4DA-R E G I O N A L A N D S E C T O R A L S T U D I E S MAV fqq5 Decentralization of the Socialist State Intergovernmental Finance in Transition Economics / 7-,; EDITED BY .$-t RICHARD M. BIRD ROBERT D. EBEL CHRISTINE 1. WALLICH N a_F Decentralization of the Socialist State Intergovernmental Finance in Transition Economies WORLD BANK REGIONAL AND SECTORAL STUDIES Decentralization of the Socialist State Intergovernmental Finance in Transition Economies EDITED BY RICHARD M. BIRD ROBERT D. EBEL CHRISTINE 1. WALLICH C) 1995 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, N.W., Washington, D.C. 20433 All rights reserved Manufactured in the United States of America First printing May 1995 The World Bank Regional and Sectoral Studies series provides an outlet for work that is relatively limited in its subject matter or geographical coverage but that contributes to the intellectual foundations of development operations and policy formulation. Some sources cited in this paper may be iiformal documents that are not readilv available. The findings, interpretations, and conclusiolIs expressed in this publicationl are those ofthe authors and should not be attributed in anv manner to the World Bank, to its affiliated organizations, or to the members ofits Board of Executive Directors or the countries they represent. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, wvhen the reproduction is for noncommercial purposes, without asking a fee. Permissioni to copy portions for classroom use is granted through the Copyright Clearance Ceinter, Inc., Suite 910, 222 Rosewood Dr., Danvers, Massachusetts 01923, U.S.A. The complete backlist ofpublications from the World Bank is showzvn in the annual Index of Publications, which contains an alphabetical title list and indexes of subjects, authors, and countries and regions. The latest edition is available free of charge from Distribution Unit, Office ofthe Publisher, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Publicationis, The World Bank, 66, avenue d'1ena, 75116 Paris, France. Richard M. Bird is professor of economics and director of'the Internationlal Centre for Tax Studies, University of Toronto. Robert D. Ebel is an economist in the World Bank's Europe and Central Asia/Middle East and North Africa Technical Department. Christine *. Wallich is lead economist in the Central Europe Department of the World Bank's Europe and Central Asia Region. Cover design by Sam Ferro Library of Congress Cataloging-in-Publication Data Decentralization of the socialist state / Richard M. Bird, Robert D. Ebel, and Christine 1. Wallich, eds. p. cm. - (World Bank regional and sectoral studies) Includes bibliographical references and index. ISBN 0-8213-3186-8 1. Intergovernmental fiscal relations-Europe, Eastern. 2. Fiscal policy-Europe, Eastern. 3. Decentralization in government-Europe, Eastern. 1. Bird, Richard Miller, 1938- . 11. Ebel, Robert D. III. Wallich, Christine, 1952- . 111. Series. HJ1000.7.D43 1995 336.4-dc2O 95-2528 CIP Contents Contributors vii Preface ix Acknowledgments xvi 1. Fiscal decentralization: From command to market 1 Richard M. Bird, Robert D. Ebel, and Christine I. Wallich 2. Financing local government in Hungary 69 Richard M. Bird, Christine 1. Wallich, and Gdbor Peteri 3. Financing a large municipality: Budapest 119 Robert D. Ebel and Peter Simon 4. Local and intergovernmental finances in Poland: An evolving agenda 153 Luica Barbone and James F. Hicks 5. Intergovernmental fiscal relations in Bulgaria 183 Jorge Martinez-Vazquiez 6. Decentralization and local government finance in Romania 223 Felix A. Jaik(obl 7. Fiscal decentralization and intergovernmental relations in Albania 251 David Sewell and Christine I. Wallich 8. Subnational fiscal decentralization in Ukraine 281 Jorge Martinez-Vazquiez, Charles F. McLure, Jr., antd Sally WVallace V Vi DECENTRALIZATION OF THE SOCIALIST STATE 9. Intergovernmental fiscal relations in the Russian Federation 321 Roy Bahl and Christine I. Wallich 10. Special issues in Russian federal finance: Ethnic separatism and natural resources 379 Charles E. McLure, Jr., Christine I. Wallich, and Jennie I. Litvack Maps Hungary 68 Budapest, Hungary 118 Poland 152 Bulgaria 182 Romania 222 Albania 250 Ukraine 280 Russian Federation: Republics, krais, oblasts, and okrugs facing page 324 Russian Federation: Non-Russian population, 1989 facing page 382 Russian Federation: Selected natural resource endowments facing page 390 Bibliography 405 Index 425 Contributors Roy Bahl is professor of economics and director of the Policy Research Center at Georgia State University. Luca Barbone is principal economist in the World Bank's Europe and Central Asia Region, Central Europe Department. Richard M. Bird is professor of economics and director of the International Centre for Tax Studies, University of Toronto. Robert D. Ebel is a senior economist in the World Bank's Europe and Central Asia/Middle East and North Africa Technical Department. James F. Hicks is principal urban finance economist in the World Bank's Transportation, Water and Urban Development Department. Felix A. Jakob is a senior economist in the World Bank's Europe and Central Asia Region, Eastern Europe Department. Jennie I. Litvack is country economist for Viet Nam in the World Bank's East Asia and the Pacific Region. Jorge Martinez-Vazquez is professor of economics and associate direc- tor of the Policy Research Center at Georgia State University. Charles E. McLure, Jr., is senior fellow at the Hoover Institution, Stanford University. Gabor Peteri is a researcher at the Know-How Program for Local Governments, Budapest, Hungary. vii viii DECENTRALIZATION OF THE SOCIALIST STATE David Sewell is a senior economist in the World Bank's Europe and Central Asia/Middle East and North Africa Technical Department. Peter Simon is a researcher in the World Bank's Europe and Central Asia/Middle East and North Africa Technical Department. Sally Wallace is assistant professor of economics and senior associate of the Policy Research Center at Georgia State University. Christine 1. Wallich is lead economist in the World Bank's Europe and Central Asia Region, Central Europe Department. Preface This volume of studies on fiscal federalism and intergovernmental finance in the transition economies of Eastern Europe and the former Soviet Union emerged out of our interest in intergovemmental finances in market economies and our realization, as our work in Eastern Europe and the former Soviet Union proceeded, that intergovernmental fiscal problems in transition economies are both very differenttfrom those of mar- ket economies and very similar to each other. The objective of this book is to provide a systematic review and analysis of intergovernmental fiscal issues and policies in seven countries in this region as they make the transition from socialism. Key themes of the book are that local finance is not a local matter and that the ongoing reforms in subnational finance taking place in virtually every transition country are more important than is generally recognized. The literature on the transition from command to market to date has tended to focus on the "big picture" issues of rapid price reform, enterprise privatization, governance, and macroeconomic stabilization policies. But the fiscal decentralization that is now taking place in all of the transition economies of Eastern and Central Europe and in most of the former Soviet Union also has major implications that have not yet received much scrutiny. How the ensuing issues in intergovernmental finance are resolved will influence privatization and the transition to a market economy; macroeconomic stabilization; and the maintenance of the social safety net that is often critical to maintaining support for reform. An important objective of this book is to fill this gap in the litera- ture on transition. ix X DECENTRALIZATION OF THE SOCIALIST STATE * * * Research on this subject began with fieldwork on Hungary in 1991, imme- diately after Hungary's first local elections and just weeks after its Law on Local Self-Government became effective. This law provided for full "autonomy" for Hungary's local governments. The breadth of spending responsibilities assigned to local governments was notable and included many spending areas that had not been local responsibilities in Hungary before, such as social welfare, education, and health. But the law gave localities few taxing powers. For macroeconomic reasons, the central gov- ernment preferred to maintain control over the major national tax bases for its own budget. With the local sector kept on a tight fiscal leash, the autonomy provided for in the law was more apparent than real. How- ever, local governments inherited from the prereform period substantial assets in the form of enterprises, which they saw not only as potential revenue eamers but also as entities that could potentially help them carry out public spending functions, supplementing the budget. This initial study was followed by further research and fieldwork in Albania, Bulgaria, Poland, Romania, Russia, and Ukraine, where similar processes were taking place. Common themes emerged. In many coun- tries macroeconomic strains on the central government were leading to greater assignment of spending functions-including socially important spending-to lower-level governments, often themselves newly created. In Russia, for example, most spending on the social safety net, including unemployment insurance, as well as capital investment, was shifted "downstairs" in 1992. On the other hand, the more productive revenue sources invariably were reserved for the central government. In Hungary, for instance, local governments are assigned five local taxes that together might generate at best some 1 to 2 percent of GDP. Everywhere transfer systems remained ad hoc, negotiated, and often counterequalizing. In many countries the continued local public owner- ship of enterprises has the potential to threaten effective privatization. In Russia, for example, the entrepreneurial capital city, Moscow, hesitates to privatize key local enterprises, fearing the employment losses that may come about, and has made an important new equity investment in McDonald's to generate foreign exchange revenues. Everywhere, the general lack of local fiscal discretion, responsibility, and accountability, along with a general absence of revenue autonomy and ability to set local tax rates, threaten efficient provision of local public services. Poland's new laws on fiscal decentralization, for instance, fix all local tax rates and tax bases from the center. And everywhere, as enterprises close or privatize, the divestiture of their social assets-schools, clinics, kindergartens, and sometimes more-is wreaking havoc with local bud- gets as the new local governments are pressed to take on these tasks. Preface xi The research presented in this volume suggests that the design of a well-functioning intergovernmental system may often be the key to suc- cessfully achieving such major reform goals of the transition economies as macroeconomic stability, privatization, and maintenance of a social safety net that is robust enough to make reform palatable to the popula- tion. It also suggests that a broader analytical framework than is usually involved in the study of intergovernmental finance is needed to analyze decentralization issues in the transition economies. Traditional analysis argues that responsibilities for stabilization and income distribution are primarily, and properly, central government concerns. In this model, resource allocation is the primary concern of local governments, which should be financed mainly from user fees and taxes on local residents, with transfers making up any gap. The appropriate assignment of local expenditures and taxes and the design of transfer systems are, of course, key issues in transition economies too, but other aspects are also critical, as set out in chapter 1 of this book, and a broader framework of analysis is therefore needed. What should this broader framework include? First, it must take account of the likelihood of continued local government ownership of enterprises on a significant scale and factor this into plans for both priva- tization and intergovernmental reform. Second, it must focus more close- ly on local governments' role in stabilization, since stabilization policies have both affected and been affected by the emerging intergovernmental reforms. Of course, subnational spending or deficits cannot on their own create macroeconomic problems, since subnational governments cannot print money. But due to the generally "soft" budget constraints that still apply to banks and enterprises and to many government agencies them- selves, local government deficits in many transition economies can be and often are monetized. And third, to the extent the social safety net has been assigned to subnational governments, analysis must address distri- butional issues in local finance more carefully than is commonly the case. How to change the intergovernmental legacy of the past in order to accommodate the needs of the present and to anticipate the demands of the future in these and other respects is the focus of the book. * * * This volume on fiscal federalism and intergovernmental finances in Eastern Europe and the former Soviet Union is thus a companion piece to Russia and the Challenge of Fiscal Federalism (Wallich 1994), recently published by the World Bank in its Regional and Sectoral Studies series. This earlier volume stressed the importance of appropriate design of intergovernmental fiscal systems for ensuring a robust safety net, facili- tating fiscal balance and stabilization, and promoting privatization in Xii DECENTRALIZATION OF THE SOCIALIST STATE Russia. The present book explores many of the same themes in a wider country and comparative context. For Russia, facing considerable centrifugal forces, the contribution of fiscal federalism to nation building could not be overlooked. Politics and history are also important motivating forces behind the demands for greater fiscal autonomy and decentralization in the countries dis- cussed in this volume. The present book, however, does not attempt to address in any detail the dimensions of the political transition in Eastern and Central Europe and the former Soviet Union; its primary focus is on economic aspects. Nor does it deal directly with such "big picture" reform issues as the sequencing of reforms, financial sector reform, or price liberalization, although-as we argue-intergovernmental finances may have a significant impact on the overall success of a reform program. Moreover, while there is significant empirical work in most of the country chapters, this is not a data-intensive study. For most of the countries discussed here, it is too early for such studies, and much of the data needed for concrete analysis of policy options at the subnational level is absent. Indeed, even gathering the sometimes sparse and seldom fully satisfactory information reported here took substantial fieldwork in most countries: information may exist, but it is seldom available in accessible form even in the country itself. .r * * The aim of the book is to provide as accurate a picture from a policy per- spective as is possible at this early stage of the emerging process of fiscal decentralization in Eastern Europe and the former Soviet Union. Countries such as Hungary and Poland that have moved quite far down the path of local government reform and that may now be embarking on a "second generation" of reforms, as well as those that are only now beginning the process, such as Albania and Romania, ate considered. The order of the chapters is designed to highlight this focus. Chapter 1 is a comparative overview of the nature of intergovernmental finance in transition economies; it describes how these systems differ from those in market economies and elaborates on the need for a broader-than-traditional framework for analyzing subnational finances in transition economies. The relatively advanced case of Hungary is then discussed in chapter 2, followed by a detailed look in chapter 3 at how the Budapest municipal government is operating within Hungary's emerging intergovernmental system. Chapter 4 discusses Poland, which, like Hungary, is making important progress toward public sector reform. The next four chapters, 5 through 8, examine countries (Bulgaria, Romania, Albania, and Ukraine) in which the reform of inter- governmental finance is at earlier stages. Finally, chapters 9 and 10 take Preface xiii a close look at recent developments in Russia, by far the largest and most complex of the countries examined here and the only one with a truly federal character. The Russia chapters complement and expand on earlier work by Wallich (1992a, 1994). Each country chapter evaluates the strengths and weaknesses of the emerging structures and suggests possible directions for further reform. This book, with chapters contributed by fourteen authors, was writ- ten as a team effort, and with a broadly common framework of analysis. While each country case differs in its emphasis, reflecting the particular circumstances of the country in question, the chapters follow a common analytical framework to the extent possible. Each discussion touches first on the traditional public finance areas of expenditure assignment, revenue assignment, and intergovernmental transfers. Then, those aspects that are outside the traditional framework of analysis in this field, including priva- tization, stabilization, and the social safety net, are addressed, with differ- ent emphasis in each country according to the importance of the issues in shaping the emerging intergovernmental and subnational fiscal structure. The analysis focuses on the period since 1989, the year in which major political change began in most of Eastern Europe, with the decentraliza- tion movements following almost immediately. For most countries, the empirical work is necessarily based largely on the 1992-93 period. Although the speed of change is dizzying across Eastern Europe and the former Soviet Union, and it may seem that the timeliness of the material contained here is on occasion limited, this does not necessarily impair its relevance. As shown in this book, much of the change in intergovern- mental relations in each country has, after the initial reforms, been evo- lutionary in nature. In some countries, the changes have been more apparent than real. In others, older forms and practices of local govern- ment are resurfacing. The analysis of even the brief period covered here may thus be of much more extended relevance. Similarly, although the focus of the discussion in this book is on Central and East European countries, similar tendencies and problems may well also be manifest in the other republics of the former Soviet Union and the former Yugoslavia, as well as in similar countries else- where. Kornai, in his well-known work, The Socialist System (1992), lists twenty-six "socialist" countries in 1987. If we add to this list the fifteen republics of the former Soviet Union and the six of the former Yugoslavia (and deduct the now-vanished East Germany), there would appear to be at least forty-six countries that only five years ago had more or less "socialist" governmental structures. By 1994, however, it is probably fair to say that all but two of these countries (Cuba and the XiV DECENTRALIZATION OF THE SOCIALIST STATE Democratic People's Republic of Korea) are now in some "transitional" stage, so that some or all of the analysis of this book should be applica- ble. In essence, therefore, this book represents an interim analysis on what seems likely to prove an ongoing and important problem in many transition economies. This book is aimed primarily at readers with an interest in the transition economies of Eastern Europe and those of the former Soviet Union- Russia and Ukraine, in particular. It should also be of interest, however, to those working on the Baltics and other former U.S.S.R. states, includ- ing the economies of Central Asia, where the "Soviet model" of inter- governmental finances was transplanted in the middle decades of this century and where new forms also must be found. Finally, it should be relevant to other countries in transition in Asia, such as China and Viet Nam, which adopted Soviet structures. In both these countries intergov- ernmental reforms are moving apace and, as in Eastern Europe, have been found to have important ramifications for poverty and well-being, macroeconomic policies, and privatization of enterprises. The basic methodological approach underlying the book is field- work. In the rapidly changing circumstances of these countries, the "facts" on intergovernmental fiscal affairs are not found neatly compiled and predigested in a quiet library in the central capital. Often, indeed, central officials, caught up in their own transitional problems, proved to have little knowledge of what was really happening "out there." The authors therefore visited rural and urban areas in each country. Resource-rich and resource-poor areas were visited, as were, in some countries, localities with important concentrations of disadvantaged, ethnic populations, such as those in poorer parts of Romania and in the Russian Federation. The regional character of many countries and the existence of ethnic minorities makes the design of intergovernmental policies especially challenging. Special efforts were made to visit regions in decline as a result of sectoral restructuring and to examine the impact of enterprise divestiture of social assets on local budgets. One region in Russia, for example, found its entire industrial base gone with the clos- ing of the military industries. These enterprises-not the city budget- had also financed all the city's social sector expenditures, including health and education. Authors talked to various central departments and officials in capital cities and in provincial or county seats, as well as to a wide variety of subnational officials, ranging from mayors to tax collectors, in large and small cities and villages. They gathered informa- tion where they could, compared conflicting stories and numbers, and tried to make sense out of the often changing and usually unclear reality Preface xv of intergovernrnental relations in every country. The result, we think, is as accurate and solid a presentation of intergovernmental and subna- tional finance in Central and Eastern Europe and the former Soviet Union as is possible at this time. This book provides the starting point for all those who wish to fol- low us into this fiscal labyrinth. More important, we hope it will alert policymakers in transition countries and the many international experts advising them of the potentially critical role of intergovernmental finance in facilitating a successful transition to a market economy. As this book shows, important subnational fiscal changes are under way in virtually every transition country, and individuals concerned with ensuring suc- cessful transition from a command to a market economy neglect the intergovernmental dimension at their peril. Local finance is not a trivial matter that can be tidied up after the "big" decisions are made on such matters as privatization and national tax reform. On the contrary, how local fiscal matters are dealt with may often prove to be a surprisingly central determinant of the extent and success of such reforms. Richard M. Bird Robert D. Ebel Christine I. Wallich Acknowledgments This book is based on work funded and supported by the World Bank's research program, headed by Michael Bruno. It has also received support from the Infrastructure Division of the Europe and Central Asia/Middle East and North Africa Technical Department, the Central Europe Department of the Europe and Central Asia Region, and the Public Economics Division of the Policy Research Department. We are indebted to the World Bank for its support throughout the research program and in the preparation of the book, and especially to Wilfried Thalwitz, Kemal Dervis, and Anil Sood of the Europe and Central Asia Region for their support and encouragement. It was a special pleasure to work with all the contributors to the book, on a personal and a professional level. We thank Roy Bahl, Charles McLure, Jorge Martinez-Vazquez, Sally Wallace, Luca Barbone, James Hicks, Felix Jakob, Jennie Litvack, Gabor Peteri, David Sewell, and Peter Simon for making our collaboration with them both intellectu- ally stimulating and personally rewarding. Invaluable support and helpful collaboration were received by all authors from their colleagues in each country at the central and sub- national govemment levels, as well as in research institutes and acade- mia. In Albania we would especially like to thank Aleksander Meksi, prime minister of Albania; Zuher Beci and Malo Hizri of the staff of the Council of Ministers; Robert Ceku, vice minister of finance; Skender Dika, mayor of Shijaku; David Gentry, U.S. Treasury adviser to the Republic of Albania; and Gjerj Konda, director, Department of Economic Development and Aid Coordination, Ministry of Finance. xvi Acknowledgmietis xvii In Russia we thank Alexander Pochinok, chairman of the Tax Subcommission of the Russian Duma; Natasha Kalinina, Institute of Forecasting, Russian Academy of Sciences; Stanislas Korolev, vice minis- ter of finance, Russian Federation; Sergei Alexashenko, vice minister of finance, Russian Federation; Alexander Loubianoy, director, Territories Department, Ministry of Finance of the Russian Federation; Andre Sitnikov, Supreme Economic Council; Sergei Vasiliev, vice minister of the economy; Vladimir Lexin, Supreme Economic Council; and Victor Vlasov, Moscow Oblast Administration. In Hungary we offer special thanks to Aladar Madarasz, financial counselor to the Municipality of Budapest; Gabor Sz6kely, deputy mayor, Budapest; Nagy Lajosn6, Papp Ferencn6, and Bodor Janosn6 of the Budapest fiscal staff; Katalin Pallai, special adviser to the Municipality of Budapest; J6zsef Hegedus and Ivan Tosics, Metropolitan Research Institute (Varoskutatas), Budapest; J6zsef Thuma, deputy secretary of state, Ministry of the Interior; and Jozsef Sivak, director of the Local Governments and Regional Policy Department, Ministry of Finance. In Romania we thank Dima Petre, deputy general director, Ministry of Finance; Dana Craciunescu, project manager, Ministry of Public Works and Regional Planning, General Division of Municipal Services; and Illeana Pascal, director, Secretariat of National Center for Human Settlements (HABITAT). In Bulgaria we thank Plamen Petrov, head of the Finance Department, Sofia Metropolitan Council; Bojko Georgiev, head of the Local Budget Department, Ministry of Finance; Nikola Karadinov, minis- ter of regional development, Housing Policy and Construction of Bulgaria; Belin Mollov, head of department, Territorial Administrative Structure and Local Authority, Ministry of Regional Development and Construction; Yavor S. Stoyanov, technical assistance manager, Local- Government Advisory Group, National Center for Regional Development and Housing Policy; Anton Andronoff, mayor, Stara Zagora; James G. Budds, United States Agency for International Development (USAID) fiscal adviser to Stara Zagora; Evguenia Tzenova, Ministry of Finance; and Rossitza Vassileva, Central Tax Administration, Ministry of Finance. Among Polish experts we thank Michal Kulesza, formerly plenipo- tentiary for administration reform in the Council of Ministers of Poland, and the many participants in a seminar on reform of the state machinery held in Warsaw in January 1994. Among Ukrainian experts we would like to thank Victor Padalka, chief of the Financial Department, city of Kiev; Semen Moroz, head of the Budget Department, Ministry of Finance; Victor Tentyuk and Oleh Ilyn, deputy ministers, Ministry of Finance; Mikhail Progrebinsky, Center for XViii DECENTRALIZATION OF THE SOCIALIST STATE Economic Reform; and Robert Kravchuck and Jo Beth Mertens, U.S. Treasury advisers. Even this long list leaves out many who generously shared their expertise and insights, and from whom the book benefited greatly. Their knowledge of their country, its people, and the economy added much to the richness of our understanding. Among professional colleagues and friends closer to home, special appreciation also goes to Roy Kelly, International Tax Program, Harvard Law School and Harvard Institute of Intemational Development; Lionel Feldman of Lionel D. Feldman Consulting, Ltd., Toronto; Juliana Pigey, independent public and local government consultant; Susan Kosinski, United States Agency for International Development (USAID); Nicolas Levrat, Council of Europe; James Wooster, U.S. Treasury adviser in resi- dence at the Bulgarian Ministry of Finance; James Wozny, U.S. General Accounting Office; James Alm, Department of Economics, University of Colorado; Jeffrey Owens, Organization for Economic Cooperation and Development (OECD); Mathew Glasser, PADCO Consulting; and Andrew Pike, American University. Additional good advice, valuable comments, and insightful sugges- tions were received from many Bank and IMF staff and consultants. In particular, the work has benefited from the expertise and insights of M. C. Andrews, Nancy Birdsall, Mario Blejer, Charles Blitzer, Zeljko Bogetic, Robert Buckley, Milka Casanegra-Jantscher, Martha de Melo, William Dillinger, Dominique Dwor-Frecaut, Richard Florescu, Ved Gandhi, Kristin Gilbertson, Ellen Goldstein, Vincent Gouarne, Richard Hamilton, Timothy Heleniak, Yukon Huang, Dani Kaufmann, Timothy King, Krisztina Kiss, Jeni Klugman, Peter Knight, Michael Koch, Huda Kraske, Russell Krelove, Philippe Le Houerou, Fred D. Levy, Robert Liebenthal, Ira Lieberman, Johannes Linn, Mark Lutz, Charles McPherson, Costas Michalopolous, Helga Muller, Gur Ofer, Gulhan Ovaliaglu, Anthony Pellegrini, Remy Prud'homme, Bertrand Renaud, Andrew Rogerson, Naveen Sarna, Marcelo Selowsky, Anwar Shah, Carlos Silvani, Everardus Stoutjesdijk, Emil Sunley, Vito Tanzi, Teresa ter-Minassian, Margaret Thalwitz, Yoshine Uchimura, and Heng-Fu Zou. Outstanding research support was provided by Ritu Basu, Caroline Freund, and Mira Kuczynska. The manuscript was ably processed by Gemma Langton. We also thank our editors and compositors at American Writing Corporation, and the World Bank's Publications Department for their excellent work. Richard M. Bird Robert D. Ebel Christine I. Wallich 1 Fiscal decentralization: From command to market Richard M. Bird, Robert D. Ebel, and Christine I. Wallich A change of government is not a change of system, merely one of the pre- conditions for it. The change of system is a historical process that seems likely to require a long period of time. Janos Kornai, The Socialist System Dramatic reform is taking place in Central and Eastern Europe. New institutions and economic infrastructure are being created to provide the foundation for a pluralist, democratic society and a well-functioning market economy. The most discussed aspect of this reform is privatiza- tion-the move from a command economy to one of liberalized markets and free economic agents. A second aspect, equally critical to the transi- tion to a market economy, is the decentralization of government itself. Some type of subnational government structure existed in most of the transition economies under the socialist regime. But fiscal systems were highly centralized, with the subnational level acting as an administrative unit or department of the center with no independent fiscal or legislative responsibility. Komai (1992) sets out the former socialist system in consid- erable detail. It is revealing that this lengthy study hardly refers to the exis- tence of subnational state administrations, noting only that they are tightly controlled in all respects by the central bureaucracy. Subnational govern- ments were essentially deconcentrated units (or branch offices) of the cen- tral government and had little or no financial autonomy. This was true even in countries such as the U.S.S.R. (and Czechoslovakia) that were for- mally called "federations." Policymaking was controlled and centralized, and local government had virtually no independent tax or expenditure powers-its budget was seen only as the handmaiden of the central plan. I 2 DECENTRALIZATION OF THE SOCIALIST STATE Extensive political and fiscal decentralization is now under way in many countries in Central and Eastern Europe and the former Soviet Union. Politically, this decentralization represents both a reaction from below to the extensive central control of the past, and an attempt from above by the center to further privatize and to relieve its strained fiscal situation. Economically, decentralization is motivated in part by the recognition that it is imperative to utilize public resources more effi- ciently than in the past. Overcentralization characterized these countries' public administrations, just as it did the rest of economic activity. Decen- tralization, if appropriately designed and implemented, may lead to improved public service provision. Decentralization is thus a key dimension of the national transition from a command to a market economy. Like the broader economic tran- sition, it will require many, often difficult, reforms. Not only must the structure of tax and transfer mechanisms be reconsidered and expendi- ture responsibilities realigned among different levels of government, but views as to what governments can and should do must change. The total level of public sector activity must be dramatically reduced, but at the same time the new subnational governments must be allowed to build staff and institutional capacities in a manner that makes them account- able for their fiscal decisions. These ongoing reforms in subnational and intergovernmental finance are of considerable importance. Intergovernmental reform and the strengthening of subnational (intermediate as well as local) govern- ments are essential to support the evolving public and private sectors. Subnational governments account for a growing share of public sector activity in most of the transition economies as general government activ- ity is scaled back and subnational governments take responsibility for many services formerly provided by the central government.' In Russia subnational budget expenditures absorbed 38.6 percent of the consoli- dated national budget in 1992 and 42.9 percent in 1993. And local gov- ernment expenditures as a percentage of consolidated government expenditures in Hungary rose from 22.3 percent in 1988 to 30.4 percent in 1993 (table 1.1).2 The structure of intergovernmental relations is close- ly related to such critical policy issues as efficient resource mobilization, privatization, the social safety net, and stabilization. Within the fiscal sphere, for example, tax reform, deficit control, and intergovernmental finance are intertwined: if one of these elements is poorly designed, the entire fiscal structure may be compromised. Similarly, if the incentives built into intergovernmental arrangements are perverse, preserving cen- tral dominance or undermining the development of private markets, disillusionment with the reform process may set in, threatening both democracy and the market economy. Fiscal decentralization: From command to narket 3 Why local finance matters This chapter synthesizes the main themes and findings of the country- specific chapters that follow. The overriding theme is that a well- functioning intergovernmental fiscal system is key to most of the transition economies' major reform objectives. In all countries, intergov- ernmental finances affect the degree to which the public sector can mobilize resources. In addition, in the transition economies, intergovern- mental finances can also influence the achievement of macroeconomic stability, the provision of a social safety net, and privatization. And in countries with a limited tradition of self-government, developing effec- tive political and fiscal institutions at the subnational level may be criti- cal to building a nation. Systemic linkages In most countries even systematic discussion of decentralization and subnational finances requires only brief reference to such issues as stabi- lization, income distribution, and privatization. Fiscal decentralization in the transition economies, however, requires explicit recognition and examination of the interplay among these traditional functions of public sector activity. Stabilizationi. Subnational finances have important links with macro- economic stabilization policies in transition economies. Sound fiscal poli- cies (with supporting monetary policies) are key to macroeconomic stabilization. Large public sector deficits are being generated in many transition economies as falling tax revenues combine with the attempt to Table 1. I Central and subnational government expenditures, selected transition economies, 1993 Government expenditure as a percentage of GDP Subnational expenditure ------- -ias a percentage of general Country General Central Subnationol govemment expenditure Albania 44.C Bulgaria 41.2 - 99 24.1 Hungary 57.2 31.9 17.4 30.4 Poland 48.1 31.8 5.9 12.3 Romania 35.6 21.7 3.7 10.4 Russia 39.6 20.3 17.0 42.9 Ukraine 73.4 57.5 20.9 28.5 - Not ava able Source World Bank and var ous country sources, see table notes at the end of this chapter 4 DECENTRALIZATION OF THE SOCIALIST STATE continue the high spending practices common under the old regime. Subnational revenue and spending outcomes can support, or sabotage, efforts to control deficits and hence limit resort to inflationary finance. While of course policy reform in many areas is needed to meet stabiliza- tion goals, it is important not to neglect the intergovernmental dimension. In most transition countries, fiscal pressures have led central gov- ernments to use every instrument at their disposal in the attempt to con- trol the deficit. Some have taken steps to "shift the deficit down"-that is, to shift expenditure responsibilities onto subnational governments or to redirect revenues from the subnational level. Shifting the deficit onto subnational governments may increase rather than reduce macroeco- nomic pressures, however, since soft budget constraints are as common for governments as for enterprises in the transition economies. Subnational governments may respond to budgetary pressure by accu- mulating arrears, by unrestricted borrowing, by developing undesirable extrabudgetary sources of revenue, or by failing to provide needed ser- vices unless additional central funding is forthcoming. They may require the many enterprises they still own to finance public expenditures, espe- cially if the enterprises have easy access to bank credit. Pushing the deficit down may thus not reduce it but simply repress it or convert it into less transparent forms. Of course, subnational governments cannot on their own cause serious macroeconomic problems since they cannot print money. But quantitatively significant subnational deficits may in one way or another end up being monetized bv extending central bank credit to local enterprises, locally owned banks, or local governments, and hence may exacerbate the macroeconomic situation in a particularly nontransparent way. Inicomiie distribiutioni anzd the social safety net. Important parts of the social safety net such as social welfare, family assistance, and consumer subsidies, and in some countries pensions and unemployment, have been delegated to some extent from central to subnational governments in Albania, Bulgaria, Hungary, Russia, and Ukraine. It may sometimes make sense for subnational governments to administer certain social programs, but they cannot deliver such services effectively if they are chronically underfunded (Oates 1990). Where an adequate safety net is critical to a successful transition to a market economy, and its mainte- nance is therefore a national policy objective-as seems true in all the transition economies-subnational governments cannot reasonably be expected to take on these tasks without being provided with corre- sponding revenues, for example through intergovernmental transfers. The correct design of such transfers then becomes crucial to the success- ful achievement of national social policy goals. Fiscal decentralization: From command to market 5 Privatization. Privatization is also linked to subnational finance. Many countries have already transferred ownership of some state enter- prises to the subnational level. Subnational governments therefore have a major role to play in promoting or inhibiting privatization. But in many cases subnational governments still rely on "their" enterprises for rev- enues from profits and taxes. Not only would they lose some or all of these revenues if privatization occurred, they would also have to assume the substantial ancillary expenditures, assets, and their associated liabili- ties-for nurseries, housing, schools, vocational education, and so on- that will be spun off by the enterprises as they privatize. Owning assets that produce revenue is attractive to revenue-strapped communities, so privatization may be undermined unless the intergovernmental fiscal system accommodates these shifts in expenditure responsibilities, for example by providing the needed revenues in some other way. Continued local ownership of enterprises-and vested interests in their revenues-may also lead to domestic protectionism and the creation of