/S305S Dec. e4 Economic Development Institute v; ,.of The World Bank Macroeconomic Management and Fiscal Decentralization Edited by Jayanta Roy EDI SEMINAR SERIES EDI SEMINAR SERIES Macroeconomic Management and Fiscal Decentralization Edited by Jayanta Roy The World Bank Washington, D. C. Copyright © 1995 The International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing December 1995 The Economic Development Institute (EDI) was established by the World Bank in 1955 to train officials concerned with development planning, policymaking, investment analysis, and project implementation in member developing countries. At present the substance of the EDI's work emphasizes macroeconomic and sectoral economic policy analysis. Through a variety of courses, seminars, and workshops, most of which are given overseas in cooperation with local institutions, the EDI seeks to sharpen analytical skills used in policy analysis and to broaden understanding of the experience of individual countries with economic development. Although the EDI's publications are designed to support its training activities, many are of interest to a much broader audience. EDI materials, including any findings, interpretations, and conclusions, are entirely those of the authors and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. Because of the informality of this series and to make the publication available with the least possible delay, the manuscript has not been edited as fully as would be the case with a more formal document, and the World Bank accepts no responsibility for errors. Some sources cited in this paper may be informal documents that are not readily available. 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, when the reproduction is for noncommercial purposes, without asking a fee. Permission to photocopy portions for classroom use is granted through the Copyright Clearance Center, Inc., Suite 910, 222 Rosewood Drive, Danvers, Massachusetts 01923, U.S.A. The complete backlistofpublications fromtheWorld Bank is shown in the annual IndexofPublications, which contains an alphabetical title list (with full ordering information) and indexes of subjects, authors, and countries and regions. The latest edition is available tree of charge from the Distribution Unit, Office of the Publisher, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Publications, Banque mondiale, 66, avenue d'l6na, 75116 Paris, France. At the time of writing, Jayanta Roy was principal economist in the Macroeconomic Management and Policy Division of the World Bank's Economic Development Institute. Library of Congress Cataloging-in-Publication Data Macroeconomic management and fiscal decentralization / edited by Jayanta Roy. p. cm. Papers presented at a Senior Policy Seminar on Intergovernmental Fiscal Relations in China, held in Dalian, China, Sept. 12-17, 1994. Includes bibliographical references. ISBN 0-8213-3409-3 1. Intergovemmental fiscal relations-Case studies-Congresses. 2. Decentralization in government-Case studies-Congresses. I. Roy, Jayanta, 1943- . 11. Senior Policy Seminar on Intergovernmental Fiscal Relations in China (1994: Dalian, China) HJ193.M33 1995 339.5-dc2O 95-34038 CIP Contents Foreword v Acknowledgments vii 1. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 1 Alan Roe 2. Intergovernmental Fiscal Relations and Economic Performance 39 Ved P. Gandhi 3. Intergovernmental Fiscal Relations in Australia 49 Paul Bernd Spahn and Anwar Shah 4. Comparative Federalism: Trends and Issues in the United States, China, and Russia 73 Roy Bahl 5. Russia's Dilemma of Fiscal Federalism 103 Christine I. Wallich 6. China's Reform of Intergovernmental Fiscal Relations in the Light of European Experiences 125 Paul Bernd Spahn 7. Intergovernmental Fiscal Relations: The Cases of India and Indonesia 163 Amaresh Bagchi 8. Decentralization and the Government Deficit in China 195 Rajiv Lall and Bert Hofman 9. Local Taxation: Principles and Scope 221 Paul Bernd Spahn 10. Intergovernmental Fiscal Relations in Canada: An Overview 233 Anwar Shah iii Foreword This publication consists of a set of papers prepared for a Senior Policy Seminar on Intergovernmental Fiscal Relations in China held in Dalian, China in September 1994. The seminar was organized jointly by the Dongbei University of Economics and Finance (DUFE), the Economic Development Institute (EDI) of the World Bank, and the Ministry of Finance of the People's Republic of China. It was directed by Jayanta Roy of EDI, as part of an ongoing program of activities in Macroeconomic Management and Policy. Thirty participants from China attended the seminar, including senior representatives from several provincial administrations. They were joined by senior officials from the International Monetary Fund (IMF) and the World Bank, and by distinguished experts from India, the United States, Germany, and Canada. Comprehensive tax reforms were undertaken in China in January 1994 with a view to broadening the tax base and simplifying the system of tax sharing between central and provincial administrations. The seminar was designed to alert participants from all levels of the Chinese government to the major elements in a strategy for effective intergovernmental fiscal arrangements and to help them build consensus for change. This book, which is intended to be of use to policymakers and practitioners, reviews the fundamental principles underlying a system of good intergovernmental fiscal relations; examines the development of the Chinese system of fiscal relations up to and including the reforms of January 1994; evaluates intergovernmental fiscal arrangements in a range of countries (including both developed and developing countries) with the purpose of drawing lessons for China; and presents views of Chinese and foreign experts on China's fiscal relations problems. The primary focus of the book is on center-state relationships; it only briefly addresses state-local relationships. Jayanta Roy edited and prepared the manuscript for publication. The summary proceedings of the seminar, as reported in Chapter 1, were prepared by Allan Roe who acted as seminar rapporteur. Vinod Thomas, Director Economic Development Institute v Acknowledgments The authors would like to thank Peter Knight, Christine Wallich, Shahid Yusuf, Ramgopal Agarwala, and Eliana Cardoso of the World Bank, and Ved Gandhi of the IMF for their support and advice in preparing the seminar. Kee-Cheok Cheong deserves special thanks for his valuable input at every stage of this project. Barbara de Boinville copyedited the papers, and Susmita Ghosh carried out with skill and patience the elaborate task of preparing and processing this book for publication. vii 1 Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar Alan Roe A Senior Policy Seminar on Intergovernmental Fiscal Relations in China was held in Dalian, China, September 12-17, 1994. The context for this seminar was the comprehen- sive tax reforms recently undertaken in China that will significantly broaden the tax base and simplify the complex system of tax sharing between central and provincial admini- strations. These reforms could strain many aspects of the relationships between national, pro- vincial, and local fiscal administrations. For example, the achievement of macroeconomic stability and a more equitable distribution of income between and within provinces ne- cessitates more resources being placed at the disposal of the central government. How- ever, this will need to be achieved at a time when the burdens on the subnational authorities are intensified by having to absorb new expenditures for social welfare as well as other expenditures associated with reform. In that sense, the timing of the semi- nar, just a few months after the latest fiscal reforms were introduced, was extremely timely. It also complemented the work already undertaken by the International Monetary Fund on the same topic. The seminar broadened both the scope of the issues under dis- cussion (the IMF's current focus is mainly on the establishment of a grants system) as well as the geographical extent of the Chinese participation.1 The seminar was designed to alert participants from the central and provincial levels of the Chinese government to the major elements in a strategy for effective intergovern- mental fiscal arrangements and to help them build consensus for change. To this end, the seminar involved in-depth explanations and discussions of experiences with intergov- ernmental fiscal arrangements in other :elevant countries, as well as presentations and workshop sessions focused explicitly on various aspects of the fiscal problem in China itself. The seminar was organized jointly by the Dongbei University of Economics and Fi- nance (DUFE), the Economic Development Institute of the World Bank (EDI), and the 1. For a synthesis of the earlier IMF-World Bank seminar on a similar theme, see E. Ahmad, Jon Craig, and Zuliu Hu, Conference on Intergovernmental Fiscal Relations: Summary Conference Proceed- ings, IMF (mimeo), October 25,1993. 1 2 Macroeconomic Management and Fiscal Decentralization Ministry of Finance of the Government of the People's Republic of China (MOF). It was directed by Mr. Jayanta Roy of the EDI. There were thirty participants from China, in- cluding senior representatives from thirteen provincial administrations-namely Jilin, Hebei, Shanxi, Anhui, Jiangxi, Henan, Hunan, Guangdong, Guizhou, Quhai, Ningxia, Liaoning, and Inner Mongolia. The proceedings comprised four separate elements, and the report presented here has adhered to this four-way division of material as closely as possible even though this re- sults in some minor contraventions of the chronological order of the seminar itself. The first element was the definition of the basic principles underlying a good system of inter- governmental fiscal relations. This discussion, led by Ved Gandhi of the IMF, is summa- rized in the first section of this chapter. The second element comprised presentations and papers, mainly from Chinese ex- perts, describing the gradual development of the Chinese system of fiscal relations up to and including the reforms of January 1994. This is summarized in the second section. I then cover the third element of the seminar-a series of presentations by interna- tional experts about intergovernmental fiscal relations in other countries, including Aus- tralia, Canada, Germany, India, Indonesia, Russia, the United States, and Switzerland. Finally, the views of Chinese experts on particular aspects of the country's fiscal rela- tions problems are considered. These views emerged from workshops during the five days of the seminar. Comparable reflections of the foreign experts regarding some of the same issues are also represented. In his introductory remarks, Mr. Liu Jubin, vice-minister of finance, drew attention to the previous initiatives to adapt the Chinese system of intergovernmental arrangements since the opening up of the economy in the late 1970s. These had included the 1980 sys- tem of "revenue-expenditure assignment and contracting," which was adjusted further in 1985. Then, in 1988, the government introduced a revised system that included "incremental revenue contracts" and "gross revenue sharing." Mr. Jubin argued that these systems had been reasonably well-adapted for their own times: they certainly provided both central and local governments with incentives to raise more revenue and moderate expenditures and, in this way, contributed to the broader reforms of the country. Reviews of these systems, however, identified important weaknesses. They criticized the lack of standardization and concluded that the systems were poorly adapted to the next stages of development of China's socialist market economy. Hence, further reforms were introduced at the beginning of 1994. These were integrated into the broader reforms of the Chinese tax system and sought, among other things, to clarify the functions of the different levels of government, to improve budget management, and to promote the aims of rational resource allocation and sound macreconomic management. Mr. Jubin noted that the reforms had been well received both nationally and internationally and were al- ready being reflected in faster revenue growth at central and local government levels. However, he recognized that some established market economies had been operating with systems of fiscal relations for as long as 100 years. Relative to this, China has taken only the first steps. He also noted that the basic framework now in place is still not a fully standardized one and that there are numerous practical problems in developing it further and consistently with the distinctive characteristics of the Chinese economy. In his opin- ion the search for these solutions should incorporate ongoing study of international practices in the area of intergovernmental fiscal relations. These practices should then be evaluated in relation to the circumstances in China. