 584 702 743 961 769 570 Total revenues 854 1,312 1,493 1,509 1,928 1,548 1,646 Total expenditures 922 1,357 1,562 1,625 1,885 1,550 1,285 Fiscal balance ­158 ­144 ­150 Country Case Studies: Colombia 295 downturn, the central government took initiatives to reduce its transfers. Bogotá's capital revenues stem from dividends from its enterprises, income from financial assets, asset sales, and reductions in the capital maintained in certain companies. Bogotá has cut not only current spending but also capital spending to maintain a sound fiscal and financial position. In 1999 the city was planning to sell its telecommunications company (Empresa de Teléfonos de Bogotá, or ETB), but the financial crisis triggered by the Brazilian devaluation and the ripple effects on Latin American economies adversely affected the deal. Dur- ing the past decade the city financed most of its investments through a pay- as-you-go scheme, making it possible to maintain stable debt service levels. Projections for 2001 showed that the city would achieve an operating surplus for the eighth consecutive year and also enjoy a fiscal surplus that would allow it to make its debt payments (US$80 million) while saving money for further investments. Debt Profile. Because the city has not financed its capital investments through borrowing, its debt has remained sustainable. The composition of its debt stock changed during 1995­2001. The share of external debt in- creased, reaching a peak in 1999 (figure 16.4), but external debt declined in both relative and absolute terms in 2000 and remained stable in 2001. The external debt consists of a syndicated loan arranged in 1997, the bond is- sue, and multilateral loans. Bogotá's debt service payments have remained smooth in recent years as a result of its conservative debt policy (figure 16.5). The city estimates that its ratio of interest payments to operating surplus will peak in 2002, at 30 percent, and then decline to 24 percent by 2004, well below the 40 percent ceiling established by Law 358 of 1997 (the "traffic-light" law). City author- ities generally are more concerned about hedging interest rate risks than hedging currency risks because only 3.6 percent of the debt stock bears in- terest at fixed rates while almost 55 percent of the debt is denominated in Colombian pesos. The authorities are also taking refinancing risk into ac- count. More than 35 percent of the debt outstanding is due during 2002­04, and almost 90 percent is due during 2002­06. Accordingly, city officials are planning to refinance the debt by contracting loans with multi- lateral agencies and issuing bonds in the domestic market. Recent Developments As noted, the central government took several initiatives in 2001 and 2002 to limit its transfers to subnational governments. How great an impact the 296 Subnational Capital Markets in Developing Countries Millions of U.S. dollars Percentage of current revenue 400 70 378 350 369 361 60 300 320 310 310 302 50 250 261 281 295 244 40 200 183 30 150 External debt stock 20 100 114 Domestic debt stock 96 Debt stock as a percentage of 50 current revenue 10 0 1995 1996 1997 1998 1999 2000 2001 Source: World Bank based on Fitch Ratings. Figure 16.4. Debt Stock, Capital District of Santa Fe de Bogotá, 1995­2001 Millions of U.S. dollars 180 Debt service 160 Interest payments 140 Principal payments 120 100 80 60 40 20 0 1995 1996 1997 1998 1999 2000 2001 Source: World Bank based on Fitch Ratings. Figure 16.5. Debt Service, Capital District of Santa Fe de Bogotá, 1995­2001 Country Case Studies: Colombia 297 cut in transfers will have on Bogotá's finances remains unclear. Conserva- tive estimates forecast losses of Co$644 billion for Bogotá during the transi- tion period (2002­08) for the new regime established by the national gov- ernment. This figure represents a 14 percent decline in transfers during the full transition period. Beyond the changes under the new transfer scheme, Bogotá has seen a downward trend since 1997 in the national transfers it has received, in both relative and absolute terms (for example, shared revenues covered half of the city's education expenditures in 1990 but only 38 percent in 2001). Because the city must rely increasingly on its own revenues, author- ities are committed to improving tax collections, cutting certain costs (in 2001, for example, the city eliminated 4,058 permanent positions), and ex- panding the tax base. Credit Ratings The U.S. dollar bond issue was globally rated by Standard & Poor's (BB) and Fitch Ratings (BB+). Their analyses reflected concerns about the economic recession and violence affecting the country at the time of the issue, but both agencies concurred that the city has shown a great commitment to maintaining prudent financial management. According to the rating analyses, the ratings took into account the fol- lowing positive factors:6 · Satisfactory fiscal operations. · Proven ability to manage severe economic downturns. · Manageable and affordable debt position. · Valuable assets. · The city's status as Colombia's main economic center. The ratings also reflected some negative factors: · A weak local economy that suffered the effects of the national eco- nomic recession. · Significant pension liabilities. · The potential adverse effects of the reform of the government trans- fer system. · Country risk. · The increasing service needs of a growing population. 298 Subnational Capital Markets in Developing Countries According to one of the rating agencies, Bogotá deserves an investment- grade rating. However, because the sovereign's rating imposes the ceiling, the country's weak financial situation and its macroeconomic conditions undermined the credit status of the city. The devaluation of the Colombian peso, the economic recession, and the violence in the country all had ad- verse effects on the issue. In recent years the international bond market for emerging market economies has been very volatile, and several Latin American economies have experienced financial crises and problems that have affected the en- tire region. Nonetheless, Bogotá was able to launch a successful bond issue. The transaction suggests that even in the face of trying national and re- gional conditions, strong subnational borrowers can gain access to the market. The city administration's reputation was a major factor in the success of the issue. Even after several years of economic recession the city's manage- ment was able to cut spending and boost revenues to offset the decline not only in its own revenues but also in those received from the central gov- ernment. The city's experience in the domestic bond market helped it pre- pare the bond offering. In addition, its financing policy for capital expendi- ture helped maintain relatively low levels of debt, considered a positive factor at the time of the issue. Its comprehensive investment plan and maintenance of valuable assets also were viewed positively by investors. Notes This chapter relies on information provided by the World Bank, Fitch Rat- ings Colombia, Standard & Poor's, the Colombian Securities and Exchange Commission, and the Colombian Ministry of Economy and Public Credit. 1. Departments are the main territorial divisions of Colombia. They were created in 1831, when the country was divided into five departments: Cundinamarca, Boyacá, Magdalena, Cauca, and Itsmo. 2. The number of students enrolled, the number of school-age children not attending school, the number of patients seen by health units, and the number of potential patients based on population. 3. In 1997 inflation (based on the consumer price index) was 18.5 per- cent. 4. World Bank consultant reports as reported in www.findeter.gov.co. 5. Much of this introductory section draws from World Bank sources. 6. Based on credit reports by Fitch Ratings and Standard & Poor's. Chapter 17 Latin American and the Caribbean Mexico Using credit ratings can be an effective means of instilling a culture of creditworthiness. Steven Hochman and Miguel Valadez Lessons Mexico has traditionally been a highly centralized state, with the states and local governments having centrally assigned duties and limited fiscal autonomy. Except for the local property tax, local revenue options are limited, and the states especially are heavily dependent on federal transfers. However, reforms in re- cent years are improving financial flexibility at the municipal level and increasing capacity to borrow in private markets. In the 1990s Mexico's federal government inadvertently in- volved itself in the decisionmaking for subnational borrowing through pledged transfers and the implicit guarantee of local government bailouts that came with them. Accordingly, credi- tors took little time to conduct thorough evaluations of subna- tional finances, and some local governments borrowed beyond their means. The 1994­95 financial crisis exposed these defi- ciencies and necessitated a costly federal bailout program that forced a rethinking of subnational lending parameters. 299 300 Subnational Capital Markets in Developing Countries To avoid a recurrence of the fiscal indiscipline and to remove it- self from the local lending equation, the Mexican government instituted reforms that induced subnational governments to ac- quire internationally recognized credit ratings. The mandate placed the onus on banks, and thus subnational borrowers, by requiring that loans be supported by risk-weighted reserves that raised the cost of borrowing. Loans without credit ratings were assigned the highest reserve ratio. In addition, further reform to the intergovernmental transfer regime added clarity to local fi- nances. These positive steps, particularly the institution of credit rating requirements, have sparked the beginnings of a credit rating culture and spurred a nascent domestic capital market for sub- national debt. Indeed, subnational governments have discov- ered that they can finance large projects more cheaply through bond issues than through bank loans. To encourage prudent lo- cal borrowing, the government has created a conservative trust fund structure for local government debt issues, a structure that is viewed favorably by international credit rating agencies and has boosted the ratings for several issues. As the case study shows, the strong mechanisms inherent in the trust fund raise the certainty of repayment and lower risks. Thus despite the remaining institutional deficiencies in intergov- ernmental relations and judicial processes, a borrowing frame- work that demonstrates a political will to repay has allowed a vi- able market for subnational debt to begin to operate. Mexico is the world's thirteenth largest economy, eighth largest exporter of goods and services, and fourth largest producer of oil. Far-reaching stabiliza- tion and structural reform efforts since the late 1980s have been rapidly transforming the Mexican economy and putting it on a faster growth track. Despite the massive setback from the 1994­95 financial crisis, the economy grew by an average of nearly 3 percent a year in the 1990s after virtually stagnating in the 1980s. The initially export-led recovery after the 1994­95 financial crisis has brought the economic growth trend close to 5 percent. Country Case Studies: Mexico 301 Mexico has benefited from its increasing integration with the North Ameri- can economy, especially that of the United States. Trade liberalization, par- ticularly through the North American Free Trade Agreement, has clearly con- tributed to Mexico's rapid economic transformation.1 Decentralization In the past two decades the relationship between the federal and subna- tional governments in Mexico has changed significantly. The enactment of the Fiscal Coordination Law in 1980, the decentralization of public services initiated in 1992, the financial bailouts of states and municipalities in 1995 and 1997, and, most recently, the introduction of credit ratings as a factor in obtaining loans all have reshaped the institutional framework. For decades Mexico has been constitutionally a federation. However, until the 1980s there had been a trend of increasing centralization (see Giu- gale and others 2001). This trend has been reversed most noticeably since the mid-1990s, when the country began devolving significant spending re- sponsibilities to the local level. Nonetheless, the federal government still dominates the fiscal landscape, raising about 94 percent of all revenues and accounting for about 70 percent of all direct spending in the country (table 17.1). As a result, the states and municipalities rely greatly on transfer pay- ments from the central government. Table 17.1. Spending and Own-Source Revenues as a Share of GDP by Level of Government, Mexico, Selected Years, 1991­97 (percent) 1991 1994 1997 Own spending Federal 8.4 11.5 11.5 State 3.0 3.2 4.9a Municipal -- 1.2 -- Own revenues Federal -- 16.9 15.8 State -- 0.2 1.0a Municipal -- 0.3 -- -- Not available. a. Data are for state and municipal governments combined. Sources: Giugale and others 2001; Amieva-Huerta 1997. 302 Subnational Capital Markets in Developing Countries Even as Mexico progresses toward greater local autonomy, fiscal respon- sibility, and accountability, there continues to be a noticeable lack of insti- tutional elements to strengthen the state and local sector. Mexico has 31 states in addition to the federal district and more than 2,400 municipali- ties. These range widely in skills and resources. Many lack training pro- grams, reliable information systems, agencies for coordination, and a legal and statutory framework (Giugale and others 2001). Without these, devo- lution remains immature and fragile. Decentralization policies have ap- plied almost exclusively to the state level. Most municipalities received few new responsibilities even after the reforms of 1998, when a large share of the new resources allocated to municipalities were directed to federally mandated expenditures. Meanwhile, large municipalities take on many critical tasks without additional funding from the center. Moreover, despite the increased devolution of spending, effective decen- tralization for Mexican states decreased throughout the 1990s. On average, states receive 85­95 percent of their revenues from federal transfers. Most transfers from the central government to the states are earmarked for specific purposes, typically to finance federally mandated employees in municipali- ties or for matching grants programs. The center still mandates how states are to fulfill their fiscal obligations, which is inconsistent with states' increasing political and economic power. In addition, the fiscal transfer regime is seen as too complex and opaque, based on historical inputs rather than performance or caseloads, and subject to political manipulation.2 Until the late 1990s states rarely had a clear picture of how much funding they would receive, and the discretionary nature of transfers discouraged efficiency. Revenues and Responsibilities: The Fiscal Coordination Law of 1980 Mexico's Fiscal Coordination Law provides for a revenue-sharing system in which all states and municipalities participate. This system enables states and municipalities to receive a share of the federal revenue collected from various sources, the most important being the value added tax and oil rev- enues. About 20 percent of the federal revenue collected goes into the Gen- eral Fund for Shared Revenues (Fondo General de Participaciones), which is distributed to the states under a formula that takes into account popula- tion, the collection effort for certain taxes (impuestos asignables), and a compensatory mechanism that effectively subsidizes poorer states. While states may spend these shared federal revenues, called participaciones, as they please, they must pass on to their municipalities at least a fifth of the shared revenues they receive. Country Case Studies: Mexico 303 Decentralization initiatives have led to a notable increase since 1995 in spending responsibilities shared by the states and the federal government, particularly in health and education. Expenses in these areas are covered primarily by specifically appropriated funds, or aportaciones. State responsi- bilities also include administration, state infrastructure, and security, while water supply and treatment are often municipal responsibilities. In addi- tion, the federal government provides discretionary financing for invest- ment in basic infrastructure programs. There are important distinctions among these three categories of funding. Shared revenues are a recurrent revenue source but subject to fluctuation with the level of tax collections. From the perspective of state governments, they represent a flexible resource that may be used for any purpose. Appropriated funds are also recurrent but are subject to yearly appropriations. Because the use of appropriated funds is federally determined, they are a less flexible revenue source for states than are shared revenues. Discretionary financing is nonrecur- ring, and the amount available for a state depends on the effectiveness of its lobbying efforts. These resources are the least flexible for states, since the funds are directed to specific projects and often must be matched by state funds. The most important tax levied by states is the payroll tax, while the most important one for municipalities is the property tax. States and mu- nicipalities also collect fees and user charges and earn interest income from financial investments. In addition, state governments use federal taxes col- lected at the state level that are transferred fully (such as taxes on new ve- hicle registrations) or partially (such as those on alcoholic beverages); there are also significant state taxes in northern border states (on cross-border ac- tivity, such as trade and tourism) and oil-producing states (on oil). Despite the gains in financial autonomy, states and municipalities con- tinue to have only weak revenue-raising powers, with few revenue sources, low rates, poor financial record keeping, and inefficient revenue collection procedures. Municipalities have traditionally relied on the property tax, which is levied at low rates and often (along with user charges) subject to relief.3 Nonetheless, in the larger municipalities property taxes can account for 20 to 40 percent of revenue.4 The potential exists for better collections, particularly for property taxes and some excises and user charges, if the po- litical will for reform can be mustered. Overview of Subnational Borrowing Until recently subnational borrowing was a product of Mexico's top-down intergovernmental relations, with the central government largely setting 304 Subnational Capital Markets in Developing Countries the rules and making the decisions--on an ad hoc basis and through nego- tiations between players belonging to the same political party (Giugale and others 2001). However, the increased political competition in Mexico and the devolution and greater subnational autonomy led to a need for stricter and more transparent rules for governing subnational borrowing. In 2000 the Mexican Treasury promulgated a new subnational borrowing frame- work. The new regulations eliminated discretionary federal transfers, re- quired lending institutions to adopt prudential limits and get risk assess- ments (ratings) on state debt, and provided incentives for regular financial reporting by states and municipalities. As Mexico moved toward greater decentralization, its total subnational debt doubled between 1994 and 1998, but the debt is concentrated in a few subnational entities. Three states--the Federal District, Mexico, and Nuevo León--together account for 65 percent of the outstanding debt (figure 17.1). The Problem of Indiscipline: Credit Markets and Debt before and after the 1994­95 Financial Crisis Federal transfers--general revenue-sharing funds and specifically appropri- ated funds--typically represent roughly 90 percent of total revenue for state governments and perhaps 70 percent or more for all but the most 1994 Nuevo León 1995 Federal District State of Mexico 1996 1997 1998 0 5,000 10,000 15,000 20,000 25,000 Millions of pesos Source: World Bank. Figure 17.1. Borrowing by Three State Governments, Mexico, 1994-98 Country Case Studies: Mexico 305 property-rich municipal governments. Subnational governments typically finance their major capital spending requirements through bank loans, both from commercial banks and from the government development bank, Banobras. The Mexican Constitution prohibits state and municipal governments from borrowing from foreign sources or in foreign currency. Denied access to international credit markets--markets that customarily insist on credit ratings as a lending requirement--Mexican governments that borrowed had little incentive until recently to seek independent evalu- ations of their credit standing. Federal Revenue-Sharing Funds as Collateral Bank loans to Mexican states and municipalities have generally included a collateral pledge of the borrower's federal revenue-sharing funds as a debt guarantee. Lenders and borrowers alike viewed the involvement of the fed- eral government in the process as an implicit guarantee, a perception that led some states and municipalities to borrow beyond their means and banks to lend without proper credit assessments. These factors exacerbated the financial turmoil experienced by most state governments in the fiscal crisis of 1995. The importance of federal revenue sharing for state revenues and its rel- ative reliability as a revenue source contributed to its use as collateral for state borrowing. The institutional arrangement supporting this practice was based on article 9 of the Fiscal Coordination Law, which authorized lenders, in the event of nonpayment by a state or municipality, to direct the federal government to deduct pledged shared revenues from state rev- enues and use them to pay the overdue debt service. When a government failed to pay its debt, the lender invoked the collateral pledge and inter- cepted that