interjurisdictional trade barriers to protect local monopolies, perhaps par- ticularly in countries as regionally diverse as Russia. Both the efficiency with which these economies perform and their future growth may be substantially affected by the intergovernmental fiscal system. Nation building. Finally, some countries such as Russia and Ukraine are explicitly "nation building," in the sense of attempting to counteract significant centrifugal forces at the subnational level. As experience has demonstrated in many less homogeneous countries, how the intergov- ernmental fiscal system works may prove to be crucial to national cohe- sion.3 The desire for more participatory democracy in the transition economies has heightened subnational demands for local political con- trol and unfettered local autonomy. Such demands are a natural out- come of the end of central planning. But the inadequate fiscal base is placing local aspirations and local fiscal capabilities increasingly at odds in some countries, giving rise to growing tensions. The choices are not easy, particularly in the more regionally diverse countries like Russia. Giving restless subnational governments the pow- ers, resources, and revenues to go it alone may placate them temporari- ly. But it may not be possible to go very far in this direction without seriously damaging national interests. On the other hand, failing to go far enough may lead to frustration and rebellion. In Russia, for example, separatist tendencies fueled in part by fiscal pressures and imbalances are causing fundamental political problems (see chapter 10). Regionally diverse countries may require some degree of decentralization to sur- vive politically: but deciding just how much, and how, to decentralize is always a matter of difficult political and economic calculation. Nowhere 6 DECENTRALIZATION OF THIE SOCIALIST STATE is this more true than in the formerly socialist economies, where many of the key factors affecting such calculations are, unsurprisingly, in transi- tion. Although this book necessarily focuses more narrowly on the design and development of intergovernmental fiscal arrangements, the importance of this wider political context must be kept in mind. Needfor a broader framework Intergovernmental fiscal relations, far from being purely a local matter, are thus key to reform goals in nearly all the transition economies. T raditional analysis of fiscal federalism examines the fiscal functions of subnational and central governments in terms of their respective (and largely separate) roles and responsibilities for stabilization, income dis- tribution (such as the social safety net), and resource allocation (Oates 1972; Musgrave 1983). This focus is not broad enough to address impor- tant aspects of local and intergovernmental finances in the transition economies for at least two reasons. First, as this literature has developed, this approach neglects the role of subnational governments with respect to stabilization, the social safety net, and privatization. Second, it does not fully address the legacies of a command economy that transition economies share and must address. Revenue sistems. Central governments faced with the need to reduce deficits (and reduce their recourse to inflationary finance) have tended to maintain the negotiated tax-sharing and intergovernmental transfer systems of the past. But this approach is becoming less acceptable in countries where demands for transparent regional treatment and equal- ization are strong and where subnational governments are seeking greater autonomy. It is also incompatible with the efficient provision of local public services, which requires a system of intergovernmental tax sharing and transfers that is both firm enough to serve as a basis for action and flexible enough to be compatible with ongoing structural changes and reform. The new subnational governments need to develop adequate rev- enue systems. Recent legislation in the transition economies has assigned specific taxes to subnational governments with the stated goal of reducing their dependence on the center for transfers.4 Invariably, however, subnational governments in most of these countries are still highly dependent on central transfers, including shares of central taxes, which are almost always allocated on the basis of the "origin" of the rev- enue, that is, to the jurisdiction in which the tax is collected. The tradi- tion of subnational governments receiving revenue from local public enterprises provides the historic basis for this practice of derivation- Fiscal decentralization: From commtnand to market 7 based tax sharing. Because of differences in regional and industrial structure and therefore the taxable base, however, the result is often to exacerbate disparities in access to public sector resources. Subnational revenue needs must be explicitly considered in nation- al plans for tax reform, since failure to consider the intergovernmental dimensions of national tax policy changes can lead to problems. In Russia and Hungary, for example, where a share of the major national taxes has been assigned to the subnational level, the central government has recently made major policy and design changes to these taxes with- out considering their impact on subnational budgets. Such changes may wreak havoc on the aggregate subnational revenue flow and have erratic effects across subnational jurisdictions. Despite the focus on autonomy in the new laws on local self- government (table 1.2), no transition economy provides its subnational governments with much fiscal discretion in the sense of permitting them either to set their own tax rates or define the tax base. Both tax bases and rates continue to be determined by the center, even for such minor levies as a dog tax. No country has assigned a major revenue base to the sub- national level-although, as discussed below, many share important national revenue sources with subnational governments. The paucity of locally controlled tax resources in most transition countries, combined with the reluctance of local officials to exercise their taxing powers even when they have them, makes it almost inevitable that hard-pressed local governments will turn to other means of gaining revenue. They will demand increased transfers, attempt to borrow, or exploit the assets they have acquired during the process of decentralization and privatization. Table 1.2 Laws on local self-government, selected transition economies Country Nome of low Dote odopted Albania Law 7572 on the Organizatior June 1 992 and Operation of Local Government Bulgaria Law on Local Self-Government September 1 99 1 and Local Administration Hungary Law 65 of 1 990 on Local August 1 990 Self-Government Poland Law on Local Authorities March 1 990, amended May 1990 and December 1 990 Romania Law 69/1991 on Local November 1991 Public Administratior Russia Law on the Rights of Local July 1 992 Self-Governments Ukraine Law on Local Self-Government 1992 Source: Council of Europe 1994. 8 DECENTRALIZATION OF THE SOCIALIST STATE For decentralization to realize its potential in economic terms, however, subnational governments should, at the margin, both have to raise their own revenue and be in a position to do so (Bird 1993).5 Public sector pricing. Another problem arises with regard to public sector prices (user charges). In the past, central governments in the socialist countries used low and fixed prices (including rents and other urban user fees) and wage controls as part of their income distribution program. Market reforms require the new governments in these coun- tries to get the prices right and to use tax policies and targeted subsidies to redistribute income. Such reforms raise special problems for subna- tional governments in some transition countries, however, because the central government continues to mandate many of the prices of services for which they are responsible, such as rents, transport, and water. Expenditures. A more systematic approach is also needed for public spending. The current strategy of devolving expenditures to the subna- tional level while holding back on transfers is unlikely to be successful. Indeed, net expenditure reductions at the subnational level may prove dif- ficult to achieve. As table 1.1 shows, the subnational sector has significant expenditure responsibilities in many transition economies. In addition, state enterprises still undertake a wide range of social sector outlays that may have to be taken over by subnational governments. Subnational gov- ernments are also shouldering much of the burden of replacing and refur- bishing obsolete and poorly maintained physical infrastructure. In relative and perhaps even absolute terms, subnational spending thus seems more likely to grow than to shrink in most transition economies. Transfers. The level, design, and effects of intergovernmental fiscal transfers constitute key elements of the emerging intergovernmental and local finance systems of the transition economies. There are few formula- based transfers. Most transfers continue to be negotiated, both in aggre- gate and for individual localities. Decisions need to be made regarding the overall size of such transfers (the distributable pool) and the distrib- ution formulas that allocate transfers across localities. Both the severe fiscal pressures on central governments in the transition economies and the vital tasks to be performed by subnational governments in the emerging structure of the public sector need to be taken into account. Practical measures of revenue capacity, expenditure needs, and perhaps fiscal effort need to be developed for use in such formulas. Central and local institutions must be developed to ensure that transfers are put to the best possible use, particularly in the smaller municipalities. Such institutions might include, for example, a central agency-or some form Fiscal decentralization: From command to market 9 of intermediary body-to monitor the performance of local governments and to provide them with needed technical assistance and support. Borrowing. Borrowing by subnational governments is also a matter of concern. As discussed later, some countries have granted virtually unlimited borrowing powers to subnational governments, apparently without considering the possible macroeconomic repercussions of fail- ing to monitor and control such borrowing. Although the undeveloped state of municipal credit markets makes abuse of this borrowing power more a potential than a real problem, in some countries (such as Bulgaria) the local debt problem is being "solved" by not honoring pay- ments due to suppliers, creating arrears. State-ozned assets. The potential for misuse of locally owned state assets is already apparent. Such assets may represent a windfall-at least in those cases where they are not accompanied by central policies, such as rent or price control, that turn them into a liability (Alm and Buckley 1994). But they are often misused, whether privatized or not. A local dis- trict heating plant that is privatized but remains a local natural monopoly raises regulatory problems. Who deals with these problems, and how? What should be done with the proceeds of privatization? When should such proceeds be used to finance recurrent budget expenditures? If an asset is not privatized and is used to generate recurrent revenues for the subnational government, the temptation to sustain and strengthen the monopoly position is again a problem. Another risk is the high probability that many speculative ventures will fail. And many subnational govern- ments are seeing scarce managerial skills dissipated in efforts to squeeze money out of such ventures rather than focusing on the more essential government task of providing local public services efficiently. In any system fiscal decentralization is invariably an ongoing and dynamic process. Reformers in Central and Eastern Europe and the for- mer Soviet Union, however, face a special challenge in having to distin- guish between problems that are structural and those that stem from the transition to a market economy. They must design an intergovernmental framework that is firm enough to serve as a basis for action, for example, regularizing transfers and tax flows, but still flexible enough to coexist with the ongoing structural adjustments in the economy relating to sta- bilization, redistribution, and privatization. Recent trends in decentralization One of the most striking features of the political and fiscal decentraliza- tion in the transition economies is how recently it was implemented and 10 DECENTRALIZATION OF THE SOCIALIST STATE how quickly it has proceeded. Legislation on local self-government has been introduced in every country in the region (see table 1.2). In most countries, this legislation was introduced shortly after the breakup of the Soviet bloc, reflecting the importance of local autonomy movements in countries emerging from the central command system. Only Russia, the single formally federal state, has addressed its center-state relations in a constitutional context, with the 1994 Constitution defining the powers of Russia's eighty-nine oblasts (intermediate-tier governments) vis-a-vis the federal government, and equalizing the status of all intermediate governments vis-a-vis each other (see chapter 9). Clear assignment of expenditure responsibilities and taxes to dif- ferent levels of government is desirable to promote efficiency, equity, and accountability, and to ensure predictability. On the whole, how- ever, given the difficulty of amending constitutions and the evolving nature of these economies, it is likely that most countries will continue to take the legislative rather than the constitutional route with respect to decentralization. Expenditure and tax assignments should be clear and fair, but they do not need to be permanent: flexibility within a well- defined structure is likely to be a better approach than prematurely attempting to lock all the ingredients of the shifting economic and fiscal puzzle in place. A time of tranisfornatiott The local self-government laws listed in table 1.2 represent a major departure from the past. Theoretically, they guarantee the political, administrative, and financial independence of subnational authorities. Often guided to some extent by the Council of Europe's model legisla- tion, these laws address issues ranging from the guarantee of free and open elections and the protection of political boundaries to the nature of central limitations on administrative oversight and the right of local authorities to cooperate and form associations (Council of Europe 1985). Typically, the legislation affirms decentralization, local financial autono- my, and the administrative independence of the subnational level from central control. The language of the laws differs, but the general intent appears to be to free subnational governments from centralized control and allow local democracy to flourish. Most countries have also enacted additional legislation on subna- tional revenue sources, the budget process, and the like. In Hungary, for example, which is further along in this process than most other countries in the region, laws have been passed on local taxes and fees, local elec- tions, and the self-government of the capital city and capital city dis- tricts. Nonetheless, key elements of the legal framework needed for Fiscal decentralization: From commnand to market I I efficient subnational government are still missing in most countries, and critical questions-such as the assignment of expenditure responsibili- ties and the role and nature of institutions intermediate between central and local governments-remain undecided. Moreover, no country has yet established an adequate political or administrative structure for designing and guiding the decentralization process further. For example, responsibility for local affairs in Hungary is split between the Ministry of Interior and the Ministry of Finance. In Russia the formation and implementation of subnational fiscal policies is the responsibility of the Ministry of Finance and four parliamentary (Duma) subcommissions. The State Tax Service (administration) also has important intergovernmental fiscal responsibilities. In Bulgaria the regional governors are appointed by the Council of Ministers, while the municipal level is managed by municipal councils. In Romania the for- mulation of decentralization policy and implementation of local gov- ernment reform is the responsibility of both the Department of Local Government Affairs of the Office of the Prime Minister and the Department of Tax Administration and Local Budget Management of the Ministry of Finance and Economy. In contrast, Poland and Albania have some unified leadership since they have granted a coordinating function to the Council of Ministers and have established an intergov- ernmental task force (which in Albania's case is temporary). Reorganizing governmlent Decentralization is a broad term encompassing several distinct govern- mental arrangements. DevoIution is the fullest form of decentralization: independently established subnational governments are given the responsibility for delivery of a set of public services along with the authority to impose taxes and fees to finance the services. Devolved governments have considerable flexibility to select the mix and level of services to provide. Some financial support may be provided by the central government. Deconcentration refers to the decentralization of central government ministries. Deconcentration with authority means that regional branches of central government offices are created with some ability to make independent decisions. Deconcentration without authority occurs when regional offices are created with no independent capacity for decision- making. All deviations from normal practice must be approved by the center. This sort of administrative decentralization within the central government is not discussed further in this book. Delegation is intermediate between devolution and deconcentration. Subnational governments (not branches of the central government) are 12 DECENTRALfZATION OF THE SOCIALIST STATE given responsibility for delivering certain services but are subject to more or less detailed supervision by the central govemment, which also often provides some or all of the finance for the services in question, usually through some form of fiscal transfer. In principle, devolution should increase the efficiency of service delivery by allowing citizens to determine more directly the mix and level of services and by providing greater incentives for local resource mobilization. People are more willing to pay for services closely attuned to their demands, and local officials are, it may be argued, more readily able than central bureaucrats in some distant capital to determine what they are able and willing to pay for. Such efficiency gains from decen- tralization benefits may be particularly significant in countries with diverse economies, geography, and tastes for services.7 On the other hand, delegation may be efficient when subnational governments can better administer programs of national interest (educa- tion, health) in ways that better match local economic and social circum- stances. The central government determines, in essence, how much should be spent in these areas and perhaps sets minimum or standard levels of provision, but the detailed decisions on exactly what is done are to some extent decided by local govemments. The detailed design of intergovernmental fiscal transfers and the degree and nature of central government monitoring may critically determine the balance of central and local influence in such delegated areas of responsibility. In transition economies, few of the potential benefits of decentral- ization-whether by devolution or delegation-have yet been attained, for two reasons: the persistence of old methods and flaws in the new dispensations. In Romania, for example, although the Public Finance Law grants subnational governments budgetary autonomy, it also gives the central government the power to change local budgetary alloca- tions.8 Moreover, deference to central commands, even if no longer legally binding, persists in many areas. As discussed in chapter 6, local governments in Romania deliver few services themselves. Instead, they contract with local public enterprises (regies autonomes) to deliver most local public services. While a 1990 law9 makes these enterprises respon- sible to a board composed of local officials, the enterprises appear to respond more to central than to local control. Finally, as in most of the transition economies, Romania has a large number of small, poor local governments with almost no resources of their own. Autonomy does not mean the same thing to a small, poor rural commune as it does to a large, relatively rich city. Similar problems-conflicting legal interpretations, tangled rela- tions with enterprises, and wide diversity in local government capacity -exist in varying degrees in most of the transition economies, casting Fiscal decentralization: From command to market 13 doubt on the significance of local autonomy. Even where such problems are absent, local autonomy may not be very meaningful if it is not accompanied by fiscal capacity-and it may not be desirable unless it is accompanied by some degree of fiscal responsibility. To make local autonomy meaningful, subnational governments need adequate locally controlled revenues. Decisions about the provision of services with primarily local benefits must be made by local govern- ments, and these governments must have the resources to carry out such decisions. As shown in detail in the country chapters, subnational expen- diture discretion is still limited in important ways in all the countries cov- ered here. Some of these limitations on local discretion may be useful, but others clearly are not. In most countries the subnational revenue base is inadequate and largely centrally controlled, with the result that subna- tional governments remain undesirably dependent on ad hoc central transfers. Greater spending discretion for those expenditures assigned to the subnational level and greater flexibility for subnational governments to raise their own revenues are thus needed in most countries if subna- tional governments are to be made more accountable to local residents and local public services are to be delivered more efficiently. To make local autonomy desirable, subnational governments must not only be accountable to local citizens for spending from revenues raised from them; they must also be accountable to the central govern- ment to the extent they are acting as central agents and are financed by transfers. One way to make local autonomy both meaningful and desir- able is to provide technical assistance and training activities that enhance the capacity of subnational governments to deliver services effectively and to generate more own-source revenues. Another is to design grants, financial reporting systems, and accountability mechanisms to this end. The structure ofgovernment Before the reform period all the transition countries surveyed here except Albania had a multitier organization of government, with the central level implementing its control over the local level through an intermediate tier of government. The central government was respon- sible for the overall plan and budget, and the intermediate level-called oblast in the Russian Federation, Ukraine, and Bulgaria, judet in Romania, megye in Hungary, and voivod in Poland-was the agency of administrative control through which the center dictated and imple- mented policy. This middle tier oversaw the expenditures of local gov- ernments and was the channel for central fiscal flows to the localities. However, from a fiscal perspective, none of the countries discussed in this book was truly organized as a federation. 14 DECENTRALIZATION OF THE SOCIALIST STATE Russia currently has a three-tier system of government, comprising the central (or federal) government, the intermediate-level regions, and the municipalities (table 1.3). Although the legal structure suggests that Albania, Bulgaria, Hungary, Poland, and Romania have a two-tier gov- ernment structure (central and local), matters are not so simple. Romania, for instance, has two tiers of local government (the judets and the municipalities). All the area of a judet is divided into municipalities, but in principle each type of local government is independent of the other. The same is true in Albania. In Bulgaria and Hungary the old intermediate tiers continue to exist, but they have so few functions that in effect there are really only two levels of government, central and local. On the other hand, in Budapest (Hungary), Sofia (Bulgaria), and the cities of Russia, there are subcity (district or ward) councils and governments with elected councils and their own budgets and assets, so the "local" level of government itself has two tiers. The average population size of these middle-tier subnational gov- ernments varies widely, often reflecting the size of the country itself (table 1.4). The average size of a Bulgarian oblast is about half that of a Russian oblast and roughly two-thirds larger than the Romanian judet and the Hungarian county. In all countries, however, there is consider- able dispersion in the size of both intermediate and local governments, as shown in table 1.4. Table 1.3 Structure and organization of government in selected transition economies Prereforma Emerging structure Unitory/ Unitory! Country federol Number of tiers federal Number of tiers Albania Unitary 1: Central Unitary 2: Central and bashki/communa Bulgaria Unitary 3: Central, regional. Unitary 2: Central and obshtini and local Hungary Unitary 3: Central, megye, Unitary 2: Central and helyi onkor and local manyzat Poland Unitary 3: Central, voivodship, Unitary 3: Central, voivodship, and local and gminy Romania Unitary 3: Central. regional. Unitary 2: Central and and local judet/municipii Russia Federalb 4: Central, Union republics, Federal 3: Federal, oblasts, and oblasts, and rayons rayons Ukraine Federalb 4: Central, Union republics. Unitary 4: Central, oblasts, rayons. oblasts, and rayons and rural settlements a. The pencd before passage of relevant act in table 1.2. b. Refers to the Sov et period. Source: Information is drawn Forr this volume. Fiscal decentralization: From command to market 15 A rolefor a middle tier? Few questions are more in flux in most transition countries than the role of intermediate levels of general-purpose government. Most countries seem to associate this intermediate tier with the former control economy and want to eliminate it. In some countries, the intermediate tier is limit- ed to overseeing the constitutionality and legality of local government operations. In others, its main role is to coordinate central government policy at the regional level. In Hungary, for example, there are now 8 regions, 19 counties, and 3,148 localities. The newly created regions, headed by a central office called the Commissioner of the Republic, have only two functions: infor- mation gathering and dissemination, which involves collecting data on local government activities and providing information to local govern- ments to facilitate the implementation of national policies; and monitor- ing of local government actions (such as taxes) to ensure they are constitutional. The counties, formerly the main territorial control instru- ment of the state, have had their functions severely restricted. As dis- cussed in chapter 2, counties retain only some minor fee revenues and are responsible in principle only for interjurisdictional functions that serve several localities. On the other hand, localities have extensive expenditure responsibilities, the right to spend national transfers and shared taxes as they see fit, and almost unlimited power to borrow, own, and dispose of property and to manage, establish, or sell public enter- prises. In response to local pressures for greater self-determination, the number of local governments has more than doubled from the 1,523 local councils in place before local government reform. Romania's situation is quite different (see chapter 6). Although the forty-one judets of the old order continue to exist, there are now two parallel administrative structures in each judet. One, headed by an appointed prefect, is essentially the deconcentrated territorial adminis- tration of the center. The second, headed by the elected president of the judet council, is the judet government itself, which is responsible for "works of regional interest," but which, like the Hungarian counties, has no hierarchical authority over the 2,948 municipalities. In the popu- lated urban areas, the municipalities have the lead role. In areas charac- terized by small towns and villages, however, the judet is the coordinating entity. And in some areas judets and municipalities share responsibilities, dividing local and interlocal aspects-although it is not always clear just what this means. Poland's forty-nine voivodships are not really subnational govern- ments: rather, they are deconcentrated public service intermediaries for certain central expenditures (mostly in health and social welfare), 16 DECENTRALIZATION OF THE SOCIALIST STATE although other central expenditures are carried out locally by line ministries. All local government functions are carried out directly by the 2,549 gmina governments. As discussed in chapter 4, a new intermediate tier of government (probably about 300 units called powiats) has been proposed but not yet been approved or implemented. In the Russian Federation a different pattern is emerging, with the oblast level taking on a larger role and emerging as an important unit in its own right. Indeed, if the fiscal reforms at the subnational level pro- posed in recent legislation (see chapter 9) are implemented, the Russian Federation will have decentralized substantial expenditure responsibili- ties and, to a lesser degree, revenues to the oblast level and in many ways might take on the character of a true federal state. To date, however, the eighty-nine autonomous regions, oblasts, okrugs, krais, and republics have only nominal autonomy: they are neither fiscally independent nor legally able to determine their own expenditure and tax policies. In contrast, as discussed in chapter 8, the government structure in Ukraine remains strictly hierarchical. It has four tiers, with the central Table 1.4 Type, size, and number of subnational government units, selected transition economies, 1990-94 Type of Country government/unit Name Number Albania District Rethe 37 Mu.nicipality/commune Bashki/communa 356 Bulgaria County Oblast 9 Municipality Obshtina 255 Hungary County Megye 19 Municipality Helyi onkormanyzat 3.148 Poland Province Voivodship/wojewodztwo 49 Municipality Gmina 2,459 Romania District Judet 41 Municipality Municipii/oras/comuna 2,948 Russia Province Oblast 89 Municipality/county Rayon 2,000 Ukraine Province Oblast 27 Municipality Rayon/city 480 rayons and 139 cities Rural settlement and city distnct 30,000 - Not avai able. Fiscal decentralization: From command to market 17 government at the top followed by 27 oblast-level governments, 619 rayon-level governments, and about 30,000 rural settlements and city districts. Tax administration is performed for all levels by a central government agency. Although each level of government formulates its own budget, they are not autonomous units because the share of rev- enues that each level of government receives is determined by the gov- ernment directly superior to it. Albania seems, on paper, to have three layers of government: 356 conimunes and municipalities, 37 district governments, and the central government. But in fact, the districts have no hierarchical authority over the municipalities and the communes. A draft constitution speci- fies the role of a judiciary designated to settle intergovernmental dis- putes and define the relationship between different levels of government. Although Albania has begun to decentralize, problems have arisen (as discussed in chapter 7) because current laws remain vague on how local governments are supposed to finance and coordi- nate their hefty responsibilities. Capital city: population: Population (percentage Country of total urban Averagea Minimum Maximum popuiation population) 96,1 18 23,789b 414,3670 3,400,000 Tirana 8,862 1 ,292b 270,000b 270,000 (21) 843,000 653,285c 1,289,99 1 8,900.000 Sofia 29,753 1,818' 5 1,068' 1.313.000 (20) 430,474 212,000 949,749 10,300,000 Bucapest 2,598 <200 243,300 212,000(31) 738,673 3,900 2,235,000 38,400,000 Warsaw 15,616 858 848,51 4 2,235,000 (9) 499,976 237,900 2.201,000 22,700,000 Bucharest 6,954 <500 2,036,900 2,201,000(18) 1,572,494 25,000 9,048,000 149.000,000 Moscow 69,976 9,048,000 (8) 1,831,926 2,638,000 52,100,000 Kiev 79,906 - 2.638,000 (8) 1,649 - Note. For footnotes and sources, see table notes. 18 DECENTRALIZATION OF THE SOCIALIST STATE Is the size of government units efficient? The reaction against intermediate governments as undesirable reminders of the old regime is understandable, and so is the desire to replace them with something more acceptable. Most of the transition countries have created a large number of small and probably not fiscally viable local government units (see table 1.4). The average population of local governments in Hungary (excluding Budapest) is about 2,600. In Romania half the population lives in local government units with fewer than 3,000 people; such units account for 90 percent of all local govern- ments. In Bulgaria the average population is 28,000 (excluding Sofia), but one-fifth of its 255 municipalities have populations of fewer than 5,000. And although Albania and Poland's local governments are less fragmented, many are arguably too small to provide services efficiently. Fragmentation may place government near the people, but it also complicates intergovernmental relations.10 The local governments in many transition economies are simply too small to provide efficiently all the public services theoretically assigned to them. Moreover, it is practi- cally impossible for the central government to maintain direct bilateral relations with so many units, and the extensive fragmentation may reduce the political voice of local government as a whole. Even if formal intermediate tiers of government are not created, some mechanism for coordination and consultation is likely to prove necessary. Possible models abound, such as the municipal federations found in the Nordic countries, under which localities organize on a voluntary basis to pro- vide certain local services, such as transport for commuters and certain types of education (Soderstrom 1991). For example, Finland's 460 local governments are organized on a voluntary basis into 100 municipal fed- erations to provide health care (soon to be reduced to 21). All the Nordic countries have "secondary communes," county-like bodies with their own elected officials, which are responsible for providing specified ser- vices to an area encompassing several localities. Similarly, in the United States there are many "special districts," entities with substantial admin- istrative and fiscal independence created to carrv out such functions as flood control, irrigation, fire protection, and water supplv over areas encompassing several local governments. Given the small size of local governments in many transition economies, there would seem to be a strong case for encouraging volun- tary cooperation (for example, through the grant system) or for assigning some services (secondary education and hospital care) to larger special- purpose units. Alternatively, a case may be made for a more effective intermediate tier of general-purpose government in some countries. As shown in the detailed country discussions later in this book, the whole Fiscal decentralization: Fromn command to market 19 question of the appropriate role and design of intermediate subnational governments remains very much up in the air in most countries. The changing role of government: From ownership to service provision In the command economies, governments at both the national and sub- national levels were involved in almost every aspect of economic activ- ity. The state owned all enterprises, from steel plants to the local laundry. Central government ministries organized production, allocated labor, planned exports, and so on. The first stage of decentralization in some countries-Hungary led the way-began as early as the 1960s, when ownership of some smaller industrial units and the retail sector were transferred to subnational governments (Jeffries 1993). Although recent reforms aim to decentralize ownership to the pri- vate sector, the transition is far from complete. All the countries studied here appear committed to privatizing state assets, liberalizing prices, and removing from the state its role as direct producer and provider of economic goods; but many productive assets-industrial assets, agricul- tural and urban land, housing stock, commercial property-remain in government hands, often those of local governments. Under the socialist system, enterprises owned by local governments had their investments financed from local budgets, and local governments derived revenues from the operations of enterprises they owned. Budgetary distinctions between enterprise functions and government functions were unclear. Government often asked enterprises to construct and finance such facili- ties as roads, schools, clinics, and even sports stadiums and clubs; "donations" to the budget for onetime outlays were also common. It made little difference whether such outlays were undertaken directly by the enterprise or through the more circuitous route of profit transfer and funding through the government budget: the burden on the enterprise and the net budgetary position was the same. All this has now changed, or is in the process of changing. Transforming the role of national and subnational governments in transition economies is a major challenge. In principle, subnational gov- ernments should become service providers and regulators, not owners. The wrenching reforms most countries are undertaking to achieve macroeconomic stabilization and to compress the state budget could produce perverse effects at the local level, however. Unless an adequate framework for intergovernmental finance is put in place, budgetary con- straints may create additional hurdles to disassociating local govern- ments from their enterprises. Subnational governments will be tempted to rely on locally owned enterprises to provide public services and to protect locally owned enterprises to ensure their profitability. 