75732 UNTIL DEBT DO US PART UNTIL DEBT DO US PART Subnational Debt, Insolvency, and Markets Editors Otaviano Canuto and Lili Liu © 2013 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org Some rights reserved 1 2 3 4 16 15 14 13 This work is a product of the staff of The World Bank with external contributions. Note that The World Bank does not necessarily own each component of the content included in the work. The World Bank therefore does not warrant that the use of the content contained in the work will not infringe on the rights of third parties. The risk of claims resulting from such infringement rests solely with you. The findings, interpretations, and conclusions expressed in this work are those of the authors and do not neces- sarily reflect the views of The World Bank, its Board of Executive Directors, the governments they represent, or any other institutions with which the external authors may be affiliated. The World Bank does not 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 any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immuni- ties of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 Unported license (CC BY 3.0) http://­ creativecommons.org/licenses/by/3.0. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Canuto, Otaviano, and Lili Liu, Eds. 2013. Until Debt Do Us Part: Subnational Debt, Insolvency, and Markets. Washington, DC: World Bank. doi: 10.1596/978-0-8213-9766-4. License: Creative Commons Attribution CC BY 3.0 Translations—If you create a translation of this work, please add the following disclaimer along with the attribu- tion: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. All queries on rights and licenses should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org.   ISBN (paper): 978-0-8213-9766-4   ISBN (electronic): 978-0-8213-9767-1   DOI: 10.1596/978-0-8213-9766-4 Cover art: Michael S. Geller, Sea Surface Full of Clouds (1), 2009, oil on canvas, 36'' 3 36'' Cover design: Drew Fasick Library of Congress Cataloging-in-Publication Data Until debt do us part: subnational debt, insolvency, and markets / Otaviano Canuto and Lili Liu, Editors.     pages cm   Includes bibliographical references and index.   ISBN 978-0-8213-9766-4—ISBN 978-0-8213-9767-1 (electronic)   1. Debts, Public. 2. Bankruptcy. 3. Fiscal policy. I. Canuto, Otaviano. II. Liu, Lili (Economist)   HJ8015.U58 2013   336.3’4—dc23 2012046037 Contents Acknowledgments xv About the Editors and Contributors xix Abbreviations xxxi An Overview 1 Otaviano Canuto and Lili Liu Part 1 Subnational Debt Restructuring 31 1 Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 33 Alvaro Manoel, Sol Garson, and Monica Mora 2 Restructuring of Legacy Debt for Financing Rural Schools in China 81 Lili Liu and Baoyun Qiao 3 Managing State Debt and Ensuring Solvency: The Indian Experience 109 C. Rangarajan and Abha Prasad 4 Subnational Debt Management in Mexico: A Tale of Two Crises 145 Ernesto Revilla v vi Contents Part 2 Subnational Insolvency Framework 177 5 Colombia: Subnational Insolvency Framework 179 Azul del Villar, Lili Liu, Edgardo Mosqueira,  Juan Pedro Schmid, and Steven B. Webb 6 France’s Subnational Insolvency Framework 221 Lili Liu, Norbert Gaillard, and Michael Waibel 7 Hungary: Subnational Insolvency Framework 261 Charles Jókay 8 United States: Chapter 9 Municipal Bankruptcy— Utilization, Avoidance, and Impact 311 Michael De Angelis and Xiaowei Tian 9 When Subnational Debt Issuers Default: The Case of the Washington Public Power Supply System 353 James Leigland and Lili Liu Part 3 Developing Subnational Debt Markets 377 10 Transition from Direct Central Government Onlending to Subnational Market Access in China 379 Lili Liu and Baoyun Qiao 11 The Philippines: Recent Developments in the Subnational Government Debt Markets 417 Lili Liu, Gilberto Llanto, and John Petersen 12 Russian Federation: Development of Public Finances and Subnational Debt Markets 455 Galina Kurlyandskaya 13 South Africa: Leveraging Private Financing for Infrastructure 495 Kenneth Brown, Tebogo Motsoane, and Lili Liu 14 Caveat Creditor: State Systems of Local Government Borrowing in the United States 539 Lili Liu, Xiaowei Tian, and John Joseph Wallis Index 591 Contents vii Boxes 1.1 The 1993 Federal Immediate Action Plan 43 1.2 The Real Plan, 1994 44 1.3  The Fiscal Responsibility Law (Complementary Law No. 101, May 4, 2000) 52 3.1 State Borrowings 113 6.1  Subnational Financial Rules: Fundamental Principles, France 231 7.1  Major Acts Regulating the Municipal Sector, Hungary, 1990–2011 264 7.2  Powers of the Trustee: Article 14 (2) of Hungary’s Municipal Debt Adjustment Law 282 7.3  Restrictions on a Municipality Undergoing Debt Adjustment in Hungary 283 13.1 Section 139 Intervention in the Greater Johannesburg Metropolitan Municipality 499 Figures 1.1 National, State, and Municipal Tax Collection and Disposable Revenue in Brazil, 1970–2010 38 1.2 SELIC Interest Rate in Brazil, August 1994–April 2011 45 1.3 Consolidated Primary Fiscal Balance in Brazil, 2001–10 46 1.4 Primary Fiscal Balance of States and Municipalities in Brazil, 1991–2010 47 1.5 Net Public Debt in Brazil, 2001–10 56 1.6 Net Debt of States and Municipalities in Brazil, 1991–2010 56 1.7 Main Sources of State Revenues in Brazil, 2000–09 (2000 = 100) 59 1.8 Main Sources of Investment Financing for States in Brazil, 2000–09 62 1.9 Main Sources of Revenue for Municipalities in Brazil, 2000–09 (2000 = 100) 62 viii Contents 1.10 Main Expenditures of Municipalities in Brazil, 2000–09 (2000 = 100) 64 3.1  Composition of Financing Pattern of State Deficits (as of End-March) 115 3.2  Deficit and Debt as Share of GDP and Interest Payments as Share of Revenues 117 3.3 Extended Debt as Share of GDP 121 3.4 Primary Deficit as Percentage of GSDP 123 3.5  Box Plot Showing Debt-to-GSDP Ratio and Interest Burden Ratio 124 3.6 Interest Burden in Selected States 124 3.7  Differential between GDP Growth Rate and Interest Rate on State Debt 133 4.1 Subnational Debt in Mexico, by State, 2011 150 4.2  The Macroeconomic Impact of the 1994–95 Tequila Crisis in Mexico 156 4.3  Deteriorating Subnational Credit Scores in Mexico during the Global Financial Crisis 161 4.4  Fall in Transfers Relative to Previous Year in Mexico during the 1994–95 Tequila Crisis and the 2008–09 Global Financial Crisis 164 4.5  Federal Transfers: Evolution of Expectations in Mexico during the Global Financial Crisis, 2009 165 5.1 Government Structure in Colombia 182 5.2  Regional Transfers to Subnational Governments as a Share of Current Central Government Revenues, 1994–2010 184 5.3  Subnational Government Direct Debt as Percentage of GDP, 1990–2010 190 5.4 Restructuring Process under Law 550/1999 200 5.5  Departments and Municipalities under Debt Restructuring Agreement, 1999–2010 207 5.6  Department and Municipality Debt Balance as Percentage of GDP, 2000–10 211 6.1 Subnational Governments in France 224 6.2 Sources of Municipal Revenues in France 228 6.3 Sources of Departmental Revenues in France 228 6.4 Sources of Regional Revenues in France 229 Contents ix 6.5 Budgetary Control and Appeal Proceedings, France 235 6.6 SNG Borrowing and Debt Repayment in France 241 6.7 Distribution of Ratings Assigned to French SNGs 242 7.1 Bonds, Bank Loans, and Other Debt, Hungary, 2005–11 276 7.2 Cash and Liquid Financial Assets, 2004–11 278 7.3 The Debt Adjustment Process, Hungary 281 8.1 Annual Chapter 9 Filings, 1980–2011 322 8.2 Chapter 9 Filings by Type of Municipality, 1980–2007 323 10.1 Revenue and Expenditure of Subnational Governments in China, as Percentage of Total Government Revenue and Expenditure, 1985–2010 384 10.2 Top Five Countries Issuing Subnational Bonds, 2000–09 (Excluding the United States) 385 10.3 Rapid Urbanization in China, 1990–2010 399 10.4 Annual Land Transfer Fee in China, 2004–09 400 10.5 GDP Growth Rates and Subnational Revenue, 1991–2010 401 10.6 Formation of Debt Equilibrium, China 406 11.1 Distribution of Total Income, All Local Government Units, 2009 422 11.2 Composition of Revenues by Type of Local Government Unit, 2009 423 12.1 Composition of Consolidated Subnational Borrowing in the Russian Federation, 1995–2000 462 12.2 Subnational Defaults by Type of Debt Operation 464 12.3 Federal and Subnational Debt as a Share of GDP 477 12.4 Regions’ Share of Total Regional Bond Debt Outstanding at the End of 2008 478 12.5 Outlook on Credit Ratings: Russian Federation Regions and Subnationals in European Countries, End-December 2010 484 13.1 Trends in the Municipal Borrowing Market, South Africa, 2005–10 512 13.2 Metropolitan Municipality Capital Expenditure, South Africa, 2004/05–2009/10 513 13.3 Metropolitan Municipality Borrowing, South Africa, 2004/05–2009/10 514 x Contents 13.4 Outstanding Debt of Metropolitan Municipalities, South Africa, 2004/05–2009/10 515 13.5 Debt Composition of Metropolitan Municipalities, South Africa, 2004/05–2009/10 516 13.6 Debt Service Costs, South Africa, 2004/05–2009/10 517 13.7 South African Municipal Infrastructure Investment Requirements, 2010–19 520 Tables 1.1  Public Sector Borrowing Requirements (PSBR) in Brazil, 1999–2010 55 1.2  Share of GDP, Population, and Long-Term Debt of Borrower States in Brazil, 2009 57 1.3  Share of GDP, Population, and Long-Term Debt of Municipalities in Brazil, 2009 58 1.4 State Revenue in Brazil, 2000–09 58 1.5 State Expenditures in Brazil, 2000–09 61 1.6 Expenditures of Municipalities in Brazil, 2000–09 63 2.1  Statistics on Rural Junior Secondary and Primary Schools and Students, 2009 85 2.2  Assignment of Major Educational Responsibilities among Levels of Government in China 89 2.3  Financing Sources for Rural Junior Secondary and Primary Schools in China, 2008 90 2.4  Composition of Educational Expenditures of Rural Primary Schools, Province of Henan, 1999 and 2002 91 2.5  Own Revenues as a Percentage of Total Expenditures, by Level of Government in China, 2003 94 2.6  Compulsory Debt as a Percentage of Subnational Budgetary Expenditure for 14 Regions in China, 2009 95 B3.1.1 Sources and Features Attached to State Borrowings 113 3.1 Weighted Average Spreads during 2010–11 116 3.2  Deposit Rate of Major Banks for Term Deposits of More Than One-Year Maturity 119 3.3 States’  Vulnerability Matrix 125 Contents xi 3.4 Debt Forgiveness by Finance Commission 128 3.5 States’ Overdrafts and Access to Cash-Credit 131 4.1 Subnationals Resources in Mexico 148 4.2 Subnational Debt Structure in Mexico, 2011 152 4.3 Two Crises: Implications for the Subnational Debt Market in Mexico 159 4A.1 “Ramo 28.” Nonearmarked Transfers (Participaciones Federales), Mexico 169 4A.2 “Ramo 33.” Earmarked Transfers (Aportaciones Federales), Mexico 170 4A.3 Mexican States’ Total Local Revenue: Own-Source and Coordinated Federal Taxes 171 5.1  Total Revenues of Subnational Governments, Percentage of Total, 2006–10 186 5.2 Department Revenues 187 5.3 Municipal Revenues 188 5.4  Fiscal Performance of Municipalities under 550 Debt Restructuring Agreements Compared to the National Average for All Municipalities 191 5.5  Composition of Subnational Debt as Percentage of GDP, 2000–10 192 5.6  Channels for Control of Deficits and Debt: Lender-Borrower Nexus and Timing of Controls and Sanctions 194 5.7 Ex-Ante and Ex-Post Fiscal Legislation 195 5.8 Indebtedness Alert Signals 196 5.9 Colombian Superintendencies 205 5.10 Total Debt Restructured under Law 550, 1999–2010 206 5.11 Entities Restructured under Law 550, 1999–2010 206 5.12 Entities under Law 617 and/or Law 358, 1997–2010 210 6.1 New Powers Devolved to SNGs in 2004, France 225 6.2 Deadlines for the Regular Budget Process, France 232 6.3 Debt Data for French SNGs Rated by Fitch 243 6.4  Main Ratios Used by the French Central Government to Detect Financial Distress 249 7.1  Debt of Hungarian Municipalities Calculated at Prevailing Euro Exchange Rates 275 xii Contents 7.2  Municipal Debt Adjustment Filings, Hungary, 1996–2010 289 7.3 Main Causes of Bankruptcy in Hungary 292 8.1 State Authorization of Chapter 9 Bankruptcy 317 10.1 Urban Infrastructure Development in China: Selected Indicators, 1990–2009 387 10.2 Subnational Bonds Issued in China, 2009–11 395 10.3 Municipal Bonds Directly Issued by Four Cities in 2011 396 10.4 Relative Shares of Expenditure at Different Government Levels in China, 2003 402 11.1 Distribution of Total Expenditure, All Local Government Units, 2001–09 421 11.2 Local Government Borrowing and Debt Limitations: The Philippines 424 11.3 Local Government Finances: Key Ratios by Type of Unit, 2010 427 11.4 Budget Surpluses of Local Government Units, 2005–08 428 11.5 Structure of Philippine Local Government Debt Markets 433 11.6 Outstanding Loans and Bonds of LGUs (as of September 10, 2010) 434 11.7 LGU and Other Entity Outstanding Debt (with LGUGC Guarantee), 2010, by Type of Unit 438 12.1 Fiscal and Debt Rules for Subnational Governments 472 12.2 Annual Growth of Subnational Revenue, Including Fiscal Transfers, 2004–08 476 12.3 Intergovernmental Fiscal Transfers from the Russian Federation to Regions as a Percentage of Each Type of Transfer in Total Transfers, 2003–10 477 13.1 Secondary City Long-Term Borrowing, South Africa, 2004/05–2009/10 518 14.1 Year of First General Law for Municipalities and First Home Rule Law, United States 546 Contents xiii 14.2 State Constitutional Provisions Governing Local Debt and Borrowing Provisions, United States, 1841–90 550 14.3 State Constitutional Provisions on Local Government Debt Issue, United States 551 14.4 Government Debt by Level of Government, Nominal Amount, and Shares, United States, 1838–2002 556 14.5 State Monitoring of Local Fiscal Conditions, Home Rule, and Local Debt Restrictions, United States 563 14.6 Difference-in-Differences Estimates, Local Share of State and Local Totals, United States 572 14.7 Local Total Revenue, Expenditure, and Debt per Capita, United States, 1972, 1992, 2007 574 14.8 State Total Revenue, Expenditure, and Debt, United States, 1972, 1992, 2007 575 14.9 Combined State and Local Total Revenue, Expenditure, and Debt, United States, 1972, 1992, 2007 576 Acknowledgments This volume is the result of a team effort. We have many people to thank. First, we thank the authors. Many of them are leading practitioners and experts in public finance in the context of multilevel government systems. We are indebted to them not only for the quality of their contributions but also for the quests they have pursued tirelessly on the challenging ques- tions that this volume tries to address. Second, the findings of the volume have been shaped by the Sub-­ Sovereign Finance Forum – Debt, Insolvency, Markets held in June 2011 at the World Bank Headquarters in Washington, DC. The Forum comprised senior officials and practitioners from developed and developing coun- tries—including the countries covered by this volume—as well as acade- micians, international technical assistance providers, and World Bank and International Monetary Fund managers and staff. The discussions at the Forum have informed the development of the volume. Third, this volume is a part of the global knowledge program of the Economic Policy and Debt Department, PREM Network, of the World ­ Bank. We thank Jeffrey D. Lewis, Director; Carlos Braga, former Director; and Sudarshan Gooptu, Sector Manager of the Department, for their support. Fourth, our special thanks go to the Public Private Infrastructure Advi- sory Facility’s (PPIAF)–Subnational Technical Assistance Program, which financed significant portions of the Sub-Sovereign Finance Forum, the pro- duction of this volume, and the background research. We thank Adriana de Aguinaga de Vellutini, Manager; and James Leigland and Paul Reddel, former Managers of PPIAF, for their support. xv xvi Acknowledgments Fifth, we are grateful to colleagues and external reviewers who have helped the authors sharpen the messages and distill the lessons shared in this volume. Various chapters have also drawn from discussions with stake- holders in the field and during World Bank missions. In particular, we thank the following colleagues and external reviewers for their comments. Part 1 Subnational Debt Restructuring José Roberto Afonso, Economist of the National Bank of Economic and Social Development and Economic Advisor, Brazilian Congress, and Pablo Fajnzylber, José Guilherme Reis, and Rafael Barroso of the World Bank (chapter 1); Xiaoyun Zhang, Professor and Head of the Public Finance Group, Research Institute for Fiscal Studies, Ministry of Finance, China, Lezheng Liu, Associate Professor of Economics, Central University of Finance and Economics, Beijing, China, and Chorching Goh and Min Zhao of the World Bank (chapter 2); Dr. Vijay Kelkar, Chairman of the Thirteenth Finance Commission, India, Professor D. K. Srivastava, Director, Madras School of Economics, India, and Deepak Bhattasali of the World Bank (chapter 3); Arturo Herrera, Paloma Anos Casero, David Rosenblat, Jozef Draaisma, and Andrea Coppola of the World Bank, and Steven Webb, con- sultant (chapter 4). Part 2 Subnational Insolvency Framework Lars Christian Moller, Christian Yves Gonzalez, and Jose M. Garrido of the World Bank (chapter 5); Alban Aucoin, former Special Advisor to the ­ Chairman, ADETEF, French Ministry of Economy and Finance, France, Danièle Lamarque, former Director, Cour des Comptes, France, Michael De Angelis, Lecturer of Business Law, University of Rhode Island, United States, and Francois Boulanger of the World Bank (chapter 6); authorities of the Hungarian government, Jose M. Garrido and Riz Mokal of the World Bank, Michael De Angelis, and Mihaly Kopanyi, a consultant and former World Bank staff member (chapter 7); Joel Motley, Managing Director, Pub- Capital Advisors, United States, James E. Spiotto, Partner, Chapman and lic ­ Cutler LLP, United States, and Jean-Jacques Dethier, Jose M. Garrido, and Matthew D. Glasser of the World Bank (chapter 8); and Michael De Angelis, and Professor John Joseph Wallis of Economics Department, University of Maryland, United States (chapter 9). Acknowledgments xvii Part 3 Developing Subnational Debt Markets Xiaoyun Zhang, Professor and Head of the Public Finance Group, Research Institute for Fiscal Studies, Ministry of Finance, China, Lezheng Liu, Associ- ate Professor of Economics, Central University of Finance and Economics, Beijing, China, and Chorching Goh, Catiana Garcia-Kilroy, Zhi Liu, and Min Zhao of the World Bank (chapter 10); and Yan Zhang, Victor Vergara, José M. Garrido, and Lawrence Tang of the World Bank (chapter 11); Kaspar Richter, Stepan Anatolievich Titov, and Pavel Kochanov of the World Bank ­ (chapter 12); Sandeep Mahajan and Matthew Glasser of the World Bank ­ (chapter 13); Michael De Angelis, Lecturer of Business Law, University of Rhode Island, United States, Martha Haines, former Municipal Finance Director, U.S. Securities and Exchange Commission, Noel Johnson, Assistant Professor, Economics Department, George Mason University, United States, Isabel Rodriguez-Tejedo, Assistant Professor, University of Navarre, Spain, and Jean-Jacques Dethier and José M. Garrido of the World Bank (chapter 14). We thank Joel Motley, Managing Director, Public Capital Advisors, United States, for his comments on the chapters in Part 3. Sixth, we would like to thank various government officials and other people with whom we had discussions in the course of developing this volume. Chapters 2 and 10 benefited from discussions with the Ministry of Finance, China. The authors of chapter 3 would like to thank Ying Li for her valuable analytic and technical inputs. The author of chapter 4 would like to thank the extraordinary research assistance of Fernanda Márquez-Padilla and Rodrigo Sanchez-Gavito, and Tuffic Miguel and Emilio Pineda for shar- ing in fruitful discussions on the subject. Chapter 5 on Colombia benefited from discussions with central and local government officials, bankers, rating agency analysts, and superintendency officials during a World Bank mission to Colombia, October 4–6, 2010, and follow-up discussions on July 13, 2011, with Ministry of Finance and Public Credit officials on an earlier draft. Chap- ter 6 on France draws on a World Bank mission to France in July 2009. The mission team met national and subnational officials, representatives from Agence Française de Développement and from the rating agencies, and uni- versity professors. Chapter 7 on Hungary benefited from contributions by Laszlo Osvath, Gabor Peteri, and Miklos Udvarhelyi of LGID Ltd, and from a series of interviews with national and subnational government officials and members of the financial services community in Hungary. Chapter 11 on the xviii Acknowledgments Philippines draws from a World Bank mission to the Philippines in ­ January 2011, where the mission team met government officials, public financial insti- tutions, and private banks. Chapter 12 on Russia benefited from dialogue with analysts of rating agencies Fitch and Standard & Poor’s. The author would like to thank Yulia Gerasimova for her research assistance. Chapter 13 benefited from background work by DNA Economics of South Africa. The authors of chapter 14 would like to thank Jessica Hennessey and Isabel Rodriguez-Tejedo for sharing their research findings. Finally we would like to thank Ying Li for her extraordinary efforts on ensuring data consistency in various chapters. Finally, we are also grateful to colleagues who helped us move this com- plex project from manuscript to publication. They are Diane Stamm, our outstanding language editor; Stephen McGroarty, Mary Fisk, and ­ Alejandra Viveros, who guided us through each stage of the production process; Mayya Revzina, who guided us on the assignment of copyrights; Detre Dyson, Debbie L. Sturgess, and Tania Tejeda, who helped format the vol- ume; and Ivana Ticha, who efficiently put together the Sub-Sovereign Finance Forum and managed various stages of the process. About the Editors and Contributors Otaviano Canuto is Vice President and Head of the Poverty Reduction and Economic Management Network of the World Bank, a division of more than 700 economists and public sector specialists working on economic policy advice, technical assistance, and lending for reducing poverty in the Bank’s cli- ent countries. He assumed his position in May 2009, after serving as the Vice President for Countries at the Inter-American Development Bank since June 2007. Dr. Canuto provides strategic leadership and direction on economic policy formulation in the area of growth and poverty, debt, trade, gender, and public sector management and governance. He is involved in managing the Bank’s overall interactions with key partner institutions including the Inter- national Monetary Fund. He has lectured and written widely on economic growth, financial crisis management, and regional development. He has pub- lished more than 70 articles in economic journals and books in English, Por- tuguese, and Spanish, and he has spoken often on development policy and global economic issues. Dr. Canuto holds a PhD in Economics from the Uni- versity of Campinas in São Paulo, Brazil, and an MA from Concordia University in Montreal in Canada. He speaks Portuguese, English, French, and Spanish. Lili Liu is a Lead Economist in Public Sector and Institutional Reform, ­ Europe and Central Asia, at the World Bank. Until the fall of 2012, she was with the Economic Policy and Debt Department as the cluster leader on public finance at the subnational government level. She sits on the World Bank Urban and Transport Sector Boards, and co-chairs the Decentralization and Subnational Regional Economics Thematic Group, a Bankwide network with over 350 members. Previously, she led high-level policy dialogue and lending operations for India and other countries, covering development xix xx About the Editors and Contributors strategy, macroeconomic policy, public finance, trade, and infrastructure. She has authored numerous publications on public finance, subsovereign ­ finance and their linkages to macroeconomic frameworks, intergovernmen- tal fiscal systems, capital market development, and infrastructure finance. She has also led advisory services to the World Bank’s operations in many developing countries. Dr. Liu is a frequent speaker at international confer- ences and to visiting government delegations to the World Bank. She has a PhD and an MA in Economics from the University of Michigan, Ann Arbor, and a BA from Fudan University, Shanghai. Kenneth Willy Brown is Deputy Director-General: Intergovernmental Re- lations Branch of the South African National Treasury. Before entering the public sector, Mr. Brown had a career in teaching. Mr. Brown joined the ­National Treasury in 1998 as a Deputy Director: Financial Planning, and in 2001, he assumed the position of Director: Provincial Policy, which un- derpins the national transfers to provinces. He has also served as Chief Director: Intergovernmental Policy and Planning, and oversaw sector poli- cies that affect provinces and local governments. As Deputy Director- General, he oversees provincial and local government finances including subnational transfers. Mr. Brown holds an MA in Economics from the University of Illinois, Urbana-Champaign; a BA with Honors in Economics from the University of the Western Cape in South Africa; and a Primary Teacher’s Diploma. Michael A. De Angelis is a Lecturer on the Faculty of the College of Business of the University of Rhode Island. He is a former partner in the New York law firm Mudge Rose Guthrie & Alexander, where he specialized in public finance. Mr. De Angelis’s expertise is in U.S. and transitional and develop- ing economy government finance (sovereign and subsovereign), ­ including general obligation, utility, infrastructure, education, health care–related nonprofit organizations, bond funds, guaranties, and other project and securitized financing. He has served as counsel to issuers, underwriters, bor- rowers, guarantors, and lenders, and he participatied in the ­ development of financial products for governmental infrastructure financing transac- tions. The recent emphasis of Mr. De Angelis’s work has been on the de- velopment of the regulatory and institutional legal framework necessary for capital markets for sovereign and subsovereign government borrowers ­ About the Editors and Contributors xxi in transitional and developing economies, including over 25 countries and the Western Africa Economic and Monetary Union. Mr. De Angelis holds a JD and BA from Boston College, Boston, Massachusetts, and an MPA from Cornell University, Ithaca, New York. Azul del Villar has worked since 2002 in the Public Sector Group for Latin America at the World Bank on policy development and technical assistance loans with national and local governments, mostly to support their regional development, decentralization, fiscal sustainability, and public financial man- agement. She has participated in economic studies related to fiscal discipline, budget reform, debt management, and public spending efficiency. While at Fitch Ratings’ global infrastructure and project finance group from 2007 to 2010, she conducted analyses for project finance rating of debt issuances in the transportation and energy sectors. She worked in the Mexican Ministry of Foreign Affairs’ Organisation for Economic Co-operation and Development Department from 1999 to 2001, where she analyzed the recommendations of the Fiscal Affairs, Public Governance and Management, and Economic Growth Committees to be implemented in Mexico. Ms. del Villar earned a BS in Economics from the Universidad Iberoamericana in Mexico City and an MA in Economic Policy Management from Columbia University in New York. Norbert Gaillard is a French economist and independent consultant and is Visiting Professor at the Graduate Institute in Geneva. He has served as a consultant to the International Finance Corporation, the World Bank, the State of Sonora (Mexico), the Organisation for Economic Co-operation and Development, and the European Parliament. He has authored several research articles and book chapters on sovereign debt and credit rating agencies, the most recent ones of which are “The End of Gatekeeping: Underwriters and the Quality of Sovereign Bond Markets, 1815–2007,” coauthored with Marc Flandreau, Juan H. Flores, and Sebastián Nieto- ­ Parra; and “To Err is Human: Rating Agencies and the Interwar Foreign Government Debt Crisis,” coauthored with Marc Flandreau and Frank Packer. Dr. Gaillard has published two books: Les Agences de Notation (La Découverte, Paris, 2010), and A Century of Sovereign Ratings (Springer, New York, 2011). He received his PhD in Economics in 2008; his dissertation dealt with sovereign rating methodologies. He was a Fulbright Fellow at Princeton University, Princeton, New Jersey, during 2004–05 and 2006–07. xxii About the Editors and Contributors Sol Garson teaches Financial Management of States and Municipalities in the Public Policies, Strategies, and Development Program of the Federal University of Rio de Janeiro, Brazil. She is also a consultant specializing in public sector finances and budgeting. From 1975 to 2001, Dr. Garson was an economist at the National Bank of Economic and Social Development, during which she was seconded to the City of Rio Janeiro in 1993, and was Secretary of Finance of the City from 1996 to 2000. As President of the ­ Association of Secretariats of Finance of the Capital Cities from 1997 to 2000, she coordinated discussions about tax reform, the Fiscal Responsibil- ity Law, and other subjects of interest to Brazilian cities. In 2007, she served as Undersecretary of Fiscal Policy at the Secretariat of Finance of the State of Rio de Janeiro. Dr. Garson holds a PhD in urban and regional planning, and her publications include the book Metropolitan Regions: Why Don’t They Cooperate? (Letra Capital, Rio de Janeiro, 2009). Károly (Charles) Jókay is an expert in municipal finance and bankruptcy and teaches courses in municipal finance, public budgeting, and public man- agement in the Department of Public Policy at Central European Univer- sity in Budapest, Hungary. He has municipal finance and creditworthiness experience in Central and Eastern European countries, including Bosnia and ­ Herzegovina, Hungary, the former Yugoslav Republic of Macedonia, Roma- nia, and Serbia. He specializes in policy reform; municipal finance and bud- geting; operational implementation of fiscal decentralization, such as the regulation and management of municipal debt; and the improvement of grant-based policies. A regular consultant to the World Bank, Dr. Jókay has completed projects on municipal bond disclosure standards, public utility transformation and regulation in the ­ municipal services sector, and munici- pal debt regulation. He has a PhD and an MA in Political Science from the University of Illinois, Urbana-Champaign and a BA in Economics from the University of Michigan, Ann Arbor. Dr. Jókay established a family foundation to support the education of poor, rural children in the High School of the Reformed Church in Pápa, Hungary. Galina Kurlyandskaya is a Russian expert in the field of public finance and in- tergovernmental relations. She is the Director General of the Center for Fiscal Policy in Moscow, a Russian think tank on government finance and intergov- ernmental relations in transitional economies. Dr. Kurlyandskaya is providing research-based policy advice and technical assistance to central, regional, and About the Editors and Contributors xxiii local governments in the area of fiscal policy, public finance management, and intergovernmental relations both in Russia and in other developing countries. She received her PhD in Economics from the Institute for World Economy and International Relations, Russian Academy of Sciences, Moscow. James Leigland is the Technical Advisor to the Private Infrastructure Devel- opment Group, where he manages the Technical Assistance Facility. Prior to retiring from the World Bank in 2010, Dr. Leigland was Team Leader for Subnational Technical Assistance at the Public–Private Infrastructure Advi- sory Facility (PPIAF), a global, multidonor trust fund managed by the World Bank. His other positions at PPIAF since 2005 include Program Leader for Africa and Acting Program Manager. Before joining the World Bank, Dr. Lei- gland was Consulting Team Leader at the Municipal Infrastructure Invest- ment Unit, a project development company created by the South African government, which structured 45 private financing deals from 1998 to 2005, including public–private partnerships, privatizations, and municipal bond issues. Dr. Leigland served as Senior Urban Policy Adviser for East Asia at the U.S. Agency for International Development in the mid-1990s, and is a former faculty member at Columbia University in New York City, where he earned his PhD in Political Economy. He is the author of 35 professional publications on infrastructure, privatization, and subnational finance. Gilberto M. Llanto is Senior Research Fellow at the Philippine Institute for Development Studies. He was formerly Deputy Director General (Under Secretary) of the National Economic and Development Authority and Executive Director of the Agricultural Credit Policy Council. He was presi- ­ dent of the Philippine Economic Society during 2004–05 and was on the editorial board of the Philippine Journal of Development. He is currently a member of the Konrad Adenauer Medal of Excellence Committee, which gives awards to outstanding local government units in the Philippines. His research interests include local finance, housing finance, public finance, and growth economics. He has a PhD in Economics from the School of Econom- ics, University of the Philippines, Manila. Alvaro Manoel is Senior Economist in the World Bank’s Economic Policy and Debt Department where he advises on a wide range of issues, including debt management, debt sustainability analysis, subnational government public fi- nances, macro vulnerabilities, and fiscal policy. Before joining the Bank, he xxiv About the Editors and Contributors was Senior Economist at the Fiscal Affairs and African ­ Departments at the International Monetary Fund (IMF), where he participated in IMF ­ program discussions and technical assistance on fiscal issues in more than 20 coun- tries. Dr. Manoel also served at the Ministry of Planning and ­ Budget in Brazil where he was team leader of the Economic Advisers Council and coordi- nated a government task force that designed and drafted the ­ Fiscal Respon- sibility Act (1998–2000) and then negotiated its provisions with Members of the Brazilian Congress. Dr. Manoel also held the position of Deputy Secre- tary of the National Treasury at the Ministry of Finance in ­ Brazil. He holds a PhD in Economics from the University of São Paulo, Brazil, and is a former professor of Economics at the University of Brasília, Brazil. Mônica Mora is an economist with the Institute for Applied Economic Research (IPEA) in Brasilia, Brazil. She joined the IPEA in 1998, working in the ­ Macroeconomic Department. Her research and expertise are in the areas of public finance, fiscal policy, and macroeconomic coordination. Ms. Mora has also taught economics at the Brazilian Institute of Capital Markets since 2004. She holds both a BA and an MA in Economics from the Universidade Federal do Rio de Janeiro, Brazil. Edgardo Mosqueira has been a Senior Public Sector Specialist at the World Bank since 2003. His expertise and experience are focused on public policy, public management, decentralization, and land policies and management, and he has experience in public sector and governance reforms in Argentina, Bolivia, Colombia, Croatia, Ecuador, Mexico, Peru, and Serbia. Mr. Mosqueira served as Peruvian Minister of the Presidency, Minister of Labor, and in other high-level government positions in Peru. Before joining the ­Peruvian govern- ment, he was senior researcher at the Hernando de Soto Peruvian Institute for Liberty and Democracy. He was also Dean of the Universidad Peruana de Ciencias Aplicadas School of Law and an expert member of the United Nations Development Programme Commission on Legal Empowerment of the Poor. He holds an MA in International Public Policy from the Johns Hopkins University School of Advanced International Studies, Baltimore, Maryland and a JD from the Pontifical Catholic University of Peru, Lima. Tebogo Motsoane is a Senior Economist with the South African National Treasury (SANT). He joined SANT in 2004, at which time he worked on About the Editors and Contributors xxv ­ ubnational fiscal transfers. His research and expertise are in the areas of s local government finance and municipal debt finance. Mr. Motsoane has contributed to or authored several papers and has presented papers at vari- ous forums on municipal debt finance. His publications include “Leverag- ing Private Finance,” a chapter he contributed to the 2008 and 2011 Local Government Budget and Expenditure Review, published by SANT; and “Pri- vate Sector Investment in Infrastructure,” a chapter he contributed to the May 2009 Municipal Infrastructure Finance Synthesis Report, published by the World Bank. Mr. Motsoane holds a Diploma in Cost and Management Accounting from the University of Johannesburg, South Africa. John E. Petersen was an expert in public finance and a leading economist an- alyzing the municipal market. He was a member of the Municipal ­ Securities Rulemaking Board, professor at the George Mason School of Public Policy, Fairfax, Virginia, and a highly regarded expert on the municipal bond market. Petersen began his career as a capital markets economist at the Federal Dr. ­ Reserve Board. He was Director of Public Finance for the Securities ­ Industry Association and then Director of the Center for Policy Research and Analysis for the National Governors Conference in the 1970s, after which he ­ became Senior Director of the Government Finance Officers Asscciation’s Govern- ment Finance Research Center. In 1992, Dr. Petersen founded the Govern- ment Finance Group, a ­ consulting firm that provided international financial advisory and research services. From 1998 through 2002, he was a division director at ARD/Government ­ Finance Group. He had a BA in Economics from Northwestern ­ Wharton School University, Evanston, Illinois, an MBA from the ­ at the University of Pennsylvania, Philadelphia, and a PhD in Economics from the University of Pennsylvania. Dr. Petersen passed away in early 2012. Abha Prasad is Senior Debt Specialist with the Economic Policy and Debt Department of the World Bank. At the World Bank, as part of a core team, she has developed and has been leading the work on the Debt Manage- ment Performance Assessment tool, including for subnationals. She has been working on debt and debt management issues in developing countries (Bangladesh, Bhutan, Guyana, the Maldives, Moldova, Nicaragua, Nigeria, Tanzania, and Vietnam,) and in subnationals (Brazil, India, Indonesia, Nigeria, and Peru). Previously, she was Director of the Internal Debt Management Department of the Reserve Bank of India, where she helped manage the xxvi About the Editors and Contributors debt of India’s federal government and 28 subnationals. She provided policy advice for framing India’s fiscal responsibility legislation and restructuring state government debt in 2004. She has published on many issues includ- ing debt, subnational finances, banking, and operational risk management in debt management. Ms. Prasad holds Master’s degrees from the University of Delhi, India; and Georgetown University, Washington, DC. Baoyun Qiao is a Professor of Economics and Dean, China Academy of Public Finance and Policy, Central University of Finance and Economics, in Beijing, China. Dr. Qiao has been invited by many governments and ­ international or- Thailand, and the Asian ganizations, including China, Indonesia, Pakistan, and ­ Development Bank and the World Bank, to deliver keynote speeches and presentations on intergovernmental fiscal relations, fiscal ­decentralization, local governance, subnational debt and management, and subnational in- frastructure financing. Dr. Qiao has published ­numerous books and papers, and his research concentrates on the areas of public finance, taxation, in- tergovernmental fiscal relation, budget management, local governance, and economic development. He served as the Vice President of the Chinese Economist Society during 2007–08. Dr. Qiao graduated from Georgia State University, Atlanta, where he earned a PhD in Economics. C. Rangarajan is the Chairman of the Prime Minister’s Economic Advisory Council, India. Formerly, he held several distinguished positions as Chairman of the Twelfth Finance Commission, the Governor of Andhra Pradesh, and the Governor of the Reserve Bank of India. He was part of the core team that was responsible for initiating the liberalization process in India. He is a distinguished former member of the Indian Parliament. Dr. Rangarajan has taught at several institutions including the University of Pennsylvania in Philadelphia, New York University in New York City, and the prestigious Indian Institute of Management, in Ahmedabad. In 2002, the Government of India awarded him the Padma Vibhushan, India’s second-highest civil- ian award. His work on the Finance Commission for the distribution of ­ resources between the center and state governments in India is an exem- plary example for decentralizing countries in managing intergovernmental relations and transfers. He has published on many subjects including fiscal federal transfers, monetary policy, agricultural growth, banking, macroeco- nomics, and forecasting. About the Editors and Contributors xxvii Ernesto Revilla is Chief Economist of the Mexican Ministry of ­ Finance. ­ Before assuming his current position, he was Chief of the Tax Policy Unit at the Ministry, in charge of the design and implementation of revenue policy and programs, economic analysis for tax policy decisions, and estimates for the budget and fiscal policy choices. Mr. Revilla also served as Director General of Fiscal Federalism at the Ministry, where he was in charge of federal transfers to subnationals and the design of policy related to the Mexican intergovernmental fiscal relations framework; as Chief of Staff of the Undersecretary of Revenues; and as Deputy Director General of Securities Markets. Prior to joining the Ministry, Mr. Revilla served at the World Bank Headquarters in Washington, DC. As a member of the Bank’s Young ­ Professionals Program, he was an econo- mist in the Financial Sector Development ­ Vice Presidency and then in the East Asia and Pacific