government's flow of federal funds. That left some govern- ments with insufficient funds to pay for essential services. As a result, they sought additional financial support from the federal government, which was forced to come to their rescue. In the most recent rescues of note, mounted in 1995­97 in response to the fiscal crisis following the Mexican peso devaluation of late 1994, all states were bailed out. Many states have since refrained from borrowing, but a few have borrowed heavily. Subnational Debt Profile in the Mid-1990s At the end of 1994 subnational (state and municipal) debt in Mexico to- taled 25 billion pesos, an amount equivalent to 72.7 percent of the shared revenues received by the states that year. More than half the debt was at- 306 Subnational Capital Markets in Developing Countries tributable to borrowers in four major states that had high ratios of debt to shared revenues: Sonora (254 percent), Nuevo León (125 percent), Jalisco (116 percent), and Mexico State (115 percent). While municipalities and smaller states tended to have lower ratios of debt to shared revenues, some had similarly high ones: Querétaro (215 percent), Quintana Roo (136 per- cent), Baja California Sur (121 percent), and Campeche (101 percent). Even among states with low ratios of debt to shared revenues, some were vulnerable because they had very short debt maturities: San Luis Potosí (2.7 years), Durango (3.8), Chihuahua (4.4), and Colima (4.7 years). The aver- age debt maturity at the end of 1994 was only 6.6 years. Assuming a con- stant payment schedule, subnational governments would have had to de- vote more than 11 percent of their annual shared revenues on average just to cover their principal payments. Further complicating the debt profile, nearly all the debt carried floating interest rates, leaving states and munici- palities with sizable interest rate exposure. Subnational Debt Relief Programs, 1995­98 This debt profile points to a high degree of vulnerability. Adverse develop- ments in late 1994 that persisted through 1997 created a situation that made debt payments unsustainable. On 20 December 1994 the Mexican peso was devalued as the exchange rate against the U.S. dollar went into freefall, sending the nation into a fiscal crisis. Short-term interest rates rose sharply in 1995, peaking at nearly 75 percent in April. An economic crisis caused federal tax revenues--and thereby the pool of shareable revenues-- to contract sharply; inflation--and thereby the cost of providing govern- ment services--rose rapidly. During 1995 most states and many municipalities, including some with relatively little debt, missed principal or interest payments or both. In some cases the default period lasted only a few weeks; in others it extended over a year. The defaults resulted from the combination of heavy debt, shrinking revenues, and soaring interest payments. Some also may have occurred in part because of a belief that the federal government would step in and pro- vide financial assistance. In late 1995 the federal government put together the first of two debt re- lief programs for states and municipalities. It offered the program to all sub- national governments regardless of their level of debt. Most states and many municipalities joined the program, which involved converting old debt into a new, inflation-adjusted unit of account (Unidad de Inversión, or UDI) that carried fixed interest rates and extending debt maturities. This arrangement Country Case Studies: Mexico 307 spread debt service payments over a longer period, though at the expense of increasing the peso cost of the debt (because UDIs were adjusted for infla- tion). In return, state and municipal governments agreed to restore fiscal dis- cipline, increase transparency, and improve their financial reporting. In addition, the federal government provided direct financial assistance to many states. The amount of this aid, often earmarked for debt payment, varied with states' financial need. In 1998 the federal government spon- sored a second debt relief program that lowered the interest rates charged on UDIs and further extended debt maturities, up to 18 years. As a result of the debt relief programs and the better revenue perfor- mance after the crisis, the debt profile of Mexico's subnational govern- ments improved substantially. The average ratio of debt to shared revenues declined from 72.7 percent in 1994 to just 38.7 percent by the end of 2001. The average debt maturity almost doubled, from 6.6 years before the finan- cial crisis to 12.5 years in 2001. Reform of Financial Legislation: Reasons, Objectives, and Preliminary Outcomes To avoid a need for rescues of subnational borrowers in the future, the fed- eral government searched for a way to accomplish the following: · Encourage banks to give greater weight to the evaluation of intrinsic credit factors in their decisions on lending to state and local govern- ments. · Give state and local governments added incentives to keep their fi- nances in order and avoid excessive borrowing. · Reduce the likelihood of financial problems arising at the state and local levels that would require federal intervention. Collateral Procedures Since the end of the debt relief program in 1998 the federal government has put in place several reforms aimed at preventing a need for new bailouts. As a first step it modified article 9 of the Fiscal Coordination Law, ending a bank's ability to request a direct transfer from the federal Treasury of a state or municipal government's shared revenues. The aim was to re- duce the federal government's involvement in the credit relationship be- tween lenders and government borrowers. 308 Subnational Capital Markets in Developing Countries Instead, state governments and the banks were to determine beforehand what collateral procedures would apply if arrears emerged. However, the new arrangement did not work as expected. Commercial banks, reluctant to participate, curtailed their lending to state governments and municipali- ties. As a transitional mechanism, the federal government accepted a tem- porary "mandate" from the states to transfer pledged shared revenues, a modified version of the original scheme that did not remove the federal government from the process. In late 1999 the federal government notified states and municipalities that it would terminate the mandate arrangement in March 2000 and an- nounced its intention to develop a new mechanism that would minimize the federal government's role. The mechanism, a master trust agreement (Fideicomiso Maestro), would enable subnational governments to use their shared revenues as debt collateral by channeling a share of these funds di- rectly to the trust. Subnational Credit Ratings In December 1999 Mexico's Treasury introduced new bank regulations, the latest in a series of steps to enhance transparency in credit and capital mar- kets and encourage state and local governments to assume greater responsi- bility for their own affairs. The regulations, which took effect in April 2000, require that a bank lending to a state or local government set aside capital reserves according to the risk-weighted credit exposure represented by the loan. Independently issued credit ratings serve as the measure of risk. The new regulations relate each state or local government's credit rating to that of the federal government and require banks to set aside reserves deter- mined by the rating gap that results. The larger the gap, the higher the cap- italization requirement. The regulations do not require state or local governments to obtain credit ratings. However, borrowers without a rating are penalized, since banks must apply the highest capital reserves--and in all likelihood will charge the highest interest rates--for these loans. The use of ratings was in- tended to encourage banks to give greater weight to credit factors in their lending decisions and to give state and local governments added incentives to keep their finances in order and reduce the likelihood of a new federal bailout. Since the new regulations took effect, most states and many municipali- ties have obtained credit ratings. By late May 2002 all but three of Mexico's 32 federal entities (31 states and the Federal District) had been assigned Country Case Studies: Mexico 309 credit ratings by at least one and, in most cases, two internationally recog- nized rating agencies (one state has three ratings). Some cities also have been assigned ratings, and these reveal important differences in creditwor- thiness between state and municipal governments. The three agencies as- signing the ratings are Moody's Investors Service, Standard & Poor's, and Fitch Ratings. A handful of negotiable debt offerings--certificates whose payment re- lies on state or municipal financial backing--also have been rated. As gov- ernments have grown increasingly aware that, for a large project, a certifi- cate issue can offer lower interest costs than a bank loan, more certificate offerings are being prepared. A capital market for state and municipal gov- ernments is developing in Mexico. Since the initial assignment of ratings, some have been raised and others lowered. When a rating is assigned or changed, the rating agencies publish press releases or reports explaining what factors support the rating and what trends may affect the rating in the future. Using these explanations and other data, some market observers have published predictions of future rating assignments for issuers not yet rated. Growing numbers of subnational governments are submitting their fi- nancial statements to independent audits. State and municipal finance offi- cials--and lenders--are developing the habit of asking, "What can be done to improve this rating?" or "If we borrow this much more, or if we take these steps involving government finances or debt, how would that affect the rating?" These are signs that a new credit culture is developing among state and municipal governments in Mexico. While it is still early, it appears that the Mexican government's goals in requiring credit ratings--promoting a new credit culture and removing the federal government from the credit relationship between state and local governments and their lenders--are being realized to an extent beyond some of the most optimistic expectations. Recent Subnational Borrowing Experience, 2000­02 The chief federal restrictions on subnational borrowing in Mexico are the ban on foreign currency loans and the requirement that the proceeds of borrowing be used solely for capital investment.5 State-enacted debt laws also regulate state and municipal borrowing, requiring approval by the state congress for state borrowing in most cases and establishing parame- ters for short-term borrowing. Municipal borrowing typically requires only 310 Subnational Capital Markets in Developing Countries local legislative approval if the loan is payable within the term of the bor- rowing administration, but longer-term debt issuance requires both munic- ipal and state approval. The recent legislation to improve financial and, by extension, subna- tional borrowing mechanisms has led to the first local government bond is- sues. Three subnational entities have issued debt under the master trust fund structure that the federal government proposed in 2000 (table 17.2). The Aguascalientes and San Pedro bond issues received ratings on a par with the national rating and are direct, fully binding obligations of the ju- risdictions. The issues were assigned a comparatively high rating for two main reasons. Both municipalities have relatively large own-source rev- enues (San Pedro's are among the largest in Mexico, and Aguascalientes has robust property tax revenue). Even more crucial, however, is the secure structure provided by the trust fund arrangement. For both municipal bond issues, the trustee of the fund is given rights to 100 percent of the municipality's shared revenues from the federal govern- ment, and all these revenues are pledged so that they can be used as a guar- antee for issue repayment. Legal provisions add further security. The state government, which distributes shared revenues to municipalities, is con- tractually obligated to redirect the funds from the municipal treasury to the trust fund. The Fiscal Coordination Law reinforces the obligation for Table 17.2. Subnational Bond Issues, Mexico, 2002 Amount Date of Type of (millions Term Entity issue instrument Rating of pesos) (years) Municipality of Certificado Bursatil Aguascalientes 11 December (Capital Market Moody's: Aaamx; (Aguascalientes) 2002 Certificate) S&P: AAA 90 5 State of Morelos 11 December Certificates of Fitch: AA+(mex); 2002 Participation Moody's: Aa2mx 216 7 Municipality of Certificado Bursatil San Pedro 24 July (Capital Market Fitch: AAA (mex); (Nuevo León) 2002 Certificate) Moody's; Aaamx 110 7 Total 416 Source: Serrano Castro 2002. Country Case Studies: Mexico 311 timely revenue transfers. In addition, both the Nuevo León and Aguas- calientes state governments have a history of good fiscal health and timely payment of shared revenues. Moreover, any modification to the trust's rights to shared revenues must be approved by all creditors under the trust. Both issues state that additional debt can be acquired only if debt and debt service limits have not been reached (these vary with the jurisdiction) and that the new debt must follow the same trust fund structure. Added security for all three issues in any events that threaten the repay- ment schedule is provided by a trigger for advance trapping of cash for the trust (for San Pedro and Aguascalientes, at 1.5 times the monthly amount required in the repayment accounts). Bondholders can respond to serious threats to their security (such as attempts to invalidate the trust contract or provide false information) by appropriating the full share of shared rev- enues allowed by the bond contract to accelerate full repayment. For all three bond issues, cross-collateralized reserve funds limit the risk of nonpayment due to revenue shortfalls. One of the more interesting dif- ferences among the issues is the payment structure. Aguascalientes uses a bullet structure that pays periodic interest until the maturity, when the full principal is paid. Both Morelos and San Pedro use amortized structures with a three-year grace period on principal. A more significant difference relates to the Morelos issue. This issue stands out not only because it does not specify debt limits but because it has a lower rating--in part because the trust for this issue does not have access to 100 percent of shared rev- enues even in the event of nonpayment. The maximum that the trust can request is 30 percent of the revenues pledged to and received by the state's master trust fund. That amounts to 16.4 percent of the state's shared rev- enues. Accordingly, the repayment contingencies for the Morelos issue are somewhat weaker and, with a two-tier trust fund structure, subject to com- peting financing needs. Nonetheless, nonpayment risk is quite low because the legal structure commits 30 percent of the revenue of the master trust fund to the issue offering. Moreover, the contract cannot be changed with- out the approval of senior lien creditors and the state congress. This con- trasts with the other two issues, for which proposed changes require unani- mous creditor approval. Significantly, the purpose of the San Pedro and Morelos issues is to refinance or retire outstanding loan obligations rather than provide direct project financing. These examples of local borrowing in Mexico show that even with defi- ciencies in enforcement and institutional development, subnational bor- 312 Subnational Capital Markets in Developing Countries rowing in the bond markets is possible. What is needed is a credible pay- ment mechanism that demonstrates the political will to ensure timely and complete repayment of debt obligations. Notes 1. As a reflection of the opening of the economy, Mexico's trade (im- ports plus exports) as a share of GDP tripled between 1980 and 2000, reach- ing 40 percent. 2. At least until 1999 politically favored states were able to receive ad hoc transfers that thwarted the incentives to manage well and enhance lo- cal revenue (Giugale and others 2001). 3. The effective property tax rate in the mid-1990s was estimated to range between 0.03 and 0.05 percent. Rural areas are taxed at half the effec- tive rate. The rates are grossly inadequate, but authorities do not want to deal with the political problems of raising them. See Amieva-Huerta (1997, p. 575). 4. Cities have been given greater ability to control land use and to deter- mine property tax values and rates under recent constitutional amend- ments (article 115, approved in 1999), powers traditionally exercised by the states. But property tax rates are still subject to state approval. The reforms are seen as providing the larger cities more revenue raising power and flexi- bility and more discretion over revenues. See Aldrete-Sanchez (2000). 5. This section is based in large part on Moody's Investors Service rating reports for the relevant municipal bond issues. Chapter 18 Sub-Saharan Africa South Africa Despite sophisticated financial markets, the country is slow to reinvigorate its municipal bond market amid rapid changes in its political and fiscal structure. Matthew Glaser and Roland White Lessons In marked contrast to other countries of Sub-Saharan Africa, South Africa has a sophisticated private financial market. Mu- nicipal borrowing--through bonds and from intermediaries-- has been a feature of local government funding for years, though before the early 1990s such borrowing was implicitly or explicitly guaranteed by the state. Aggregate lending volumes have stagnated and declined in recent years, however, primarily as a result of the interplay between a deficient policy and regu- latory framework and poor budget discipline and financial man- agement practices in local governments. South Africa has taken measures to address these deficiencies. However, these measures, combined with ongoing reforms in the organization of the local government system (such as changes in boundaries), have led to a lack of stability, creating an uninviting investment environment for private lenders. 313 314 Subnational Capital Markets in Developing Countries Today, prospects for growth in municipal borrowing are uncer- tain even though the potential demand for loan finance far ex- ceeds existing volumes. South Africa has many of the basic conditions for expanded local borrowing, including a sophisti- cated and liquid financial sector, local authorities with substan- tial fiscal capacity, and a sound policy and legal framework that is soon to be introduced. Other factors, however, such as weak- nesses in budgetary and financial management and the nature of certain structural and regulatory reforms, militate against growth in local borrowing. The next five years will be critical in determining the long-term outlook for municipal borrowing in South Africa. South Africa is unique in many ways. It has the institutions and policies that many countries seek. It has efficient and vital capital markets for na- tional government, public enterprise, and corporate bonds. It has substan- tial experience with municipal securities, a large and liquid financial sector eager to lend to municipalities, and clearly stated, market-oriented policies on the verge of being enacted into law. Despite these strengths, some key indicators are headed the wrong way. The number of lenders to municipalities is shrinking. Private lending to municipalities is stagnant, and the government-owned lender is actively competing for the business of large and creditworthy municipalities while the market's structure is becoming steadily desecuritized. All this points to a need for clear, stable, and effective legal and finan- cial arrangements within which municipalities can plan. The disruptions of the post-apartheid transition period since 1994 have been unavoidable, and South Africa has managed this transition fairly well. Long-term bor- rowing and lending, however, depend on long-term predictability. Succes- sive changes in municipal borders, powers, and functions have made it dif- ficult for municipalities or investors to anticipate the future. These changes have made municipal borrowing expensive and have caused many private lenders to withdraw from the market, at least until conditions stabilize. Clear remedies for defaults have not yet filled the vacuum created by the disappearance of implicit government guarantees. With the finalization of Country Case Studies: South Africa 315 legislation expected during 2003, the new municipalities and their legal framework will be settled. The municipalities then must be allowed time to find their footing. Any further uncertainty could prevent municipalities from obtaining long-term credit for infrastructure. Efforts are also needed at the municipal level. If the supply of "bank- able" projects and municipal debt securities is to grow significantly, munic- ipalities must develop the basic skills and experience in accounting, plan- ning, reporting, and marketing that support wise borrowing choices. Demand for credit should be the natural consequence of careful and in- formed municipal capital planning. Borrowing may be the most powerful tool in a municipality's financial toolbox, for it can lay the foundation for economic development and a virtuous cycle of growth. If used unwisely, however, it can leave crippling debt for the next generation (box 18.1). Once the legal framework is finalized and stabilized and the basic skills and experience are developed, investors will have no reason not to come to the table. Assuming that borrowing does expand, financial crises will even- tually occur in some municipalities, as they do in any country. How well South Africa deals with these crises will indicate the likelihood of long- term success. If sound financial emergency mechanisms are in place and if they prove effective, some