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 3 He noted that his own Ministry had already collaborated with the World Bank, the IMF, and other international agencies through symposia and seminars on particular as- pects of the problem. He fully expected that this latest seminar would provide an excel- lent opportunity for Chinese officials to learn more about the fiscal arrangements of other countries and to consider their suitability for China. He expressed his thanks to all the organizers for their extensive preparation and for arranging such a strong team of inter- national and domestic experts to lead the discussion. These sentiments were endorsed by the seminar director, Jayanta Roy, who noted that the importance of an optimal structure of intergovernmental relations in large economies had been recognized in the work program of the Economic Development Institute of the World Bank for several years. That work had culminated in a seminar in New Delhi in 1991 on Intergovernmental Fiscal Relations and Macroeconomic Management in Large Countries; at this seminar China had been represented. A similar seminar that focused on Russia had been held in April 1993. In reviewing the content of the present seminar, Mr. Roy agreed with Mr. Liu Jubin that foreign experts and the existing literature on the subject can provide important les- sons and principles about the design of an efficient multitiered system to meet national objectives. However, the foreign experts and literature cannot set the national objectives and cannot present a fully worked out system designed to meet them. Mr. Roy very much hoped that the Chinese experts present at the seminar would learn from the for- eign experts 'best practices" as well as the pitfalls but would then regroup to continue the work of designing a system specifically suited to Chinese circumstances. Fiscal Relations and Economic Performance: Some General Principles In the opening keynote address, Ved Gandhi from the Fiscal Affairs Department of the IMF sought to examine the effects that a system of intergovernmental fiscal relations can have on economic performance both in general terms and in the specific circumstances of China.2 In particular, he posed the question of whether the exceptional Chinese growth performance of 12 to 13 percent per annum in the previous two or three years might be damaged by a failure to modernize the existing fiscal arrangements. He began by defining a few basic terms used during the seminar (see Annex at the end of this chapter). He then addressed two specific questions. First, can the lack of an appropriate intergovernmental fiscal structure seriously affect the continuation of rapid economic growth in China? And is it possible to design a "good" intergovernmental fis- cal structure that will be conducive to continuing strong economic performance? Second, do countries generally and China in particular have intergovernmental fiscal structures that meet the tests of economic principle against which the design of the "good" system is assessed? Gandhi suggested that this was not the case generally or in China in par- ticular. Partly this is because of the conflicting demands involved in the design of such systems. 2. It should be noted that a later session of the seminar, and especially the presentation by An- war Shah, provided further definitions of the basic principles underlying a sound system of fiscal federalism. That presentation is not reflected in the write-up here because most of the points raised are covered more fully in his already published paper on this subject. See Anwar Shah, The Reform of Intergovernmental Fiscal Relations in Developing and Emerging Market Economies, Policy and Re- search Series, No. 23, World Bank, Washington, D.C., 1994. 4 Macroeconomic Management and Fiscal Decentralization Fiscal Relations and Economic Growth in General Gandhi pointed out that there are four main factors that influence directly the perform- ance of an economy. These are the microeconomic efficiency of resource allocation; the degree of macreconomic stability that is achieved; the extent to which the fiscal system of an economy is believed to be fair and equitable; and the degree to which the growth po- tential, conditioned by these first three factors, is actually achieved. He explained how fiscal interrelations could negatively affect growth through each of these four factors. First, the tax system could encourage a significant migration of the re- sources of either labor or capital between provinces and localities. This in turn might arise if the fiscal system resulted in wide differentials in the taxable capacities of different locations, as well as in their access to grants, subsidies, and other transfers from higher levels of government. The same problem might also be associated with a situation in which provinces and localities were assigned expenditure responsibilities, the benefits of which spilled over beyond their own geographical jurisdictions to result in expenditure levels considerably in excess of available revenues. Second, the achievement of macroeconomic stability could be problematic if the cen- tral government was not given the responsibility for those categories of public expendi- tures that could be adjusted reasonably quickly through the course of the business cycle. Similarly, the central government's responsibility for macroeconomic stability might be more difficult to achieve if the provincial and local governments were given the right to borrow on their own account or to run up large arrears. This is not to say that such bor- rowing should not be permitted-only that the possible consequences and the need for some controls should be well understood. Third, the equity in the fiscal system could be greatly impaired if redistributive taxes (the personal income tax, wealth and inheritance taxes) and redistributive expenditures (social welfare expenditures) were assigned too heavily to subnational governments rather than to the central government. As a final example, growth prospects would be impaired if provincial and local gov- ernments were assigned the responsibility for taxes on those essential raw materials and natural resources that happened to be located in their geographical locations, and if they also sought to levy excessively high tax rates on such resources. The burden of these rates would then be exported to manufacturing industries in other regions. Gandhi argued that if the economic principles underpinning these four examples were the only factors conditioning intergovernmental fiscal structures, then such struc- tures would tend to have relatively uniform features across different countries. In prac- tice, differences between countries, in terms of both economic and other characteristics, operate to prevent such uniformity. Strong Features of the Chinese System The Chinese system as reformed in early 1994 has several important qualities as judged against his general principles. First, most of the public services with benefits widely dis- tributed across the population (for example, national defense, internal security, external trade, other aspects of foreign relations, macroeconomic management) have been as- signed to the central government. So, too, have those expenditures where the geographi- cal spill-over effects are likely to be large (for example, higher level education and health, scientific research, technical renovation, new product development, and regional devel- Intergovernmnutal Fiscal Relations in China: Report of a Senior Policy Seminar 5 opment). On the other hand, those expenditures where the benefits accrue essentially to the local populations (for example, lower level education, basic health, and investment in locally owned enterprises) have been made the responsibility of lower level govern- ments. A second good feature is that the taxes applying to mobile economic bases, including some that can usually generate very large revenues (for example, enterprise profits tax and value-added tax) have been assigned to the central government. By contrast, taxes on immobile factors of production (such as land and property taxes, agricultural and hus- bandry taxes), have been assigned to subnational governments. So, too, have taxes with a mainly local incidence (such as the slaughter taxes, stamp taxes, city maintenance and construction taxes and taxes on contracts). Third, arrangements are in hand to ensure that the revenues from relatively buoyant taxes such as the value-added tax, and the securities transaction tax, are shared with pro- vincial and local governments even though they are collected centrally. This is an emi- nently sensible arrangement and one that is extremely important for the balancing of provincial and local government budgets. Fourth, the recent separation of the National Tax Service from Local Tax Services, and the withdrawal of the rights of subnational governments to grant tax exemptions and reliefs from central govemment taxes, is an important step that will considerably enhance the efficiency of tax collection. Fifth, the central government in China retains the power to legislate the rates and bases of most taxes assigned to subnational governments, and this seems desirable on the grounds of economic efficiency. It should also help to ensure that the burdens of provin- cial and local taxes, which can be exported beyond their own borders, will remain mod- erate. Finally, by restraining the powers of provincial and local governments in China to fi- nance their budget deficits through bank and other borrowing, the authorities seem to have improved greatly the prospects of effective macroeconomic control and inflation control by the central government. What Are the Remaining Problems? Notwithstanding the soundness of the basic structure referred to in the introductory re- marks by Mr Liu Jubin, Gandhi suggested that a number of problems remain to be re- solved. The first problem he referred to concerns the lack of clarity about the matching up of the expenditure responsibilities of the different levels of government with their corre- sponding tax assignments plus shared taxes. Serious "vertical imbalances"-chronic defi- cits for one or more levels of government matched by chronic surpluses elsewhere- could arise from the present system. If this indeed were the case, it would have major implications for the provision of essential services by some provincial and local govern- ments, as well as for the operations and maintenance expenditures on the investment capital needed for economic growth. Second, "horizontal imbalances"-imbalances between the fiscal positions of different governments at the same level (for example, local or provincial) - were also foreseeable. This being so, it was still unclear what the system of corrective transfers might be and also how the transfers might affect allocative efficiency and economic growth. 