Region. Mr. Revilla holds degrees in ­economics from the Instituto Tecnológico Autónomo de  México (ITAM) and the University of Chicago in Illinois. He currently teaches Advanced Macro- economics at ITAM. Juan Pedro Schmid is currently the Country Economist for Jamaica at the Inter-American Development Bank (IDB), responsible for a broad range of ­ issues including monitoring of the macroeconomy, preparation of the country strategy, and programming and analytical work. Before joining the IDB, Dr. Schmid worked for three years as an economist at the Eco- nomic Policy and Debt Department of the World Bank. While his main area expertise was debt relief under the Heavily Indebted Poor Country initia- of ­ tive, he also contributed to work in debt management, growth analytics, and fiscal policy. Dr. Schmid holds an MA in Economics from the University of Zurich and a PhD in Economics from The Federal Institute of Technology, both in Zurich, Switzerland. Xiaowei Tian is a staff member in the Department of Finance of Gansu Province, China. He has worked on public expenditure, budgetary man- agement, and fiscal policies on agriculture and ecological protection. Mr. Tian was a consultant in the Economic Policy and Debt Department of the World Bank’s Poverty Reduction and Economic Management Network. He worked on public finance, the impact of the global crisis on government finance, particularly at the subnational level, and on the subnational debt xxviii About the Editors and Contributors market development in major economies. He holds an MS in Economics from the University of Illinois at Urbana-Champaign. Michael Waibel is a University Lecturer in Law and a Fellow of the Lauterpacht Centre for International Law at the University of Cambridge, ­ Cambridge, United Kingdom. His main research interests are public interna- tional law and international economic law, with a particular focus on finance and the settlement of international disputes. At Cambridge, he teaches in- ternational law, World Trade Organization law, and European Union law. He has also taught at Harvard College in Boston, Massachusetts; the London School of Economics (LSE) in London, United Kingdom; and the University of St. Gallen in St. Gallen, Switzerland. He is the author of the monograph “Sovereign Defaults before International Courts and Tribunals,” published by Cambridge University Press in 2011. Dr. Waibel holds an LLM and a JD from the Universität Wien, Vienna, Austria; an MSc in Economics from LSE; and an LLM from Harvard Law School. He is a member of the New York Bar and holds a diploma from the Hague Academy of International Law in the Netherlands. He has worked for several international organizations includ- ing the European Central Bank, the International Monetary Fund, and the World Bank. John Joseph Wallis is Professor of Economics at the University of Maryland, College Park, and a Research Associate at the National Bureau of Economic Research. He is an economic historian and an institutional economist who focuses on the dynamic interaction of political and economic institutions over time. As an American economic historian, he has collected large data sets on government finances and on state constitutions and has studied how political and economic forces changed American institutions in the 1830s and 1930s. In the past decade, his research has expanded to cover a longer period, a wider geography, and more general questions of how societies use institu- tions of economics and politics to solve the problem of controlling violence and, in some situations, sustain economic growth. Dr. Wallis is the co-author of Violence and Social Orders: A Conceptual Framework for Interpreting Recorded Human History, written with Douglass North and Barry Weingast (Cambridge University Press 2009). Dr. Wallis received his PhD from the Uni- versity of Washington, Seattle, in 1983 and spent two years as a postdoctoral fellow at the University of Chicago. About the Editors and Contributors xxix Steven B. Webb has for the past 24 years worked for the World Bank as lead economist, adviser, and consultant in research, evaluations, and opera- tions in Africa, Latin America, the Caribbean, and other regions. His special- include political economy, public finance, central banks and monetary ties ­ policy, decentralization, and economic history. His publications include nu- ­ merous articles and several books: Hyperinflation and Stabilization in Wei- mar Germany; and with co-authors, Achievements and Challenges in De- centralization: Lessons from Mexico; Public Sector Reform—What Works and Why?; Voting for Reform: The Politics of Adjustment in New Democra- cies; and In the Shadow of Violence: Politics, Economics and the Problems of Development. Dr. Webb has also taught in the Economics Department at the University of Michigan, Ann Arbor, and was a visiting senior econo- mist at the U.S. Department of State. He received a BA in economics from Yale University in New Haven, Connecticut, and an MA in History and a PhD in Economics from the University of Chicago in Illinois. Abbreviations ARI average regular income ARO anticipated revenue credit operations BLGF Bureau of Local Government Finance BPA Bonneville Power Administration CGCT  General Code of Territorial Entities, Code Général des Collectivités Territoriales DAF Fiscal Support Agency, Dirección General de Apoyo Fiscal DBSA Development Bank of Southern Africa DCED  Department of Community and Economic Development, Pennsylvania DGCL  Direction Générale des Collectivités Locales, General Directorate of Subnational Entities DGCP  General Directorate of Public Accounting, Direction Générale de la Comptabilité Publique DOF Department of Finance a euro ESF Emergency Social Fund EU European Union FC Finance Commission FEIEF  Fund for the Stabilization of the Federal Revenue for the Federal Entities, Fondo de Estabilización de los Ingresos de las Entidades Federativas FIFA Fédération Internationale de Football Association FINDETER Subnational Development Financial Entity, Financiera de Desarrollo Territorial, SA xxxi xxxii Abbreviations fisc  the combination of a government’s fiscal activity, which includes revenues, expenditures, and debts FONPET National Pension Fund, Fondo Nacional de Pensiones en las Entidades Territoriales FPE State Participation Fund FRL Fiscal Responsibility Law, Brazil FRL fiscal responsibility legislation, India FUNDEF Financial Fund for Educational Services GDP gross domestic product GFI government financial institution GJMC Greater Johannesburg Metropolitan Council GNP gross national product GSDP gross state domestic product GTS General Transfer System, Sistema General de Participaciones IBGE Brazilian Institute of Geography and Statistics ICMS  value-added tax on goods, intermunicipal transportation, and communications services IFI International Financial Institution IGP-DI General Price Index, Domestic Availability IMF International Monetary Fund INFIS  Institutes for Territorial Promotion and Development, Institu- tos de Fomento y Desarrollo Territorial IPI value-added tax on industrialized products IRA internal revenue allotment IRRF income tax withheld at source ISS service tax KWK Kloha, Weissert, and Kleine LGC Local Government Commission, North Carolina LGU Local Government Unit LGUGC Local Government Unit Guarantee Corporation LWUA Local Water Utilities Administration MAV national railways MDFO Municipal Development Fund Office MEC Member of the Executive Council MFMA Municipal Finance Management Act MFPC  Ministry of Finance and Public Credit, Ministerio de Hacienda y Crédito Público Abbreviations xxxiii MHCP Ministry of Finance, Ministerio de Hacienda y Credito Público MOF Ministry of Finance MSRB Municipal Securities Rulemaking Board MVM power grid NCR net current revenue NPD  National Planning Department, Departmento Nacional de Planeación NSSF National Small Savings Fund OCIP Orange County Investment Pool OECD Organisation for Economic Co-operation and Development OTB Off-Track Betting Corporation OTP  former monopoly state-owned savings bank privatized in 1995, Hungary PACER Public Access to Court Electronic Records PAF  Incentive Program for the States’ Restructuring and Fiscal Adjustment PAI Federal Immediate Action Plan (1993) PFI private financial institution PICA Pennsylvania Intergovernmental Cooperation Authority PPP public-private partnership PSBR public sector borrowing requirements PSUs public sector undertakings PUDs public utility districts RBI Reserve Bank of India RCA Regional Chambers of Accounts RCL net current revenue RDF rainy day funds RLR real net revenue S&P Standard & Poor’s SEC U.S. Securities and Exchange Commission SELIC Special Settlement and Custody System SELIC rate prime interest rate: average charged on daily operations backed by treasury bills and bonds, registered at SELIC SEMs public-private partnerships, sociétés d’économie mixte locales SGP Central Government Transfers SNG subnational government SOC Superintendency of Corporations xxxiv Abbreviations SOE state-owned enterprise STN National Treasury Secretariat TSS Tax Sharing System UDC Urban Development Corporation UDIC Urban Development and Investment Corporation URV real value unit USAID United States Agency for International Development VAT value-added tax WPPSS Washington Public Power Supply System WMA ways and means advances ZAC  Centers of Urban Development, Zone d’Aménagement Concerté Note: All dollar amounts are U.S. dollars unless otherwise stated. An Overview Otaviano Canuto and Lili Liu Subnational Debt and Insolvency State and local debt and debt of quasi-public agencies have grown in importance. Three structural trends have contributed to the rising share of subnational finance, including subnational debt, as a share of general public debt (Canuto and Liu 2010a). First, decentralization in many countries has given subnational gov- ernments (SNGs)1 certain spending responsibilities, revenue-raising authority, and the capacity to incur debt. With sovereign access to finan- cial markets, SNGs are seeking access to these markets as well. Second, rapid urbanization in developing countries requires large- scale infrastructure financing to help absorb influxes of rural popula- tions.2 Borrowing enables SNGs to capture the benefits of major capital investments immediately, rather than waiting until sufficient savings from current income can be accumulated to finance them. Infrastruc- ture investments benefit future generations, which should bear a por- tion of the cost. Subnational borrowing finances infrastructure more equitably across multigenerational users of infrastructure services because the debt service can match the economic life of the assets that 1 2 Until Debt Do Us Part the debt is financing. Infrastructure services thus can be paid for more equitably by the beneficiaries of the services. Third, the subnational debt market in developing countries has been going through a notable transformation. Private capital has emerged to play an important role in subnational finance, and subnational bonds increasingly compete with traditional bank loans. Notwithstanding the temporary disruption of the subnational credit markets during the 2008–09 global financial crisis, the trend toward more diversified sub- national credit markets is expected to continue. SNGs or their entities in various countries have already issued bond instruments (for exam- ple, in China, Colombia, India, Mexico, Poland, the Russian Federation, and South Africa). More countries are considering policy frameworks for facilitating subnational debt market development (for example, ­ Indonesia), while others are allowing selected subnational entities to pilot-test transaction and capacity-building activities (for example, Peru). With debt comes the risk of insolvency. When SNGs follow unsus- tainable fiscal policy, it can jeopardize the ability to service their debt, the services they manage, the safety of the financial system, their country’s international creditworthiness, and overall macroeconomic stability. Too often, the central government gets dragged in to provide bailouts, which can disrupt its own fiscal sustainability and reward the populist fiscal tactics of the recipient SNGs. Several major emerging markets experienced subnational debt crises in the 1990s. Newly decentralized countries face potential fiscal risks. To many observers, runaway provincial debt in the Provinces of Men- doza and Buenos Aires was a factor behind Argentina’s sovereign debt default in 2001. Brazil experienced three subnational debt crises in the 1980s and 1990s. In India, many states experienced fiscal stress in the late 1990s to the early 2000s, with increases in fiscal deficits, debt, and contingent liabilities. The 1994–95 Tequila Crisis in Mexico exposed the vulnerability of subnational debt. Subnational insolvency is a recurring event in history. In 1842, eight U.S. States and the Territory of Florida defaulted on their debt, and three other states were in perilous financial condition ( Wallis 2005). During the Great Depression, 4,770 local governments ­ defaulted on US$2.85 billion of debt (Maco 2001). As capital mar- kets and their regulatory framework matured, the default rates of U.S. An Overview 3 local governments declined. Yet there are recent episodes, including the default of the Washington Public Power Supply System in 1983, and the bankruptcy of Orange County, California, in 1994 and of Jefferson County, Alabama, in 2011. ­ The 2008–09 global financial crisis has had a profound impact on subnational finance across countries, as a result of slowing economic growth, the rising cost of borrowing, and deteriorating primary bal- ances. The impact has been mitigated in various countries by fiscal stim- ulus, monetary easing, and increasing fiscal transfers. However, looking forward, pressures on subnational finance are likely to continue—from the potentially higher cost of capital, the fragility of global recovery, refi- nancing risks, and sovereign risks (Canuto and Liu 2010b). Aligning Fiscal Incentives Subnational debt crises have led governments across countries to search for frameworks to restructure subnational debt and to undertake legal, regulatory, and institutional reforms that will sustain subnational debt finance in the long run. In a multilevel government system, the reforms need to resolve three challenges (Liu and Webb 2011). The first chal- lenge applies to governments at any level, whereas the second and third are relevant mainly in countries with multilevel governments. The first challenge is the short time horizon of public officials, who have shorter terms of office than citizens’ life spans. Public officials face the risk of being forced out of office if results are painful in the short term. The mobility of citizens and businesses between local jurisdic- tions means that excess borrowing could drive residents away and leave those remaining with more debt per person than they had anticipated. The second challenge is free riders. The interests of individual sub- national governments may diverge from the common national interest when factors such as electoral pressures motivate subnational govern- ments to follow unsustainable fiscal policy. An individual government would bear only part of the cost of its misbehavior, but would still receive all of perceived benefit accrued, only if (most of) the other gov- ernments continued to follow good fiscal behavior. So, there might be a prisoners’ dilemma—a situation where the equilibrium of isolated indi- vidual choices leads to suboptimal outcomes for all.3 4 Until Debt Do Us Part The third challenge is moral hazard. Subnational borrowers might have an incentive not to repay their creditors, and creditors might lend without risk differentiations, if they perceive that defaulting debtors could be bailed out by the central government. In a country with multilevel governments, the national govern- ment exists for the purpose (among others) of protecting the common interest, and typically has special powers such as running the central bank and regulating the financial sector. The national government also provides transfers to the SNGs, giving it additional leverage over SNGs and their fiscal behavior. However, the constitution and rules (such as on revenue sharing) may constrain the national government’s power over the SNGs. Political considerations, such as the national political cycle or subnational political cycles, may bias the deci- sions of the national government away from the optimal (Braun and Tommasi 2004). For instance, when a state government of the same political party as the national government faces a close election, the national government might be inclined to “condone” the state’s fis- cal misbehavior by offering a debt bailout or rescheduling guarantee. Also, under some configurations of political institutions, the national executive might “purchase” blocks of legislative votes by giving SNGs fiscal favors. The incentives in the political system affect the need for effective subnational fiscal control institutions. To the extent that the constitu- tion and party system lead to more centralized power, the country may have less need for special institutions to coordinate fiscal discipline across governments over time and among SNGs. Decentralization and market decontrol, however, increase the need for coordination of fiscal discipline. The subnational debt crises or fiscal stress of the 1990s in several major developing countries led to reforms in subnational borrowing frameworks, including the development of ex-ante fiscal rules and debt ­ limits. The search for insolvency resolution has also intensified, since ex-ante rules have been shown not to be sufficient on their own without ex-post mechanisms. Insolvency mechanisms should increase the pain of circumventing ex-ante regulation for creditors and debtors, thereby enforcing preventive rules. An Overview 5 Key Design Issues in Subnational Debt Restructuring The country experiences in this volume reveal several design issues with respect to debt restructuring frameworks: (a) how to balance the tension between the contractual rights of creditors and the need for maintaining public services in the event of subnational insolvency; (b) how to define the respective roles of different levels and branches of government in resolv- ing insolvency; (c) how to develop a collective framework for debt reso- lution; and (d) a basic choice among a judicial, administrative, or hybrid approach. The country cases show that country-specific circumstances— historical, constitutional, and economic context, and entry points for reform—influence the framework design in each country. The framework design ultimately needs to address the challenges of fiscal incentives facing SNGs in a multilevel government system. A sound framework should reduce the moral hazard of subnational defaults, discourage free riders, bind all SNGs to pursue sustainable fis- cal policies, and extend the short-term horizon of SNGs to minimize the impact of unsustainable fiscal policy on future generations. Public and Private Insolvency Insolvency of subnational governments differs from that of private cor- porate entities. The core difference is the public nature of the services provided by SNGs. Thus, debt restructuring inevitably involves a dif- ficult balance between interests of the debtor (and the citizens it serves) and the creditors (and savers). While a corporation can be dissolved, this route is barred for SNGs. When a private corporation goes bankrupt, all of its assets are potentially subject to attachment. The ability of creditors to attach assets of SNGs is constrained in many countries. In the United States, a judicial doctrine typically holds that only proprietary property is attachable. “Proprietary property,” subject to debt foreclosure, was defined by the U.S. Supreme Court as “held in (the municipality’s) own right for profit or as a source of revenue not charged with any public trust or use”4 (McConnell and Picker 1993). Who Has the Authority over What? Fiscal adjustment by debtors requires difficult political choices to bring spending in line with revenues and to bring borrowing in line 6 Until Debt Do Us Part with debt service capacity. In a decentralized system, tension exists between the role of the national government in enforcing collec- tive fiscal discipline of SNGs and the fiscal autonomy of SNGs. Can a higher-tier government force spending cuts and tax increases in a lower-tier government? Can courts influence spending priorities and tax choices that are normally reserved for legislative and executive branches? How do a country’s legal framework and political reality define the roles of different tiers and branches of the government? These are among the key issues that the case studies in this volume try to address. Subnational fiscal adjustment is also complicated by the legis- lative mandates of the central government vis-à-vis subnational ­ governments and the intergovernmental finance system (Ianchovi- china, Liu, and Nagarajan 2007). Unable to issue their own currency, subnationals cannot use seigniorage finance. SNGs may not freely adjust their primary balance due to legal constraints on raising their own revenue, dependence on central government transfers, or the cen- tral government’s influence on key expenditure items such as wages and pensions. Many other policies that affect economic growth and fiscal health of the subnational economy may also be determined largely by the central government. Debt restructuring and debt discharge are complex processes but can be distilled into two basic questions: whether the creditors and the debtor can reach agreement on debt resolution; and who holds the cramdown power5 when both sides fail to reach an agreement (Liu and Waibel 2009). In Brazil and Mexico, the national government led SNG debt restructuring, and there were no debt write-offs. In Hungary and the United States, the courts hold cramdown power when local govern- ments and creditors negotiate. Clarity of Rules and Collective Enforcement Without an insolvency framework, subnational debtors and their creditors resort to ad-hoc restructuring negotiations. The need for a collective framework for resolving debt claims is driven not only by conflicts between creditors and the debtor, but also by competing interests among creditors and competing demands by constituents of the debtor. Individual creditors may have different security provisions An Overview 7 for the debt owed to them and may demand preferential treatment and threaten to derail debt restructurings voluntarily negotiated between a majority of creditors and the subnational debtor—the “holdout prob- lem.” Individual ad-hoc negotiations can be costly and harmful to the interests of a majority of creditors (McConnell and Picker 1993). The holdout problem is not as serious if debts are concentrated in a few banks. A collective framework for restructuring takes on more impor- tance as the subnational bond market develops and grows to include thousands of creditors. Lack of clear rules for insolvency is likely to raise borrowing costs, and may limit market access for creditworthy borrowers. South African policy makers viewed clear rules for insolvency as critical to the growth of a broad-based competitive subnational capital market. In the United States, utilization of Chapter 9 of the Bankruptcy Code has carried a strong stigma for a defaulting municipality, to offset debtor moral haz- ard. Municipalities are thus wary that capital markets would interpret the filing for federal bankruptcy protection as a strong signal of finan- cial mismanagement, to which lenders are likely to react by charging a risk premium. The tension between maintaining essential services and creditors’ contractual rights would imply that the pain of insolvency needs to be shared between creditors and the debtor. The insolvency mechanism needs to balance these competing interests and guide the priority struc- ture of settling competing claims. The priority structure will depend, first, on the distributional judgment of the society concerned and, sec- ond, on the effect of a chosen priority structure on the capital market and its impact on new financing (Liu and Waibel 2009). Judicial vs. Administrative Approach The two approaches to subnational insolvency procedures discussed in this volume are the judicial and the administrative.6 Various hybrids also exist. In judicial procedures, courts make decisions to guide the restructuring process. The judicial approach has the advantage of neutralizing political pressures during the complex restructuring. However, the courts’ ability to influence fiscal adjustment of SNGs is limited because mandates for budgetary matters usually rest with the executive and legislature. In some administrative interventions, 8 Until Debt Do Us Part by contrast, a higher-level government intervenes in the entity con- cerned, temporarily taking direct political responsibility for many aspects of financial management and restructuring the subnational’s debt obligations into longer-term debt instruments. The choice of approach varies across countries. In Hungary, the desire to neutralize political pressure for bailing out insolvent subnationals favors the judicial approach. South Africa’s legal framework for munici- pal bankruptcy is a hybrid, blending administrative intervention with the role of courts in deciding debt restructuring and discharge. Colom- bia has a formal administrative process, where central government rep- resentatives facilitate restructuring negotiations between subnational borrowers and their creditors, and supervise the implementation of the agreement on fiscal adjustment and debt workouts. In Brazil, the federal government restructured the subnational debt in the late 1990s condi- tional on the SNG undertaking fiscal reform and adjustment packages. Similarly, the federal government in Mexico restructured states’ debts after the Tequila Crisis, and a few years later introduced regulations on the lenders that effectively constrained the borrowers as well. In India, the federal government used a debt swap instrument as an incentive to encourage states to enact their own fiscal responsibility laws. Reforms to Align Fiscal Incentives and Develop a Robust Framework Reforms in subnational borrowing frameworks and debt restructur- ing mechanisms have gathered momentum in developing countries since the late 1990s. The objectives of reforms are broadly similar— strengthening fiscal management and preventing future insolvency. Often, these proceed in tandem with broader public finance reforms, macroeconomic stabilization, and the development of a robust medium-term fiscal framework and transparency. The reform paths and sequences that these countries chose reflect their own historical context, legal framework, and reform dynamics. This volume surveys the reform experience of selected countries in strengthening subnational fiscal discipline and developing a framework for the resolution of subnational debt stress. The first two sections of the volume focus on two types of debt restructuring. One type is the An Overview 9 national-government-led debt restructuring, which includes the expe- riences of Brazil, India, and Mexico. The review also includes China’s experience in central-government-led restructuring of the rural edu- cation legacy debt, so that local governments could gain stronger fis- cal capacity for education service delivery. Another type focuses on the framework that spells out, in advance, the procedure in place in the event of a subnational default. It compares the experiences of Colombia, France, Hungary, and the United States. Subnational insolvency is not limited to developing countries. The reform experience of developed countries offers important lessons. The third section of the book discusses the experiences of China, the Philippines, Russia, and South Africa in developing their subnational credit markets. This topic is highly relevant to aligning fiscal incen- tives for SNGs and developing a robust regulatory framework. When the central government refrains from bailouts, creditors serve as an enforcer of fiscal discipline on SNGs by pricing risks of defaults. Reduc- ing default risks is not the same as minimizing the use of debt instru- ments. As already noted, debt instruments are essential for financing large-scale infrastructure and supporting economic growth. Competi- tive supply of subnational credits lowers borrowing cost and extends loan maturity. We also include the United States, which has the largest subna- tional capital market in the world, with outstanding SNG (states and local governments and their special purpose vehicles) debt of US$3.4 ­ trillion and an annual average issuance of US$450 billion. However, the United States was not endowed with a mature, well-functioning market from the outset. Over its long history, the U.S. subnational cap- ital markets experienced episodes of widespread defaults in the 1840s, 1870s, and 1930s. The reforms of legal frameworks and institutions have been gradual and path dependent, in the sense that later reforms built on earlier reforms. The United States experience offers lessons for developing countries, but a developing country could not simply duplicate the institutions that currently govern subnational borrow- ing in the United States. Nonetheless, the United States does offer ­ relevant lessons including the importance of tying revenue sources to borrowing, transparency in markets for government credit, and creat- ­ ing interest among creditors in strengthening borrowing rules. 10 Until Debt Do Us Part National-Government-Led Subnational Debt Restructuring Part 1 of the volume reviews subnational debt restructuring in Brazil, China, India, and Mexico. Brazil, India, and Mexico all have a federal system, but the origins of their SNG debt crises differ, as do the formula- tions of their restructuring programs. The chapter on China focuses on the restructuring of rural school debt for better alignment of the inter- governmental fiscal system and supporting inclusive economic growth. These countries offer lessons on the common-pool and moral hazard problems inherent in debt restructuring and how they have moved toward rule-based transparent frameworks. SNGs in Brazil experienced debt crises during the 1980s and 1990s, a period of macroeconomic instability, oil shocks, and a bal- ance-of-payments crisis. Circumventing the strict controls imposed on SNGs, public sector borrowing increased substantially to finance capital investments. The federal government provided bridge loans to assist SNGs in rolling over their external debt, but SNG debt stress was unabated. The federal government restructured SNG external debt in 1989 and SNG debt owed to federal entities in 1993. With a substantial part of SNG debt unresolved and persistent pressures placed on the primary balance, the fiscal and debt position of SNGs continued to deteriorate. As hyperinflation was brought under con- trol, the SNG debt obligations rose rapidly in real terms. The federal government initiated a third round of debt renegotiations in 1996. During the first two rounds of debt restructuring, state politicians suffered minimal consequences and their creditors suffered almost none. The cycle of failure in discipline and cooperation came to a halt in the third round of debt restructuring, as the deeper political and economic incentives had changed after a national macroeconomic adjustment program ended hyperinflation and stabilized the economy. Beyond macroeconomic stabilization, the federal Real Plan sought to reorganize the entire public sector and privatize banks and public enter- prises. The federal government offered SNGs incentives to restructure their debt, but also required them to undertake comprehensive struc- tural and fiscal reforms, including privatization of state-owned banks and enterprises. Three of the four largest debtor states supported the reforms and formed the core of a critical mass of states ready to An Overview 11 cooperate in fiscal restraint, making it worthwhile for additional states to join the reform. The success of the Real Plan, together with the third round of SNG debt restructuring, created the political and economic conditions for enactment of the Fiscal Responsibility Law (FRL) in 2000. The law established limits and placed restrictions on key fiscal variables and assigned responsibility for enforcing the obligations and fiscal transpar- ency requirements. Throughout the 2000s, state and municipal finances improved significantly. Total public net debt as a share of gross domes- tic product (GDP) declined from 52 percent in 2001 to 39 percent in 2010, with the decline at all three levels of government. The improve- ment in the SNG fiscal accounts is associated with Brazil’s improved macroeconomic fundamentals, but has come at a cost in reduced SNG ­infrastructure investment. Local governments in China resorted to borrowing to finance school facilities to meet the goal of universal nine-year compulsory education. In 2000, China achieved the goal, a historic accomplishment. However, the rural education debt became a significant fiscal burden on local govern- ments. With the public policy goal in the late 1990s of inclusive economic growth, the debt financing of nine-year compulsory education in the rural areas was replaced by grant financing for all children. The new policy needed to address the legacy debt and its write-offs. With a strong fiscal position, the central government could easily have written off the entire debt. This option was not chosen because it would have encouraged moral hazard. The debt restructuring program in 2007 shared the fiscal responsibility for debt write-off among three tiers of government. The central government grants used an output- based rather than an input-based formula, which took into account both the required expenditure to achieve basic provision of education and the local government fiscal capacity. This output-based approach was designed to prevent perverse incentives for local governments to increase the size of their debt or to reduce their service of debt in anticipation of more grants or bailouts. A local government that bor- rowed excessively would not gain extra advantage, and another local government that borrowed less or paid off its debt would not be in an unfavorable position. 12 Until Debt Do Us Part The Constitution of India forbids states from borrowing abroad and requires them to obtain central government permission for domestic borrowing. The central government places limits on states’ borrowing through the annual discussions with states on financing state develop- ment plans. While limiting explosive growth and systemic insolvency of state debt, the system did not prevent deterioration of fiscal con- ditions as indicated by high levels of debt over gross state domestic product in many states in the late 1990s. The outstanding state debt to GDP peaked at 32.8 percent during 2003–04, up from 20 percent during 1997–98, and interest payments as a share of revenue receipts increased from 16.9 percent to 26 percent over the same period. Fac- tors contributing to the deteriorating fiscal accounts across Indian states in the late 1990s include the rapid increase in expenditures on salaries, retirement benefits, pensions, and subsidies; increased bor- rowing to support the growing revenue deficit (current expenditure in excess of revenue including fiscal transfers); and growth in contin- gent liabilities associated with fiscal support to state-owned public enterprises. Since the early 2000s, fiscal reform has focused on moving toward a more flexible, market-linked borrowing regime within sustainable overall borrowing caps imposed by the central government and self- imposed state-level deficit caps. The federal government enacted the Fiscal Responsibility and Budget Management Act in 2003, which applies to the national government only, but some states had also adopted their own FRLs before the enactment of the federal FRL (for example, Karnataka and Punjab in 2002), and many states have since 2003 adopted FRLs in line with the national law. FRLs became man- datory after the Twelfth Finance Commission, and the federal govern- ment offered a sizable incentive to restructure high-cost debt to states for passing FRLs. During the centralized system in Mexico before the 1990s, subna- tional debt was implicitly guaranteed by the federal government. Impor- tant controls and consequences were outside the formal rules and were based on political party connections. In the 1990s, Mexico’s federal gov- ernment inadvertently involved itself in the decision making for subna- tional borrowing through pledged transfers and the implicit guarantee of bailouts that came with them. Accordingly, creditors took little time An Overview 13 to conduct thorough evaluations of subnational finances, and some local governments borrowed beyond their means. The main vulner- abilities of the subnational debt profile were the high ratio of debt over the shared revenues received by the states, and refinancing risks stem- ming from short debt maturity and floating interest rates. The 1994–95 Tequila Crisis resulted in a rapid currency depreciation, a sharp rise in interest rates, and sharp declines in the pool of shared revenues, all of which led to a state debt crisis. The development necessitated a costly federal bailout program that forced a rethinking of subnational lending parameters. The federal bailout program rescheduled subnational debt into long- term inflation-indexed debt at affordable but positive real interest rates and granted four years of assistance payments. To avoid a recurrence of the fiscal crisis, each state had to agree to a fiscal adjustment program designed by the Secretariat of Finance, which monitored compliance prior to disbursement of the annual tranches of assistance, and brought most states to a good financial situation by the end of the 1990s. The indexed debt that the banks were forced to accept helped them avert total ruin and collapse of the system, but illiquidity of the assets and low return inflicted a penalty on the borrowers as well. The federal government also ended its policy of formally guaranteeing subnational debt, although as a transition it agreed to accept and execute contractual mandates by which the borrowers pledged their revenue-sharing transfers as collateral for the debt service. During 1999–2000, the federal government effectively required credit ratings for subnational governments and brought in a new subnational borrowing framework through tightened regulations on the lending side. The federal constitution left little scope for direct regula- tion of the subnational borrowers. We have learned from the experience of Brazil, China, India, and Mexico that each debt restructuring regime needs to be based on the origin of the debt problem and the specific historical and institutional context of the debt stress. Debt restructuring plans must pay attention to their incentive effects. Rule-based debt restructuring reduces ad-hoc bargaining and adverse incentives; hard budget constraint prevents moral hazard; and burden sharing provides proper incentives and avoids free-riding behavior, while also recognizing the incentive role played by higher levels of government to leverage reform. 