diminution in loan volumes still may occur, but that would probably be followed by a resumption of steady growth in South Africa's municipal credit market. Local Government. Before the advent of democracy in 1994 South Africa had a variety of local government systems, with about 1,300 municipalities throughout the country. In urban areas separate white and black local au- thorities were subsidiary to the four provinces that then existed. White lo- cal authorities included the core cities and virtually all the economic activ- ity. These local authorities had their own councils, staff, and revenue sources, including property taxes and revenue from utility services. Black local authorities, often located nearby and providing cheap labor for the core cities, had limited services and widespread poverty. Other areas where black people lived were included in "independent homelands" and "self- governing territories," where a variety of administrative and traditional au- thorities provided local governance and limited services. Since 1994 the legal and financial underpinnings of municipalities have undergone a series of changes aimed at democratizing and deracializing mu- nicipalities. In 1994 the "homelands" were reincorporated into the republic, and negotiations were initiated to determine local government boundaries. This process resulted in the creation of 843 municipalities after local govern- 316 Subnational Capital Markets in Developing Countries Box 18.1. A South African Parable Not long ago a district councilor asked his municipality's ac- counting firm for help. Collections of budgeted revenues had been falling steadily, while expenditures and responsibilities for providing services had increased with the addition of new terri- tory. Every month the municipality was spending more money than was coming in. Unpaid suppliers threatened to withhold services, and local banks refused to extend more credit. Fortu- nately, he told the accountant, the municipality had been of- fered a lifeline--a euro-denominated loan at only 3 percent in- terest from an overseas development agency. The councilor wanted help building a case for South Africa's National Treasury to guarantee the loan, as required by the development agency. With the loan proceeds, the municipality could launch tourism projects necessary for its economic development. The hard truth is that this municipality can ill afford a 3 percent euro interest rate (which could amount to 50 percent a year in South African rand if the exchange rate were to fall at the rate it did in 2001). In fact, municipalities are legally barred from bor- rowing in a foreign currency. Nor was the municipality able to convince the National Treasury to guarantee the loan: govern- ment policy. It is clear that municipalities' access to credit must depend on their own creditworthiness. Before the municipality borrows, it must increase its revenues, cut its expenditures, or both, even if that means delaying important projects. The loan the overseas agency had offered to the district council is part of a "low-cost" lending program intended to help South African municipalities build infrastructure and pursue economic development projects. The agency's project officer in South Africa is under heavy pressure from his agency and his govern- ment to place the project funds and demonstrate concrete suc- cesses. Well-intentioned development programs that make credit avail- able to the noncreditworthy do South African municipalities no service. Some development programs do more to support em- Country Case Studies: South Africa 317 ployment and careers in development agencies than to help build sustainable systems and structures. This loan would be a negative-sum transaction: The donor has spent large amounts to create and staff the lending program, and lending the money at the 3 percent euro interest rate would create additional cost. The municipality would be asked to assume additional debt when it cannot even meet its existing obligations, a step that would probably accelerate its developing financial crisis. Unfor- tunately, neither the council nor the management staff has the training and experience to recognize that the "low-cost" loan could turn out to be quite expensive. The municipality, reluctant to believe that the infusion of cash would not relieve its budget crisis, continues to search for a sympathetic ear in the national government. ment elections in 1995. Neighboring white and black urban areas were amalgamated, with the intention that revenues generated in the core cities could be used to extend services to underserved areas. In 1996 a new consti- tution established a decentralized system of government featuring au- tonomous local, provincial, and national spheres of government. In 2000 a second step in the consolidation of municipalities reduced the number from 843 to 284 and, in many cases, integrated rural and urban areas. These 284 new municipalities consist of three groups: · Local municipalities (232). · District municipalities (46), which typically include several local mu- nicipalities within their borders. · Metropolitan municipalities (6), which include South Africa's largest cities. This series of changes has brought clarity and certainty to the institu- tional framework for the six metropolitan municipalities. For the 278 local and district municipalities, however, which cover the same territory, an im- portant step remains--sorting out their respective powers and functions. Although recent amendments to the Municipal Structures Act have created 318 Subnational Capital Markets in Developing Countries a legal framework within which district municipalities will eventually pro- vide most services, in many cases services are still provided by local munic- ipalities. By ministerial regulation, legal authority remains mainly with lo- cal municipalities for the present time. How the eventual transfer of responsibility from local to district munici- palities will occur, and what it will mean for fiscal powers, are being debat- ed. The uncertainty associated with this ongoing transition makes it diffi- cult for local and district municipalities to plan capital spending strategically and to borrow to finance their capital investment plans. Local Government Revenues and Expenditures Municipalities spend a little less than a quarter of the total budgets of all three spheres of the South African government. In the 2000/01 financial year aggregate municipal spending was budgeted at some 61.8 billion rand (R), while national government spending was budgeted at R 84.3 billion, and provincial spending at R 110.5 billion Municipal revenues in South Africa come from own-source revenues (lo- cal government taxes and tariffs) and from intergovernmental transfers, mostly from the national sphere. Own-Source Revenues While municipalities generate about 92 percent of their own revenues in the aggregate, the experience of large urban centers differs from that of other municipalities. The six metropolitan municipalities, with strong rev- enue bases, generate some 97 percent of their own revenues, while munici- palities with annual budgets of less than R 300 million generate only 65 percent of their own revenues in the aggregate. Many poor and rural mu- nicipalities generate less than 10 percent of their own revenues. Most of the own-source revenues of municipalities come from tariffs for utility services such as water, sewerage, and solid waste disposal. National policy, reflected in legislation, calls for these services to be self-financing.1 In some cases they generate a surplus, and in others, losses. Much depends on the ability of the served population to pay and the seriousness with which the municipality pursues collections. Many municipalities provide electricity service to their residents, though this function is to be trans- ferred to new regional service entities. This prospect causes concern among municipalities that make a profit on electricity service or that rely on the threat of cutting it off to collect other taxes and tariffs. Country Case Studies: South Africa 319 The second biggest source of municipal revenues is the property tax, but this tax is available only to local and metropolitan municipalities. With the December 2000 advent of "wall-to-wall" municipalities, property taxes now may be imposed on essentially all property in the country. This represents a significant expansion of the tax base compared with that of apartheid-era local authorities, which generally included only urban areas. Historically, some municipalities imposed taxes on land value only, though most im- posed taxes on both land and improvements, often using different rates. National legislation is expected to soon provide uniform regulations to re- place the patchwork of apartheid-era provincial ordinances, but such legis- lation will leave tax policy decisions largely to local councils.2 For district municipalities own-source revenues come mainly from the regional services council levy, a business tax also used by metropolitan mu- nicipalities. It is generally recognized that this tax is in need of reform. Intergovernmental Transfers The national government transferred some R 6.5 billion to municipalities in the 2000/01 financial year. These transfers, and their share in the nation- al budget, have been increasing and are expected to continue to grow for at least the next three years. The transfers come from many small programs that South Africa's National Treasury has been working to consolidate. There are three basic types of transfers, and ultimately there may be as few as three transfer programs: · Unconditional transfers, generally determined by a poverty-based for- mula and often described as subsidies for providing basic municipal ser- vices to people who cannot afford to pay the full cost. These transfers ac- count for 57 percent of the national transfers to local government. The largest is the "equitable share" transfer, guaranteed by the Constitution. · Conditional transfers intended to help municipalities build infra- structure. The largest conditional transfer is the Consolidated Munic- ipal Infrastructure Programme grant. Infrastructure-related transfers make up 35 percent of the national transfers to local governments. · Conditional transfers intended to help municipalities improve their capacity or restructure their operations. These account for 8 percent of the national transfers to municipalities. In addition to consolidating the transfer programs, the National Treasury is committed to making the transfers as predictable as possible to facilitate 320 Subnational Capital Markets in Developing Countries local planning and capital investment decisions. This predictability is partic- ularly important for poor municipalities that rely heavily on national trans- fers for general operating revenues and, potentially, to secure borrowing.3 Provincial transfers to local government, made at the discretion of each province, are less well documented. The total in the 2000/01 financial year was estimated at R 1.2 billion. These transfers are usually tied to arrange- ments under which a municipality delivers a service on behalf of the province, though they also have been used to provide assistance to finan- cially troubled municipalities. Role of Municipal Borrowing in Financing Capital Investment All municipalities in South Africa--metropolitan, district, and local--have infrastructure responsibilities. Municipalities are responsible for local ser- vices such as potable water supplies, wastewater and solid waste disposal, city streets and street lighting, and, in many cases, electricity. All these re- sponsibilities require physical facilities, which in turn require capital in- vestment.4 Extending services to unserved and underserved areas has received the most attention recently and, given South Africa's history, is the most press- ing need. However, at least three other types of investment needs also must be considered. First, services above the basic level must be available to those who can afford to pay for them. Second, if South Africa is to create the conditions for increasing employment and thereby lifting more people out of poverty, well-chosen investments must be made in economic infra- structure that will help generate private direct investment. Third, and often overlooked in current debates, ongoing investment is needed to upgrade infrastructure that has reached the end of its useful life. While some at- tempt has been made to quantify the "backlog" investment needed to ex- tend services to underserved areas,5 little has been done to quantify the need for strategic or replacement investment. Even the "backlog" analysis may be of little use; demand for infrastructure probably has no practical limits, and the experience of industrial countries suggests that backlog in- vestment grows in proportion to a country's wealth rather than reaching some theoretical ultimate state. How are South African municipalities to finance such capital investment if they lack the current resources to do so? Like municipalities elsewhere, they can look for private equity investors, apply for intergovernmental grants, or turn to the municipal debt market. Country Case Studies: South Africa 321 Private Equity Investment through Public-Private Partnerships Public-private partnerships are one important channel through which pri- vate equity investment can contribute to the provision of local infrastruc- ture. In the three years that South Africa's Municipal Infrastructure Invest- ment Unit has been tracking the local infrastructure sector, projects using public-private partnerships have attracted some R 1.69 billion in private in- vestment (including the projected capital investment over the lifetime of the contracts). The public-private partnerships that South African munici- palities recently have entered into can be broadly divided into three groups: · Short-term partnerships that do not involve capital investment and usually require the municipality to make payments to the contractor for services rendered. · Long-term partnerships requiring fee payments to the municipality or investment in municipal infrastructure. · Divestiture arrangements under which the municipality transfers a facility to a private firm, though it may retain some regulatory role. Notable examples of public-private partnerships have been formed in the municipalities of Nelspruit, Richards Bay, and Johannesburg. In 1995, as a result of the redrawing of municipal boundaries, Nel- spruit's land area increased eightfold and its official population increased tenfold to 240,000, but its income grew by only 38 percent. Many newly incorporated areas had never received water and sanitation services. To ex- tend service to all residents, Nelspruit needed to make large-scale invest- ments in infrastructure. However, many residents of the new areas are very poor and can contribute little toward the cost of new infrastructure. To deal with these problems, in 1999 Nelspruit granted a concession for water and sanitation services, the largest long-term municipal public-private partner- ship in South Africa. The contract calls for a private firm, the Greater Nel- spruit Concession Company, to take over, manage, maintain, build, reha- bilitate, and, after 30 years, transfer back to the municipality all of Nelspruit's water and sanitation assets. Every resident is to receive basic ser- vice within five years. By early 2002, R 35 million had been invested, but the project has not attracted private finance; most of the funding has been put up by the government-owned Development Bank of Southern Africa. The main barrier to private investment appears to be the possibility that the national government will impose tariff caps. Although the government has never done so, South African legislation provides for this possibility. 322 Subnational Capital Markets in Developing Countries In 2000 the municipality of Richards Bay signed a 20-year concession contract for the operation, maintenance, and development of its airport. The contract involves R 13 million in payments to the municipality, which will be used to repay debts associated with the facility. Another R 7 million will probably be invested in upgrading runways, depending on the results of an independent assessment later in the contract period. In addition, 20 percent of the concession firm's equity and 20 percent of its dividends will go to a trust fund for the development and support of local communities, particularly traditional communities near the airport. In 2000 Johannesburg sold Metro Gas, a gas distribution business serv- ing approximately 15,000 business and residential customers, to U.S.-based Cinergy Global Power for R 110 million. The new owner is expected to in- vest another R 276 million in the facility over 10 years, making the deal ar- guably the biggest municipal privatization in South Africa. Experience has shown that the ability of municipalities to make wise and effective use of public-private partnerships depends on their ability to identi- fy and articulate their needs, negotiate with potential partners, live up to the commitments they make in their agreements,6 and manage the contracts they establish with private service providers. Local politics in South Africa, as elsewhere, can be turbulent, and public-private contracts have sometimes become political touchstones. Implementing any public-private partnership necessarily involves some tension between the municipal council's short- term interests in keeping tariffs low and service levels high, and the conces- sionaire's interest in earning a return for investors, so it is essential that con- tracts be clear and thorough. It is also critical that key decisions related to the partnership--such as tariff and collection policy--have broad support from the community. Finally, it is important that the community be able to give the private investor reasonable assurances, based on the community's legal standing and the commercial viability of the services involved, that the revenue streams for repaying the investment will be adequate. The prognosis for private equity projects in South Africa is unclear. For investors, it may depend on South Africa's willingness to clarify tariff is- sues.7 It also will depend on whether a significant number of municipal projects can be identified that will generate reliable cash flows. For munici- palities, policy considerations may come into play, with some municipal councils preferring to retain ownership and control over essential munici- pal assets. For others, political interests may be at stake. South Africa's pow- erful labor unions, for example, often see public-private partnerships as a threat. All these limitations suggest that public-private partnerships will Country Case Studies: South Africa 323 provide only a fraction of the investments needed--and that most infra- structure investment must be funded from other resources. Intergovernmental Transfers for Infrastructure In the 2000/01 financial year the national government provided only about R 2.4 billion in infrastructure transfers to local government. Infra- structure grants are made through a number of separate (generally sectoral) programs and are tied to specific projects. South Africa's National Treasury is committed to consolidating these grant programs and allowing munici- palities more discretion in deciding how to allocate funds and what infra- structure they most need to build, and efforts to ensure this are under way. The aim is to avoid the bottlenecks and unintended results that sometimes have occurred under the current system, which may make grants available for extending one service to an area but not other services. With the consolidation of these grant programs into one or two, it might be possible to reshape the infrastructure transfers into predictable revenue streams that the municipalities could then leverage through bor- rowing. In contrast to lump sum grants, this approach would allow more municipalities to receive simultaneous streams of revenue, helping the lo- cal government sphere build infrastructure more quickly. Municipal Borrowing Public-private partnerships will clearly finance only a fraction of South Africa's infrastructure investment needs in the foreseeable future. Infra- structure transfers are also small relative to investment needs. In most cas- es, therefore, municipalities will have to finance infrastructure from taxes and tariffs. Borrowing against these revenue streams, and possibly against infrastructure transfer streams as well, would allow municipalities to build infrastructure more quickly and distribute the financial burden more equi- tably across the generations that will use it. South African municipalities generally understand that borrowing is not a new or separate source of revenue and that borrowed capital and interest must be repaid with revenues from taxes, tariffs, and intergovernmental transfers. The good news is that municipalities in South Africa, unlike those in many other African countries, have significant recurring revenue streams available for leveraging. Borrowing, provided it is done wisely, can help these municipalities deliver tangibly on the promise of democracy. The South African policy on municipal borrowing, as laid out in the government's 1998 White Paper on Local Government and its 2000 Policy 324 Subnational Capital Markets in Developing Countries Framework for Municipal Borrowing and Financial Emergencies, clearly calls for such borrowing to be based on a market system, with lenders pricing credit to reflect the perceived risks. Potential Size of the Municipal Debt Market. Outstanding long-term mu- nicipal debt (to the public and private sector) in South Africa was estimat- ed to be around R 19 billion by mid-1997. Though relatively stable for years, this figure declined slightly after June 2000, in the run-up to the De- cember 2000 municipal elections. In the aggregate, municipalities clearly have the financial capacity to responsibly service a great