20 DECENTRALIZATION OF THE SOCIALIST STATE The regulatory tasks facing both subnational and central govern- ments have changed. Administrative interventions, such as the detailed regulation of prices, the administration of waiting lists for housing, and the allocation of land and material supplies, are no longer appropriate. What is needed is a regulatory structure that provides the conditions under which the market can function, such as regulation to avoid monopoly, prevent food adulteration, curtail financial fraud, promote transparency and consistency in zoning and land use, and so on. The national government must establish a clear legal framework-setting, for example, environmental standards-not only for the private sector but also for the decentralized local public sector. Subnational govern- ments may also need to regulate areas such as land use, business licens- ing, sanitation conditions, and localized environmental problems (Oates 1990). To the extent these governments are competitors as well as regu- lators of the private sector, their actions will be suspect. Macroeconomic dimensions of intergovernmental finances Intergovernmental financial reforms are taking place in a constrained fiscal context, at a time when tax reforms and major changes are taking place in national revenue systems. Current stabilization concerns may in some instances unduly dominate the design of a sound, long-term inter- governmental system. Simultaneity of national tax reform and intergovernmtental reform Fiscal decentralization is taking place at the same time as national tax systems are undergoing major changes. National tax reforms have been initiated in almost every country in transition to make tax systems more compatible with a market economy and to provide a secure source of revenue for essential state budgetary outlays (Gray 1990; Tanzi 1991, 1993a, 1993b). More recently, some tax changes have been driven by the desire of some transition economies to harmonize their fiscal regimes with the European Union. While there are considerable variations of detail from country to country, most national tax reforms have involved: - Replacing profit remittances to government as enterprise owner with corporate income tax (CIT) to government as tax collector. * Introducing a value added tax (VAT) to replace the socialist "turnover tax," a cascading tax which consisted of thousands of commodity-specific rates that acted as wedges between adminis- tratively set retail and wholesale prices. * Introducing a personal income tax (PIT) to replace the wage con- trols that served as the implicit tax on labor income prior to reform. Fiscal decentralization: From command to market 21 * Introducing property and land taxation. * Introducing or revising payroll taxes to finance social security, unemployment, and health systems. * Eliminating a number of taxes aimed at regulating enterprise behavior, such as the excess wage tax, wage bonus tax, and excess profits tax. The simultaneous implementation of decentralization and national tax reform has often provided little time to anticipate the effects of rev- enue reform on subnational finances. In Hungary, for example, the PIT and VAT were introduced in 1988, two years before the Law on Local Self-Government was passed, and the 1988 Budget Act determined that a share of the new personal income tax-the yield of which was unknown-should be the major revenue source for local government under the old system. In Russia revenues from both the just-introduced PIT and the VAT were assigned to the oblast level in 1991, again with little basis for judging revenue yields or adequacy, and before spending functions were clearly assigned. Major changes have since been made to both the VAT (extending the base to imports to increase the yield, and changing the rates) and the PIT (tripling the personal exemption, there- by halving oblast PIT revenues), with profound effects on oblast bud- gets. In Ukraine, as in Russia, major new taxes are being introduced at the same time that local self-government laws are assigning them to or sharing them with oblast and local governments. Despite the fact that subnational spending accounts for more than 40 percent of total public sector outlays and one-fifth of GDP in some countries (see table 1.1), in most cases national tax reforms do not seem to have taken the fiscal needs of subnational governments into account. Given the inadequacy of the "own source" revenues specifically assigned to subnational governments in most transition countries, these governments are likely to remain dependent on shares of central taxes (or transfers) for years to come. National tax policy is also subnational tax policy, and this fact should be more explicitly taken into account than seems to have been true so far. The mlacroeconomzic contextfor decetralization Transition economies have had to address stabilization and liberaliza- tion concerns simultaneously. Intergovernmental fiscal reforms are tak- ing place in a weak macroeconomic context, often coupled with a weak central government fiscal position (table 1.5). In Hungary in 1990, the year the new system of local self-government was introduced, the con- solidated deficit of the general government was 4 percent of GDP, with inflation at 32 percent. In 1992 the deficit was 7 percent, and by 1994 it 22 DECENTRALIZATION OF THE SOCIAI.ST STATE had risen to an estimated 9 percent of GDP, with local and national needs competing in an environment of increased fiscal austerity. In the Russian Federation the fiscal deficit was more than 15 percent of GDP in 1991 and 1992, it remained high at nearly 8 percent of GDP in 1993, and a deficit of more than 10 percent of GDP has been estimated for 1994; infla- tion also remains high." Bulgaria's deficit in 1993 was almost 10 percent, and Albania's was even higher, at 16 percent. In Romania, although the budget deficit has been small, the high inflation rate reflects an underly- ing instability, and it seemns likely that fiscal accounts understate the deficit.12 Such pervasive fiscal constraints have had a major impact on the design of intergovernmental fiscal systems. The move to private entrepreneurship has contributed to a more vital economy, but it has also reduced central government revenues. Many smaller private enterprises elude the tax net, and the burgeoning informal economy presents major challenges to tax compliance and tax administration. State-owned enterprises no longer provide a secure rev- enue base. Tax reforms are desirable for market efficiency, but it will take time for tax administration to adapt to the radically new environ- ment. Budget balance is also complicated by the fact that the expendi- ture budget in most countries continues to be burdened by outlays for subsidies, generous cash benefit programs, and new and sometimes large support of enterprise and bank restructuring and recapitalization. Intergovernmental reforms are taking place in the context of a con- centrated effort to reduce the size of government. In Hungary expendi- tures shrank from about 62 percent of GDP in 1988 to about 57 percent in 1993. In Russia the general budget was more than 60 percent of GDP in the late 1980s and about 40 percent in 1993; the target for 1994 was just 20 percent. In Albania spending fell from 62 percent of GDP in 1991 to 44 percent in 1993 (see table 1.1).13 Table 1.5 Fiscal balance in selected transition economies, 1993 (deficit/surpsuosa a percentage olFGDP) Genera' Centroa SubnGtuonal Country government governmenr government Albania -16.0 -16.0 -- Bulgana -9.2 -9.2 0 0 Hungary -6.0 5.15 -0 3 Poland -2.5 -2.9 0.0 Romania -0.4 --1.5 -1.7 Russia -7.9 -6.7 0.6 Ukraine -6.2 -6.2 - Not available. Note General government may nclude budgets of agences and autonomous entities and is therefore not, in a I cases, the sum of central and subnational government shares. Source: World Bank and various country sources: see table notes. Fiscal decentralization: From command to market 23 Not surprisingly, stabilization concerns often dominate the national and intergovernmental agenda. Reducing fiscal imbalances, both at the center and at the subnational level, is fundamental to the adjustment programs of most transition countries. The current situation is in sharp contrast to the prereform period, when strict central controls ensured subnational finances had few macroeconomic consequences (Blejer and Szapary 1989; Bahl and Wallich 1992). Under the old system, revenue sharing served only as an administrative device to simplify a system of central resource allocation, and expenditures were guided by planning norms. The result was that the budgets of subnational governments were always in balance, in an accounting sense at least, with required adjustments made simply by transfers from the central budget. The new concern for macroeconomic stabilization has led central governments to view fiscal decentralization as an opportunity to reduce central expenditures in two ways. First, "spinning off" expenditure responsibilities to the subnational level reduces the central deficit. Second, reducing fiscal transfers, purportedly to make subnational gov- ernments more independent, also reduces central outlays. Shifting expenditure responsibilities. In some countries budgetary responsibility for social expenditures and the social safety net is being transferred to subnational government. In Hungary, for example, respon- sibility for welfare expenditures was transferred to the localities in 1993 under the Social Assistance Law. In Ukraine, too, the social safety net is a subnational responsibility. In Russia the central government transferred social expenditures equivalent to some 6 percent of GDP to localities in the 1992 budget, in effect pushing the deficit down. The hope seems to have been that subnational governments would perform the politically painful cutting required, even though the demand for these services is likely to grow with the worsening economic situation. And in 1993, again in Russia, responsibility for key national, interjurisdictional investments (such as in transport) was transferred to the subnational sector. Even some of the asset transfers to subnational governments that have occurred appear to have been motivated partly by budgetary con- cerns. Some of the transferred assets-notably housing and some enter- prises-are really liabilities given the heavy burden of maintenance and operation of these units and the fact that rental income (adjusted in 1993 in the Russian Federation for the first time since 1928) does not cover even a small part of costs (Alm and Buckley 1994). Reducing intergovernmental transfers. Fiscal difficulties have led some central governments and their finance ministries to focus on reducing intergovernmental transfers. Such reductions are often promoted on the principle of budgetary independence-that is, that 24 DECENTRALIZATION OF THE SOCIALIST STATE subnational governments should be financially self-sufficient- although they have not been matched by giving those governments any additional revenue-raising authority. A good case may be made for reducing the old discretionary trans- fers. This system perpetuates the socialist structure of deconcentration and works against the development of responsive and efficient subna- tional governments. In practice, however, most transfer reductions have not been accompanied by any attempt to rationalize the system through a reassignment of tax bases to subnational governments, or to efficiently design and target grants (for example, along the lines sketched later in this chapter). In all countries central transfers to subnational governments remain large (table 1.6), reflecting both the center's reluctance to give up a tool for controlling subnational governments and the failure of some local authorities to modernize and strengthen the limited revenues they do control. In Romania and Bulgaria, for example, the central government has maintained its political control and thwarted local creativity through its control of old line institutional mechanisms (the judets in Romania) or simply by coming down with a heavy hand (Bulgaria): see chapters 5 and 6 for further discussion. In Hungary, however, part of the problem appears to be with some local officials who behave as if the laws on local self-government and local taxes will implement themselves and who put off the need to act today to strengthen, say, the property tax, which could be a major source of revenue five years from now. Improving overall budgetary outcomes While budgetary flexibility is clearly desirable from the central govern- ment's short-run perspective, transfers should not be seen as a compress- ible part of the national budget, as they are in some countries. Many of Table 1.6 Local government revenue sources, selected transition economies, 1993 (percentoge of government revenue) Nontcx/other Tronsfers and Country Own tax revenue shored revenue Albania 0.0 0.0 100.0 Bulgaria 1.2 4. 1 94.7 Hungary 4.0 19.9 76.1 Poland 25.7 18.8 55.5 Romania 17.0 6.0 77.0 Russia 8.8 1.6 89.6 Ukraine 0.7 2.5 96.8 Source: See table notes. Fiscal decentralization: From command to market 25 the services provided by subnational governments are essential to politi- cal stability and economic development. Many local governments cannot provide these services at an adequate level from their own resources alone. Even from a short-run stabilization perspective, cutting transfers may be inadvisable. Underfunded subnational governments may cope with budgetary pressure by using such economically undesirable sources of revenue as profits derived from the exploitation of income-earning assets and from direct public ownership of local businesses. At the same time, in some countries subnational governments' open-ended responsi- bility for social assistance may result in emergency recurrence to the cen- tral government for additional funds, the unsustainable accrual of arrears, or undesirable short-term borrowing. Subnational government arrears are currently a major problem in Bulgaria, Romania, Russia, and Ukraine. Budapest is borrowing to finance recurrent expenditures (see chapter 3). Russia's oblasts are sometimes several months behind in pay- ing mandated adjustments to pensions and teachers' salaries. The only way to keep down such presumably undesirable developments in the coming years in most transition countries is likely to be to maintain some form of intergovernmental transfers. Deficit controls. Direct control over subnational deficits is one wav for central governments to ensure that subnational governments do not create macroeconomic pressures. Such limits are used in Russia, for example. Provisions of this kind may make sense in a framework of "hard" budget constraints, but the combination of deficit controls and soft budget constraints in the enterprise sector can create perverse out- comes."4 In response to an apparent surplus in overall subnational bud- gets in Russia and Bulgaria, for example, the central government transferred expenditure responsibilities to the subnational level, read- justed tax-sharing levels, and minimized transfers. But since Bulgarian subnational governments cannot legally run deficits or borrow, even for liquidity, and since borrowing abilities and authorities in Russia are lim- ited, budgets by definition must contain a surplus sufficient to meet the cash requirements of monthly local outlays."5 Reducing revenue shares or transfers will not eliminate this surplus; since subnational budgets cannot be in deficit, such measures will lead only to measured expendi- tures below "normal" levels and increased cumulative arrears. In this environment, a simple requirement for subnational governments to bal- ance their budgets is not sufficient. Restricting borrowing. A good case can be made for permitting local governments to borrow for certain capital investments. A striking feature of the current arrangements for subnational finances in some countries, 26 DECENTRALIZATION OF THE SOCIALIST STATE however, is the virtually unrestricted legal access given to subnational governments for tnlimited borrowing including, in some cases, foreign borrowing (table 1.7). In Hungary, for example, the Law on Local Self- Government grants all local governments unrestricted domestic borrow- ing rights for current and capital expenditures. Local governments in Albania can also borrow without restrictions for capital investment. Most of the other countries impose some limitations, however. In Poland, for example, localities are not allowed to borrow more than 5 percent of cur- rent budgeted expenditures. In the command economy period, subnational governments' bor- rowing was determined by the overall credit plan, and the central govern- ment guaranteed repayment to banks, just as it did for state enterprises. Under current circumstances, however, such generous local access to loan finance seems out of place. Restricted and limited own-source revenues may tempt local governments to overuse debt finance. The precarious macroeconomic situation in many transition economies makes the case for limiting local government's access to borrowing even stronger. Borrowing for recurrent costs, and free access to credit through such routes as local government ownership of banks, seem wholly undesirable. Few of the new subnational governments have experience with invest- ment financing, and most are not yet capable of preparing complete and meaningful projects. In most countries, intergovernmental fiscal flows are still in flux, and subnational governments have limited autonomy to set Table 1.7 Subnational government borrowing, selected transition economies Country Locol borroving authority Local borrowing activity Albania Unlimited borrowing authority for No information investment purposes onlya Bulgaria Limited borrowing authority but From Ministry of Finance governments accumulate arrearsb Hungary Unlimited borrowing authority Some municipalities have past debt. Minimal new borrowing Poland Limited borrowing authority Debt service not to exceed 1 5 percent of revenuec Romania No borrowing authority but Virtually no activity governments accumulate arrearsb Russia Limited borrowing authority, Minimal new borrowing limited by decree in 1993 (for liquidity purposes) Ukraine Limitec borrowing authonty from Minimal new borrowing comestic sources (for liquidity purposes) Note: For Indicated footnotes, see table notes. Source: World Bank est mates. Fiscal decentralization: From command to narket 27 user fees and little experience (or room) to tax on their own. It is thus difficult to judge their creditworthiness. Most financial institutions in transition economies are not yet capable of evaluating long-term risks or handling long-term financial instruments. All these factors argue for cau- tion in allowing local governments to borrow. Nevertheless, to the extent that benefits from some projects are enjoyed in the future, borrowing for local infrastructure projects has a sound theoretical base. Financing capi- tal expenditures from current revenues would unnecessarily limit the pace and efficiency of subnational investments. The challenge is to define appropriate financing mechanisms that do not threaten macroeconomic stability.16 In most countries, however, this task, while important, must wait on the more immediate challenge of developing a sound structure for financing subnational current expenditures. Privatization and subnational governments As noted earlier, an important feature of the fiscal landscape in transition economies is the continued significant ownership role of subnational governments at a time when national priorities support privatization. One reason this role persists is that in many countries a key element in the decentralization process has been the center's transfer of assets to subnational authorities. Another reason is the inadequate revenue base given to localities and their continued financial dependence on enter- prise revenues. For privatization to succeed, subnational governments must get out of the business of business and into the business of govern- ment. This process raises questions relating to the use of transitory rev- enues from privatization and how best to take advantage of the assets held by subnational governments. Asset transfers to the subnational level One of the most important features of fiscal decentralization has been the transfer of assets to subnational governments (table 1.8). Sub- national governments are increasingly becoming owners of enterprises, housing, vacant urban land, and retail establishments as these assets are transferred from the center. In Hungary the 1991 Act on Property Transfer transferred ownership of the housing stock to local govern- ments. Local governments in Ronmania have been given all formerly state-owned properties within their territories, ranging from public domain property such as parks to private domain property such as enterprises. Enterprises in Russia have long been owned by each level of government, and housing and land were transferred to municipal and rayon-level governments beginning in 1991. In both Poland and 28 DECENTRALIZATION OF THE SOCIALIST STATE Albania small-scale enterprises have been transferred to subnational governments, which have been charged with privatizing them. Asset transfers can be a mixed blessing, however. Indeed, one moti- vation for some transfers of public utilities, some blocks of public hous- ing, and certain other enterprises may have been to avoid the fiscal burden (such as subsidies, maintenance, and repair) that ownership car- ries. Subnational governments thus face major maintenance and subsidy burdens unless costs are recovered quickly. The net effect of asset trans- fers may be a long-term fiscal drag on local financial maneuverability, as has been argued may be true in Budapest (Alm and Buckley 1994). Entrepreneurial subnational governments In some countries there is evidence that property transfers are increasing the entrepreneurial ambitions of subnational governments as they use their new property rights to activate idle property. A worrisome pattern in the transition economies has been the involvement of governments at different levels in joint economic ventures in purely private, market- oriented activities. As Romania's Law on Local Public Administration states (and it is not unusual in this respect): "Local governments are authorized to create, contribute equity [to], and/or take shares in com- mercial enterprises formed with their assets [and] belonging to or trans- ferred to their ownership, including the possibility of forming joint ventures with foreign or domestic partners." Table 1.8 Ownership of housing and of commercial and economic property, by level of government, selected transition economies Domestic industriol Country Housing Reto;l units enterprises Public utilities Albania Subnational Central and Central and Central and subnationalb subnationala subnational Bulgaria Subnational Subnational Central and Central and subnational' subnational Hungary Subnational Central ano Central and Central and subnationald subnational subnational Poland Subnational Subnationale Centr-al and Central and subnationalf subnational Romania Subnationalg Subnationalg Central and Subnational (local) subnational contractsh Russia Subnational Subnational Subnational Subnational Ukraine Subnational Subnational Central and Central and subnational subnational Note: Subnationai includes ocai and/or provincial (oblast. judet, and so on) governments. For indicated footnotes, see table notes. Source: World Bank. Fiscal decentralization: From commanid to market 29 Localities appear excessively optimistic about their abilities to enhance revenue from entrepreneurial activities. Some hope to establish joint ventures with a domestic or foreign partner or another state enter- prise, using local assets such as land as their equity share. Some locali- ties also see potential in developing and servicing empty land to enhance its value as equity. Real estate development and financing is one of the riskiest business areas in market economies, and it is not like- ly to be a better bet in the transition economies. In Hungary tourism lodges, recreational facilities, and golf courses were cited as examples of such ventures-in-process, and in localities that inherited important real properties potential was seen in developing these as contributions to such industrial joint ventures as bakeries, construction firms, and food processing. In Russia all the rayon-level governments visited in the course of fieldwork were developing and exploiting their asset base through domestic joint ventures, anticipating revenue from the projects once completed. Similarly, in Poland providing private services as a means of earning additional local revenue appears to have strong sup- port, with housing and real estate development, food processing, and furniture manufacturing cited as examples. Subnational governments have good reasons for being involved with these new market ventures. Available land can be privatized only with difficulty, so the cost of contributing this resource to a joint venture may be perceived as being low. Also, manv local public officials want to create business employment opportunities in their jurisdictions, which often are economically depressed and lack private entrepreneurship.17 But the hope for profits from such ventures could quickly turn into demands on local budgets to cover losses. Public officials tend to be poor business managers. It is hard for a government to close a failing business, especially when the local employment situation is unhealthy (Kornai 1986). There is also the risk that subnational governments and their busi- nesses will undercut private competitors, possibly by employing regula- tory devices that protect local monopolies, with the result that the role of government in the economy will not diminish. And subnational govern- ments are subject to the ownership risks of any business. The rate of small business failure is high in developed market economies: there is no reason to expect that the transition economies can beat these odds. More important, entrepreneurial activity by subnational govern- ments is fundamentally inconsistent with privatization and represents a bottleneck to true decentralization-that is, decentralization not from the state to subnational govemments, but from government to the private sec- tor. Local government entrepreneurial activity does not promote privati- zation or true decentralization. Such "entrepreneurialism" inhibits or even reverses the privatization process-the long-term goal of removing gov- ernment from those activities that can be handled by the private sector. 30 DECENTRALIZATION OF THE SOCIALIST STATE An essential element in privatization is thus the creation of ade- quate revenue instruments and flexibility at the subnational level. If the only resource available to subnational governments in response to bud- getary pressure is profits from direct public ownership of local business- es, it is the source governments will use. Municipalities should not develop local monopoly enterprises to secure the revenue they need to function, just as they should not be forced to establish small taxes on a wide range of products and activities that are expensive to administer, generate widespread resistance, and typically yield little. Providing access to the national tax base (through transfers), granting the ability to levy local surcharges on national taxes where appropriate, and transfer- ring sufficient revenue sources to subnational control are essential if decentralization is to support privatization. Asset sales versus giveaways The sale of assets (housing, enterprises, land) is a potential source of significant revenue to subnational governments in most transition economies. But there is considerable controversy over how this process should take place and whether assets should be sold or given away. Different countries are employing different methods, ranging from auctions to spontaneous privatization to voucher systems (Hinds and Pohl 1991; Kikeri, Nellis, and Shirley 1992; Milanovic 1992b; Frydman 1993). In most of the countries under study, assets at the subnational level are being sold, not given away. In Russia privatization of the local housing stock and locally owned factories is at the discretion and under the direction of their oblast and rayon owners. In Hungary the sale of large enterprises is the responsibility of the State Property Management Company, while privatization of housing and many locally owned enterprises is a local responsibility. Local governments are responsible for the sale of small enterprises and locally owned housing in Albania and Poland. Maximizing the yield from asset sales in a manner that does not undermine privatization is a major challenge. Should they be sold now or later? Selling assets now, when markets for ownership and property rights are underdeveloped, may mean that they do not bring a good price. But local governments need revenue, and privatization cannot wait. Yields may be higher later, but other financing sources will also be available later. In addition, many locally owned properties cost more to operate and maintain than they realize in rents, constituting a drain on subnational public finances. Getting rid of this burden is a high priority. Unfortunately, the sale of locally owned assets, especially housing, has been slow in most countries, in part as a result of the absence of credit Fiscal decentralization: From command to market 31 (especially mortgage finance) and the lack of an adequate legal frame- work (Gray and associates 1993). Using privatization revenues What should subnational governments do with the proceeds from the privatization of public enterprises and the sale of other assets? It is important to remember that these are nonrecurring revenues and do not represent permanent sources of financing. Should such revenues be capitalized in special funds, invested in capital projects, or used to finance current expenditures? In practice, no restrictions have been placed on the use of revenues from privatization in any country. As a result, funds have commonly been used as an additional source of revenue in the recurrent budget. This is not all bad: it may be more distributionally attractive and economically efficient to raise revenues from privatization than from taxes. But such revenues are not unlimited. Sometimes these revenues have been kept in extrabudgetary accounts. A more conservative principle of public finance would call for the investment of these proceeds in long-term assets to allow local governments to draw income from assets over time. Alternatively, revenues could be used to repay debt (Newbery 1991).18 In Hungary revenues from privatization involving foreign joint ventures and generating foreign exchange have been largely channeled through the State Property Management Company to repay Hungary's external debt, although fiscal pressures emerging since 1994 may lead to demands for privatization revenues to finance the deficit. In the Russian Federation, where most privatization has been domestic, it has been recommended (Lieberman and Nellis 1994) that the stock of domestic government debt be reduced using this approach. In Poland, where revenues from privati- zation are being used by the central government to finance recurrent oper- ations and to compensate for inadequate tax yields, hardly any revenues accrue to subnational governments. In contrast, half of the (smaller) pro- ceeds from the privatization of state enterprises in Ukraine will accrue to subnational governments in 1994, although previously all proceeds went to the central government. Similarly, in Albania subnational governments have recently been allotted 28 percent of privatization revenues. Expenditure assignment The primary economic role of subnational governments relates to the delivery of local and regional goods and services. In principle, each function of government should be assigned to the lowest level of gov- ernment that encompasses all or most of the service benefits (Oates 32 DECENTRALIZATION OF THE SOCIALIST STATE 1990)-provided that it can effectively manage the activity. This does not mean, however, that the level of government that delivers a good or service should also be the one that produces it. Government provision is not the same as government production. Many "local" public goods might be better produced by higher-level governments or perhaps bv private producers. Local variations in tastes and costs mav make it more efficient for public sector activities to be carried out in as decentralized a fashion as possible. As noted earlier in this chapter, expenditure choices are more likely to reflect local preferences when made by local officials who have been democratically elected, and who are spending funds raised from the citizens who elected them. A decentralized system may also promote innovation, as local governments (and even the central government) can learn from each other's successes and failures. The efficiency case for dccentralization On efficiency grounds, central authorities should generally be respon- sible for defense, postal services, telecommunications, national roads and highways, and ports and civil aviation. Such services provide bene- fits bevond a single jurisdiction and are typically covered by the national budget. Judicial functions, national law enforcement, foreign affairs and diplomacy, industrial and environmental legislation, and interregional commerce are also usually seen as central responsibilities. The public services that are best delivered locally, with local deci- sionmakers deciding which services are provided, to whom, and in what quantity and quality, include transportation services such as Table 1.9 Expenditure responsibilities between levels of government, selected transition economies, 1993-94 Education Water and Country Priyar' Higher Housfnig sewerage Heo/th Albaniaa C C C S C Bulgaria SN C SN C SN Hungary SN C SN SN SN Poland SN0 C S SN C Romania' C C SN SN S Russia SN SN SN SN SN Ukraine SN SN SN SN SN S-Shared respons bilitY. SN-Pr mary responsibitliy is with the subnational governments. C-Primary responsibility is with the centra government. Note: Many of the subnationa ry prov ded services are provided by locally owned enterprises. a. Many of these services are be ng transferred to subnational governments. Fiscal decentralization: From commanid to market 33 construction and maintenance of local roads and provision of public transit, security services such as fire protection and enforcement of local laws, public utilities such as those for water, sewerage, and various forms of energy distribution, and general urban services such as provi- sion of parks, recreation facilities, markets, and slaughterhouses. These services mainly benefit the local area and are often provided by local governments. Alternatively, some can be provided privately or con- tracted out to private providers. Where there are significant spillovers, such as for education, health, and social welfare, expenditure assignments are less clear. Local schools and clinics clearly benefit residents of the local area, but a well-educated and healthy population also has national benefits. Because of such bene- fit spillovers, some components of education and health are sometimes national responsibilities (Musgrave and Musgrave 1989). When such ser- vices are local, the central government may help (or require) local gov- ernments provide services to a certain standard, for example, through grants conditional on the provision of the requisite level of service. In most of the transition economies of Eastern Europe, decentralization has led to a substantial transfer of these responsibilities to the subnational level (table 1.9), but the related transfers received from higher levels of government to carry out these and other spending functions are almost wholly unconditional (see later discussion). Often, shared responsibilities may be carried out either through a central-local arrangement or an association of subnational govern- ments. Sometimes both levels of government share spending responsi- bility within a sector. For example, in primary education, textbook selection and teacher training and certification might be at the national Social Public Public Public services Roads Police utilities amenities enterprises SN S C C S S SN S C S S S SN C SN S SN S SN S SN S SN S S S C SN SN S S SN SN SN SN S S C SN SN SN S b. Although full transfer to local governments was to be completed in 1993 oniy about one quarter of the gminy had negot ated agreements by that tme. The federal government s now recons dering the transfer. c. The Law on Local Admin stration was supposed to transfer many of the responsib Itjes to local govern- ments, but few were actually transferred. Source: Information is drawn from this volume. 34 DECENTRALIZATION OF THE SOCIALIST STATE level, capital planning at the middle and local tier, and recurrent financ- ing at the local level. Similarly, whereas water delivery is a local responsibility, treatment plants may require the creation of a special- purpose subnational government. Another issue involves financing responsibility and service delivery. Governments need not always directly provide services for which they are responsible. In market economies many local governments contract for ser- vices to be delivered by the private sector; garbage collection is a common example. Subnational governments may contract with higher levels of gov- ernment-in Canada, for example, the federal police provide contracted services to some provinces and municipalities, while other subnational governments have their own police forces. Similarly, central responsibility for such functions as poverty relief can be delegated to local governments. Services are often delivered most efficiently at the local level, even if financed in whole or in part by central transfers. For instance, the Albanian social assistance program (ndihme economike) is centrally funded but locally managed. The assignment of functions in transition economies is compli- cated by the fact that many goods and services once supplied by govern- ments are now being privatized. The extent to which government-supplied goods and services are public rather than private is debatable even in market economies, many of which are privatizing such functions as trans- portation and communications utilities. Even traditional public monopo- lies-postal systems, for example-have been considered for privatization in countries such as Chile, Finland, and Sweden. In principle, many public services can be delivered locally. In prac- tice, however, spending responsibilities vary considerably and there is no single best way to allocate responsibilities. Decisions about which level of government provides a service are often in practice as much or more a function of a country's history and political heritage as they are based on considerations of economic efficiency. Current expeniditure assignments in tranzsition economies Since subnational governments acted as units of the centralized adminis- tration under the socialist regime, subnational expenditures were includ- ed in the unified budget of the central government. All responsibilities were with the central government, either directly or through its local administrative units. Recent legislation in some countries under review has assigned certain expenditures to the subnational level of government (see again table 1.9). In Bulgaria, Hungary, and Poland, for example, local governments have specific expenditure responsibilities in education (pri- mary and kindergarten), transportation (local or urban streets), environ- ment (garbage collection, disposal, industrial waste), and housing and Fiscal decentralization: From command to market 35 related services. Local governments are also assigned a wide range of gen- eral spending responsibilities in areas such as nonschool education (nurs- eries, reform schools, orphanages), health, and social welfare. As a rule, however, these spending responsibilities are defined vaguely. The laws do not spell out, for example, whether responsibility for primary education implies responsibility for capital spending and rehabilitation of school holdings or just for recurrent spending. Moreover, spending responsibili- ties are almost never quantified in terms of their cost, making it difficult to address concerns over the adequacy of the local revenue base. Russia and Ukraine go further in the direction of vagueness. These countries have no legal definition of expenditure assignments at the sub- national level. The budget has thus far been the vehicle by which spend- ing authorities are codified. Since both countries have explicitly assigned revenues to the subnational sector, this approach puts the cart before the horse: subnational spending decisions are being driven by revenue avail- ability instead of the other way around. As it happens, tradition and iner- tia have established expenditure assignments in Russia and Ukraine that in many cases seem sensible. Thus, the federal or central level is responsi- ble for functions with a national benefit area, oblasts for regional func- tions, and local governments for expenditures with a local benefit area. But these principles are increasingly being violated as subnational gov- ernments are being forced to take on functions with national benefit areas (such as large-scale public investments and social welfare). Romania and Bulgaria allow even less local expenditure autonomy than Russia and Ukraine. In Romania expenditures that have strong local benefit characteristics are still controlled centrally. The central government is entirely responsible for education and health, reportedly to ensure a minimum standard in the provision of these important ser- vices. The main spending responsibilities assumed by local govern- ments in Romania are for water, local transportation, and district heating, as well as various minor activities in social assistance, munici- pal service, culture and arts, public investment, and maintenance.19 Public security and fire protection are also provided throughout Romania by local branches of a central ministry. In Bulgaria central mandates on, and preemptions of, local authori- ty are pervasive. Although many important local services are financed through local budgets, there is no local control over the provision of most services. For example, local officials have little discretion over per- sonnel expenditures in their budgets-they can neither set payroll levels nor effectively control the size of the workforce. In education, the central government appoints school directors who control the hiring of person- nel-even though the mayor is legally responsible for such appoint- ments. Localities in Bulgaria (as is the case in Russia and Ukraine) are 36 DECENTRALIZATION OF THE SOCIALIST STATE also subject to central mandates for payroll wage increases, which are often made without prior notice and without much attempt to determine the impact of the increases on local budgets. In some cases local officials are not even aware of such mandates until they become effective. These policies are not limited to such big-ticket items as payrolls. In the case of municipal hospitals, for example, the Ministry of Health mandates some expenditures down to the most minor details-from the number of hos- pital beds to the inventory of medicine and materials. A recent study reveals that some Bulgarian hospitals are operating at only 17 percent of capacity; yet local governments, which must finance these facilities, are unable to do anything to reduce the substantial excess capacity that appears to exist (Anderson and others 1994). Public expenditures atnd public enterprises No discussion of expenditure assignment in transition economies would be complete without a discussion of the public expenditures assigned to the enterprise sector. Subnational governments in most transition economies continue to look to enterprises to finance many essential pub- lic services. Enterprises provide their emplovees and families with many services that elsewhere are the