14 Until Debt Do Us Part Subnational Insolvency Systems Part 2 of the book discusses the development of subnational insol- vency systems in Colombia, France, Hungary, and the United States. All four countries developed a framework for insolvency proceedings in response to subnational debt crises. But the frameworks differ across the countries, reflecting historical contexts, constitutional frameworks, entry points for reform, and institutional developments that are path dependent. While Hungary and the United States opted for court pro- ceedings for insolvency, Colombia and France chose to use adminis- trative proceedings. All four countries confronted key design issues, whether they were federal (the United States) or unitary (Colombia, France, and Hungary). Part 2 concludes with the 1983 default case of the Washington Public Power Supply System in the United States, the largest subnational bond default in modern U.S. history. This case illus- trates the dynamic interactions among stakeholders that have ramifica- tions for regulatory reform and market development. In the late 1980s and 1990s, the trend toward political decentraliza- tion in Colombia was accompanied by more freedom for subnational borrowing. SNGs experienced debt stress in the late 1990s to early 2000s, exacerbated by the economic downturn. Contributing factors included weak bank lending supervision, excessive reliance on transfers, and per- mission to borrow for current expenditure, which blunted incentives for fiscal discipline. Although the SNG debt level was not high by inter- national comparison, the arrears had been increasing by the late 1990s, and SNG capacity for debt service had weakened, due primarily to the decline in SNG own revenues and fiscal transfers. The SNG debt stress led to substantial public finance reform. ­ Colombia enacted several laws, mostly between 1998 and 2003, that regulate the origination of SNG debt and encourage fiscal responsibility. Law 550 (1999) deals explicitly with bankruptcy proceedings for SNGs. The essence of the proceedings is to evaluate and reconcile compet- ing claims against subnational debtors, according to a defined priority structure. The procedures in Colombia are administered by the Super- intendency of Corporations (SOC) in coordination with other central government institutions. Created in the 1930s, the SOC’s unique role arose within a historical context in which the court system was weak. An Overview 15 The SOC administers bankruptcy procedures for both corporations and most government entities. The implementation of Law 550 and other fiscal legislation has taken place in the context of improving macroeconomic performance of the country since 2003. There has been little divergence between the law and its practice. The protection offered by bankruptcy Law 550 enables insolvent SNGs to reach orderly debt restructuring agreements with creditors. Focusing on debt workouts, Law 550 has limited ability to address the root causes of fiscal stress and debt. Other complementary laws—mainly Laws 358, 617, and 819—work in several ways by limiting borrowing, promoting fiscal transparency, strengthening the budgetary process, and helping to finance debt restructuring. During 1982–83 and 2003–04, two waves of decentralization in France devolved more powers to the three levels of SNGs: the munic- ipalities, the departments, and the regions. This new institutional framework has enabled SNGs to enjoy a greater degree of autonomous expenditures, to raise their own taxes, and to borrow from financial markets, within ex-ante rules established by the central government. However, SNGs are subject to ex-post controls by the Prefect and the Regional Chambers of Accounts, and to ongoing controls by the Public Accountants. The ex-ante fiscal rules and the regulatory framework for managing SNG fiscal risks were established after a period of unregulated borrow- ing by SNGs following the decentralization and subsequent debt stress experienced by some SNGs in the early 1990s. The regulatory frame- work combines the laws and regulations with three sets of institutions, while preserving considerable SNG fiscal autonomy. The laws and pru- dential rules regulate debt, liquidity, and contingent liabilities. The state exercises strong supervision and monitoring of SNG financial accounts through the Prefect, the Regional Chamber of Accounts, and Public Accountants. By law, SNGs cannot go bankrupt and public assets can- not be pledged as collateral. If an SNG is insolvent, the central govern- ment will intervene, enforcing fiscal adjustment and facilitating debt negotiations among the creditors and the borrower. SNG accounts may be placed under the control of the Prefect and the Regional Chamber of Accounts for several reasons, including failure to present a balanced budget, deficits exceeding 5 percent of operating revenues, and failure to 16 Until Debt Do Us Part make provisions in the budget for compulsory expenditures including debt services. State supervision has helped to substantially reduce SNG insol- vency risks, although several debt restructurings occurred in the last two decades. Nonetheless, the lack of a clear, established legal structure for priority payments creates uncertainties. Off-budget entities, such as public-private partnerships, pose contingent fiscal risks, a common challenge across countries. The 1990 Law on Local Government in Hungary granted local gov- ernments unfettered freedom to manage their finances. They borrowed for commercial activities and long term to finance short-term operating deficits. The macroeconomic deterioration in the mid-1990s exposed the seriousness of subnational financial distress. Imprudent lending by public banks without proper evaluation of SNG creditworthiness was attributed to the assumed central government guarantee for subna- tional debt. Several local governments successfully lobbied for central government grants. This threatened to set a bailout precedent, raising concerns of adverse incentives for debtors and creditors. Several options were debated at the time, including informal restructuring negotiations between creditors and local governments. The deteriorating financial performances of local governments caused a concern about creating contingent liabilities for the central govern- ment. The government eventually opted for a formal insolvency mech- anism. Transparency and predictability were viewed as central to an effective subnational insolvency mechanism. The Law on Municipal Debt Adjustment, approved by the Hungarian Parliament in 1996, gives courts the central role in fiscal and debt adjustment for insolvent local governments. The implementation experience has exceeded the expectations of the framers of the law. The legal procedure is transparent, moral haz- ard of bailouts has been minimized, and essential services have been maintained. The debt adjustment procedures have given participating municipalities a clean slate to move forward. However, many insolvency cases were resolved through informal negotiations. While bilateral nego- tiations are an integral part of all insolvency regimes, the nontranspar- ency and potential asset stripping could negatively affect less-informed or smaller creditors and the public interest. Discussions are ongoing An Overview 17 among stakeholders on making the pre-bankruptcy negotiated restruc- turing more transparent. In the United States, after the initial refinancing of national and state debts incurred during the Revolutionary War, when the national gov- ernment assumed the existing state debts, national government involve- ment in state and local government finances was minimal until the Great Depression and New Deal programs of the 1930s. It was not until 1933 that the national government began significant grant and transfer pro- grams to the states (Wallis 1984). Since the 1930s, national, state, and local finances have been more closely intertwined, but the national gov- ernment has generally maintained a no-bailout policy and has left the structure and regulation of subnational borrowing to state governments. As chapter 14 on U.S. state systems of local government borrowing shows, there are 50 different subnational finance systems in the United States. In response to widespread municipal defaults during the Great Depression, the U.S. Congress in 1937 adopted a municipal insolvency law known as Chapter 9 of the U.S. Bankruptcy Code. Chapter 9 is a debt restructuring mechanism for political subdivisions and agencies of U.S. states. The widespread subnational financial distress during the Great Depression revealed the practical drawbacks of the mandamus (a court order obliging municipalities to service debt obligations) and informal protracted negotiations between municipal debtors and credi- tors (McConnell and Picker 1993). Chapter 9 delineates the procedures whereby a debt restructuring plan acceptable to a majority of creditors can become binding on a dissenting minority. The design of Chapter 9 was guided by the U.S. constitutional provi- sions that reserve the control over state and local government finances completely to the states. State consent is a precondition for municipali- ties to file for Chapter 9 in federal bankruptcy court. Chapter 9 is not the primary subnational insolvency mechanism in the United States. Only about half of states authorize their political subdivisions to file for Chapter 9 relief. The unique federal structure of the United States also profoundly influences the specific design of Chapter 9, where the federal courts have limited ability to impose conditions on the debt adjustment plan of an insolvent municipality. Most of the institutions that govern subnational government borrowing in the United States are embodied in state constitutions, state laws, and state administrative agencies. 18 Until Debt Do Us Part The U.S. corporate insolvency laws have had influence on other countries. While the U.S. municipal insolvency framework offers a valu- able reference for other countries, the framework itself cannot be cop- ied without care. Chapter 9 was conceived with the narrow objective of resolving the holdout problem. It is based on a respected, independent, and competent judiciary that has the authority to reject a municipality’s Plan of Adjustment. In many developing countries, intergovernmen- tal systems are still evolving, lending to SNGs may still be dominated by a few public institutions, and judicial systems may lack capacity. The development of a subnational insolvency mechanism must be sequenced with other reforms. To develop a legal framework to resolve financial distress, many countries face similar objectives and challenges, namely, the interest in the functioning of local government autonomy, safeguarding essential public services and the assets that provide such services, transparent procedures, the interests of creditors, and functioning subnational capi- tal markets. Part 2 of the volume also includes the largest municipal bond default in the United States since the 1950s. In 1983, the Washington Public Power Supply System (WPPSS) defaulted on US$2.25 billion in out- standing bonds. The debt issued by the WPPSS was ruled by the State Supreme Court as invalid and unenforceable. Therefore, filing for Chapter 9 was never an option. The default, a rare event in scale and frequency in the modern U.S. subnational debt market, offers a win- dow into the interactive roles of market, the courts, the regulators, the debtor, the creditors, the federal and state governments, and taxpayers. The WPPSS default shows that, even in a developed country, the issuance of debt for infrastructure has endemic risks such as the lack of transparency and disclosure, poor project management, and construc- tion delays. But even if the debt issued is valid and legally enforceable, the mounting problems in construction delays, cost escalations, and dif- ficulty in refinancing existing debt would have made it difficult for the WPPSS to pay back bondholders. None of the bailout proposals was seriously considered by Congress. The government took minimal enforcement action against actors in the WPPSS drama, because the principle of self-regulation outweighed the cost of enforcing regulations. Although there was little government An Overview 19 action, the amount of private damage litigation was unprecedented and resulted in many failed careers and business collapses. Very few indi- vidual market participants gained from the WPPSS disaster, but the market not only weathered the storm—it became stronger. The U.S. municipal market showed little evidence of damage resulting from the WPPSS default. Not only did the market quickly return to normal after the WPPSS default, but the period during which the WPPSS drama unfolded, from 1975 to 1985, was one in which total annual municipal bond issuance grew tenfold—a dramatic decade of growth in the his- tory of the modern market. Subnational Credit Market Development Part 3 of the book focuses on subnational credit market development in China, the Philippines, Russia, South Africa, and the United States. Developing countries face long-term challenges in developing liquid, deep, and competitive subnational credit markets. In general, bank loans continue to dominate the supply of credit to SNGs in developing countries, and public financial institutions continue to dominate credit supply in some countries. Subnational securities markets in developing countries in general are small in scale and lack liquidity and second- ary markets. The United States has the largest, most liquid, and most competitive subnational capital market, but the market development has interacted with a series of institutional reforms through its history. Although the lessons from the U.S. experience are highly specific to the history of the American states, there are some general lessons for devel- oping countries. China has been investing about 10 percent of its GDP annually in infrastructure, with SNGs taking on a large share of investments and rapidly transforming the urban infrastructure landscape. SNGs relied on central government onlending and their own off-budget vehicles— Urban Development and Investment Corporation (UDIC), borrow- ing directly from the financial markets mainly through loans but also bonds—and land assets-based finance to develop urban infrastructure. The limitations of these financing instruments became evident to policy makers in the mid-2000s. With central government onlending, SNGs have no market interaction with creditors, and the borrowing power 20 Until Debt Do Us Part and payment obligations are not linked. UDIC’s off-budget debt is non- transparent. Financing infrastructure through land lease is not sustain- able in the long run, because of the up-front collection of leasing fees. China has undertaken reforms since 2009 to allow the issuance of provincial bonds and later the piloting of municipal bonds. Policy makers recognized that important preconditions for the issuance of bonds by provinces did not exist in 2009. It takes time to develop credit rating systems, and SNGs had no market access experience. The reform thus took a learning-by-doing approach. The central government acted as the issuing agency, with SNGs participating in the auctions. From 2009 to 2011, RMB 600 billion (US$90 Billion) of provincial bonds was authorized and issued. In 2011, reform took a further step: the State Council approved piloting of direct bond issuance by four cities (RMB 23 billion) without the central government acting as the issuing agency. The reform helped SNGs finance the subnational matching part of investment projects in which the central government co-invested in response to the 2008–09 global financial crisis. The new debt instru- ment significantly lowered the financing costs for SNGs, enabled them to start acquiring market access skills, and linked the SNGs as debtors with their debt service obligations. Piloting municipal bonds without the central government as the issuer is one step further for SNGs to access the market. The issuance of SNG bonds has been supported by developing legal, institutional, and market infrastructure. The reforms in fiscal manage- ment (including the single Treasury account and expenditure reforms) and separating management from ownership of public enterprises have laid the groundwork for the piloting of provincial bonds. The new bond instrument to finance capital outlays under newly developed budgeting procedures will facilitate the development of a framework for medium- term capital budgeting for infrastructure investments. The audit of, and the ongoing efforts to better classify, UDIC debt will facilitate the devel- opment of different bond instruments with different risks and securi- tization profiles. Further regulatory reforms can support sustainable market access, as would complementary reforms in strengthening inter- governmental fiscal systems, enhancing fiscal transparency, and deepen- ing financial markets. An Overview 21 The Philippines is an emerging economy that continues to chart its own course in developing its subnational debt markets. The Philippines has been innovative in its efforts to extend the legal possibilities for local governments to take initiative in the use of credit and in the design of credit market techniques to make that possible. The Local Government Code, with its broad array of borrowing powers granted to local govern- ment units, and the creation of the Local Government Unit Guarantee Corporation to bolster local credits, are pioneering efforts. The subnational debt market is small and levels of indebtedness are low. The risk of default is minimized by an intercept mechanism used to secure such debt. As a result, the lack of a formal insolvency system is not a key challenge for developing a competitive subnational credit market. There are more fundamental structural challenges, including institutional and political economy factors, which deter subnational governments from accessing private sector credit markets, and there is a lack of competition for subnational debt instruments. The development of competitive subnational credit markets needs to address both demand- and supply-side constraints. On the demand side, it is critical to strengthen the local finance and accountability sys- tems for citizens to demand better services. On the supply side, remov- ing constraints to private bank participation in subnational credit markets will increase competition and help lower the cost of financing. Some financing instruments may help forge closer links between local governments’ own revenues and their capacity to access the market, which in turn strengthens local accountability. The recent experiments encouraging greater partnerships between the local governments and the private sector credit markets could pave the way for a more com- petitive and diversified subnational credit market. The subnational debt market in the Russian Federation began to develop in the early 1990s. Unfunded federal mandates and political decentralization contributed to the growing demand for debt instruments including foreign currency debt. At the same time, there was a complete lack of debt regulation, and SNGs lacked experience in managing debt risk. Debt was issued to finance recurrent expenditures, mostly with short-term maturities. With a rapidly deteriorating macroeconomic envi- ronment in Russia in the late 1990s, refinancing risks facing SNGs rapidly rose. Fifty-seven of 89 regions defaulted on their debt from 1998 to 2000. 22 Until Debt Do Us Part Improved macroeconomic fundamentals during 2000–08 and sub- stantial legislative reforms—significant amendments to the Tax Code, the adoption of the Budget Code, and the 2006 legislation on local self- government—contributed to positive changes in intergovernmental relations and incentives to formulate new principles of financial man- agement for the regions and municipalities. The Budget Code contains provisions for regulating the subnational debt, including the provisional limits on deficit, debt and debt service, regulations of external bor- rowing, guarantees, and structure and types of debt instruments. With revenue growth, the financial positions of the regions and municipali- ties strengthened considerably. The debt load of the Russian regions remained low at the end of 2007. The 2008–09 global financial crisis struck Russian public finances in 2009, though the impact varied across SNGs. There were, however, no regional defaults owing to support from the federal government and the liquidity accumulated by the regions in prior years. Since 2011, subna- tional fiscal positions have improved along with a gradual recovery of oil prices and the Russian economy. The debt markets have recovered and borrowing costs have declined. However, activity in the domes- tic bond market remained moderate until 2011, when the market expanded. There are continuing challenges in subnational debt market development. For example, most SNGs have short-term debt profiles dominated by one-year bank loans, implying higher refinancing risk; bank financing of the regions is dominated by a few state-controlled banks; and there is a lack of comprehensive accounting for the contin- gent liabilities of government enterprises. In the post-apartheid era, South African municipalities faced the challenges of large-scale infrastructure investments to make up for huge backlogs left by the apartheid regime, rapid urbanization, and the need to accelerate economic development. The government’s 1998 White Paper on Local Government stressed the importance of lever- aging private sector finance to meet the infrastructure requirements of municipalities. This was followed by extensive stakeholder consul- tation between 1998 and 2003, leading to enactment of the landmark Municipal Finance and Management Act (MFMA). As part of the finan- cial management, the act provides a comprehensive set of ex-ante rules regulating municipal borrowing. The act also spells out a procedural An Overview 23 approach for dealing with municipalities in financial distress, which is important for lenders. South Africa engaged in lengthy political consultations to develop insolvency procedures. Two constitutional amendments paved the way for a municipal insolvency mechanism. The South African case demon- strates the complexity of subnational borrowing and insolvency legisla- tion and the path dependency of reforms. It illustrates the importance of building political consensus among various interest groups. Broad support may require concerted effort over a number of years. South Africa took two years to develop the basic policy framework (1998– 2000), another year for cabinet approval (2001), followed by two years of parliamentary debate on the constitutional amendments and on the Municipal Finance Management Act (2001–03). All metropolitan municipalities have, in the last decade, borrowed funds from the banking sector, capital markets, or both to finance infrastructure development. Since 2005, activity in municipal credit markets has risen quickly. Long-term borrowing increased rapidly in the run-up to the 2010 FIFA World Cup, changing the landscape of municipal finance from a high level of dependency on fiscal transfers to one where borrowing plays an increasingly important role in financ- ing capital expenditure. Private lenders credit the MFMA as the most important factor in promoting market activities. There are continuing challenges, however, including the lack of a fully developed secondary market, the incompatibility of short-to-medium-term debt maturities with long-term assets of infrastructure, and the need to crowd-in more private financing in the market. The United States has by far the largest local government capital market in the world, with the longest history of market development, achieved through a series of incremental changes in institutions over a long period of time. All the governments below the state level—what Americans call “local government”—are not sovereign, but rather are created by and subject to the laws of each respective state. Local govern- ments borrow significant amounts of money to finance infrastructure investment and have very low rates of default. Local governments in most states face restrictions on how they borrow and what they can borrow for, and, in some states, how much they can borrow. For the most part, these restrictions are on the procedures that local governments must 24 Until Debt Do Us Part ­ ollow to approve borrowing and how debt service obligations are related f to specific revenue sources (particularly in the case of revenue bonds). A central feature of the American experience is the importance of ex-ante and passive insolvency systems. Twenty-three of 50 states pro- hibit their municipalities from filing Chapter 9 in federal courts. Only one-third of the states have a system in place for monitoring local gov- ernments, and less than 20 percent have institutions and policies that enable or require state action in the face of a local government fiscal crisis. The lack of active state programs does not mean that local gov- ernment borrowing and debt servicing are not actively monitored by the larger society. Instead, it highlights how the interaction of ex-ante institutional rules, voters, capital markets, and courts play key roles in monitoring and limiting local government borrowing. A distinctive feature of the American state systems is that they estab- lish a close relationship between borrowing and taxation. Most of the states’ constitutional reforms in the North that followed the states’ debt crisis in the 1840s required states (and local governments after the 1870s) to raise current taxes when they issued debt (Wallis 2005). Forcing voters and taxpayers to simultaneously raise taxes when they borrowed money increased the burden on borrowing, and led voters to pay closer attention to the benefits of the expenditures and debt that local governments pro- posed. A similar set of incentives was set in motion with the development of special districts and revenue bonds at the end of the 19th century. The project that the bond proceeds will finance is securitized by the revenue streams of the project, and the beneficiaries (including future genera- tions) of the project will pay user fees to finance the debt service. The U.S. experience shows the importance of creating clear inter- est among creditors in strengthening both the rule of law and incen- tives for private market development. This is in marked contrast to a system where lenders assume that the central government would be, and often was, ultimately responsible for repaying debts. Passive insolvency systems also clearly require the existence of a strong and credible rule of law in order to work. Establishing the legal precedent that local taxpayers were not responsible for servicing debts that were incurred in an unauthorized manner or through defective procedures was a long, drawn-out process undertaken at the end of the 19th century. An Overview 25 The result of the framework for debt issuance has not been that local governments borrow wildly in unauthorized ways and then default, but rather a steady increase in the capacity of private capital markets to assess the creditworthiness of local governments and inform potential borrowers of the actual conditions under which local debt is issued and will be repaid. The institutional developments such as ex-ante debt rules (1840s), the bond counsel (the late 19th century), and the Municipal Securities Rulemaking Board (the mid-1970s) all make the provision of information to private market participants more credible and transpar- ent. The national government has not violated the sovereign powers of states to tax, spend, and borrow as they wish, nor have they impaired the ability of states to establish systems for their local governments. In principle, states possess the authority to unilaterally change the structure of any local government. In practice, however, states moved toward “general” laws governing local governments. This was an institu- tional change that arose endogenously in the American setting. If individ- ual local governments could approach the state for special treatment, or if the state can single out individual local governments for special treatment (either positive or negative), then the incentives to create and enforce credible rules would be eroded. If all cities know that the same rules apply to all of them equally, then all cities collectively have a strong incentive to make sure that a state enforces the rules equally across all cities. Lessons Learned Structural trends of decentralization and urbanization are likely to con- tinue in developing countries, requiring massive infrastructure invest- ments at the subnational level. A range of middle-income countries, and low-income countries in transition to more open market access are con- templating expanding subnational borrowing and debt financing for infrastructure investments. The experiences of countries covered in this volume offer valuable lessons. As shown by Canuto and Liu (2010a), subnational credit risks are intertwined with broader macroeconomic and institutional reforms. Macroeconomic stability and sovereign strength set an effective cap on the credit ratings of SNGs and influence the availability and cost of funds. Debt sustainability of SNGs is determined by the interplay of the 26 Until Debt Do Us Part existing debt stock, economic growth, cost of borrowing, and primary balance. Macroeconomic framework and policies strongly influence the interplay of all these factors. The history of subnational debt crises shows that unregulated borrowing, particularly in an unstable macro- economic environment, is extremely risky; unfettered market access by subnational borrowers can outpace the development of sound revenue systems and adequate securitization. Deficits and debt arise from the joint decision of governments mak- ing fiscal policy and their creditors. These decisions are made in light of not only the rules governing issuance of the debt, but also the expecta- tions about what will happen to the debtor and the creditors if payment difficulties arise—who will lose money or who will be forced into painful adjustment. The decisions of that lending moment become a fait accom- pli conditioning the subsequent decisions. This points to two impor- tant dimensions of control of government borrowing. First, the type or timing—ex-ante controls or ex-post consequences; and second, whether the ex-ante controls and ex-post consequences act on borrowers or lenders. Ex-ante constraints on subnational borrowers include procedural rules for incurring debt, limits on debt and deficit ceilings, rules for borrowing in international markets, and regulation of subnationals’ borrowing based on fiscal capacity criteria. To complement the ex- ante constraints and to make them credible, there need to be ex-post consequences for failures in fiscal prudence. Without lenders there is ­ no ­ borrowing or debt, so their constraints and incentives deserve equal attention. Relying on constraints only on borrowers means that lenders still have incentives to push loans and may find reckless officials willing to borrow despite the rules. Relying only on ex-ante constraints, without ex-post consequences, gives irresponsible borrowers and lenders an incentive to get around the ex-ante rules and execute transactions that will later get bailed out. Relying only on ex-post consequences allows irresponsible (and large) entities to build up such large debts that they could threaten macro- economic stability. Debt restructuring needs to pay close attention to its incentive effects. Rule-based debt restructuring reduces ad-hoc bargaining and adverse incentives; hard budget constraint prevents moral hazard; and burden sharing provides proper incentives and avoids free-riding An Overview 27 behavior, while also recognizing that higher levels of government can create incentives for reform. The purpose of borrowing and insolvency controls is not to mini- mize the use of debt financing, but rather to promote sustainable debt financing through a competitive and diversified subnational credit sys- tem. Such a system can help ensure the lowest cost of capital and a sus- tainable supply of credit. Debt financing is valuable for infrastructure development where the maturity of assets is generally longer than the current terms of taxation and transfers. The dynamics of subnational-central government interaction pro- vide reform momentum. On the one hand, one or a few subnational governments can serve as catalysts for fiscal reform and as a demonstra- tion for national reform. On the other hand, the national government can offer fiscal incentives to encourage subnational fiscal adjustment. One common trait of successful debt restructuring for SNGs is the com- mitment of the central government to its own fiscal prudence. The design for regulating debt and insolvency needs to be consistent with the broader cultural, economic, legal, constitutional, and social context of the country. Subnational fiscal adjustment and debt restruc- turing operate within a country’s specific intergovernmental system that defines the respective authority of each level of government, and within a country’s political system that defines the respective authority of each branch and level of government. Capacity and entry point for reform matter. The maturity of the legal system and the capacity of the judiciary influence the choices of procedure. Regulations on debt and insolvency cannot compensate for inade- quacies in the design of overall intergovernmental fiscal relations. The intergovernmental fiscal system underpins the fundamentals of the ­ subnational fiscal structure. Without increased fiscal autonomy and greater own-source revenues, subnationals will rarely be in a position to borrow sustainably on their own. In addition, an intergovernmen- tal fiscal transfer system that routinely fills deficit gaps will undermine the incentives for a balanced budget. The regulations on debt and insolvency cannot substitute for other reforms such as budgetary and financial management, taxation reform, and governance reforms. The incentive signals of insolvency mechanisms require a more competitive subnational capital market. 28 Until Debt Do Us Part It is critical to understand the interaction of rules, enforcement, and capital markets. In the case of government borrowing, decisions to spend in the present must be matched with decisions to tax and service debt in the future. Well-functioning capital markets are a way for societies to pool the best information about conditions today and changes tomorrow. When governments possess the discretionary ability to change the rules between today and tomorrow, it becomes difficult for the capital markets to assess either the returns from financing infra- structure spending or the risks that debts will not be repaid. The importance of closely tying borrowing decisions to revenue decisions as a feature of good institutional design cannot be overstated. Debts have to be repaid, and debt issuance that is tied to tax increases or dedicated revenue sources is much more likely to be repaid. The experi- ences discussed in this volume show the importance of moving to rules- based systems in which the higher-level government treats all lower- level governments according to the same rules. No matter what the rules, their ad-hoc or discretionary application is likely to be plagued with moral hazard and common-pool problems. Notes  The findings, interpretations, and conclusions expressed in this work are those of the authors and do not necessarily reflect the views of The World Bank, its Board of Executive Directors, the governments they represent, or any other institutions with which the external authors may be affiliated. 1. The term subnational in this book refers to all tiers of government below the federal, or central, government. The category also includes special purpose vehicles or investment companies created by SNGs. At the national level, estimations of future infrastructure investment require- 2.  ments vary greatly by income level. Estache and Fay (2010) discuss methodolo- gies for quantifying these requirements and estimate that low-income countries should spend 12.5 percent of GDP on investment and maintenance to meet demand, whereas lower-middle-income and upper-middle-income countries should spend 8.2 and 2.3 percent, respectively. 3. Inman (2003) develops the prisoners’ dilemma model formally for this situation and shows how restrictive are the conditions under which the ­ ­ market successfully establishes subnational fiscal discipline if the central government ­ takes a hands-off, no-bailout approach. The conditions include c ­ompetitive An Overview 29 ­uppliers of local public services, a stable central government, clear and s enforceable accounting standards, a well-managed aggregate economy, and an informed and sophisticated local government bond market. 4. This might include, for example, an unused vacant lot outside the corporate limits or a private residence taken for failure to pay taxes (McConnell and Picker 1993, 432). To “cramdown” is the ability to force dissenting minority creditors to accept an 5.  agreement between a majority of creditors and the debtor. In some places, there is no system, so “ad hoc” is a third system. In other places, 6.  defaults are dealt with as political problems, and there is no (or little) judicial or administrative capacity to deal with the problem. Bibliography Braun, Miguel, and Mariano Tommasi. 2004. “Fiscal Rules for Subnational Govern- ments: Some Organizing Principles and Latin American Experiences.” In Rules and Practice in Intergovernmental Fiscal Relations, ed. G. Kopits. Washington, DC: International Monetary Fund and World Bank. Canuto, Otaviano, and Lili Liu. 2010a. “Subnational Debt Finance: Make It Sustain- able.” In The Day after Tomorrow: A Handbook on the Future of Economic Policy in the Developing World, ed. Otaviano Canuto and Marcelo Giugale, 219–238. Washington, DC: World Bank. ———. 2010b. “Subnational Debt Finance and the Global Financial Crisis.” Premise Note, Poverty Reduction and Economic Management Network, World Bank, Washington, DC. Estache, A., and M. Fay. 2010. “Current Debates in Infrastructure Policy.” Commission on Growth and Development Working Paper 49, World Bank, Washington, DC. Ianchovichina, Elena, Lili Liu, and Mohan Nagarajan. 2007. “Subnational Fiscal Sustainability Analysis: What Can We Learn from Tamil Nadu?” Economic and Political Weekly 42 (52): 111–19. Inman, Robert P. 2003. “Transfers and Bailouts: Lessons from U.S. Federalism.” In Fiscal Decentralization and the Challenge of Hard-Budget Constraints, ed. J. Rodden, G. Eskeland, and J. Litvack, 35–83. Cambridge, MA: MIT Press. Liu, Lili, and Michael Waibel. 2009. “Subnational Insolvency and Governance: Cross-Country Experiences and Lessons.” In Does Decentralization Enhance Service Delivery and Poverty Reduction?, ed. Ehtisham Ahmad and Giorgio Brosio, 333–75. Cheltenham, U.K.: Edward Elgar. Liu, Lili, and Steven Webb. 2011. “Laws for Fiscal Responsibility for Subnational Discipline: International Experience.” Policy Research Working Paper 5587, World Bank, Washington, DC. Maco, Paul S. 2001. “Building a Strong Subnational Debt Market—A Regulator’s Perspective.” Richmond Journal of Global Law and Business 2 (1): 1–31. 30 Until Debt Do Us Part McConnell, Michael, and Randal Picker. 1993. “When Cities Go Broke: A Concep- tual Introduction to Municipal Bankruptcy.” University of Chicago Law Review 60: 425–35. Wallis, John J. 1984. “The Birth of the Old Federalism: Financing the New Deal.” Journal of Economic History 44 (March): 139–59. ———. 