6 Macroeconomic Management and Fiscal Decentralization Third, it was his understanding that many of the important price subsidies for food grains, cooking oil, meat, and other essential consumer goods continue to be adminis- tered and controlled by governments at the subnational level. These responsibilities could give rise to budgetary problems for at least some subnational governments. Nor- mally, the main responsibility for redistributive policies lies with central governments. A related concern is that personal income taxation, as well as inheritance and gifts taxes, have been assigned to subnational governments, which probably have some authority to grant exemptions. Gandhi suggested that the redistributive effects of such taxes would be eroded as individuals at the margin changed localities in order to escape the full brunt of such taxes. In addition, any remaining powers to grant exemptions and reliefs would be likely to encourage undesirable competition between provinces and lo- calities to attract a fixed total supply of labor and capital. A fifth problem related to the treatment of natural resource taxes in China. It was his understanding that the revenues from these taxes are shared with provinces and locali- ties partly on the basis of their geographical origin. He wondered whether this was a de- sirable arrangement. In particular, was it right for the population of a particular region or locality to benefit disproportionately from such revenues just because of where the re- sources in question were located? Finally, he questioned the present system of grants and subsidies in China and asked whether such a system is adequately transparent and is effective from the point of view of encouraging subnational tax effort, improved fiscal discipline, and strengthened budget accountability. He conceded that he was not aware of the answers to all the problems that he had posed and was naturally less well informed than the local participants about the par- ticular historical, political, and cultural factors in China that needed to be factored into their resolution. However, he presented his ideas as a useful organizing agenda for the discussion that would follow in the remainder of the seminar. Discussion from the floor noted that the good and bad features of the Chinese as- signment of taxes, as noted by Gandhi, had been substantially preserved by the January 1994 reforms. The major change had been that the important new value-added tax (VAT), which is likely to attract 40 percent or more of total tax revenues, will share those reve- nues between central and provincial administrations. The share of revenues accruing to the central government, which until 1992 had been about 39 percent of the total, was ex- pected to increase. Why then had the provincial authorities accepted the new arrange- ments? Apparently, the higher tax base and prospective revenues from the VAT will re- sult in a higher guaranteed revenue base for subnational administrations, as well as a rising share for them of any revenues over and above that base. A more serious problem was felt to be the absence of any real scientific basis for as- sessing the expenditure needs of the local administrations. This assessment is done at present by incremental adjustment to a base-year allocation, which itself was not scien- tifically established. A related issue of concern was that some of the tax revenue sharing arrangements, in- cluding the new arrangements for VAT sharing, were ad hoc. Some participants argued that little serious effort has been made to link these sharing arrangements to any assess- ment of the taxable capacity of different provincial and local governments, or to their ex- penditure responsibilities. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 7 On the basis of the discussion of Gandhi's paper, participants earmarked three main topics for in-depth consideration by working groups: the manner in which expenditure responsibilities should be assigned between central and provincial government admini- strations; the manner in which taxation responsibilities might best be assigned between the various tiers of government; and the manner in which an intergovernmental transfer system might be designed to support the needs of vertical and horizontal equalization between and within the provinces. Working groups were set up to consider each of these matters, and their reports back to the main seminar are described later in this chapter. The Main Features of the Reformed Chinese System The next important task of the seminar was to define and discuss the present Chinese arrangements in more detail, including the nature of the reforms recently introduced. This task fell to Xu Shanda, Chen Gong, Ding Xianjue, and Liu Minhyuan. The Chinese System Liu Minhyuan began by providing a broad perspective on the Chinese arrangements. From the earliest days of the People's Republic, the major taxes were levied on a nation- wide basis including some that would have been better regarded as subnational taxes. Only a few minor taxes (for example, the slaughter tax and the tax on the licensing of ships) were assigned specifically to provincial and other subnational authorities. This was because the imperatives of a highly planned economy saw no obvious role for any real autonomy in subnational taxation. During the period 1953 to 1980, tax revenues in total amounted to only 50 percent of total fiscal revenues and, within this, subnational taxes amounted to only a few percentage points. Both Liu Minhyuan and Chen Gong explained that with the general economic reforms of the 1980s, matters began to change. Beginning in 1983, a comprehensive reform of in- dustrial and commercial taxation was carried out. This had the effect of raising the over- all tax share in total fiscal revenues quite considerably. The reform also included the in- troduction of the fiscal rationing system. Until 1988, revenues from some taxes were shared between different jurisdictions according to a predetermined formula. In essence, the system provided that the revenues from specific taxes would be shared between cen- tral and provincial governments according to tax sharing contracts that varied across re- gions and also over time. The idea was that the subnational governments should have just enough revenue to provide a reasonable basic level of service in their areas, with the balance of revenues being surrendered to the center. In 1988 the system was put on a more negotiated basis whereby different provinces were able to establish different mar- ginal retention rates for the tax revenues concerned. This system, however, had many problems, not the least of which was the declining share of central government revenues, especially after the 1988 reforms (51.6 percent of total revenues went to the center in 1981 but only 38.6 percent by 1992). Another problem was the regional imbalances that the system tended to promote: the more prosperous regions kept and invested all revenues above a certain base, while the poorer regions failed to reach that base and needed special subventions. In addition, the incentives within the system encouraged too many quick payoff investments in terms of revenue (for example, tobacco industries yielding high excises) rather than more basic industries. 8 Macroeconomic Management and Fiscal Decentralization Finally, the system lacked uniformity across regions as well as certainty in the revenues generated. Xu Shanda and other Chinese participants explained that the focal point of the 1994 reforms is the designation of three important taxes-the VAT, the new securities trading tax, and the resources tax-as shared taxes. All other taxes are now assigned explicitly to either the central government or to the subnational governments. The other major inno- vation has been the establishment of separate national and local tax services, effective since July 1994. As regards the shared taxes, it was explained that 25 percent of the 1994 VAT reve- nues will be apportioned to the provincial governments with the rest to the center. Al- though the main securities market activity is in Shanghai and Schenzhen, the tax re- sources that are generated from this activity come from many different parts of the country and hence a 50:50 sharing of the securities tax revenue between the center and local governments has been agreed. As regards natural resources, only the tax take from off-shore oil will be regarded as national, while all other resource-based taxes will be shared. As regards the main assigned (nonshared) taxes, the central government has been assigned customs duties and the new excise taxes as well as almost all other indirect taxes. The local govermnents have been assigned a number of minor taxes as well as the revenues from the enterprise income tax for those enterprises that are owned locally. The new system is regarded as something of a compromise but one that should leave no province worse off than in 1993. This is because parity with the 1993 outcome will be achieved by a central government transfer to any province that, on the basis of the new system, achieves less revenue than it retained after tax sharing in 1993. At the same time, the new tax sharing system, especially for the VAT, preserves some incentives for the faster growing provinces. If a province's actual VAT and excise duty collections under the new system exceed its actual 1993 transfers to the center, the center will transfer back an amount equal to 30 percent of the VAT plus excise duty increase compared to 1993. Presumably, this means that the local share of the VAT revenues in some provinces will gradually rise above the 25 percent level. In addition to eliminating some of the inefficiencies of the previous system, the new system is expected to lead to a significant increase in centralization: the share of central taxes in the total is expected to rise from 27 percent to about 80 percent by the end of the decade. However, the part transferred back to the subnational level would also rise from a presently negative amount to more than 30 percent over the same period. This would mean a rise in the retained or accrued revenue at the center from less than 40 percent of the total tax-take to well over 50 percent by the year 2000. 3 A corresponding explanation of expenditure assignments was provided in the paper by Ding Xianjue. He noted that there were some broad listings of the expenditure re- sponsibilities assumed by central and local governments respectively. In the postreform period these assignments had been amended to give the local governments a great deal more power in areas such as enterprise promotion and local economic development more generally. However, the system was characterized by considerable ambiguity at the mar- gin and by several manifest weaknesses that needed to be resolved. It was quite common, in relation to functions such as disaster relief, environmental welfare, and new settle- 3. This explanation is derived in part from China: Macroeconomic Stability in a Decentralized Econ- onty, Country Economic Memorandum, World Bank, August 1994. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 9 ment, to observe considerable confusion about where the expenditure responsibilities actually lay. In recent years, when the fiscal resources of the central government have been under strain, the center has failed to discharge some of the functions that are its notional re- sponsibility-for example, key construction projects-and also has resorted to passing down certain of its functions to the subnational governments. The practice of the central government deciding on policies, and the subnational governments providing the money to pay, unfortunately has become more common. This problem has been compounded, Ding Xianjue explained, by the practice of faster growing provinces financing quick pay- off industrial projects and at the same time looking for central government funding of certain infrastructure projects. Meanwhile, lower revenue provinces are unable to find the funds to discharge even their basic functions. The net result of this essentially ad hoc decentralization of expenditure functions has been significant loss of central control in relation to macroeconomic management. While these problems were widely recognized, it was noted both in the presentations and in the discussion that the 1994 reforms had not really progressed far in defining ex- penditure responsibilities. However, the speakers offered a number of suggestions, many of which conformed to the principles presented earlier by Ved Gandhi. One was that with the evolution of the market culture in China, the government at all levels should withdraw more and more from direct investment in competitive industries. This would leave more resources for infrastructure projects. Certain major highway, railway, avia- tion, and shipping projects could be designated as national ones requiring central gov- emnment financing. Similarly, infrastructure projects providing narrowly local benefits should be mainly financed by local governments. In areas such as education, health, in- temational relations, scientific research, and cultural promotion, the assignments could follow the classical principles discussed above. Finally, other weaknesses were identified: in recent reforms the absence of clear ar- rangements to allow local government borrowing consistent with national macroeconomic objectives; the absence of clear proposals for a grants mechanism that could handle the horizontal imbalances not resolved through revenue sharing; and re- maining limitations on revenue autonomy at the local government level. Decentralization and Macreconomic Control In a later session of the seminar, Rajiv Lall of the World Bank linked these assessments of the present arrangements in China to an analysis of how the Chinese system of fiscal re- lations had affected the country's problems of macroeconomic management and control. He also assessed the extent to which the recent reforms might be expected to improve upon what had gone before. Although China has enjoyed exceptionally high growth rates in recent years, growth has been increasingly associated with a risk of rising and destabilizing inflation. It was the view of the World Bank that inflation in China is conditioned by deep structural causes. The present arrangements for, and reforms of, fiscal interrelations are an impor- tant part of the problem and its possible solution. In particular, the high level of fiscal decentralization in the past had contributed to a low degree of tax buoyancy, and this in turn had intensified the inflationary pressures coming from the government's own budget. 