government's responsibility. For example, they build and support hospitals, construct and maintain housing, build and run kindergartens and preschools, and make donations toward pub- lic transport and subnational extrabudgetary funds. The capital spending of larger public enterprises extends to roads, sewerage, water lines, and environmental protection. Most enterprises provide housing for employ- ees and many build and operate child-care facilities, such as summer camps. Enterprises provide money for construction of such facilities and often contribute labor and materials and participate in upkeep.20 With privatization, such expenditures cannot continue to be an enterprise responsibility. In all likelihood, they will be transferred to subnational governments. Owing to the financial difficulties of many enterprises, this transfer is already occurring in some cases. Enterprises cannot continue to provide such services and compete successfully in an increasingly privatized and market-oriented economy. But who is to assume what? Central governments appear to have neither quantified this problem nor considered possible solutions. Nor have subnational governments been given the fiscal authority to solve the problem them- selves. If subnational spending responsibilities are to be increased, sub- national revenues must be expanded either through strengthened local own-source revenues or a rationalized system of intergovernmental transfers. Alternatively, of course, the services previously provided by enterprises may simply cease to be provided. Fiscal decentralization: From command to market 37 Financing subnational governments The need to redesign the public sector is particularly evident with respect to the revenue system that has been inherited from the command era. Current user prices for public services, for example, are a classic combination of subsidy, inefficiency, and inequity. The lack of experi- ence with raising local own-source revenue hinders the proper exercise of the local fiscal discretion that is called for in the various laws on local self-government. And the lack (apart from Hungary and Poland) of a transparent system of intergovernmental transfers creates incentives for subnational government to revert to-rather than free themselves from-central regulation and control of their fiscal decisions. Until recently the only form of local own-source revenue in most transition economies came from enterprises' profit remittances. Tax reforms have largely replaced these remittances with profit taxes, though some countries still provide for some local "ownership" rights to enter- prise profit taxes.2" In Romania, for example, local governments (judets and municipalities) have the right to both profit and dividend taxes levied by the central government on locally owned public enterprises. Prior to the recent tax reform in Poland, both profit and dividend taxes went to the subnational government of origin and these governments (voivodships and gminy) also received shares of similar taxes levied on national enter- prises that had significant activities in their territories. This process was so complicated that the Polish tax administration spent almost as much time allocating enterprise taxes among jurisdictions as it did collecting them. Even where modern systems of tax administration have been adopted, as in Hungary, there is still a tendency for local governments to feel they have a primary claim on tax revenues generated within their jurisdiction. This "source entitlement" approach to the allocation of rev- enue is familiar in such contexts as the international taxation of corpo- rate profits and the taxation of natural resources. But it is seldom so visible at the local level as in some of the transition economies. Source- based taxation (also known as derivation-based) is unusual at the subna- tional level in market economies because of the inequities and inefficiencies that are likely to result from allowing subnational govern- ments to keep large shares of the taxes levied on firms producing for a national or world market. Source-based systems will give rise to the same problems and inequities in the transition economies. Financing through user charges Governments in the command economies used fixed prices and wage controls as essential elements of their distributional toolkit. Health 38 DECENTRALIZATION OF THE SOCIALIST STATE services and education were free. As a part of the social contract, govern- ments also provided some private goods either free or at a low price, including rents, many urban services, transport, utilities, food, and some- times clothing. These measures attempted to rectify fundamental or per- ceived distributional problems through inefficient pricing of scarce resources and, as in other cases where such policies have been followed, probably resulted in little equity being purchased at a high price in effi- ciency terms.22 With subsidies available to all, leakages to the nonpoor represented a major fiscal cost. Moreover, the failure to target meant that the level of benefits available to the poor was frequently inadequate to keep them from poverty (Milanovic 1995). As part of their adjustment programs, transition economies are thus being urged to correct public sector prices and to use national fiscal policies and targeted allowances and subsidies to achieve their distributional aims. The benefit approach to local government finance suggests that, where possible, services should be paid for by those who benefit from them (Bird 1976). Put another way, the first rule of local finance should be: "Wherever possible, charge." For efficiency, charges should be levied on those who receive the benefits where such beneficiaries can be clearly identified, unless there is a strong policy argument to the contrary. The importance of charging for public services goes beyond the revenues that can be collected from such levies. Only when the correct prices are charged (roughly marginal cost) will the correct amount and type of ser- vice be provided to the right people-that is, those willing to pay for it. Correct pricing helps public officials make sensible judgments on match- ing scarce resources with rising demand. When the true cost (and the related user charge) rises so high that quantity demanded falls, it is a clear signal that service costs have exceeded the perceived benefits of the service. Decisions must then be made on whether to maintain high prices and serve fewer people, cut back on a service's costs and standards to keep it affordable, or subsidize the service from general revenues. In the absence of price-generated demand information, such decisions must be made through bureaucratic rather than market-oriented processes. Public sector pricing is one of the last areas to be reformed in most of the transition countries. While price reforms have liberalized the prices of most goods in the private sector, both central and subnational govern- ments are having trouble adjusting prices for public services, even for pri- vate goods provided by the public sector such as housing rents, water, heating, and transport fares. For example, among the transition economies of Eastern and Central Europe, only Albania and Croatia, in mid-1994, had household electricity tariffs set at levels (comparable to West European lev- els) that approached economic (long-run marginal) costs. Adjusting such prices at the subnational level is essential for both achieving efficiency and Fiscal decentralization: From command to market 39 enhancing subnational revenues. Doing so, however, implies significant changes in the cost of living and in welfare, so changing public sector prices will result in large and unpopular distributional shifts if not proper- ly coordinated with other reforms. Such shifts could upset the fragile democracies that are emerging in the transition countries. In Poland, for example, it has been estimated that increasing the price of electricity to its appropriate economic level of long-run marginal cost could result in a dou- bling of its current price. In the short run, and assuming no reduction in demand, energy would consume as much as 20 percent of some house- holds' budgets (Freund and Wallich 1995, forthcoming).23 Not surprisingly, most subnational governments in transition economies-like those elsewhere (Bird and Slack 1993)-make less use of user charges than seems desirable, and many of the charges that are levied are poorly designed. Weak accounting systems that do not clearly identify costs (especially depreciation that covers the cost of capital) are another reason these services are commonly underpriced. In addition, newly elected subnational governments find it difficult to impose sub- stantial price increases on those who have just elected them. Central pricing mandates. A major problem even for reform-minded subnational governments in many transition economies is that user charges continue to be centrally controlled. In Albania, for example, water charges are set at uniform rates across the nation even though provision costs differ across districts. In Russia transport fares were fixed in nominal terms at a uniform rate for all localities until late 1993, creating costly subsidies for most municipalities. Some progress is being made in Hungary, however, where authority for setting sewer and water management fees has recently been transferred from the central government to municipal companies. Social pricing mechanisms. Raising public sector prices and user charges to the appropriate levels mav require special transitional approaches and pricing mechanisms. For example, lifeline rates, which set prices for small volumes of consumption at low levels, and senior cit- izen rates, which provide pensioners on fixed incomes with low-cost access, may help make price reforms more politically palatable (Bahl and Linn 1992). For major price increases, some price compensation scheme-perhaps a one-time payment-may be needed to compensate households for the impact on incomes. If well-designed, such schemes would facilitate cost recovery and allocative efficiency while cushioning the impact on incomes, thereby facilitating the price rise. Replacing the panoply of subsidized goods and services provided by subnational governments with targeted subsidies to those in need 40 DECENTRALIZATION OF THE SOCIALIST STATE will also require the development of new information and tracking sys- tems. Some observers think that effective targeting will be relatively easy in the transition economies because of the large "command and control" data base that existed in these countries (Milanovic 1995; Jarvis and Micklewright 1996, forthcoming). This hope may be naive, however, since substantial flaws appear to exist in these data bases-in Poland several million people seem to be "off the books"; Hungary and Romania have important low-income ethnic groups that are similarly elusive. Recent initiatives to undertake more broadly based household expenditure and income surveys should help fill the gap. In Poland and Hungary such surveys have helped to identify the poor and their socioe- conomic characteristics, and have been used to analyze the impact of changing cash benefit systems on the poor (World Bank 1994c). Although central information systems have disintegrated in the transition economies, local information on social needs appears to be better. The best results in terms of "targeting" might thus be achieved through locally administered social assistance schemes, as in Albania and Hungary, though such schemes generally require substantial cen- tral financial support. Suibn7ational taxcs Where it is impractical to impose fees or levy user charges, services should be financed by taxes levied on local residents to the extent they benefit from the services. "Local taxes" in this sense are those over which local authorities have some control-for example, they can set the rate or determine the base, in addition to receiving the revenue. Such taxes need not be locally administered. Some countries permit subnational governments to share national tax bases by allowing sub- national surcharges ("piggybacking") on such- centrally administered levies as the personal income tax. Like local taxes, such piggybacking arrangements increase the autonomy of local governments and also, in principle, the efficiencv with which they are likely to spend the money, compared with so-called tax-sharing arrangements, which simply direct a share of national taxes to local governments without the latter having any say, or responsibility, for the tax levied. Of course, channel- ing significant reventues to the subnational level restricts the center's ability to provide national goods and services and to support structural adjustment. Moreover, permitting subnational governments to deter- mine the taxes applied to mobile factors-of-production bases may impair macroeconomic flexibility and affect the efficient allocation of resources. Care should therefore be taken in deciding which taxes are assigned to which level of government. Fiscal decentralization: From command to market 41 Assigning taxes. Ideally, subnational governments should have access to relatively stable sources of revenue, the burden of which is not easily exported beyond local boundaries. In most countries, the central govern- ment controls those taxes considered to be most redistributive, such as personal income taxes with progressive rate schedules. In most of the tran- sition economies, however, some personal income tax revenues are cur- rently directed to subnational governments, although in no case do those governments have any control over the rate or base of the tax (table 1.10). Tax administration. A special problem in some transition economies arises from the fact that the tax administration is decentralized in the sense that subnational governments collect revenues on behalf of the center and transfer them upward. Strictly speaking, in these cases the tax administra- tions are deconcentrated arms of the federal or central administration. However, evidence in some countries, such as Russia, suggests that local tax officials may have strong ties and loyalties to the local governments- which often provide their housing and fringe benefits-and hence are responsive to local policy. In Russia and Ukraine all revenues other than customs and trade taxes are collected by local administrations. This dele- gation of collection responsibility, with its potential for noncompliance, is unlikely to prove a viable means of financing the central government in the absence of strong central political control. The fiscal deathblow to the Soviet Union came when the republics failed to remit resources to the Union budget. The same risk of oblasts withholding tax revenues threat- ens the central budget in Russia and Ukraine today (chapters 8, 9, and 10). WhiclI taxes to assigni? Some taxes are less suited to subnational gov- ernments than others. In particular, to minimize potential tax distortions, a high degree of national uniformity seems desirable with respect to the corporate income tax and the value added tax. The border controls need- ed to implement "local value added taxes" as a national value added tax with differential subnational surcharges would impede interjurisdiction- al trade and likely be impossible to administer effectively (McLure 1983d; Cnossen 1987; Tait 1988). The European Union has been trying for decades to deal with this problem and has not yet succeeded. The prob- lem becomes even worse across, say, 3,000 municipalities or even 89 oblasts. Moreover, even with identical rates, the value added tax is not suited as a local tax. If it is levied as an origin tax, some regions would see their revenue base enhanced or compromised, depending on the stage of the production process taking place within their borders.24 And if it is imposed as a "destination" tax, those regions with a heavy export orientation would lose (because of the zero-rating of exports), while those with high import ratios would benefit. 42 DECENTRALIZATION OF THE SOCIALIST STATE Table 1. I 0 Local government revenue shares, selected transition economies, selected years (percent) Albania Bulgaria Hungary Revenue source 1993-94 1993 1994 Central taxes Turnover tax/value added tax (VAT) 0 0 or I Qoo 0 Corporate income tax 0 10 or I 00' 0 Personal income tax 0 50 301 Natural resources tax - n.a. n.a. Export taxes 0 0 n.a. Tax on assets of locally owned state-owned enterprises n.a. I00 n.a Excise taxes n.a. Small amounts 0 shared with center Local taxes Property tax 60d 1IC Coo Auto/road tax n.a. 50e Small business taxes n.a. - 100 Agricultural tax n.a. 100 n.a. Poll tax n.a. n.a. I CC0 Other Significant user Cost recovery All tourism fees for services, charges and taxes and such as fee on user fees miscellaneous garbage local taxes n.a. Not applicab e. - Not available. Assigning corporate taxes to the subnational level also raises resource allocation concerns. While not uncommon in federal systems in market economies (Canada, Switzerland, and the United States), considerations of both administrative complexity and allocative efficien- cy suggest that subnational corporate levies should be avoided in the transition economies (Boadway 1992). Permitting subnational govern- ments to set corporate tax rates (or adjust the tax base) is likely to lead to substantial differentiation in enterprise taxation, influencing enterprise location decisions in perhaps undesirable ways-for instance, richer regions could set lower rates and hence exacerbate regional differences. Countries with subnational company taxes also have to decide how to prorate revenues when the company operates in more than one juris- diction. There is no objective scientific basis for determining the geo- graphic origin of income of a multi-jurisdictional enterprise. In the United States complex formulas use various sales, employment, and asset values to apportion profits across states and to determine what can legiti- mately be taxed in each jurisdiction, with different states using different Fiscal decentralization: From command to market 43 Poland Romania Russia Ukraine 1993 1993 1993-94 1994 0 0 20-509 20 (VAT) 5 0 651 50r 15 n.a. 100 50 n.a. n.a. 60-80' n.a. n.a. 25-65) 100 100 100 100 0 0 60 20, 100 100 100 700 100 100 100 100 n.a. 100 n.a. loop n.a. 100 n.a. n.a. n.a. n.a. Miscellaneous Miscellaneous 21 miscellaneous Timber duty, state local taxes local taxes taxes duty, water fee. other local fees. and 16 other miscellaneous taxes Note. For indicated footnotes, see table notes. Source Various; see table notes. formulas (McLure 1984). In Canada some of the resulting problems are avoided because all provinces use the same formula-although it took many years of negotiation to reach this result (Bird and Brean 1986). Some variant of the Canadian approach may perhaps work for the larger intermediate govemments in countries such as Russia and Ukraine, but it is unthinkable for the small, fragmented local-government sector that is more characteristic of Eastern Europe and much of the former U.S.S.R. In contrast, it may be possible to utilize local personal income taxes even for small local governments. Much of the literature on tax assign- ment argues that the personal income tax, which is often implicitly assumed to be one of the most important taxes in revenue terms, should be ietained by the central government, largely for redistributional and stabilization reasons (Musgrave 1983; Boadway 1992). Nevertheless, a variant of the piggybacking approach used to finance local governments in the Scandinavian countries (Bird and Slack 1993) is suggested for Hungary in chapter 2 of this book. Similar surcharges are suggested in the later chapters for Poland, Russia, and Ukraine, and might well be 44 DECENTRALIZATION OF THE SOCIALIST STATE considered in other transition countries. Combining a uniform national base and national collection with locally set rates makes such "local" income taxes administratively feasible. Compared with countries such as the United States, there is a relatively low level of population mobility in countries such as Hungary, owing in part to housing problems and in part to tradition, so the potential effects of variable local income taxes on location are reduced. Moreover, since the personal income tax is a rela- tively new, and as yet minor, source of revenues in transition economies, the immediate implications for central redistributive and stabilization policy of giving up some "tax room" to subnational governments are limited-although in the long run matters may be different. In some ways, excise taxes and single-stage retail taxes seem to be prime candidates for subnational taxation, especially if the taxing region is large enough to avoid revenue loss from consumers crossing the border to regions with lower tax rates. Such taxes thus seem more suitable for larger intermediate governments than for small local governments. As sources of revenue for subnational governments in transition economies, however, excises and retail sales taxes are notable by their absence. In the case of excises, one reason is that such taxes provide a sur- prisingly important source of national revenue-often more than the personal income tax-and the center does not want to give them up.25 Another explanation lies in the way excises are administered: because excises are levied on the (often single-monopoly) manufacturer-say, of cigarettes or alcoholic beverages-the excise tax base (production) is not broadly distributed across regions. Substantial administrative reform would be necessary to allocate revenues on the basis of consumption, let alone to levy taxes varying with the place of consumption (which would, for example, create "border" problems such as smuggling).26 But without such administrative reforms, the excise tax would accrue to only the few producing localities. Apart from the case of motor vehicles (dis- cussed below), there seems little near-term potential in subnational excise taxes in most transition countries. As for retail sales taxes, no East European country has yet intro- duced such a tax and none seems likely to do so. Retail sales taxes are, in principle, feasible at the subnational level, even when the national gov- ernment has a value added tax: as experience in Canada-currently the only country with this setup-suggests, however, the result is likely to be an unduly costly and complex tax system (Mintz, Wilson, and Grendon 1994; Bird 1994). In transition economies the task of implementing a robust national-level value added tax is central to tax reform (Tanzi 1993a, 1993b). It is unlikely that any transition country would or should be interested in complicating matters still further by experimenting with subnational retail sales taxes on top of the value added tax at this time. Fiscal decentralization: From commtiand to market 45 The taxation of natural resources presents a special case (see chap- ter 10). Natural resources are completely location-specific and thus could be taxed by the subnational jurisdiction in which they are located without any fear of their moving. Other economic efficiency considera- tions argue for national taxation, however (McLure and Mieszkowski 1983). In particular, the abundant revenue from natural resources may lead resource-rich jurisdictions to set other taxes below the level needed to pay for public services or to pay dividends to residents, resulting in overexpansion of public services or excessive labor or capital migration from other parts of the country. The equity argument against subnation- al taxation of natural resources is also a powerful one. Since natural resources are typically unevenly distributed across regions, assigning this tax to the subnational level would generate substantial differences in fiscal capacitv and be highly inequitable. In Russia, for example, three oblasts (with 2 percent of Russia's population) produce 80 percent of Russia's oil and gas (see chapter 10). Such arguments restrict the tax assignment choices available for sub- national governments. But there are a few options. The first and most efficient is taxing bases that are relatively inmmobile. This is the funda- mental case for a real property tax, particularly on residential property. Local owtn taxes. As table 1.10 shows, most of the taxes specifically assigned to subnational governments in transition countries are what might be labeled "nuisance" taxes rather than potentially robust own- revenue sources. In Russia, for example, the combined revenue from twenty-one local taxes assigned in the 1991 Law on the Basic Principles of Taxation is not likely to yield even 0.5 percent of GDP. These taxes include such gems as taxes on the resale of used computers, on horse racing, on the use of logos in advertising, and on dog owners. In Hungary there are five local taxes, which, in addition to the property tax, include a poll tax, a local business tax, and a bed tax in tourist areas. On the whole, it seems fair to say that subnational governments in most transition economies have been assigned residual tax sources that are generally not likely to yield much revenue. One tax that has some growth potential is local taxation of automo- biles.27 For the transition economies, such a tax tends to promote a rough degree of equitv-bv definition, it is mostlv the well-off who have auto- mobiles. In addition it can serve as a user charge for public services since it taxes the people who contribute to pollution and congestion and wear out roads. The explosion of the automobile population makes this a large and growing tax base. Few transition economies, however, have given motor vehicle taxation the attention it deserves. In Hungary, one of the four countries with a local vehicle tax at present, only 7 percent of 46 DECENTRALIZATION OF THE SOCIALIST STATE vehicles were subject to the small and poorly designed duties levied by local governments. Albania, where private ownership of automobiles was illegal until 1991, imported 25,000 cars in 1993, yet has an import tax of just $400 a vehicle and a yearly tax of only $50 a car-though neither is a local tax. Many countries have extensive centrally mandated exemptions from their existing vehicle taxes, such as exempting vehicles used by state-financed public institutions, those used by war veterans, and so on. Vehicle taxes could become a revenue source for at least the larger localities. All vehicles should be subject to taxes that at least offset the social and environmental costs of vehicle use. While such taxes should probably be designed and imposed uniformly across a country to mini- mize administrative problems, there is no reason revenues should not be partly assigned to the local governments in which the vehicle is registered. Tax sharing. Tax sharing refers to a system in which one govern- ment (usually the central government) collects revenue from a tax and shares it with subnational governments, usually on the basis of the geo- graphic origin of the revenues. Such sharing is common in transition economies. In Hungary, Poland, Russia, and Ukraine-and in all of the former U.S.S.R. republics-some or all of the personal income tax is assigned to the subnational level. In the former U.S.S.R. this is a carry- over from the days when the (unimportant) personal income tax was a local tax on the wage bill (rates and base were centrally defined) assigned to the oblast level and passed on to rayons and cities. Tax sharing has emerged as a response to the growing revenue needs of subnational governments in transition economies. While subna- tional governments were initially assigned only a few minor taxes, these revenues did not come close to financing the significant expenditure responsibilities with which they were charged. Thus, in countries (Hungary, Russia, and Ukraine) where the subnational sector is expect- ed to finance an increasing share of public expenditure, access to major national tax bases was needed. Russia now shares all of the personal income tax, part of the value added tax, part of the corporate tax, and most excises with oblast governments, in proportions that have been ris- ing over time. In contrast, in Hungary, where the personal income tax is also shared, the portion of this tax flowing to the local level has shrunk as the fiscal pressures on the center have increased. The personal income tax share to local governments was initially 100 percent in 1991, while in 1994 it was just 30 percent, shared on a derivation basis. Similarly in Bulgaria, the local (obshtini) share of the personal income tax fell from 100 percent in 1991 to 70 percent in 1992 to 50 percent in 1993. Tax sharing has the advantage of simplicity, and subnational govern- ments are guaranteed some degree of revenue certainty. Its drawbacks are that it does little to enhance either accountability or efficiency. Localities Fiscal decentralization: From command to market 47 receive revenues regardless of their tax effort, and have no responsibility for setting the tax rate or base. If they view these revenues as costless, as seems likely, their incentive to spend efficiently is obviously low. The result may be undue expansion of subnational spending. The extensive revenue sharing in transition economies is entirely on a derivation basis. Taxes are retained by the jurisdiction in which they are collected. Alternatively, shared taxes might (like the value added tax in Gennany) be shared through a formula based on such factors as popula- tion, per capita income, or urbanization. The roots of derivation-based sharing in transition countries are twofold-historical and pragmatic. The important ownership role of subnational and local governments under the old system gave them an "entitlement" to enterprise profits and, con- sequently, to the direct and indirect taxes paid by "their" enterprises. The concept of entitlement to revenues on an ownership or origin basis remains strong and accounts for the present-day reliance on this method. In addition, derivation-based sharing (or revenue retention in the case of Russia, Ukraine, and other former Soviet republics) is obviously relatively easy to administer and monitor since it really amounts to little more than the former system of regional allocation of revenues in a different guise. Derivation-based sharing, as a rule, channels resources to higher income areas where tax bases and therefore revenue collections are largest. It is thus inherently counter-equalizing. This may be a problem in countries where regional inequities are serious, and where the inter- governmental system lacks other instruments (such as transfers) to address such imbalances. In Hungary, for example, there are some townships that derive no revenues whatsoever from the personal income tax, since all their residents earn incomes that fall below the tax's threshold. Other aspects of Hungary's transfer system are designed to offset this, ensuring more equitable service provision. Most transition countries have much less sophisticated transfer systems. The larger the fraction of subnational resources that accrue on a derivation basis, the more the richer localities will benefit. Regional dif- ferentials in service provision will emerge as a result. The alternative of sharing revenues on the basis of an equalizing formula (as in most feder- al countries, for example; see Shah 1994) penalizes better-off regions whose industrialization and growth potential are greatest. Transition economies must balance this tradeoff between growth and equalization. Choosing the degree of equalization is an inherently political judgment. In some countries there are vast disparities between regions, and some degree of equalization is a political priority. In other countries the need for political unity may require less equalization. In Russia, for example, frustrated better-off and rapidly growing regions may simply opt out of the revenue-sharing system if too many resources are taken from them and shared with poorer regions (see chapter 10). 48 DECENTRALIZATION OF THE SOCIALIST STATE Surcharges. As noted earlier, another option for subnational financing is to levy local surcharges (piggybacking) on national taxes. Piggybacking on the national tax base is administratively simple for subnational govern- ments since they can usually rely on the superior central administrative machinery for collection. For transition economies, surcharges on the per- sonal income tax may represent an attractive alternative to the current derivation-based sharing of this tax.28 Instead of receiving a centrally man- dated share of the personal income tax (as is now the case in a number of countries), subnational governments would instead receive "their" share by imposing a surcharge on the national income tax, up to some ceiling, to be collected by the national government and remitted to each subnational government. As discussed further in chapter 2, the center could reduce its personal income tax rate to the level implied by the central share of the income tax to create "tax room" for a local surcharge; depending on the local surcharge introduced, the combined central-local rate could be as much as the previous national rate. The case for the personal income tax (or a surcharge on it) as a tax for the subnational level is pragmatic, and to some degree historical. In many East European transition economies, the personal income tax, or its precursor, the wage tax, was assigned to subnational governments. In a context where the choices among local revenue bases are limited-as noted earlier, the corporate income tax, the value added tax, natural resource taxes, most excise taxes, and retail sales taxes should probably be ruled out-the personal income tax is one of the few robust revenue sources that remain, one that is capable of generating significant rev- enues and likely to grow over time, especially as wages are liberalized. In most transition countries, workers are not highly mobile and few problems of revenue apportionment should arise. To the extent any dif- ferential local personal income tax falls on the local population, it may be considered a benefit tax. Merely assigning personal income tax rev- enues to subnational governments in whole or in part does little for local autonomy or accountability. Permitting a local surcharge should, in con- trast, enhance both accountability and local autonomy. Since the central tax base and rates will remain unchanged and uniform across all juris- dictions, any central income distributional objective being pursued through the personal income tax should not be affected. International experience certainly demonstrates that local income taxes and local income tax surcharges are feasible (Bird and Slack 1993). Perhaps the best-known examples of local income surtaxes are those in the five Nordic countries (Denmark, Norway, Sweden, Finland, and Iceland). In these countries, with some important variations, the local tax is basically levied at a flat, locally established rate on the same tax base as the national income tax and is collected by the central government. In Fiscal decentralization: From command to market 49 contrast, in Belgium (as in most Canadian provinces) the local surcharge is levied as a percentage of the national tax liability rather than the national tax base. A similar system exists in Switzerland, where most cantons-the intermediate level of government-allow local govern- ments (communes) to levy surcharges at locally established rates on the cantonal income taxes, taxes that are (like most U.S. state income taxes) in no way harmonized with the central income tax. The Canadian provin- cial income surtax accrues in part to the provinces participating in the tax collection agreement, and is assessed and collected by the national gov- ernment. However, the provinces themselves set the rates of "their" tax. In effect, the provinces have contracted with the national government to take advantage of its comparative advantage in tax collection. How big a rolefor property taxes? In some countries, especially those in the Anglo-Saxon tradition, the property tax is the most significant subnational tax (Bird and Slack 1993). Although relatively expensive to administer, property taxes have good efficiency characteristics compared with most other potential sources of local revenue.29 To the extent locally provided services increase property value, the beneficiaries of public expenditures are made to pay. From another perspective, the residential property tax may be broadly equitable in the sense that the better-off generally have better housing and hence usually pay relatively higher taxes. In practice, how- ever, the performance of the property tax as a revenue source has been mixed. Outside the Organization for Economic Cooperation and Devel- opment (OECD) countries, especially in the developing world, most sub- national governments do not raise much from the property tax, in part because assessments are quickly eroded by inflation and do not auto- matically increase with economic growth (Dillinger 1992). In addition, cadastral survevs and registration are generally poor, and collection and enforcement efforts weak. Institutional shortcomings in implementing property taxes. Property taxes in some form have been introduced in many transition econo- mies, including Hungary, Poland, Romania, Russia, and Ukraine. Albania approved a property tax in the beginning of 1994 and Bulgaria has a draft law, though it has yet to receive serious legislative attention. Indeed, the property tax is the only potentially significant tax that tran- sition countries have assigned directly to subnational governments. But much needs to be done if this tax is to fulfill its potential (Kelly 1994). Current yields are minimal, and yields in the near term will be insignif- icant, partly because housing ownership remains limited in many 50 DECENTRALIZATION OF THE SOCIALIST STATE countries and housing privatization has lagged. Moreover, most transi- tion countries do not have a basic cadastre in place. The property tax is not likely to be a buoyant revenue source until these and other prob- lems are ironed out. Structuiral problems with the property tax. In transition economies the property tax rate is usuallv set by the central government, and tax bases are often riddled with exemptions. In Romania, for example, the proper- ty tax is in principle levied on all privately owned real estate. But tax rates are differentiated and, in an undesirable holdover from the old regime, vary according to the social characteristics of the property own- ers: 1.0 percent on urban property owned by workers and employees of commercial societies; 1.5 percent on property owned by individual busi- ness, independent professionals, self-employed individuals, and family associations of no more than ten employees; and 0.75 percent in rural areas. Taxable value is reportedly based on insurance assessments, sales contracts, and construction costs, and appears to be grossly underesti- mated. Moreover, all publicly owned buildings are exempt and, in a fea- ture guaranteed to minimize the yield of this tax in a country privatizing its housing stock, all houses and apartments bought by individuals from the state are exempt for ten years. In Hungary the central government mandates full exemptions (car- ried over from the socialist regime) lasting ten to twenty years for any privately owned housing that has had improvements (such as a new bath or roof). The center also defines the tax base and the maximum tax rate the locality can impose: either 3 percent of "corrected sales value" (which, though defined as 50 percent of market value, is usually about 25 percent because of the manner in which market value is calculated) or a flat nominal amount per square meter.3"1 As a rule, area measures and in some cases fertility of land are at present the only data available to local governments for valuation. The exemptions and nominal rates ensure that, as in Romania, little revenue will be collected from this source in the near future. The new Property Tax Law in Albania assigns 60 percent of revenues to subnational governments, but the central gov- ernment has been granted complete control over value assessment, rates, and collection. Such structural defects, combined with the difficult and costly task of updating property valuations and keeping them current, mean that although property taxes may become an important source of subnational revenues in the long term, they are unlikely to produce substantial rev- enue in the near future. Many improvements will be needed for the property tax to become an important component of subnational finance in transition economies.3' Fiscal decentralization: From command to market 51 Intergovernmental transfers Since it is unusual for the revenue-raising capacity and the expenditure needs of subnational governments to be perfectly matched, fiscal transfers are usually required to close the resulting gap. Such fiscal deficits could also be closed in principle by shifting tax powers to subnational govern- ments, assigning expenditure responsibilities back to the central govern- ment, or reducing subnational expenditures and service standards. Transfers may be preferable to these approaches, in part because they allow local governments to provide services and the central government to set standards. Transfers can be used to influence spending patterns at lower levels of government in line with higher-level priorities, to limit inequalities