2005. “Constitutions, Corporations, and Corruption: American States and Constitutional Change, 1842 to 1852.” Journal of Economic History 65 (1): 211–56. Part 1 Subnational Debt Restructuring 1 Brazil: The Subnational Debt Restructuring of the 1990s— Origins, Conditions, and Results Alvaro Manoel, Sol Garson, and Monica Mora Introduction During the 1980s and 1990s, Brazilian subnational governments (SNGs) experienced extreme fiscal difficulties that resulted in increasing debt, giving rise to major renegotiations with the federal government near the end of the 1980s, in 1993, and during 1997–2000. In each period, the federal government undertook massive debt restructuring with states and municipalities. Each crisis had different underlying macro- economic conditions and political economic dynamics, and the subse- quent restructuring thus produced different results. The conditions and results are related to the evolving system of fiscal federalism in Brazil. The lessons drawn from these crises and the resulting debt restructur- ing are relevant not only for Brazil, as it continues to recover from the 2008–09 global financial crisis, but also for developing countries where fiscal decentralization is under way and debt continues to be an impor- tant instrument for financing economic growth. The importance of Brazilian SNGs—that is, states and municipalities1— in the provision of goods and services gained ground in the 1980s and 1990s, ­ as decentralization and urbanization deepened. In 1970, 56 percent of the total population of 93 million lived in urban areas. In 2000, 81.2 p ­ ercent 33 34 Until Debt Do Us Part of the almost 170 million people did so, and the rate of urbanization reached 90.5 percent in the Southeast region, the most developed in the country. Throughout this period, SNGs accounted for nearly half of pub- lic sector spending and for most spending in education, health care, infra- structure, and public security. The debt crises in the 1980s and 1990s evolved in the context of development and macroeconomic policies of the military govern- ­ ment (1964–85). Public sector borrowing increased substantially to finance capital investments associated with rapid urbanization. These investments were undertaken mainly by newly created institutions such as public agencies and enterprises at both the federal and subna- tional ­levels. Federalism in Brazil revived in the 1980s with the return to democracy. The 1986 Congress, with strong representation of SNGs, crafted provisions for the 1988 Constitution that gave states significant authority and resources, including a much broader revenue base for the state-level value-added tax (VAT), but did not specify their spending responsibilities or set rules for fiscal prudence. SNGs experienced debt stress during the 1980s, a period of macro- economic instability that included oil shocks and a balance-of-payments crisis. Although the federal government provided bridge loans to assist SNGs in rolling over their external debt, the debt stress of SNGs con- tinued unabated. The three subsequent debt restructurings (1989, 1993, and 1997–2000) dealt with different types of debt—external debt, debt owed to federal entities, and all other debt including market ­ securities. The macroeconomic stabilization program in the early-to-mid-1990s, which was centered around the Real Plan, shaped the direction of the third debt restructuring. The Real Plan sought to reorganize the entire public sector, including the privatization of banks and public enter- prises, and to reduce hyperinflation. The most notable feature of the third round of debt restructuring was that it dealt with the underlying reasons for the subnational fiscal imbal- ance by focusing on the quality of fiscal adjustment and reforms. The federal government offered SNGs incentives to restructure their debt, but also required them to undertake comprehensive structural and fis- cal reforms, including control over personnel spending and privatiza- tion of state banks and state-owned public enterprises. The success of the Real Plan, together with the third round of SNG debt restructuring, Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 35 created the political and economic conditions for enactment of the Fiscal Responsibility Law (FRL) in 2000. The FRL established limits and ­ placed restrictions on key fiscal variables and assigned responsibility for enforcing the obligations and fiscal transparency requirements. Throughout the first decade of the 2000s, state and municipal ­ finances improved significantly. SNGs began to generate a primary sur- plus of about 1 percent of gross domestic product (GDP), reversing the deterioration of the previous two decades. The limits imposed by the FRL on public spending, debt, and debt services were crucial to achiev- ing these improved fiscal results and are central to the subnational debt restructuring agreements and the rules of budgetary and financial exe- cution. Total public net debt as a share of GDP declined from 52 percent in 2001 to 39 percent in 2010, with the disaggregation declining in all three levels of government. The improvement in the SNG fiscal accounts is associated with Brazil’s improved macroeconomic fundamentals. This chapter reviews the Brazilian SNG debt restructuring imple- mented at the end of the 1980s and during the 1990s, and assesses its evolution within the macroeconomic context of crises, stabilization, and reforms. Section two provides a historical context for the origins of the SNG debt crisis. Section three defines the macroeconomic framework that gave rise to the high indebtedness of SNGs and details the succes- sive rounds of debt renegotiation. Section four analyzes the evolution of SNG fiscal performance in the post-renegotiation era. Section five gives special attention to the 2007–10 period, and assesses the impacts of the 2008–09 global economic crisis on subnational debt and finance and the main fiscal federalism challenges emerging in the recent period. Section six draws some conclusions. Fiscal Federalism, Centralization, and Decentralization: Historical Context Fiscal federalism in Brazil has developed in alternating waves of cen- tralization and decentralization of power. In 1891, the first constitution of the republic adopted a federal structure that decentralized revenues and gave states control of export taxes (Varsano 1996). It also gave governors, supported by regional oligarchies, control of the electoral process. The period 1930–45 was marked by a nondemocratic regime 36 Until Debt Do Us Part that favored fiscal centralization. The Constitution of 1946 marked the return to democracy and a more decentralized fiscal structure. This was followed by a military dictatorship from 1964 to 1985—a period of cen- tralization and authoritarianism—and redemocratization after 1985, with SNGs increasing their share of revenues, but also accumulating mounting debt.2 Economic and Political Crises in the 1960s and 1970s and Authoritarian Rule In 1964, a military government took power in the midst of a deep eco- nomic and political crisis and adopted a strategy based on two essential elements. The first comprised measures to stabilize the economy and promote fiscal and financial reforms. These reforms created a new tax system, which raised the overall tax burden and improved tax collection, and adopted modern financial instruments, creating (a) special funds made up of earmarked revenues, (b) the Central Bank of Brazil and the National Monetary Council, and (c) a mechanism for indexing financial assets to inflation. The second strategic element reshaped the public administration by granting autonomy to the so-called indirect administration—agencies, foundations, public enterprises, and mixed public-private companies. A large part of the production of goods and services was transferred to them. Centralized planning and allocation of resources covered key sectors relating to national integration (for example, transportation and telecommunications) and basic materials (for example, petrochemicals, paper, and cellulose). After the initial period of economic stabilization in the mid-1960s, Brazil entered a phase of accelerated growth of around 11 percent a year between 1968 and 1973, lower inflation, higher exports, and less volatile exchange and interest rates. The accelerated growth, however, highlighted the strong dependence of the country on capital goods and imported inputs, particularly petroleum and its derivatives. The growth of foreign debt and direct foreign investments increased debt servicing and the repatriation of profits abroad. To finance large-scale projects, the federal government also established special funds such as the Fed- eral Fund for Urban Development and special public agencies such as Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 37 the National Housing Bank. The use of foreign capital to finance not only investment but also working capital initiated a cycle of increasing indebtedness, which eventually culminated in a national debt crisis in 1982. A political crisis arose in 1974, when the government-backed National Renewal Alliance Party suffered significant defeats in state and national elections. On the economic side, oil prices skyrocketed: the first oil shock increased oil prices by more than 300 percent from 1973 to 1974. The pressure on the balance of payments was magnified by a drop in exports that resulted from a sharp fall in global economic growth and rising interest rates. In mid-1979, the second oil shock led to ris- ing interest rates, followed by higher costs of borrowing in the London market. Brazil, which had obtained foreign resources at floating interest rates, experienced a sharp drop in the supply of foreign credit and rising costs to roll over debt. A huge balance-of-payments crisis was followed by tight domestic fiscal and monetary policies with significant impacts: (a) the federal tax burden increased from 18.4 percent of GDP in 1980 to 20.6 percent in 1983; (b) strict control of public spending negatively affected public investments; (c) GDP fell 6 percent during 1981–84; and (d) inflation accelerated from about 100 percent in 1982 to more than 200 percent annually during 1983–85, which resulted in an erosion of public rev- enues and a significant increase in the stock and service of public debt, which was indexed to inflation (see Hermann 2005). Redemocratization in the Early 1980s: Rise in SNG Share of Revenue and Debt Distress The elections of 1982 marked the first step in the redemocratization of Brazil. The return to direct elections at the state level in 1982, after 22 years of governors being appointed indirectly, allowed the new gov- ernors to recoup their sources of power, creating alliances with local governments due to the coincidence of state and municipal elections. The results strongly favored the opposition—10 of 22 opposition gov- ernors were elected—including in the largest and most powerful states. While the central government lost political influence, the governors benefited from the economic recovery in 1984 and 1985.3 As a result, 38 Until Debt Do Us Part they were able to fulfill their mandates with high levels of investment (see Villela and Rezende 1986, 215). A 1983 constitutional amend- ment increased the transfers to SNGs and raised the tax basis of states. The states, which received 21.3 percent of disposable fiscal revenues in 1983 (corresponding to 5.7 percent of GDP), received 27 percent of revenues in 1986 (7.1 percent of GDP). At the same time, the munici- palities increased their participation from 8.9 to 12.1 percent of avail- able revenues, jumping from 2.4 to 3.3 percent of GDP, as shown in figure 1.1. The wave of democratization in the context of the macroeconomic crisis of the late 1970s to early 1980s encouraged the population to turn to the SNGs for services, even those that were the responsibility of the federal government. During discussions in the National Constituent Assembly during 1987–88, governors and mayors fought for a larger ­ share of public revenues. The 1988 Constitution greatly expanded the Figure 1.1  National, State, and Municipal Tax Collection and Disposable Revenue in Brazil, 1970–2010 40 35 30 25 % of GDP 20 15 10 5 0 70 72 74 76 78 80 82 84 86 88 90 92 94 98 20 0 20 2 04 20 6 08 10 96 0 0 0 20 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 19 20 Year Total tax collection Federal tax collection States’ tax collection Municipal tax collection Federal disposable revenue States’ disposable revenue Municipal disposable revenue Sources: Afonso 2011; National Bank of Economic and Social Development; Secretariat of Fiscal Affairs. Note: GDP = gross domestic product. Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 39 social responsibilities of the government; most of the expanded services provision was left to states and municipalities, however, which increased their expenditure obligations and, as a consequence, the borrowing ­ requirements. As a result, in 1991, SNG participation in disposable fis- cal revenue rose as high as 29.8 percent for states (7.5 percent of GDP) and 15.9 percent for municipalities (4 percent of GDP). Later develop- ments, mainly higher tax rates in the form of higher social contributions (which are not shared with SNGs), allowed the central government to increase its share of disposable revenue. Despite the increase in revenues, the SNGs’ fiscal difficulties deep- ened, as demand for services rose and debt accumulated. The credit crunch, the difficulty of rolling over debt, and the strong indexation of debt to inflation deepened the fiscal vulnerability of the SNGs, which resorted to short-term credit called anticipated revenue credit opera- tions (ARO), which by law is expected to be repaid in the same fiscal year, based on estimated tax revenue for the period. Nevertheless, this type of credit operation was often rolled over and accumulated year after year at much higher interest rates than regular credit or loans. In 1989, credit based on ARO accounted for 97 percent of state and munic- ipal debt (see Lopreato 2000, 127). In this difficult environment, the fiscal stance of the states and municipalities deteriorated significantly. In addition to their weak plan- ning and fiscal management capacity, states and municipalities could not count on adequate sources of financing for their increased spend- ing. Thus, several SNGs, in particular the states, ended the decade in massive fiscal distress. The subnational debt crisis, which erupted in the 1980s and the 1990s, was largely the result of the institutional reforms of the mili- tary government, which allowed debt to become a key source of fund- ing for governments. The regulations, which facilitated the access of public (and private) sectors to external resources in an environment of strong international liquidity, led to an “economy of debt.” While the term “economy of debt” explains a good part of the subnational debt story during this period, the responsibility of some states and munici- palities must also be understood. The most notorious example is that of the Municipality of São Paulo, whose debt was greater than that of the majority of states. 40 Until Debt Do Us Part Until the mid-1960s, the ability to raise funds was hampered by a legal ceiling on the nominal interest rate, which discouraged lenders in an environment of high inflation rates. Military government reforms included adjusting the value of bond debt for inflation (Almeida 1996). Removing the legislative cap on nominal interest rates was one of the most important reforms initiated by the military regime, because it helped develop a modern capital market in the country. Borrowing became a mechanism to circumvent the strict controls imposed on states and municipalities. The 1967 Constitution empow- ered the Senate to authorize SNG domestic and external credit opera- tions. The rules for contracting operations were defined by the central bank, which was responsible for regulating the financial and banking systems. In 1975, Senate Resolution No. 93 set strict rules for contract- ing credit operations. However, foreign operations and the so-called “extra limit” were not regulated. To induce the states and municipali- ties to act in line with the federal government’s interests, the social and urban projects considered as priorities by the federal government were funded by those “extra limit” operations. Having passed through the arbitrary sieve of the federal government, and authorized by the Senate, those operations became significant. After the second oil shock in 1979, regardless of the payment capacity of SNGs, the state and municipal public companies were induced to borrow from abroad as part of the federal strategy to reverse the public sector’s balance-of- payments deficit. To deal with the debt crisis unleashed by the Mexican moratorium in 1982 and the consequent need to turn to the International Mon- etary Fund (IMF) as the balance of payments deteriorated and for- eign reserves dwindled, the federal government adopted a tight fiscal policy in 1983–84. The tax burden was increased to 27 percent of GDP, and public investments were severely reduced. Nevertheless, inflation continued to erode tax revenue and swell public debt. SNG budget- ary and financial management was strictly conditioned by macroeco- nomic policies defined at the federal level. During the 1980s, SNGs were awarded “bridge loans” through the National Treasury so that they could roll over their external debt. In this way, SNG external debt was transformed into “domestic” debt, with the federal government as the main creditor. Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 41 SNG Debt Renegotiation and Its Macroeconomic Context Brazil experienced three major subnational debt renegotiations from the late 1980s to the late 1990s. These renegotiations must be under- stood within the broader context of macroeconomic and political reform. The focus of each renegotiation grew progressively broader. The first round focused narrowly on changing the terms of debt repay- ment. The second round accompanied major efforts to stabilize the macroeconomic situation and make the government more efficient, as the Brazilian economy moved from state control to open markets. The third round consolidated the fiscal adjustment with enactment of the FRL in 2000. This section briefly describes the main rounds of renego- tiation and analyzes their links with macroeconomic stabilization and fiscal consolidation.4 Addressing SNG Debt Distress: Three Major Renegotiations The first round of SNG debt renegotiation in 1989 focused on resolving SNG external debt. This problem had been fueled by the federal gov- ernment itself, which had sought to raise funds to fill the balance-of- payments gap. Based on Law No. 7976 of December 27, 1989, the federal government refinanced SNG debt, which included federal government loans in order to honor foreign debt guaranteed by the National Trea- sury or contracted until the end of 1988. In 1991, following the inauguration of new governors and the failure of the President Collor Plan II,5 authorities from all levels of govern- ment began to discuss a new approach, which led to the adoption of Law No. 8388 of December 30, 1991. However, the institutional crisis resulting from the impeachment of President Collor in 1992 postponed implementation of the law until 1993, when the second round of debt renegotiation started. The second round of SNG debt renegotiation in 1993 concentrated on the debt owed to agencies and entities controlled by the federal gov- ernment. Law No. 8727 of November 5, 1993, included outstanding amounts of loans contracted during the military period of 1964–85 for investments coordinated by the federal government.6 The new law restructured the terms of payment by extending the maturity of loans to 42 Until Debt Do Us Part as long as 20 years and limiting debt service to no more than 11 percent of real net revenue (RLR).7 In the event that debt service obligations exceed the 11 percent limit, the federal government would provide addi- tional loans to cover the excess, and the state would pay back the loan in 10 years. Furthermore, the law stipulated that interest rates would be kept as originally established. Although states requested that public securities (state Treasury bills and bonds) be included, they were excluded. In fact, both federal and subnational governments were concerned about the rapid growth of public securities and the increasing difficulties of refinancing them in the late 1980s and early 1990s. Constitutional Amendment No. 3 of March 17, 1993, limited the ability of SNGs to issue public securities until the end of 1999. SNGs were only allowed to issue public securities in the amount necessary to refinance the principal indexed by inflation, with one exception related to the issuance of SNG securities to pay for judicial writs.8 The third round of debt renegotiations was initiated in 1996. The first and second rounds restructured only part of subnational debt. With a substantial part of SNG debt unresolved and persistent pres- sures placed on the primary balance, the fiscal and debt position of SNGs continued to deteriorate. As hyperinflation was brought under control, the SNG debt situation worsened, with debt obliga- tions rising in real terms. SNG debt as a share of GDP increased from 9.2 percent in 1993 to 10.7 percent in 1996. The debt service pressures were even more severe. Law No. 9496 of September 11, 1997, autho- rized the third round of renegotiations by establishing criteria for consolidating, refinancing, and assuming the national debt and other securities held by states and the Federal District. The main difference between this round and the previous two was that a major fiscal con- solidation program underpinned the debt renegotiation. Although a bill relating to state debt was passed in 1997, legislation related to municipal debt was enacted only in 1999. Macroeconomic Stabilization The decade-long effort to renegotiate subnational debt unfolded within the broader context of national efforts to achieve macroeco- nomic stabilization and put Brazil onto a more sustainable growth Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 43 path. The second round accompanied two complementary mac- roeconomic adjustment programs: the Federal Immediate Action Plan (PAI) in 1993 (box 1.1), which sought to reorganize and reori- ent the entire public sector including SNGs; and the Real Plan in 1994 (box 1.2), which aimed to control inflation (monthly inflation rates were about 45 percent in the second quarter of 1994). The implemen- tation of these two macroeconomic programs helped shape the direc- tion and substance of the third, and most consequential, round of debt renegotiation. The main elements of the Real Plan in 1994 included the introduction of a new currency (the real), the de-indexation of the economy, an ini- tial freeze on public sector prices, the tightening of monetary policy, and the floating of the currency, with a floor specified for its value vis-à-vis­ the dollar. The Real Plan managed to break the vicious cycle of high inflation and to de-index the economy. Eichengreen (2007) argues that defending a particular exchange rate with capital mobility implies implementing domestic policies aimed at stabilizing the system. In Brazil, it was no different. The monetary policy during the Real Plan sought to maintain exchange rate stability, con- tributing to the success of the stabilization plan. Thus, in addition to the expected effects on production and consumption associated with a Box 1.1  The 1993 Federal Immediate Action Plan The 1993 Federal Immediate Action Plan (PAI) reorganized the public sector by untying earmarked revenues and reducing expenditures. The PAI was based on the assumption that inflation was caused by, among other factors, the financial and administrative disarray of the Brazilian state. The main provisions of the PAI provide for the following: Raising tax revenue by implementing a financial transactions tax and efforts to combat tax •  evasion Controlling public spending and using resources more efficiently •  Changing the nature of the relations between the national government and the SNGs, reducing •  unconstitutional transfers, and restructuring SNG debt Restructuring the public banks (both federal and state), which aimed to privatize government •  shares Deepening the privatization process, including the electricity and railroad sectors, and complet- •  ing the privatization of state-owned enterprises in the petrochemical and steel sectors Creating the Emergency Social Fund, which temporarily reduced the share of earmarked rev- •  enues at both the federal and subnational levels. 44 Until Debt Do Us Part Box 1.2  The Real Plan, 1994 The implementation of the 1994 Real Plan, with the PAI as its backbone, consisted of two steps. The first step encompassed monetary reform to combat inflation by introducing the real value unit (URV) to disrupt the indexing system of prices and incomes. Implemented on January 3,­ 1994, the URV followed the daily pro-rata variation of a set of three price indexes. All contracts were to be expressed in the new unit of account. To coordinate the inflationary expectations, the URV was pegged to the dollar. Use of the URV prior to adoption of the new currency was intended to realign relative prices, reduce the redistributive effects, and coordinate agent expectations. On July 1, 1994, the currency reform was completed, transforming the URV into the new currency, the real. The entire money stock was replaced, and the real became the medium of exchange. Wages, in turn, were converted to the new currency, calculated based on the average in URV between March and July. The second step adopted an anchor with the purpose of coordinating expectations. After a first attempt to use a monetary anchor, it became clear that a stronger instrument was needed to stabilize expectations. The choice fell naturally to the exchange rate. To meet the goal of coordinating inflationary expectations, the exchange rate regime was redesigned three times between 1994 and 1999: When the exchange rate targeting regime was adopted, the exchange rate was fixed between •  October 1994 and February 1995, despite an official system of bands. When the Mexican crisis arose in December 1994, the significant reduction in capital flows to •  emerging markets and the loss of foreign reserves by the central bank led to the adoption of a more widely fluctuating band. When the huge appreciation of the real against the dollar became evident, the exchange rate •  band was systematically adjusted to ensure the gradual devaluation of the real. Despite the changes, the central bank essentially set the exchange rate during this period, coordinating expectations and preventing the resurgence of inflation. Despite the evidence of excessive valuation, it was believed that a possible devaluation would negatively affect inflation. So even when the system showed signs of exhaustion, the central bank declined to redraft its essence. Only under a speculative attack in early 1999 did the central bank, unable to sustain the parity, drop out. restrictive monetary policy, higher interest rates also attracted specula- tive capital in search of greater profitability. As shown in figure 1.2, the Special Settlement and Custody System (SELIC)9 rate rose to more than 80 percent a year in February 1995. While the Real Plan focused on price stabilization, the PAI and the Emergency Social Fund (ESF) introduced greater budgetary flexibility, generated proceeds from privatization, and created more sources of rev- enue. Therefore, the PAI and the ESF helped create the conditions for the primary surpluses obtained by the consolidated public sector in the 2000s (figure 1.3). Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 45 Figure 1.2  SELIC Interest Rate in Brazil, August 1994–April 2011 90 80 70 60 Annual rate, % 50 40 30 20 10 0 c/ 7 ct 7 g/ 8 r/ 0 1 b/ 1 c/ 2 Ap 010 ct 2 g/ 3 Ju 004 r/ 5 b/ 6 c/ 7 ct 7 g/ 8 Ju 009 Ju 994 Fe 1996 Ju 1999 Ap 1995 01 Fe 200 Au 200 De 200 O 200 De 199 De 200 O 199 O 200 Ap 200 Fe 200 Au 199 Au 200 Ap 200 r/2 2 1 2 2 b/ n/ n/ g/ r/ / / n/ / n/ Au Month/year Source: Central Bank of Brazil. Note: SELIC = Special Settlement and Custody System. However, the PAI and ESF were insufficient to ensure short-term fis- cal balance, partly because a substantial share of SNG debt remained, even after the second round of debt renegotiation. Following the adoption of the Real Plan, especially during 1995–98, SNGs, particularly the states, began to run primary fiscal deficits (figure 1.4).­Importantly, state representatives ended their terms in 1994, while municipal representatives were in the middle of their four- year terms, which implied loose fiscal policy at the state level. Following political cycle, state public finances worsened, likely due to (a) the the ­ end-of-inflation factor (see next paragraph), and (b) the fact that state governments granted generous wage increases at the end of 1994 and especially in 1995, with full financial impact on the following years. The primary deficits during 1994–98 highlighted the structural imbalance of the SNGs. The sharp reduction in inflation in 1994 removed a public sector instrument that governments at all levels had used to delay payments, which reduced their real value while revenues were indexed. In addition, the rise in interest rates accelerated the growth of state and municipal securities debt, which was charged the 46 Until Debt Do Us Part Figure 1.3  Consolidated Primary Fiscal Balance in Brazil, 2001–10 4.0 3.7 3.8 3.5 3.4 3.3 3.3 3.2 3.2 3.2 3.0 2.8 2.5 % of GDP 2.0 2.0 1.5 1.0 0.5 0 01 02 03 04 05 06 07 08 09 10 20 20 20 20 20 20 20 20 20 20 Year Source: Central Bank of Brazil. Note: GDP = gross domestic product. SELIC rate—that is, the growth of securities debt was mainly due to the accumulation of interest on the principal. The primary surplus had to cope with higher interest rates, and aggregate SNG debt reached about 15 percent of RLR in 1995 and 10 percent of RLR in 1996 (Mora 2002). The majority of state and municipal issuers of public securities, who had serious difficulties generating adequate primary surpluses, could not even afford to pay the interest on the public securities. As a conse- quence, they began to roll over the entire debt based on federal Senate authorizations,10 and to use short-term ARO to finance normal bud- getary operations. The amounts were large in relation to the fiscal year budget. Even those that could pay at least a portion of interest flocked to the Senate to request the rollover of debt; they had no incentive to reduce their debt. In this context, SNG debt became explosive, requiring a third round of renegotiation that encompassed all the debt stock that had not yet been refinanced. Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 47 Figure 1.4  Primary Fiscal Balance of States and Municipalities in Brazil, 1991–2010 1.5 1.4 1.1 1.0 1.0 1.0 0.9 0.8 0.8 0.8 0.8 0.7 0.7 0.6 0.6 0.5 0.5 % of GDP 0.2 0.1 0 –0.2 –0.2 –0.5 –0.5 –0.7 –1.0 91 93 94 95 96 97 98 92 99 00 01 02 03 04 05 06 07 08 09 10 19 20 20 19 19 20 19 19 20 20 20 19 20 19 19 19 20 20 20 20 Year Source: Central Bank of Brazil. Note: GDP = gross domestic product. Does not include state- and municipal-owned companies. Third Round of SNG Debt Renegotiation and Fiscal Consolidation The central role of fiscal balance in the design of the Real Plan required SNGs to adhere to federal efforts to stabilize the economy and bring down inflation. Fiscal policy was a key ingredient, and the third round of debt restructuring sought to induce SNGs, particularly states, to gen- erate the primary surplus required to ensure long-term fiscal sustain- ability and macroeconomic stability. The renegotiation opened up the possibility that the federal govern- ment would give more incentives to SNGs to adopt the principles of the PAI plan (box 1.1). The 1996 renegotiation took place within the context of a discussion of the role of government in the economy. The privati- zation process at the federal level questioned the prevailing ­conception of state interventionism. Telecommunications and electricity, until then public monopolies, began to be managed by private companies, subject to regulatory agencies. As a condition for debt refinancing, the states would need to undertake fiscal and financial restructuring, ­ reorienting 48 Until Debt Do Us Part the role of the government to make it compatible with the changes occurring at the federal level. In addition, the restructuring programs were aimed at tackling the potential sources of imbalance, in order to create the conditions for the gradual repayment of debt. The contracts between the federal govern- ment and the states included a rigorous fiscal and restructuring adjust- ment program (the Incentive Program for the States’ Restructuring and Fiscal Adjustment, PAF),11 encompassing fiscal targets relating to the following indicators: (a) the ratio of debt to RLR, (b) the primary fis- cal balance, (c) the public sector wage bill, (d) own-revenue ­ collection, (e) privatization and concession of public services and utilities, (f) administrative reform, and (g) the ratio of capital expenditures to RLR. The state debt would be indexed to the general price index (General Price Index, Domestic Availability, IGP-DI)12 and charged an annual inter- est rate of 6 percent, provided that the state paid 20 percent of the debt within two years. These resources would be obtained by privatizing SOEs. Refinancing agreements and contracts covered almost all existing debts: • The total amount renegotiated reached 11.2 percent of GDP and was highly concentrated in the four largest debtors. The states of São Paulo and Rio de Janeiro accounted for 5.9 and 1.7 percent of GDP, respectively. The states of Minas Gerais and Rio Grande do Sul also had significant financial liabilities. These four large debtors absorbed about 90 percent of the amount renegotiated. • Depending on the initial conditions in each state and the possibil- ity of making an important down payment with proceeds from the privatization of state assets, the terms of the contracts varied: limits on the ratio of debt service to revenue, for example, could vary from 12 to 15 percent. Most of the states signed a 30-year contract, some of which had 15-year terms to maturity. The interest rate in most of the contracts was IGP-DI plus 6–7.5 percent a year. To fulfill their contractual obligations, state governments conducted a rigorous fiscal adjustment and began to generate primary surpluses. This extensive process of fiscal and financial restructuring also included the following: Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 49 • The privatization of SOEs, especially in the electricity sector • The restructuring and privatization of public banks owned by the states, which adhered to the Incentive Program for the Reduction of Participation of the States in Public Sector Banking • The reorganization of state public finances, with greater fiscal responsibility. In addition to spending cuts, states were encouraged to take loans to modernize tax administration, contributing to an increase in tax collec- tion, especially the VAT on goods, intermunicipal transportation, and communications services (ICMS). The ICMS revenue collection also benefited from an increase in the rate levied on the transport, electric- ity, and telecommunications sectors, characterized by low elasticity of demand. The fiscal adjustment strategy also increased the fiscal space for debt payment. If the states did not fulfill their commitments, the federal government was authorized to withhold transfers from the State Participation Fund (FPE) and even the states own tax, the ICMS. The PAF and its mechanisms for monitoring and enforcement were critical to the success of the renegotiation. The PAFs, signed by the gover- nors of 25 states,13 had annual targets for the following three years. Levels of compliance have been high because the debt renegotiation contracts have a specific clause that allows the National Treasury to stop making legal trans- fers to states that do not comply, and even to sequestrate states’ own tax rev- enue in case of nonpayment of the agreed portion of their debt. Each year the goals and commitments of the previous year are evaluated and the tar- gets updated where appropriate. Targets not achieved are subject to fines. These procedures will follow the refinancing contracts until their discharge. Therefore, it is reasonable to conclude that after the implementation of the Real Plan, the federal government had a clear interest in renegoti- ating the state’s debt because it was linked to other economic measures such as privatization of public utilities and state financial institutions, and implementation of programs to modernize tax administration at the state level. In addition, considering the macroeconomic context, the debt renegotiation had a crucial impact on improving the solvency and liquidity of the Brazilian banking system: large commercial bank assets which were held against the states (previously classified as high risk) were exchanged for high-yield federal assets (National Treasury bonds). 50 Until Debt Do Us Part The negotiation with municipalities was conducted separately start- ing in 1998. Unlike state debt, municipal debt was not addressed in Law No. 9496 in 1996, because the financial situation of municipalities was far less worrying than that of states. Nevertheless, municipalities such as Campinas, Guarulhos, Osasco, Rio de Janeiro, and São Paulo, which held debt securities, saw their situation worsening, since the assumption of state debt by the federal government meant that municipal securities now had to compete “alone” with federal securities in the market. Research conducted by the city of Rio de Janeiro14 indicated that the 50 municipalities with the highest RLR encompassed 29 percent of the population, 55 percent of RLR, and 82 percent of long-term municipal debt. The ratio of debt to RLR on average was 0.72. Only 28 municipalities had a ratio higher than 1. Furthermore, in April 1998, the securities debt of the two capital cities—São Paulo and Rio de Janeiro—constituted 25 percent of total SNG debt still to be renegotiated.15 The federal government issued Provisional Measure No. 1891 in January 1999, which extended to the municipalities the general conditions granted to the states without requiring fiscal programs, but tightened restrictions on debt contracting. The main terms of restructuring the municipalities’ debt were16 as follows: • An interest rate of 9 percent a year, which could be reduced to 7.5 percent a year, or 6 percent a year if the municipality extraordi- narily amortized an amount equivalent to 10 or 20 percent of the outstanding balance within 30 months of signing the contract • Exemption of municipalities from fiscal adjustment programs, which the states were obliged to participate in under the PAF. Around 180 municipalities signed contracts with the federal government restruc- turing their debt under these new conditions, which affected about 10 percent of all operations with the states. For larger municipali- ties, the possibility of extraordinary amortization only postponed the imposition of the 9 percent interest rate.17 In the macroeconomic sphere, the devaluation of the domestic cur- rency in 1999, amid a currency crisis,18 implied the end of the exchange rate anchor. This was followed by the adoption of an inflation-targeting regime to shield the Real Plan. When this system was introduced, SNG fiscal accounts were already improving. Still, policy makers wanted to Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 51 ensure that the changes would be lasting and to avoid any retreat from maintaining fiscal balances. After all, the main objective of the new macroeconomic policy was to stabilize the macroeconomic situation. The introduction of inflation targeting was accompanied by the adop- tion of a floating exchange rate regime and a vigilant effort to maintain fiscal austerity. The 2000 FRL and Long-Term Fiscal Sustainability The success of the Real Plan and the third round of subnational debt restructuring created the economic and political conditions for the approval and implementation of the FRL in 2000.19 The FRL established fiscal limits and restrictions on key fiscal variables, such as personnel expenditures and debt stock, and defined the responsibility for enforc- ing the fulfillment of obligations and transparency requirements (see box 1.3). The FRL was essential to strengthening the process of fiscal and financial restructuring of the government, including SNGs, and ensuring long-term fiscal sustainability. The approval of the FRL was the culmination of a series of key and incremental reforms in which sequencing and pace were critical.20 Among the most important were the following: • The creation in 1986 of the National Treasury Secretariat (STN) in the Ministry of Finance, with broad responsibilities for pub- lic finances, especially with regard to managing federal assets and liabilities21 • The 1988 Constitution, which expanded social responsibilities of the government, particularly for subnationals • The closing or privatization of state commercial banks in early 1990 • The 1998 Constitutional Amendment No. 19, which established new rules related to public administration and which made public the debate about issues like wage ceilings for public servants and pension systems. Also, the implementation of the Real Plan strongly influenced the fis- cal path to be followed by states and municipalities. Furthermore, according to Liu and Webb (2011, 10), “Even without a strong party system, a powerful president can enforce subnational fis- cal discipline. President Cardoso in Brazil became a strong president in 52 Until Debt Do Us Part Box 1.3  The Fiscal Responsibility Law (Complementary Law No. 101, May 4, 2000) The FRL was part of the efforts to strengthen fiscal institutions. The law applies to the federal ­ government, the states, the Federal District, and the municipalities; the legislature (including the audit courts); the judiciary and the attorney general’s office; and to their respective agencies, foundations, and funds. The FRL addresses the following issues: ­ Planning and budgeting. Besides defining basic parameters for the annual budget bill to be •  prepared by the executive, the Budget Guidelines Law defines fiscal targets relating to the pri- mary balance and debt, and to rules for fiscal management. It includes a detailed assessment of the government’s contingent liabilities. The Multiyear Plan, the Budget Guidelines Law, and the Annual Budget Law must be consistent. Debt. The FRL presents a detailed definition of consolidated long-term public debt, public secu- •  rities, credit operations, and guarantees; sets strict provisions on indebtedness and issuance of public debt by the central bank, and prohibits creditor debt-restructuring operations among the various levels of government. In accordance with Article 52 of the Federal Constitution, the Sen- ate passed a resolution establishing limits on indebtedness by level of government. If an SNG is in breach of the debt ceilings, new financing and discretionary transfers to the SNG are banned. Personnel expenditures. The FRL sets separate ceilings for personnel spending, including pensions •  and payment of subcontractors. Spending is limited to 50 percent of net current revenues for the federal government and 60 percent for states and municipalities. The law also establishes limits for the executive, legislative, and judicial branches. In case of noncompliance, the jurisdic- tion will not be allowed to engage in new credit line operations, and SNGs will not be allowed to receive transfers or credit guarantees from the federal government. Control and transparency. Budget reports and compliance with the limits set by the law must •  be reported every two or four months. The legislative branch of each level of government, sup- ported by its respective court of accounts, monitors the fiscal targets and ceilings. Procedures for record keeping and consolidation of accounts were significantly improved to reveal compli- ance with the provisions of the law. Article 167, Section III, of the Federal Constitution establishes a “golden rule” to prevent the use of borrowing to finance current expenditures: the amount of new loans contracted is limited to the amount of capital expense. Law No. 10028 (October 31, 2000) establishes penalties for public officials not complying with the FRL. the late 1990s even in the context of weak party loyalties and used his office (and reputation as an inflation fighter, from when he was minister of finance) to press successfully for fiscal discipline at the national and subnational levels.” All of these institutional changes informed the debate and helped the main agents agree to more fiscal responsibility and transparency, which were consolidated in the 2000 FRL. The preparation, discussion, approval, and implementation of the FRL followed a strategy in which Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 53 all key players—federal and subnational governments, the Parliament, the financial sector, the media, and public finance experts—were edu- cated about the need to change and enforce the framework for public finances in Brazil. The past “traumas” of fiscal distress, especially at the subnational level, also made the agents more open to introducing more controls and limits.22 SNG Fiscal Performance after Debt Renegotiation Agreements Since the first round of negotiations in 1989, the Brazilian economy has undergone profound changes, including a substantial improvement in macroeconomic indicators, especially in the fiscal stance. The insti- tutional reforms allowed, among other things, a new pattern of inter- governmental relations in which states and municipalities were able to achieve sound fiscal outcomes. As noted, following implementation of the Real Plan, the country adopted an inflation-targeting, flexible exchange rate regime and began to generate significant primary surpluses. Inflation was reduced and has remained below 5.9 percent since 2005. Starting in 2004, the economy experienced solid growth, with annual rates above 5 percent (except in 2009 when GDP fell 0.6 percent). In 2010, following the accumulation of foreign reserves, the public sector became a net creditor in foreign currency. Not only was SNG debt reduced, but federal government debt also fell, from 38 percent of GDP in 2002 to 27.5 percent in 2010. The rescheduling of SNG debt and the institutional changes that occurred in the post-FRL period led the fiscal adjustment occurring in the past decade. Since the debt renegotiation in 1999 and 2000, the ratio of SNG debt to GDP declined steadily, from more than 20 percent in the beginning of 2000 to around 13 percent in 2010. To assess the factors that contributed to this performance, we examine revenue and expenses during 2000–09.23 It is critical to examine the consolidated public finance data for both states and municipalities and to include indirect institutions such as SOEs, foundations, agencies, and autar- Excluding them runs the risk of underestimating the extent of chies.24 ­ contingent liabilities. 