10 Macroeconomic Management and Fiscal Decentralization Looking more closely at this diagnosis, Lall noted that the inflation rate in China in July and August 1994 was about 3 percent a month and that the primary cause of this had been the monetary financing of a large government deficit. However, the total of budget- ary expenditures no longer provides an accurate guide to the extent of government ac- tivity in the economy. Prior to China's economic reforms, that total was more meaningful because "public sector" and "government" activity were largely synonymous. Now that state-owned enterprises (SOEs) have achieved various degrees of autonomy from central and subnational government, the government's involvement in economic activity needs to be measured in terms less broad than the total public sector borrowing requirement but broader than the narrow budget deficit. In his opinion the most appropriate concept for this purpose is the consolidated government deficit, which reflects the financing needed to cover both the government's own expenditures as well as the government- directed expenditures by the rest of the "public sector." This consolidated deficit has risen in recent years and now amounts to about 9 percent of GDP. At the present time, it is predominantly financed using monetary methods (that is, borrowing from the Central Bank). However, a deficit of this size and financed in this way is not consistent with the achievement of control over inflationary pressures. Nor would be a deficit of the same size but financed through the issue of government bonds, Lall argued. Simulations have shown that this would rapidly increase the interest ele- ment in total government expenditure and would lead to unsustainable deficits, even assuming that economic growth continues at a high rate. This being the case, it is clearly important to ask why the government deficits have become so large and what might be done to reduce them. The first explanation suggested by Lall is the sharp fall in the share of budgetary revenues in GDP, which went from 34.4 percent in 1978 to 15.4 percent by 1993. This, in turn, is associated strongly with the fall in budgetary contributions from the SOEs-only 4 percent of GDP by 1992 compared with 20 percent in 1978. This is a direct consequence of economic reform and the greater autonomy and profit retentions allowed to enterprises. A second explanation lies in the progressive devolution of administrative authority to lower levels of government, a fea- ture of China's reform experiences since 1978. Lall argued that the complex system of revenue sharing described above had shifted the de facto control over tax policy to the subnational authorities. This had included active subnational authority negotiations with local enterprises leading to tax incentives/exemptions not authorized by the center. In addition, local tax bureaus were far less vigorous in their collection and remittances of taxes to the center than they were supposed to be. These problems, taken in isolation, need not have caused the large increases in the consolidated deficits that have occurred. But, as Lall emphasized, central government expenditures have not been reduced in line with the reduction in the available revenues. On the contrary, weak control over the expenditure side of the budget has been an im- portant cause of the deficit problem. Three aspects of the postreform period are signifi- cant here. First, the investment planning/control of the central government, through the State Planning Commission (SPC), has deteriorated. This is because of the progressive transfer of responsibility for SOEs to the local authorities (80 percent of output was con- trolled in this way by 1988) and also because of local authorities' limited incentives to exercise the same tight control over SOE investment approvals that was exercised for- merly by the SPC. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 11 Second, the "investment hunger" manifest by many local authorities -specially in the priority sectors including social and physical infrastructure as well as basic industries (for example, steel and coal) and so-called "pillar industries" (automobiles and electron- ics)-has led to a big increase in the priority investment program of the government from 8.4 percent of GDP in 1989 to over 11 percent by 1992. This tendency has been exacer- bated by the fact that neither central nor local governments bear any direct investment risks: the consequences of bad decisions have been, in effect, socialized via losses in non- commercial banks. The whole issue of what actually constitutes a "priority" also has been clouded by the decentralization process. Third, the SOEs, in spite of reforms, still bear a large part of the burden for providing basic social services such as health care, child care, and some schooling, and they are also required to maintain excessive work forces. This results in an ongoing and justified SOE claim on the government for special assistance in addition to any help needed to cover genuine operational deficits. Lall noted that explicit and implicit government subsidies to the SOEs at the present time amount to at least 3.5 percent of GDP-equivalent to 1.5 times the narrow budget deficit. In short, the decentralization process of the reform period combined with the lack of real progress on many aspects of SOE reform accounts for a significant part of the large and rising consolidated deficit of the central government. In this context, Lall felt sure that the reform measures recently put in place repre- sented a good framework to tackle at least a part of the problem he had described. The tax reform measures, in particular, were extremely ambitious in their scope-more so, it can be argued, than those in any other country. The reform measures recognize that macroeconomic stability in China calls for a substantial increase in the ratio of tax reve- nues to GDP. Lall agreed only partly with a questionner who said that raising the taix:GDP ratio in the short term might be difficult given the continuation of a large social burden on the SOEs-traditionally a main source of tax revenues. Lall pointed out that economic activity in China was continuing to diversify away from the SOEs and that substantial additional revenues, even from the enterprise sector, were a real possibility. So far the results of the tax reforms seem to be encouraging. Nominal tax revenues in the period January to July 1994 turned out to be 24 percent higher than in the corre- sponding period in 1993. However, nominal GDP is also rising fast, and so the size of any increase in the tax : GDP ratio is still unclear. In addition, certain aspects of the reforms give rise to concern. First, the effect of the reforms on government revenues was likely to be dampened by the special treatment to be allowed to soften the impact on certain adversely affected en- terprises. For example, there is a grandfathering arrangement to hold down the sales tax burden for foreign enterprises and joint ventures. Second, the scale and the breadth of the reforms mean that their implementation poses a tremendous challenge. Lall was not certain that the preparations, training, and equipping of the new national and local tax services were yet adequate for the task in- volved. Third, the increasing expenditure responsibilities of the government in the future and, in particular, the absorption of the social burdens from the SOEs, argued for a further broadening of the tax base. Lall suggested that prime candidates for consideration in this regard might be environmental taxes, the personal income tax, revenues from divesti- 12 Macroeconomic Management and Fiscal Decentralization tures, property taxes, and increases of several fuel taxes, some of which are very low by international standards. Fourth, Lall agreed with earlier speakers that the reform proposals seemed to be much weaker or more vague in relation to expenditure control than in relation to taxation. The resolution of the problems associated with the control of public investment, in particular, required a stronger framework than at present involving a much clearer definition of the national public investment program (NPIP). Japan's Fiscal Investment and Loan Program (FILP) might provide one useful model to consult in this regard. In any event, the new framework ought to include a clear definition of the role of government even in relation to those investments projects falling outside the budget. It would also be desirable to have the funding of all public investment approved by the National Peoples' Congress. Under the present arrangement, the State Investment Plan is approved only by the State Council and is subject to numerous changes through the year. As a final example of the gaps in reform so far, he repeated the need to assign expen- diture responsibilities more clearly between the various levels of government.4 This is not a trivial task. The Chinese authorities might begin by identifying the best practice, least cost provision of government functions between provinces and the center based on international experiences. National consensus then needed to be built regarding the divi- sion of responsibilities. This should be codified in a new law on intergovernmental fiscal relations. These changes on the expenditure side should be accompanied by arrangements to provide the local governments with more revenue authority to ensure that they can meet their expenditure responsibilities. If properly done, this should enhance the accountabil- ity of the local governments and discourage them from searching for extra budgetary sources of revenues, most of which would be inefficient compared with more traditional sources of tax revenue. Comparative Arrangements in Russia and the United States: The Main Lessons for China A series of sessions sought to put most aspects of the Chinese fiscal relations problem into a comparative context. Roy Bahl of Georgia State University opened this next stage of discussion by presenting a paper that compared the situation in China with the situa- tions in Russia and the United States. He began by considering the general advantages and disadvantages of fiscal decen- tralization. China is already quite decentralized, although it has some strong centralist tendencies. He noted that there is, above all, an economic efficiency argument for decen- tralization. So long as preferences differ across subgroups of the population, national welfare is more likely to be maximized if local communities are able to express their own preferences and so have a say in persuading local governments to provide a mix of serv- ices that match these preferences. The noneconomist might couch this argument in terms of "getting the government closer to the people," but the recommendation is broadly the same. 4. For a more complete assessment of the possible gaps in the reforms, see Chapter 8 by Rajiv Lall and Bert Hofman. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 13 However, this theoretical argument for decentralization is likely to be far stronger in already industrialized countries than in transitional ones. One reason for this is that the theoretical model of fiscal assignment has been developed in the context of industrial countries and is strongly reliant on democratic processes of budget-making and on the high degree of responsiveness of both the tax effort and the expenditure mix to changes in relative prices and income. In these circumstances, the efficiency losses associated with interference from a higher level of government can be considerable. In transitional economies, such as those in Russia and China in the 1990s, the effi- ciency gains from decentralization may not be so obvious or large. This is partly because voter preferences are not so readily translated into budget outcomes and also because of the limited dependence of some parts of govermment on elections and so on voter prefer- ences. Decentralization is more likely to be justified the greater is the electoral account- ability of local govemments; the greater is the subnational taxing power to capture sig- nificant proportions of community income increments; the greater is the subnational dis- cretion in shaping the budget and setting tax rates; and the greater is the efficiency of the local government in administering an efficient tax system and collecting the taxes. This tableau of conditions conforms with present circumstances in the United States quite well but does not, he suggested, fit the current situation in China. At the same time, the arguments for fiscal centralism are stronger in transitional economies. There are for four main reasons. First, periods of major economic transition are often extremely difficult from the viewpoint of macroeconomic stabilization, and this is an area of economic management that must be assigned to the center. Second, invest- ment capital is likely to be scarce, and this argues for its central allocation, at least for the provision of new public infrastructure. Third, a limited total tax capacity may argue against giving subnational govemments too much tax discretion in case they compete too effectively with the center for access to the major tax bases. Finally, in those cases-the majority-where the inherited regional and other income disparities of transitional economies need to be changed, a strong central government coordination of the equity aspects of fiscal management is likely to be important. Russia and China, Bahl noted, face difficult choices as regards their spatial equalization policies. Russia must make the hard choice between fiscal equalization, central government fiscal solvency, and the appeasement of the potential break-away provinces. China is forced to choose between funneling more resources to the lower income provinces or leaving the retained revenues at a higher level in the faster growing coastal provinces. However, both countries still retain a high degree of fiscal centralism and so can still make such decisions. By contrast, the United States, which is much more decentralized fiscally, has only limited power to affect a regional redistribution of resources. An Overview of Main Differences A detailed comparison and analysis of the main differences between the Chinese, Rus- sian, and U.S. fiscal arrangements are set out in Chapter 4. In Table 1-1, the most impor- tant differences, are presented. 5 5. For an excellent summary of cross-country tax and expenditure assignrments see "The reform of Intergovernmental Fiscal Relation in Developing and Emerging Market Economies," Anwar Shah, World Bank 1994 (Appendix C, D, and E). 14 Macroeconomic Management and Fiscal Decentralization Table 1-1. Chinese, U.S., and Russian Fiscal Systems Compared Characteristic of systern China United States Russia Degree of local autonomy Limited (no scope to set High; states have wide Limited discretion on tax tax rates or to borrow) powers but cannot affect rates and bases interstate commerce Local government activity 50 - 60 percent 40 percent 47 percent as share of total govern- ment activity Expenditure assignments Not at all clear More or less clear who Not at all clear to local governments should do what Revenue sharing Tax sharing (derivation Mixed approach depend- 85% tax sharing using basis); equalizing grants ent on political process. derivation basis; 15% ad largely ad hoc About 60% formula hoc determined; 40% ad hoc Degree of transparency More after 1994 reforms Ad hoc Limited Expenditure per capita 6 to I Connecticut (best = 2.5 11 to I (variation across prov- times Arkansas (worse) inces: ratio of "best" to "worse") Does central government Tax sharing system is No real relationship No - policies accentuate equalize? equalizing (richer between income levels inequalities provinces pay more); and center to local earmarked grants sys- transfers tem is not (one part of system neutralizes the other) System of tax administra- n.a. Federal and states totally n.a. tion independent Recent trends Moving to centralize Becoming more central- Moving to decentralize ized n.a. Not applicable. The fifty states in the U.S. system have a very high degree of fiscal autonomy. This de- centralization is in contrast especially to Russia but also to China. In the United States, the states are quite free to choose their own tax structures as long as these do not violate the federal constitution. Hence, tax rates, tax bases, and user charges may be set by the states without federal approval. In general terms, expenditure autonomy is also high, and states can determine which expenditures to provide. This general arrangement is overridden, however, by federal government mandates, some of which restrict expenditures on, for example, health and environmental grounds and others of which require certain expenditures by the state governments to match the provision of federal grants. Likewise, state governments have mandated certain actions by the lower levels of local government. The fiscal autonomy of local government in general is more restricted than that of the state governments, although this also varies from state to state. The U.S. trend for the past fifteen years has been toward reducing federal transfers to both state and local governments. "General revenue sharing," formerly the only program of general assistance to state and local governments, has been eliminated. The 1986 fed- eral tax reform eliminated the deductibility provision for state and local government general sales tax against the liability of individuals for federal income tax, thereby ending the shifting of a substantial part of the burden of such taxes on to the federal government. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 15 The net effect is that the price a state resident must now pay for each dollar's worth of public expenditures is higher than before. This has resulted both in greater accountability of state government officials and in some increased hesitancy about fiscal decisions by state and local governments. Although there have always been large interstate variations in per capita fiscal expen- ditures, the federal grant system is not designed systematically to equalize these dispari- ties. This is in sharp contrast to the arrangements in, for example, Australia and Germany (see discussion later in this chapter). Additionally, the state-by-state competition for in- dustry and scarce mobile resources, which Ved Gandhi had earlier suggested might be wasteful, is not restricted in any significant way by the U.S. arrangements. The present Russian system summarized in Table 1-1 offers a very significant contrast to the U.S. system. It operates as a three-tier system: the federal government, ninety-one provinces/states such as oblasts, and a lower level of government known as rayons. The revenues of the subnational units are derived mainly from four shared taxes, but pre- dominantly from the enterprise income tax and VAT, with the rates and bases set by the federal government. The sharing is on a derivation basis (that is, point of collection), with the sharing rates set by the Parliament. These can vary from oblast to oblast, but they are set in an ad hoc fashion and have been subject to frequent amendment in the past few years in response to changing political pressures. The assignment of public expenditures in Russia is made on the basis of the geo- graphical dimension of benefits-an arrangement deriving from the system operated formerly in the Soviet Union. In brief, public service activities whose "benefit area" is the whole nation (for example, defense, international trade expenditures/subsidies) are as- signed to the federal government. Those services with an identifiable regional benefit area (such as universities and certain specialized hospitals) are provided at the oblast level or a similar level. Elementary schools, parks, and other services with a narrower benefit area are provided by the lower levels of government. Among other things, this means that most forms of social welfare are provided at the subnational level. Bahl noted that while this conceptual assignment of expenditures fits with good prac- tice and accepted norms as explained by Gandhi, the actual practice in Russia often fails to do so. Because of a serious lack of clarity in the assignment of specific services to spe- cific levels of government, there is often some duplication of provision or provision of services at the "wrong" level. This problem has been compounded in the recent transi- tional period by ambiguity about the assignment of responsibility for the ownership of certain commercial and industrial enterprises as well as ambiguity about the expend i- tures associated with that ownership. Recent trends in the fiscal arrangements have been driven by the dictates of the tran- sition to a market economy and have been colored significantly by the need to restore macroeconomic stability. The latter has constrained the government in its ability to run fiscal deficits, but at the same time there has been a need to hold the Russian federation together by new revenue sharing devices. The consequence has been frequent changes in the details of both revenue sharing and expenditure assignment arrangements. Some of the more recent changes have violated accepted principles in very funda- mental ways. For example, early in 1992 the central government shifted the responsibility downward for most of its price subsidy and income maintenance programs, but it failed to provide the necessary transfers-approximately 5 percent of GDP-to fund these pro- grams. No plans have been developed about how these extra costs will be matched by 16 Macroeconomic Management and Fiscal Decentralization available revenues at the oblast level. Somewhat later in the same year, the responsibility for many investment expenditures, including several with obvious national significance, were also shifted down to the oblast level. This change was motivated apparently by the need to balance central government budgets but also to claw back some of the gains that the oblasts had derived short term by the introduction of fixed sharing rates. Whatever the rights and wrongs of these particular changes, the system, in recent years, has evolved on the basis of a bargaining approach rather than as the consequence of the concrete implementation of clear principles. Bahl argued that such arrangements cannot continue for much longer. If the central government continues to eschew its ex- penditure responsibilities in the interests of a balanced budget, it may be contributing to the realization of its own worst fears-namely, the disintegration of the Russian federa- tion. On the subject of equalization, Bahl noted that the Russian approach to tax assign- ment on a derivation basis necessarily means that high-income territories will derive more revenue than low-income territories. This, in turn, means that some forms of sub- vention must be provided to protect the position of territories that would otherwise be unable to provide adequate public services. In practice, the federal government in Russia uses three types of discretionary actions to influence the fiscal resource positions of the lower level authorities. Specifically, it approves the budget of each oblast, it determines the subventions that each will receive, and it sets the VAT retention rates. It would be reasonable to suppose, he argued, that in such a system the oblasts with the greatest expenditure needs and the lowest fiscal capacities would enjoy the highest retention rates. In reality, multiple regression analysis indicates little evidence of such systematic relationships. Those systematic relationships that can be found suggest a higher retention rate in cases where the rate of urbanization is lower and where there is a smaller concentration of elderly people. However, the most significant result is that oblasts having lower per capita incomes do not benefit disproportionately from fiscal redistribution. The system, in other words, cannot be regarded as an equalizing one. Further Aspects of the Russian System Roy Bahl used his comparison of the U.S., Russian, and Chinese systems to identify a number of main issues