deal more long- term debt. Quantifying the potential size of the South African municipal debt mar- ket is inevitably a speculative exercise, but some indication of that size can be gleaned from municipal capital budgets. For the 2000/01 financial year these totaled some R 13.7 billion. Budgeted amounts may be higher than actual spending, but the previous year's actual capital expenditures are esti- mated at R 10.3 billion. If half of all capital spending were debt financed and the other half "pay as you go," this would suggest a potential debt ser- vice capacity of R 5.1­6.8 billion a year. Based on these debt service capacity figures and a 10 to 12 percent annu- al interest rate on 20-year financing (a reasonable rate for low-risk debt in the South African capital market), total municipal debt capacity could be expected to be between R 38 billion and R 85 billion--two to four times the current outstanding debt of South African municipalities in 2000. If munic- ipal budgets continue to grow, debt service capacity also will grow. Natural- ly, much depends on assumptions about interest rates, the term of the debt, and the degree of leverage. Still, it is not unreasonable to conclude that the financial capacity of South African municipalities could support a municipal debt market around three times the current size. Trends in Municipal Borrowing. Actual lending in the municipal sector, however, has fallen far short of the performance that these figures imply. Long-term private lending to South Africa's municipalities has been essen- tially flat for at least four years. National Treasury data, collected since 1997, show that municipal debt owed to the private sector generally re- mained between R 11 billion and R 12 billion during 1997­2000. At the same time debt owed to public sector institutions, including the Develop- ment Bank of Southern Africa, grew significantly, from R 5.6 billion to R 8.1 billion (figure 18.1). This increasing reliance on public sector lending to municipalities is worrisome given South Africa's goal of expanding private investment. The Country Case Studies: South Africa 325 14 Total private sector 13 Total public sector 12 11 10 9 8 7 6 5 4 M J S D M J S D M J S D M J S D 1997 1998 1999 2000 Source: South African National Treasury. Figure 18.1 Outstanding Municipal Debt, South Africa, 1997-2000 (billions of rand) Development Bank of Southern Africa accounted for more than 30 percent of outstanding municipal debt by the end of 2000, a share that had nearly doubled since 1997. Most of its portfolio is with large, relatively robust met- ropolitan municipalities. Several of these municipalities have reported re- cent price competition by the Development Bank for their borrowing needs in cases where private lenders have been ready and willing to lend. In the short term, having a discount lender willing to "beat any price in town" be- cause of historical or current advantages conferred on it by the state (such as a lower cost of capital) is advantageous for municipal borrowers. In the long term, however, this will undermine the development of private lend- ing. Private lenders will have no incentive to spend time considering a po- tential loan if they are consistently undercut by a government-owned lender. Most new private lending since 1997/98 originated through a single specialized entity, the Infrastructure Corporation of Africa. The company's appetite for debt, through originating new loans and acquiring existing debt, has helped offset the exit from the market by other actors, and its 326 Subnational Capital Markets in Developing Countries market share has grown even faster than that of the Development Bank of Southern Africa. Like the bank, the Infrastructure Corporation of Africa ex- tends most of its municipal loans to large metropolitan municipalities. To- gether, the two institutions now account for about half of all outstanding municipal debt. Insurance companies have sold most of their municipal debt holdings, and pension funds have cut theirs significantly. This increas- ing concentration of municipal debt stock in the hands of a few lenders does not bode well for the South African government's goal of "a vibrant and innovative primary and secondary market for short- and long-term municipal debt" (South Africa 1998). Another undesirable trend is the changing nature of the debt stock. Mu- nicipal securities, which are (at least potentially) freely tradable on South Africa's capital markets, have steadily declined, while loans, which are less mobile and generally remain in the originator's portfolio, have increased markedly (figure 18.2). Because securities can be traded, term risk is less- ened where there is a market for the bondholder to sell the bond if neces- 14 Securities Annuity and other long-term loans 12 10 8 6 4 2 0 M J S D M J S D M J S D M J S D 1997 1998 1999 2000 Source: South African National Treasury. Figure 18.2 Outstanding Municipal Debt by Form, South Africa, 1997­2000 (billions of rand) Country Case Studies: South Africa 327 sary. This liquidity brings more potential investors into the picture, which is highly desirable in South Africa. The shift to loans can be attributed to two main factors. First, the reli- able and public accounting, budgeting, and financial information that in- vestors and rating agencies need is not readily available for most munici- palities. Thus investors' due diligence requires analysis and often proprietary recasting of municipal financial statements. That leads to high transaction costs in originating loans and transferring them among in- vestors. This situation favors large, specialized investors with experience lending to municipalities over casual investors that otherwise might be willing to buy a relatively small amount of rated municipal debt as part of their portfolios.8 Second, there is a lack of clear remedies in a municipal default, and some institutional investors have dealt with this legal gap by structuring highly se- cured loans that are specific to the originating institution.9 Some of these spe- cialized structures could be securitized, but the excess of capital supply over municipal demand means that there is little incentive to go to the effort. Assessment This discussion raises an obvious question: Why has such a large discrepan- cy arisen between the potential size of the municipal debt market and actu- al lending activity? If there is so much scope for additional borrowing, why has it not happened? Four main factors appear to be responsible. Local Government Reform Local government in South Africa has been subjected to continuous reform since 1994, a process that has involved all key parts of the sector--institu- tional, fiscal, and organizational. Two aspects of this reform have affected municipal lending activity. First, the process of change has created a great deal of uncertainty for investors, discouraging exposure to municipal risk. Second, some of the reforms themselves, such as those related to boundary demarcation, have adversely affected the structural basis of many munici- palities' financial positions (depressing ratios of revenue to population, for example), reducing their credit capacity. While the effects have been rela- tively minor for large metropolitan areas, they have been significant for many secondary cities and towns, which previously had been viable credit risks and which, after the metropolitan municipalities, represented the largest sector of the market for municipal credit. 328 Subnational Capital Markets in Developing Countries Thus both the fact of continuous reform and the nature of that reform have curbed lending. That is not to argue against such reforms. However, it does suggest that the impact of the reform and the length of the process can have long-lasting adverse effects on the ability of local governments to finance and deliver much-needed infrastructure and that such reform should be carefully assessed and guided in light of these effects. Budgetary Performance and Financial Management A combination of poor budgetary performance and weak financial man- agement has undermined the creditworthiness of a significant number of local authorities in South Africa. Some local governments are well man- aged and pursue disciplined fiscal policies. However, these tend to be the exceptions, and poor fiscal management and discipline are common throughout the municipal sector, even in the larger urban centers. Johan- nesburg, for example, ran into major financial difficulties in the second half of the 1990s, though its performance has since improved consider- ably. Many municipalities run budget deficits--while disguising them with formal budgets that unrealistically inflate revenues to achieve budget "bal- ance" as required by law. Most municipalities have limited human re- sources and systems capacity and a flawed municipal accounting system that undermines their ability to provide financial data that investors can regard with confidence. Legal and Regulatory Framework for Municipal Borrowing South Africa still has not enacted a sound, comprehensive legal and regu- latory framework for municipal borrowing. Uncertainties relating to processes and the rights and recourses of borrowers and lenders in the mu- nicipal sector remain, discouraging lending. For example, because rich, white local authorities rarely defaulted on debts under the apartheid regime, and because municipal debts were viewed as guaranteed by the na- tional and provincial governments, South Africa's law on remedies in the event of municipal defaults is underdeveloped. This has led to prolonged uncertainty and ineffective remedies in some recent municipal financial crises. Work initiated in 1998 to develop a comprehensive municipal bor- rowing framework resulted in a "best-practice" policy framework and draft legislation promulgated by the Cabinet in mid-2000. Constitutional and political difficulties, however, halted the progress of this legislation, and by early 2003 it still had not been enacted. Country Case Studies: South Africa 329 Activities of Public Sector Lenders The Development Bank of Southern Africa has advantages over private lenders as a result of its connection to the government and its ability to compete directly with these institutions for municipal clients. These ad- vantages have led to complaints that the bank is "squeezing out" such players from the market and thereby suppressing the development of pri- vate activity in the sector. Challenges Beyond the four factors discussed in the previous section, others also may have played some part in stifling the development of the municipal debt market in South Africa, such as high real interest rates and poor capital planning by local authorities. In early 2003 there were several reasons to hope that South African municipalities' access to private credit could im- prove in the short to medium term: · The December 2000 demarcations resolved long-pending amalgama- tion and boundary questions. The division of functions between lo- cal and district municipalities still needs to be clarified if they are to access capital markets autonomously. However, for metropolitan mu- nicipalities (and for local and district municipalities willing to coop- erate on debt issues), the chronic uncertainty relating to boundaries is over. · The December 2000 municipal elections put municipal councils in place for five-year terms. This placed councils in an excellent strategic position to assess their infrastructure needs and debt capacity and plan for the responsible use of debt as part of their strategies for ser- vice delivery and economic development. · South Africa's National Treasury has begun providing three-year in- dicative allocations for most intergovernmental transfer programs. For municipalities that cannot rely on own-source revenues from tax- es and tariffs, predictability in intergovernmental transfers is key. Clear indications of future transfers could enable these municipalities to access credit at whatever scale fits their capital needs. · A November 2001 constitutional amendment empowers municipal councils to make legally binding commitments relating to future budgets and revenues that will secure debt. Before this amendment, 330 Subnational Capital Markets in Developing Countries the weight of legal opinion was that a municipal council could not make such binding commitments, a restriction that would clearly limit investors' willingness to invest in long-term municipal debt. · Legislation to give effect to the government's policy framework on municipal borrowing, including in the event of municipal default, is included in the Municipal Finance Management Bill, which was ap- proved by South Africa's Cabinet in 2001 and was due to be enacted in 2003. All these factors point to a potential for significant expansion of South Africa's municipal debt market. Three main challenges must be addressed if this is to occur. Finalizing the Legal Framework South Africa's government set out a clear vision for a legal framework for municipal borrowing in its 1998 White Paper on Local Government and its 2000 Policy Framework for Municipal Borrowing and Financial Emergencies, but not all the policies described in these documents have been enacted into law. The most important legislation is the Municipal Finance Management Bill. This bill has three key parts: · Finance management. The bill regulates the budgeting, accounting, and financial reporting of local governments, requiring clear and consistently formatted information about municipalities' financial condition. This information should facilitate municipal borrowing by enabling lenders, rating agencies, and other players to make in- vestment decisions more quickly and efficiently. · Borrowing. The bill regulates short- and long-term municipal borrow- ing, implementing the elements of the government's policy frame- work that relate to borrowing. Key provisions of the bill limit short- term borrowing to cash flow management within the financial year; limit long-term borrowing to financing property, plant, and equip- ment; and allow municipal councils, under certain conditions, to pledge assets and future revenue streams to secure debt. A constitu- tional amendment paving the way for these security provisions was adopted by Parliament in November 2001. · Financial emergencies. The bill creates a process, including an agency within the National Treasury, to deal with municipalities in financial crisis, implementing the financial emergency provisions of the policy Country Case Studies: South Africa 331 framework. The goal is to restore a municipality to financial health as soon as possible while balancing the interests of citizens, the munici- pal council, creditors, and other stakeholders. It remains to be seen whether the provisions of the bill, once enacted, will provide a frame- work that is sufficiently robust and efficient to build investor confi- dence in municipal debt. In addition to enacting the Municipal Finance Management Bill, a few other loose ends need to be taken care of if South Africa is to create a leg- islative framework that enables municipalities to freely access private cred- it. These include the following: · Drafting disclosure regulations and providing mechanisms for dis- seminating information. If active trading in municipal securities is to occur, potential buyers of municipal bonds must have ready access to reliable information that is material to investment decisions. · Clarifying the ability of municipalities to commit to future tariffs or to tariff setting mechanisms. The recent constitutional amendment may help, but uncertainties remain. The tariff capping provisions of the Water Services Act and Municipal Systems Act that are trouble- some to private equity investors are of concern to debt investors. · Clarifying the powers and functions of local and district municipali- ties in a way that limits the potential for future uncertainty and change. · Reviewing old legislation for inconsistency with policy and revising it where necessary.10 Strengthening Local Government Capacity and the Budget Culture Some South African municipalities prepare and use capital and operating budgets and financial reports. Only a few, however, have developed com- prehensive capital investment programs that address their needs since the December 2000 amalgamations. These basic planning and accounting processes should be in place before a municipality goes to the capital mar- kets, because any municipality considering infrastructure borrowing should be in a position to understand how debt service and operational expenses for infrastructure will affect its budget. The municipality must be able to realistically project the revenue from the new investment. To achieve efficiencies and plan strategically, the municipality should be able to analyze different infrastructure options and financing scenarios. Mu- 332 Subnational Capital Markets in Developing Countries nicipalities that lack these skills are not equipped to make the best deci- sions for their community. Both municipal councilors and managers need these skills, though at different levels of detail. A council that can ask the right questions is more likely to get the information it needs to make good decisions. The December 2000 amalgamations exacerbated the effects of the lack of financial experience and capacity in some municipalities. This round of urban-rural consolidation blended an average of three municipalities and significant unincorporated territory into one new local municipality. This change meant that the new municipalities would have to consolidate fi- nancial information of varying quality from several sources, a process that could take a year or more. Even then it would be several years before mu- nicipalities or potential investors could discern trends in revenues or ex- penses. Ultimately, municipalities that want to borrow at reasonable rates, and have a choice of investors, must be able to produce a reliable record of financial performance. The amalgamations also pose a challenge in identifying capital needs. Most municipalities include newly incorporated areas whose infrastructure needs must be considered systematically as well as previously incorporated areas whose needs must be reprioritized in the context of the new munici- pality. South Africa's Municipal Systems Act requires municipalities to de- velop integrated development plans that include capital plans. However, many municipalities have not yet completed integrated development plans; among the plans that have been developed, not all are of high quali- ty or represent a true community consensus on needs and priorities. Another concern is the "culture of nonpayment" in some parts of South Africa, a legacy of resistance from the apartheid era. In a few municipalities council members have encouraged citizens not to pay their tariffs and tax- es. In many more, council members have failed to take the lead in helping citizens understand the need to pay for services. The practice of budgeting realistically and spending within the limits of available resources must be- come embedded in both the political and the management cultures of local government in South Africa. Many of South Africa's municipalities need sustained technical assis- tance, training, and experience to identify their capital needs and financ- ing options and to effectively articulate their need for credit. To borrow wisely and efficiently and to be able to pay their debts when due, munici- palities in South Africa, like those everywhere, must have strong skills in the following areas: Country Case Studies: South Africa 333 · Budgeting and accounting. · Identifying, analyzing, and prioritizing community needs. · Planning an appropriate mix and sequence of projects and funding options. · Developing specifications suitable for competitive procurement of construction and financing. · Managing procurement issues. · Managing projects during and after construction. · Marketing the municipality, its projects, and its debt instruments to investors. · Legal drafting and negotiation. Developing these skills will take time and effort, but the payoff will be good government, well-chosen projects, and appropriate financing. Shortcuts could result in poor projects, expensive financing, and little support in local communities or the country as a whole for further municipal borrowing. Foreign and domestic development agencies seeking to make a sustain- able contribution to South Africa's municipalities would do well to consid- er mentoring and support to develop these basic skills. South Africa's Na- tional Treasury has launched a pilot program that is bringing experienced municipal finance managers from other countries to work with their South African colleagues. These managers will stay for one to two years, helping to get newly amalgamated municipalities' budgets and accounts in order and into compliance with the reporting requirements of the Municipal Fi- nance Management Bill. This kind of ground-level support and capacity building is essential for financially healthy municipalities, for sound, infor- mation-based local policy decisions, and for wise borrowing. Facing South Africa's Ambivalence about Markets These challenges--dealing with imperfections in the legal framework and building municipal capacity--will be familiar to anyone who has worked on municipal debt policy anywhere. A more difficult issue needing to be addressed in South Africa is the society's ambivalence about the market-ori- ented policies being pursued. This ambivalence mirrors, and is reinforced by, global debates about economic integration and deregulation. South Africans have mixed feelings about their private financial institu- tions. Many are proud of their "first-world" capabilities. The South African economy boasts well-functioning stock and bond markets, commercial and investment banks, insurance companies, rating agencies, and regulatory 334 Subnational Capital Markets in Developing Countries bodies. Many others, however, see these institutions, which developed un- der an oppressive regime, as instruments and beneficiaries of that repres- sion. A deep-seated mistrust of capitalism and resentment of the role that some capitalists played in the apartheid era persist in some quarters. This history has made it difficult for many South Africans to embrace market- oriented financial institutions.11 Although government policy endorses the need to attract private fi- nance, there is little confidence that the private sector will come to the table.12 There are concerns that private lending to subnational govern- ments will develop slowly or not at all, even if the correct policies are put into place and the necessary capacity built. Moreover, there are concerns that even if markets provide finance for large and financially secure munic- ipalities, small and poor municipalities will be left out. These concerns have had several consequences. First, they have made it difficult for the government to push the necessary legislation and constitu- tional amendments through Parliament as quickly as had been hoped. Sec- ond, concerns about the reliability of financial markets have led some in government to consider various forms of artificial stimulus for subnational borrowing, including national government guarantees and debt insurance sponsored by the Development Bank of Southern Africa. Third, there is some support within the South African government for nonmarket ap- proaches to subnational debt--such as government on-lending--especially for provincial governments (there is presently no provincial borrowing in South Africa, which is effectively prohibited). These nonmarket approaches would probably prove problematic in exe- cution. With both direct government lending and guarantees, there will in- evitably be defaults, imposing future costs on the national government. These contingent costs are difficult to predict and quantify. Governments throughout the world have a poor record of managing loans to subnational borrowers--the rate of default on government or government-guaranteed loans usually exceeds that on commercial loans to the same entities. An- other concern is that shifting private investors' focus from the creditwor- thiness of the borrower or the project to the creditworthiness of the nation- al government will result in loans being made to subnational governments that cannot afford them, increasing their financial stress. These cures for market imperfections could easily be worse than the disease. South Africa's ambivalence about markets, while understandable given the country's history, must be overcome if its municipalities are to attract private investment. Choosing the route of relying on capital markets Country Case Studies: South Africa 335 would mean that the focus would have to be on improving the framework, skills, and information that those markets need to function effectively. Maintaining this focus will be more difficult than taking shortcuts such as guarantees, but it will be a more sustainable and strategic choice. Notes 1. Municipal Systems Act, No. 32 of 2000. See especially subsection 74(2). 