54 Until Debt Do Us Part Fiscal Turnaround: Deficit and Debt Trend Throughout the first decade of the 2000s, state and municipal finances improved significantly. SNGs generated a primary surplus of about 1 percent of GDP, reversing the deterioration of the previous two decades (table 1.1). The limits imposed by the FRL on public spending, debt, and debt service were crucial to these results. In addition, increases in bank credit operations for the public sector were subjected to controls and limits established by the National Monetary Council. Total public net debt as a share of GDP declined from 52 percent in 2001 to 39 percent in 2010, with all levels of government showing improvement in this indicator: for the federal government, it declined from 31.7 to 26.4 percent; for states, from 18.1 to 11 percent25; and for municipalities, from 2.2 to 1.8 percent (figure 1.5). These improvements were the result not only of GDP growth, but also of rising tax revenues and declining ratios of SNG net debt to GDP (figure 1.6). States have benefited significantly from the increase in revenues occur- ring after 1998. This is due in part to the fact that ICMS collection has been positively affected by higher taxes on certain products, such as telephony in several states; to the higher prices of petroleum products generally; and to improved tax collection efficiency (see Giambiagi 2008). ­ Municipalities experienced a substantial real increase in tax revenues, par- ticularly the tax on services (ISS). Total municipal tax revenues increased 50 percent over inflation from 2001 to 2010. The ISS grew 105 percent during the same period, while national GDP grew 40.7 percent. The reduction in state public debt can also be assessed by the ratio of net debt to net current revenue. This ratio was greater than 1 in 17 states in 2000, but in only 7 states in 2010. States with the majority of SNG debt, such as São Paulo and Rio de Janeiro, lowered their ratio of debt to net current revenue, respectively, from 1.9 and 2.1 in 2000 to 1.5 and 1.4 in 2010. In general, SNG budgetary execution became stronger, and the burden of debt service became looser. As seen earlier, the distribution of state debt is highly concentrated, favoring the richer states, which were the major beneficiaries of the debt renegotiation process. The three richest states—São Paulo, Minas Gerais, and Rio de Janeiro—accounted for about two-thirds of total debt in 2009 (table 1.2). Like state debt in Brazil, municipal debt is also highly concentrated. The four state capitals and five cities with medium and large populations Table 1.1  Public Sector Borrowing Requirements (PSBR) in Brazil, 1999–2010a % of GDP 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Consolidated nominal balanceb 5.3 3.4 3.3 4.4 5.2 2.9 3.6 3.6 2.8 2.0 3.3 2.6   States and municipalities c 3.2 2.0 2.0 3.8 1.5 1.8 0.2 0.6 0.5 1.2 -0.1 1.3    State governments 2.7 1.8 1.9 3.2 1.2 1.4 0.2 0.4 0.5 1.0 -0.1 1.1    Municipal governments 0.5 0.3 0.1 0.5 0.2 0.3 0.0 0.1 0.1 0.2 0.0 0.2 Consolidated nominal interestc 8.2 6.6 6.7 7.7 8.5 6.6 7.4 6.8 6.1 5.5 5.4 5.3   States and municipalities d 3.4 2.7 3.0 4.7 2.5 2.8 1.4 1.6 1.7 2.3 0.7 1.9    State governments 2.9 2.3 2.7 4.0 2.1 2.4 1.1 1.4 1.4 1.9 0.5 1.6     Real interest n.a. n.a. n.a. 0.3 1.1 0.7 0.8 0.8 0.5 1.0 0.5 0.6     Monetary correction n.a. n.a. n.a. 3.8 1.0 1.7 0.3 0.5 0.9 1.0 0.0 1.1     Municipal governments 0.5 0.4 0.4 0.7 0.4 0.4 0.2 0.3 0.2 0.3 0.1 0.3     Real interest n.a. n.a. n.a. 0.2 0.2 0.2 0.2 0.2 0.1 0.2 0.1 0.1     Monetary correction n.a. n.a. n.a. 0.5 0.1 0.2 0.0 0.1 0.1 0.2 0.0 0.2 Consolidated primary balance b -2.9 -3.2 -3.4 -3.2 -3.3 -3.7 -3.8 -3.2 -3.3 -3.4 -2.0 -2.8   States and municipalitiesc -0.2 -0.6 -1.1 -1.0 -1.0 -1.0 -1.1 -1.1 -1.1 -1.1 -0.8 -0.6    State governments -0.2 -0.5 -0.8 -0.8 -0.9 -0.9 -0.9 -0.9 -1.0 -0.9 -0.6 -0.5    Municipal governments 0.0 –0.1 –0.3 –0.1 –0.1 –0.1 –0.2 –0.1 –0.2 –0.2 –0.1 –0.1 Source: Central Bank of Brazil. Note: GDP = gross domestic product. n.a. = not applicable. a. Includes federal companies Petrobrás and Eletrobrás during 1999–2001, not affecting data for states and municipalities. b. Surplus in terms of PSBR concept. c. Includes companies. 55 d. The breakdown of monetary correction/real interest for SOEs is not available, so the total amount of nominal interest was added to monetary correction. 56 Until Debt Do Us Part Figure 1.5  Net Public Debt in Brazil, 2001–10 70 60 50 % of GDP 40 30 20 10 0 1 2 2 3 3 4 4 4 5 5 6 6 6 7 7 8 8 9 9 9 0 0 c /0 y/0 t/0 r/0 g /0 /0 /0 v/0 r/0 /0 /0 l/0 c/0 y/0 t/0 r/0 /0 /0 /0 v/0 r/1 p/1 a c a n n p b u g n n p e De p a c a M O M Au Ja Ju No A Se Fe J De M O M Au Ja Ju No A S Month/year Total debt Federal government and central bank States Municipalities Source: Central Bank of Brazil. Note: GDP = gross domestic product. Figure 1.6  Net Debt of States and Municipalities in Brazil, 1991–2010 25 20 15 % of GDP 10 5 0 91 992 993 94 995 96 997 98 99 00 001 002 003 04 05 06 07 08 09 010 19 1 1 19 1 19 1 19 19 20 2 2 2 20 20 20 20 20 20 2 Year State and municipal government net debt State and municipal net debt (includes companies) Source: Central Bank of Brazil. Note: GDP = gross domestic product. Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 57 Table 1.2  Share of GDP, Population, and Long-Term Debt of Borrower States in Brazil, 2009 % of states’ long-term State % of GDPa % of populationa debt, 2009b São Paulo 33.1 21.6 38.5 Minas Gerais 9.3 10.5 14.3 Rio de Janeiro 11.3 8.4 12.9 Rio Grande do Sul 6.6 5.7 9.4 Goiás 2.5 3.1 3.2 Bahia 4.0 7.6 2.4 Alagoas 0.6 1.6 1.6 Mato Grosso do Sul 1.1 1.2 1.6 Other states 31.5 40.2 16.1 Total 100.0 100.0 100.0 Sources: STN; Federal Institute of Geography and Statistics (IBGE). Note: GDP = gross domestic product. a. GDP and Population: 2008. b. Long-term debt stock on December 31, 2009. (which are home to 13.9 percent of the total population and generate 22.5 percent of GDP) account for 84.2 percent of long-term public debt and around 20 percent of judicial writ debt (table 1.3). The Municipal- ity of São Paulo alone accounts for two-thirds of the total. In general, the liabilities from judicial writs are greater for municipalities than for states. Currently, municipal debt with tax and legal or judicial obliga- tions in arrears contributes 9 percent of the outstanding balance of long-term debt, while judicial writs account for 27.7 percent. State Revenues and Expenditures The improvements experienced in the 2000s in state deficits and debt were the result of strong revenue growth and the consolidation of expenditures. On the revenue side, the growth of states’ own revenue surpassed the growth of real GDP by 30 percent, contributing to a gen- erally upward trend in the ratio of total revenue to GDP, as shown in table 1.4, which also identifies the source of revenues.26 The volatility of the main sources of budgetary revenues (including loans and other credit operations) is worth considering because it helps explain why investments are unstable at the subnational level (figure 1.7). 58 Until Debt Do Us Part Table 1.3  Share of GDP, Population, and Long-Term Debt of Municipalities in Brazil, 2009 % of municipal long-term Municipality/state % of GDPa % of populationa debt, 2009b São Paulo/SP 11.8 5.8 63.3 Rio de Janeiro/RJ 5.1 3.3 11.0 Campinas/SP 1.0 0.6 2.4 Belo Horizonte/MG 1.4 1.3 2.1 Salvador/BA 1.0 1.6 2.0 Guarulhos/SP 1.1 0.7 1.4 Contagem/MG 0.5 0.3 0.8 Joinville/SC 0.4 0.3 0.7 Mauá/SP 0.2 0.2 0.5 Other municipalities 77.5 86.1 15.8 Total 100.0 100.0 100.0 Sources: STN; IBGE. Note: BA = Bahia, GDP = gross domestic product, MG = Minas Gerais, SP = São Paulo. a. GDP and Population: 2008. b. Long-term debt stock on December 31, 2009. Table 1.4  State Revenue in Brazil, 2000–09 % of GDP 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Total revenue 12.9 12.7 13.1 12.4 12.5 13.0 13.0 12.7 13.2 13.2   Taxes 7.8 8.2 8.1 8.1 8.2 8.4 8.3 8.2 8.6 8.5    ICMS 6.7 7.1 7.0 6.9 7.0 7.1 7.0 6.9 7.2 7.1    Other taxes 1.0 1.1 1.2 1.2 1.2 1.3 1.3 1.3 1.4 1.5  Grants from other   governments 2.5 2.6 2.9 2.3 2.4 2.6 2.7 2.6 2.8 2.6    Current grants 2.4 2.5 2.6 2.2 2.3 2.6 2.6 2.5 2.7 2.5    Capital grants 0.1 0.2 0.2 0.1 0.1 0.1 0.1 0.1 0.1 0.2   Credit operations 0.3 0.1 0.3 0.2 0.1 0.1 0.1 0.1 0.1 0.4   Sale of assets 0.7 0.1 0.1 0.1 0.0 0.1 0.1 0.1 0.0 0.1  Miscellaneous   revenues 1.7 1.6 1.8 1.8 1.7 1.8 1.7 1.7 1.6 1.6 Sources: STN; state financial statements; IBGE. Note: GDP = gross domestic product. Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 59 Figure 1.7  Main Sources of State Revenues in Brazil, 2000–09 (2000 = 100) 4.0 3.5 3.0 2.5 % of GDP 2.0 1.5 1.0 0.5 0 01 02 03 04 05 06 07 08 09 10 20 20 20 20 20 20 20 20 20 20 Year Source: STN. During the 2000s, states’ own taxes and current grants from the fed- eral government grew steadily until 2009 and the onset of the global financial crisis (figure 1.7). States’ own taxes collected grew 7 percent on average annually.27 The ICMS represented more than 80 percent of states’ own tax revenues. The growth in ICMS was partially due to the growth of sectors such as telecommunications, and to the price of petroleum-related products, in addition to the states’ efforts to improve tax collection. Current grants from the federal government increased 36.2 percent during 2000–09 as a result of the decentralization of the provision of health services and the expansion of the FPE, through which 21.5 percent of the federal income tax and the value-added tax on industrialized prod- ucts (IPI) is granted to the states.28 During 2000–09, the FPE increased its share of current transfers from 48.4 to 56.9 percent. Although benefiting a few oil-producing states such as Rio de Janeiro, the rise in oil prices brought a large increase in the federal transfer of royalties. Total transfers were almost twice as high in 2008 as they were in 2003. 60 Until Debt Do Us Part The sources of budgetary revenue directly related to investments— capital grants and credit operations—declined until 2007 (figure 1.7). Capital grants generally are not mandatory and may be restricted by the targets set by the federal government,29 while current transfers are totally earmarked to specific social areas, leaving no margin for additional capital investments. In addition, credit operations are sub- ­ ject to limits on total public borrowing set by the central bank. From 2007 onward, both sources began to expand, even after the world finan- cial crisis. The improved macroeconomic conditions and fiscal perfor- mance allowed the states to raise new debt while reducing their ratio of debt to revenue. On the expenditure side, the FRL limits personnel expenditures to no more than 60 percent of net current revenue. The reduction in person- nel expenditures and the simultaneous increase in revenue (especially current revenue) improved this ratio: in 2000, 13 states were over the 60 percent limit compared with only a few states in 2009, reflecting a temporary deterioration caused by the global financial crisis. The con- trol of wage bill outlays, together with the reduction in investments30 as a percentage of primary revenue, allowed states to obtain primary surpluses to fund their debt service. In addition, as a result of the reduc- tion in debt stock, debt service fell as a share of GDP, from 1.2 percent in 2000 to 1 percent in 2009. Debt service in a given year may not include the total debt service incurred that year, since the debt renegotiation agreements placed a limit on the ratio of debt service to net current revenue. The major borrowers may have been accumulating residual amounts of debt ser- vice, which they capitalized as part of the debt stock, according to the debt renegotiation contracts. Table 1.5 presents expenditures as a share of GDP during the 2000s. The increase in grants to municipal govern- ments was due not only to the growth in tax revenue, but also to the creation of a financial fund for education services.31 At the subnational level, investments are highly dependent on fis- cal year current surpluses and eventual revenues from the sale of assets, which makes them susceptible to the economic cycle. Besides that, new administrations generally follow a political cycle, cutting current (mainly capital) expenses as soon as they take office and increasing spending at the end of their term.32 For example, in 2003 and 2007, new Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 61 Table 1.5  State Expenditures in Brazil, 2000–09 % of GDP 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Total expenditure 12.9 13.0 13.4 12.5 12.5 12.9 13.1 12.4 12.9 13.3   Personnel 5.8 5.8 6.1 5.8 5.6 5.7 5.8 5.6 5.5 5.7    Public servants 3.7 3.7 3.4 3.3 3.1 3.2 3.3 3.3 3.3 3.4    Retired/pensions 1.9 1.8 2.1 2.0 1.9 1.9 1.9 1.8 1.8 1.8   Other personnel   expenses 0.2 0.3 0.6 0.6 0.6 0.6 0.6 0.5 0.4 0.5   Grants to other   governments 2.0 1.9 2.4 2.2 2.3 2.4 2.2 2.3 2.5 2.5   Investments + acquisition   of financial assets 1.4 1.3 1.3 0.9 1.0 1.1 1.2 1.0 1.3 1.5   Debt service 1.2 1.1 1.1 1.1 1.0 1.0 1.1 1.0 1.0 1.0   Miscellaneous expenses a 2.5 2.8 2.4 2.5 2.6 2.7 2.8 2.5 2.6 2.7 Sources: STN; state financial statements; IBGE. Note: GDP = gross domestic product. a. Goods and services/social security transfers. administrations took power, which helps explain the relatively small amount of investments. During the 2008–09 global financial crisis, investments rose sharply due mainly to the expansion of capital grants and credit operations as part of countercyclical macroeconomic policies (figure 1.8). In 2009, the National Bank of Economic and Social Devel- opment offered federal loans to states to fund critical projects in the Federal Program for Growth Acceleration.33 Municipal Revenues and Expenditures Unlike the states, which generated a primary surplus in the early 2000s only after the third round of debt renegotiation, municipalities began to generate primary surpluses in 1997. This was an inaugural year of new municipal administrations, when expenditures are generally lower. The primary balance increased during 1999–2001, when debt renegotiations took place, and remained positive in the following years. Municipal revenues increased substantially during 2000–09 (figure 1.9). Tax collection grew at an annual average rate of 4.4 percent. Since 2004, the ISS has been growing no less than 10 percent per year in real terms, with the exception of 2009, which was during the global financial cri- sis, when the rate (still) was 3.3 percent higher than in the previous 62 Until Debt Do Us Part Figure 1.8  Main Sources of Investment Financing for States in Brazil, 2000–09 1.6 1.4 1.2 1.0 % of GDP 0.8 0.6 0.4 0.2 0 00 01 02 03 04 05 06 07 08 09 20 20 20 20 20 20 20 20 20 20 Year Capital grants Credit operations Investments + acquisition of financial assets Source: STN. Figure 1.9  Main Sources of Revenue for Municipalities in Brazil, 2000–09 (2000 = 100) 160 140 120 100 80 60 40 20 00 01 02 03 04 05 06 07 08 09 20 20 20 20 20 20 20 20 20 20 Year Taxes Current grants Capital grants Credit operations Sources: STN; municipal financial statements; IBGE. Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 63 year. Faster economic growth, changes in legislation, and moderniza- tion of tax management contributed to this improvement. Grants from the federal and state governments, by far the major group, increased during 2000–09, at an average rate of 4.8 percent per year. During the period, municipalities counted on additional resources for education (the FUNDEF and the Fund for Maintenance and Development of Basic Education and Teacher Training), and for health, since the provision of health services was decentralized. The composition of expenditures was relatively stable, consisting mainly of personnel and other miscellaneous expenses, most of which are current expenses incurred for the provision of services. Table 1.6 identifies the main items of expenditure. It also includes a memoran- dum item—the ratio of personnel expenses to net current revenues— that remained relatively stable throughout the period of analysis, with a slight increase in 2009. Figure 1.10 highlights the major municipal expenditures for this period. At the local level, the political cycle explains the smaller share of investments in total expenditures in the initial years of a new adminis- tration, as occurred in 2001 and 2005. The ratio of investments to total Table 1.6  Expenditures of Municipalities in Brazil, 2000–09 % of GDP 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Total expenditure 7.0 7.0 7.4 7.0 7.2 7.1 7.7 7.6 7.8 8.0   Personnel 3.0 3.0 3.2 3.1 3.2 3.3 3.4 3.3 3.4 3.7     Public servants 2.3 2.3 2.3 2.3 2.3 2.3 2.4 2.4 2.4 2.6     Retired/pensions 0.3 0.3 0.4 0.3 0.3 0.3 0.3 0.3 0.4 0.4    Other personnel   expenses 0.4 0.4 0.6 0.5 0.6 0.6 0.7 0.6 0.6 0.7  Investments + acquisition   of financial assets 0.8 0.7 1.0 0.8 0.8 0.6 0.9 0.9 1.1 0.8   Debt service 0.2 0.2 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3   Miscellaneous expenses a 2.9 3.0 2.9 2.9 2.9 3.0 3.2 3.1 3.1 3.2 Memo   Personnel/NCR (%) 43.1 42.1 43.5 44.8 43.2 41.3 42.8 42.8 41.7 45.6 Sources: STN; municipal financial statements; IBGE. Note: GDP = gross domestic product, NCR = net current revenue. a. Goods and services/social security transfers. 64 Until Debt Do Us Part Figure 1.10  Main Expenditures of Municipalities in Brazil, 2000–09 (2000 = 100) 200 180 160 140 120 100 80 00 01 02 03 04 05 06 07 08 09 20 20 20 20 20 20 20 20 20 20 Year Personnel Investments + acquisition of financial assets Debt service Miscellaneous expenses Sources: STN; municipal financial statements; IBGE. Note: GDP = gross domestic product. expenditures was 10 percent in 2001, rising to an average of 12 percent during the three remaining years of the term. In 2005, the ratio fell to 8.8 percent, recovering to an average of 12 percent in the remaining years of 2005–08 administrations. The lower ratio of debt to revenue results in a lower burden of debt service, which averaged 0.26 percent of GDP during 2000–09. The ratio of debt service to net current revenue, per the FRL target, averaged close to 3.5 percent. Debt service may not include the total amount incurred in the year, since limits are placed on payments to net current revenue. The sharp increase in debt service is a “base effect,” however, because the municipal debt renegotiations took place mainly during 1999–2000, and debt payments were resumed in 2001, with the National Treasury as the major creditor. To sum up, during 2000–09, except for some specificities such as the growth of debt until 2003, public sector borrowing requirements, debt, and personnel expenditures improved steadily. Growth in revenue and control of personnel expenses, particularly in the states, strengthened the fiscal accounts of SNGs, and this improvement was fundamental to Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 65 tackling the crisis of 2008–09.34 One important lesson is that the subna- tional fiscal framework, which was established at the end of the 1900s with the debt renegotiations, has not only allowed SNGs to improve their fiscal stance, but has also shown a strong resilience and a reasonable level of flexibility, overcoming without major changes two economic down- turns. First, during 2001–03, right after enactment of the FRL, ­ Brazil faced an energy supply crisis (especially electric power), which had a significant impact on economic developments, and a market confidence crisis associated with the 2002 election. Second, the f­iscal framework also stood firm during the 2008–09 global financial crisis, which is analyzed in the next section. Impact of the Global Financial Crisis and Challenges to Subnational Fiscal Sustainability Brazil showed strong resilience during the 2008–09 financial crisis. Blanco, Barbosa Filho, and Pessôa (2011) conclude that Brazil’s adoption of far-reaching structural reforms and price stabilization initiatives in the 1990s, followed by the consistent pursuit of sound macroeconomic poli- cies in the 2000s, strengthened the country’s resistance to external shocks. Nevertheless, the global financial crisis affected the Brazilian econ- omy and state and municipal finances. The immediate and perhaps most important impact was on financial markets—a reduction in external credit and a sharp depreciation of the domestic currency. The downturn in economic activity in the world’s major economies had a significant impact on the price of commodities exported by Brazil: exports fell 10.2 percent in 2009 compared to 2008, reducing tax revenue, particu- larly the ICMS of exporter states. The government’s active management of macroeconomic policies— fiscal, monetary, and external policies, in particular—mitigated the effects of the crisis. After contracting 0.3 percent in 2009, real GDP rebounded in 2010, growing 7.5 percent and, because of the continuation of the global economic crisis, economic growth slowed in 2011 with GDP increasing only 2.7 percent. Inflation increased to around 6 percent a year but is still close to the upper band of the inflation targets defined by the cen- tral bank. Gross official reserves recovered, and credit to the private sector returned to trend levels in 2010. 66 Until Debt Do Us Part The decline in economic activity and the implementation of a set of tax exemption measures aimed at sustaining aggregate household consumption led to a fall in tax revenue in 2009. Tax as a share of GDP declined from 34.4 percent of GDP in 2008 to 33.6 percent in 2009, as GDP contracted 0.6 percent and total tax collection fell almost 3 percent. The federal government increased its spending on salaries for civil servants and raised the minimum wage, affecting the payment of pension benefits. As a consequence, the consolidated primary fiscal surplus fell to 2 percent of GDP in 2009, well below the average level of 3.5 percent during 2004–08. Inflation as measured by IGP-DI declined 1.4 percent in 2009, reducing the nominal interest rate and the need for public sector borrowing.35 The countercyclical fiscal policy included many measures that sup- ported state and municipal revenues: (a) an increase in capital grants due to federal investments programmed under the Federal Program for Growth Acceleration; (b) a reduction in income tax collection and cuts in the IPI tax, which helped support the auto industry, prevent job losses, and reduce SNG transfers through the FPE and Munici- palities Participation Fund; (c) compensation for the loss of current grants—states received additional credit lines from the National Bank of Economic and Social Development to keep investments and to match federal capital transfers related mainly to the Federal Program for Growth Acceleration; and (d) receipt by municipalities of general grants to keep the amounts transferred from the Municipalities Participation Fund the same as in 2008. Public banks expanded the supply of credit to help exporters finance their costs in foreign currency and to help corpo- rations and consumers offset the decline in private credit.36 Impact of the Crisis on States and Municipalities The impact of the crisis differed across states and municipalities and across participants in each group. It is necessary to examine the impact not only of the crisis but also of federal policies to offset the economic downturn.37 The immediate impact of the economic downturn on the fiscal accounts of states was to reduce ICMS collection and transfers from the FPE, the latter due to the decrease in federal revenues from the income tax and tax on industrialized products. For all the states, these Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 67 two sources of revenue have represented around 50–70 percent of dis- posable income (after transfers to municipalities).38 The impacts of the global crisis on state finances began to be felt only in late 2008 and were not enough to erase the revenue gains in 2007 and most of 2008: • In 2008, the ICMS, the main source of states’ own revenues, grew at an average real rate of 9.5 percent compared with 3 percent in 2007 (8 out of the 27 states had increases over 10 percent). FPE trans- fers from the federal government performed even better, increasing 12.9 percent compared with 3 percent in 2007 (uniform for all the states). In 2009, the reduction in revenues from the ICMS (2.5 per- cent as a national average) and in the FPE (8.9 percent) was only a reduction relative to the excellent performance of 2008. In 2010, the ICMS tax collection had recovered to its precrisis level. •  Fiscal indicators for a sample of 10 states during 2007–09 confirm these results.39 In 2009, all but one state in the sample had revenue losses compared with 2008, but they were in a stronger position than in 2007 as measured by net current revenue (RCL).40 The investment boom during 2008–09 was stupendous. In 2007, investments were financed mostly by “other internal sources,” which included the cur- rent surplus (after payment of debt service). The expansion of invest- ments was an important element in the expansion of capital grants and credit operations, although the composition of sources differed among states. The federal government, through National Monetary Council resolutions, provided additional resources to states through loans to compensate the loss of transfers from the FPE. The increase in funds raised through new indebtedness was small compared with the outstanding stock of debt; thus, the ratio of net debt to net cur- rent revenue was not seriously affected. Total credit granted in 2009 was about 2.5 percent of the outstanding contractual debt at the end of 2009. Moreover, the growth in net current revenue and decline in the IGP-DI helped reduce the ratio.41 The impact of the global crisis was less intense on municipalities than on states. While states generally lost current revenues during 2008–09, municipalities were able to offset the losses from ICMS transfers through increases in other transfers and own taxes. Fiscal indicators for a sample of six municipalities (Belo Horizonte, Cuiabá, Porto Alegre, 68 Until Debt Do Us Part Rio de Janeiro, Salvador, and São Paulo) show that in 2009, total rev- enue remained constant, mainly due to the following factors: •  The property tax and the tax on services grew faster than inflation (5 and 3.3 percent, respectively), allowing total tax revenue to grow in real terms. • The federal transfers targeted to health services increased 10.2 percent  in real terms, compensating the slight real reduction (0.5 percent) in other current transfers. • The federal government adopted a policy of using voluntary transfers  to compensate the loss of transfers from the Municipalities Participa- tion Fund (around 25 percent of current grants), in order to main- tain the nominal amount received by each municipality during the previous year. • The resources of the Fund for Maintenance and Development of  Basic Education and Teacher Training grew largely as a result of state contributions. In contrast to state investments, municipal investments, as reflected by investments of the six capital cities, either decreased or grew just slightly in 2009.42 In addition, the stability of revenues resulted in a sta- ble ratio of personnel expenses to net current revenue. Challenges to Subnational Fiscal Sustainability The fiscal performance of Brazil’s SNGs during the last decade was impressive. However, SNGs are facing challenges that may have an impact on their long-term fiscal sustainability. Here we highlight three issues that bear a more immediate relationship to the debt restructuring framework discussed in previous sections:43 (a) the indexation rules for debt that was renegotiated at the end of 1990s and the accumulation of residuals, (b) the narrow fiscal space and low current public investment, and (c) the potential for reducing debt service cost through more com- petitive subnational credit markets. We deal with each in turn. As noted, the SNG debt renegotiation indexed the subnational debt to the IGP-DI, plus a “real” interest rate of about 6 percent. This arrangement was reasonable at the time of the renegotiation because of the then prevailing macroeconomic conditions. The interest rate (or the SELIC rate, which is the basic funding cost of the federal government) Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 69 was much higher than the conditions agreed. The exchange regime had a fixed exchange rate, and the inflation targeting regime was not yet part of the monetary instruments. Finally, there was no umbrella fiscal law, like the FRL, establishing fiscal limits and conditions. Now, the macro- economic context has changed with new, improved mechanisms such as a floating exchange rate, and inflation-targeting regimes resulting in a higher level of macro stability. More important, the SELIC rate has steadily declined, especially since the global economic crisis, and the cost of the renegotiated contracts is much higher than the basic interest rate in the domestic economy. As a result, states would like to refinance their debt in the market, but the contracts forbid doing so.44 The state authorities have initiated negotiations in the Congress in the hope of changing the conditions of the original contracts. The IGP- DI is highly volatile, however, and therefore not a good reference for the asset and liability management of federal and SNG governments. A related challenge that is derived from the debt renegotiation contracts is the 13 percent cap on net revenues. Although this cap is a good coun- tercyclical feature, which was useful for the most indebted states, it has generated a residual in several states—building up capitalized interest, which is a potentially serious financial problem. According to the leg- islation, this residual must be repaid in 10 years starting in 2030. The residuals also depend on GDP growth and the trend of the IGP-DI. For some states, such as Rio de Janeiro, residuals could disappear if GDP growth averages 4 percent during 2010–20 and the IGP-DI converges to inflation (according to Levy 2009). Official projections for the Munici- pality of São Paulo estimate that if the growth of net current revenue is about 1 percent per year, the ratio of debt to net current revenue will be about 327 percent in 2030, and the ratio of debt service to net cur- rent revenue may reach 51 percent, indicating an unsustainable path (according to Prefeitura da Cidade de São Paulo 2011). The narrow fiscal space and the various regulations for new indebt- edness have posed challenges in financing public investments at the sub- national level. The brunt of SNG fiscal adjustment has fallen on public investments, resulting in a deterioration in infrastructure and threaten- ing economic growth. While the limits imposed on SNG debt financing as part of the fiscal adjustment program have been successful in turning around the SNG fiscal imbalance, a challenge ahead is to identify ways 70 Until Debt Do Us Part of increasing infrastructure investments and finance while maintaining fiscal discipline. In the long term, economic growth is a key determinant of fiscal sustainability. Finally, another challenge is to develop competitive credit markets for subnational public investments. More competitive subnational credit markets help lower the cost of financing and extend maturity. Cost of financing is another key determinant of fiscal sustainability. Canuto and Liu (2010) show that the subnational debt market in developing countries has been undergoing a notable transformation. Private capital has emerged to play an important role in subnational finance in coun- tries such as Poland, Romania, and South Africa. Subnational bonds increasingly compete with traditional bank loans. Notwithstanding the temporary disruption of credit markets during the crisis, the diversifi- cation of subnational credit markets is expected to continue. SNGs or their entities in various countries have already issued bond instruments (for example, in China, Colombia, India, Mexico, Poland, the Russian Federation, and South Africa). More countries are considering policies to foster the development of subnational debt markets (for example, Indonesia and South Africa), while others are piloting transaction and capacity-building activities to the same end (for example, Peru). Com- petitive financial markets, with a variety of buyers and sellers and a vari- ety of financial products, can keep borrowing costs down.45 In Brazil, public financial institutions are currently the primary pro- viders of credit to SNGs.46 As part of debt restructuring, SNGs have been prohibited from issuing bonds. This has helped bring subnational debt onto a more sustainable path. Looking forward, a key challenge is to identify ways of increasing competition in the subnational credit market to reduce the cost of financing and help fiscal sustainability in the long term. International experience shows that prudent regulatory frameworks for subnational capital markets are important to manage the risks of defaults, and the process of developing such frameworks is gradual and incremental. Conclusions The SNG debt crisis in Brazil was the result not only of autonomous decisions by the SNGs, but also of decisions by the federal government Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 71 to implement economic development plans amid financial internal and external constraints, mainly during the authoritarian political regime of 1964–85. The SNGs had an important role in both attracting external resources that eased the pressure on external accounts and implement- ing large projects to accelerate urbanization. Coupled with the exter- nal fundraising of SNGs—led by central government macroeconomic policy—the recourse to borrowing was a way to circumvent the strict controls on fiscal management, with a strong negative effect on the fis- cal accounts of SNGs, particularly the states. The control of subnational borrowing in the form of debt renegotia- tion agreements in 1997 imposed strict rules on the financial and fiscal management of SNGs and assured their adherence to the guidelines set by the federal government for the conduct of macroeconomic policy. The latest round of renegotiations during 1997–2000 paved the way to macroeconomic stability. Building on the success of the Real Plan, in 1999 Brazil adopted a regime of inflation targeting and flexible exchange rates and initiated a strong fiscal adjustment program that generated significant primary surpluses. Inflation was reduced and has stayed below 5.9 percent since 2005, partly explained by the strong expansion of imports at progres- sively lower prices due to the appreciation of the Brazilian currency during 2003–10. Starting in 2004, the Brazilian economy began a sus- tained path of economic growth. Brazil was resilient in the face of the 2008–09 global financial crisis, owing to the combination of sound macroeconomic management and a highly favorable external scenario throughout the 2000s (except for the international liquidity crisis in 2009). Both SNG and federal government debt levels were reduced, with federal government debt falling from 38 percent of GDP in 2002 to 26.4 percent in 2010. Following the accumulation of reserves, the public sector became a net creditor in foreign currency. The country achieved trade surpluses, mainly between 2005 and 2007, averaging US$43.7 billion a year. The country’s institutional consolidation and modernization, result- ing in approval of the Fiscal Responsibility Law in 2000, is crucial to understanding the trajectory of relative fiscal consolidation and reduc- tion in the ratio of debt to GDP achieved in the last 11 years. However, the institutional reforms do not end with approval of the FRL. The FRL 72 Until Debt Do Us Part principles, procedures, and requirements would be ineffective if not followed by other initiatives such as continued efforts to improve rules and procedures for registering and reporting on public accounts and the adoption of uniform criteria to demonstrate compliance with legal limits. These efforts have given more credibility to financial statements, particularly from SNGs. This chapter shows that in all government spheres—federal and subnational—fiscal adjustments and consolidation have been greatly facilitated by the growth of government tax revenue. Since the debt restructuring, almost all states and a large number of municipalities have been modernizing their tax administration. Revenue growth, how- ever, has not translated into expanded public investment, leaving urgent projects in urban infrastructure with insufficient funding. In contrast, the provision of basic public services has expanded—increasing the coverage of health services and education, for example—but this has increased the need to fund higher current expenditures. Notwithstanding the success of fiscal reform and debt restructur- ing, Brazilian SNGs face challenges that may impact the sustainability of subnational finance and economic growth. The three challenges that have more direct bearing on debt restructuring are the indexation rules and potential accumulation of residuals, narrow fiscal space to finance the public investment gap and its impact on economic growth, and the need to foster competitive subnational credit markets for lowering financing costs for public investments. Notes  The findings, interpretations, and conclusions expressed in this work are those of the authors and do not necessarily reflect the views of The World Bank, its Board of Executive Directors, the governments they represent, or any other institutions with which the external authors may be affiliated. 1. Brazil is a federal republic, encompassing a federal government (the union), 26 states, a Federal District (Brasília), and 5,564 municipalities. Ranked fifth in world population, Brazil had 191 million inhabitants in 2010. States and municipalities vary greatly in size. Unless otherwise indicated, the states include the Federal District. Serra and Afonso (2007) remind us that the federative framework in Brazil is a 2.  system that is still evolving and point out main aspects in terms of vertical and horizontal decentralization and government powers. Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 73 3.  The restrictive fiscal policy and reduced reliance on imports as a result of investments made during 1974–79 (the second National Development Plan), coupled with the world economic recovery, allowed GDP to grow 5.4 percent in 1984 and around 4.5 percent in 1985 and 1986 (Giambiagi et al. 2005). 4. Rezende and Afonso (2002) find that two important facts have shaped the way Brazilian fiscal federalism currently looks: (a) the transition from authoritarian rule to democracy, following the demise of the military regime in 1985; and (b) the policies adopted in the 1990s to achieve domestic and external balances. Rigolon and Giambiagi (1999) describe the context in which the renegotiation took place and emphasize that it was part of the administrative measures aimed at reducing SNG indebtness. 5. The Collor Plan II attempted to deal with the persistent inflation and increasing difficulty in placing public bonds. The plan aimed to eliminate overnight oper- ations and other forms of price indexation. A Financial Investment Fund was created as a captive market for government securities. Fiscal austerity measures included the blocking of federal spending by the ministries and state compa- nies. Despite lower inflation, political resistance to the economic team doomed the plan (see Gremaud, Vasconcellos, and Toneto 2010). 6. Law No. 7976 of 1989 refinanced the direct and indirect debt of SNGs derived from loans that had been granted by the national government in order to meet commit- ments due to external credit operations, guaranteed by the National Treasury. 7. RLR is used to calculate the SNG’s debt payment limit and the ratio of finan- cial debt to RLR. It is also used as a parameter for the states’ fiscal adjustment and restructuring programs. It is calculated as a moving 12-month average of revenue collected, excluding revenue from credit operations, sale of prop- erty, voluntary transfers, donations received to meet capital expenses, and, in the case of states, transfers to municipalities, due to legal or constitutional participations. 8.  Judicial writs, called precatórios, are legal requests for payment of a certain amount by the federal, state, or municipal treasuries. They cannot be appealed. At the end of judicial enforcement, a letter is submitted to the president of the court requiring payment of the debt. The requests received by the court until July 1 of each year are included in the budget proposal, to be paid the following fiscal year. SNGs often have significant amounts of these debts in arrears. 9.  The SELIC rate—Brazil’s prime interest rate—is the average interest rate charged on daily financing of operations and is backed by government securi- ties registered in the SELIC. 10.  According to the Federal Constitution, the Senate is entitled to define amounts and conditions of debt issued by the union, states, and municipalities. 11.  The debt negotiations have adopted a “contractual approach.” See Grembi and Manoel (2011) for an analysis of Latin America. 12.  The IGP-DI is a weighted index of the Wholesale Price Index, the Consumer Price Index, and the National Construction Cost Index, with weights equal to 74 Until Debt Do Us Part 6, 3, and 1, respectively. The “Domestic Availability” version was created in 1969 with the aim of isolating the effects of coffee price oscillations by assigning a lower weight to export products. 13. The states of Amapá and Tocantins had no significant debt so were not included. 14. Based on Ministry of Finance data. 15. The states of São Paulo and Bahia had already completed their renegotiation. 16. At the end of 1998, the Municipality of Rio de Janeiro presented a proposal for restructuring its debt with the federal government. The proposal suggested the same restructuring terms that applied to the states (Law No. 9496) (see Prefei- tura Municipal do Rio de Janeiro 1998). 17. Recently, the Municipality of Rio de Janeiro restructured its debt through a World Bank loan of US$1.045 billion to repay up to 25 percent of the debt renegotiated with the federal government. This prepayment reduced the inter- ­ est from 9 to 6 percent. 18. See IMF (2003, 9) for a detailed explanation of the crisis. According to the IMF, the Brazilian crisis, like others in emerging markets during the 1990s, would be better “described as capital account crises to distinguish them from the more conventional crises which have their origins mainly in the current account.” 19. The Fiscal Responsibility Law was approved on May 4, 2000, as Complementary Law No. 101 (see box 1.3). Passage of a complementary law requires a higher threshold of voting (75 percent of both houses). 20. Leite (2011) argues that some of these factors are key to explaining why the FRL was passed in the Brazilian Congress. 21. Setting up the National Treasury at the Ministry of Finance was a milestone in terms of comprehensive analysis and control of public finance in Brazil. Among other responsibilities, it included (a) executing the budget (cash flow), (b) overseeing subnational public finances, (c) monitoring financial aspects of government banks and SOEs, (d) public debt management (registering, con- trolling, reporting), (e) managing federal government financial assets, and (f) accounting and reporting on federal government accounts. For several years, the National Treasury was also in charge of “internal control” activities for the federal government. 