in the quality and quantity of regional services, and to ensure that neces- sary services are provided (Boadway and Hobson 1993; Shah 1993). Designing a transfer system is difficult. Transfers often have a num- ber of objectives: correcting vertical imbalance (the gap between the expenditure responsibilities assigned to the subnational level and the revenues assigned to or shared with this level), reducing horizontal imbalance between rich and poor provinces to equalize expenditures among jurisdictions, stimulating subnational tax effort (encouraging localities to raise more own-source revenue), and influencing subnation- al spending decisions in accordance with central preferences. Rules versus discretiont in transfers. Should a transfer system be defined by rules, or should it be discretionary? If the central govern- ment's macroeconomic objectives were the only consideration, the discre- tionary approach would have much to recommend it. But transparencv favors a formal, rules-based approach that allows subnational govern- ments to plan their fiscal operations and makes the central government's policy aims clear (Bahl 1986). Specifying the transfer of a set percentage of national revenues, or of a particular national tax or taxes, has substan- tial advantages for both central and subnational budgeting.32 With this system, central government is somewhat insulated from pressure to increase its support of local governments, and subnational governments can budget with greater certainty, knowing that the level of central sup- port will vary with a known formula and is not subject to discretionary central policy. In most transition economies the aggregate volume of transfers is determined entirely by the central government. In Hungary aggregate transfers are determined as part of the annual budget process, and are strongly affected by the fiscal stringency facing the central government. The situation is the same in Poland, although since aggregate transfers are (as discussed in chapter 4) formula-driven, transfers have been 52 DECENTRALIZATION OF THE SOCIALIST STATE relatively transparent and predictable. Russia and Ukraine had no explicit transfers until recently: ex post subventions were negotiated between deficit regions and the Ministry of Finance; the Ministry also determined an agreed or appropriate level of expenditures as the basis for calculating any gaps. For example, transfer decisions in Russia until 1994 were determined on the basis of negotiated spending levels (agreed between the Ministry of Finance and each oblast in the course of the budget negotiation process) and oblasts' fiscal gaps after accounting for their agreed share in selected central taxes, and other revenues. Once the aggregate volume of transfers is determined, the next step is determining its distribution across subnational governments. Conceptually-if not always politically-this step too should be as trans- parent as possible. One approach, discussed further with respect to Hungary in chapter 2, would be to estimate minimum subnational expenditure needs related to objective demographic or economic indica- tors (population, per capita income, or social indicators), to assess the local own-source, assigned, or shared revenues available to finance these needs, and to provide a transfer related to the difference. Depending on how much interregional equalization of expenditures (and thus services) is desired, such transfers may fill all or part of the estimated revenue gap. In principle, the estimates of expenditure needs in this approach should be independent of the actual policies of any individual govern- ment: they may, however, be based on the average policies of all govern- ments, as shown in chapter 2. The idea is to relate the estimates to underlying demographic and economic factors that serve as an objective indicator of the "workload" for each spending function. This approach should be contrasted with the "norms" currently used for regional budget allocations in a number of transitional countries. These "norms" (as out- lined in the discussions of Bulgaria, Hungary, Russia, and Ukraine) typi- cally relate to the existing installed capacity for particular services (number of hospital beds and the like). As a basis for distributing inter- governmental aid, such norms clearly reward jurisdictions that have had high rates of capacity/expenditure in past periods. For example, a norms- based approach may distribute funds on the basis of the number of hospi- tal beds in a jurisdiction regardless of whether or not the beds were used (that is, "needed"). In contrast, a needs-based distribution scheme might target aid using a demographic indicator such as the number of elderly, which is highlv correlated with the likelihood of going to a hospital and using one of those beds. Similarly, the number of students registered in an elementary school would reflect capacity, whereas the number of children of elementary school age would more accurately measure need. Despite the theoretical attractiveness of estimating "needs," how- ever, experience suggests that such estimates are difficult, and few Fiscal decentralization: From command to market 53 countries (one example is Australia) have made them (Shah 1993). For the most part, countries with a formula-based transfer system have attempt- ed mainly to distribute funds on a more or less equal per capita basis, sometimes adjusted for the size of the jurisdiction, while taking into account differential local capacities to raise revenues from sources under their own control. How this might be done is spelled out in more detail for Hungary in chapter 2. In any case, rules-based allocation formulas are not yet common in the transition economies. Apart from Hungary and Poland, where an explicit formula for the normative grant includes an equalization com- ponent and a complex formulation of expenditure needs, most intergov- ernmental fiscal flows continue to be essentially discretionary and negotiated. Empirical studies of the transfer systems in Bulgaria, Romania, Russia, and Ukraine (see chapters 5, 6, 8, and 9) were unable to determine a significant relationship between needs-based allocation criteria and income levels across oblasts. In none of the empirical studies were the transfers explicitly equalizing across regions. Simplicity in the design of the transfer system. If a formula is used to determine transfers, measurements of fiscal capacity and perhaps also of expenditure needs are required. Such calculations are difficult under any circumstances, and particularly so in the changing circumstances of the transition economies. In some countries it may be difficult even to deter- mine which regions are rich and which are poor-that is, where expendi- ture needs are high in relation to means. Compressed wages, distorted prices, regional inflation differentials, and in some cases shortages of goods may make it hard to interpret statistics on per capita income. For these and other reasons, transition economies require simplicity in for- mula design. Some countries may be ready for a formula based on per capita income and other sophisticated indicators. In others, it may be pos- sible to identify broad indicators of fiscal capacity and need. Some coun- tries may use demographic indicators such as population size, number of pensioners, and levels of urbanization. Others may base their formulas in part on key indicators of locally assigned public service needs, such as miles of roads or deficiencies in school or hospital space. Experiments will be needed to determine which transfer formula is most appropriate in light of the needs and aims of the new subnational finance system. (For a preliminary exploration along these lines, see chapter 2.) Transition economies could also, of course, retain their existing sys- tems of expenditure "norms" to estimate expenditure needs. One draw- back of this approach is that it requires constant updating, which can be expensive and complex. In addition, as noted above, since the expendi- ture norms in the countries that use this system were designed mainly to 54 DECENTRALIZATION OF THE SOCIALIST STATE cover the recurrent costs of existing facilities, different standards of facil- ity provision across regions result in different recurrent budgets. Transfers based on these factors will therefore perpetuate significant dis- parities in the level of public service provision across the country. A casefor conditionality? A striking feature of transfers in the transi- tion economies is that they are, without exception, fully unconditional. In Hungary, which has the best-developed transfer system, grants can be spent any way the local government sees fit. Similarly, in Romania, Russia, and Ukraine, transfers are lump sum, and subnational govern- ments can spend them free of all conditions and constraints. In principle, there is little case for bestowing such large unconditional transfers on local governments. In a nonfederal svstem, local authorities should be fundamentally responsible to the central authorities (or more accurately to taxpayers at large) when they are spending central funds (Bird 1993). So far, however, there seems to have been little consideration in the tran- sition economies of the possibility of introducing a degree of condition- ality into transfers-perhaps in part to avoid the danger of recurring to the centrally dominated model of past intergovernmental relations. A good case can be made for some conditionality in intergovern- mental fiscal transfers in transition economies: the central government has a legitimate interest in what is done with its grants and the country has a legitimate need to ensure that services such as education and health are available throughout the country at some minimum standard. In Hungary, for example, without conditionality the present system per- mits larger towns with secondary schools to discourage students from smaller towns without such schools from attending. Indeed, since the norm provided in Hungary's grant system for secondary education does not cover its costs of provision, towns have an incentive to do just this. In most Central and East European countries the central goyernment, as was the practice under the socialist regime, has continued to supervise the implementation of subnational budgets to make sure that specified targets in local budgets are achieved. But such "command and control" conditionality is not in keeping with the fiscal autonomy of the self- government movement and is increasingly unacceptable. Systems need to be developed that ensure subnational governments spend grant funds to provide agreed services of a specified quality and level.33 Standards might include, for example, adherence to national cur- ricula, or hiring only credentialed medical personnel or teachers. Another approach is to impose conditions on transfers. For instance, support for health services might depend on whether residents from other localities have access to treatment; support for education might require adopting a prescribed curriculum. Fiscal decentralization: From command to market 55 Matching special-purpose gran7ts. Central governments often make grants other than general transfers to subnational governments specifical- ly to influence their spending patterns. A matching grant may provide a prescribed level of support to "match" subnational spending for a specific purpose. If nonresidents enjoy a percentage of the benefits of a subnation- al govermnent's spending, in principle a matching grant equivalent to that percentage would ensure that the good is supplied efficiently. Such grants may also generate increased tax-raising efforts by local governments as the opportunity cost of not raising revenues goes up. Unfortunately, so lit- tle is known about spillovers that this rule provides little practical guid- ance to the design of matching grants. Another problem with most matching grants is that they provide greater benefits to wealthier local governments that can afford more extensive discretionary expenditures. Differential matching rates could handle this problem (Feldstein 1975), but again it is difficult to determine what such rates should be. With the exception of Hungary (which uses transfers for infrastruc- ture completion projects), systematic conditional matching grants hard- ly exist in the transition economies. Local infrastructure and capital outlays are typicallv still directly funded and determined by the center. The development of an appropriate design for the grant system and a system for infrastructure financing remains a major challenge for the transition economies. Although a conditional matching grant system may be a key ele- ment of intergovernmental finances in the future, the lack of budgetary autonomy in some countries and local tax authority in all of the coun- tries studied makes implementing such a system in the transition economies more complex. Matching grants encourage local govern- ments to increase spending on the targeted good through the price effect. When the price of the designated service relative to other services falls, local governments have incentive to shift spending to that service; however, in order to do so they must have power over their expendi- tures and the ability to improve tax effort. In Romania no significant ser- vices are the responsibility of the local governments and in Bulgaria local governments have almost no spending discretion. In Russia and Ukraine the picture is somewhat more mixed, but there is far from full spending discretion. A conditional grant system in these countries can therefore influence neither their spending pattern nor their tax effort. Without tax authority, governments have little ability to enhance tax effort and without spending discretion, little incentive to promote tax effort. This is a problem in all of the transition economies; the small size of local governments' own-source revenues implies matching grants will not do much in the way of encouraging localities to increase their tax effort (Martinez-Vazquez and Bogetic 1995, forthcoming). 56 DECENTRALIZATION OF THE SOCIALIST STATE Nonmatching special-purpose grants. Like matching grants, nonmatch- ing grants lead to increased public service provision, but unlike match- ing grants they do not alter the marginal cost of the service. Such conditional block transfers are one way of ensuring a minimum stan- dard of provision: in effect, the subnational government acts as an agent of the central government. Again, however, the effectiveness of this device depends very much upon the capacity of the central government to monitor performance. Continuing problems: Underfunded and transfer-dependent? The fiscal situation facing most central governments in transition economies is such that few major revenue sources have been ceded to subnational governments. Only Hungary and Russia have given subna- tional governments a significant share of certain national-level revenues. Subnational governments in transition economies are generally under- funded and overly transfer-dependent. Vehicle taxes and especially the property tax may have significant long-term revenue potential, but the subnational own-tax package in most transition economies leaves much to be desired. Most subnational tax rates and tax bases are established by the central government. Many taxes thus established have rates fixed in nominal terms, making them highly vulnerable to erosion through infla- tion. Others are minor nuisance levies that may yield less than they cost to collect. None provides an adequate fiscal foundation for an active and viable subnational government. The transfer systems in transition economies are also vulnerable to criticism. They discourage local responsibility, so recipients are tempted to run bigger deficits to increase transfers. To the extent they involve negotiation, they may also be criticized for lack of transparency and accountability. The future development of transfers in transition economies should incorporate simple formula-based criteria to discour- age negotiation and, where politically desirable, to promote equaliza- tion. Subnational governments need a hard budget constraint on their expenditures-as well as sufficient budgetary flexibility to accommo- date the growth of subnational finances. The distribution formula should be based on feasible and acceptable measures of capacity and need, and transfer payments should to some extent be made conditional on performance. Most transition economies need to derive measures of local fiscal capacity and to develop systems for monitoring subnational fiscal performance. Close attention should be paid to supervision, accounting, and monitoring of transfers, associated subnational expendi- tures, subnational borrowing and, in the future, the use of matching funds (such as for investment purposes). Fiscal decentralization: From command to market 57 The institutional requirements for intergovernmental systems Intergovernmental relations will require oversight after the current legislation is implemented, particularly in technical and nonpolitical functions. An appropriate institutional forum for intergovernmental coordination fulfills several needs. It could analyze local budgets and report social, economic, and financial data on local revenues and spending outcomes. It could provide technical assistance to subnation- al governments to build up financial and management skills. It could also help local governments act cooperatively where they have com- mon interests, such as providing services where joint cost savings are possible. Finally, it may also represent local governments' interests with the center. Intergovernmental coordination The lack of appropriate central government structures to monitor and support decentralized subnational governments is a common problem in Central and Eastern Europe. Despite the involvement of a number of central government institutions in a number of countries, many essential tasks are being left unattended. Central authorities in all countries need a better understanding of both the current situation and the likely effects of any proposed changes. Among the tasks for which responsibility must be clearly established are the monitoring and assessment of subna- tional finances, both in aggregate and for individual subnational govern- ments. This could be the responsibility of the Ministry of Finance, a special subnational government department, or an agency with semi- independent status and some subnational government involvement. Fiscal accounting and reporting. Regular and detailed financial data should be maintained on subnational governments. Ideally, uniform financial reporting and budgeting systems should be established, possi- bly with varying degrees of complexity for different categories of local governments. An appropriate agency needs to be made responsible for regular and timely data collection and processing, the execution of a census of governments, and ideally, publication of the public finance data that are needed for monitoring, revenue estimation, and economic research and analysis. The development and implementation of financial reporting and information systems generally requires substantial training and follow- up efforts. Developing such informational and technical infrastructure is neither quick nor cheap, but it is essential if important public sector activities are to be decentralized without the center losing sight of the 58 DECENTRALIZATION OF THE SOCIALIST STATE activities taking place in the subnational part of the public sector. Imple- mentation of these systems may require additional central institutions to develop and maintain the reporting systems, to train and support local government officials, and, as mentioned, to run the statistical side of the operation. Training is also needed to upgrade the technical capacity of subna- tional governments to perform efficiently and effectively the expendi- ture functions for which they are responsible and with which they may be entrusted in the future, whether autonomously or as delegated agen- cies of the central government. Technical assistance could come from the relevant central ministry, regional agencies, universities, private sources, or some combination of these. In view of the heterogeneous nature of the services that have been or may be decentralized among the countries, it is unlikely that the same solution will be appropriate in all cases. Each functional area-including the budgeting and financial reporting discussed earlier-has different needs, problems, and possi- bilities, and will likely have to be treated differently. Policy analysis capability. An especially high-priority task for the needed central monitoring unit is analyzing and evaluating intergov- ernmental fiscal transfers. As noted, many questions must be consid- ered with respect to the objectives, design, and effects of such transfers. More open and informed discussion of these matters is needed in many countries, particularly when the process is as novel as it is in the transi- tion economies. Regular publication of data will help, but no central government can be expected to carry out objectively and publish the economic and political analysis that is needed in any country where intergovernmental financial issues are important. Accordingly, most countries could usefully establish a small (and perhaps nongovernmen- tal) research institute to focus on subnational government problems. Development and quantification of options for redesigning systems of intergovernmental relations in the transition economies requires con- crete empirical work. The system design and performance monitoring of intergovernmental systems will be a long-term, data-intensive exercise. Needed data are not available in many countries. (Investment data for Russia's oblasts, for example, have not been published since 1975.) Income data are often hard to interpret. Although the task is formidable, if decentralization is to work in the sense of delivering desired services as efficiently as possible, governments across the region should develop such databases as disaggregated tax collection and tax base statistics, expenditure composition and expenditure needs measures, socioeco- nomic characteristics of the population, and the stock of phvsical infra- structure and the state of its maintenance. Fiscal decentralization: From command to market 59 Conclusion What do the concerns expressed above imply for the design of subnation- al and intergovernmental fiscal systems in the transition economies? While the details with respect to particular countries may be found in the rest of this book, the general goals for the design of an ideal system of subnational government financing in transition economies are as follows: * Supporting the emergence of a subnational government role con- sistent with market-oriented reforms. * Ensuring correspondence between subnational expenditure responsibilities and financial resources (including transfers from the central government) so that functions assigned to local gov- ernments can be carried out effectively. * Increasing the autonomy of subnational governments by provid- ing incentives for them to mobilize revenues of their own. * Supporting the center's macroeconomic management policies. * Giving expenditure discretion to subnational governments in appropriate areas to increase the efficiency of public spending and to improve the accountability of local officials to their con- stituents in the provision of local services. * Establishing a system of intergovernmental transfers that is transparent, based on objective and predictable criteria, and not subject to negotiation and ad hoc bargaining. M Minimizing administrative costs to conserve scarce administra- tive resources. * Providing some equalizing transfers to offset differences in fiscal capacity among some localities and to ensure that poor localities can offer sufficient amounts of key public services. * Incorporating mechanisms to support public infrastructure development and its appropriate financing. As argued earlier, a broader framework than is customary is need- ed to analyze fiscal decentralization and subnational and intergovern- mental fiscal issues in the transition economies. This framework must incorporate such elements as the likelihood of continuing structural changes in the economy and continuing political shifts, the need to undertake intergovernmental reforms while coping with stabilization pressures and the increased importance of the social safety net, the like- lihood of continued (local) public ownership on a significant scale, the financial implications of such ownership and its possible conflicts with privatization, and continued vestiges of price and wage controls and other rigidities. Including these elements in the framework of system design is essential if either decentralization or some of the other key transition 60 DECENTRALIZATION OF THE SOCIALIST STATE reforms are to be achieved fully or soon. Each country in transition is different, and each will require close examination to determine the struc- ture of the incentives, constraints, and opportunities that must be com- bined to design as hard a budget constraint as is acceptable. It will be a major challenge in the transition economies to develop an intergovernmental fiscal system that is an optimal combination of rules and flexibility-one that is flexible enough to be compatible with macro- economic stabilization and the major structural shifts that are taking place in the economy, and firm enough to provide a stable framework for the effective operation of both central and subnational governments in the longer term. Countries in transition should, given the clear political pressures to decentralize and the potential economic advantages-and dangers-of doing so, move empirically and cautiously toward the design and imple- mentation of new subnational and intergovernmental finance systems. The issues are complex, technical, political, and location-specific, and knowledge of the effects of various alternatives is often lacking. In these circumstances, good analysis based on sound knowledge of local condi- tions may have a major impact on improving the design of the systems that are being put into place and, hence, improving the lives of those undergoing one of the major upheavals of our time-the move from a command economy to a market economy. This book is offered as a first step in this ongoing process. Notes 1. In this book, the term "subnational government" includes all levels of government below the national level: it thus includes such intermediate bodies as (in different countries) the oblast, county, or voivod, as well as the cities and other local governments that are in some sense "below" these intermediate gov- ernrments. The term "local government" refers to the lower level of general gov- ernment-such as cities, municipalities, or villages-where an intermediate level exists. As discussed later in this and other chapters, the role and structure of intermediate in relation to local governments is unclear in a number of countries. In some countries local governments may in turn be divided in certain instances into smaller "districts" or "settlements"; again, the relationship between these sublocal units and local governments is not always clear, although this matter is not discussed in detail in this book, except with respect to Budapest in chapter 3. 2. Data for Russia are from the Ministry of Finance and those for Hungary from the IMF's Government Finance Statistics yearbook (GFS) and Hungary's Ministry of Finance. 3. A small sampling of the extensive economic and political literature bear- ing on this theme might include May (1969), Dafflon (1977), Breton and Scott (1978), Courchene (1984), Bird (1986b), Burgess and Gagnon (1993), and Banting, Brown, and Courchene (1994). Fiscal decentralization: From command to market 61 4. See table 1.10 for examples. 5. It should perhaps be noted that this argument applies strictly only in non- federal countries. As Bird (1994) has argued, in federal countries one aspect of the basic federal contract may be that some or all subnational governments are, in effect, "licensed" to be inefficient-at least when viewed from a central govern- ment perspective. With the possible exception of Russia-which is also the only formal federation among the countries considered here-none of the transition countries considered in this book appears to be "federal" in this sense. 6. Wherever possible the data used in this book have been taken from pub- lished sources available to the general research community such as national sta- tistical yearbooks and the IMF's Governmenit Finance Statistics. In many cases, however, unpublished data, made available by central or, in some instances, sub- national governments, have had to be used. The availability, consistency, and comparability of data on subnational finance in many countries clearly leaves much to be desired. Every effort has been made to verify and update the data used here, but many gaps and problems inevitably remain. 7. It is sometimes argued that devolution may provide a better overall framework for the development of private market forces since devolved govern- ments have greater incentives to compete for businesses and residents by select- ing an efficient set of revenue sources and service levels, but this point is subject to considerable controversy (see Kenyon and Kincaid 1991) and cannot be dis- cussed further here. 8. Article 51 of the Public Finance Law states that "the distribution of the revenues and expenditures or categories of local budgets ... is established by the decision bodies of each county [district] and of the city Bucharest." But Article 22 of the same law states explicitly that the central government has decisive power in the formation of local budgets. Local budgets are submitted to the Ministry of Finance each year no later than July 1. The Ministry of Finance examines the draft budgets and has ten days in which to "ask for the necessary modifications" to establish budget balance. Most important, "in case of divergence the [central] Government is to decide." 9. Law on the Restructuring of State-owned Economic Units (1990). 10. The larger metropolitan areas in transition countries, like their fellows throughout the world, often have special problems arising in part from govern- mental fragmentation and in part from their more complex and often different internal governmental structure. Chapter 3 discusses some aspects of the latter problem in the case of Budapest, but otherwise special "large city" problems are not treated in this book. 11. The Russian deficit for 1991 is hard to measure owing to the treatment of Union-level expenditure responsibilities taken over by the Russian Federation beginning in November 1991, following the dissolution of the U.S.S.R. Some sources cite higher numbers than reported here, referring to a "notional" deficit, which assumes that Russia had financed Soviet Union expenditures for the full year. 12. Almost certainly the central bank in Romania carries out heavily quasi- fiscal operations, and the banking system finances expenditures that are fiscal in nature. 62 DECENTRALIZATION OF THE SOCIALIST STATE 13. Sources for the national budget expenditures are the ministries of finance of Hungary, Russia, and Albania. 14. For examples of dysfunctional incentives for economic liberalization as a result of fiscal decentralization policies in China, see Wong (1991, 1992). 15. The subnational budget surplus in Russia, measured on the reported "cash basis," amounted to 1.2 percent of GDP in the first quarter of 1992, when the same rules applied in that countrv. See Wallich (1994) and Anderson and others (1994). 16. For a discussion of this problem in the more general context of develop- ing countries, see Dillinger (1992). 17. A note of caution is also needed with respect to the apparent desire of many local governments to use fiscal incentives to bolster private business in their jurisdictions. The record of such policies is not very encouraging in any country (see Wasylenko 1981 and 1991 for evidence from the United States), and efforts in this direction are all too likely to degenerate into a zero-sum game for local governments unless great care is exerted. 18. Newbery (1991) argues against a rapid, "fire sale" or giveaway approach because many of the assets may have been financed with debt (as noted in the country chapters in the present book, some local governments in transition economies have substantial liabilities), in which case the proceeds should be used to pay off this debt, rather than burden current and future local taxpayers with debt service. 19. As discussed in chapter 6, many of these services are actually provided by local public enterprises. 20. This system can have advantages for enterprises. In Russia, for example, local tax offices reportedly permit illegal tax breaks to firms that finance their own kindergartens or other social services. 21. See Tanzi (1993a, 1993b) for a discussion of national tax reforms that are in process in most countries. In the former U.S.S.R., all enterprise taxes went to the level of government that owned the enterprise. 22. For a review of the literature on this subject in developing countries, see Bird and Miller (1989) and Jimenez (1989). 23. Calculated by the authors based on price data for 1993 and on the Polish Household Budget Survey. 24. This problem, like that of evasion, has proved to be very serious in Brazil, the only country in which subnational units levy value added taxes. See Longo (1990) and Purohit (1991). 25. See the discussion of Bulgaria in Martinez-Vazquez and Bogetic (1995, forthcoming). 26. Countries such as Canada and the United States, with important subna- tional levies on easily smuggled products such as tobacco, have experienced sig- nificant administrative problems with these taxes. Recently, for example, the province of Ontario in Canada was forced to cut its tobacco taxes in half to match a similar cut in the neighboring province of Quebec (which in turn lowered its taxes to match the lower taxes in neighboring New York state). 27. For a general discussion of automobile taxation, see Bahl and Linn (1992). 28. As suggested in the earlier discussion, and discussed at more length in chapter 8 with respect to Ukraine, the value added tax and the corporate income tax do not lend themselves as easily to surcharges. Fiscal decentralization: From command to market 63 29. There is, of course, an extensive (and to some extent controversial) litera- ture on the incidence of local taxes, especially local property taxes, which cannot be discussed here; for a recent summary, see Bird and Slack (1993). The discus- sion in the text concentrates on residential property taxes: taxes on business prop- erty-apart from the extent to which they serve a "benefit" tax purpose-raise problems similar to other taxes on business such as the corporate income tax and the value added tax, and, as in those cases, a good argument can be made for restraining subnational access to this potential revenue source (see chapter 2 for further discussion). 30. There is little to be said for using area as a basis for urban property taxes. The case is different in rural areas, however, where taxes such as that in Romania-a differentiated rate ranging from 35 lei to 400 lei per hectare, depend- ing on fertility and location-may make sense, provided the fixed levies are adjusted annually in accordance with an appropriate price index. Unfortunately, those who have received land under land reform in Romania are exempt from land taxes for eight years, thus weakening this potential source of local revenue. 31. The same will be true of the benefit levies (charges) that are sometimes imposed on the assessed value of real property, or on some characteristic of that property such as area, frontage, or location, to "charge" for improvements in the stock of public capital. 32. Assigning a percentage of a specific tax may undesirably bias national tax policy decisions. 33. Of course, such legal requirements are to some extent only pro forma. The fungibility of money and the ability of local governments to alter other expenditures and taxes mean that requiring a grant to be spent on a particular activity does not necessarily imply that total (centrally funded plus locally fund- ed) expenditure on the activity goes up by the entire amount of the transfer. Indeed, in most cases it will not. Nonetheless, the empirical evidence (from Gramlich 1977 to McDowell 1994) suggests that transferred funds to some extent "stick"-that is, they are for the most part spent on the designated functions. Table notes Tables 1.1 and 1.5 Source for Alb1ania: Government of Albania; World Bank and vIMF staff esti- mates. Source for Bulgaria for general government: "Actual" data are from National Statistical Institute, IMF, and World Bank staff estimates. Nominal GDP in Bulgaria is 350 billion leva. Source for Bulgaria for subnational government: Ministry of Finance, Government of Bulgaria, "Aggregated Actual Municipal Budgets." Sourcefor Hungary: Report on the Implementation of the 1993 Budget of the Hungarian Republic, August 1994. Sourcefor Poland: Ministry of Finance, Government of Poland; World Bank staff estimates. Sourcefor Roniania: World Bank staff estimates. 64 DECENTRALIZATION OF THE SOCIALIST STATE Source fir Russia: Ministry of Finance, Government of Russia; IMF staff cal- culations. Source for Ukraine: IMF 1993, table 24; Ministry of Finance, Government of Ukraine. Table 1.4 a. The average figure is the simple arithmetic mean of the country popula- tion, excluding the capital city population. b. According to Feldman (1993), a three-tier system of local government has been established in Albania: districts, municipalities, and communes. The dis- tricts include the municipalities and comrnunes in all cases except the District of Tirana. The Municipality of Tirana is separated from the District of Tirana. The largest district in Albania in terms of population is Elbasan in Central Albania; and the smallest district is Has. Based on this source and United Nations (1992), it can be inferred that Tirana is the largest municipality, with a population of 245,000 in the city-proper. c. Based on Council of Europe (1993), the most populated oblast in Bulgaria in 1991 was Plovdiv, and the least populated was Michailovgrad. Among municipalities, Samokov had the biggest population, while Chelopech had the smallest. Source: Chapter 7; Council of Europe 1993; United Nations 1992, p. 28, and table A.12, p. 134; World Bank 1994b, vol. 2, table on "Regional Statistics: Population," p. 7; World Bank 1994h, tables 1 and 31; Feldman 1993. Table 1.6 Total revenue: For each of the above countries, total local government rev- enue consisted of own-source taxes and nontax revenues and shared taxes and transfers. Bulgaria: Own taxes: own-source property tax and inheritance tax. Nontax revenues: own-source fee and other own-source nontax revenues. Shared revenues and transfers: shared revenues (profit tax, income tax, and turnover tax) and other transfers. Source: Chapter 5, table 5.5. Hungary: Note: These are 1993 estimates. Own taxes: all local taxes. Nontax revenues: includes profits of locally owned enterprises, fees, old taxes, duties, and own revenues from sale of assets. Shared revenues and transfers: shared personal income tax, grants (norma- five and specific), transfers from within and outside general government, and social security fund transfers. Source: Chapter 2, table 2.3. Fiscal decentralization: From command to market 65 Poland: Note: Available information on local government revenue includes data on consolidated revenues earnings of "budgetary local enterprises" only. Revenue earnings of other local government-owned enterprises, incorporated in a different form, such as water and sewerage companies, public transportation enterprises, and so on, are not reflected in gmina accounts. The latter are known to account for the bulk of municipal enterprises. Own taxes: own taxes such as real estate tax and other taxes levied by the local governments (gminy). Nontax revenues: includes revenues from enterprises, stamp duties, lease or sale of assets, interest income, and other nontax revenue sources. Shared revenues and transfers: include local government taxes that are shared with the center, such as the personal income tax and corporate income tax, plus directed and general subsidies from the Treasury. Source: Chapter 4, table 4.2. Romania: Own taxes: Own-source direct and indirect taxes. Nontax revenues: own-source capital revenue and local activities revenue. Shared revenues and transfers: budgetary transfers (tax sharing and subsi- dies) and extrabudgetary transfers. Source: Chapter 6, table 6.3; World Bank estimates based on Romanian Ministry of Finance data. Russia: Note: These are actual 1992 figures provided by the Russian Ministry of Finance. Own taxes: include land and property tax and other local taxes. Nontax revenues: include stamp duties and other nontax revenues, and rev- enues from privatization. No information was available on the total revenue earned from stamp duties and other nontax revenues. Shared revenues and transfers: include personal income tax, enterprise income tax, value added tax and excise taxes, natural resource tax, and subven- tions and transfers to autonomous regions. Source: Chapter 9, table 9.2. Ukraine: Own taxes: include motor vehicle taxes and other local taxes. Nontax revenues: include timber and state duties, state enterprise privatiza- tion revenues, water fees, and nontax revenues. Shared taxes and transfers: corporate profit tax, excise tax, land tax, personal income tax, and net transfers from central government. Source: Chapter 8, table 8.4. Table 1.7 a. Currently, much of the capital expenditure on local infrastructure in Albania is being financed by foreign grants or loans on concessional terms. 