2. A draft Property Rates Bill was published for comment in August 2000 and is likely to be adopted in 2003. 3. These transfers must be both well targeted and predictable. But there can be a tension between these two goals. As new information becomes available that would help improve targeting, the equity of adjusting target- ing must be weighed against the need to ensure that commitments can be met for infrastructure that has been begun or financed in good faith. 4. With some minor exceptions, municipalities in South Africa are not responsible for social services (such as health and education), which are provided by provinces, or for policing, which is a national function. 5. The 1996 Municipal Infrastructure Investment Framework estimated the basic need at R 38.5 billion and the full service need at R 75 billion. 6. In some cases politicians and activists have actively discouraged resi- dents from paying utility bills, and councils have been reluctant to cut off service to enforce payment. In Fort Beaufort the municipality sued to void the concession agreement and was successful in setting aside the contract. 7. The Water Services Act provides that the minister for water affairs and forestry can regulate municipal tariffs for water services. The Municipal Sys- tems Act provides similar authority to the minister for provincial and local government with respect to all municipal services. The potential for such regulations to interfere with negotiated arrangements between a munici- pality and a concessionaire limits investor interest in revenue-based public- private partnerships in South Africa. 8. The apartheid-era local authorities that issued the now-disappearing securities did not necessarily have good-quality financial information ei- ther. However, with their financial strength and implicit national and provincial support, their bonds were seen as safe investments for individu- als and institutions. 9. Examples include tax-structured transactions in which a financial in- stitution benefits from depreciation on municipal assets; pledges by munic- 336 Subnational Capital Markets in Developing Countries ipalities to banks of property tax revenues derived from the banks' own property; and deposits by municipalities with financial institutions that, with interest, equal the principal amount of the loan at maturity, protect- ing investors' principal. 10. Buried legislation creates an entry barrier, at least to the uninitiated. Recently a financial institution that had not previously lent to South African municipalities negotiated a loan with one of the country's biggest cities. The lender commissioned a South African law firm to review the ap- plicable legislation. The firm found a Transvaal ordinance from 1903 and an unused exchange control regulation that required National Treasury ap- proval of the proposed loan. Existing lenders to the municipality, either unaware of or unconcerned by these laws, had never asked for Treasury ap- proval, and Treasury did not want to be in a position of approving--and perhaps implicitly endorsing--local borrowing decisions. 11. Investors from such countries as the United Kingdom and the Unit- ed States are not necessarily viewed more favorably. While it is accepted that foreign investors found the stigma of white racism unappealing, they are nevertheless seen as benefiting from the perceived stability and low la- bor costs of that era, and their investments in South Africa as propping up the apartheid regime. 12. In an interesting contrast, in Eastern Europe and the former Soviet Union a broad perception that communism had failed left the typical per- son willing to accept more market-oriented solutions. In South Africa capi- talism was seen as part of the problem, and the expected coming of democ- racy was often linked to diminution of the power of capitalists. Chapter 19 Sub-Saharan Africa Zimbabwe A centrally prescribed and unsustainable credit market succumbs to political and economic turmoil. Roland White and Matthew Glaser Lessons Although Zimbabwe neighbors South Africa, its recent experi- ence in local government borrowing is very different from that country's. In Zimbabwe local government borrowing has been premised on a policy and regulatory regime--most evident in central government guarantees and prescribed assets for insti- tutional investors--that is inimical to the development of sus- tainable municipal credit markets. The country's two largest cities have been able to successfully issue a limited volume of securities for many years, but this has come at the cost of accu- mulating liabilities for the central government and significant fi- nancial losses for investors. When a default by Harare in 1998 prompted the government to withdraw its implicit guarantee of local bond issues, interest rates immediately jumped, making local government borrowing prohibitively expensive. Not until 2001 did the central government again begin to underwrite mu- 337 338 Subnational Capital Markets in Developing Countries nicipal debt issues, this time explicitly. Moral hazard has thus become embedded in the subnational lending system. The economic and political crisis that Zimbabwe now confronts makes it difficult to talk about prospects for sustainable subna- tional borrowing in the country. The most recent bond issues suggest that borrowing has ceased to be an affordable source of funding for long-term investment programs. The policy and regulatory environment for subnational borrowing in Zimbabwe has fostered a "false" municipal bond market in which investors rely on the financial position of the central government rather than the creditworthiness of the municipal issuer. The sustain- ability problems that this has created have been greatly exacer- bated by the deterioration in the country's economic and politi- cal climate. Starting in early 2001 economic and political conditions in Zimbabwe be- gan to deteriorate dramatically. In February 2002 the official annualized in- flation rate stood at around 110 percent, GDP had shrunk by 5 percent over the previous year, and one in four jobs in the formal economy had disap- peared. Under conditions such as these, the prospects for sustainable, cost- effective subnational borrowing are remote. Whether the municipal fi- nance market will regain any of the momentum it began to show in the 1990s--though in a problematic policy environment--remains unclear. Institutional Framework for the Financial Sector Until recently Zimbabwe had a relatively stable financial system based on a long tradition of high savings and a well-diversified financial sector. Com- plementing the substantial and relatively sophisticated banking system are near-banks (nonbank financial institutions), insurance and pension funds, a well-established stock market, and a range of other market participants. The system has become increasingly competitive, though with a growing amount of risk. The government's National Economic Structural Adjust- ment Program, implemented in the early 1990s, introduced some econom- Country Case Studies: Zimbabwe 339 ic deregulation, enhancing competition, allowing market conditions to in- fluence financial prices, and increasing the diversity of market participants and financial services (World Bank 1998). During the 1990s assets held by deposit-taking institutions grew significantly in real terms (table 19.1). Financial Institutions and Instruments The financial sector remains segmented, with legal restrictions on the kinds of transactions that different types of institutions can undertake. · Commercial banks. The seven commercial banks in operation in 1999 were the largest institutions in the banking system. At that time they held 32 percent of the assets in the financial system and 60 percent of loans and advances, and 49 percent of deposits, in the banking sys- tem. Permitted to undertake most types of financial intermediation (except financial leasing and hire purchase), commercial banks are important lenders to subnational governments for short-term facili- ties, such as overdrafts and bridge financing. · Merchant banks. With six existing in 1999, merchant banks perform the second largest volume of financial intermediation but cannot of- fer checking accounts. They tend to cater to larger enterprises by pro- viding tailored services and trade financing. · Discount houses. There were six discount houses in 1999. Their prima- ry role is to act as a market maker for commercial and merchant bank liquidity and as the main receiver and dealer of treasury bills from the central bank, the Reserve Bank of Zimbabwe. (Until 1999 dis- Table 19.1. Assets of Deposit-Taking Institutions, Zimbabwe, 1992­97 (millions of Zimbabwe dollars) 1997 Type of institution 1992 1993 1994 1995 1996 (Sept.) Commercial banks 8,062 14,900 20,275 24,669 32,648 40,533 Merchant banks 1,961 6,126 7,756 10,361 17,051 18,349 Discount houses 716 1,347 1,674 3,781 3,161 2,958 Finance houses 1,491 1,498 1,908 2,866 4,270 5,420 Building societies 2,916 3,892 5,808 10,088 13,843 15,666 Post Office Savings Bank 2,315 2,977 3,662 4,466 6,227 6,927 Total 17,460 30,740 41,082 56,230 77,200 89,852 Source: Reserve Bank of Zimbabwe 1997. 340 Subnational Capital Markets in Developing Countries count houses had exclusive access to the Reserve Bank and would purchase treasury bills and rediscount them into the rest of the finan- cial market. Now, however, all financial institutions have access to the Reserve Bank.) The discount houses play an important market- making role in the municipal finance market, assisting subnational governments in issuing long-term bonds for capital development. In 2001 one of the discount houses was licensed to assist local authori- ties in seeking funding for infrastructure and to create a market for lo- cal government securities (such as municipal bonds and municipal treasury bills). Discount houses held around 3.3 percent of the assets of deposit-taking institutions in 1997. · Finance houses. The four finance houses existing in 1999 function largely as fixed asset financing arms of commercial banks, financing equipment and vehicle loans and collateralized lending. The finance houses hold about 3.1 percent of the assets in the financial system (excluding the Reserve Bank of Zimbabwe). · Building societies. There were five building societies in 1999. These in- stitutions accept share, savings, and fixed deposits and negotiable certificates of deposit. They lend for residential and commercial mortgages, purchase treasury bills, provide other loans to the govern- ment, place funds in the money market, and finance low-income housing projects. In 1997 they held approximately 17 percent of the assets in deposit-taking institutions. · Post Office Savings Bank. A government-owned institution, the Post Office Savings Bank mobilizes funds from small savers throughout the country and offers both savings and fixed deposit facilities. In 1997 it held 7.7 percent of the assets in deposit-taking institutions. · Development finance corporations. The development finance corpora- tions specialize in term financing, often with support from the gov- ernment or donors. These include the Agricultural Finance Corpora- tion, Zimbabwe Development Bank, and Small Enterprises Development Corporation. Instruments traded in the market include treasury bills and government bonds, bank acceptances, certificates of deposit, and a limited amount of commercial paper. Treasury bills dominate the market and have had a crowding-out effect on competing investments. Maturities range from 30 to 91 days for treasury bills and from 1 to 16 years for fixed-income debt stock. Country Case Studies: Zimbabwe 341 Prescribed Assets Regime The government of Zimbabwe, to regulate interest rates in its favor, has maintained a "prescribed assets" regime. Institutional investors (pension funds and insurance companies) in Zimbabwe are required to keep 45 per- cent of their holdings in prescribed assets, which have generally included long-term bonds (with maturities of more than six years) of the central government, parastatals, and local governments (this share has been re- duced from 55 percent in 1997 and 65 percent before that).1 This require- ment originally had some validity, as a way to ensure safe investments by financial institutions. However, its continued use guarantees the public sec- tor preferential access to domestic financial markets. Throughout the 1990s there were insufficient prescribed assets in the market to satisfy the statutory requirements. This, coupled with the gov- ernment guarantee implicit in the prescribed asset requirements, has made pricing these assets difficult. A recent World Bank (1998) review of the fi- nancial sector concluded that . . . [T]he contradiction between investors holding fewer prescribed assets than required and the inability of [the] Government to place long-term stock suggests that market participants are reticent to hold long-term instru- ments except at very high yields and [that the] Government is reticent to pay such a premium for long-term funds. The availability of prescribed assets has tended to fluctuate. During shortages municipal bonds, for example, have often been oversubscribed when first issued, depressing interest rates (Phelps 1997). There is no sec- ondary bond market in Zimbabwe because most institutional investors find it difficult to meet the required 45 percent share for prescribed assets and therefore adopt a "buy-and-hold" strategy. Intergovernmental Structure Zimbabwe has a two-tier system of elected government: the central govern- ment (with line functions devolved to both the provincial and the district level) and local governments. The local governments consist of 57 rural district councils and 23 urban councils--city and town councils, town boards, and local boards. The councils are elected on the basis of ward constituencies, with all adults having the right to vote. Their functions and responsibilities are set 342 Subnational Capital Markets in Developing Countries out in the Rural District Councils Act and the Urban Councils Act. Local governments are responsible for administering the areas under their juris- diction. However, the central government maintains strategic control of boundaries and the hiring and firing of local government senior staff. The urban councils have a broad range of responsibilities--street light- ing, street cleaning, physical planning, emergency services, municipal po- lice, low-cost housing, primary education, primary health care, solid waste management, road maintenance and expansion, and water and sewerage services. They also operate enterprises, including farms and beer gardens. Subnational Funding Sources Local authorities have three main sources of funds for financing current and capital expenditure: internal revenues, external revenues, and borrowing. Internal Revenues Local governments derive internal revenues primarily from taxes and ser- vice charges. They do not have complete autonomy over these revenue sources: to increase taxes, property assessments, and service charges for low-income areas, local governments must obtain the approval of the min- ister of local government, whose ministry oversees their operations on be- half of the central government. · Assessment rates and supplementary charges. Assessment rates, applied to the ratable value of property within the municipal boundaries, are levied on commercial properties and on residential properties in high-income areas. Flat unit charges on properties (known as "sup- plementary charges") are levied on residential properties in low-in- come areas. Any increases of more than 20 percent in either of the charges must receive prior approval by the minister of local govern- ment, often leading to delays that result in revenue shortfalls. Assess- ment rates and supplementary charges account for around 20 percent of subnational income. · General service revenue. Local authorities generate increasing revenues from charges for "economically viable" services, such as sewerage and refuse collection. Sewerage contributes around 7 percent of subna- tional income, refuse removal and housing income around 6 percent each, and income from health and welfare and other general services around 24 percent. Country Case Studies: Zimbabwe 343 · Trading income. Profits from water distribution are by far the largest source of trading income. Some councils also run such trading opera- tions as farms and beer gardens, although these typically contribute only a minor share of budgets. · Other internal revenue sources. From time to time local authorities also derive income from the sale of fixed property, primarily land. External Revenues External revenues come through intergovernmental transfers, which have recently declined precipitously to insignificant levels (table 19.2). In trans- ferring responsibility for health care services to local authorities, for exam- ple, the central government initially agreed to fund a share of the salary costs through grants. Within a year, however, these grants had ceased to flow, and the health funds of local authorities now operate at significant deficits. In general, intergovernmental transfers are a much less important source of revenue for larger, urban councils than for rural councils. Harare, for example, received no intergovernmental transfers in 1995­97, and for Kwekwe such transfers accounted for less than 1 percent of revenue during that time. Borrowing Local authorities can turn to three main sources for borrowing: · Government loans. Government loans have accounted for around 90 percent of local authorities' total borrowing for capital development over the past two decades. The funds for these loans have been ob- tained mostly from aid agencies, then on-lent to local authorities at Table 19.2. Local Government Revenues by Source, Zimbabwe, 1995­98 (percentage of total) Revenue source 1995 1996 1997 1998 Assessment rates and supplementary charges 24 25 22 31 General services 35 44 36 56 Trading income 22 7 6 8 Intergovernmental transfers 18 24 35 6 Source: Zimbabwe Central Statistical Office accounts. 