that need resolution in the ongoing Chinese reforms. However, before we consider these issues it is useful to summarize the further analysis of the Rus- sian system provided in a separate presentation by Christine Wallich. She emphasized that a successful resolution of the Russian problem of inter- governmental fiscal relations lies at the very heart of that country's attempts to make the transition to a market system with well-functioning democratic arrangements. Failure to resolve the fiscal problems would affect not only Russia's efforts to achieve the privatiza- tion of the majority of its enterprises but also its prospects for macroeconomic stabiliza- tion, for regional coherence, and for the viability of the federal political structure. At the core of the problem, she argued, was the serious lack of correspondence be- tween the expenditure responsibilities of the different levels of government and their ac- cess to revenues. The intergovernmental fiscal system that is now evolving gives subna- tional governments new expenditure responsibilities as well as new budgetary rights and financial resources. However, many of the laws passed since 1991 to enact these changes have not been implemented, and much of what is currently happening is authorized if at Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 17 all by transitional rules and decrees. On the spending side, the subnational governments still have less than full budgetary discretion, and on the revenue side, the old tax sharing system still has not been phased out. The Ministry of Finance effectively determines the resources to be transferred to individual oblasts, and arrangements in this area generally remain ad hoc and contentious. As regards expenditure assignments, Wallich reiterated Bahl's earlier remarks about the serious absence of clear-cut rules in Russia. She noted that the dictates of macroeconomic stabilization had resulted in an implicit strategy of "pushing the deficit downward." The unarticulated hope was that the subnational governments would some- how succeed in cutting expenditures even though the national government had failed to do so. This had applied in particular to social safety net expenditures since early 1992 and from mid-1992 to most categories of capital spending. These reallocations, she noted, are against most principles of sound expenditure assignment and compromise the pros- pects of efficient service delivery by shifting matters of national priority to the local level and, at the same time, failing to accompany the new mandates with adequate funding. Of course, the realities of the transition period have increased, not reduced, public ex- penditure needs at the local levels and especially for social safety net purposes. Caught with insufficient revenues to cover their new expenditure responsibilities, many oblasts have accumulated significant payment arrears and have also delayed the remittances of tax revenues to the federal level. Extra-budgetary funds (fines, funds from auctions, some loans) that the local authorities do not need to share with the federal government have grown rapidly in relative importance in response to the increased local-level expenditure responsibilities. In a situation where the subnational governments account for almost half of total budgetary outlays, these various ad hoc outcomes are clearly inimical to Russia's efforts to improve budgetary management and macroeconomic stability. Wallich stressed that these developments also impede privatization objectives. Be- cause the oblast governments derive significant extra-budgetary funds from the enter- prises that they own and because they also benefit from the social and other expenditures that local enterprises presently finance, it is in their interest to oppose privatization, espe- cially of the stronger enterprises. The Russian system differs from most Western systems in that revenue is "shared upward" from the rayons and oblasts that collect it to the federal budget. Wallich noted that this system was clearly one factor contributing to the break up of the former Soviet Union in 1990 when Ukraine, Russia, and then the other Republics ceased to make their revenue payments to the Union budget. Likewise, the persistence of the same system in the Russian federation leaves that system itself exposed to breakaway pressures. The Basics Principles law of 1991, which sought to reform this system, assigns revenue sources to one or another level of government. In practice, as Bahl explained, most taxes remain shared, and most subnational revenues are derived from four shared national taxes with all the sharing being done on a derivation basis. All tax administration and collection remain federal responsibilities but are executed locally by local tax offices. With the setting of all tax rates and bases the responsibility of the federal level, subna- tional governments have no fiscal discretion except in relation to the extra-budgetary charges referred to earlier. Nor is there any system to ensure a correspondence between the current tax assignments/shares and the new expenditure responsibilities. In some oblasts, the sharing of personal and corporate tax revenues result in "over-financing," 18 Macroeconomic Management and Fiscal Decentralization which there is no legal basis to claw back. Likewise, in oblasts that are underfunded, there is no legal right to subventions from the center to make up the shortfalls. One possible but partial solution to this problem that has been discussed would in- volve moving from a tax sharing system to a tax assignment system in which each local administration would receive the revenue from its own assigned taxes. However, Chris- tine Wallich had serious doubts about whether the assignment approach could really en- hance the autonomy of the subnational governments while they lacked the discretion to set tax rates and establish tax bases. So long as this power was left with the federal authorities, local authorities would remain vulnerable to the revenue volatility of their assigned taxes. At the same time, revenue assignments would be likely to encourage even more domestic protection and interoblast trade barriers in order to protect local monopolies and the revenues that they were capable of generating. Wallich drew attention once again to the serious inadequacies of the Russian system for fiscal equalization. The new Russian legislation makes no provision for this. As in the system inherited from the former Soviet Union, transfers between the different levels of government still play only a limited role. Until 1994, most subventions that were paid were provided on an ad hoc and ex post basis to oblasts arguing some special needs. From 1994 onward, this bargaining system is to be based on a more transparent ar- rangement, with subventions going to all oblasts whose per capita revenue is below the national average. Some 22 percent of federal VAT revenue is being set aside for this pur- pose, and it is reckoned that forty-three needy oblasts will receive transfers as a conse- quence. This arrangement will be supplemented in the case of the most seriously needy oblasts. The difficulties of the present system of intergovernmental fiscal relations in Russia are further complicated by the demands of some territories for special fiscal treatment. In some cases, greater autonomy is claimed because of the presence of non-Russian ethnic groups such as the Tatars. Others cite the presence of valuable natural resources; the local governments feel that the benefits from such resources should accrue disproportionately to them. Finally, some strong industrial areas are looking for increased fiscal autonomy to ensure the fullest possible benefits from this strength. Wallich noted that in the ab- sence of more transparent fiscal arrangements, the better-off oblasts may feel that they are cross-subsidizing the others, even though this may not be the reality of the situation. Issuesfor China The presentations by Roy Bahl and Christine Wallich and the discussion that accompa- nied them suggested issues requiring further consideration for the Chinese case-recent refoms notwithstanding. (i) What is the overall goal of the Chinese government in establishing its system of in- tergovernmental fiscal relations? In particular, what are the main motivations for, and the weights to be given to, decentralization versus centralization in that system? (ii) Can the expenditure responsibilities of the different levels of government be clearly assigned? In China as well as in Russia, tradition and inertia have characterized past efforts in this area, and while this continues, it is very difficult to design a sound overall system of intergovernmental fiscal relations. Both countries have put the cart Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 19 before the horse by concentrating initially on tax and transfer systems rather than on expenditure assignments. (iii) Will or can the Chinese national government grant subnational authorities some freedom to set the tax rates and bases for certain large taxes as, for example, in the United States? For the moment, both the Russian and the Chinese systems are highly "unified," meaning that subnational authorities have been assigned only minor nui- sance taxes that do not generate much revenue. (iv) Which taxes, if any, will be assigned to the subnational authorities, and which ones will be shared? The present Chinese system of regarding the VAT as the main shared tax may not be appropriate. The VAT might be more usefully treated as a central government tax. Other candidates for tax sharing or assignment were listed by Ved Gandhi. (v) How will horizontal and vertical imbalances be addressed, and will transfers from and to the center be based on a formula approach? The Russian situation indicates the dangers of not making the arrangements transparent. For the moment, neither China nor Russia seems to have faced up to equalization as an issue, in spite of the very substantial inequalities that prevail. Possibly this is the correct approach in a transition era where particular poles or regions of growth are crucial to overall strong performance and so to investment growth and fiscal resource mobilization. However, China has made some important recent moves away from the contracting system and in the direction of a formula-based transfer system. (vi) What approach if any will the central government take to problems in the relation- ships between provincial and lower level local governments? In cases where there is clearly a significant degree of nonequalization of fiscal positions within a particular province, will there be any role for the central government? In Russia, the federal government seems to want to retain power even for allocating resources to the lowest levels of government. This implies limited trust in the oblasts in this regard. The problem is particularly important because in China, as in Russia, the inequalities within certain provinces are far more substantial than those between provinces. (vii) How will tax administration be streamlined to make it generally more efficient but also to ensure that it is adequately structured to deal with whatever new system of intergovernmental arrangements emerges? The initial steps already taken in China to establish a unified National Tax Service are an important potential improvement on the former "bottom-up" system. As Russian experience indicates, that system renders central government revenues extremely vulnerable. In particular, what role will the subnational tax administrations play in the event that there are few if any purely subnational taxes in the new system? (viii) How will China modernize its personal income tax arrangements and ensure com- patibility of this tax with the system for taxing enterprises? &x) What arrangements will be made to deal with the thorny problem of natural resource taxes? For the moment, both Russia and China seem to be operating on the basis that the territories where the resources are located should benefit disproportionately from the associated fiscal revenues. However, this approach can be very divisive, and there is now an (academic) consensus that the bulk of such revenues should accrue initially to the center. 