66 DECENTRALIZATION OF THE SOCIALIST STATE b. Even though local governments in these countries (Bulgaria and Romania) are permitted to borrow in principle, in reality no borrowing exists- instead, governments carry arrears (nonpayment of bills). c. In 1991 Poland's subnational government borrowing as a percentage of total revenue was 1.4 percent. The government finance law for 1994 states that central government guarantees will be provided for some local borrowing within the prescribed limits of local debt services of 15 percent of programmed local rev- enues. The law also allows for facilitation of local government access to capital markets, through promotion of credit ratings of local governments and their investment programs. Generally speaking, local governments (gminy) do not have access to financial markets. Loans to gminy from the commercial banking system totaled a paltry US$8 million at end-1992. Table 1.8 a. Local governments in Albania have recently been given control of some small and medium-size enterprises. It is evident that small and medium-size enterprises, which are subject to Decree No. 248 (1993) on "Measures for Acceleration of Privatization of Small and Medium Enterprises," are being moved expeditiously into private hands. According to the decree, local govern- ments are to receive 28 percent of the net proceeds of privatization of small and medium-size enterprises. b. Telecommunications and electric utilities are under central jurisdiction in Albania. c. In Bulgaria, budgetary authority and responsibility for several important public services are not clearly divided among central and local governments. The Law on Local Self-Government transferred the ownership of all infrastructure related to local services to municipal governments. However, water, sewerage, and electricity continue to be provided by state enterprises. The city of Sofia is a special case because the water and sewerage services are provided by municipal enterprises. d. Ownership of telecommunications and energy utilities in Hungary is cen- tral; urban transport and water is shared between central and local government, while ownership of sanitation utilities is local. e. Retail units in Poland are mostly privately owned. f. Water and sewerage utilities in Poland are usually locally owned, while ownership of electric utilities is central. g. Ownership of retail units in Romania is almost exclusively local. h. In Romania local governments contract with regies autonomes (local pub- lic enterprises) for the delivery of most local public services. Table 1.10 a. The "Property Tax Law," passed in Albania in mid-March 1994, requires the central government to undertake property valuation, set rates, and administer the tax, with the proviso that 60 percent of the revenues are to be distributed to Fiscal decentralizationi: From command to market 67 local governments according to the source of the revenues. In this respect there is little sense in which the proposed tax is a local property tax. Bulgaria: b. Local governments received the entire tax paid by municipal and private enterprises, while the tax paid by state enterprises went to the central govem- ment. The turnover tax was replaced by a value added tax in January 1994. c. Municipalities kept 100 percent of tax levied on municipal and private enterprises operating in their jurisdiction and a 10 percent surtax levied on state enterprises operating in their jurisdiction. Hungary: d. The personal income tax share was reduced from 100 percent in 1990. e. Auto tax sharing is 50 percent; a road tax is not assigned. f. The poll tax is levied at the communal level. Local governments retain 100 percent of the tax. Russia: Note: All taxes in Russia are shared on an origin basis with the oblast where revenues are generated. g. The shared value added tax ranged from 20 percent to 50 percent, depending on the oblast (until 1994), and is a homogeneous 25 percent since 1994. h. Oblasts receive up to 25 percentage points of a 38 percent corporate income tax (first quarter 1994). i. Oblasts receive 60 percent of natural resource taxes. Okrugs received 80 percent in 1993. j. Producing oblasts received 25 percent to 65 percent of the export tax on petroleum, depending on the oblast (third quarter 1993). k. Oblasts receive 100 percent of all excises except excise on vodka and petroleum, for which they receive 50 percent and 30 percent, respectively. The 60 percent figure is based on actual revenue share in first quarter 1993. Ukraine: Note: In 1994 the government used for the first time homogeneous sharing rates for all oblasts and increased the number of regulating taxes. 1. In 1993 the value added tax was more heavily used as a regulating tax with nineteen different oblast sharing rates, ranging from 23.1 percent to 100 percent. m. National enterprise income/profits tax shares were set at three levels in 1993: 25 percent, 50 percent, or 100 percent. n. Oblast shares from excise taxes were set at three levels in 1993: 10 percent, 50 percent, or 100 percent. o. Real estate taxes imply land taxes only. In 1993, 100 percent of land taxes were retained by the local governments. Property taxes were abolished in 1993. Prior to that they were exclusively assigned to local governments. p. Local governments collect 100 percent of the company profit falling on enterprises owned by the local government. HUNGARY ® County (megye) centers * National capital : County (megye) boundaries / A1 , International boundaries Miskolc __'-_ Budapest is a municipality that hais ) 3 j' county (megye) status. It is also the capital _ Salg6tari6n Nyiregyhaza ofPestCounty(Megye). 2 EgerI \ Don fSe 2 J' 0 ~~~~~~' 11 r . tEger DorA- R/ -_, .O< (N>^,; Gyor : ES/TE; ' 0/.> ,- Debrecen C' %QN Tatabanyo1 BUDAPEST t , * A~~~~~~~~~~~~~~~~~~~~~IAI 5mbothely 3 ' A ~~~~~~~Szekesfehe"rva SzolInok C S .I Veszprem 5\ ,rt ;F; .%W'&\f Sz4kes F. ----r !ono - .- C6II.- Zaloegeruzeg . * | jr-Kecskemet' 47 \ @oto>' . , / > _B6k6scsaba | Kaposv6r Szekszard X- "d *1 ., :t DA >. n > n , ' Szeged "N -~~~~~~~- Thr ~ ~_ $Judre.crSdnoen dBwi eaol > _,,_Iff \~~~~ K SAC; NY ~~~~~~~~~~~~~~~~0 55 100 =n og-l o h o z 0z otlliw lon no mn KILO ETERS 2 Financing local government in Hungary Richard M. Bird, Christine I. Wallich, and Gdbor PNteri Hungary's bold, far-ranging reform of its system of subnational finances, undertaken against a background of national fiscal reform, is dramatically altering the expenditure responsibilities and revenue authorities of subnational governments. It is also changing their political relations with the central govemment. The new system of local government finance has many merits- both political and economic. It involves Hungarians throughout the country in a positive and democratic way with their local governments, at least in principle. It could also improve the efficiency of government by subjecting government actions to the scrutiny of responsible local officials and voters. But because the system is so new, there is still much to be learned and some serious decisions to be made, both by officials and by citizens, to ensure that the potential gains are realized. The new system of local govemment finance has two objectives: to free local authorities from the heavy hand of central control and to make them more responsible and accountable. The first objective is to be accomplished by removing all central control over local spending, whether financed by central transfers or by own revenues, and the sec- ond by providing new sources of locally controlled revenue. But the fun- damental inadequacy of the new local taxes raises the specter of undesirable outcomes from this well-intentioned experiment. Among the possible consequences: increased local demands for transfers from an already hard-pressed central budget, increased regional divergence in the provision of basic social infrastructure and services (particularly 69 70 DEC(ENTRALIZATION OF THE SOCIALIST STATr to the poorest elements of society), and increased pressure on local gov- ernments to engage in unwise entrepreneurial activities. This chapter outlines the changes introduced in the system of local finance as a result of the Law on Local Self-Government, one of the first laws approved by Hungary's newly elected Parliament in 1990. The Law outlined the basic structure of a modem local government system in the context of a legal, administrative, and fiscal environment that was highly uncertain, in part because it was passed only three months after the nation- al elections and much of the old institutional order still prevailed. Over the next four years, other laws elaborated and modified this system. The laws define the new assignment of expenditures between the central govern- ment and the 3,148 new local governments. They also set out the new local revenue sources. And they establish the economic foundation, property rights, and entrepreneurial functions of the localities. After presenting the basic elements of the new local and intergovernmental finance system, the chapter outlines policies that should avoid undesirable outcomes of the current system and help Hiungary achieve its goal of a smaller, more effi- cient local government sector. The new system of local government Under the system of national public finance in place until 1990, Hungary had a unitary system of government finance. Local governments had little independent revenues and few independent spending functions. The government was organized in a multitier system, with the center impos- ing its control over 1,523 local councils through 19 county councils. The local councils had no legally separate identity and were governed by the county councils. T'he local councils undertook a wide range of expendi- tures as agents of the central govemrment. Some fees and duties-such as tourist fees, stamp duties, and license fees--were collected locally, but the central government fixed the rates. Any added revenue that was needed to cover spending was negotiated with the central governrnent and chan- neled, mostly through the counties, from the central budget. The Law on Local Self-Government eliminated the middle tier of government, and at the same time the number of local governments mul- tiplied-to 3,148 in 1993 (see map of Hungary, p. 68)-as many of the local councils broke themselves up into discrete units.1 The new units have an average of only 3,482 inhabitants-2,834 if Budapest is exclud- ed-and three-quarters of them have fewer than 2,000 inhabitants. Many of these local governments are simply too small to provide efficiently all the public services demanded from them. Although the number of municipal associations is growing, only recently have local municipalities started to deal jointly with common problems, and there is still almost no Finiancing local government in Huingary 71 cooperation in service delivery or in administrative functions (Peteri and Wright 1994). As in other countries of the region, Hungary may have something to learn from countries with two-tiered structures at the local level (Bird and Slack 1993) and from those localities in Scandinavia that have cooperated in regional service provision (Soderstrom 1991). The Law on Local Self-Government redefined the rights and responsibilities of the two levels of local government (figures 2.1 and 2.2). The responsibilities of the regional bodies (the nineteen counties) have been dramatically scaled back (Davey 1990; Peteri and Wright 1994). The newly created Commissioners of the Republic-there are eight-coordinate, supervise, and review the constitutionality and legal- ity of local decisions. But they have no fiscal functions. A parallel system of thirty-six deconcentrated central ministerial organizations was creat- ed at the regional and county level to operate and supervise regional activities that range from supervising public health and education to providing information. The Commissioners of the Republic are sup- posed to coordinate these activities, but little coordination is yet evident. Counties remain responsible for interjurisdictional spending that serves more than one locality, but their revenue raising has been reduced. Most fees and duties previously retained by the county are now retained by localities, and counties no longer serve as a conduit for local finance from the central government. Governed by indirectly elect- ed bodies, the counties have no independent political basis. Cities with county status do not send representatives to county councils, and coun- ties have no power over these cities. Since local governments can take over any county service, counties end up providing only services that are costly-such as special homes for the elderly. Figure 2.1 Prereform administrative structure of central and local governments, Hungary, 1990 Central government and ministries City councils (19) Local councils (I .523)- Source Authors. 72 DECENTRALIZATION OF THE SOCIALIST STATE Figure 2.2 Administrative structure of central and local governments, Hungary, 1995 Parliament .-.-President Central ministries (13) Commissioners of the Republic (8) I 6 > 1 < < I ~~~~~~~~~~~~~Capital Village Deconcentrated with Cities Counties city munici- organs (36) 1 county (196) (19) (Budapest) palities ringh.ts (20) ( M (2 93 1 ) Note: Solid lines indicate subordination: dotted lines indicate supervision; dashed lines indicate the Ministry of Finance is the principal source of funding for local governments. Source: Authors. Local governments (referred to here as localities) are now directly responsible for most traditional local government functions. The corre- sponding shift in expenditure responsibilities has been paralleled in part-but only in part-by a shift in taxing authority. Localities receive grants from the central government to help them meet spending respon- sibilities, as laid out in the Law on Local Self-Government. Though unconditional, the grant is related in part to spending norms linked to these responsibilities. Localities are also authorized to own, borrow, and dispose of property and to establish, manage, and sell public enterprises. Local government finance in a macroeconomic context In the previous system of finance and budgeting the finances of subna- tional governments had few macroeconomic consequences. Transfers to localities and transfers to extrabudgetary funds and institutions were set to be consistent with the budgetary stance of the day. Localities might have wanted higher expenditures and transfers, but their access to funds was fixed. Under that system transfers to local government were about 10 per- cent of central government consolidated expenditures (table 2.1). With the Law on Local Self-Government, transfers rose sharply in 1991 to more than Table 2.1 Local finance in the national fiscal context, Hungary, 1985-94 (billions of forints) 1993 1994 Local finance category 1985 /986 1987 1988 1989 1990 199 i 1992 (estimated) (projected) Expenditures Total general consolidated expenditurea 632.9 719.0 768.4 892.3 1,041.5 1,194.5 1,345.6 1,778.1 2,004.9 2,403.3 Local expenditure as a percentage of total general consolidated expenditure 27.1 25.0 22.3 22.3 24.3 21.5 27.6 28.2 30.4 27.0 Transfers to local governmentb 66.5 75.1 78.2 97.6 123.3 1 13.3 190.7 219.5 256.6 298.4 Transfers as a percentage of total general consolidatedexpenditure 10.5 10.4 10.2 10.9 1 1.3 9.5 14.2 12.3 12.8 12.4 Revenues . Local revenue 147.6 156.9 169.5 205.2 243.0 300.5 391.7 508.4 577.7 637.7 Local own-source revenuec 40.3 48.1 53.0 42.8 53.2 62.0 67.4 91.0 106.0 106.8 Local own-source revenue as a percentage of local revenue 27.5 30.7 31.1 20.9 21.9 20.6 17.2 17.9 18.3 16.7 Transfers as a percentage of GDP 6.4 6.9 6.7 6.9 7.2 5.4 7.7 7.6 7.3 8.5 Consolidated deficit/surplus -11.8 -31.8 -43.5 -0.6 -22.6 9.7 -53.5 -158.3 -211.4 -275.5 Memo items Inflation in consumer price index (percent) 7.0 5.3 8.6 15.5 17.0 28.9 35.0 23.0 22.5 21.0 Average exchange rate (Ft/$) 50.1 45.8 47.0 50.4 59.1 63.2 74.7 79.0 91.9 107.5 - Not available. a. Capital and current only. b. Excludes personal income tax share and transfers from the social secunty fund for health care payments. c. Excludes personal income tax Source: For 1985-90, IMF 1992c, for 1991-94, Ministry of Finance, Government of Hungary. 74 DECENTRALIZATION OF THE SOCIALIST STATE 14 percent of consolidated expenditures and financed about 51 percent of local expenditures. In 1993 budgeted transfers (including those from the social security fund for health) rose because of additional spending respon- sibilities for health services, housing, and transport and accounted for near- ly two-thirds of local expenditures. The level in 1994 is largely unchanged. The system is still new, so it is premature to judge its long-term impli- cations for the national budget. Nonetheless, local dependence on transfers and the limited taxes assigned to the local level could give rise to macro- economic difficulties. Localities may have to turn to the central govern- ment for additional help with their newly assigned responsibility for social assistance, the demands for which would grow with a worsening econom- ic situation. There is also a backlog of unmet infrastructure improvements for which localities may, within the framework of the Law, have to turn to the center for financial support-unless other ways of financing infrastruc- ture are developed. If the center can hold the line, subnational government deficits will not imply any increase in the overall deficit. If, however, these demands are met by the center, they can contribute to aggregate demand pressures and inflation, if ultimately they are monetized. Hungary's macroeconomic position is uncertain. After three years of negative growth, the economic climate is starting to show signs of improvement, with manageable inflation and real growth in 1994 and forecasted for 1995. Real GDP fell by about 21 percent between 1990 and 1993 and real growth of 2.5 percent is expected in 1994. And rising prices, which put an additional strain on the economy during the transi- tion, are beginning to stabilize. Following an acceleration in 1990 and 1991, inflation fell to an average rate of about 23 percent in 1992 and 1993, and declined further in 1994. Still, Hungary's bleak economic situation continues to make stabi- lization difficult. The huge loss in output, and hence taxable income, has reduced government revenues and increased the budget deficit and unemployment. Since 1990, expenditures as a percentage of GDP have risen and revenues have fallen, resulting in an expansion of the deficit from 0.5 percent of GDP in 1990 to 6.4 percent in 1993. The fiscal deficit in 1994 is projected to deteriorate to more than 9 percent of GDP. The current account has also worsened, to some 10 percent of GDP. Unemployment, once negligible, was about 12 percent in 1994. The recent output expansion is encouraging, but insufficient to alle- viate the financial constraints facing the government. The change in gov- ernment in May 1994 makes the direction of the economy even more uncertain. The speed of reform remains to be seen, as does the govern- ment's response to inflation and unemployment. Regardless, local and intergovernmental fiscal policy will continue to be developed in the now very constrained fiscal environment of the central government. Finaancing local government in Hupngary 75 Given the assignment of expenditures and revenues in Hungary, the need for central transfers and shared taxes will continue to be large. In 1993 transfers (mainly the so-called normative grant) and shared taxes accounted for more than 70 percent of localities' total revenue receipts. Many of the services that local governments provide (education, health, infrastructure) are essential for Hungary's development. It is impractical to think that small local governments will be able to finance such services adequately with their presently assigned own resources. Even the large capital city of Budapest will find this difficult. And with Hungary's weak economy, local governments' open-ended responsibility for social assis- tance may lead them to turn to the center for additional funds, threaten- ing fiscal balance. Transfers to local governments are not an easily compressible chunk of the national budget. As argued in chapter 1, what is required is a grant system that is clearly defined, transparent, and that balances the needs of both central government and the localities in a consistent fashion. Some countries have opted for a formula to determine the aggregate transfer volume, such as a specified percentage of total national revenues. This approach is a compromise. The center gives up some revenue but insulates itself from unexpected and possibly escalating demands from localities. The localities avoid surprises such as stabilization-induced cutbacks in transfers. For Hungary, regularizing the size and distribution of trans- fers and strengthening local finances by broadening the local tax base should be priority reforms. The central government may also want to limit its commitment to providing open-ended targeted (matching) grants for local investments to a level consistent with macroeconomic constraints. Local government expenditure responsibilities Under the Law on Local Self-Government localities in Hungary have both mandatory and optional expenditure responsibilities. Among the tasks legally mandated for local governrunents are the provision of potable water, primary education, basic health and social services (including social assistance), public lighting and maintenance of public roads and public cemeteries, and guarantees of the rights of national and ethnic minorities. Neither the standard nor the level of provision is described in the Law, and the exact nature of the mandate is thus unclear. The Law also defines a range of other activities that are within the scope of local competence but are not mandatory. These include housing management, garbage collection, fire protection, public security, development of sports facilities, and so on. Indeed, localities can undertake any task not explicitly assigned to another body or government level. In practice, Table 2.2 Local government expenditures, Hungary, 1988-93 (billions of forints) 1988 1989 1990 j99/ja 992 1993 Share Shore Shore Shore Share Share Expenditure category Amount (percent) Amount (percent) Amount (percent) Amount (percent) Amount (percent) Amount (percent) Total expenditure 201.4 100.0 252.9 100.0 256.9 100.0 370.8 100.0 501.2 100.0 608.5 100.0 Education 68.2 33.9 79.4 31.4 107.3 41.8 122.3 33.0 152.7 30.5 185.4 30.5 Health 36.6 18.2 48.4 19.1 16.5 6.4 74.2 20.0 99.2 19.8 121 .7 20.0 Social security and welfare 8.8 4.4 17.7 7.0 18.3 7.1 22.2 6.0 50.4 10.1 37.2 6.1 Housing and community amenitiesb 32.3 16.0 16.5 6.5 27.4 10.7 17.1 4.6 13.3 2.7 24.9 4.1 Recreational, cultural, and religious affairs 4.5 2.2 12.8 5.1 15.8 6.2 14.8 4.0 20.3 4.1 25.9 4.3 9 Transportation and communication 13.6 6.8 27.9 11.0 - - 22.3 6.0 17.1 3.4 15.6 2.6 Other expenditures' 37.4 18.6 50.2 18.8 71.6 27.9 97.9 26.4 148.2 29.6 197.8 32.5 Current expenditures 146.7 72.8 194.3 76.8 201.1 78.2 311.1 83.9 408.3 81.5 497.0 81.7 Capital expenditures 54.7 27.2 58.6 23.2 55.8 21.7 59.7 16.1 92.9 18.5 111.5 18.3 Memo items Inflation in consumer price index (percent) 15.5 17.0 28.9 35.0 23.0 22.5 Average exchange rate (Ft/$) 50.4 51. 1 63.2 74.7 79.0 91.9 - Not available. a. Estimate. b. For 1991-93, includes only housing and water. c. Other expenditures include public order and safety, fuel and energy, agnculture, general public services, and other expenditures. Source: For 1988-90, IMF 1992c; for 1991-93, Ministry of Finance, Govemment of Hungary. Financing local government in Hungary 77 localities not only deliver public services, they also invest in infra- structure provision, such as energy and telecommunications, and foster employment and economic development. Local spendingfunctions The largest expenditure category is education, which accounts for more than 30 percent of total local expenditures (table 2.2). Primary education is mandatory for all localities. (Indeed, having a primary school and a gener- al practitioner define the minimum scale of a village [paragraph 52, Law on Local Self-Government].) Preprimary education and after-school day- care take a big part of total education outlays. A centrally set fee schedule applies to such services, but in some localities it covers less than 10 per- cent of costs. Moreover, central guidelines exempt some families from paying fees-families with more than three children, for example, pay reduced fees. This confusion of social objectives and financial policies con- cerns some localities, which believe they are better able to identify those in need. Secondary, technical, and vocational schools, while not mandatory, are typically financed by county governments or larger towns. Each of these functions is supported to a varying degree by the central normative grant channeled to the locality providing the service. In the health sector, localities act as agents of the National Health Insurance Fund in providing health services, including hospitalization in the larger towns. Services are determined by the national government, and local governments are reimbursed for the cost of providing services and medicine. Investment outlays are also a local responsibility, though financial support is available from central investment grants. Along with recent changes in the social security system (self-governance and separa- tion of health care and pension funds from the budgetary accounts), health service financing has also recently been modified. For instance, doctors can now privately contract to provide municipal health services. Housing is also a local government responsibility. The Law on Local Self-Government transferred ownership of some properties to localities-including parks, recreation centers, utility companies and their lands, commercial enterprises formerly owned by local councils, and other businesses. The Property Transfer Act of 1991 transferred much of the public housing stock (social flats) and commercial build- ings to local governments. But two years later the Act on Housing con- strained local autonomy in housing by defining the conditions for privatization of the housing stock and other premises, including a rent freeze that lasted until mid-1994. After the rent freeze ended, some local governments set new, market-based rents. However, the local elections of December 1994 made such actions politically unpopular. 78 DECENTRALIZATION OF THE SOCIALIST STATE With the transfer of communal housing and other assets to localities, maintenance on houses and properties has become a local responsibility. But rents, fixed until recently by the central government, remain well below the market. Losses of the local housing maintenance corporations were partly offset by central government subsidies until 1990, but then were fully phased out. For Hungary's cities, the fiscal implications were considerable (Alm and Buckley 1994). Localities are reacting in different wa's. Some are pushing to sell housing rapidly to reduce the recurrent costs on the budget. Some are making costly improvements to the prop- erties-hoping for a higher future sale price. Indeed, many localities appear to be developing vacant urban land in search of future profits. Considerable responsibility for administrating social welfare and sev- eral forms of social assistance has also been delegated to localities through the Law on Local Self-Government and the more recent Law on Social Assistance (1993). Responsibilities include the management of long-term social care facilities, such as homes for the elderly and for the handi- capped. Local governments receive funds from the national budget-the normative grant discussed later in this chapter-to support the mainte- nance of some of these institutions. Localities also finance home-care ser- vices for the old and for the handicapped, also supported through the normative grant. Other forms of social assistance prescribed by the laws constitute new municipal functions. Under the 1993 Law on Social Assis- tance local governments must now provide assistance for the unemployed who are not eligible for unemployment benefits. This outlay is partly com- pensated with an intergovernmental grant of 50 percent of the assistance provided. Discretionary housing subsidies are paid out of local budgets; child allowances for third children and maternity benefits are locally administered but centrally funded. The central government provides assistance for unemployment benefits, retirement pensions, and similar types of assistance. Local governments are then responsible for meeting the needs of those inade- quately served by the central welfare system. Central and local govern- ments' responsibilities for social assistance vary, with the central government responsible for setting guidelines for minimum incomes of pensioners and the unemployed and local governments tasked with establishing need, determining eligibility, establishing assistance levels, and funding outlays. There may thus be substantial differences across localities in the level of welfare provided, as well as in eligibility crite- ria, depending on the resources available to the locality (FAbian and Straussman 1994). It has been said that Hungary has 3,200 different social assistance policies. In general, in areas where the incidence of poverty is high, the per capita benefits provided are lower than in areas where poverty incidence is low (Jarvis and Micklewright 1996, forth- Financing local government in Hu1ngary 79 coming; World Bank 1994c). In villages local government councils make determinations based on applications brought by needy villagers. In larger towns a more formal process involves social workers and the Ministry of Social Welfare. To help local governments carrv out their responsibilities under the 1993 Law on Social Assistance, the allocation of social assistance funds to localities, which in the past followed a uni- form per capita allocation, now uses a weighted formula incorporating the percentage of inactive population, unemployment, and income capacity in each jurisdiction. Nonetheless, it is claimed that these weights have not resulted in sufficiently differentiated funding levels, and many local governments are having difficulty in delivering the programs mandated by law. Social assistance payments have risen from 0.2 percent of GNP in 1990 to 1.5 percent in 1993 (World Bank 1994c). The demands on central resources to enable localities to fulfill these tasks may further burden Hungary's economy. Local public service delivery by the private sector The nonprofit sector is increasing rapidly in Hungary. In 1992 there were 8,900 foundations and 18,000 unions, compared with 400 foundations and 8,500 unions in 1989. This sector often delivers local public services. Based on agreements with the local government, nonprofit organizations have taken over public functions, including schooling, and receive nor- mative grants from the central budget equal to the grants that would be provided to the local government for the functions they perform. Most (70 percent) of the institutions funded in this way are religious in nature. Grants to nonprofit organizations have increased substantially during the past three years: in 1992, 2 percent of the normative grant for schools was allocated to nonprofits; in 1993 this share increased to 3 percent. Are spending assignments appropriate? For the most part, the expenditure functions assigned to local govern- ments in Hungary appear logical. Given the small size of many Hungarian localities, however, there is a strong case for assigning some services-such as secondary education and hospital care-to larger units of government. These could be special-purpose governments or voluntary associations of local governments, such as in some Scandinavian countries (Soderstrom 1991). The rationale is that the ben- efit area for these services is likely to encompass several localities. Another important reason for involving larger units in the provision of services such as education and health may be the desire for more uniform quality standards. 80 DECENTRALIZATION OF THE SOCIALIST STATE Local government revenues Like most countries, Hungary assigns more expenditure functions to local governments than can be financed from the revenue sources directly allocated to these governments. The result of this vertical imbalance is that local governments generally depend on transfers from higher levels of government-the more so, the more spending they are charged with. Another problem of local finance in Hungary is that not all local governments are equal: there are big cities and small, urban localities and rural ones, rich units and poor. Providing local services fairly and efficiently in the absence of a well-designed revenue and transfer system can give rise to horizontal imbalance-differences between localities in the services they can afford (even when they levy Table 2.3 Local government revenues, Hungary, 1990-93 (billions of forints) 1990 Share Qf total Revenue source Amount budget (percent) Own revenuea 62.0 20.3 Of which, local taxes n.a. n.a. Own revenue from sale of assets n.a. n.a. Shared personal income tax (PIT) 74.5 24.4 Of which, PIT equalizing funds 4.5 1.5 Vehicle tax n.a. n.a. Grants 113.3 37.1 Normative grants 74.0 26.0 Per capita-based grants - - Expenditure norm-based grants - - Specific grants 29.0 10.0 Targeted and addressed grants n.a. n.a. Grants for disadvantaged localities n.a. n.a. Other grant funds n.a. n.a. Transfers from central government n.a. n.a. Social security fund transfers 50.6 16.6 Credit, bonds, and borrowingb 5.1 1.7 Total current revenues 305.6 100.0 Balance from previous year 5.9 Total revenues 3 1 1.5 Local government revenue as a share of GDP (percent) 14.6 Memo items Inflation in consumer price index (percent) 28.9 Average exchange rate (Ft/$) 62.2 - Not availabe. n.a. Not app cable. Financing local government in Hungary 81 exactly the same taxes). Hungary's system of local government finance deals only in part with these two imbalances. Moreover, the design of local taxes and the heavy reliance on transfers give rise to major effi- ciency concerns. Local own-source taxes The Law on Local Self-Government provides for local own-source taxes, shared taxes, and grants (table 2.3). The 1991 Act on Local Taxes assigns only five own-source taxes to local government. These taxes can be introduced by local governments, but are not mandatory. No other taxes or fees can be established at the local level, and maximum tax rates and the tax bases are fixed by the central government (table 2.4). 1991 t 992 1993 Shore of totoa Share of total Shore of total Amount budget (percent) Amount budget (percent) Amount budget (percent) 67.4 17.0 91.0 17.6 106.0 17.3 9.5 2.4 17.2 3.3 27.1 4.4 15.6 3.9 28.5 5.5 48.9 8.0 47.0 11.9 63.0 12.2 51.4 8.4 7.0 1.8 4.5 0.8 3.8 0.6 n.a. n.a. 2.3 0.4 2.5 0.4 190.7 48.1 219.5 42.5 256.6 41.8 148.6 37.6 169.8 32.9 207.5 33.9 58.9 14.9 67.4 13.0 83.1 13.5 89.7 22.6 101.9 19.8 124.7 20.3 42 1 10.6 49.7 9.6 49.0 8.0 17.3 4.4 22.6 4.4 18.3 3.0 5.0 1.3 3.1 0.6 1.6 0.3 19.8 5.0 24.0 4.6 29.1 4.7 3.9 1.0 23.4 4.5 23.1 3.8 67.1 16.9 80.8 15.7 91.6 14.9 4.8 1.2 7.5 1.5 35.7 5.8 396.5 100.0 515.9 100.0 613.6 100.0 20.4 39.6 63.5 416.9 555.5 667.0 16.0 18.6 17.5 35.0 23.0 22.0 74.7 79.0 91.9 a. Includes profits, fees, old taxes, and duties. b. Included in revenues under Hunganan accounting system. Source Ministry of Flnance, Government of Hungary: Bukva and Peten 199 1. 82 DECENTRALIZATION OF THE SOCIALIST STATE Table 2.4 Local taxes, Hungary, 1994 Bose of tax assessment Maximum rote of tax Exemptions Tax on buildings and property * Useful area of the building Ft 300 per square meter * All poor social flats in square meters * All properties less than 100 square meters * Living spaces with less than 25 square meters per person * All temporary lodgings * Garages of less than 16 square meters * "Corrected sales value" 3 percent of value * Properties owned by any of the building' income tax-exempt entity (churches, foundations, non- profit organizations, historical buildings, educational institu- tions. health care facilities) * All properties with prior exemptions Tax on undeveloped plots and land * Area of the plot in square Ft 100 per square meter meters * "Corrected sales value" I percent of value * Land owned by transport of the plota or telecommunications companies Communal taxes Communal tax on private persons * Premises serving dwelling Ft 3,000 per year per or other purposes and premise downtown plots Communal tax on entrepreneurs * Average number of Ft 2,000 annually per person people employed employed. If the activity is not (proprietors included) carried out during the whole by the entrepreneur year, the tax is calculated by months Tax on tounsm * Tourists and nonperma- Ft 100 per person each night * Children under 16 nent residents staying more or 4 percent of hotel charges, * Students than forty-eight hours to be collected by hotel or * Employed relatives homeowner * Tenants in social institutions or homeowners * Vacation residences Ft 300 per square meter of building's basic area Finiancing local governnent in Hungary 83 Table 2.4 (continued) Bose of tax assessment Maximum rate of tox Exemptions Tax on gross receipts/ local practice of industry Gross sales receipts of 0.8 percent of corrected * All final-level retail sales products sold or services net business turnover performed, net of value added tax paid a. The corrected value is 50 percent of the centra government's assessed price, which corresponds on average to 50 percent of actual observed market price. Source: Bird and Walich 1992: amendment (I1992) to Act on Local Taxes. Business tax. This local tax is a gross turnover tax that is levied at all stages of production/sale (except retail sales) at a maximum rate of 0.8 percent or 5,000 forints (Ft) a day for temporary business activities. It can be levied on all enterprises on gross sales revenue net of value added and other consumption taxes, less the purchase value of goods sold and the value of services provided by subcontractors. Communal tax. The communal tax can be levied on household dwellings or on businesses. If on dwellings, the occupant pays Ft 3,000 per dwelling annually, regardless of the number of inhabitants. This tax is suited to localities that have many government-owned social flats and cannot rely on property taxes for revenue. If levied on businesses, the enterprise pays a maximum annual rate of Ft 2,000 per employee. Land tax. The land tax is essentially an annual property tax on unimproved, privately owned land. It is levied at a maximum rate of Ft 100 per square meter, or 1 percent of the so-called corrected value of the land, where corrected value is defined (in regulations) as 50 percent of the government-determined assessed value. This value is typically equivalent to about 50 percent of actual observed market values. Publicly owned land used for purposes not exempt from income or other taxes is also taxable. Land owned by transportation companies and telecommunications companies is exempt. Property tax on buiildings. Local governrnents can levy an annual prop- erty tax-at a rate of 3 percent of corrected value or Ft 300 per square meter-on all privately owned buildings, such as flats, summer houses, garages, workshops, and other residential housing. As mentioned, the corrected value is about 25 percent of market value, (With more than 50 percent of housing in private hands, market transactions take place with some frequency.) The property tax base suffers from numerous centrally mandated exemptions, many of them extending well into the next century. 