344 Subnational Capital Markets in Developing Countries concessional rates through the General Loan Development Fund and National Housing Fund. · Bank overdrafts and short-term loan finance. Overdrafts and short-term borrowings are capped at the local authority's income from rates in the previous year (unless otherwise authorized by the minister of lo- cal government) and are intended to be temporary. These funds may be used for capital expenditure only if borrowing power for that ex- penditure has been obtained from both the Ministry of Finance and the Ministry of Local Government. Between 1994 and 1997 subna- tional revenue grew by 26 percent while outstanding debt grew by 78 percent. Much of this deficit is being financed by overdraft facilities. · Long-term finance. Local authorities can raise long-term finance from private investors by issuing bonds or by borrowing directly from fi- nancial intermediaries, including the Local Authorities Pension Fund. Local Government Borrowing The Urban Councils Act and Rural District Councils Act provide broad bor- rowing powers to local governments. Local governments are permitted to borrow for capital works or improvements, acquisition of fixed property, certain kinds of advances, payment of compensation (excluding that for permanent employees), liquidation of previous loans, relief of general dis- tress (caused by a natural disaster, for example), and acquisition of plant, equipment, and vehicles. Regulation and Oversight Before a local government can borrow, however, the local council must pass a formal resolution of its intention to borrow, give public notice of its intention, including the purpose for which the borrowed funds will be used, and invite comments from its constituents. Before borrowing from the central government, the Local Authorities Pension Fund, a municipal provident fund, a medical aid society, or another local authority, the coun- cil must obtain permission from the Ministries of Local Government, Pub- lic Construction, and National Housing. If the council intends to issue bonds, stock, or debentures, it must obtain authority from the Ministries of Finance and Local Government. Four of Zimbabwe's five major cities-- Harare, Bulawayo, Gweru, and Kwekwe--have issued bonds in local capital markets. Indeed, Harare and Bulawayo have been raising bond finance since the late 1960s or early 1970s. By contrast, short-term overdraft facili- Country Case Studies: Zimbabwe 345 ties require central government approval if they exceed the previous year's income from rates. The requirement for central government approval, along with the pre- scribed assets regime, has provided a foundation for the primary market for municipal debt in Zimbabwe because it has led to a market assump- tion of an implicit central government guarantee of such debt. The phrasing of issues has often reinforced this impression. For example, the prospectus prepared by Kwekwe for its 1998 issue stated that "[w]hile the Government does not explicitly guarantee this stock issue, it has the moral responsibility to ensure that the local authority meets its external obligations." Prud'homme (1999, p. 14) has argued that the local gov- ernment bond market in Zimbabwe is really a "false" market in which "[f]or all practical purposes, municipal bonds . . . are basically a variety of central government bonds, and are seen as such by the financial com- munity." Borrowing Trends Faced with deteriorating revenue bases and increasingly unfunded man- dates in the social sectors, local governments have been forced to in- crease their debt over time (table 19.3). Of the total debt, an average of 47 percent is held in bond issues in the domestic markets (registered stock) and 53 percent is from other sources (including the central gov- ernment and private sources, such as the Local Government Pension Fund). Long-term borrowing from private financial institutions is there- fore an important source of finance, particularly for the five largest coun- cils (table 19.4). Table 19.3. Gross Public Debt of Local Authorities, Zimbabwe, 1994­97 (millions of Zimbabwe dollars) 1994 1995 1996 1997 Long-term borrowing 1,376.7 1,559.4 1,767.5 2,293.4 Registered bonds 553.5 543.5 629.5 815.6 Central government 713.4 871.1 945.6 1,287.4 Private 109.7 144.8 192.4 190.4 Short-term borrowing 4.8 50.7 43.8 169.5 Total 1,381.5 1,610.1 1,811.3 2,462.9 Source: Zimbabwe Central Statistical Office 1999. 346 Subnational Capital Markets in Developing Countries Table 19.4. Bond Issues by Local Governments, Zimbabwe, 1990­2001 Total amount Total (millions of amount Coupon Inflation Zimbabwe (millions of rate rate Term Year City Issues dollars) U.S. dollars) (percent) (percent) (years) 1990 Harare 1 46 17 15.3 19 1991 Bulawayo 1 10 1.98 17 10 1993 Harare 1 90 12.98 32 27.6 5 1994 Harare 1 120 14.30 17 22.3 7 1994 Bulawayo 1 100 1.92 18 22.3 10 1996 Harare 1 100 9.93 18 21.4 17 1996 Bulawayo 1 100 9.93 20 21.4 10 1997 Bulawayo 1 100 8.04 14 18.9 15 1998 Kwekwe 1 28 1.15 28 30 10 1998 Gweru 1 50 2.05 23 30 10 2001 Hararea 10 249 1.84 13.4­18 60.8­112.4 1­13 2001 Bulawayo 2 250 1.388 28­32 60.8­112.4 a. All these bonds carried an explicit government guarantee. The official Zimbabwe dollar­U.S. dollar exchange rate was pegged in January 2001. The U.S. dollar figures here reflect the parallel market rate. Source: Reserve Bank of Zimbabwe. According to data collected from the local authorities and one of the dis- count houses, local government bonds in issue in September 2001 stood as follows: City Millions of Zimbabwe dollars Harare 539 Bulawayo 630 Gweru 50 Kwekwe 36 1,255 However, municipal bonds constitute a very small share of the instru- ment base of the lending institutions (about 5 percent of their portfolios), and the instruments are rarely traded. The secondary trades that have oc- curred have been transacted at significant discounts, probably reflecting the market risk perception of municipal bonds. Nearly all the bonds are held by pension funds, with banks preferring a short-term investment hori- zon. As a result, it has become extremely difficult to price existing stocks and to compute the cost of funds for new issues. Country Case Studies: Zimbabwe 347 The implicit government guarantee has always ensured a lower cost of borrowing for local governments compared with the yield curve. This has been changing, particularly since a much-publicized default by Harare on the redemption of locally registered bonds in August 1998. Most financial market participants have access to very little information on municipal bonds or securities in issue and consequently have very little information on the operations and performance of urban councils. This lack of information limits the market's capacity to price municipal securi- ties and, more important, limits the financial sector's ability to design and offer services to the municipal sector. There is interest in developing loans for urban councils, but municipal bonds account for only about 3 percent of institutional investors' portfolios on average (PADCO and Techfin Re- search 1999). Design of Municipal Debt The Urban Councils Act and Rural District Councils Act specify that "all loans made by authorities are secured and charged upon the assets of the Council and all securities granted by the Council in respect of such loans shall rank equally without priority." This pledge is similar to the security for general obligation bonds in the United States. However, this statutory lan- guage appears to prevent the issue of revenue bonds in Zimbabwe--bonds for which the issuer pledges a particular revenue source. Nor do the acts an- ticipate the use of innovative security structures that would dedicate a sepa- rate revenue stream to an escrow account for future debt offerings (Johnson and Kimberley 1999). Even so, this type of security enhancement is under review by Harare, where potential investors in a proposed bond offering are suggesting that the water fees of the 20 largest users be paid directly to an es- crow agent as security for payment of the debt service on the new bonds. The general pledge or parity bond provision is modified by other provi- sions of the Urban Councils Act relating to the creation of separate estate and parking accounts into which a council's income from specified sources flows. Money in these accounts can be directed only to specified costs and expenditures. Although these accounts may be considered dedicated in- come streams for purposes of a revenue bond financing, there is no case law to support this conclusion or the conclusion that these accounts are as- sets of the council and therefore subject to the general pledge provision (Johnson and Kimberley 1999). In some cases, however, a dedicated revenue stream other than property taxes (for example, revenue from water fees) might be more attractive to in- 348 Subnational Capital Markets in Developing Countries vestors than a general obligation pledge of the local government. More- over, given the decline in the general balance sheets of local governments, a project investment financed from a dedicated revenue stream might be more acceptable to investors than the general obligation bonds now autho- rized. Revenue bonds have therefore attracted growing interest from local governments, though no such bonds have been issued yet. All bonds have been for general development purposes, issued on a pari passu basis (with a ranking equal to that of all others). The legislation allows subnational governments to establish and operate a consolidated loan fund, used to separately account for all money bor- rowed and principal and interest payments made. Only Harare has estab- lished such a fund; other local governments use sinking funds to repay loans. They pay annual installments into the sinking fund, sufficient to pay off the debt over the period for which the money was borrowed, and then use the fund for final (bullet) redemption. The amount of issues is usually lower than the amount authorized by the Ministries of Local Government and Finance. The authorizations usual- ly remain valid for several years (in principle, for an unspecified number of years). A local government then can opt to use up only part of its right to issue bonds, either because it does not want to run the risk of failing to sell all its bonds or because it does not want to commit itself to excessively high interest payments in the coming years. For example, Kwekwe, which obtained authorization to float bonds for 55 million Zimbabwe dollars (Z$) in 1998, chose to go to the market for only Z$28 million that year and planned to go to the market for the rest at a later date. Bond Issuance and Trading Local governments that have issued bonds prepare the issue in-house or have discount houses assist them by drawing up the prospectus, undertak- ing the initial canvassing, and processing the initial public offering. Trad- ing takes place mainly over the telephone, and settlement is conducted on the same trading day. There is no central depository, nor is there a well-de- fined settlement system. Because of liquidity problems, some financial in- stitutions have encountered situations in which transactions have been confirmed only after funds have been cleared. In principle, bonds are allocated to prospective investors through bid- ding. Interested investors fill in a stock auction application form indicating the price they are prepared to pay for the bonds of a nominal value of US$100 and the number of bonds they want, enclosing a check for the ap- Country Case Studies: Zimbabwe 349 propriate amount. The bonds are then allocated to those offering the high- est price. In practice, the prices offered appear to be very close to the nomi- nal value of the issue. The Kwekwe issue of Z$28 million, for example, sold for Z$28.92 million. Most bond issues have reportedly closed at somewhat below par, at about 95 to 97 percent. The prospectus for a bond issue states the purpose of the issue, the amount to be raised, the interest rate, the issue price, the opening and clos- ing dates for applications, and the maturity date. It explains where, when, and how to file applications. Payment is often 10 percent on application, and the balance three months later. The prospectus also gives details about the allocation of bonds and any refunds of payments made on application. Interest is normally calculated on a daily basis at the stated rate and paid semiannually. The prospectus states the redemption date, on which pay- ment is made against surrender of the bond certificate. The prospectus ex- pressly states that the debt will rank pari passu (equal) with all existing debt and form a charge on the rents, rates, and general revenue of the local au- thority and that it will be further secured on all its assets. The level of disclosure in a typical prospectus (generally four to five pages long) for a local authority bond offering in Zimbabwe falls far short of international norms and often fails to meet listing requirements of the Zimbabwe Stock Exchange. The quality of prospectuses reflects an underly- ing assumption by investors that the creditworthiness of the issuer is im- material because of the sovereign guarantee implicit in the prescribed as- sets regime and approval requirements for borrowing. Moreover, in Zimbabwe there is no authority that verifies the facts provided in prospec- tuses. No bonds issued in Zimbabwe have been supported by an independent credit rating. Indeed, with the central government guarantee, there has been no need to verify the financial information. In 1999­2002, 17 of the urban councils were independently rated. Contrary to expectations, the rating agency Duff & Phelps (1999) found that some of the smaller local governments "reflect lower costs structures, more efficient collection proce- dures, and more pragmatic financial planning than their large counter- parts." The rated councils have used the associated management reports as guides for improving their internal operating efficiency. Even though all local municipal long-term bonds are listed on the Zim- babwe Stock Exchange, there is virtually no trading. The Stock Exchange and its members tend to ignore municipal bonds. If secondary trading in local debt securities develops, there will be a need for continuing disclosure 350 Subnational Capital Markets in Developing Countries to allow the market to reassess the value of securities on the basis of devel- opments since the initial prospectus. Even an initial prospectus that prop- erly and fully disclosed relevant information would not contain informa- tion relevant to investors years after the bonds were first sold. Just as for initial disclosure, standards for continuing disclosure do not need to be es- tablished by statute. Assessment and Prospects Toward the end of the 1990s and through 2001 central government fund- ing of local authorities in Zimbabwe began to decline precipitously. Grants fell both because of policy decisions by the government and because of its increasingly strained fiscal position. Public sector loans declined mainly as a result of the closure of donor-funded projects that had been the source of the intergovernmental loans and because of the diminished donor interest in the country. Moreover, while no accurate data are available on the re- payment performance of local authorities, there have been reports of an in- creasing number of effective defaults on public sector loan obligations--a common phenomenon in Africa and indeed globally. Under these circumstances--and given the advantages of the policy en- vironment for local borrowers due to the prescribed assets regime and cen- tral guarantees--it might have been expected that local governments would make greater recourse to the bond market. However, this has not occurred. Real borrowing in 1998­2001 was substantially lower than in the 1980s and early 1990s. This decline probably reflects the deterioration of the financial sector's position resulting from the deterioration in the wider economy as well as the central government's difficulties in fully enforcing its prescribed assets policy. It probably also reflects factors specific to the municipal bond sector, such as the fallout from the Harare default in 1998. It is impossible to disentangle these effects and weigh their relative impor- tance. Moreover, attempting to establish what would have happened in the municipal debt market if Zimbabwe's macroeconomic situation had not de- teriorated would be a largely speculative exercise. Only a small number of local governments have floated bonds in Zim- babwe, and until 1997 only Harare and Bulawayo had done so. The value of bonds in issue constitutes a relatively small share of capital investment financing by local governments. For example, Prud'homme (1999) has cal- culated that in 1990­96 bonds accounted for only around 20 percent of capital investment by local governments. During this period subnational Country Case Studies: Zimbabwe 351 governments made capital investments amounting to US$298 million while issuing US$60 million in bonds. This amount seems relatively small when compared with the reported investment needs of local governments--around US$55 million a year (PADCO and Techfin Research 1999). A World Bank­funded program fi- nancing infrastructure investments in all 23 urban centers in Zimbabwe disbursed a similar amount (also around US$60 million) in 1994­98, so the nascent municipal finance market in Zimbabwe was able to match this large public sector program in dollars invested. The domestic financial market has the potential to make significant fi- nancing available for the investment needs of local governments, but cur- rent stocks of municipal bonds have been acquired largely on the back of government guarantees and the prescribed assets regime. Private investors have shown a reluctance to take municipal paper on its own merits. Fur- ther development of the municipal finance market in Zimbabwe (feasible only when macroeconomic stability is recovered) would depend on a num- ber of measures, including introducing a more transparent and predictable intergovernmental fiscal transfer system, amending the legislation to allow for the ring-fencing of infrastructure projects through the issuance of dedi- cated revenue bonds, introducing credit enhancement measures such as municipal bond insurance, and strengthening financial management and disclosure by local governments. A False Municipal Bond Market As Prud'homme (1999) has argued, the central authorization, central guar- antee, and prescribed assets policies that have provided the foundation for private lending to local governments have created a "false" municipal bond market in Zimbabwe: investors have lent on the basis of these poli- cies rather than on the basis of the risk presented by a local authority. The importance and impact of this policy environment became clear in the Harare default of 1998. The central government did not make good on its implicit guarantee, with the consequence that all Harare bond issues since then have had to carry its explicit guarantee. The only other city that has been able to raise bond finance since that time (Bulawayo) had to pay a large premium for the privilege and commit to special arrangements that may or may not be replicable. Broader economic circumstances aside, the policy regime governing sub- sovereign borrowing in Zimbabwe has created an unsustainable situation. Prescribed assets have created general problems for the financial sector. Ac- 352 Subnational Capital Markets in Developing Countries tions by the central government aimed at strengthening the municipal fi- nance sector (authorizations and implicit and explicit guarantees) have in- creased rather than diminished the government's liabilities while making it almost impossible to price municipal risk. Even if the macroeconomic situ- ation in Zimbabwe improves dramatically, this policy environment needs to be fundamentally reformed if a sustainable municipal debt market is to emerge. Market Discipline The policy regime also appears to have contributed to a decline in the qual- ity of local financial management and budgetary discipline. The period 1996­98 was characterized both by fairly high levels of borrowing from the public and private sectors and by deteriorating financial positions of many larger urban councils and growing local fiscal deficits, funded by short- and long-term debt.2 In 1995­97, for example, Kwekwe's annual budget deficit grew from Z$1.7 million to Z$13.8 million, and its aggregate debt burden from Z$2.3 million to Z$17.8 million (Steffenson and Trollegaard 2000). Harare has consistently been both the most active local borrower and the one with the worst financial record (Prud'homme 1999).3 Thus in a regulatory environment where the ability of a local authority to borrow on the markets has been determined by the actions and financial position of the central government rather than the soundness of the au- thority's own financial position, access to debt finance has failed to exert a disciplinary effect on local government borrowers. As Prud'homme (1999) has noted, "Recourse to the bond market does not seem to have been [driv- en by] financial wisdom and discipline. [Nor has] the ability of a city to go to the market [been] constrained by the soundness of its financial position and the existence of healthy accounts." Bond Issues and Inflation Until 1998, as a result of the prescribed assets regime, local governments were generally able to access the market at interest rates close to or below the inflation rate, borrowing at zero or negative interest rates and thereby benefiting from a hidden tax on financial institutions. Interest rates rose sharply on new issues after the 1998 Harare default. Once again in 2001, when the central government reduced and strictly regulated interest rates in an effort to contain its own soaring domestic debt, local governments benefited substantially from the spread between the real cost of money and the rates on their bond issues.4 Prud'homme (1999) has calculated that in Country Case Studies: Zimbabwe 353 real terms, with an annual inflation rate of 20 percent, a 15-year bond of US$100 is reimbursed for about US$6. This is clearly not a sustainable situation for financial institutions in the long term. Of course, the policy of prescribing financial assets is a concern that extends beyond the municipal sector. However, like many other mon- etary and fiscal policy questions, it requires serious attention if a sustain- able municipal debt market is to emerge in Zimbabwe. Only after funda- mental reforms in the overall policy environment will it be possible to address the issues that relate specifically to municipal borrowing: the inter- governmental fiscal structure, the powers of local governments to budget and issue debt, the regulation of market participants, and procedures in the event of default. Serious and detailed reform in all these areas is essential before a sustainable debt market can emerge in Zimbabwe, but under to- day's circumstances this is unlikely to happen in the short or even the medium term. Notes 1. These long-term bonds are known as "registered stocks" in Zimbabwe. 