22. Tavares, Manoel, and Afonso (1999) describe the FRL project and the anteced- ents of the new fiscal rules. 23. The main sources of information in this section are the Federal Institute of Geography and Statistics (IBGE), the Central Bank of Brazil, the STN, the Council of States Secretaries of Finance of the Ministry of Finance, and the financial statements of several states. Data pertaining to budget execution of the states for 2000–09, made available by the STN based on information pro- vided by the states, were consolidated. Despite differences in methodology and coverage among the different sources, the results obtained for the indicators are consistent, allowing comparison of the data. Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 75 Central Bank of Brazil data for net public debt and main fiscal indicators distin- 24.  guish direct and indirect administration (state and municipal governments) from the figures for SOEs. STN data consolidate figures for revenues, expenditures, and gross debt for direct and indirect administration with those related exclusively to dependent companies. Data for municipalities during 2000–09 cover around 95 percent of Brazilian cities. The debt stock is the debt of both state governments and state companies. 25.  Data on SNG revenue and expenditures are based on information made avail- 26.  able by the STN and by states and municipalities through official sites. Time series presenting the evolution of revenues, shown in figure 1.9, eliminate the impact of inflation. Beginning in 2000, taxes include federal income tax withheld at source and 27.  the payment of tax debts in arrears (dívida ativa). In 2000, both accounted for 4.5 percent of total tax revenue. Therefore, comparisons can be made for only 2000–09. Reports from the Brazilian Internal Revenue Service (Receita Federal) indi- 28.  cate that the ratio of income tax to GDP was 12.6 percent in 2009, up from 5.5 percent in 2000. The ratio of the IPI to GDP was 1.7 percent in 2006, up from 0.9 percent in 2000, due to a temporary exemption from the IPI granted to the automotive sector. To comply with the targets for primary balance set by the Congress, the federal 29.  government may have to reduce the amount of capital grants to SNGs. Investments include the acquisition of financial assets, mostly related to the 30.  issuance of bonds of independent state companies. The Financial Fund for Educational Services (FUNDEF) was created by Con- 31.  stitutional Amendment No. 14/96 instituting intergovernmental financial cooperation for improving elementary education. The fund is financed by ear- marking percentages of transfers from the revenue-sharing system to guarantee a specified minimum amount of spending per student enrolled in public ele- mentary schools throughout the country. FUNDEF is funded by (a) 15 percent of the municipal and state share of the ICMS; (b) 15 percent of the Municipali- ties Participation Fund; (c) 15 percent of the FPE; and (d) 15 ­ percent of the ­ municipal and state share in the Compensation Fund for Exports. FUNDEF funds are distributed according to the number of students enrolled in munici- pal- or state-owned elementary schools. If the money collected from these sources is not enough to guarantee the minimum spending established by law, the federal government is responsible for providing ­ supplementary ­ transfers. In 2007, the earmarking of revenues was expanded, with the creation of the Fund for Maintenance and Development of Basic Education and Teacher Training, to 20 percent from 15 percent, and other revenues were included, mainly state revenues. The distribution criteria were also changed to take into account, among other things, students in higher education, a service provided by the states. 76 Until Debt Do Us Part 32. In Brazil, federal, state, and municipal officials serve four-year terms. The head of government—the president, the governor, and the mayor—can be reelected once. Federal and state terms coincide, and municipal terms begin and end two years thereafter. Municipal administrations took office in 2001, 2005, and 2009. Federal and state administrations took office in 2003 and 2007. 33. In 2009, the state of São Paulo sold the state bank, Nossa Caixa, to finance investments. 34. Piancastelli and Boueri (2008), after examining the evolution of state fiscal accounts after 10 years of debt renegotation, conclude that the renegotiation was a major effort to improve the fiscal stance of the public sector. 35. Central bank data; http://www.bc.gov.br. 36. See the Ministry of Finance website; http://www.fazenda.gov.br. 37. Data used in this analysis can be found on the STN website and on the web- sites of several states and municipalities. Information about 2010 municipal finances came from FRL reports, available on the websites of municipalities and the STN. Some indicators may not be available in some years for all the states and municipalities. The Finanças do Brasil (FINBRA) database for states and municipalities is available at http://www.stn.fazenda.gov.br/estados_muni cipios/index.asp. 38. According to IBGE, the Brazilian territory encompasses five geographic regions: North, Northeast, Center-west, South, and Southeast. The North and Northeast are the poorest, and the South and Southeast are the richest. The nation’s capi- tal, Brasília, is in the Center-west Region. 39. States in the sample are Bahia, Ceará, Goiás, Maranhão, Minas Gerais, Paraná, Pernanbuco, Rio de Janeiro, Rio Grande do Sul, and São Paulo. 40. RCL is calculated as current revenue (taxes, intergovernmental transfers, and other) minus constitutional transfers to municipalities (obligatory) minus the revenue of social contributions to the public servants’ pension fund. 41. As indicated, the debt stock is subject to monetary correction by the IGP-DI, which varied negatively in 2009 (−1.43 percent). 42. This was the first year of a four-year (2009–12) administration, and, follow- ing the political cycle, current expenditures, mainly investments, are reduced, allowing the administration to accumulate cash to spend during the last years of its term. 43. There are other factors influencing SNG fiscal sustainability. These include the intergovernmental fiscal system, the expenditure framework, and taxation reform, which are beyond the scope of this chapter. 44. For example, Article 35 of the FRL prohibits any form of debt renegotiation between public entities. Pellegrini (2012) also analyzes several constraints, such as legal, technical, and fiscal, which impede normal renegotiation of the current debt. 45. Various countries have been moving toward more diversified instruments, including bonds. Total SNG bond issuance in developing countries reached US$45.1 billion during 2000–07 and US$102.8 billion during 2008–10Q1, with Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 77 China being the largest and dominant issuer, followed by Russia (Canuto and Liu 2010, based on data from DCM Analytics). Central bank data on domestic debt of the four largest states in 2010 (includ- 46.  ing direct and indirect administrations) indicate that private institutions pro- vide only 0.5 percent of total funding for domestic debt. They have a notable presence only in Minas Gerais and São Paulo, where they hold about 20 and 10 percent, respectively, of the outstanding SOE debt. Bibliography Abrucio, Fernando Luiz. 2002. Os barões da federação: Os governadores e a redemoc- ratização brasileira. 4th edition. São Paulo: Hucitec. Afonso, José Roberto. 2011. Crise, estado e economia brasileira. São Paulo: Agir. Almeida, Anna Ozório de. 1996. “Evolução e crise da dívida pública estadual.” Texto para Discussão 448, Instituto de Pesquisa Econômica Aplicada (IPEA), Brasília. Bacha, Edmar. 1995. “Plano Real: Uma avaliação preliminar.” Revista do BNDES 2 (3): 3–26. Blanco, Fernando, Fernando de Holanda Barbosa Filho, and Samuel Pessôa. 2011. “Brazil: Resilience in the Face of the Global Crisis.” In The Great Recession and Developing Countries: Economic Impact and Growth Prospects, ed. Mustapha K. Nabli. Washington, DC: World Bank. Canuto, Otaviano, and Lili Liu. 2010. “Subnational Debt Finance and the Global Financial Crisis.” PREM Economic Premise 13: 1–7. Eichengreen, Barry. 2007. Globalização do capital: Uma Historia do Sistema Mon- etário International. São Paulo: Editora 34. Garson, Sol. 2010. “Gestão fiscal do Espírito Santo 2002–2008: Pavimentando o caminho para o crescimento da economia estadual.” In Espírito Santo: Institu- ições, desenvolvimento e inclusão social. Vitória: Instituto Jones dos Santos Neves. Giabiagi, Fabio, Fernando Blanco, and Wagner Ardeo. 2008. “Proposta de recriação de uma dívida estadual e Municipal com o Setor Financeiro Privado.” Revista do BNDES 14 (29): 365–98. Giambiagi, Fábio. 2008. “18 anos de política fiscal no Brasil: 1991/2008.” Economia Aplicada 12 (4): 535–80. Giambiagi, Fábio, and Ana Claudia Além. 2007. Finanças públicas: Teoria e prática no Brasil. 3rd edition. Rio de Janeiro: Editora Campus. Giambiagi, Fábio, André Villela, Lavinia Barros de Castro, and Jennifer Herrman. 2005. Economia brasileira contemporânea (1945–2004). Rio de Janeiro: Editora Elsevier. Gremaud, Amaury Patrick, Marco Antonio Sandoval Vasconcellos, and Rodinei Toneto Jr. 2010. Economia brasileira contemporânea. 7th ed. São Paulo: Atlas. Grembi, Veronica, and Alvaro Manoel. 2011. “Fiscal Rules for Subnational Govern- ments? Evidence from Latin America.” World Bank, Washington, DC. Hermann, Jennifer. 2005. “Auge e declínio do modelo de crescimento e endividamento: O II PND e a crise externa (1974/1984).” In Economia Brasileira Contemporânea 78 Until Debt Do Us Part (1945–2004), ed. Fábio Giambiagi, André Villela, Lavinia Barros de Castro, and Jennifer Herrman, 93–115. Rio de Janeiro: Editora Elsevier. IMF (International Monetary Fund). 2003. “The IMF and Recent Capital Account Crises: Indonesia, Korea, Brazil.” Evaluation Report, Washington, DC. http:// www.imf.org/external/np/ieo/2003/cac/pdf/all.pdf. ———. 2011a. “IMF Executive Board Concludes 2011 Article IV Consultation with Brazil.” Public Information Notice 11/108, Washington, DC. http://www.imf .org/external/np/sec/pn/2011/pn11108.htm. ———. 2011b. World Economic Outlook: Slowing Growth, Rising Risks. Washington, DC. http://www.imf.org/external/pubs/ft/weo/2011/02. Leite, Cristiane Kerches da Silva. 2011. “Federalismo, processo decisório e orde- namento fiscal: A criação da Lei de Responsabilidade Fiscal.” Texto para Discussão 1593, Instituto de Pesquisa Econômica Aplicada, Brasília. ­ Levy, Joaquim V. F. 2009. “Stability and Innovation in Brazilian Subnational Debt.” Paper presentation at the Subnational Fiscal Reform and Debt Management Forum, World Bank, Washington, DC, April 29. Liu, Lili, and Steven Webb. 2011. “Laws for Fiscal Responsibility for Subnational Discipline: International Experience.” Policy Research Working Paper 5587, World Bank, Washington, DC, March. Lopes, Francisco. 1986. O choque heterodoxo: Combate à inflação e reforma mone- tária. Rio de Janeiro: Campus. Lopreato, Francisco Luiz C. 2000. “O endividamento dos governos estaduais nos anos 90.” Economia e Sociedade 15 (dezembro): 117–58. Modenesi, Andre de Melo. 2005. Regimes monetários: Teoria e experiência do Real. Barueri: Manole. Mora, Monica. 1997. “Processo de endividamento dos estados: Problemas e limites à descentralização e à autonomia.” Dissertação de Mestrado, Universidade Fed- eral do Rio de Janeiro, Rio de Janeiro. ———. 2002. “Federalismo e dívida estadual no Brasil.” Texto para Discussão 866, Instituto de Pesquisa Econômica Aplicada, Brasília. Pellegrini, Josué A. 2012. “Dívida Estadual.” Textos Para Discussão 110, Núcleo de Estudos e Pesquisas do Senado, Brasília. Piancastelli, Marcelo, and Rogério Boueri. 2008. “Dívida dos Estados 10 Anos Depois.” Texto Para Discussão 1366, Instituto de Pesquisas Econômicas Apli- cada, Rio de Janeiro. Prefeitura da Cidade de São Paulo. 2011. Dívida com a União: MP 2185-35. Secre- taria de Finanças, Report of December 31, 2011, São Paulo. Prefeitura Municipal do Rio de Janeiro. 1998. “Proposta de reestruturação das dívi- das municipais.” Rio de Janeiro. ———. 2011. “Emissão de Títulos.” Texto para Discussão, Rio de Janeiro. Rezende, Fernando, and José R. Afonso. 2002. “The Brazilian Federation: Facts, Challenges and Prospects.” Working Paper 149, Center for Research on Eco- nomic Development and Policy Reform, Stanford University, Stanford, CA. Brazil: The Subnational Debt Restructuring of the 1990s—Origins, Conditions, and Results 79 Rigolon, Francisco, and Fábio Giambiagi. 1999. “A Renegociação das Dívidas e o Regime Fiscal dos Estados.” In A Economia Brasiliera nos anos 90, ed. F. Giam- biagi, and M. M. Moreira, 111–44. Rio de Janeiro: National Bank of Economic and Social Development. Serra, José, and José R. Afonso. 2007. “Fiscal Federalism in Brazil: An Overview.” CEPAL Review 91, Santiago, Chile. Tavares, Martus, Alvaro Manoel, and José Roberto R. Afonso. 1999. “Uma proposta para um novo regime fiscal no Brasil: O da responsabilidade fiscal.” Compen- dio de Documentos, XI Seminário Regional de Política Fiscal, CEPAL, Brasília. Varsano, Ricardo. 1996. “A evolução do sistema tributário ao longo do século: Anotações e reflexões para futuras reformas.” Texto para Discussão 405, Insti- tuto de Pesquisa Econômica Aplicada, Brasília. Villela, Renato, and Fernando Rezende. 1986. “Alterações na política financeira: Desenvolvimento urbano e problemas metropolitanos; Aspectos financeiros.” In Alternativas político-institucionais para a região metropolitana de Salvador: Dimensão nacional, ed. Ana Maria Brasileiro, 212–34. Rio de Janeiro: Convênio IBAM/CONDER. 2 Restructuring of Legacy Debt for Financing Rural Schools in China Lili Liu and Baoyun Qiao Introduction China started to promote nine-year compulsory education in rural areas throughout the country in the mid-1980s. To achieve that goal, the Compulsory Education Law, enacted in 1986, mandated that the cen- tral and local governments at all levels guarantee the operational and capital expenditures to implement compulsory education. In practice, the local governments, especially the county-level governments, played a main role in financing rural compulsory education. With limited fis- cal resources, local governments financed rural compulsory education funds through a multitude of channels. In particular, to achieve the national goal of accomplishing universal nine-year compulsory educa- tion by 2000, local governments (towns and villages) resorted to bor- rowing to finance school facilitates to help meet the minimum facility standards for all schools, although the borrowing was not allowed by the 1994 Budget Law of China. In 2000, nine-year compulsory education became universal in China, a historic accomplishment. However, the accumulation of debt that resulted from borrowing to finance rural compulsory education had become a significant fiscal burden on local governments. The aggregate 81 82 Until Debt Do Us Part rural education debt was about RMB 110 billion1 at the end of 2007, about 3.9 percent of aggregate subnational own fiscal revenue, exclud- ing transfers. Of the outstanding compulsory education debt, about RMB 80 billion was used to finance capital expenditure, and about RMB 30 billion was used to finance operational expenditures. The negative impact of rural compulsory education debt on the rural economy and social development had become serious. The rural Tax-for-Fee Reform in 20012 was intended to reduce financial burdens (such as taxes, sur- charges, and fees) on rural households, but it also further limited local fiscal resources. At the same time, the rural reforms increased the demand for fiscal resources to finance rural infrastructure and social services. Restructuring the legacy debt for financing rural schools thus became a priority of the central government in the mid-2000s. With the pub- lic policy goal in the late 1990s of inclusive economic growth, the debt financing of nine-year compulsory education in rural areas would be replaced by grant financing for all children. It was necessary to resolve the widespread indebtedness of rural local governments to make the debt-to-grant financing transition, and the debt-to-grant financing would need to address the legacy debt and its write-offs. Resolving the legacy debt and making a transition to grant finance should be viewed as part of building a sustainable fiscal system frame- work to ensure the delivery of basic public services, including imple- mentation of the Compulsory Education Law. With less fiscal capacity as a result of the rural Tax-for-Fee Reform, the county-level govern- ments could choose new approaches to increase the fiscal burden of rural residents. China’s central government, through the Ministry of Finance, initi- ated a program of restructuring the rural legacy school debt in 2007. This was a pioneering effort—the first time the central government undertook debt restructuring of subnational governments nationwide. A key design issue in any debt restructuring is how to avoid moral haz- ard. International experience shows that unconditional write-offs create soft budget constraints. To address the problem of moral hazard, the Ministry of Finance’s plan had two distinct features. First, the fiscal resources to finance the debt write-off were distrib- uted equally among the central government, provincial governments, Restructuring of Legacy Debt for Financing Rural Schools in China 83 and subprovincial governments; that is, the central government, pro- vincial governments, and subprovincial governments each contributed one-third of the financial resources. Second, the distribution of central government grants was based on a formula that took into account the required expenditure to achieve basic provision of education and the local government fiscal capacity to deliver the results. Thus, the distribution of central grants to a par- ticular local jurisdiction was not directly related to the size of the debt of that jurisdiction. This output-based approach was meant to prevent perverse incentives for local governments to increase the size of their debt or to reduce their service of debt in anticipation of more grants or bailouts. This chapter analyzes the moral hazard aspect of the rural debt restructuring within China’s current intergovernmental fiscal frame- work. The remainder of the chapter is organized as follows. Section two summarizes the compulsory educational system and what responsibili- ties are assigned to each government level in China. Section three exam- ines the features of rural education legacy debt. Section four focuses on China’s strategy to deal with moral hazard. Section five explains how China restructured the legacy debt for financing rural schools. Section six offers conclusions. The Responsibility and Revenue Assignment for Education in China China has a unitary system with five levels of government. The cen- tral government is at the top under which are four tiers of subnational government: provincial, prefecture (city), county, and township.3 The central government determines the establishment and the geographic division of the provincial, prefecture, and county governments, and the provincial government determines the establishment and the geo- graphic division of the township governments. The provincial government, directly under the central govern- ment, consists of the governments of the provinces, autonomous regions, and large municipalities with provincial status. A p ­ refecture or a prefecture-level city under the jurisdiction of a province is an administrative division between a province and a county. At the end ­ 84 Until Debt Do Us Part of 2009, China had 333 prefectures or prefecture-level cities. A county or county-level city is the general administrative division at the local level. At the end of 2009, there were 2,858 counties or county-level cities in China. The township is the basic administrative division in the coun- tryside. At the end of 2009, there were 40,858 townships in China. According to the Chinese Constitution, the People’s Congress and people’s government at the levels of province, prefecture (or city), county, and township are the local legislative and executive organs of power, respectively. Law on Nine-Year Compulsory Education The Law on Nine-Year Compulsory Education, enacted by the central government on July 1, 1986, mandated requirements and deadlines for attaining universal education (to be tailored to local conditions) and guaranteeing school-age children the right to receive at least nine years of education (six years of primary education and three years of secondary education). People’s Congresses at various local levels were, within certain guidelines and according to local conditions, to choose the steps, methods, and deadlines for implementing nine-year com- pulsory education in accordance with the guidelines formulated by the central authorities. The program sought to bring rural areas, which had four to six years of compulsory schooling, into line with their urban counterparts. The Decision on the Reform of the Education System by the Central Committee of the Party in 1985 divided jurisdictions into three catego- ries: (a) cities and economically developed areas in coastal provinces and a small number of developed areas in the hinterland (approximately 25 percent of China’s population); (b) towns and villages with an aver- age level of development (approximately 50 percent of China’s popula- tion); and (c) economically backward areas (approximately 25 percent of ­China’s population). In fact, some jurisdictions in the first category had achieved univer- sal nine-year education by 1985, and all jurisdictions in the category had achieved universal nine-year education by 1990. The jurisdictions in the second category had achieved universal nine-year education by 1995, and technical and higher education were projected to develop at the same rate. The jurisdictions in the third category should attain the Restructuring of Legacy Debt for Financing Rural Schools in China 85 basic education target for all students with a timetable consistent with the pace of local economic development, and the central government would support the educational development. The central government also would assist educational development in minority nationality and remote areas. Currently, there are over 260,000 rural primary and junior secondary schools with about 76 million students (see table 2.1). Distribution of Educational Responsibilities among Levels of Government Fiscal decentralization reforms in China after the economic reform started at the beginning of the 1980s provided local governments with significant fiscal autonomy in various areas such as the determination of their own spending priorities and policies on relevant aspects of local budgets. In 1994, the central government introduced the Tax Sharing Sys- tem (TSS) reform, with the two major goals of increasing the share of combined government revenue in total gross domestic product and the central share in combined government revenue (World Bank 2002). This reform introduced clear and stable assignments of tax rev- enues between the central and provincial governments, and created Table 2.1  Statistics on Rural Junior Secondary and Primary Schools and Students, 2009 Nine-year Junior primary- Total Number of secondary secondary enrollment Item schools (unit) schools schools (person) Rural junior secondary education 30,178 22,921 7,257 19,345,061   Run by education departments   and collectives 28,590 22,240 6,350 18,498,085   Run by private institutions 1,222 557 665 646,478   Run by other departments 366 124 242 200,498 Rural primary education 234,157 56,555,439   Run by education departments   and collectives 231,360 54,941,004   Run by private institutions 2,395 1,260,235   Run by other departments 402 354,200 Source: National Bureau of Statistics 2010. 86 Until Debt Do Us Part separate tax administration services at both levels of government. The TSS reform introduced the value-added tax as the major government revenue source, and established a uniform tax-sharing system that provided that the central government would receive 75 percent of the value-added tax and subnational governments 25 percent. The taxes assigned to the center and the major taxes shared with the subnational governments were collected by the newly created National Tax Services, which operated in all provinces. The new system pro- vided for separate local (subnational) tax services for the collection of the taxes assigned to local governments. The headquarters of the local tax services, the State Administration of Taxation, was empowered to supervise local tax services and prohibit the use of tax exemptions by local governments. Given the main focus of the TSS reform on the taxation side, how- ever, there was no apparent change in either policy or practice in terms of expenditure assignment between the central government and subna- tional government and among the four tiers of subprovincial govern- ments. In fact, the TSS restated the prereform expenditure assignment and provided only basic guidelines to define expenditure responsibili- ties between the central and subnational governments. For example, the State Council Regulations on the Implementation of the TSS in 1994 defined expenditure responsibilities of central and subnational govern- ments as follows: Central budgets are mainly responsible for national security, interna- tional affairs, the costs of operating the central government, the needs for adjusting the structure of the national economy, coordinating regional development, adjusting and controlling the macro economy, and others. Detail items include: national defense, cost of military police, international affairs and foreign aid, administration costs of the central government, centrally financed capital investments, the tech- nical renovation of central-government-owned enterprises and new product development costs, the expenditure to support agriculture, arts and culture, education and health, price subsidies and others. Subnational budgets are mainly responsible for the costs of running subnational governments, and the need for local social economic devel- opment. Detail items include: the costs of running subnational govern- ment, the needs of local economic development, a part of the costs of Restructuring of Legacy Debt for Financing Rural Schools in China 87 running the military police and militia, locally financed capital invest- ments, the technical renovation of local government-owned enterprises and new product development, the costs of support to agriculture, urban maintenance and construction, and the expenditure to support arts and culture, education and health, price subsidies and others. These guidelines illustrate that both the central government and subnational governments not only have extensive expenditure respon- sibilities, but also that these responsibilities overlap and are not spe- cific. Adding to the lack of clarity in expenditure assignments is the fact that the subprovincial governments do not have explicit expenditure assignments. The expenditure assignments for subprovincial govern- ­ ments are basically at the discretion of their provincial governments. To improve expenditure at the subprovincial government level, the M ­ inistry of Finance in December 2002 issued “Suggestions on I ­mproving Sub-­ provincial Fiscal Relations,” to provide further g­ uidelines on subprovin- cial expenditure assignment. However, considerable challenges remain in clarifying expenditure assignment. Education, for example, is mainly the responsibility of subnational governments. For compulsory education, the role of the central govern- ment is that of policy maker and overall planner. In addition, the cen- tral government has the responsibility for establishing special education funds for subsidizing compulsory education in poor, minority areas, and teacher education in all areas. The provincial government has the overall responsibility for formulating the development plan for compul- sory education and providing assistance to counties to help them meet recurrent education expenditures. The responsibility for actually imple- menting compulsory education programs lies with the cities or districts of large cities in the case of urban areas, and with counties in the case of rural areas. The provision of compulsory education services in rural areas is one of the major concerns of the central government because of the gener- ally worse service conditions, especially in poor rural areas. Some initia- tives, especially the “Decision on Strengthening Rural Education,” issued by the State Council in September 2003, expanded the expenditure responsibilities of the central government on compulsory education. This basic service was defined as a shared responsibility with the goal of supporting students from poor families by waiving their payments for 88 Until Debt Do Us Part textbooks, tuition, and miscellaneous fees, and by subsidizing housing expenditures for elementary and secondary education students. In general, several important decision-making powers were decen- tralized to county governments in implementing the State Council 2003 directive. For example, the county governments were able to close schools or reduce their size in their jurisdictions. This decentralized power eventually influenced the education sector in significant ways. In addition, upper-level governments (the central and provincial govern- ments) became more involved in financing education. There were two driving forces behind the changes in the assign- ment of responsibilities for upper-level governments. First, as discussed below, the reduction in revenues that resulted from the Tax-for-Fee Reform rendered some local governments fiscally unable to finance education (Xiang and Yuan 2008), so more transfer funds were required to keep education services running smoothly. Second, reducing dispari- ties in education expenditure was deemed to be a desirable public policy goal. Therefore, more transfer funds should go to poorer areas. County governments have become the most important—but not the only—players in the current regime of education finance. Other levels of government, such as the prefecture, province, and central govern- ment, also have roles to play; even township governments and village self-governing bodies have some responsibilities in particular areas (Huang 2006). The power to make education policy is still centralized in the hands of provincial and central governments. Nevertheless, county governments have been given considerable latitude for making decisions on the daily operations of educational services, as shown in table 2.2. Revenue Assignment of Rural Compulsory Education Although there are various financing channels, compulsory education in general is mainly financed by the government, particularly by the government’s budgetary expenditure, as shown in table 2.3. The finance system of compulsory education in China has evolved since the early 1950s, in three major stages. The first stage was 1950–93. During that period, primary and second- ary education services were mainly provided by subprovincial govern- ments. The central government provided financial support, mostly upon Restructuring of Legacy Debt for Financing Rural Schools in China 89 Table 2.2  Assignment of Major Educational Responsibilities among Levels of Government in China Central Provincial Prefecture Terms government governments governments County governments Policy Establishing     1.      1.  Making policy Coordinating     1.  Planning local     1.  the for education educational school system educational development planning structure system 2.  Planning 2.  Implementating 2.  Paying teachers 2.  Curriculum, for the educational 3.  Managing approval of reorganization examination principals and textbooks of primary and evaluation teachers 3.  Determining and secondary 4.  Guidancing teacher-pupil schools teaching ratios 3.  Examining and activities evaluating 5.  Evaluating rural schools schools 4.  Approving size 6.  Proposing of teaching teacher-pupil staffs ratios Financing Transferring Providing support None Purchasing funds to to poor equipment and help poor counties books and minority areas Teacher Earmarking Using transfer Providing transfer Paying teachers salaries transfer funds funds to funds to poor for teacher support poor areas salaries counties Operational None Determining     1.  Helping county Providing funds costs teacher- governments for operation of student finance schools ratios and operational corresponding costs operational expenditures for rural areas 2.  Transferring funds to help poor areas Construction Earmarking Providing Helping county Implementing plans costs subsidies for subsidies government for building the repair of to county finance building schools dangerous governments to of school classrooms in finance school facilities poor areas facilities Source: Compiled by the authors. 90 Until Debt Do Us Part Table 2.3  Financing Sources for Rural Junior Secondary and Primary Schools in China, 2008 RMB millions Rural junior secondary Rural primary Financing sources schools Percent schools Percent Total Percent Government 135,722.3 95.03 222,362.4 96.76 358,084.7 96.09   Budgetary 128,372.0 89.88 213,637.4 92.96 342,009.4 91.78 Private school 210.4 0.15 210.7 0.09 421.1 0.11 Donations 823.0 0.58 1,233.2 0.54 2,056.2 0.55 School charge 4,412.2 3.09 4,058.6 1.77 8,470.8 2.27   Tuition 1,173.4 0.82 1,073.3 0.47 2,246.8 0.60 Other 1,656.0 1.16 1,954.2 0.85 3,610.2 0.97 Total 142,823.9 100.00 229,819.1 100.00 372,643.1 100.00 Source: National Bureau of Statistics 2010. request of the provincial governments. Education services were treated as local public goods, and education expenditure was financed mainly by budgetary funds, education surcharges, donations to education, and student fees. Minban teaching staff (nongovernmental employees) were commonly used in rural schools as instructors, and school facilities were financed mostly by local residents. The second stage was 1994–2003. During that period, the education sector continued to decentralize. The 1994 TSS reform had an impor- tant impact on education finance. As noted, the TSS reform had two major goals: increasing the share of combined government revenue in total gross domestic product, and increasing the central share in com- bined government revenue. Given that larger shares of fiscal resources went to the upper-level governments, especially the central govern- ment, and that no changes were made in the assignment of expenditure responsibilities, local governments, especially township governments, began to experience increasing difficulties in providing educational ser- vices. During this period, off-budgetary resources, in particular school fees and charges, played a bigger role in financing education than in the first stage. There were other policy shocks that added to the fiscal pressures fac- ing local governments. For example, during 1996–97, many Minban teachers became government employees, in accordance with explicit Restructuring of Legacy Debt for Financing Rural Schools in China 91 policies set by the State Council. Local governments were already strug- gling financially, and the accompanying increases in salary payments added to the burden. In addition, the Nine-Year Compulsory Education plan required all schools in rural areas to comply with regulations con- cerning school facilities. To comply with these regulations, local govern- ments incurred large increases in construction expenditures. As shown in table 2.4, with the expanding expenditure per student in rural primary schools in Henan province, salary payments increased significantly, and construction was more concentrated in 1999 to meet the requirement that schools implement the Nine-Year Compulsory Education plan. Meanwhile, a 1999 survey by the Ministry of Finance revealed that the minimum financial gap for the local government to implement the Nine-Year Compulsory Education plan was RMB 35 billion (Hu 2002). In short, during this period, the increases in expenditures and con- tracted budgetary revenues rendered some local governments unable to provide enough resources to fulfill their educational responsibilities. Many local governments suffered significant shortages in education funds, with the following consequences: (a) teachers in many provinces could not get paid on time—in fact, the total amount of unpaid teachers’ salaries ­ (Beijing, Shanghai, Tianjin, Zhejiang, and Tibet were not included) was RMB 13.6 billion (Liao 2004); (b) students were charged high fees in order to finance the regular operation of schools; and (c) local residents were heavily taxed to finance the new school facilities (see, for example, Li 2006). Table 2.4  Composition of Educational Expenditures of Rural Primary Schools, Province of Henan, 1999 and 2002 RMB per student Year 1999 2002 Change Education expenditure 364.45 545.76 181.31 Teacher salaries 224.42 450.23 225.81 As share of total expenditure 61% 82% 21% Operational expenditure 97.16 81.59 –15.57 As a share of total expenditure 27% 15% –12% Construction expenditure 42.87 13.94 –28.93 As a share of total expenditure 12% 3% –9% Source: Henan Education Department 2003. Note: Bolded text indicates expenditures. 92 Until Debt Do Us Part The third stage began in 2003 and continues to the present. To reduce the tremendous financial burdens facing rural households, the central government in 2001 initiated the Tax-for-Fee Reform. The reform can- celed various charges and fees levied on rural households by rural local governments, and improved rural tax structure and administration.4 While reducing the financial burdens on rural households, the reform also eliminated the revenue base for township governments, since their major revenue bases, in particular five agriculture-related taxes and budgetary funds, were no longer available. The central government also reassigned the responsibility for basic education services to the county- level government, and increased the central contribution for financing primary and secondary education.5 However, the newly assigned responsibility to the county govern- ments was not accompanied by revenues, since the agriculture-related taxes and other fees were also main sources of the county government budgets. In response to these budgetary constraints, county govern- ments first closed a number of primary and secondary schools, then started to charge higher student fees to finance operational costs for the schools that remained open. Under the new arrangement, teacher salaries were increased and were more likely to be paid on time, while the operational costs were financed jointly by budgetary funds and stu- dent fees. The financial arrangements for construction costs, however, remained unsettled (Liao 2004). The unevenness between revenue availability and expenditure responsibilities since the 1994 TSS reform created serious challenges in service provision at the local level (Jia 2008). Other central government policies also contributed to the financial challenge faced by local gov- ernments. In particular, implementation of the Nine-Year Compulsory Education plan required that all rural schools satisfy certain conditions, including good physical facilities and a qualified teaching staff.6 While this policy led to significant increases in local expenditure needs for education and construction costs, in particular, the central government made no new provision of funds for local governments (Zong 2010). In response to the imbalances in their public finances, subprovincial governments developed two strategies. On the expenditure side, both the quality and quantity of some pub- lic goods provided by local governments decreased. For example, in Restructuring of Legacy Debt for Financing Rural Schools in China 93 some areas, local governments failed to pay teachers on time. In other areas, teachers were perhaps relatively luckier; they could receive their paycheck on time, but with a condition, say, that their salaries would be reduced by 20 percent. Failure to pay teachers was not an isolated occur- rence; rather, it reportedly took place across the country, except in some rich areas like Beijing and Shanghai (Zhang 2005). On the revenue side, it appears to have been difficult for rural local governments to increase revenue collection on their own, since the tax bases and tax rates are largely controlled by the central government. But local governments were able to collect additional funds to finance education and other local public goods through a variety of fees and charges, formally and informally. Given that no horizontal accountabil- ity mechanisms were in place, in the face of financial stress it can be rea- sonably expected that local governments would have increasingly and consistently exerted their powers to obtain money from rural house- holds (Liao 2004). Serious pressures from this situation redounded to the central gov- ernment in a variety of ways. To protest unpaid and delayed salaries, teachers first loudly voiced their concerns through their representatives in the People’s Congress at different tiers of government. In addition, more teachers quit their jobs in rural schools or moved to schools in richer areas for higher and more stable salaries (Cai 2002; Wang 2004: Zhou, Liu, and Tian 2003). As a consequence, an adverse selection prob- lem for teachers developed across rural schools; those teachers taking jobs at or remaining in rural schools were in many cases not the teachers those schools needed. The loss of qualified teachers reduced the quality of educational services provided in rural areas and raised serious con- cerns among parents about the quality of education. These concerns eventually filtered up to the central government. In particular, restruc- turing of legacy debt for financing rural schools needed the immediate attention of the central government. Fundamentally, county governments assumed the key role in bud- geting education finance following the 2003 State Council directive. County governments pool all revenues from own and transferred funds, and decide on their use to finance education and to distribute funds to all schools in all three categories in their jurisdictions. The roles played by upper-level governments are mostly supportive, although in some 94 Until Debt Do Us Part instances they contribute a significant share to financing general edu- cation expenditure. Higher-level governments pay special attention to teacher salaries by earmarking funds, but do little in relation to other categories of expenditures.7 Significant Features of Rural Compulsory Education Debt In general, there is a mismatch between responsibility and revenue assignment in China; in particular, the aggregated own revenue of county and lower governments amounts to only 42 percent of their expenditure, as shown in table 2.5. Although the central government required subnational governments to provide rural compulsory educa- tion, the county governments, which are responsible for its implemen- tation, and other tiers of subnational government, had limited resources to do so, which forced them to borrow funds to finance the services (Xiang and Yuan 2008). However, as mentioned, borrowing is not per- mitted under the Budget Law of China, and this further complicated the debt issue and encouraged county governments to borrow off budget. The result was that not only was debt concentrated in the county governments, but there was also asymmetric information between the county governments and the upper levels of government, especially the central government, about the size of debt and its service cost. In addition, the debtor-creditor relationship was informal, in general, because the borrowing governments did not have the legal status to borrow (Shi 2004). More important, most county governments were not able to pay the debt by relying on own resources. These factors increased the difficulty of solving the problem of the rural education debt (Wang 2007). Table 2.5  Own Revenues as a Percentage of Total Expenditures, by Level of Government in China, 2003 Level of government (consolidated) Average Minimum Maximum Provincial 53 8 75 Prefecture 55 2 85 County and lower 42 0 90 Source: Ministry of Finance. Restructuring of Legacy Debt for Financing Rural Schools in China 95 Table 2.6 presents the share of compulsory debt in subnational bud- getary expenditures for 14 provinces at end-2005. Since most of the debt is concentrated in county-level governments,8 the share of debt in county-level expenditures would be much higher. The debt mainly con- sisted of arrears on funds owed to construction companies that built the schools and to suppliers, and on funds for teacher wages and pensions. A small part of the debt was incurred by local government off-budget vehicles to construct schools, that is, funds borrowed from financial institutions. Restructuring the rural compulsory education debt became a prior- ity of the central government in mid-2000. The main reasons include (a) it was an important step toward improving rural human capital and achieving equality between urban and rural education, (b) writing off the rural compulsory debt was the prerequisite to building a sustainable finance system for rural compulsory education, and (c) it was needed Table 2.6  Compulsory Debt as a Percentage of Subnational Budgetary Expenditure for 14 Regions in China, 2009 RMB billions Compulsory education debt Provinces (by end-2005) Budgetary expenditure Percent Inner Mongolia 3.92 192.684 2.03 Jilin 2.31 147.921 1.56 Helongjiang 2.434 187.774 1.30 Jiangsu 5.75 401.736 1.43 Anhui 3.079 214.192 1.44 Fujian 2.312 141.182 1.64 Jiangxi 4.094 156.237 2.62 Hubei 3.074 209.092 1.47 Hunan 4.803 221.044 2.17 Sichuan 9.092 359.072 2.53 Ningxia 0.901 43.236 2.08 Guizhou 3.994 137.227 2.91 Shan’xi 3.065 184.164 1.66 Guansu 2.333 124.628 1.87 Total 51.161 2,720.189 1.88 Source: Ministry of Finance. 