84 DECENTRALIZATION OF THE SOCIALIST STATE Under current law, for example, improvements to a property (such as a new roof or addition of a bathroom) exempt the entire property from tax for ten years. Also exempt are temporary lodgings, garages of less than 16 square meters, all so-called social flats (for the poor), all properties of less than 100 square meters, all living spaces of less than 25 square meters per person, and properties owned by any entity exempt from income tax. Records show widespread exemptions and almost exclusive reliance on area measures, not values. Tourism tax. The tourism tax is an extension of the earlier tourism fee, which could be levied only in resort areas. The current tax can be levied by all localities at Ft 100 per person for each night spent in a hol- iday hotel or private hostel, or 4 percent of the charge paid for the accommodation. It is payable only for stays of more than forty-eight hours. Students, employed relatives, tenants in social institutions, and children under 16 are exempt. The tax can also be levied annually at Ft 300 per square meter for vacation residences. Local taxing discretion Except for the business tax, which was new in 1991, these local taxes represented an extension of the taxes the local councils collected and administered under the previous system. The decision to assign these taxes to local government was driven by the objective of greater local independence and a separation of taxing functions between local and central governments. The possibility of building on the strengths of superior central tax administration and on the broader central tax base, accompanied by tax devolution or a system of surcharges, appears to have been unacceptable in a political environment where localities wanted to shed any semblance of central control. Localities can levy any or all of these taxes. For 1991, the transition- al year between the old system and the new, the old local taxes remained in place-at centrally mandated rates. (These taxes are described below in the discussion of other revenues.) It was expected that by end-1991, when the preexisting taxes expired, localities would establish tax rates (within the centrally mandated ceiling) for the new taxes under their new revenue authority (Sivak 1992). This was opti- mistic: local governments proved reluctant to introduce new local own-source taxes. In 1991 local taxes provided only 2.4 percent of local budget rev- enues, the same share as the preexisting taxes in municipal budgets. In 1992 the share of local taxes in local budgets (total revenues, including grants and shared taxes) was only 3.3 percent, and in 1993 this share Financing local government in Hungary 85 increased to only 3.9 percent (see table 2.3). About half the local gov- ernments have levied at least one of the new local own-source taxes, with most revenue gained from business tax (70 percent in 1992). The property tax on buildings provided 12 percent of local own-source tax revenue, and communal taxes 8 percent, in 1992. By mid-1992, 1,233 local governments had introduced some local own-source taxes, most frequently in localities with population between 5,000 and 50,000. As might be expected, the yield from business tax is highest in bigger cities (75 percent of it is collected in cities with population of more than 50,000). By contrast, per capita local own-source tax revenues- excluding the business tax and the communal tax on entrepreneurs- are higher in smaller municipalities. Local own-source tax revenues in 1992 were considerably less than projected by the Ministry of Finance (table 2.5), but by 1993 they had exceeded the projections. The choice of tax, additional exemptions, and rates (subject to a ceil- ing) is left to localities, subject to three conditions. First, two taxes may not be levied on the same object. This is interpreted in various ways by local government, but appears to mean that the same base cannot be taxed twice: a vacation home, for example, might be taxed under the property tax, the tourist tax, or the communal tax-but not more than one of them. Second, preexisting, centrally mandated preferences and exemptions must be respected. Third, locally set rates cannot exceed the centrally mandated maximums set in the Act on Local Taxes. All local taxes are treated as deductions from both the personal and corporate income tax base. In implementing these new taxes, local governments had to grapple with several issues. Hungary's tax burden at the national level was among the highest in Europe, and localities were loath to overburden residents with new taxes-citing rent arrears as evidence of low capacity Table 2.5 Local tax yields, Hungary, 1992 and 1993 (millions of foonts) 1992 1993 Tax type Estimate Actual Estimate Actual Buildings 5,000 2,119 2,500 2,600 Land 2,000 394 700 500 Communal tax on social flats 1,000 1,353 5,000 1,600 Tourism tax - 855 800 900 Business tax 12,000 12,110 13,000 21,300 Other taxes - 400 - 200 Total 20,000 17,231 22,000 27,100 Memo item Average exchange rate (Ft/$) 79 92 - Not avaiJable. Source Ministry of Finance, Department of Local Finance, Government of Hungary. 86 DECENTRALIZATION OF THE SOCIALIST STATE to pay. There was also concern that higher local taxes might increase the demand for social assistance. Administration and collection were also issues, with many localities favoring the exemption of small business. Local governments had little time-or administrative capacity-to set up effective tax systems. Nonetheless, the extent to which the localities have exploited their new freedom and tax discretion remains disap- pointing. Initiatives thus far have been too limited, almost as though localities were still in the command economy mode, waiting for direc- tion from the top and for implementation to be initiated, and decisions made, on their behalf. Duties,fees, and preexisting taxes In addition to the taxes defined under the Act on Local Taxes, localities continue to levy a range of other centrally regulated taxes, fees, licenses, duties, and penalties. In 1992 Ft 9.3 billion were collected from these levies. Fees on inheritances, property transfers, and gifts range from 5 to 10 per- cent of the value of the transfer, with land transfers paying a reduced rate of 2 percent. The county and the local government share these fees, which are one of the few independent sources of county government revenue. Fees are also paid for the use of public facilities (books, meals, dor- mitory rents, library fees, after-school daycare). Set by the central gov- ernment, these fees typically do not cover service costs. Shared taxes In the preparation of the Act on Local Taxes, consideration was given to sharing a number of central taxes with localities. Because of the unequal spatial distribution of revenues from the value added tax, sales taxes, and enterprise income tax, it was decided that the personal income tax would be the basic shared tax. In 1990 localities received 100 percent of the personal income tax collected two years before. These revenues were allocated to the place of residence of the taxpayer-in other words, on a derivation, or origin, basis. The tax is paid to localities with a two-year delay because the central tax administration takes this long to sort returns according to the location of the taxpayer's residence (which may differ from the taxpayer's workplace or tax office location). Local governments derived some 12 percent of their total revenues in 1991-92 from their share of the personal income tax, the major shared revenue for localities. In 1991 the allocation formula was changed; while local governments still received 100 percent of the personal income tax, 50 percent was allocated on the basis of derivation, and the remaining 50 percent was added to the amount available to all localities under the Financing local government in Hungary 87 normative grant. Starting in 1993 local government's derivation-based share was decreased to 30 percent, such that the personal income tax contributed only 8.4 percent of local revenues (see table 2.3). For 1995 the share is increased to 35 percent. These determinations are made in the context of the annual Budget Act. Since there was unequal regional distribution of personal income tax revenues (some localities-for example, where there is a concentra- tion of pensioners or low-income taxpayers-would have received none), the 1993 Budget Act also guaranteed that smaller localities would each receive a minimum per capita grant (which was larger in towns than in villages) financed from central government revenues. While this additional grant totaled only 1.5 percent of the normative grant allocated in 1991, 89 percent of localities-most with fewer than 2,000 inhabi- tants-received this equalization supplement. Another shared tax, the vehicle tax, was introduced in 1992. Motor vehicles-except those owned by public institutions or nonprofit organi- zations and used for public transportation or communal services-are taxed according to their weight. Local governments are the taxing authorities, but tax proceeds are shared equally with the central budget. With the vehicle tax only 0.5 percent of local revenues and the adminis- tration rather costly, this is not a cheap revenue source for localities. Grants Localities receive six types of grants from the central govemment. The first and most important is the normative grant, which is unconditional and broadly intended to redress the imbalance in the fiscal system. The second, introduced in 1991, consists of targeted grants for investment in a prespecified list of investment activities. The third type are the addressed grants for completion of specific unfinished investment projects, many of which were started under the previous regime. The second and third types of grants, approved annually by Parliament, replaced similar grants channeled through and administered by the former county coun- cils (discussed later in connection with investment finance). The fourth type of grants are centralized appropriations for specific purposes (the- aters, public transportation in Budapest, and so on). Fifth are grants for distressed areas. And sixth is the personal income tax equalizing grant (described above), as well as such special payments as compensation for centrally mandated salary increases. Normative grants. Since 1990 the center has given local governments a normative grant, the localities' single most important source of funds (see table 2.3). The grant, which is used for operations and maintenance 88 DECENTRALIZATION OF THE SOCIALIST STATE expenditures, is fully unconditional, in accord with the autonomy grant- ed local governments in spending and economic management. Although its allocation across localities is based on a complex formula that incor- porates twenty-four needs-based measures, in keeping with mandated local expenditure categories (called norms), and two per capita-based measures, there are no requirements on how localities actually allocate their funds within these categories. The per capita measures allow for some horizontal equalization. The grant is paid in twelve equal monthly installments. Neither the magnitude of the normative grant nor its allocation is set forth in the Law on Local Self-Government, which prescribes only that a grant must be given and that a formula, rather than discretion, be used for allocating grant funds. In this respect the normative grant is an important improvement over the previous system of bargaining. It is also distinguished from the former system in that localities now see the transfer as a right, not a subsidy. The Budget Act fixes the amount of the normative grant each year, following certain principles. Aggregate local spending is first estimated. Then the estimated revenues accruing to all localities from their own and other sources are subtracted to generate a measure of the fiscal gap. That is, personal income tax share, local own-source tax collections (as estimat- ed by the Ministry of Finance), targeted grants, addressed grants, social security funds, and revenue from fees, duties, and other centrally regulat- ed revenue sources are deducted from estimated local spending. This gen- erates a measure of the resources needed for the local sector as a whole. In a second stage the allocation of the aggregate transfer across locali- ties is determined. For example, the norms, or estimated expenditure needs of the localities, are adjusted to account for such factors as inflation. Expenditure norms were adjusted by 16 percent in 1991 and 18 percent in 1992. Although norms are thus intended to be set in real terms, both the per capita transfer and the norms may be readjusted to stay within the state's budget. There was no adjustment for inflation in 1994, and the norms were left at 1993 levels despite significant inflation during 1993. A change in design made the 1991 normative grant almost double in nominal terms over 1990, while recorded inflation was 35 percent. However, the localities' direct personal income tax share was reduced by 50 percent, so local nominal revenues from the grant and personal income tax together increased by only 32 percent (see table 2.3). In 1992 the normative grant was increased by 19 percent, while the official infla- tion rate was 23 percent. There is thus no guarantee that either the nor- mative grant or total local revenues are maintained in real terms. Under the Law on Local Self-Government, the normative transfer and its amount have been subject to ongoing modification. In 1990, with Financing local government in Hungary 89 localities receiving 100 percent of the 1988 personal income tax (PIT) (equal to 58 percent of the 1990 PIT), better-off localities (whose tax base was largest) benefited since the PIT was distributed largely on the basis of residence. In 1991, 50 percent of the 1989 PIT was added to the norma- tive grant transfer pool. Simultaneously, the norms governing its alloca- tion increased from twelve to twenty-two-to include, for example, kindergartens, primary schools, and $1.20 per capita to support theaters. In effect, the proportion of total funds allocated by residence fell and the share distributed for expenditure needs rose. Since 1991, apart from small changes, the factors used in the grant formula have been much the same (table 2.6). Four norms were added in 1993, driven by the need to accommodate political interests. There are major philosophical differences between local and central government about the functions of the norms. Localities take the view that all mandatory local expenditure functions defined in the self-government law should be supported by either central grants or delegated revenue capacity (such as the personal income tax). The center takes the view that it should not be financing all locally assigned expenditures-and that localities would have adequate financing if they fully used the own- source tax powers available to them. The truth lies somewhere in between. Own-source revenues are far from adequate to finance local functions-some 15 to 20 percent of GDP. But localities have not fully exploited the revenue sources they have been given. The norm-based grant looks very complex at first glance, but it has just two basic elements. The first are population-based weights consisting of lump-sum transfers to each locality plus a component based on the number of inhabitants and the population in defined age cohorts. Intended to provide a financial basis for small towns and some degree of equalization, this element accounts for about 40 per- cent of the total normative grant. The second element consists of "expenditure-related" norms relating to localities' expenditures on their main areas of responsibility, such as primary schools, care for the old, and care for the handicapped. Thus, localities will receive Ft 27,500 for each child enrolled in a municipally run kindergarten, Ft 70,000 for each mentally handicapped student, and so on (see table 2.6). Allocations were initially made only for local governments that provided such schooling or care. These norms, despite their apparent precision, are based neither on actual expenditures nor on actual costs, however, and the proportion of actual outlays they cover differs widely by sector and locality. These normative expenditure-related weights cover about 60 percent of the nationwide average cost of providing the services, with wide variations from service to service and place to place. In 1991 the per student grant Table 2.6 Normative grant allocation criteria, Hungary, 1991 and 1993 1991 1993 Total Share of Total Share of allocotion total grant allocation totol grant (millions expenditure (millions expenditure Expenditure norm Amount (forints) of forints) (percent) Amount (forints) of forints) (percent) Lump-sum grant to village municipalities 2 million per municipality' 5,804 3.97 2 million per municipality' 5,386 2.06 Lump-sum grant 2,000 per inhabitant 21,095 14.41 3,950 per inhabitant 41,612 20.02 Matching grant per forint of tourist tax collected 2 770 0.53 2 828 0.40 Housing subsidies 1,400 per inhabitant 14,776 10.10 4,680 per 20-30 year old 6,999 3.37 Subsidy for economically inactive citizensd 3,000 per inact ve person' 13,822 9.44 n.a. n a. n.a. Underdeveloped municipalitesc n.a. n.a. n.a. 2,200 per person 1.691 0.81 Social tasksc n.a. n.a. n.a. 2,200-3,600 per inhabitant 27.686 13.32 Institutional child care 210,000 per child 6,042 4.13 245,000 per ch Id 6,354 3.05 Homes for the elderly 147,000 per resident 5,786 3.95 186,400 per resident 7,455 3.60 Daycare for the elderly and the handicapped 24,000 per person 1,188 0.81 30.850 per person 1,247 0.60 Other homes for the elderly and the handicapped 40,000 per resident 110 0.08 87,000 per resident 232 0.11 Accommodations for the homeless' n.a. n.a. n.a. 87,000 per homeless person .. Institutional care for young handicapped and retarded children 172,000 per child 1,468 1.00 275,000 per child 2,273 1.09 Kindergarten (owned by municipalities) 15,000 per child 5,662 3.87 27.500 per child 10,462 5.03 Kindergarten for nationalities and minorit es 5,000 per child 70 0.05 5,500 per child 175 0.08 Elementary and primary schools 30,000 per student 33,938 23.19 41,000 per student 41,952 20.19 Elementary music schools 19,000 per student 1,383 0.94 25,100 per student 2,248 1.08 Mentally handicapped at elementary schools 56,000 per student 1,729 1. 18 70,700 per student 2,852 1.37 Secondary schools 44,000 per student 5,471 3.74 62,500 per student 9,750 4.07 Vocational secondary schools 54,000 per student 10,895 7.44 66,000 per student 13,1 17 6.31 Skilled-worker training schools 33,000 per student 6,879 4.70 42,100 per student 7,710 3.71 Workshops for apprentices and students at training schools 36,000 per student 1,286 0 88 40,600 per student 2,045 0.09 National, ethnic, and bilingual education 14,000 per student 676 0.46 16,500 per student 1,591 0.07 Residents of dormitories and hostels (elementary and secondary schools) 53,000 per student 4,556 3.1 1 66,000 per student 5,343 2.57 Local culture and public education 100 per resident 1,060 0.72 250 per inhabitant 2,634 1.26 Sports activities' n.a. n.a. n.a. 50 per inhabitant 527 0.25 Maintenance of theaters 450 per viewer 894 0.61 490 per viewer 5,162 2.48 Lump-sum grant for counties for education and other activities at regional level' 50 million per county 1,000 0.68 n.a. n.a n.a. Total 146,360 100.00 207,782 100.0 n.a. Not applicable. . Did not accrue until 1 994. Note. The average exchange rate in 1991 was Ft 74.7 = US$ 1: the average exchange rate in 1993 was Ft 91.9 = US$ I . a. Applies to municipalities with population of more than 200. For towns with fewer than 200 residents, grant is 10,000 per capta. b For population younger than 17 and older than 60 c. Norm added since 1991. d. Norm dropped since 1991. Sourcc: Ministry of Finance, Government of Hungary, 1993. 92 DECENTRALIZATION OF THE SOCIALIST STATE for secondary schools covered some 90 percent of service provision costs, that for primary schools, less than 50 percent, and that for old age homes, more than full cost. The reason for the differentiation allegedly relates to some presumptions made by the center about local priorities in service provision. For services provided largely to nonresidents or for services that might otherwise get low priority (such as care for the hand- icapped), coverage is fuller. The norms themselves do not distinguish between small and large jurisdictions, yet there appear to be differences in service provision costs between larger towns and villages. Generally speaking, the norms cover a smaller fraction of costs in cities, such as Budapest, on the grounds that their resource mobilization capacity is greater. Future reforms are intend- ed that will establish these costs more accurately and adjust the normative elements more by type and size of locality-despite the unconditionality of the grant-to improve the perception of equity and transparency. Grants for distressed localities. Beginning in 1991 the budget also allo- cated Ft 5 billion for localities that, through no fault of their own, cannot achieve basic budgetary balance. The beneficiaries include localities whose population structure (pensioners, the unemployed) implies a lim- ited tax base or personal income tax share. These funds, intended to meet only recurrent budgetary shortfalls, are for municipalities that can- not meet even minimal maintenance and expenditure needs. Funds can- not be used for investments or infrastructure, and localities whose budgets include investment expenditures are not eligible. The maximum allowed each locality was Ft 150,000 (population less than 1,000) or Ft 200,000 (population more than 1,000). Of 480 applications, 184 were accepted, providing 1.4 percent of local revenues. In the 1992 budget only Ft 2.7 billion went to distressed localities, with Ft 500 million set aside as a reserve for catastrophes or unexpected events. The balance was allocated by government decree (not by Parliament), with most going to small localities (population less than 1,000) on a per capita basis. The same principle was followed in 1993, although the grant volume was reduced. Localities with a recurrent bud- get deficit are eligible for this grant, but they must undergo a detailed approval process. Parallel to the decrease in this grant, an extrabud- getary Fund for Regional Development was established. Other grants. Finally, a large number of central budgetary payments -seventeen in 1993-go to localities as centralized appropriations for such items as social assistance, subsidies for public transportation in the capital, and allowances for central heating. These grants are almost entirely discretionary. Financing local government in Hungary 93 Local revenues: Issues and options The basic principles of local revenue assignment apply in Hungary, as elsewhere. First, local revenues should be collected from local residents, preferably in relation to the perceived benefits from local services. Of course, to the extent that nonresidents receive benefits, taxes on nonresi- dents may be justified. Second, where transfers are needed to finance expenditures, they should allow for some equalization and be used to encourage local governments to provide services with regional spillovers, but they should not be so generous as to discourage local rev- enue mobilization. From both these perspectives, reforms of Hungary's system seem warranted. Reforming local own-source taxes The new own-source taxes assigned to local governments in Hungary are not correctly aligned to benefits and suffer from other design prob- lems. Much the same can be said for local revenue from user charges and the old duties (to the extent they continue under the new system). The property tax is desirable and has significant revenue potential, but only in the longer run. The only information many local govern- ments (Nyiregyhaza City is an exception) have as a base for such taxes is square meters of area. Information on assessed value is rare although market transactions are becoming more common. The centrally man- dated exemptions specified in the law (and carried over from the old system) ensure that not much revenue is likely from this source for sev- eral years. Moreover, the property tax rate differential established in the law between vacant and built-up property is questionable, and local governments are not given any freedom to set their own tax rate on res- idential property, which is essential to make this tax a mainstay of responsible local finance. (The ability of local governments to tax busi- ness property should be restricted to restrain tax exporting-that is, shifting the tax burden to those outside the community.) Some recast- ing of the national framework for local property taxes is needed-and assistance in developing an adequate valuation base for such taxes must be provided-if the property tax is to become an important part of local finance (Kelly 1994). The primary new local tax with revenue potential is the tax on business turnover. One possible justification for such a tax is to charge enterprises for the use of local infrastructure, although user fees that recover costs and the property tax, which taxes immobile buildings and land, are in principle more desirable ways of achieving this. Moreover, with the explicit exclusion of retail sales from the base, this 94 DECENTRALIZATION OF THE SOCIALIST STATE tax encourages tax exporting and local manipulation of the tax system. Tax exporting (unless it matches benefit exporting) is the antithesis of rational local finance. However, until other sources of revenue can be tapped the business tax will be a necessary source of local finance. If it is to be retained, however, certain distortionary effects should be reme- died. For example, although the tax is designed to fall on net turnover, banks and insurance companies are taxed on gross turnover and thus are taxed more heavily than other commercial enterprises. The high proportion of social housing in Hungary-a proportion that will decline slowly-may justify levying a local communal tax on public housing tenants. If social assistance and general income support are provided primarily through national channels, the communal tax on social flats is consistent with the distributive rationale for local govern- ments not to misprice their services. But the communal tax will probably prove more difficult to administer in Hungary's increasingly mobile society than seems generally to be realized. And such a tax, never likely to yield much revenue, will play, at most, a minor role in local finance. As for the property tax, local governments should be given more leeway in specifying the rates. With respect to the tourist tax, in principle tourists should be taxed to the extent that they receive benefits from local services. Tourist estab- lishments (including second homes and cottages) should be subject to general local business and property taxes. In certain circumstances a special tax on tourists might even be justified (Bird 1992a), but it seems unlikely that most Hungarian villages fit such circumstances. There is no reason, however, to encourage an inordinate expansion of the tourist tax-to a degree that tourists bear a tax burden that outweighs the bene- fits received-by making it especially attractive (as is currently done through the bonus feature in the normative grant; see table 2.6), for local governments to impose taxes on nonresidents. Opportunity for greater own-source revenue of Hungarian local governments lies in the systematic exploitation of the potential for local user charges and the development of an adequate base for local property taxation, perhaps combined with a more adequate local vehicle tax. For the immediate future, limited local access to business taxation may prove necessary-until the property tax system is adequately devel- oped. The low-rate business tax needs to be reformed to cover retail sales and must be designed to treat all activities more uniformly. Since it is unlikely that even the richest local governments can be expected to finance most of their expenditures from local sources, local governments will continue to need access to one or more of the national tax bases-such as the value added tax or the personal income tax- either through revenue shares, surcharges, or the transfer system. Financing local government in Hungary 95 Reforming the transfer system Transfers, in many ways the heart of subnational finance, are intrinsically neither good nor bad. What matters are their policy outcomes-for alloca- tive efficiency, distributional equity, and macroeconomic stability. Intergovernmental fiscal transfers have several purposes in countries with decentralized governmental structures. First, such transfers close the fiscal gap between levels of government-helping countries achieve verti- cal fiscal balance and ensuring that the revenues and expenditures of each government level are about equal. Second, transfers achieve horizontal fiscal balance (equalization) among local governments. Third, transfers can be designed, if desired, to stimulate local fiscal effort-or at least not to discourage it. Finally, transfers can influence local spending decisions in accord with central preferences-for example, encouraging local gov- ernments to choose an efficient level of public services, where there are benefit spillovers. From these perspectives, several changes seem desir- able in the current structure of Hungary's central-local fiscal transfers. The normative grant has three important characteristics. The grant is unconditional-local governments can spend the money however they see fit. The total amount distributed to all local governments is entirely discretionary, set in the annual budget exercises. And the distri- bution formula contains both per capita- and needs-based components. Because the central government and taxpavers in general have a legitimate interest in what is done with grants to local governments, and the nation as a whole has a concern for ensuring that services such as education and health are available throughout the country at minimum standards, there is a case for limited conditionality of normative grants. For instance, there could be requirements for spending grant funds on education or health-or for local governments receiving grants to pro- vide services of a specified level and quality. But since local govern- ments can alter other expenditures and taxes, requiring a grant to be spent on a particular activity does not necessarily imply an increase in spending on that activity-resources are fungible. Still, there is consider- able evidence that conditionality is likely to increase the amount spent on designated services (Bahl and Linn 1992). In Hungary, to the extent the normative grant is allocated by indicators related to social problems (such as the number of registered unemployed), it appears widely approved that the grant be used for social expenditures. But there is no legal requirement to this effect. In the current political environment where local autonomy remains a sensitive issue it is probably too soon to make grants conditional. Still, the national government should make a reasonable effort to improve local financial reporting-for example, making the provision of financial reports a condition for receiving 96 DECENTRALIZATION OF THE SOCIALIST STATE grants-and to improve the information base on the provision of local public services. In the longer run, performance-based grants would seem appropriate. The size of the normative grant also requires consideration. It is entirely determined by the national government, which sets the transfer each year in the national budget. Because it is based on need, and less an estimate of local revenues, it has the appearance of a gap-filling transfer. This feature reduces the accountability of local governments and should be eliminated. Many countries use a nondiscretionary mechanism that fixes the size of the grant transparently. Such a system has substantial advantages for central and local budgeting. Since the amount of local transfer is predetermined, the central government is insulated from pressure to increase its support of local governments. And rather than be at the mercy of discretionary central policy, local governments can budget with greater certainty if they know that the level of central sup- port will vary with, say, income tax or value added tax collections, which can be more easily predicted. Regardless of how the volume of the normative grant is set, its dis- tribution formula should be altered. There now are two main elements in Hungary's formula: equalization (the per capita norms) and expendi- ture needs.2 A third essential element in a general grant formula is explicit allowance for the revenue-raising capacity of local governments. To incorporate these three elements, the basic formula for Hungary's normative grant should be some version of G, = eE, - tRi, where G is the amount of the grant, i refers to a particular municipality, E is some mea- sure of expenditure need (for example, the current normative grant for- mula, or need defined in relation to population, or another indicator that seems appropriate and can be satisfactorily measured), e is an assumed expenditure level for each unit of measured need, R is a measure of rev- enue capacity, and t is the rate at which this capacity is assumed to be tapped (or taxed) on average, across the nation. The critical elements in this formula in Hungary are its simplifica- tion and the measurement of tax capacity. The simplification involves changing the current eE, now allocated by twenty-four norms as proxies for needs. The broader study that was the basis for this chapter identified and explored several ways of simplifying the grant (Bird and Wallich 1992). The relationship between grant distribution and Hungary's demo- graphic indicators (closely related to some of the existing norms) was explored using regression analysis. The regressions indicated that the normative grant's current distribution across localities, presently based on twenty-four norms, could be approximated with a simplified formula based on only three variables: total population, the age cohort under 18, and the age cohort over 60 (see annex 2.1 at the end of this chapter). Financing local government in Hungary 97 The tax capacity element (tR) in this formula will shift grant funds from high-tax capacity to low-tax capacity localities. As shown later, it will also stimulate all localities, regardless of their estimated tax capaci- ty, to tax their capacity at the assumed rate-if they do not, their rev- enues will be reduced by the amount they fall below the assumed rate. If a locality chooses to levy higher taxes than those assumed in the grant formula, it can keep the extra revenue. Calculations of tax capacity, difficult in any circumstances, are par- ticularly difficult in Hungary today and it is hard to say the extent to which localities exploit their tax capacity. Estimating revenues collected in relation to tax bases is one approach. Among the tax indicators one might examine-as data become available-are collections from the new duties, from the shared personal income tax, the property tax, and the business turnover tax. In the absence of data on the tax base, alter- native estimates of local revenue capacity based on instrumental vari- ables and regression analysis could be explored. The broader study that is the basis for this chapter used population, and industrial employ- ment and personal income tax transfers (as a proxy for income levels), as independent explanatory variables proxying the tax base. Local tax revenues levied in 1989 were used as a proxy for taxes under the new tax regime and served as the dependent variable. These estimates of tax capacity indicated that smaller localities are exploiting (what small) capacity they have, while larger municipalities' taxes are below capaci- ty (Bird and Wallich 1992; see also annex 2.1). Analysis based on more recent and complete data is needed to draw firm conclusions on these matters and will become more important as the revenue sources assigned to the local level increase in scope. Three optionsfor reforming local governmentfinance Three options for reforming local finance were simulated to explore different dimensions of the redesign of Hungary's grant and local tax system. Each reform option meets the fundamental reform objective of improving local own-source taxes, simplifying the normative grant, and introducing a tax capacity measure into the grant. What differenti- ates them is treatment of the personal income tax (PIT) transfer. Option 1 reforms local own-source taxes and modifies the normative grant to incorporate some measure of tax capacity, leaving the PIT transfer as is. Option 2 distributes all of the PIT transfer by the same formula as the normative grant. Option 3 adds, instead, a proposal to make the PIT a truly local tax by converting it into a surcharge instead of a shared tax (for a discussion of the simulation technique employed, see annex 2.1). 