2. An analysis of the financial statements of 13 of the largest urban councils for fiscal 1998 confirmed the weak financial position of most of the councils. Of the 13 assessed, only 2 reported (modest) surpluses in fiscal 1998. Of the 11 councils in deficit, 4 were carrying deficits of more than Z$20 million. 3. In June 1999 the entire Harare council was dismissed by the central government on the basis of an extremely adverse report on the city's finan- cial position by Deloitte and Touche. 4. For example, in January 2002 interest on 90-day negotiable certifi- cates of deposit was 23 percent, while the annualized inflation rate was 112.4 percent. Chapter 20 Middle East and North Africa Morocco In a developing financial market with good potential, private investors are reluctant to lend to local governments that have little fiscal autonomy. Samir El Daher Lessons In Morocco, a unitary state with highly centralized governance, a national lending agency has dominated local government bor- rowing. The country is working to decentralize its governance and is still developing its domestic financial markets, which have had little experience in lending to subnational govern- ments. Municipalities depend heavily on centrally collected and administered revenues and have little flexibility in setting local rates. Even so, local governments are important service providers, with large capital spending needs. Morocco relies on a municipal development fund, the Commu- nal Infrastructure Fund, as a vehicle for ensuring access to cred- it for municipalities that are too small or too heavily dependent on the central government to tap credit markets directly. The country's experience with its lending program has been fairly positive, but resource constraints threaten to crimp future 355 356 Subnational Capital Markets in Developing Countries growth. The Communal Infrastructure Fund, almost the only source for long-term credit to subnational governments, has re- lied for resources largely on government funding and on lines of credit from official bilateral sources and multilateral institu- tions. Now diversifying its funding sources, it has been tapping domestic financial markets for long-term credit. The fund has is- sued medium- and long-term securities on domestic markets without government guarantee and has been steadily improv- ing its management of financial risks and conforming with pru- dential regulations. Deficiencies in financial management and reporting by subna- tional governments hamper their access to private credit, and centrally provided credit has limitations. Prescriptions of ways to address these shortcomings and limitations provide insight into the choices policymakers need to confront to make subna- tional access to credit more feasible. In the more than 20 years since Morocco introduced the basic laws for mu- nicipalities--the Law on Municipal Organization and the Law on Organiza- tion of the Finances of Local Governments--it has experimented with and gradually extended the fiscal and administrative framework of decentraliza- tion. During this period the country has strengthened decentralization by reforming local revenues, adopting formula-based intergovernmental transfers, and establishing a stable system of credit financing. Decentraliza- tion both poses great challenges and offers great potential--because of tight budgets, disparities in access to basic services, and the increased de- mand for basic services that has arisen with Morocco's rapid urbanization. The challenges of development require a good framework for budgetary processes, intergovernmental fiscal relations, the assignment of expendi- tures and revenues, and the development of subnational credit markets. In Morocco urban municipalities depend for 40 percent of their rev- enues on three centrally administered taxes: two shared taxes--an urban tax and a business tax--and a local property-based tax. They derive another third of their revenues from intergovernmental transfers, and the rest from Country Case Studies: Morocco 357 local fees and taxes. The share of revenues from own sources for local gov- ernments in Morocco appears to be comparable to that in many OECD countries. Local governments in Morocco depend less on central transfers than do municipalities in Latin America, for example, where transfers ac- count for more than half of revenues on average. However, these govern- ments have limited fiscal autonomy, with little discretion in determining the base or rate for shared and local taxes or even for user charges and fees for municipal services. An estimated 70 percent of local government rev- enues are subject to centrally set rates. If local governments are to meet their financing needs, they will have to rely more on local revenues. Most own-source revenues are fees and charges for services or for periodic activities (such as annual fairs), and mo- tor vehicle taxes generally accrue to the central level. Thus the most imme- diate prospect for increasing local revenues appears to be raising existing property taxes. Local governments also need greater freedom to determine local fees and charges for services. Morocco has recently reformed its intergovernmental transfer system to remove the perverse incentives for deficit spending and unwise borrowing. By eliminating the budget deficit subsidy, the reforms have imposed a hard budget constraint on local governments and ensured that they would bear the full cost of borrowing. The reforms have also introduced a formula for distributing value added tax revenue that has improved the equity of fiscal transfers to local governments as well as the predictability and transparen- cy of local revenues. Subnational Government Borrowing While Morocco has been liberalizing its financial sector and reforming its local government sector, the ability of subnational governments to access private credit remains largely untested. That ability depends on the quality of the fiscal and financial management, the budgetary and control systems, and the planning and implementation of their investment programs. One of the main determinants of subnational governments' creditworthiness and access to private funding under market conditions, however, is the le- gal and regulatory framework in which they generate revenues, manage as- sets, and finance and provide services. Morocco's relatively developed financial markets provide a potentially good base of long-term funding for subnational governments. Total bank credit in the country averaged around $15 billion, or about 50 percent of 358 Subnational Capital Markets in Developing Countries GDP, over the past five years. However, with the exception of a small num- ber of direct bond issues by Casablanca,1 local government borrowing has taken place predominantly through loans from the publicly owned, spe- cialized financial intermediary, the Communal Infrastructure Fund. Al- though no legal restrictions prevent subnational governments from bor- rowing from other financial institutions, borrowings from commercial banks have been limited, mainly because of regulatory and institutional impediments affecting the risk associated with subnational borrowers. Foremost among these impediments are the limited autonomy and author- ity of subnational governments in mobilizing revenues and managing ex- penditures. To foster a more efficient subnational finance system, the central gov- ernment has embarked on reforms to help expand the bankable demand for private credit and increase the participation of private financial institu- tions in subnational investment funding. Success in these reforms will be critical given the pressing need to reach beyond budgetary resources to meet the large investment funding requirements. The hope is that, in an appropriate enabling environment, commercial banks would be interested in the potentially growing subnational finance market--previously too fragmented and unknown to justify large commitments and investments in systems development and know-how. Main Issues in Subnational Borrowing In Morocco credit financing accounted for about 8 percent of the total re- sources of subnational governments, and debt service for 19 percent of their current expenditures, over the five-year period from 1996­2000. While there is both potential and scope for further expansion of subnation- al credit markets, this expansion would need to be supported by regula- tions and instruments adapted to the investment financing needs of subna- tional governments. It also would require easing the demand-side constraints that impede their access to credit. These constraints relate mainly to the following: · Subnational governments' lack of autonomy in fiscal decisionmaking coupled with poor planning and operating capabilities. · Inadequate, ad hoc, or inconsistent economic selection criteria and benchmarks used by the central government in its review of subna- tional investment programs. · The inefficiency of asset management by subnational governments. Country Case Studies: Morocco 359 · The absence of performance standards for locally provided monopoly services, which has resulted in poor quality of service delivery. · The inadequacy of cost recovery practices and their lack of differenti- ation between revenue-generating services and public goods. · Inadequate collection levels for user fees. In addition, mobilizing private resources for subnational infrastructure in Morocco is likely to require a greater role for private providers of ser- vices--through greater use of concession arrangements, for example. The provision of services remains too centralized and uncompetitive. Reforms aimed at improving local revenue mobilization and financial management focus on accounting, payment, financial control, and audit systems and on the incentives needed for timely collection of resources and reduced payment delays. Particularly important is greater capacity to gen- erate own-source revenues through rational bases and rates for local taxes and adequate collection levels. The recently approved formula-based distri- bution of value added tax revenue has improved the predictability of part of the revenues of local governments, though more needs to be done for other sources of shared revenues. In addition, the central government has been working to improve the efficiency and transparency of its extensive administrative, financial, and budgetary controls, with a view particularly to reducing perceptions among potential private lenders that the govern- ment implicitly guarantees subnational borrowings. Regulatory Environment for Subnational Borrowing On the supply side an increase in subnational borrowing in the domestic fi- nancial markets will depend on adjustments to laws and regulations related to borrowing authority and the issuance, registration, and servicing of debt. The governance and transparency of financial markets need to be im- proved through rules that avoid creating distortions by remaining neutral between market participants and between instruments (such as between loans and bonds). Moreover, the regulatory and supervisory framework for subnational borrowing needs to be strengthened through sound bankrupt- cy laws and prudential regulations. Bankruptcy of subnational governments needs to be governed by regula- tions on debt adjustments, bankruptcy initiation, debt workout plans, and allowed expenditures during bankruptcy proceedings. The protection of creditors' rights needs to be clarified by addressing such things as the au- thority to pledge assets as security for borrowing, to set up independent 360 Subnational Capital Markets in Developing Countries trusts to handle the recurrent revenues of projects, and to establish sinking funds to amortize repayments of principal. In particular, the authority to pledge assets needs to be defined by regulations that differentiate between categories of subnational assets, such as core assets that may not be sold or used as debt collateral and noncore assets that may be used as collateral. In addition, measures are needed to improve the reliability of asset registra- tion and valuation. In the area of prudential regulations the central government has im- posed limits and controls on subnational government borrowing, primarily through regulations relating to the receipt of loans from the Communal In- frastructure Fund. Debt service payments associated with subnational gov- ernment borrowing must be treated as obligatory expenditures recorded in local budgets. Funds are not sequestered, but the local government must approve the budget, and debt service is a priority expenditure. Although this budgetary commitment does not guarantee against default--since ex- penditures related to debt service still need to be authorized by the local government--it does make default less likely. Moreover, regardless of the implied comfort, private creditors can set their own exposure limits. The prudential rules of the Communal Infrastructure Fund, for example, limit eligibility for borrowing to subnational governments whose debt service does not exceed 40 percent of their combined own-source and value added tax revenues. Debt service is estimated to be in the range of 20 to 25 per- cent of total local budgets. Several reforms are needed to encourage private participation in subna- tional borrowing. An important focus of the regulatory reforms should be information to prospective creditors about the financial situation of subna- tional borrowers, particularly their indebtedness--whether direct liabilities or contingent liabilities such as guarantees. Disclosure rules should distin- guish between publicly and privately placed offerings, with private place- ments subject to less stringent disclosure. Disclosure also would provide a useful basis for credit rating, which in time might become a statutory re- quirement to allow creditors to assess their exposure to local governments. Audits are required, but they should be conducted by a government agency rather than an independent party. Types of Subnational Borrowing Most subnational borrowings in Morocco have been earmarked for specific infrastructure investments. Even so, the financial industry has not sought to develop specific infrastructure finance instruments that would entail a dis- Country Case Studies: Morocco 361 tinction between tax-supported (general obligation) and revenue-based (lim- ited obligation) borrowing schemes. Tax-supported borrowings allow subna- tional governments to service debt out of general revenues. They are often used to finance "public-good" projects that indirectly produce tax revenue (such as by increasing real estate values or stimulating economic growth and business investments) that can help meet debt service. In Morocco the prospects for effective tax-based funding schemes are not promising. Be- cause subnational governments lack the authority to determine the bases and rates for locally collected taxes and to ensure adequate collection levels, they are unlikely to capture the benefits of many public-good projects. Revenue-based borrowing, not yet used in Morocco, allows the borrower to meet debt service obligations from the revenues of the project financed by the debt. As a means of credit enhancement, these revenues are pledged to creditors as senior debt collateral. This requires segregating the revenues in an account out of which debt holders would be paid on a priority basis. Under existing laws, however, it is unclear whether subnational govern- ments in Morocco can pledge future revenues as security for borrowing. Revenue-based borrowings can be structured on a nonrecourse basis, in which creditors' claims are limited to the project's revenues or, at most, the project entity's assets. Because these revenues and assets can be segregated from other subnational government assets, nonrecourse finance can shield the central government from implicit contingent liabilities associated with subnational debt. Communal Infrastructure Fund: Intermediary for Subnational Borrowing Using a financial intermediary that can tap credit markets on behalf of sub- national borrowers is one way to foster market access for small and medi- um-size municipalities that still cannot directly access Morocco's long-term credit markets. The Communal Infrastructure Fund, which has been almost the only source of long-term credit for subnational governments, has relied for resources largely on government funding and on lines of credit from of- ficial bilateral sources and multilateral institutions. Recently it has sought to diversify its funding sources, and as part of its strategy it has been tap- ping the domestic financial markets for long-term credit. The Communal Infrastructure Fund has issued medium- and long-term se- curities (variable rate notes and certificates of deposit), without government guarantee, at small spreads over government debt issues. It has been steadily 362 Subnational Capital Markets in Developing Countries improving its management of financial risks, conforming with the prudential regulations of the banking authorities. Its issues, amounting to a dozen so far, have been relatively evenly spaced. In the 1990s the fund's lending account- ed for some 20 to 25 percent of total investment outlays by subnational (prin- cipally municipal) governments. In the past few years subnational credit has declined somewhat in real terms (as has subnational investment), despite ex- pectations that the volume of credit would grow as a result of the reforms in the intergovernmental transfer and local revenue schemes. The subnational credit market in Morocco remains relatively small, with the Communal Infrastructure Fund's portfolio of $500 million representing about 3 percent of outstanding bank credit. Its annual loan commitments range between $100 million and $200 million. Most of the fund's lending has gone to urban municipalities, which account for about 70 percent of outstanding credit. Rural communes, with small projects and limited debt service capacity, account for about 25 percent of outstanding credit. (The balance of the fund's lending goes to municipal corporations.) However, rural communes hold more than half the outstanding loans by number. The loans extended by the Communal Infrastructure Fund in the 1990s were largely for non-revenue-generating projects--roads (32 percent of total loans), electricity (16 percent, mainly to rural communes participating in na- tional electrification programs), solid waste management (10 percent), sew- erage (9 percent), sports facilities (6 percent), urban transport (3 percent), and water supply (2 percent). This sectoral distribution shows that the fund has been a relatively important source of financing for sewerage and solid waste projects. The small share of financing for water supply reflects the fact that water investments are the responsibility of the national water authority, which does not borrow from the fund. Loans for commercial infrastructure (such as local markets) have accounted for some 15 percent of the total. The Communal Infrastructure Fund has the potential to take on an in- creasingly important role in subnational finance. It has a valuable franchise with subnational governments, specialized knowledge, a strong equity base, and a broad and diversified project portfolio. It has been the preferred lending vehicle for subnational governments and can continue to mobilize funds on behalf of local governments, mainly from official bilateral and multilateral sources. Financial measures have been taken to strengthen its autonomy, financial viability, and capital base. Government guarantees are no longer provided for its domestic bond issues, and its provisions on loan arrears are now in line with the prudential requirements for the banking sector. Its repayment experience is good, with few loans in arrears. Country Case Studies: Morocco 363 With further reforms, the Communal Infrastructure Fund might be able to play more of a developmental role in subnational finance in the context of an open, competitive financial sector. The fund is expected to make its operations increasingly sustainable and to leverage its capital by issuing debt in private credit markets and lending the proceeds to subnational gov- ernments. It could expand its menu of services to include nonlending products, such as fee-based financial advisory services and technical assis- tance to larger municipalities. In time it could expand its range of borrow- ers beyond subnational governments to other providers of local services, whether public entities or private concessionaires. As a provider of long-term credit to subnational governments, the Com- munal Infrastructure Fund could participate in and sometimes lead bank- ing syndicates for larger loans to subnational governments and concessions for local infrastructure and possibly help in underwriting subnational bond issues. In pooling the credit demand of subnational borrowers, the fund could act as a bridge between subnational governments and institutional investors. This would be especially useful where subnational governments have little potential for directly accessing private credit markets and where institutional investors are in no position to assess the risk of individual sub- national governments but might be willing to assume, through the Com- munal Infrastructure Fund, a diversified exposure to the subnational gov- ernment sector. Private Financial Institutions in Subnational Borrowing Commercial banks have shown little interest in lending to local govern- ments in Morocco. Among the main factors inhibiting such lending are the limited autonomy of subnational governments, the weak institutional and management capacity, the lack of transparency and weak auditing