96 Until Debt Do Us Part to prevent increasing the fiscal burden on rural residents. In addition, writing off rural compulsory education debt was relatively easy in terms of size and complexity, and it could provide experience and knowledge for restructuring other subnational debt (Zhang 2007). Although the size of the aggregate debt is relatively small, in the view of policy makers, how the debt was restructured would influence the future behavior of subnational governments. In particular, improper incentives could lead to moral hazard for local governments, which could lead to more problems in the future when trying to solve debt problems. Key Strategy for Dealing with Moral Hazard in Debt Restructuring9 The RMB 110 billion rural compulsory education debt accounted for an insignificant portion of the RMB 10.7 trillion in subnational liabilities at the end of 2010.10 However, the restructuring of the rural education debt represented the first effort to restructure subnational debt; thus, its design and approach would affect subsequent debt restructuring efforts. In particular, an improperly designed framework could create negative incentives for subnational governments concerning their future bor- rowing decisions. A key issue in writing off the rural compulsory education debt con- cerns moral hazard. The soft budget constraint challenge exists in the fiscal system of many countries (see, for example, Liu and Webb 2011). An improperly designed fiscal system encourages moral hazard on the part of local government and creates the soft budget constraint. Not surprisingly, the soft budget constraint issue has also been a challenge ­ for China’s fiscal system. To some extent, the compulsory ­education debt resulted from the soft budget constraint and the weakness of China’s­ intergovernmental fiscal relations. Consequently, how to deal with a potential moral hazard challenge was a serious concern to policy makers in designing the restructuring package of the compulsory education debt. The restructuring of this debt served as a pilot from which to draw lessons. Finally, restructur- ing the rural compulsory education debt in China took place within the framework of intergovernmental fiscal relations. The process of the debt Restructuring of Legacy Debt for Financing Rural Schools in China 97 restructuring would help clarify intergovernmental fiscal relations in delivering basic education and help inform future reforms to the inter- governmental fiscal system. There might have been other options to resolve the rural legacy debt. One would have the central government write off the entire debt. This option was not chosen because it would have encouraged moral hazard. Another option would have allowed provincial governments to resolve the debt of their local governments. This option was not chosen because the central government viewed the rural debt restructuring as an oppor- tunity to realign intergovernmental fiscal relations with respect to rural education; letting provincial governments resolve the problem was not feasible given the existing intergovernmental fiscal relations. The cho- sen mechanism was based on the principles of burden sharing, trans- parency, and formula-based restructuring. As mentioned, the rural debt is only a small portion of subnational government debt, and there are broader issues concerning moral hazard and opportunistic behavior of subnational governments. The experience of rural debt restructuring can offer lessons on addressing these broader issues. In restructuring the compulsory education debt, there were two types of moral hazard. The first relates to the overall borrowing size of local governments. As the debtors, the county governments might have an incentive to expand the size of total rural compulsory education debt in anticipation of seeking more central government grants to replace the incurred debt. The second type of moral hazard concerns the behavior of county governments in anticipation of a write-off. Instead of using their own revenues to contribute to the write-off of debt, the county governments might have an incentive to seek more bailout grants to write off the existing debt. To deal with these two types of moral hazard, the central government designed a strategy with two distinct features. First, the fiscal resources required for debt write-off and restructur- ing are distributed among three tiers of government: roughly one-third from the central government, one-third from provincial governments, and one-third from lower-tier governments. Thus, the fiscal burden of debt restructuring is shared. Second, the distribution of the central grants was based on an output-based rather than an input-based formula. The output includes ­ 98 Until Debt Do Us Part two set of factors. The first set formed the base on which to calculate the total grants to a jurisdiction, and the second set formed the base to determine the performance of a local government in writing off the rural compulsory education debt. Regarding the first set of factors, although the overall size of the cen- tral government grant was determined by the overall size of the com- pulsory education debt,11 the grant going to each individual j ­urisdiction was not directly linked to the size of the debt of that jurisdiction. The total grants were only a pool that provided the grants’ source; the link- age between the pool and its distribution among local governments went through the following two steps. First, as mentioned, the grants would not be directly related to the total debt of a jurisdiction; that is, the central government grants would not be tied to actual indebtedness of county or town governments. Second, the distribution of central government grants to local gov- ernments for write-offs was determined by a formula that considered four factors within local government jurisdiction: (a) number of stu- dents, (b) number of schools, (c) population density, and (d) local fiscal capacity. By this method, a local government that borrowed excessively would not gain extra advantage, and another local government that borrowed less or paid off its debt would not be in an unfavorable position. By choosing objective factors and giving full consideration to the financial difficulties of all pilot areas, the grants to all pilot areas would be cal- culated uniformly according to the formula.12 The areas having more financial difficulty would enjoy a higher proportion of subsidy. Coun- ties and towns that had not incurred debt but needed to cover a cer- tain number of schools and students could obtain funds as a positive incentive. In addition, there was a penalty rule to prohibit new debt for compulsory education. The grants to a local government would be reduced by the central government if new rural compulsory education debt emerged. Provincial government grants were required by the cen- tral government to follow the same principles. The second set of factors formed the base to determine the perfor- mance of a local government in writing off the rural compulsory educa- tion debt. To encourage subnational government effort in writing off debt, the central government established the incentive by providing Restructuring of Legacy Debt for Financing Rural Schools in China 99 a subsidy for those that had made efforts to write off the ­ compulsory ­ education debt, and did not provide a subsidy to those that had not completed the work within the stipulated time limit. That is, central government grants to a province were based on performance, and only after the debt write-offs of lower-level governments could a province receive funds from the central government.13 The requirements, for- mula, and matching methods for writing off debt were included to pre- clude possible rent-seeking behavior. All pilot provinces and their local governments were encouraged to mobilize revenues to contribute their share for writing off compulsory education debt. Subnational governments were encouraged to raise revenues through, for example, improving collection efficiency of local budgeted revenue, expenditure efficiencies, revitalizing idle school facil- ities, and mobilizing donations. Restructuring of Legacy Debt for Financing Rural Schools, in Practice Based on the above strategy, the central government launched the proj- ect to restructure the rural compulsory education debt. In December 2007, the General Office of the State Council transmitted the Advice Notice on Pilots Working on Resolving the Debt for Rural Nine-Year Compulsory Education, which was prepared by the Working Group under the State Council on the comprehensive reform in rural areas.14 The preparation work included the classification and audit of existing subnational rural school legacy debt. Implementation of the project followed the existing framework of intergovernmental fiscal relations. Under this framework, rural compulsory education was planned by provincial governments and implemented by their county governments. Thus, the pilot project for writing off rural compulsory debt was organized by provincial governments and implemented by county governments. According ­ to the State Council’s Advice Notice, all pilot provinces must have refrained from incurring new debt to finance compulsory education. All subnational governments were required to adjust their financial expenditure structure to establish reliable revenue sources for servic- ing new debt. 100 Until Debt Do Us Part Of the total rural school debt of RMB 110 billion outstanding at the end of 2007, RMB 80 billon had been borrowed to finance capital investments (for school construction). An additional RMB 30 billon was used to finance operational deficits. The implementation strategy was designed to write off RMB 80 billion of debt for financing school construction, with two steps. The first step was to write off the compulsory education debt of 14 provinces15 within two years (that is, by the end of 2009). During this step, all non-pilot-project provinces would choose two or three counties (cities, districts) to pilot the writing off of such debt. After progress was achieved in the 14 provinces and the pilot cities and counties, the sec- ond step was to extend the restructuring exercise to all other provinces, and to complete the write-off of the debt in all non-pilot provinces by the end of 2010. While the originally planned central government contribution toward writing off the RMB 80 billion compulsory education debt had been projected to be RMB 26.67 billion, the final central government contribution was RMB 30 billion, slightly higher than the originally planned one-third, and the subnational government contribution was RMB 50 billion. By the end of 2009, the central government provided RMB 14.5 billion in grants, about 90 percent of the total funds, for writ- ing off the debt of 14 pilot provinces and Chongqing, consistent with the schedule of debt write-off. In 2009, the central government launched a project of compulsory debt write-off in non-pilot areas. By the end of 2009, RMB 3.58 billion in grants had been provided to the non-pilot provinces, and in 2010, the remaining RMB 11 billion in grants was transferred to a central govern- ment account for distribution to the non-pilot provinces. By the end of 2011, the funds were almost completely disbursed. Given the success of writing off rural compulsory education debt relating to capital financing in the pilot areas, the central government in 2009 initiated a new program to write off debt that was used to finance operational deficits in the pilot areas. The same output-based formula was used. The total pool of the central grants for writing off operational debt for compulsory education was 37.5 percent of total operational debt related to rural compulsory education, the same percentage as that for the capital debt. Restructuring of Legacy Debt for Financing Rural Schools in China 101 The first 14 pilot provinces completed the task of writing off debt in 2009, which benefited about 1.7 million rural creditors. The 17 non- pilot provinces actively prepared for the write-off and started writing off debt in 2009. By the end of 2009, 31 provincial governments had paid RMB 56.6 billion of the rural compulsory education debt, among which the central government provided RMB 14.5 billion in grants, and subnational governments, including provincial governments, financed RMB 42.1 billion. By the end of 2011, almost all rural compulsory edu- cation debt had been restructured, including a RMB 30 billion opera- tional deficit. Since 2007, the central government has contributed RMB 30 billion toward writing off rural compulsory education debt, or 37.5 percent of the debt. To ensure implementation of the overall strategy, the central govern- ment has established a management system to supervise implemen- tation. In principle, all central government grants were designed to contribute to the write-off of the compulsory education debt. Mean- ­ while, for an individual local government, the remaining grants from the central government after writing off compulsory education debt could be arranged to write off other local non-education-related debt, with priority given to the rural-education-related debt. Subnational governments are required to establish a compulsory edu- cation debt control system to monitor grants from the provincial and central governments and the progress of writing off debt to make sure the grants are used effectively and follow central government require- ments. In addition, the departments of finance of provincial, municipal, and county governments are required to report the use of the grants in their budget and final financial reports to the relevant People’s Congress or Standing Committee. The provincial finance department must report the progress of the write-off of compulsory education debt and the usage of grants to the central government on a monthly base. The pen- alty may be imposed rule if new debt were contracted. In addition, the Ministry of Finance retained RMB 2 billion to deal with contingencies. It is too early to assess the impact of writing off the rural educa- tion debt, since the entire exercise was completed only at the end 2011. ­ ­ overnments Several questions will necessarily arise, including: Do local g continue to borrow to finance rural education? Given the transi- tion from debt to grant financing of rural compulsory education, has 102 Until Debt Do Us Part the grant allocation system been sufficient to finance compulsory education? Although China achieved universal compulsory education in 2007, a key question is how to ensure the sustainability and quality of educa- tion. Evaluating the debt write-off and addressing these questions will need to be done in the context of the evolving changes in the intergov- ernmental fiscal system. Chinese reform of its intergovernmental fiscal system is ongoing, as are major discussions on the assignment of expen- diture functions among the tiers of governments; the streamlining of the tiers, potentially into three tiers; and the need to continue to reform the intergovernmental revenue system to grant subnational govern- ments revenue flexibility at the margin, which is critical to underpin their access to financial markets. Conclusion International experience has shown that it is difficult to undertake debt restructuring and to write off debt liabilities while managing moral haz- ard. In a unitary system of government, the design mechanism by the central government has an important bearing on the incentive signals to subnational governments and financial markets. Important lessons can be drawn from China’s experience of formulating its strategy and framework for dealing with the write-off of rural compulsory education debt. It is important to have proper intergovernmental fiscal relations to assure the delivery of basic public services. Besides the proper assign- ment of expenditure and revenue among the tiers of government, it is necessary to provide formal fiscal instruments for subnational govern- ments to manage capital expenditure. One of the main reasons for the existence of China’s subnational government compulsory education debt was the lack of fiscal resources and fiscal instruments, such as for- mal debt financing. A suitable subnational debt management framework is a critical part of intergovernmental fiscal relations. In designing a sound sub- national debt management system, it is important to have an informa- tion system and an accounting and statistics reporting system to ensure the risks of subnational debt are transparent and reported. In addition, Restructuring of Legacy Debt for Financing Rural Schools in China 103 there needs to be a functioning audit department. More important, subnational governments in general have the incentive to overborrow because of the soft budget constraints and the common pool prob- lem. Thus, it is important to provide proper incentives to avoid moral hazard. Writing off subnational debt should proceed with the goal of improving intergovernmental fiscal relations. In China, county govern- ments should be responsible for servicing their debt, which was con- sistent with the responsibility assignment of the existing fiscal system. However, the county governments may lack incentives to pay their debt without the effective incentive provided by the central government. Although the central government has sufficient fiscal resources to write off the entire rural education debt, such a write off would lead to moral hazard problems. China established a system in which the distribution of grants was based on a transparent, rule-based, and output-based formula; the dis- tribution of the grants to a particular local government was not related to the size of the debt of that local government. If a grant to an indi- vidual jurisdiction were linked to the size of its debt, it could undermine efforts by that local government to prudently manage its debt service, and at the same time could encourage accumulation of additional debt in anticipation of larger bailouts. The system, based on the performance efforts and standard factors of local governments, such as the size of the student population and their respective fiscal capacity, encouraged local governments to achieve the goal while managing moral hazard problems. Each debt restructuring will need to examine the origin of the debt problem and the specific historical and institutional context of the leg- acy debt. Any debt restructuring will need to pay close attention to the design and its incentive effects. The experience and lessons learned from China’s rural debt restruc- turing can help generate lessons for developing a consistent strategy for debt restructuring in general. A rule-based debt restructuring reduces ad-hoc bargaining and adverse incentives, a hard budget constraint pre- vents moral hazard, and burden sharing provides proper incentives and avoids free-riding behavior, while also recognizing the incentive role played by higher levels of government to leverage reform. 104 Until Debt Do Us Part Notes  The findings, interpretations, and conclusions expressed in this work are those of the authors and do not necessarily reflect the views of The World Bank, its Board of Executive Directors, the governments they represent, or any other institutions with which the external authors may be affiliated. 1. The exchange rate at the time of writing of the RMB (renminbi) to the U.S. dol- lar was US$1 to RMB 6.30. The RMB has appreciated continuously since 2005, at about 3–5 percent per year. 2. In 2001, a package of policies called the Tax-for-Fee Reform was enacted. See the next section for details. 3.  Based on Article 30 of the Constitution of China of 2004, the administra- tive division of China is as follows: (a) the country is divided into provinces, autonomous regions, and municipalities directly under the central govern- ment; (b) provinces and autonomous regions are divided into autonomous prefectures, counties, autonomous counties, and cities; and (c) counties and autonomous counties are divided into townships, nationality townships, and towns. Municipalities directly under the central government and other large cities are divided into districts and counties. Autonomous prefectures are divided into counties, autonomous counties, and cities. 4.  The reform started with a pilot program in Anhui province and later was implemented nationwide. Before the reform, there were five agriculture-related taxes, including various taxes on agriculture and a slaughter tax, and various charges such as an education surcharge. There were three core components of the 2001 Tax-for-Fee Reform: (a) cancelations of charges and fees imposed on rural households, including the cancelation of township general fees, the edu- cational surcharge, and the slaughter tax; (b) two adjustments: (i) changes in the agriculture tax made the amount of taxable land and the tax rate fixed, and the production level should be determined by the average of the past five years; the maximum tax rate was set at 7 percent; and (ii) changes in the special agriculture tax. This tax is levied by the procedure applied to the agriculture tax but with a higher tax rate; and (c) one reform: the revenue the village can collect is essentially a surcharge on the agriculture tax, the maximum rate of which is 20 percent. This surcharge can be used to pay for three items in village expenditures: (i) salaries of the leaders in the village self-governing committee, (ii) the old-age support program, and (iii) the operational costs of the gov- erning body. After this reform, only the agriculture tax remained. In 2006, the agriculture tax was also abolished. 5.  In 2001, the State Council issued a new policy, the “Decision on the Reform and Development of Primary Education,” which assigned the responsibility of providing educational services to county governments. In May 2002, the State Council issued a complementary policy change for education. In the “Notice on Improvement of Administration of Compulsory Education in Rural Areas,” the Restructuring of Legacy Debt for Financing Rural Schools in China 105 responsibility for financing compulsory education was removed from township to county governments. The two regulations ended the township-centered system of the previous 17 years (1985–2002), and the county governments began to play a core role in providing educational services. 6. In 1993, the central government issued a Blueprint on Education Reform and Development, which stated that “two basics” should be reached before 2000. In the following years, the Ministry of Education issued several regulations on the qualifications of teachers and the status of school facilities. To meet the require- ments, township governments, administrative villages, and households made a big effort, but sizable debt was accumulated, since outlay expenditures were beyond the ability of local communities. 7. For example, 1,424 counties out of a total of 2,806 received earmarked transfer funds for teacher salaries in 2003. 8.  Following implementation of the State Council Directive in 2003, township and village government debt was transferred to the accounts of their respective county governments. 9. This and subsequent sections are based on discussions with Ministry of Finance officials. 10. “Auditing Report 2010,” No. 35, National Audit Office of China. 11. The deadline for calculating the size of debt was December 31, 2005. 12.  The grants for writing off the compulsory education debt for pilot areas were based on a standard compulsory education input gap (standard compulsory education debt) and the coefficient of grants. The formula was: Grants for a par- ticular pilot area = the standard input gap (of the area) × the coefficient of grants (of the area), where the standard input gap (of the area) = ∑[(0.85 × the number of students in rural compulsory education [county-level jurisdiction where the rural population stands for no less 60 percent of total population] × the standard input gap per student) + (0.15 × the number of the schools × the standard input gap per school × the unit cost difference)]. The standard input gap per student and the standard input gap per school are calculated, respectively, in accordance with the number of students and population density by provincial jurisdictions based on the overall input gap. The unit cost difference is calculated by county based on the factors related to construction cost, such as the geographic elevation and weather conditions. The coefficient of grants is determined by fiscal capacity. The coefficient grants for the middle and western regions follow the coefficient of fiscal difficulty applied in the general transfers. 13.  The annual grants for a pilot area were based on the schedule of writing off compulsory education debt for that area and used the following formula: The amount of grants (in year T) = the amount of debt that had been written off (in year T) / the total amount of the compulsory education debt that should be written off × the total grants for the particular pilot area. The grants were par- tially advanced at the beginning of the year based on the anticipated write-off during the year, and the remaining portion was given at the end of every year 106 Until Debt Do Us Part after completion on schedule. The funds provided by the central and subna- tional governments earmarked for writing off the debt were managed in the single account system of the National Treasury and monitored regularly. 14. The General Office of the State Council, 2007, No.70. 15. The 14 provincial jurisdictions were Anhui, Fujian, Guansu, Helongjiang, Hubei, Huizhou, Hunan, Inner Mongolia, Jiangsu, Jiangxi, Jilin, Ningxia, Shan’xi, and Sichuan. Bibliography Cai, Yue’e. 2002. Why Did Rural Teachers Quit Their Jobs? A Case Study on the Pre- fecture of Hengyang, Hunan Province. Statistics Bureau of Hunan Province, Hunan. http://www.hntj.gov.cn/csdc/news/ztdc/2002917171354.htm. Chang, Gene H. 2003. “Urbanization and Unemployment in China.” In Urbaniza- tion and Social Welfare in China, ed. Aimin Chen, Gordon G. Liu, and Kevin H. Zhang, 96–105. Aldershot, England: Ashgate Publishing. Henan Education Department. 2002. An Analysis of Education Finance of Henan Province. Henan. http://www.jyck.gov.cn/system/info/showinfo.asp. Hu, Yifan. 2002. “Difficult Issues in Rural Compulsory Education.” Caijin (11): 44–51. Huang, Shixian. 2006. “Exploration of Function of Village and Township Gov- ernments after Reform of Tax and Fees.” Academic Journal of Changbai (4): 27–30. Jia, Kang. 2008. “Build a Central-Local System with Matched Function and Finance.” China Public Finance 13: 1–3. Li, Qun. 2006. “Thinking about Funding for Compulsory Education in Rural Areas.” Journal of Education Exploration 2: 44–45. Liao, Hongfeng. 2004. “An Analysis of the Education Expenditure after Tax for Fee Reform.” Journal of Guangxi School of Public Finance 2: 43–48. Liu, Lili, and Steven B. Webb. 2011. “Laws for Fiscal Responsibility for Subnational Discipline: International Experience.” Policy Research Working Paper Series 5587, World Bank, Washington, DC. http://ssrn.com/abstract=1792244. National Bureau of Statistics. 2010. China Statistical Yearbook 2010. Beijing: China Statistics Press. Qiao, Baoyun. 2011. The Subnational Financial Platforms of China. Beijing: China Fiscal and Economic Press. Shi, Hongxiu. 2004. “The Decentralization of Lending and Borrowing Cannot Deal with Problem Assessment: The Way Out of Local Government Debt.” Journal of the Banker 5: 68–71. Wang, Wenhong. 2007. “A Study of Transfer Payment Mode of Public Finance in Rural Compulsory Education—Based on the Study of the Country-Level Finance from the Perspective of Public Finance.” Journal of Dalian University 28 (1): 123–28. Restructuring of Legacy Debt for Financing Rural Schools in China 107 Wang, Youwen. 2004. “Paying Attention to Loss of Teaching Staffs in Rural Areas.” China Education Daily, November 26. http://www.jyb.cn/gb/2004/11/26/zy/ jryw/12.htm World Bank. 2002. “China—National Development and Sub-national Finance: A Review of Provincial Expenditures.” Public Expenditure Review 22951, ­Washington, DC. Wu, Harry Xiaoying. 1994. “Rural to Urban Migration in the People’s Republic of China.” China Quarterly 139 (September): 669–98. Xiang, Jiquan, and Fangcheng Yuan. 2008. “The Effect of Reducing Revenues on Compulsory Education in Rural Areas—Current Financial Dilemma and Policy Choice of Compulsory Education in Rural Areas, February 23.” China Agricultural Economy and Information Net. http://www.caein. com/index.asp ?xAction=xReadNews&NewsID=31785. Yuan, Guilin. 2004. “Research on the Current Situation and Multi-Development Model of the ‘County-Level’ Administration Mechanism for Compulsory Edu- cation in Rural Areas.” Journal of Northeast Normal University (Philosophy and Social Sciences) 1 (1): 115–22. Zhang, Shaochun. 2007. Speech at the Executive Meeting of Writing-Off Rural Compulsory Education Debt, Chongqing, June 5. http://www.mof.gov.cn/pub/ nczhggbgs/zhengwuxinxi/lingdaojianghua/200806/t20080616_45565.html. Zhang, Shouxiang. 2005. “The Compulsory Education System in Rural China: Prog- ress, Problems and Solutions.” Basic Education Review 1: 4–8. Zhang, Zhuogong. 2004. “The Outstanding Issues and Policy Options in Dealing with the ‘County-Level’ Administration Mechanism for Compulsory Educa- tion in Rural Areas.” China Agriculture Education 1: 4–6. Zhou, Dongai, Ziling Liu, and Jianjun Tian. 2003. “Why Do Rural Teachers Quit?” Peo- ple’s Daily, April 6. http://www.people.com.cn/GB/kejiao/40/20030406/964244 .html. Zong, Xiaohua. 2010. “Empirical Test of Fiscal Decentralization, Population Move- ment and the Investment of the Local Government in Compulsory Education.” Journal of Statistics and Decision 15: 86–88. 3 Managing State Debt and Ensuring Solvency: The Indian Experience C. Rangarajan and Abha Prasad Introduction There has not been any repayment default among the Indian states,1 although fiscal stress and debt repayment pressures were experienced by many states in the late 1990s with continued deterioration evidenced in the early 2000s. The deterioration in the current account was the driv- ing force for declining fiscal health as reflected by the worsening of fiscal and primary balances. An analysis of the evolution of states’ debt, defi- cit, and interest payments reveals three distinct phases. The first, pre-1998 phase, was characterized by low current account (revenue balance)2 and fiscal deficits, with moderate debt levels. The second phase, during the late 1990s to mid-2000s, reflected significant deterioration in all key deficit indicators, with rising debt levels and interest burden. During this period, the outstanding states’ debt to gross domestic product (GDP) peaked at 32.8 percent in 2003–04, up from 20 percent in 1997–98, and interest payments as a share of revenue receipts increased from 16.9 to 26 percent over the same period. Of concern was the fiscal stress experienced by the central govern- ment over the same period (Pinto and Zahir 2004), during which the combined center-state fiscal deficit rose from 7.3 percent of GDP to 109 110 Until Debt Do Us Part 9.4 percent. Furthermore, this reflected only the direct liabilities of states; exacerbating the debt burden and repayment pressure were the contingent liabilities, in the form of guarantees issued by states to sup- port their enterprises. This was followed by the third phase and the onset of fiscal correction and reforms from the mid-2000s onward. This is reflected in the lowering of all key deficit indicators and debt and interest payments as a share of GDP. The resulting reform package had three interrelated components. First, to reverse the fiscal decline, a fiscal adjustment package was formu- lated to control the growth of current expenditures (such as wages and pension), and structural reform of the taxation system (such as moving from a turnover tax to a value-added tax) was instituted. S ­ econd, a rule- based institutional framework was developed to ensure the sustainabil- ity of the adjustment and consolidation. Third, there was a move from central government onlending to states toward market-based financ- ing, with a focus on both self-regulation (through fiscal legislation) and market discipline. The priority of fiscal consolidation was to restore the balance of revenue accounts, that is, to reduce the revenue deficit to zero. It was realized that even after lowering the primary deficit, the debt service repayment pressure and high indebtedness would continue, because about 80 percent of states’ borrowings during 2003–04 was at high-cost, non-market rates. But turning states to a sustainable fiscal path implied reducing both the stock of debt and the cost of borrowing. However, debt restructuring, write-offs, and relief would have an inherent moral hazard challenge. Being cognizant of this, the debt restructuring pro- gram was linked with broader institutional reforms, including provid- ing incentives for states to undertake difficult fiscal reforms. The Twelfth and Thirteenth Finance Commissions (FCs)3 compre- hensively examined the situation for both the center and the states, high- lighted their interdependence, and presented an overall strategy. Most subsequent reforms in terms of fiscal responsibility legislation (FRL) were based on a well-considered strategy and incentive structure. Although the steps taken were gradual, the synergistic effect of many institutional, fiscal, and legislative reforms was much greater. The reform efforts were initiated and implemented by different parts of the ­ government—the FC, the Ministry of Finance (central g ­ overnment), the states themselves, Managing State Debt and Ensuring Solvency: The Indian Experience 111 the Planning Commission, and the Reserve Bank of India (RBI). All par- ties were aware that this was not “business as usual” (World Bank 2005), and there was a sense of urgency about transforming the situation. Among the states themselves, there was a move away from competi- tive populism (Kurien 1999), which included subsidies and lowering tar- iffs, toward coordination by ending the competitive tax rate reduction and instituting the value-added tax, which proved to be highly buoyant. This, coupled with increases in the states’ share of central taxes instituted by the Twelfth and Thirteenth FCs and the high buoyancy of the center’s direct taxes, improved state finances and led to their progress towards the FRL goals. The coordination and consultation among all engaged enti- ties ensured consistency of approach and moved the reforms forward. There is ample fiscal literature on the states’ fiscal reform—fiscal rules, the quality of fiscal adjustment, expenditure and taxation reforms, power sector reform, and budget and financial management reforms.4 This chapter focuses on the states’ borrowing and debt restructur- ing process, underpinned by the move toward a rule-based framework and market discipline.5 It concentrates on the perspective of the policy maker during this period and reflects on the key challenge that was to balance the provisions of debt relief with the need to avoid moral haz- ard and enforce fiscal discipline. The rest of the chapter is organized as follows. Section two presents the states’ borrowing framework as prescribed by the constitution, and changes in the borrowing channels and lending policy for states, while incentivizing market access with a rule-based system. Section three summarizes the trends in states’ deficits, debts, and interest payments in the last two decades, and highlights interstate disparities in fiscal perfor- mance. Section four presents the major policy and institutional reforms undertaken to restructure states’ debt and discusses efforts to minimize moral hazard. Section five presents the impact and challenges of the ongoing global financial crisis. Section six offers conclusions. States’ Borrowing Regime India is a federal polity of 28 state governments and 7 union territories. The states’ borrowing regime is defined by federalism, characterized through the constitutional division of powers among the three levels 112 Until Debt Do Us Part of government—the center, the states, and the local bodies.6 The power to raise major taxes is allocated to the central government, while major expenditure responsibilities are assigned to states due to their prox- imity to local issues and needs. While states’ own revenues constitute 37 percent of total revenue receipts, their expenditures account for 55 percent of total central government expenditure (RBI 2011a). The imbalance is addressed through fiscal transfers from the center to the states, mandated by the FC. The constitutional arrangements for revenue sharing among the Indian federation and the consultative mechanism among the c ­ enter and states have tended to reduce the risk of explicit state defaults.7 Regarding the constitutional arrangement, the FC uses a formula-based approach to allocate taxes and grants, with the objective of filling the expenditure- revenue gap (deficit financing). The vertical sharing between the center and states is simplified by including all central taxes and excise duties in the divisible pool of central taxes.8 For the horizontal sharing among states, the FCs have attempted to correct the differentials in revenue capacity and cost factors inherent in the diversity of states. The pattern of transfers through the FC channel shows that the share in central taxes has persistently been the predominant component of revenue sharing since the First FC (RBI 2011a). Starting with the Ninth FC, a greater emphasis on fiscal discipline has been added to balance the gap-filling approach (RBI 2007).9 Residual imbalances in the fiscal accounts after the federal trans- fers are financed through borrowing. The main borrowing sources are domestic, external, and issuance of loan guarantees. The borrow- ing channels are multiple and the process complex but are organized around the principles of maintaining sustainability, solvency, and liquid- ity of the states (for a description, see box 3.1).10 The overall control is with the center, under Article 293(3) of the Constitution, which states that if any state government is indebted to the center, it requires the center’s permission to borrow. Further, the Constitution forbids states ­ from borrowing abroad on their own. Thus, all external borrowing must be onlent or guaranteed by the center.11 The limit on the annual amount and sources of borrowing is based on consultations among the center, the state government, the Planning Commission, and the RBI.12 Previously, after the delinking of plan13 Managing State Debt and Ensuring Solvency: The Indian Experience 113 Box 3.1  State Borrowings Borrowing channels for states are multiple and the process complex; some channels are controlled and restricted by the center and others are more autonomous. Borrowing channels controlled by the center are the following: Market borrowings. Market borrowings are controlled by the center and managed by the RBI.a •  The state securities issued