98 DECENTRALIZATION OF THE SOCIALIST STATE Option 1. Option 1 incorporates a tax capacity estimate into the pre- sent grant formula, providing implicit stimulus to fiscal effort. Using a tax capacity measure (based on local PIT collection, population, and industrial employment) to proxy the tax base, and a simplified grant formula (needs indicators based on total population and the age cohorts specified earlier), the grant distribution simulated under option 1 would favor small and medium-size localities with up to 10,000 inhabi- tants (table 2.7 and annex 2.1). They would gain because the simplified grant emphasizes per capita needs indicators, and because the limited tax capacity of smaller localities results in a higher grant. Grants to municipalities with more than 10,000 inhabitants would fall because their estimated tax capacity is greater. The broad outcome: slightly more equalization than under the present system. In another altemative, the transfer norms (eE) could also be simplified or made more transparent and better tailored to localities' different cost structures. There is some ambiguity over just how weights in the grant should be calculated and chosen-that is, the relative expenditure and per capita-related weights and the extent to which the cost of specified services should be covered. One possibility might be to cover, as closely as possible, the local cost of providing each designated service. Since no data were available on these cost structures, no estimates were made for this approach to redesigning the grant system. Option 2. Another approach would be to modify the existing shared tax-cum-transfer system by allocating both transfers and shared taxes to local governments on a formula basis. Option 2 simulates the effects of adding the share of the PIT distributed on a derivation basis (35 percent in 1995) to the total normative transfer and then distributing the combined total by the simplified needs-cum-tax capacity formula in Table 2.7 Simulated effects of alternate grant reform packages on local revenue, by city population size, Hungary (percentage change in actual revenues received in 1991 ) City size Option I Option 2b Option 3' Less than 2,000 -0.3 8.9 -4.5 2,000-4,999 14.7 27.4 12.6 5,000-9,999 7.7 15.3 5.2 10,000-49,999 -11.3 -10.2 -I2.8 50,000-99,999 -20.0 -20.4 -17.3 More than 100,000 -15.0 -29.6 -11.9 a. Personal income tax (PIT) plus simplified grant p!us equalizing grant, ess representative tax capacity. b. Simplified grant applied to full revenue transfer, less representative tax capacity. c. P T surcharge plus simplifed grant applied to all transfers, less representative tax capacity. Source: Authors simuiations (see annex 2. ). Finiancinig local government in Hungary 99 option 1. In this case the entire transfer is made via the simplified for- mula less revenue capacity. Conceptually, the size of the transfer in this formula (together with conditional matching grants) determines how much the fiscal gap is closed. The formula then determines how hori- zontally equalizing the transfer is with respect to expenditure needs and fiscal capacities. Simulations of this approach show that it provides greater equalization than the current system or than option 1 (see table 2.7 and annex 2.1). Grant distribution favors cities with populations of less than 10,000, while cities with populations of more than 100,000 lose grant money. This is not surprising, since by definition derivation-based sharing of PIT benefits better-off areas where more PIT is collected. Allocating it instead on the basis of the formula would favor more equalization. Option 3. This option assumes the same simplified normative trans- fer and tax capacity measure; in addition, local governments are assumed to impose their own PIT-via a surcharge on the national PIT. This option thus remedies the inadequacy of the local own-source tax base by supplementing it with a local share of the PIT, but provides this share in a form that encourages local fiscal responsibility and effort. Surcharges on taxes such as the value added tax or corporate income tax are also possible in theory, but would be undesirable on administrative efficiency and equity grounds. Surcharging the PIT, as outlined in chap- ter 1, is preferred because a tax imposed on local residents is more suit- able for local finance than either a corporate income tax or a value added tax whose incidence is not likely to be local. Furthermore, pragmatically, localities already share the PIT in Hungary but in a way that contributes neither to accountability nor to efficiencv. Finally, the PIT surcharge would provide localities with a more stable tax base (wage income) and is relatively simple to administer in the Hungarian case.3 This option deserves attention because it could lower income taxes in Hungary without hurting the budget deficit and induce more effi- cient local expenditure. But the advantages are not costless. The approach might make administering the national tax system more diffi- cult. It would also require an amendment of the Law on Local Self- Government, which might not be politically popular. Furthermore, income tax payers could move from high to low tax jurisdictions if the dispersion of surtax rates among neighboring towns became too wide. This efficiency cost needs to be weighed against the advantages of local accountability. Moreover, to make this approach possible in a high income tax economy such as Hungary, the central government would first have to create some "tax room" by modestly reducing its own PIT rates. Local governments could then avail themselves of this room if 100 DECENTRALIZATION OF THE SOCIALIST STATE their expenditure requirements warranted it-and if local voters valued the expenditures enough to pay the taxes. To show how option 3 mimics the current PIT transfer we ignore, for the sake of simplicity, the two-year lag in the distribution of local tax revenues (and the equalizing PIT transfer). First, the central government would lower PIT rates by, say, 30 percent. (The reduction in the national income tax could be less, or more, than 30 percent; the remaining 70 per- cent could still flow to local governments, as at present, through the nor- mative grant.) Local government could then levy a surcharge of 42.86 percent on national taxes (0.3/0.7), collecting their 30 percent PIT share through the surcharge rather than through a transfer. The surcharge would be collected by the national government and remitted to the localities, and the national/local tax burden would be unchanged. (The two-year lag in the distribution of PIT revenue to localities means that the loss of central budget revenue from a 30 percent income tax cut would exceed the gain from eliminating the PIT transfer. To make the proposal budget-neutral, the local surcharge would have to equal the amount of the PIT transfer plus the equalizing grant.) In this example, the same normative grant and PIT transfer is paid by the national government to local governments in aggregate, and the same PIT transfer and normative grant is received by each local gov- ernment. In the formula set out earlier, eE would be the current nor- mative grant received by each locality plus the PIT revenues that would accrue to that locality if it applied a surcharge equal to 42.86 percent of the national PIT (again ignoring the two-year lag and the equalizing PIT). Although the results may look the same, the effects would be quite different. First, the local share of the PIT would now be a truly local tax. Localities would choose the rates they impose. If they levy lower taxes than assumed in the grant formula the result would be both lower total taxes and lower expenditures, since the total revenue amount (grant, PIT surcharge, and other revenues) at their disposal would be lowered by the amount they do not collect. There is thus incentive for localities to levy taxes at least as high as the tax rate assumed by the grant formula- in the example above, the taxes forgone by the national government. But since these taxes now have to be justified directly to their voters, there should be incentive to spend this money efficiently. If a locality decided to increase taxes above the current local PIT share, it would be able to increase its spending accordingly, since there is no reduction in the central normative grant as a result of additional tax effort. Again, however, there would be considerable incentive to ensure that additional funds are spent to the satisfaction of the voters-an incen- tive missing in the current system. Such a check on local spending seems Financing local government in Hungary 101 especially desirable when central transfers are likely, for political reasons, to remain unconditional. Option 3 thus gives localities access to a tax base that can finance more adequately the extensive range of services they are supposed to provide. It may even result in lower taxes. And it should increase the efficiency of local expenditures-whether taxes go up, down, or stay the same. These are major virtues. But it represents a major change and will require close consideration by authorities at all levels. And it is only a variant of the basic reforms needed in the present system-to strengthen local own-source taxes and simplify the normative grant. If option 3 were to be pursued, the amount of the central tax rate reduction and the phasing in of the surcharge are important transitional issues. It might be appropriate for the center to impose a surcharge on behalf of localities for the first year or two, giving local communities time to decide the rate they want to set. An additional transitional issue relates to the equalization implica- tions of this option. Some local governments have no income tax base or have one so small that they would not gain much from a PIT surcharge. It might therefore be inequitable to let local governments that do have a large tax base exploit it. But the unequal distribution of local tax resources is precisely what the capacity element (tR) in the proposed grant formula is designed to deal with. In poor localities revenue capaci- ty may be close to zero: so they will get a larger normative grant. In the proposed system equalization could be more fully and fairly achieved through the revised grant formula. Simulations indicate that allowing the PIT surcharge and allocating the equalization transfers through the revised grant formula, as in option 3, is slightly less equalizing than allocating it directly to poor localities, as in option 1. But the difference is negligible and could be off- set by additional variants of the grant formula. The assumption of a zero surcharge-no locality imposes any surcharge-is essentially captured by option 2, in which the PIT transfer disappears. Not surprisingly, this variant is more equalizing than the present system, benefiting villages and small cities (populations less than 10,000) because of the added emphasis on per capita indicators and on revenue capacity. Another attraction of option 3 for national authorities may be the explicit dependence of localities on the PIT-strengthening the case for strong national action to prevent the erosion (and evasion) of this tax base through such devices as nontaxable allowances. The national government should do this in its own interest; option 3 might give it more than 3,000 local allies in the political struggle to do so. These simulations are only exercises, and only indicative of broad direction- al effects. Data gaps and substantial additional work remain. The 102 DECENTRALIZATION OF THE SOCIALIST STATE three variants are only some of many that could be explored of the basic reform. Pros and cons of a personal income tax suircharge Combined with the basic reforms to local taxes and the transfer system suggested earlier, allowing local governments to impose income tax sur- charges has four advantages. First, local governments-at least the richer ones-would have access to a broad tax base to finance more adequately from local resources the range of services they are supposed to provide. Second, although the revised grant formula would encourage local fiscal effort, the result might be lower income taxes. Third, whether taxes increase, decrease, or stay the same, the accountability of local govern- ments and the efficiency of their expenditures should increase. Fourth, adjusting the parameters of the grant formula permits any degree of equalization and any level of total central-local transfers. While some of these advantages could be secured without allowing local governments direct access to the income tax, the objective of lowering overall taxes can be achieved only by these means. One main argument against local income taxation is administrative. Other arguments can also be made-that it reduces national fiscal flexi- bility or induces inefficient fiscal competition or inefficient resource allo- cation. However, when the functions local governments carry out are essential, expenditures have to be financed. And the local income tax approach seems capable of freeing the national government from the responsibility of financing such services while leaving it free to alter its owvn tax system. The central government could reduce the national income tax and allow local governments some flexibility in setting their own income tax rates. Moreover, such taxes are the prices of local public services and have fewer adverse effects on resource allocation or on fis- cal competition than the surtaxing or assignment of the value added tax or the corporate income tax (see chapter 1). On the contrary, some of their allocative effects may be considered desirable, as is the competition they induce in lower-cost provision of public services. The ownership role of local governments As a result of the Property Transfer Act, local governments own retail and comrmercial units, considerable vacant land, and most of the munici- pal housing stock. Localities also became shareholders in state-owned enterprises in proportion to the value of the land held by such enterpris- es. This portfolio may become an important revenue source for larger localities, but it is likely insignificant for most. Financing local governnment in Hungary 103 Revenues from owtierslhip Local governments receive income from the sale of state property and from rentals of flats owned by the locality. Rents in the past have been fixed by the central government, and rental property is typically man- aged by a local property management company, Ingatlan Kezelo Vallalat (IKV). Expenditures on maintenance and rental income are "extrabudgetary" and do not appear in the local budget. All rents accrue to the IKV and are earmarked for maintenance and other expenses (including property development and rehabilitation). A 1993 Housing Act allowed local governments to set rents in principle, but in practice rents remained frozen until June 1994. Hiousing sales are administered by the IKV, which receives a fee for the service. Proceeds accrue to the local budget. Developmental and sales activities are viewed in some localities as a significant source of future income, while in others there is concern about maintenance costs and the possibility of low sale prices. Many housing sales are financed by local governments at highly subsidized interest rates, and local gov- ernments are obliged to sell apartments and units such as garages and shops to tenants (Hegedus, Mark, and Pigey 1994). Localities and their agencies also receive revenue from the rental or sale of commercial properties, from the sale of locally owned enterprises, and from the sale and rental of plots (table 2.8). Data were not available on proceeds from these sources. In broad terms, rents accrue to the IKV. Sales of small plots, small or communal enterprises, and commercial properties accrue directly to the municipal budget. And sales of major enterprises or commercial properties (significant privatizations) are man- aged by the national-level Property Agency, which shares the proceeds with the locality and uses its portion to reduce the national debt. In addition to revenue from rentals and property sales, localities own commercial and industrial enterprises, from which they receive dividends or the proceeds from a sale or privatization. While privatiza- tion revenues are limited, localities have been optimistic about their abil- ity to get more revenue from this source. Many localities have established joint ventures with a domestic or foreign partner, using local assets as the locality's equity share. In localities with extensive land holdings, land has been a preferred equity contribution. Localities have also seen potential in developing and servicing empty land to enhance its equitv value. Tourism lodges, hunting parks, recreational facilities, and golf courses were examples in 1991. Localities that inherited impor- tant properties saw similar potential in developing these as contribu- tions to joint ventures, with potato processing, bakeries, construction, and wood products among the range of such enterprises encountered. Table 2.8 Local government assets: disposal, sale, and revenues, Hungary, 1992 Retention of Budgetary Asset Administroting body proceeds stotus Notes Residentiol housing Rentals IKV IKV Off municipal budget Rents controlled by central gov- emment. Local control over rents awaits passage of Housing Act. Sales City council or local Locality:a Sales proceeds on budget; Sales prices confined to 41 assembly determines I KV obtains fees for I KV fees off budget corrected" value-25 percent whether to sell handling sale of market value Commercial property Rentals IKV or a municipal IKV Off municipal budget service enterprise Sales of less than City council or local Locality; Sales proceeds on budget; 1.000 square meters assembly decides whether I KV obtains fees for 1 KV fees off budget to sell handling sale Sales of more than SPMA Shared evenly by SPMA Local proceeds on budget; 1,000 square meters and locality SPMA proceeds off budget Enterprise Sales of communal Locality Locality On budget Includes utilities transport, enterprises garbage, parks, baths, and so on Sales of noncommunal SPMA Shared evenly by SPMA Local proceeds on budget; Owned by locality as a result enterprises and locality SPMA proceeds off budget of J 940s nationalization Sales of Budapest's district entet-prises More than 1,000 SPMA SPMA Under preprvatization program square meters Less than 1,000 Locality Locality square meters Enterprise Locality I 8 percent On budget Under preprivatization program proftts and dividend to dividends locality Plots Rentals Sales Locality Locality On budget Ui IKV Ingatlan Kezelo Vallalat, a local property management company. SPMA State Property Management Agency. - Not available. Note: In Budapest sales proceeds and rental income may not apply as shown in this table, as there are frequently different arrangements. a. In Budapest 40 percent of the proceeds of housing sales go to the Budapest Housing Development Fund and 60 percent go to the districts. Source: Bird and Wallich 1992. 106 DECENTRALIZATION OF THE SOCIALIST STATE Entrepreneurial local governments In contrast to taxation, which was viewed as having low revenue poten- tial, local governments appear to see entrepreneurial activity as generat- ing stable dividends for the budget and providing the localitv with access to technology and to export markets and foreign contacts. Localities apparently felt direct involvement in such ventures was neces- sary. Providing conditions in which local business might flourish-as a market-economv local government might do-was seen as insufficient for bringing private enterprise to their region. There is danger that the exploitation of new property rights-driven by the pressure to activate idle property, and faced with the difficulties of selling property or businesses in the current environment-will lead to the emergence of businesses owned by local governments, which would limit both private competitors and the tax base. In market economies the rate of small business failure is high: reputedly only one in five such busi- nesses survives its first three years. Localities in Hungary are unlikely to beat these odds. Pressures will arise to subsidize local business to main- tain employment, and the role of the government in the economy will not have diminished. Moreover, entrepreneurial activity by localities is inconsistent with privatization. And it is a bottleneck to true decentraliza- tion. The sooner the risks of business activity are understood by localities, the better. An improved framework for local finance could provide a suitable occasion for moving toward this goal. It may be distributionally attractive and economically efficient to raise revenues from privatization instead of taxes. In a high-tax economy, the distortion of additional taxes may be quite high. Obtaining revenue from appropriate nontax sources is thus both efficient and equitable (Newbery 1991). But such revenue must be appropriately used. Some localities are using proceeds from asset sales (buildings and land) to finance current operations. This approach is clearly unsustainable. Asset sales are viewed as an attractive substitute for-and are delaying the development of-local taxing capacity, while stripping localities of valu- able properties. One argument against rapid sales is that many assets may have been financed with debt-some localities have substantial lia- bilities. In such cases, the proceeds should pay off this debt. The proportion of local revenues from asset sales increased from 2.3 percent in 1991 to 3.3 percent in 1992, equal to the yield of local taxes. While asset sales yield revenues in the near term, they are not a permanent source of funds, and localities should develop the capacity to raise recur- rent revenues. Asset sales could be used to finance new local investments, for example, thus changing the form of the locality's assets (from housing to local amenities) but leaving the locality's asset base unchanged. Financing local government in Hungary 107 Investment finance and local borrowing The financing of local investment is not directly addressed in the Law on Local Self-Government. But local governments have access to several types of investment grants. They may also, in principle, borrow. Investment granits Two types of grants are in use to finance local investments: targeted grants and so-called specific grants, designed to finance uncompleted infrastructure projects initiated before the introduction of the Law on Local Self-Government. Targeted (matching) grants are available to local- ities undertaking investments consistent with central government priori- ties. Under the Law on Local Self-Government, localities have a right to grants for all investments meeting the criteria set forth in the Law. The matching requirement differs by sector and by type of local government, ranging from 30 percent to 90 percent. For urban water and sewerage systems, for example, the grant finances 30 percent of total expenditures. In small municipalities, 90 percent of such investments are financed. Extrabudgetary funds also support development at the municipal level through grants and interest-free loans. Examples include the Water Fund, the Environment Protection Fund, and the Regional Development Fund. Since 1993 the Act on Municipal Targeted and Addressed Grants has regulated the allocation of this grant. This Act specifies water, sewerage, education, health and social care, and culture for grants during 1993-95. There is a 10 percent premium for joint investments (to encourage munic- ipal cooperation) and a 20 percent premium for localities with population below 1,000. In addition, there is a supplementary fund for localities with unemployment rates twice the national average if in a depressed area. In 1992 targeted grants were directed mainly to water, sewerage systems, and wastewater treatment plants (52 percent), education (28 percent), and health care equipment and homes for the elderly (18 percent). Localities forward requests to the Ministry of the Interior, which eliminates those not conforming to the Act's criteria. The 2,800 requests received in 1991 amounted to Ft 15.8 billion, of which Ft 8.5 billion met the criteria. The Ministry paid Ft 6.2 billion, consistent with the amounts allocated in the budget (1.7 percent of local revenues). A compromise was reached to fund eligible but unfunded investments in 1992. To avoid this problem after 1995, the unit costs of investments financed under targeted grants will be defined by the relevant national government departments. Local governments are constrained in the activities that these grants finance. When the central budget is invoiced, the bank where the local government has its accounts (96 percent of the municipalities have their 108 DECENTRALIZATION OF THE SOCIALIST STATE accounts at the National Savings Bank) automatically withdraws the amount from the local government's account. Localities are not allowed to sell assets or modify the function of any building for ten years if financed by a specific grant. If they do, they must pay back the grant. This contrasts strikingly with the fully unconditional nature of the nor- mative grant to finance recurrent expenditures. Specific (addressed) grants were intended to finance investments initiated during the earlier regime. In 1992 most were allocated for investments in hospital reconstruction and health (74 percent). The rest went to water projects (13 percent) and to education and culture (12 percent). Since 1993 these grants have gone for purposes defined by the central ministries, which are responsible for requesting the funds. There is no local matching requirement, but funds must be spent on the intended investment. In 1993 a mere 21 of 366 applications for such grants were funded. Eligibility criteria are broadly defined to include projects that are large, of regional importance, and for a diverse popula- tion. The local council is accountable to the Ministry of the Interior for expenditures, and operations and maintenance are the responsibilities of the local government. In principle, only the targeted investment grants are permanent fea- tures of the local finance system. They have economic and fiscal advan- tages in allocative efficiency (spillovers) and in the efficient use of scarce central government resources. Although making local governments more susceptible to central influence and control, targeted grants in principle have the political advantage of introducing local involvement, commitment, and accountability. Moreover, matching grants can con- tribute to equalization (horizontal fiscal balance). And like all other transfers, they can help resolve any basic fiscal mismatch (vertical fiscal balance) problem. But neither theory nor evidence provides clear guidelines to determine the appropriate matching rate for particular expenditure programs-let alone how these rates should vary with the characteristics of different local governments. The central government could restructure the grant system to identify more closely both spillover and central-local priorities. The government may prefer to extend finance through loans rather than grants. Addressed grants-if not transitional-should be combined with the targeted grants program and subject to the same matching principles. The close control over investment grants designed to corre- spond to central, more than local, priorities is in striking contrast to the autonomy granted localities in their recurrent budgets. A case was made, earlier in the chapter, for some conditionality on the latter, to the extent they are financed by central transfers. The split between grants for investment and normative grants for current expenditure Financing local government in Hungary 109 (and the division of responsibilities between the Ministries of Finance and Interior) also requires coordination between normative grants and investment grants. Local borrowing The Law on Local Self-Government grants localities unlimited borrow- ing authority, but bond issues and bank loans are the only financing sources now available. Short-term borrowing for liquidity management can be initiated by the local mayor, with financing from the local finan- cial institution at market interest rates. Long-term borrowing to finance infrastructure, investments, and property improvements requires the approval of the local assembly. There are some subsidized loans, but the majority of funds are borrowed from banks and carry market interest rates and medium-term (five to eight year) maturities. The National Savings Bank accounts for 99 percent of lending to municipalities. Bonds are another potential source of funds. Five municipalities have issued bonds, guaranteed by the central government, to finance infrastructure investment in gas, telephone, water, and drainage systems. There are also proposals to establish a special Municipal Bank that would facilitate access to credit by municipalities, and some municipalities are consider- ing the creation of regional banks. In obtaining loans, localities are permitted under the self-government law to mortgage properties owned by them as collateral, with the excep- tion of core properties, such as streets and public parks and areas. County guarantees of local borrowing-common in the past-remain legally pos- sible but are not likely. Counties no longer have secure revenue sources, and their role as a guarantor has correspondingly been weakened. The indebtedness of some localities is striking, stemming from local councils' borrowings for projects under earlier investment plans. (Under the earlier system, repayments due on any borrowings approved under the national credit plan would also be guaranteed.) This debt has become the responsibility of the new local governments, and for some it is a major burden. Data on borrowing since the inception of the Law on Local Self- Government show an increase in municipal borrowing. In 1991 Ft 4.5 billion in credit was taken locally. In 1992 the total was Ft 7.5 billion- only 1.5 percent of total local budget revenues. In 1993 the share of bor- rowing rose to 4.0 percent. Although most localities are cautious in the use of the borrowing authority, it seems likely to increase. Revenue system reform and the abolition of credit planning and central guarantees of local investment put municipalities in a new posi- tion. There are three causes for concern. First, their level of investment 110 DECENTRALIZATION OF THE SOCIALIST STATE financing is low, reflecting in part a large backlog of unmet needs. Second, the revenue base that can be pledged for repayment is limited- only 4 percent of total revenues is from local taxes.4 And third, borrow- ing authority is unrestricted. The localities' unconstrained borrowing and scope for entrepreneurial activity may be a regrettable mix. In most countries localities do borrow, albeit in restricted ways. These include bond finance (unusual in developing countries because of limited capital markets and the absence of long-term finance), borrow- ing from the commercial banking system, borrowing from a central gov- ernment loan fund, and borrowing from a municipal development bank funded by the center. In most countries, the central government or its agency sets parameters for local access to credit-not least because it is generally undesirable from a macroeconomic point of view to increase overall public sector borrowing. The options and controls appropriate for Hungary require further study. Since Hungary's economic perfor- mance is likely to improve, the case can be made that intergenerational equity justifies deficit financing of public infrastructure. If localities are given the ability to set user fees, the central govern- ment might consider replacing some investment grants with loans. User fees provide demand information for public services. This information is crucial for local governments allocating investment to the public sector. They are more likely to invest in projects for which the benefits exceed the costs if true prices are known. Hence, borrowing favors self-financ- ing projects, which places responsibility for decisionmaking with the local government-where it belongs-and reduces the burden on the center and on taxpayers. Charges for local public services User fees and other charges for public services had a limited role in com- mand economies-a tradition carried over to most transition economies, including Hungary. Under the former social structure wages were low, but housing and other services were either free or below cost. Today local governments must review housing rents, transport fees, water and sewer charges, and other prices for the services they provide. There appear to be user charges for municipal services in some locali- ties-connection charges for water, gas, and sewerage were encountered. Fiscal pressures are increasingly inducing local governments in Hungary (as in nearly all market economies) to impose such charges. In 1994 the authority to set water tariffs was delegated to local governments, although it is not clear that they or local water enterprises have the capaci- ty to assume this responsibility. A thorough review of local fees and charges is needed. The appropriate policy is usually to charge a price that Financing local government in Hungaiy 111 covers economic costs (long-run marginal cost), unless externalities or other market imperfections justify a subsidy. This approach provides cor- rect amounts and types of service to consumers willing to pay. Where the beneficiary of the service is hard to identify, or the benefits are generalized over the population, or the service is characterized by large benefit spillovers, tax revenues are the appropriate financing mechanism (Netzer 1966; Bird 1976; Meier 1983). Thus, education, police services, and fire pro- tection would not lend themselves to financing via user charges. User charges are likely to be viewed by local officials as a potential source of additional revenue. But their main value is promoting economic efficiency by providing demand information to public sector suppliers and by ensuring that local supply is valued at cost by citizens. Efficiency is particularlv important for local government since the main economic rationale for local government is allocative efficiency. Whenever possible, local public services should be charged for rather than given away (unless they are pure public goods or the explicit intention is redistributive). Indeed, one reason for using the residential property tax for local finance is that property values bear, however roughly, some relation to services of local governments. Unfortunately, most countries charge less at the local level than is desirable, and many charges are poorly designed. The first rule of good local finance should be: "Wherever possible, charge" (Bird 1976). The distributive consequences of charging for local public services is not necessarily regressive (Bird 1976; Bahl and Linn 1992). Fees, benefit levies, and user charges for locally provided services could be more fully exploited. Where private beneficiaries can be identi- fied (property owners benefiting from public improvements, recipients of daycare, and so on) and there is no distributive argument or externali- ties to suggest the contrary, beneficiaries should pay the correct price for what they receive, whether in fees, user charges, or special improvement levies or contributions. To the extent that local governments can set such charges, they should be encouraged to do so. Any centrally mandated fees imposed by localities should also emphasize cost recovery. In the transition, since the tariff increases implied by cost recovery may be quite high, some compensating mechanisms-lifeline rates, senior citi- zens fees-may be needed. In some cases, actual compensating pay- ments may be needed to make the large price increases palatable. These would need to be designed carefully to ensure that prices are consistent with allocative efficiency while income effects are offset. Conclusion Local government finance in Hungary is in transition, and much remains to be done to ensure the resulting decentralization of government 112 DECENTRALIZATION OF THE SOCIALIST STATE activities will be economically efficient and politically sustainable. The Law on Local Self-Government has assigned-and localities have taken on-far more expenditure responsibilities than before. Because local taxes are too limited to pay for these new responsibilities, local govern- ments depend on transfers from the central government. They have few incentives (or opportunities) to generate revenues or to spend efficiently. Despite its defects, however, the basic structure of the system offers sev- eral possibilities for reform that would make it both more efficient and more equalizing. It remains to be seen whether the changing political scene will encourage-or even permit local government finance to devel- op more efficiently, along the lines proposed here. Annex 2.1 Simulations of local transfers in Hungary: Simplifying the normative grant and incorporating revenue incentives This annex explains how the effects of the three options for grant reform described in the text were determined.5 The first option was to simplify the grant design by replacing capacity norms with indicators of need, such as demographic variables. The second option also simplified the grant allocation formula and in addition it allocated the personal income tax (PIT) transfers according to the formula. And the third option simpli- fied the grant formula and allowed localities to place their own sur- charge on the PIT. To develop a simpler grant allocation formula to be used in each one of the options, regressions were run to estimate the weights that should be placed on different demographic indicators. Own-tax capacity for each municipality was estimated based on regressions using PIT, population, and employment as independent variables. The total revenue that would accrue to each municipality was then determined for each option based on the derived formula grant, the derived tax capacity, and the available data. The data Three databases were used. The first consists of data on 3,069 municipal- ities, including data for all elements of the normative grant, and the absolute amounts of PIT and PIT-equalization grants in 1991. As simula- tions were focused on municipalities, county local governments were excluded. The second data set contains demographic and employment data for 1989 from the Central Statistic Office local database in which 3,032 local governments were characterized. Demographic and employ- ment data for 1989 were aggregated by municipalities according to their administrative status in 1990, and 998 units were developed and used in the regressions. The third database has municipal fiscal/tax revenue Financiniglocalgovernmnent in Hunlgary 113 data from fiscal year 1990 based on a sample of the 1,586 local govern- ments that existed in 1990, of which 1,301 municipalities were used. Simplify,ing the nornative grant Simplification of the normative grant design means a significant decrease in the number of elements included and a shift away from capacity norms toward a needs indicator-based allocation model. To accomplish this task of simplification the first question that needed to be answered was How strong is the connection between local public institutional capacities and population age groups within the same municipality? To answer this question the capacity indicator of institutions of a particular type was regressed on the population that utilizes the institution (table A2.1). A significant linear connection between the capacity indicators and the demographic variables is evident. The coefficients show that for every 1,000 youngsters there are 889 students in educational institutions. The number of rooms in social institutions and the age cohort represent- ing the elderly population are also highly correlated. Since demographic variables appear to be good indicators of need, they can be used as elements of the grant formula. Based on these results, the new grant design was determined by examining the relationship between the existing normative grant and the age cohorts suggested above. Other norms did not need to be related to age groups because many were already based on such groupings. Using grant allocation as the dependent variable, the appropriate formula weights on the demographic variables were determined. Population and two age cohorts were used as independent variables, all of which were significant (table A2.2). The coefficients represent the elements of a simplified grant formula-that is, how much should be allocated to a municipality according to these need indicators. The negative constant can be explained as the minimum level of population, rela- tive to the age cohorts, necessary for the central government to allo- cate grants to the municipality on a formula basis. Using this formula results in a lower grant to municipalities formerly receiving less than Ft 5 million, higher grants to small poor municipalities, and lower grants to large rich municipalities. improving revenue effort In addition to simplifying the grant, improving local revenue capacity is an important part of local fiscal reform. Own-revenue capacity was esti- mated from fees and charges, local taxes, and actual PIT collection. Municipal tax capacity can be measured by a representative tax system or 114 DECENTRALIZATION OF THE SOCIALIST STATE by regression, based on existing fiscal information. Because of information constraints, a representative tax system approach could not be used. So a regression model was used to measure tax capacity, where the dependent variable was the aggregated prior tax yield of each municipality and the independent variables were the available fiscal and real indicators-the PIT, population, and employment variables. The formula chosen to esti- mate tax capacity for further simulations was based on all of the above indicators (equation 4, table A2.3). Table A2. I Simplifying the normative grant: Regressions of selected capacity norms against age cohorts by municipalities, Hungary, 1991 BI Age 82 Age 83 Inoctive Dependent cohort cohort populotion vanable Constant [0-18] [60-x] [0-18,60 -x] R2 I . Students' -3 1.445 0.795 .. .. 0.9962 (-3.981) (890.321) 2. Studentsb -117.700 0.919 .. .. 0.9853 (-22.019) (450.245) 3. Students' -! 6.861 0.889 .. .. 0.9945 ( 1.585) (739.533) 4. Social institutionsd 1.247 .. 0.036 .. 0.9839 (1.753) (430.224) 5. Social institutionsd -3.427 .. . 0.019 0.9873 (-5.413) (485.006) Variable vvas not used in the regression. Note: N - 3,032. t-stat st cs are n parentheses be ow the regression coeffc ent. a. K ndergartens, elementary and secondary schools, vocational and skilled workers tra ning schools. b. Same as (a). without Budapest. c. Same as (a), plus kindergartens for nationalities ard minorities. music schools, elemen.tary schools for the handicapped, workshops for apprentices, b lingual education, and dorm tories. d. Child care, day home care for the e derly and the handicapped, homes for the elderly, and institutions for young peop e who are handicapped. Source: B rd and Walich 1992. Table A2.2 Simplifying the normative grant: Regressions against selected population variables by municipalities, Hungary, 1991 Dependent 8I 82 Age B3 Age voriable Constant Population cohort[0-18] cohort[60-x] R2 Local grant -845.81 13.134 .. .. 0.9969 (- .705) (981.73) Local grant -4,411.7 .. 60.947 -6.104 0.9992 (-12.463) (175.873) (-17.0) Local grant -845.81 13.134 1.141 -0.504 0.9969 (-1.705) (981.73) (51.535) (-64.111) Variable was not used in the regression. Note. N = 3,032. t-stat.stics are n parentheses below the regression coefficient. Source: Bird and Waiiicn 1992. Finiancing localgovernment in HutgarnV 115 Formula options Given these estimates, the total revenue to each municipality from employing option 1 was calculated as the sum of the PIT, the simpli- fied grant based on the new formula, and the equalizing grant, less the estimated tax capacity. Option 2 revenues were simulated as the sim- plified grant formula applied to the full revenue transfer (PIT + PIT- equalizing grant + normative grant) less tax capacity. And option 3 revenues were estimated from the PIT surcharge plus the simplified grant applied to all transfers less the estimated tax capacity (see also text table 2.7). These simulations are only illustrative and are at most the first steps in preparations for a new local transfer system in Hungary. Many questions remain. The fiscal data are incomplete: some counties have been excluded from the calculations, and there are many estimates in the data. And, absent comprehensive data on tax bases, especially incomes and profits, the calculations of capacity and the simulations employing these data are very tentative. But the uncertainty of the esti- mates does not discredit the idea. Preliminary as it is, this work is pre- sented here in the spirit of illustration, to indicate what might be refined by further work, and to point the way toward further improve- ments in calculations, which will depend in large part on availability of data. Table A2.3 Determinants of revenue capacity: multiple regression of (old) own-source taxes against selected tax base indicators, Hungary, 1990 Industriol Dependent vonoble Constont P17 Populotfon employment R2 I. Revenue capacity/ own-source tax -243.843 0.207 -0.147 . 0.8573 (-0.7) (15.755) (-- .445) 2. Revenue capacity/ own-source tax -378.280 0.150 1.546 0.8595 (-1.239) (15.640) (4.206) 3. Revenue capacity/ own-source tax -1,177.49 0.692 3.780 0.8397 (-3.280) (9.631) (10.631) 4. Revenue capacity/ own-source tax 37.806 0. 176 -0.248 1.737 0.8603 (0. 108) (12.052) (--2.406) (4.630) Variable was not used in the regression. Source Bird and Wallich 1992. 116 DECENTRALIZATION OF THE SOCIALIST STATE Notes 1. The counties of the former middle tier still exist but they no longer play the role of an intermediate level of government. They are now parallel authori- ties and unrelated to the localities. 2. For a general discussion of the design and principles of transfers see Shah (1994), but note that he argues in favor of unconditional grants on "federalist" grounds, which does not seem particularly appropriate in the Hungarian case. 3. The problem is residence. If people live outside of the jurisdiction in which they work, the benefit principle will be violated. 4. In principal, the unconditional grant or shared revenues could be pledged, but this has not yet gained currency. 5. This annex draws on Bird and Wallich (1992). A fuller description of the method used is described in that paper. (7t / -- . HUNGARY Li t---\ t 9 ' BUDAPEST District boundories ' \ r -- Budapest municipolity boundary Roman numerals denote districts. r XI 1,~ ~ ~ ~ ~ ~ ~~XV V V 1 2 'Y I I I- / .. . \IV ~ ~ ~ ~ ~~~\I XI IX~~~~~~~~~~~~~~x l r WIX' \E 1- > I '1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 I-1i . k )~~~~~~~~~~~~~~~~~XI /~~~~~~~~~~~~~~~~~~~~~ i~~~~~~~~~~~~~~~~~~~~~~~~~~ 5 10T D j_ \ RXIII. \KILOMETERS .~ ~ ~ ~~~~~~~~~. / .. T I - K\ ) r~~~~~~~~~~~~~~ .. . I,;d, * / L nObeR '_