stan- dards, and the lack of access to timely and reliable information on subna- tional governments' financial and operational performance. Added to these are the generic difficulties of perfecting collateral interests through cumber- some and uncertain recourse to the judicial system. Accordingly, despite the banks' large liquidity and their financial advantages over the Commu- nal Infrastructure Fund (which lacks access to lower-cost deposits and can- not finance private concessions), the few attempts by banks to enter the subnational finance market have been inconclusive. Most commercial bank term lending for infrastructure has been restricted to private conces- sions and, occasionally, municipal corporations. As a result, banks still lack 364 Subnational Capital Markets in Developing Countries familiarity with the legal, institutional, and regulatory framework in which subnational governments operate. Yet commercial banks' interest and involvement in the subnational fi- nance market could grow in line with demand if the demand is at a scale suf- ficient to sustain the business development costs involved for banks entering and competing in the infrastructure sector (particularly in water, sewerage, and solid waste, where future investments by subnational governments are expected to be substantial). Depending on the depth and effectiveness of re- forms in the subnational government sector, banks also could be attracted by prospects for increased profitability resulting from the expectation of greater credit demand, lower operational costs, and more manageable credit risks. Potential Role for the Bond Markets in Subnational Borrowing The institutional savings in Morocco could potentially provide a substan- tial funding pool for investments in subnational government projects. In- deed, given the shortage of attractive fixed income securities for pension funds, insurance companies, and mutual funds, debt securities issued by subnational governments could be appealing, particularly those issued by large urban municipalities with sufficient and stable resources. First, how- ever, obstacles to issuing public debt must be removed and regulations gov- erning financial transactions by subnational governments introduced (in such areas as financial disclosure, registration, underwriting, distribution, instruments). The capital market regulations now in effect do not respond to the needs of subnational issuers, particularly with respect to procedures and disclosure requirements. Because of limited technical capabilities and high transaction costs, only a few subnational governments in Morocco--the largest urban municipali- ties--would be likely to access the bond markets directly. Nonetheless, the development of a subnational securities market could foster the broader re- forms--in disclosure, financial transparency, management quality, audit- ing, asset collateralization, and credit rating capabilities--that are needed for independent assessment of the creditworthiness of local governments. Note 1. These bonds are not listed on an exchange and were distributed solely domestically. Chapter 21 Middle East and North Africa Tunisia Meeting most local government capital needs, the municipal development fund starts to borrow in the domestic capital markets. Samir El Daher Lessons A unitary and highly centralized state, Tunisia is making efforts to decentralize its governance and to develop its domestic fi- nancial markets. Private financial institutions have had little ex- perience in lending to subnational governments, which have re- lied for credit finance almost exclusively on a municipal development fund. Local governments rely heavily on centrally collected and administered revenues and have little freedom to set local rates. The municipal development fund in Tunisia is the cornerstone of the system for financing subnational investment, carried out as part of national planning. Combining loans and grants, the fund executes the central government's policy on financing subnational capital investment. Although the fund has little ex- perience in raising capital, it recently has issued bonds on the domestic market. 365 366 Subnational Capital Markets in Developing Countries The fund's sustainability would be enhanced by diversifying capital sources, broadening the range of services to local bor- rowers, and expanding its client base. Transforming the fund into an autonomous, commercially viable specialized financial institution may offer the best prospects for sustainability. Prop- erly done, the development of the fund into a market-based in- stitution--able to mobilize long-term resources on behalf of subnational borrowers--could catalyze the emergence and growth of a local government bond market in Tunisia. In Tunisia subnational governments have relatively limited responsibilities for financing and providing services and so have relatively limited needs for long-term borrowing. The main urban services--such as water, trans- port, sanitation, and electricity--are provided by national agencies. Invest- ments in health and education are also a national responsibility. Subna- tional governments are responsible mainly for roads, drainage, public order, pollution control, solid waste collection, and street cleaning and lighting. Their capital expenditures represent about 50 percent of their cur- rent revenues on average, though the share varies depending on the juris- diction, with large urban centers devoting a bigger part of their budgets to investment spending than small rural entities. The central government plays a major role in allocating resources and credit to subnational governments. Borrowing by subnational govern- ments has occurred almost exclusively through a specialized financial in- termediary wholly owned by the government, the Fund for Loans and Sup- port to Local Communities (Caisse de Prêts et Soutien aux Collectivités Locales, or CPSCL). The CPSCL is also the main channel through which the central government transfers resources to subnational governments to fi- nance their capital expenditures. Tunisia has no local government bond market in which subnational governments can issue debt, although it does have long-term credit markets for central government and corporate debt. The CPSCL's resources consist of a relatively large equity base provided by the government and external local government lines of credit provided by bilateral and multilateral financial institutions. Country Case Studies: Tunisia 367 Tunisia has 256 local governments, representing about 60 percent of the country's population; the other 40 percent live in unincorporated areas. Local governments are managed by governing councils that elect a presi- dent. The governments have inadequate financial management capacity, with accounting systems that need to be improved. For example, while tax registers have been computerized in local governments, tax receipts are still processed manually. Local governments depend on direct taxes for 32 percent of their re- sources. These include the hotel tax; the real estate tax; the undeveloped real estate tax; and the industrial, commercial, and professional establish- ments taxes. The taxes on real estate and undeveloped real estate were re- cently reformed, improving collections. Legislative and administrative changes in assessment values and improvements in accounting and com- puterization would also boost collections. Indirect taxes, mainly license fees for several types of small businesses, account for 8 percent of local gov- ernment revenue. Many of these local levies are of little value and could be eliminated to allow local governments to focus on those with the greatest revenue potential. As in Morocco, local governments need greater freedom to fix rates in accordance with local ability to pay. Fees for public services, such as a surtax on electricity and a tax for removal of industrial waste, ac- count for 24 percent of local government revenue. In addition, subnational governments receive two types of financial transfers from the state. Transfers for operating expenditures are made through a dedicated fund, the Common Fund for Local Communities, which provided 27 percent of local government revenue in 2000. The for- mula for distributing these transfers includes an incentive for improved tax collections and an element of cross-subsidy to aid the poorer local govern- ments. Transfers for investment financing come through one of two chan- nels: the sectoral ministries, which deposit funds directly with local gov- ernments for financing certain types of equipment, or the CPSCL, which provides both grant financing and loans. The fiscal situation of subnational governments remains precarious, re- flecting their weak savings and borrowing capacity. Their current expendi- tures rose faster than their current revenues in 1992­2002. Their invest- ments, though relatively small in volume, also increased more rapidly than their current revenues, with estimates suggesting that they doubled in cur- rent value in the same period. Moreover, a decline in the savings capacity of subnational governments has undermined their finances. Even so, the outlook for growth in subnational investment--and thus in subnational 368 Subnational Capital Markets in Developing Countries borrowing--remains good. Progress, however, will depend on further insti- tutional reforms addressing such aspects as decentralizing authority and in- volving the private sector in the provision of services through concessions, management contracts with capital expenditures, and other arrangements. Framework for Subnational Finance The capital investment projects of subnational governments and unincorpo- rated rural areas are listed in the Communal Investment Plan as part of a na- tional five-year economic development plan. Once the projects of subnation- al governments are included in the Communal Investment Plan, they are entitled to receive financing from the program. The financing plan for capital investment projects is based on a formula that outlines the following: · The share to come from self-financing or a subnational government's own resources (about 30 percent). · The share to come from grants or the central government's contribu- tion (30 to 40 percent). · The loan component granted by the CPSCL (30 to 40 percent). Local projects outside the framework of the Communal Investment Plan are revenue-generating (commercial) schemes, such as slaughterhouses or public markets. The Role of Credit in Financing Subnational Investment At 30 to 40 percent, borrowing represents a significant share of the financ- ing of subnational investment projects. Debt service accounts for some 15 to 20 percent of current revenues for subnational governments. Given lim- ited savings capacity, repayments of principal need to be spread out over six years on average, so as not to undermine budgets. Credit is allocated to subnational governments according to their size, a distribution policy intended to provide small subnational governments with the basic equipment needed even at the risk of increasing their debt relative to their own resources. Large cities have more room for maneuver but nevertheless are hampered by the rigidity of their resources from local taxes and intergovernmental transfers. The capacity of subnational govern- ments to finance their capital projects in line with the investment plans set out in the Communal Investment Plan, and to repay their debt, will de- pend on a real increase in their revenue receipts. In turn, that will require Country Case Studies: Tunisia 369 increasing central government transfers and subsidies, widening tax bases, and improving tax collection. Constraints on Subnational Governments' Access to Investment Finance The great diversity among Tunisia's subnational governments--in popula- tion size, economic resources, and the range of public services provided to residents--is mirrored in disparities in the capital bases and solvency. As a result, some subnational governments are better able than others to access credit, and a small number might be deemed creditworthy by potential creditors. However, the needs of these potentially creditworthy entities are large relative to the total demand for credit by subnational governments. For subnational governments with a substantial capital base, the central government could conceivably decide not to finance investments in proj- ects that have a strong potential of attracting private financing, such as sol- id waste treatment projects. In this case such projects might be financed by the CPSCL through arrangements involving private concessionaires and delegated service providers, although reforms would be needed to lay the groundwork for such private participation. Indeed, the government is considering possible changes in the way sub- national investments are financed. One option might be to progressively reduce, for a given period, the volume of CPSCL lending in a way congru- ent with the currently weak borrowing capacity of some subnational gov- ernments. An assessment should be made of the extent to which large sub- national governments might be able to take advantage of more flexible financing plans than those provided for under the Communal Investment Plan--and greater freedom in the choice of investments. For the subnational sector as a whole, the main constraint on invest- ment financing relates to the financial situation of governments. Substan- tial financial adjustments will be crucial to enable subnational govern- ments to carry out their investment programs. The Role of the Municipal Development Fund As Tunisia's municipal development fund, the CPSCL is the cornerstone of the system for financing subnational investment. Combining credit and grants, the CPSCL carries out the government policy on financing subna- tional capital investment under a set of procedures aimed at ensuring effi- cient distribution of investment credits. Since the CPSCL was created in 1975, the institutional framework governing its operations (legal status, 370 Subnational Capital Markets in Developing Countries procedures manual, investment guide) has been amended twice, in 1992 and 1997. These amendments increased the resources available to the CP- SCL but failed to endow it with sufficient autonomy in decisionmaking. Its mandate is still part of the relatively rigid framework under which the Communal Investment Plan is implemented. As the partner of subnational governments in financing their capital investments, the CPSCL ought to have greater autonomy. The CPSCL's activities remain relatively modest in scope because large in- frastructure investments are outside the purview of subnational governments and do not receive CPSCL financing. The policy on subnational capital in- vestment under the Communal Investment Plans has been giving priority to bringing infrastructure and basic equipment up to a common level across subnational governments. The CPSCL distributes its financing among the 256 subnational governments in accordance with the government policy on financing subnational capital investment. Tunis alone accounts for 12 per- cent of debt outstanding, and the country's 10 next largest towns for another 35 percent. With the investment effort by subnational governments exceed- ing 40 percent of their current revenues, it can be realized only by combining loans and grants to achieve a high average rate of subsidy. The CPSCL's financial prospects are linked to the resources and solvency of the subnational governments that are its customers. A recent strategic study commissioned by the CPSCL from expert consultants in local gov- ernment finance looked at the CPSCL's prospects for change and sustain- ability. The study led to proposals for redistributing tax revenue and charges between the central and subnational levels of government (with an increase in financial transfers from the central government) and between subnational governments and other institutions or authorities providing public services at the local level. Strengthening the borrowing capacity of subnational governments--and the growth potential of the CPSCL--also would require increasing their tax revenue and tax collection levels. With- out such measures, the CPSCL's activities could expand only very slowly or even decline, with serious repercussions for its profit margins and financial equilibrium. An Expanded Framework for the Operations of the Municipal Development Fund A review of the financial situation of subnational governments, particularly their debt ratios and their capacity to self-finance investments, suggests that Country Case Studies: Tunisia 371 the CPSCL needs to move beyond the uniform method of financing that it has been using. Introducing several different financing "windows" would al- low the CPSCL to adapt its assistance to the varied financial situations of sub- national governments. It would also enable the CPSCL to identify new op- portunities and to broaden the range of its activities, products, and clients. The CPSCL, for example, might have two windows: · One window, providing financing under the Communal Investment Plan, involving all transfers or subsidies provided on an off­balance sheet basis and for which the CPSCL would be acting as an agent for the government against a management fee. · A second window for more commercial financial operations, for fi- nancing investments outside the Communal Investment Plan, on terms reflecting the cost of resources mobilized by the CPSCL on pri- vate credit markets. In addition, the CPSCL might consider introducing instruments and products enabling it to offer its clientele a broader range of financial engi- neering and technical assistance services. As the CPSCL evolves, there might be a possibility of a third line of activities for financing operations by other categories of borrowers or clients, such as private entities to which subnational services are subcontracted. Such financing could occur only af- ter reforms allowing private concessionaires and subcontracted service providers to participate in subnational investment programs. By ensuring a reliable channel for credit for priority projects, the CPSCL has brought about significant improvements in the financing of subnational investments, often in difficult circumstances arising from the weak technical capabilities and financial and institutional constraints of subnational gov- ernments. For the CPSCL to be sustainable, however, it will need to boost its capacity to mobilize long-term financing and broaden its capital base by di- versifying the resources on which it draws. If the CPSCL is to take on an en- hanced role as a specialized financial institution and gain access to other types of resources, subnational finances must undergo adjustments. The ca- pacity of subnational governments to finance their capital investments and repay their debts depends on a significant increase in subnational receipts. Diversification of Activities, Products, and Clients Under the proposed strategy for making the CPSCL financially sustain- able, efforts should be made to strengthen the CPSCL's authority in com- 372 Subnational Capital Markets in Developing Countries mitting resources. Requests by subnational governments for financing in- vestments outside the Communal Investment Plan should be eligible for loans from the CPSCL to the extent that they meet its eligibility criteria, particularly the criteria relating to economic and financial returns (some of the projects now under the Communal Investment Plan might not ful- ly comply with the CPSCL's standards). Moreover, these loans should be granted on terms reflecting the true cost of borrowing by the CPSCL on private credit markets. The prospects for diversifying the CPSCL's clientele are limited in the short term. Consideration should be given to making CPSCL loans avail- able to private concessionaires providing local public services, subject to limits on the share of such loans in the CPSCL's activities. This financing of concessionaires would be unlikely to crowd out other options for subna- tional governments if it were limited to resources that the CPSCL could borrow on private credit markets. Financing requests from the most solvent subnational governments are expected to attract the interest of commercial banks and other sources of capital market financing over the next few years. Even so, the development of lending activity directed at the subnational sector remains a commit- ment that few domestic financial institutions are inclined to undertake. For this reason the CPSCL is still the agency best placed to respond to the fi- nancing needs of subnational governments, even if its skills in subnational financing may still need improvement. The CPSCL has yet to acquire all the characteristics and assets of an ef- fective specialized financial institution. Its lending activities and debt out- standing with subnational governments remain modest, reflecting the small share of subnational capital spending in the national investment ef- fort. Moreover, its portfolio is subject to credit risks arising from the precar- ious financial situation of many of its borrowers. While its activities are now governed by centralized administrative procedures and financing poli- cies under the Communal Investment Plan, under a new market-oriented regime the CPSCL would need to enhance its skills in the analysis of subna- tional finances, the evaluation of credit risk, and financial engineering, par- ticularly in project structuring. The continued importance of the policy of state subsidy of subnational investments means that the CPSCL would have to continue to finance sub- national investments through a combination of loans and grants. Howev- er, this role should not preclude the diversification of its financial products. The CPSCL should aim at steady growth in its turnover, by offering a wider Country Case Studies: Tunisia 373 range of options to subnational governments and by expanding its clien- tele to include private concessionaires and larger projects.