through this channel are eligible for meeting the banks’ statutory liquidity requirements and are thus backed by “automatic” intercepts from the state treasury account (automatic debit). There have been no restructuring or defaults associated with these, and investors perceive an implicit sovereign guarantee attached to them. Loans from the center. Historically, the center used to borrow and onlend to states. This has •  now changed, with financial market developments and states’ ability to borrow on their own behalf, onlending from the center was discontinued in May 2005. Loans from banks and financial institutions. The center sets the global ceiling on the amount •  states can borrow from the banks and financial institutions, but the rate of interest is negotiated directly by the state with the concerned creditor. The rate of interest depends on the perceived credibility and fiscal position of the state. External loans. Previously, the center would onlend the proceeds in rupees at harder terms, •  adjusting exchange exposure and elongating maturities. With the recent change in lending policy, the entire loan proceeds are passed through directly by the center to states at the same terms (currency, maturity, and amortization) given by the creditor. The states bear the currency and the refinancing risk, but most do not undertake an impact evaluation of the cost-risk trade-offs of such transactions on their total debt portfolios (see table B3.1.1). Table B3.1.1  Sources and Features Attached to State Borrowings Amount controlled Automatic Creditor by the center intercepts perceives guarantee External loans Yes Yes Yes Loans from center Yes Yes — Market borrowings Yes Yes Yes Loans from bank and Yes No Partially financial institutions Provident funds No No No NSSF No No No Contingent liabilities No No Partially Note: NSSF = National Small Savings Fund. — = not applicable. Borrowing channels not controlled by the center are the following: Small savings loans and use of state provident funds.b Prior permission from the center is not •  required for these. The small savings schemes are run by the center with a social security objec- tive to encourage household savings. Eighty percent of the collections within a state’s territory are automatically passed on by the National Small Savings Fund to that state. The rate of interest (continued next page) 114 Until Debt Do Us Part Box 3.1  (continued) paid by states is currently fixed at 9.5 percent. The money is available for 25 years with a five-year grace period. Special purpose vehicles. States issue loan guarantees to special purpose vehicles, which borrow •  in the market with the backing of these guarantees. Anecdotal evidence suggests that loan pro- ceeds have been sometimes used to finance current state expenditures. Liquidity management: Ways and means advances (WMA) from the RBI. These are designed to meet temporary liquid- •  ity shortfalls. They are formula based and depend on the state’s total expenditures. If the short- fall is higher than the WMA amount, the state gets into overdraft, which is extended at a penal rate of interest to be cleared within 10 days or the account of the state is frozen. If there is sur- plus cash in the single treasury account, it is invested in 14-day intermediate treasury bills. In the interest of transparency, the number of days a state uses the facility during a fiscal year is published in the RBI’s Annual Report. Access to this short-term credit facility disciplines states to manage liquidity shortfalls prudently to avoid closure of accounts, and benefits them in avoiding arrears and payment defaults. a. The RBI manages domestic borrowings for each of the 28 states through separate agreements with each. b. A provident fund is a retirement benefit scheme; employees contribute 12 percent of monthly wages and can withdraw funds on retirement or on reaching age 55. Contributions are an unfunded liability in the public account, but balances are available to the state (see Rao, Prasad, and Gupta 2001). borrowing and plan grants, states tended to revise their objective of maximizing plan assistance by arguing for higher plan sizes, thereby committing to higher borrowing. This changed considerably with the enactment of FRL targets. Key decision parameters on the demand side include the states’ financing needs, developmental needs, repayment profile, and, since the early 2000s, its debt sustainability. On the supply side, an important factor is the absorption of liquidity from the market by both the center and states, without impinging on the supply of credit for the private sector for productive purposes.14 Since the mid-2000s, the ceiling on borrowing by a state has been capped by the fiscal targets under the state-level FRLs.15 Figure 3.1 illustrates the changing financing pattern of states’ debt during the 1990s and 2000s. This mirrors three phases, with a decline in the center’s loan intermediation and onlending (since 1998–99), an increase in the National Small Savings Fund (NSSF)16 and small savings borrowings (1999–2000), and the move toward market-based financing since mid-2000. Managing State Debt and Ensuring Solvency: The Indian Experience 115 Figure 3.1  Composition of Financing Pattern of State Deficits (as of End-March) 80 60 40 Percent 20 0 –20 8 e) 8 0 2 4 6 7 9 E) ) 5 BE –0 –0 –0 –0 –0 –0 –0 –9 00 (R ag 1( 01 06 04 03 97 05 08 07 –2 er 0 –1 –1 20 (av 20 19 20 20 20 20 99 20 10 09 20 19 95 20 – 90 Year 19 Market borrowings Loans from center NSSF Small savings, provident fund, etc. Source: Reserve Bank of India. Note: BE = budget estimates, RE = revised estimates, NSSF = National Small Savings Fund (described in box 3.1). Traditionally, loans from the center were the dominant source of funding for states. In keeping with the trend of financial sector liber- alization, the center’s loan intermediation role has been reduced since 1999–2000. The other notable change has been the rising share of NSSF and small saving loans (see box 3.1 for a description). The share of NSSF increased sharply to 69 percent during 2004–05 from 39 percent during 2001–02. This characterized a move from center-controlled borrowings to the autonomous NSSF but at higher cost (NSSF loans at 9.5 percent compared with cheaper market loans, weighted average of 8.39 percent during 2010–11). There is an inflexibility related with NSSF borrowings, since these are based more on availability and col- lection within the territory of the state than the requirement by the state to borrow. The NSSF loans are also at higher interest costs and have been more asymmetrically beneficial for the center (Thirteenth FC, 144). Given the Twelfth FC recommendations for greater autonomy and discontinuation of the financial intermediary role for the center, the lending policy was changed, with more market access for states.17 Thus, states got more freedom, but also greater responsibility to manage their 116 Until Debt Do Us Part debt. A consequence of the new lending policy was the move to mar- ket discipline and transparency to enhance credibility among the mar- ket participants. Competition gradually increased among states to avail themselves of the best market terms and obtain credit ratings. There has been evidence of some variation in the spreads among states, with some states borrowing at slightly lower rates, although the overall range of the spreads has been narrow (table 3.1). Cross-country evidence shows that spreads over central govern- ment securities should be linked to debt and deficit (fiscal) indicators of states. For example,18 Schuknecht, von Hagen, and Wolswijk (2009) concluded this for the European Union member states, and Lemmen (1999) analyzed similar issues for the subnational governments in Aus- tralia, Canada, and Germany. Poterba and Rueben (1999) found that states with tighter antideficit rules and authority of state legislatures can issue debt at a lower interest burden. However, somewhat coun- terintuitive is the case in India. Bose, Jain, and Lakshmanan (2011) indicate that the conventional deficit indicators have not been signifi- cant in determining the yield spreads during 2006–07 to 2010–11. The study, however, concludes that since the period is characterized by the prevalence of rule-based fiscal policy, it appears to have provided con- fidence to investors regarding states’ commitment to fiscal discipline. Although the impact of FRLs cannot be directly determined, it cannot be undermined. Table 3.1  Weighted Average Spreads during 2010–11 Weighted average General category states/ spreadb (basis points) union territories Special category statesa 30–40 Puducherry, Gujarat, Goa, Manipur, Nagaland, Meghalaya, Rajasthan, Tamil Nadu, Bihar Tripura 40–50 Kerala, Andhra Pradesh, Madhya Assam, Himachal Pradesh, Pradesh, Punjab, West Bengal, Jammu and Kashmir Uttar Pradesh, Karnataka, Haryana, Orissa 50–60   Sikkim, Uttarakhand Source: Rakshitra, various issues, The Clearing Corporation of India Ltd. a. Special category states are all the North-eastern states, along with Jammu and Kashmir, Himachal Pradesh, and Uttarakhand. They have distinct characteristics: a low resource base, cost disabilities due to their physical geography, sparse terrain, remoteness, and historical circumstances. These states account for only 5–6 percent of all states’ gross domestic product. b. Over the center’s benchmark. Managing State Debt and Ensuring Solvency: The Indian Experience 117 This shifting in the sources and methods of state borrowing has had a bearing on the interest payments, deficits, and debts of the states. The next section presents the changing trends of states’ fiscal deficit, debt composition, and the interest burden, and details on interstate variabil- ity in these key indicators. Trends and Composition of States’ Deficit, Debt, and Interest Burden An analysis of the evolution of states’ deficit, debt, and interest bur- den (defined as the ratio of interest payments to current receipts) during the 1990s and 2000s reveals three distinct phases (as depicted ­ gure 3.2). The first phase, in the early to mid-1990s, was char- in fi acterized by low current account and fiscal deficits, moderate debt levels, and a tolerable interest burden. The second phase, during the ­ late-1990s to mid-2000s, was characterized by a significant deteriora- tion in state finances, with all key deficit indicators, debt levels, and interest burden rising. The third phase, from the mid-2000s, was Figure 3.2  Deficit and Debt as Share of GDP and Interest Payments as Share of Revenues 5 35 30 4 GDF/GDP and RD/GDP Debt/GDP and IP/RR 25 3 20 2 15 10 1 5 0 0 –1 ) ) 20 008 8 – 9 20 –04 20 –06 20 –07 20 –03 99 99 20 000 20 –01 20 –05 19 –95 19 96 19 –97 19 –98 19 –93 19 94 20 –02 19 91 19 2 10 RE BE –9 0 9 –0 2 –0 – – 19 8– – 20 10 ( 1( 90 00 92 02 91 96 01 06 94 04 93 03 95 97 05 07 –2 9 19 Year GDF/GDP RD/GDP Debt/GDP IP/RR Source: RBI Handbook of Statistics on State Finances, various issues. Note: BE = budget estimates, GDP = gross domestic product, GFD = gross fiscal deficit, IP = interest payments, RD = research and development, RE = revised estimates, RR = revenue receipts. 118 Until Debt Do Us Part ­ haracterized by the onset of fiscal correction and reforms and mani- c fests with improvements in key fiscal indicators. Until the mid-1990s, states’ finances were relatively stable, character- ized by low current and fiscal deficits, averaging below 1 and 3 percent of GDP. Debt levels remained moderate at about 20 percent of GDP, and the interest burden hovered close to 15 percent of revenue. The turning point came during 1998–99, with a significant deterioration in the cur- rent account, which became a key driving force for the declining fiscal health of the states, with increased spending on administrative services and interest payments. The next phase, 1998–99 to 2003–04, saw the steep rise in the fis- cal deficit as a ratio of GDP—from 2.8 to 4.2 percent; the revenue deficit more than doubled from 1.1 percent of GDP to 2.5 percent. As a result, the states’ outstanding debt to GDP grew from 21.7 percent during 1997–98 to its peak of 32.8 percent during 2003–04. Inter- est payments as a share of revenue receipts (repayment burden) rose from 17.9 to 26 percent over the same period, and the primary deficit grew from 0.9 to 1.5 percent. This period, until 2003, was also char- acterized by higher interest rates, with the interest rates being gradu- ally liberalized; the average market interest rate on states’ borrowing was over 10 percent during this period. Concomitantly, the average interest burden, at 23.4 percent, was significantly higher than the 15 percent considered tolerable for a sustainable debt level (Dholakia, Mohan, and Karan 2004).19 There is vast fiscal literature on the factors leading to the deteriora- tion of state finances in the late 1990s. Factors considered critical to the fiscal deterioration include the impact of the wage revisions; inabil- ity to contain wasteful expenditure, including subsidies; reluctance to raise additional resources; and competitive reduction in taxes. Mohan (2000) pointed to the increasing debt service payments and inad- equate returns on government spending as important factors behind the deterioration in states’ fiscal conditions. Acharya (2001) and Rao (2002) attributed the worsening of revenue (current) balance during this period to the implementation of the Fifth Pay Commission recom- mendations.20 The RBI Study of State Budgets, 2002–03, while drawing attention to the growing fiscal and revenue deficit and high debt levels of states, pointed to the following causes of the deterioration in states’ Managing State Debt and Ensuring Solvency: The Indian Experience 119 fiscal condition: (a) an inadequate increase in tax receipts, (b) negative or negligible returns from public investments due to losses in public sector undertakings (PSUs), (c) large subsidy payments, (d) increased expenditure on salaries due to pay revisions, and (e) higher pension outgo. Another study, by Prasad, Goyal, and Prakash (2004), concludes that interest payments played a prominent role in the deterioration of state finances. Until the mid-1980s, interest rates on government borrowing were highly subsidized, indicative of the degree of financial repression. After the 1980s, the rates on government bonds became progressively aligned with market interest rates; during the 1990s there were increases in both bank deposit rates and policy rates (table 3.2). During the 1990s, aver- age interest rates rose, and those on state government bonds ­ averaged Table 3.2  Deposit Rate of Major Banks for Term Deposits of More Than One-Year Maturity percent Year Average interest rate Bank rate/repo rate/reverse repo 1 2.0 3.0 Mar-91 10.0 10.0 Mar-92 12.5 12.0 Mar-93 11.0 12.0 Mar-94 10.0 12.0 Mar-95 11.0 12.0 Mar-96 12.5 12.0 Mar-97 12.0 12.0 Mar-98 11.3 10.5 Mar-99 10.3 8.0 Mar-00 9.5 8.0 Mar-01 9.3 7.0 Mar-02 8.0 6.5 Mar-03 5.3 5.0 Mar-04 4.8 4.5 Mar-05 5.8 4.75 Mar-06 6.5 5.5 Mar-07 8.3 6.0 (continued next page) 120 Until Debt Do Us Part Table 3.2  (continued) Year Average interest rate Bank rate/repo rate/reverse repo Mar-08 8.3 6.0 Mar-09 8.3 5.0 Mar-10 6.8 5.0 Mar-11 8.6 6.75 Source: RBI 2011c. Note: Average interest rate refers to the midpoint of interest rates charged by commercial banks on demand deposits. In column 3, the policy rate used is the relevant policy rate at that time. The bank rate was used for the period prior to 2003, when it was in active use. For the subsequent period, the repo/reverse repo rate was used depending on the prevailing liquidity conditions in the system. over 10 percent during the 1990s (RBI). At the same time, the reli- ance on market borrowing to finance the fiscal deficits increased from 11 percent in the 1980s to 16 percent in the 1990s. Significant changes in the structure and cost of state government debt contributed to a sharp increase of about 60 percent in the repayment burden from the begin- ning to the end of the 1990s. Interest rates started softening in the mid- 2000s, and these were taken advantage of in formulating the debt swap scheme for states (discussed in “Debt Restructuring and Institutional Reform” section). The data in figure 3.2 capture only the direct and explicit state lia- bilities; exacerbating the debt burden and repayment pressure were the contingent liabilities, in the form of guarantees issued by states to support their enterprises. During the mid-to-late 1990s, there was a rapid increase in the issuance of guarantees by states to support their public enterprises, many of which could not borrow on their own credit strength.21 Although the latest data indicate that loan guaran- tees issued by states were lower at 2.8 percent of GDP by end-March 2009 compared to 3.3 percent of GDP in 2008, this does not incorpo- rate the unfunded pension liabilities or the losses of the state PSUs. The Thirteenth FC estimated that by the end of 2007–08, about 1,160 state PSUs had accumulated losses of about Rs 659.24 billion (almost 1.3 percent of GDP), particularly the implicit liabilities associated with power utility companies, because their large accumulated losses repre- sent a huge exposure for states.22, 23 The probability that these liabilities will devolve are not identical for each state, and thus cannot be treated uni- formly in terms of their fiscal impact.24 As a rule of thumb, assuming that Managing State Debt and Ensuring Solvency: The Indian Experience 121 about one-third of such liabilities devolve to the states to service, figure 3.3 presents a broader concept of “extended” debt, that is, debt inclu- sive of the likely devolvement of outstanding guarantees, to provide an assessment of the exposure and fiscal risk for the states. Extended debt is calculated as direct debt (explicit) plus one-third of the contingent liabilities25 extended by the state governments (as reported by them). This adds to the stress scenario being faced by the states. Along with the deterioration in state finances during the second phase (1998–99 to 2003–04) was the fiscal stress experienced by the central government. The combined center-state fiscal deficit rose from 7.3 percent of GDP in 1997–98 to an average of 9.3 percent over the period. Studies indicate that although India has had primary deficits, it has avoided an explosive rise in debt, mainly because of high economic growth rates relative to the interest rate paid on government debt. Milan (2011) analyzed the decomposition of India’s public debt trajectory using the method of debt dynamics and concludes that the strong rate of economic growth compared to the interest rate paid on debt helped avoid an explosive debt trajectory. The situation was similar for states’ finances, where the lower rate of interest on debt and the higher revenue buoyancy (Thirteenth FC, 126) (from both their own taxes and their share in central taxes) enabled improvements in the fiscal stance. Figure 3.3  Extended Debt as Share of GDP 40 35.5 5.0 33.4 32.9 4.5 35 30.1 28.4 27.7 27.2 4.0 30 24.5 3.5 Debt/GDP, % GFD/GDP, % 25 3.0 20 2.5 15 2.0 1.5 10 1.0 5 0.5 0 0 00 08 06 09 04 05 07 92 20 20 20 20 20 20 20 19 Year Debt/GDP Extend debt/GDP GFD/GDP Sources: Reserve Bank of India and author’s calculations. Latest data on guarantees are available only until 2009. Note: GDP = gross domestic product, GFD = gross fiscal deficit. 122 Until Debt Do Us Part Fiscal correction set in after 2004–05, with the onset of reforms that went beyond the “realm of fiscal space” (World Bank 2004, 11). These included reforms on the expenditure side to contain spending, restrict recruitment, and curb growth in administrative expenditures; and some states cut the cost of pension schemes and reduced subsidies (through power sector reforms), including closure of and privatization of selected PSUs. On the revenue side, reforms aimed to enhance revenue receipts by revising tax rates and broadening the base, while focusing on improving tax compliance. Institutional reforms reflected a paradigm shift, with the adoption of medium-term fiscal frameworks and FRL at the state level (Howes, Lahiri, and Stern 2003). Much has been written about the reforms to correct the fiscal imbal- ances and sectoral improvements, including improving the business ­ climate to facilitate growth (World Bank 2003b). The reforms under- taken specifically to restructure or reduce the debt and interest burden, along with those to enhance the credibility of states and ensure sustain- ability of debt, are discussed in the next section. Since the implementation of reforms in the mid-2000s, the declining fiscal/debt trends have been reversed. However, the 2008–09 global financial crisis has posed challenges. Another aspect to consider is that, at the macro level, the states’ aggre­ gate analysis masks state-level disparities in fiscal performance. The differentiation among state performance persists, but the dynamics change over time, with some states reversing their fiscal decline from the second phase to the third phase (for example, Karanataka and Orissa), while the record of some states continued to deteriorate (for example, West Bengal). Assessing the performance of individual states against the median for the period reveals that in terms of the primary deficit, among the nonspecial category states,26 Bihar, C ­ hattisgarh, ­ Gujarat, Haryana, Karnataka, Madhya Pradesh, Orissa, Punjab, and Uttar Pradesh have improved fiscal performance (primary balance)27 since 2004–05 (until 2009–10) compared with the deterioration d ­ uring 1998–99 to 2003–04 (compared with median values), while Goa, Jharkhand, Kerala, Maharashatra, Uttar Pradesh, and West Bengal con- tinued to have persistently high deficits even during the fiscal correction phase (from 2004 onward) (see figure 3.4). As expected, most of the states that reflected weak fiscal performance are also plagued with high debt and repayment burdens across the three Managing State Debt and Ensuring Solvency: The Indian Experience 123 Figure 3.4  Primary Deficit as Percentage of GSDP 10 Mizoram 8 Nagaland 6 Sikkim Himachal Jharkhand 4 Himachal Pradesh Pradesh Percent 75 percentile Uttarakhand 2 Median 25 percentile 0 –2 –4 8 4 0 00 –9 –1 05 91 –2 20 19 99 19 Years Source: The box plot was created by Stata using data from Reserve Bank of India. Note: The line in the middle of each box indicates the median (50th percentile); the top of the box indicates the 75th percentile and the bottom the 25th percentile. The lines above and below the box represent the adjacent values, which are within 1.5 times the interquartile range (iqr) of the nearer quartile (75th percentile for the line above and 25th percentile for the line below). The iqr is calculated as the value of the 75th percentile minus that of the 25th percen- tile. Thus, the largest value (upper end of the upper line) is identified by the 75th percentile plus 1.5 times the iqr, and the smallest value (the lower end of the lower line) is the 25th percentile minus 1.5 times the iqr. GSDP = gross state domestic product. periods. As can be seen from the box plot in figure 3.5, a large number of states have both debt and interest burden above the 75th percentile across the three periods under study. A case in point is West Bengal, which has persistently had high debt (an average of 45 percent during 2005–10) and a large interest burden (39.3 percent) continuing relent- lessly, even during the current period (figure 3.6). Analyzing vulnerability in terms of debt as a ratio of gross state domestic product (GSDP) and interest burden (payments as a share of each state’s revenue receipts) provides a useful indication of the suscep- tibility that states face. Table 3.3 plots states in a matrix that highlights states facing more vulnerability. Gujarat, Himachal Pradesh, Kerala, 124 Until Debt Do Us Part Figure 3.5  Box Plot Showing Debt-to-GSDP Ratio and Interest Burden Ratio 120 Mizoram 100 Arunachal Mizoram Pradesh 80 Percent 60 75th percentile West West 40 Bengal Median Bengal 20 25th percentile 0 8 4 0 00 –9 –1 05 91 –2 20 19 99 19 Years Debt/GSDP IP/RR Source: The box plot was created by Stata using data from Reserve Bank of India. Note: GSDP = gross state domestic product, IP = interest payments, RR= revenue receipts. Figure 3.6  Interest Burden in Selected States 50 45 40 35 30 Percent 25 20 15 10 5 0 8 4 0 00 –9 –1 05 91 –2 20 19 99 19 Years Bihar Gujarat Kerala Orissa Punjab Rajasthan Uttar Pradesh West Bengal Median Mean Source: Authors’ calculation using data from RBI reports. Managing State Debt and Ensuring Solvency: The Indian Experience 125 Table 3.3  States’ Vulnerability Matrix Debt Debt-GSDP Ratio Very high Medium Low (below Interest payment (above 50%) High (30–50%) (20–30%) 20%) Very high West Bengal (above 25%) High Himachal Gujarat, Kerala, Maharashtra (15–25%) Pradesh Punjab, Rajasthan Ratio of Medium Jammu and Andhra Pradesh, Bihar, Karnataka, Haryana, interest (10–15%) Kashmir Goa, Jharkhand, Tamil Nadu NCT Delhi payment to Madhya Pradesh, revenue Orissa, Uttar Pradesh, receipts Uttaranchal Low Arunachal Meghalaya, Tripura Assam Chhattisgarh (below Pradesh, 10%) Manipur, Mizoram, Nagaland, Sikkim Source: Authors compilation using data from RBI reports. Note: GSDP = gross state domestic product, NCT = National Capital Territory. Punjab, Rajasthan, and West Bengal reflect both debt levels of over 30 percent of GSDP and a high interest burden. The combined GSDP of these states accounts for over 12 percent of the national GDP. Their continued struggle with fiscal adjustment poses a challenge, which is further compounded by the global financial crisis (discussed in “Impact of the Global Financial Crisis and Going Forward” section). In keeping with the diverse fiscal situation in states, the Thirteenth FC recommended a state-specific approach for adjustment based on past fiscal performance (with 2007–08 the base year), and prescribed differentiated adjustment paths for different groups of states. It was estimated that to attain the aggregate target of states’ debt-to-GDP ratio of 25 percent, the aggregate fiscal deficit of states should be main- tained at 3 percent of GDP. Being an aggregate, however, this target indicator does not reflect the specific realities of individual states. For example, an abrupt reduction in fiscal deficits in states that also had high revenue deficits would lead to undesirable compression in capital expenditures. 126 Until Debt Do Us Part Thus, the Thirteenth FC, while keeping a balance between the need for customization with the requirement for adopting a uniform approach for determining targets for all states, recommended a differ- entiated approach. It was recommended that the nonspecial category states that had a revenue surplus or balance in the base year 2007–08 adopt a road map that eliminated their revenue deficits by 2011–12, and target fiscal deficit to 3 percent of GSDP. Other states with a higher revenue deficit in the base year were to adjust following a gradualist approach to avoid sudden cutbacks in capital expenditures, and elimi- nate the revenue deficit by 2014–15 and achieve a 3 percent fiscal deficit by 2013–14.28 Debt Restructuring and Institutional Reform The structural deterioration in states’ finances led to intense delib- erations among stakeholders—parliamentarians, policy makers, think tanks, and other interested parties—about reform options to not only reverse the fiscal decline and lower debt levels, but also to put state finances on a more sustainable path going forward. The fiscal ­ correction in state finances since the mid-2000s thus resulted from interrelated reforms on multiple fronts, helped in large part by the higher reveune buoyancy (Thirteenth FC) and the overall strong economic growth in India. The priority of fiscal consolidation was to restore the balance of revenue accounts—that is, reducing the ­ revenue deficit to zero. The reforms included the standard fiscal con- solidation measures through expenditure and taxation reforms. But, importantly, efforts were taken to develop a rule-based institutional framework, including fiscal responsibility laws, to ensure the sustain- ability of the consolidation. Such a rule-based system complemented the move from central government onlending to the market-based financing mechanism for meeting the states’ financing requirements. It was realized, however, that even after lowering the primary defi- cit, the debt service repayment pressure and high indebtedness would continue, since about 80 percent of states’ borrowing in 2003–04 was at high-cost, nonmarket rates. Research indicates that in addition to the important elements of fiscal consolidation, such as controlling the rapid growth of current Managing State Debt and Ensuring Solvency: The Indian Experience 127 expenditures and implementing structural taxation reforms, fiscal consolidation must include the objective of reducing repayment pres- sure by reducing interest costs (Dholakia, Mohan, and Karan 2004; Prasad, Goyal, and Prakash 2004). The Twelfth FC had also viewed the large interest payments as a major factor leading to the outstanding debt of states, and felt that reducing these payments was integral to attaining debt sustainability. With regard to the broad approach on the issue of debt sustainability, the Twelfth FC was of the view that debt relief measures were required as a prerequisite to achieve revenue bal- ance. Moreover, international experience showed that given the high indebtedness of states, it would be difficult to adhere to the fiscal tar- gets when established by the states’ fiscal responsibility law (Liu and Webb 2011). To achieve this would imply reducing both the stock of debt and the cost of borrowing.29 However, it was also recognized that debt write-offs, relief, and restructuring alone cannot ensure the sustainability of state finances. Policy makers were cognizant that waivers of loans and interest should be restricted to avoid moral hazard problems and encourage debt repayment discipline. The debt restructuring was thus linked to states undertaking reforms to increase revenue efforts, controlling expenditure, and reorienting expenditures toward supporting growth (Twelfth FC). This section focuses on the debt restructuring program and its links to incentive packages offered to states for undertaking institutional reforms. Debt Relief and Fiscal Responsibility Legislations Debt relief had been provided by the waiving of repayment and/or interest payments due, altering the terms of repayment, reducing inter- est rates, and consolidation of loans. In the 1980s and 1990s, successive FCs had given unconditional debt relief to states, although the relief had been provided only periodically, and the amount of relief was not significant (table 3.4).30 Thus, states have had to repay most of the debt they incurred. The Tenth and Eleventh FCs started to link debt relief with fiscal performance.31 However, it was not until the Twelfth FC that debt relief was linked explicitly to rule-based legislative reforms. In a pathbreaking move, the Twelfth FC recommended debt relief for states contingent upon the enactment of fiscal responsibility laws and 128 Until Debt Do Us Part Table 3.4  Debt Forgiveness by Finance Commission Finance Commission Year of report Rs (billion) GDP Rs (billion) % of GDP Sixth 1974 20 667 2.95 Seventh 1979 22 1,025 2.11 Eighth 1984 23 2,223 1.03 Ninth 1989 10 4,357 0.22 Tenth 1995 5 10,672 0.05 Eleventh 2000 34 20,050 0.17 Twelfth 2005 535 31,494 1.70 Sources: McCarten 2001; Report of Thirteenth Finance Commission 2009. Note: GDP = gross domestic product. incorporation of a fiscal correction path, with milestones for attaining fiscal targets while improving the current (revenue) balance (reducing the deficit to zero by 2008–09). To implement the recommendations of the Twelfth FC, the Debt Consolidation and Relief Facility was introduced during 2005–06, which provided debt relief through consolidation, rescheduling repay- ments for a fresh term of 20 years, and lowering of the interest rate on the debt to 7.5 percent. All states were eligible to obtain relief from the year they enacted FRL. This amounted to Rs 187 billion in terms of lower interest payments, and Rs 211 billion in terms of lower repay- ments, totaling Rs 398 billion (US$8.9 billion32) during 2005–06 to 2009–10. In addition, repayments due during 2005–10 on central loans contracted up to March 31, 2004, (after consolidation and rescheduling) were eligible for write-off subject to the reduction in revenue deficits. The debt write-off would also be subject to containment of the fiscal deficit to the 2004–05 level. Subject to these provisions, if the revenue deficit were brought down to zero by 2008–09, all repayments during 2005–10 would be written off. Carrying forward the momentum to support states toward urgent fiscal correction, the Thirteenth FC worked out a differentiated fiscal adjustment road map (described in the previous section), with a state- specific approach based on past fiscal performance (using 2007–08 as the base year) for different groups of states. A key requirement is that all states eliminate their revenue deficits (the deficit on current balance), but they can have a fiscal deficit of 3 percent of GSDP by Managing State Debt and Ensuring Solvency: The Indian Experience 129 2014–15, along with a reduced debt target of 24.3 percent of GDP in the same year (from 27 percent in 2008–09). The debt relief granted was similar to the Twelfth FC; all loans to states from the Govern- ment of India outstanding as of 2009–10 would be written off if the state enacted or amended its FRL. Moreover, interest on past NSSF loans (contracted during 2006–07) was reduced to 9 percent from 9.5 percent. The center enacted the Fiscal Responsibility and Budget Manage- ment Act in 2003, with applicability only to the national government. Some states had also enacted their own FRLs before the center (for example, Karnataka and Punjab, in 2002), and many states had since 2003 adopted FRLs in line with the national law. The Twelfth FC sub- sequently mandated that states pass FRLs to avail themselves of the benefit of debt relief, with revenue deficits (total revenue minus current expenses) to be eliminated and fiscal deficits to be reduced to 3 ­percent of GSDP by fiscal year 2009. Since then, all 28 states have passed FRLs, 33 most of which require the state to present a medium-term fiscal plan with multiyear rolling targets for key fiscal indicators, along with the annual budget, to the state legislature. Some of the FRLs, passed by states, also place limits on guarantees; others mandate the disclosure of contingent liabilities and other borrowing. Most FRLs require disclo- sure of significant changes in accounting policies. Fiscal targets adopted by Indian states are remarkably similar to each other with respect to fiscal and revenue deficits. Some states adopted additional legislation on fiscal targets, such as the Kerala Ceiling on Government Guarantee Act (2003), which was enacted the same year as its FRL. According to the Guarantee Act, the guarantee outstand- ing for any fiscal year shall not exceed Rs 140 billion,34 no government guarantee shall be given to a private entity, and the Guarantee Redemp- tion Fund shall be established. Other initiatives included the setting up of (a) the Consolidated Sinking Fund (1999) to provide a cushion for repaying market loans of states (20 states have established this), (b) the Guarantee Redemption Fund (2001) to provide a cushion for servicing any contingent liabilities because of guarantees issued by state govern- ments to its PSUs (11 states have established this), and (c) several tech- nical committees and working groups on topical issues of cash and debt management.35 130 Until Debt Do Us Part Debt Swap and Securitization: A Move toward Market-Based Financing The fiscal correction was given an impetus with the introduction of a “debt swap scheme” during 2003–04 to lower the existing interest bur- den and increase market access. Loans from the center amounting to Rs 1,000 billion (US$23 billion36) with interest rates in excess of 13 per- cent were substituted with new market loans and small savings proceeds at lower rates of interest; the outstanding debt remained unchanged. The market conditions prevailing were fortituous and the rates were significantly lower, at 7.5 percent (RBI State Finances Study 2004–05, ­ p. 24), enabling an interest savings for states of Rs 310 billion (US$7.1 billion37) and 0.75 percent per year in revenue (Twelfth FC). This direc- tion toward the market was reaffirmed by the Twelfth FC in conjunction with state debt relief, where it stated, “As regarding the future lending policy, the central government should not act as an intermediary and allow the states to approach the market directly” (Twelfth FC, 236). States issued “power bonds” to securitize the fiscal risks emanat- ing from the losses of electricity utilities arising from the gap between the cost of producing power and the tariff charged. This gap between the cost and tariff had resulted in significant losses and an accumula- tion of arrears. With the securitization, arrears and accrued interest at about 1.5 percent of GDP were cleared by states through the issuance of 15-year tax exempt “power bonds.”38 Cognizant of the moral hazard issue, this was clearly announced as a one-time settlement measure and was supplemented with reforms to ensure discipline going forward. Par- ticipating states qualified for funds on the basis of reform milestones and improvements in the reduction of commercial losses. Although state liabilities had increased by 22.8 percent during 2003–04 at the time of issuance of these bonds, many states have prepaid, and only Rs 144.23 billion (US$3.23 billion39) remained as of end-March 2011. In addition to the above debt restructuring program to link with institutional reform and move toward market access, the role of the RBI is also important. First, as the regulator of the banking sector, the RBI sets the statutory requirements for banks to hold state debt. This increases the acceptability of state securities by the market. Second, the RBI tightened the regulation for use of the overdraft facility by states. Previously, states had resorted to the facitily as a way to roll Managing State Debt and Ensuring Solvency: The Indian Experience 131 over short-term borrowing to finance structural deficits. The terms and conditions for facilty use were formula based and specified. More- over, the use of the facility by the states is disclosed to the market on an ex-post basis. For example, the market has information on the bet- ter performers compared to the chronic-deficit states (table 3.5 shows that Punjab, Uttarkhand, and West Bengal depended on this facility during 2010–11 to meet their temporary resource gap). Such informa- tion influences market sentiment and spreads, while lowering credit ratings. The intent of the FRL, debt swap, securitization, and the move toward market operation was to support the fiscal discipline reform at Table 3.5  States’ Overdrafts and Access to Cash-Credit Number of days Special WMA Normal WMA Overdraft 2009–10 2010–11 2009–10 2010–11 2009–10 2010–11 Andhra Pradesh 1 3 — — — — Haryana 7 10 5 10 — 8 Kerala 18 — 2 — — — Madhya Pradesh 11 — 11 — — — Maharashtra — — — — — — Karnataka — — — — — — Nagaland 69 — 45 — 13 — Punjab 130 133 128 132 29 13 Rajasthan — — — — — — Uttar Pradesh 8 4 8 4 — — West Bengal 95 195 15 113 8 62 Himachal Pradesh — — — — — — Manipur — — — — — — Mizoram 29 25 15 15 — — Goa — — 1 — — — Uttarakhand 69 35 26 12 9 10 Meghalaya — 1 — — — — Jharkhand — — — — — — Source: Reserve Bank of India, Annual Report 2010–11. Note: WMA = ways and means advances. Normal WMA is formula based, special WMA is after access to normal WMA but is collateralized. — = no access to facility and strong cash management position. 132 Until Debt Do Us Part the state level to reverse the structural decline of state finances from the late 1990s to the early 2000s. It will be difficult to precisely evaluate the direct impact of these reforms. In this context, a study by Liu and Webb (2011) concludes it would be difficult to precisely separate and mea- sure the effects of the FRL given the lender-borrower nexus and various channels that would influence government fiscal deficits and indebt- edness. Nonetheless, it was noted that to the extent the FRL intends to improve government finance and avoid over-indebtedness, it is worth- while ascertaining whether FRL has been associated with improved fis- cal outcomes.40 Liu and Webb (2011) choose growth of public debt before and after passing subnational FRL in several countries, including India. The measurement of the fiscal improvement or deterioration was normal- ized, since each state government might have passed its FRL in differ- ent years. The paper shows that in Indian states, the growth of debt to GSDP was slower in the post-FRL period than in the pre-FRL period for 24 of 26 states. Twenty-one of these 24 states had reversed the trend of increasing debt to GSDP in the pre-FRL period. A study on the “Dynamics of Debt Accumulation in India” (Ran- garajan and Srivastava 2008) pointed to the fact that accumulation of debt can be seen as the result of the balance between cumulated primary deficits and the cumulated weighted excess of growth over interest rate. Decomposing the change in the central government’s liabilities relative to GDP shows that a significant part of the cumulated primary deficit could be absorbed due to the excess of growth over interest rates. How- ever, this cushion is not always available, and the sharp increases in debt relative to GDP during 1997–2003 were because of both factors, that is, cumulated primary deficit and excess of effective interest rate over growth rate. One study (Milan 2011) shows that strong economic growth rela- tive to the interest rate paid on government debt helped India avoid an explosive rise in debt despite the existence of successive primary deficits. The same holds true for the aggregate performance of states; figure 3.7 shows that the fiscal correction phase in states also coin- cides with a higher GDP growth rate and lower rate of interest paid on state debt. This was also in part the result of fiscal correction, which led to a reduction in government dis-savings and debt. This, Managing State Debt and Ensuring Solvency: The Indian Experience 133 Figure 3.7  Differential between GDP Growth Rate and Interest Rate on State Debt 25 Onset of Period of 20 deterioration correction with g > r where 15 g