20 Macroeconomic Management and Fiscal Decentralization Lessons from Australia and Canada Australia Berndt Spahn noted that in China there is considerable interest in the Australian fiscal system, especially because of its scientific basis for fiscal equalization. However, Austra- lia is a proper federation of states (as is Canada) rather than a unitary state like China. It was founded on the principle that the central government should be of minimal size with most of the main government functions residing with the state administrations as in the United States. However, the expansion of social and welfare functions since the estab- lishment of these federations has changed this original conception quite considerably. In the nineteenth century Australia had two main revenue sources, customs duties and excises, and these were collected centrally. About three-quarters of the revenues were paid to the state administrations as tax reimbursement grants. This system changed over time. Inflation eroded the value of the grants, set initially in nominal terms. How- ever, states showed a reluctance to levy income and other unpopular taxes, and so large vertical imbalances emerged. In the 1930s the establishment of the Commonwealth Grants Commission was the solution to this problem. Spahn explained that today the Australian federation (or "Commonwealth") is the most centralized of the major federal countries of the Western world as far as revenue collection is concerned. Hence, it is an important counterexample among relatively de- veloped economies to the United States (considered in the previous section) and to Can- ada (considered below). Some four-fifths of total tax collections in Australia are ac- counted for by the federal government. The states collect the bulk of the remainder, with the lower levels of local government having an extremely tiny share. In spite of this, the states still carry the responsibility for most major expenditure functions (except defense, foreign affairs, and interstate affairs). Thus, the system results in substantial vertical im- balances (own-resources minus expenditure needs) at the various subnational levels of government and in a high degree of dependence by the states on the federal government. Approximately 50 percent of state revenue came from this source in 1992-93 with the vertical imbalances being filled predominantly through general revenue sharing- unconditional bloc grants-as well as through specific or categorical grants. The general revenue grants account for about 45 percent of the revenue going to the states and pro- vide for the financing of basic services. Specific purpose grants (for example, to hospitals) amount to about 35 percent of the total. Additionally, there are some grants that are paid through to the municipalities. More specifically, as regards tax assignment, the federal government in Australia co 1- lects all the major taxes including taxes on income and capital gains (some 70 percent of all tax revenues in 1988), sales tax (16 percent), and customs duties (5 percent). The state and municipal governments collect payroll taxes, stamp duties, motor vehicle taxes, gambling taxes, and other taxes. Collectively-the payroll tax excepted-these taxes con- stitute a small part of total revenue. In the past, opposition political parties have pro- posed a broader based consumption tax, but since it has not gained acceptance, an ex- ceptionally high reliance on the income tax is the result. A more likely reform will be the reintroduction of tax sharing, an arrangement that was in operation for a time from 1976 to 1985. It was then abandoned for reasons connected with the vulnerability of the states' Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 21 revenues to unilateral federal tax decisions: an outcome that echoes Christine Wallich's earlier reservations about such a system in Russia. In the absence of reform along these or other lines, grant money remains the most im- portant single source of revenues to the subnational governments. Many regard the Australian grants system as one of the more significant of the equalizing mechanisms now in use worldwide. The system has certainly been widely advocated as a possible basis for a new Chinese approach to the problem of interprovincial inequalities in fiscal resources. In broad terms, Spahn noted that the distribution of total resources among the Aus- tralian states depends on four elements: the size of each state's own tax base; the distri- bution of general revenue grants according to the recommendations of the Common- wealth Grants Commission; the distribution of specific purpose grants; and the distribu- tion of states' loan moneys, which are also centrally controlled. Within this list, the pro- portion of the total attributable to specific purpose grants-many of them requiring local matching revenues-has increased dramatically during the past twenty years. This trend is much resented by states as a source of Commonwealth intrusion into their sover- eignty-undermining their freedom to determine the size of the services for which the states nominally carry the responsibility. However, since the availability of the specific grants broadens the overall funds available to the states, this criticism of them, Spahn argued, is only partly valid. In any case, the general revenue grant constitutes the main instrument for achieving horizontal fiscal balance. The Commonwealth Grants Commission that advises on these grants has been oper- ating-albeit with some variations in approach-since 1933. In the early days, the needs of claimant states were assessed relative to two designated "standard" states-namely, New South Wales and Victoria. Grants tended to be quite small because the claimant states were those with generally quite small populations. By contrast, the present distribution model in operation since 1981 embraces all states and territories. It calculates each state's share of the total general revenue grants by as- sessing its relative per capita revenue-raising capacities, standardized against its own- resource revenue sources; its receipts of most specific grants; its relative per capita ex- penditure needs as assessed against the costs of providing standard services-costs that vary considerably depending on the population sizes and densities of different states. In effect, the calculation is of the "standardized" deficit of each state defined relative to population and per capita grant relativity. It is this that the system seeks to equalize. The specifics of this system were described in rather more detail by Anwar Shah. He explained that the goal was to provide reasonably even levels of public services as well as taxation across the federation. The total pool of revenues available for distribution is equal to 40 percent of the income tax collections of the previous year. This is an arbitrary figure chosen independently from the formula that governs distribution. The formula itself takes account of the factors already listed in the previous paragraph. Revenue needs are determined as in Canada by reference to a standard national average rate of tax ap- plied to the tax base of each state. Unlike the Canadian system that ignores expenditure needs, the Australian system recognizes explicitly the possible differences of such needs across states. Shah explained that this is done very rigorously by reference to all major expenditure categories. The general formula is linked to population and the per capita standard expenditure but adjusted by reference to, for example, scale factors, dispersion and urbanization factors, and cultural and environmental factors. To the extent possible, 22 Macroeconomic Management and Fiscal Decentralization the adjustnent parameters are determined objectively using regression and other techni- cal methodologies. In many cases, subjective assessments also need to be made. The cal- culations involve some thirty revenue categories, more than sixty expenditure categories, and more than twenty grant categories. 6 Shah's own assessment is that this system, impressive as it is, nonetheless has serious flaws and imprecisions. Many of the judgments about allocation remain purely subjec- tive. Moreover, the data needs for implementating the system are very demanding, and much of the relevant data are seriously outdated by the time they are used. Finally, the main decision to endorse allocations remains a matter controlled by Parliament and so by a political process. Spahn's assessment of the system is that it provides a secure revenue base and so sup- ports a stable environment for the planning of state-level policy decisions. The absence of a more conventional link between tax assignments and expenditure assignments has not necessarily meant that the Australian states have lost out in terms of total revenue or autonomy. Although specific purpose grants have risen in importance relative to the general purpose ones, the "fungibility" of the specific grants has meant that federal power over the states has not necessarily increased as a consequence. However, the matching requirements may have introduced some biases into the structure of state government outlays (for example, too much university-level education and not enough primary). This criticism, in Spahn's view, argues for some improvements in the present arrangements and not for their fundamental reform. Above all, the analysis of the Australian system shows that it has generally preserved state independence. The lower tiers of government are far from being at the mercy of the federal government in spite of their limited powers to levy taxes and the large vertical imbalances. Canada Anwar Shah noted that Canada is a country manifesting considerably less centralization than Australia. It has a very large land area but a relatively small population. Its prov- inces are characterized by large differentials in both their per capita incomes and their resource bases: agricultural, mineral, and industrial. Canada operates a two-tiered and decentralized fiscal system that contrasts sharply with that operated in its neighboring country, namely the United States. In particular, the Canadian provinces, the second tier of the system, enjoy a far stronger de facto influence on revenues and expenditures than do their U.S. counterparts, and they also have a strong influence over decisionmaking at the federal level. Similarly, the local administrations in Canada enjoy a fairly strong in- fluence on provincial decisionmaking. Another important difference is that while inter- state fiscal equalization arrangements are implicit and piecemeal in the United States, they are constitutionally mandated and of considerable importance in Canada. In Canada, unlike in Russia and China, expenditure responsibilities are clearly de- fined. Broadly speaking, expenditures subdivide into those that are exclusively the re- sponsibility of the federal government; those that are exclusively the responsibility of provincial and local governments; and those that are discharged as shared responsibili- 6. A more formal explanation of the system is provided in Shah, The Reform of Intergovernmental Relafions, Annex B. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 23 ties. These assignments tend to conform quite closely with the general principles dis- cussed earlier but with some important exceptions. For example, the responsibility for immigration policy is shared between the federal and the provincial governments, and unemployment assistance, although a formal responsibility of the provincial govern- ments, is largely implemented at the federal level. As these examples illustrate, there is generally more coordination and cooperation between the different tiers of the govern- ment than, for example, in the United States. A more complete summary of the structure of responsibility sharing is set out in Table 1-2. Note should be taken of the relatively high share (almost 60 percent) of the sub- national levels of government in total expenditures. Taxation assignments are equally transparent, although they have been subject to sig- nificant changes over time. In the British-North America Act of 1867 that established the initial system, the federal government was authorized to levy all taxes, while the prov- inces obtained the authority only to level "direct" taxes. However, periodic court chal- lenges to the interpretation of the term "direct" taxes have progressively extended the scope of the taxing power of the provinces-a development that was confirmed by the Constitution Act of 1982. The present system is one involving considerable overlapping of taxation responsi- bilities between federal and provincial administrations but with a great deal of consent, cooperation, and harmonization concerning bases and rates of most taxes. A summary of the arrangements as regards the corporate and personal income taxes is shown in Table 1-3. This essentially harmonized system provides scope for different rates of tax as well as for different choices about the location of the responsibility for administration and collection. For the most part, the federal administration accepts the main responsibility for fixing the bases of these two taxes. Table 1-2. Expenditure Responsibilities in the Canadian System Percentage of total expenditures Activity Provincial and local expenditures Federal expenditures Defense, police, etc. 38 62 Transport 73 27 Health 97 3 Social services 35 65 Education 97 3 Debt service 42 58 Other activities 57 43 All activities 59 41 24 Macroeconomic Management and Fiscal Decentralization Table 1-3. Federal and Provincial Responsibilities for Income Taxation in Canada Level of government Tax type (number Administration & of provinces) Base determined Rate determined collection Corporate (7) Federal Federal (27%) + Pro- Federal vincial (9-17%) Corporate (3) Provincial Provincial Provincial Personal (9) Federal Federal (17-29%) + Federal Provincial (45-69%) Personal (1) Provincial Provincial Provincial Table 1-3 also illustrates the scope that is possible in the Canadian system for the pro- vincial administrations to "piggy-back" their own local taxes on to otherwise federal sys- tems of tax. Hence, in seven provinces there is a local corporate income tax at the rate of between 9 and 17 percent in addition to the federal tax levied at 27 percent. Full harmony, however, is certainly not achieved. When the value-added tax was in- troduced, the federal government asked the provinces to accept it and apply a harmo- nized tax base and joint administration. However, only Quebec accepted this proposal and now collects a federal VAT at the rate of 7 percent in addition to its own provincial VAT at the rate of 8 percent. All but one of the other provinces (Alberta) still retain their own sales taxes. The transfer system in Canada's form of fiscal federalism is mandated by the country's Constitution. In essence, this system provides equalizing transfers to ensure that all pro- vincial governments are able to provide comparable levels of public services at compara- ble levels of taxation. The system provides a simple alternative to the Australian model described earlier, and it co-exists happily with the Canadian arrangements in which some provinces take the major responsibility for total revenue generation (Table 1-3). The system comprises three transfer programs. Each has different rules but contrib- utes jointly to the objective just defined. The first transfer program governs federal to provincial transfers for major programs such as education and health. These transfers are paid on the basis of equal per capita transfers to each province but are conditional on the provinces adhering to established standards of access and service. To qualify fully for this type of transfer as exampled by the health program, the provinces must meet certain basic conditions: comprehensive coverage (all medical services must be offered); univer- sality (the whole of the provincial population must be covered); portability of access from one region to another; public insurance coverage (although the service delivery can be either public or private); and free access (no user charges). Some dilution in relation to these five conditions is allowed, but this will normally result in some reduction of the transfer below that otherwise provided. For example, if provinces decide to impose user charges for some classes of medical provision, they will be subjected to a dollar-for-dollar reduction of transfer for all income so generated. The second transfer program is the Canada Assistance Plan designed to help finance social assistance for needy groups and individuals. It is provided to provinces on a 50:50 Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 25 matching basis, implying a large provincial commitment to all the expenditure activities that it finances. Although designed rather differently from the first transfer program, it results in broadly similar per capita transfers to each province. The third program is the Fiscal Equalization Program, which has been described as "the glue that holds the federation together." Unlike the other two transfer programs, it results in transfers that are highly uneven across provinces. Some provinces derive some 20 percent of their total revenues from this source, and others derive nothing. Shah ex- plained that the equalizing transfers are calculated separately for each revenue source. They are calibrated against the average tax rate (for each revenue source) across all provinces and by reference to a standard per capita tax base as calculated on a five province "standard" (British Columbia, Saskatchewan, Ontario, Manitoba, and Quebec). The per capita equalizing transfer (E) received by province "i" for each revenue source is given by the following formula: Ei = [(R10/Bl0 x B5/P5) - (R1,,/B10 x Bi/Pi)]/P; where R = total revenues; B = the tax base; P = population, and the subscripts have the meaning 10 = ten province total; 5 = total for the five "standard" provinces. The equalizing transfers associated with each revenue source are added together. If the resulting total is positive, the province in question receives a payment from the fed- eral government of that amount. Negative sums are ignored. The system makes the im- plicit assumption that the per capita expenditure needs are the same across all the prov- inces. Because the formula uses an average tax rate (R10/B10), it adjusts revenues mainly for different tax bases as between provinces rather than for any decision to set differential provincial tax rates. However, as Shah emphasized, the system determines both the pool of funds to be transferred and the allocations themselves in an objective manner (consider the case of Australia). In 1989-90 all but two of the provinces received some positive transfers under this scheme, but the variations in the amounts received were ex- tremely large, both absolutely and relative to total revenues. 7 Lessons from India, Switzerland, and Germany India Amaresh Bagchi explained that the foundations of India's present system of fiscal feder- alism were laid down in the pre-Independence Government of India Act of 1935. This Act assigned the revenues from most of the high-yielding tax sources-personal income tax, customs, corporation tax, and so on-to the Center at the same time it assigned to the provinces (later states) responsibilities for delivering a wide range of public services. The considerable vertical imbalances that arose from these assignments were dealt with by sharing of certain central taxes combined with a system of grants-in-aid. This division of functions and responsibilities was continued in many respects after independence, when the 1950 Constitution listed the respective responsibilities of the center-the "Union"-and the States. The Union list of responsibilities includes expendi- tures such as defense, foreign affairs, national highways, railways, posts and telecommu- 7. Ibid., Figure A.1. 26 Macroeconomic Management and Fiscal Decentralization nications, shipping and navigation, as well as responsibility for the main economic func- tions such as the currency, foreign exchange, foreign borrowing, external trade, interstate commerce, banking, and finance more generally. Although the tax means provided to the states continued to fall well short of their overall expenditure needs, the Constitution did assign the states the rights to revenues from taxes on the sale of goods, agricultural taxes, land taxes, vehicle taxes, stamp duties, and taxes on trades and professions. Major taxes remained the prerogative of the center, and the Constitution also stipulated that borrowing by any state would be subject to the approval of the center so long as the state had any debt outstanding to the central gov- ernment. The prospective expenditure: revenue gaps at the state level associated with this sys- tem were filled by arrangements for the compulsory sharing of the personal income tax and also, with the approval of Parliament, Union excise duties. Additionally, some taxes to be collected by the center were to be assigned to the states. This applied, for example, to estate duties and to some succession duties. Finally, a grant system was confirmed as an important part of the system of center to state supports. In order to ensure that the states could rely on objectivity and fairness in the applica- tion of these various principles, a Finance Commission-to be re-appointed every five years-was appointed to oversee the implementation of the arrangements for sharing taxes and transferring grants-in-aid. Bagchi also noted that once planning began in ear- nest in India and a Planning Commission was established, large "Plan Grants" emerged as an important but problematic feature of the fiscal arrangements. The retention of broadly similar arrangements for more than forty years is evidence that the devolution of federal funds based on the Indian Constitution and the work of the Finance Commission has operated quite well. However, that same experience has indi- cated several weaknesses that provide important lessons to other countries including China. The first problem to which Bagchi drew attention, a problem also evident in the Chi- nese experience, relates to the disparate trends between the levels of state and federal government expenditures. Between 1960-61 and 1993-94, the taxes and other revenues raised by, and otherwise accruing to, the states increased from 59 percent of the national total to 64 percent. But this did not prevent the states from collectively moving from a situation of budget surplus to one of significant deficit. This was partly the result of larger devolutions of expenditures recommended by the Finance Commission since the mid-1970s. The deficit has raised serious questions as to whether the assignment of ex- penditures and revenues is out of balance or whether the system of transfers itself is flawed. The "gap-filling" approach followed by the Finance Commission until quite recently provided little incentive for economy or efficiency in expenditure management or im- proved revenue effort at the state level.8 In addition, large devolutions out of particular taxes are believed to have acted as a disincentive for the center to raise more revenue and as an encouragement to turn instead to nonshared categories of taxation to meet the cen- 8. The total of tax and nontax revenues collected by the states on their own behalf fell from 61 percent of their total revenues in 1960-61 to 57 percent by 1993-94. During that same period, the taxes devolved to them from the center rose from 16 percent of the total to almost 23 percent. Grant receipts fluctuated through time but accounted for a slightly larger share of the total in 1993-94 than they had in 1974-75. Within this, Plan Grants accounted for a larger share. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 27 ter's own requirements (for example, surcharges on income and corporation taxes). A further and related problem is that a rising share of Plan Grants (controlled by the Plan- ning Commission) has encouraged the states to expand capital expenditures beyond the levels that can be supported adequately by recurrent expenditures. This has resulted in the neglect of the maintenance of existing assets. The second main problem referred to by Amaresh Bagchi is that horizontal equaliza- tion has been only partially achieved through the transfer system. Specifically, the richest states in India achieve own-revenue per capita that is up to 5.2 times higher than that of the poorest states and a revenue:expenditure ratio that is 2.3 times higher.9 Despite at- tempts at equalization, the richer states have much larger surpluses in their revenue budgets, they have been able to deploy more resources under successive Plans, and the gaps have been widening. Bagchi suggested two main reasons for this. First, tax devolu- tion constitutes the dominant component of the transfers recommended by the Finance Commission and overshadows the grants in this regard. Second, the equalizing charac- teristics of the devolved taxes are much weaker than those of grants. Even states with large surpluses in their budgets before devolution can receive substantial shares of the transfers and so mount even larger expenditures than would otherwise be possible. Although the Finance Commission has endeavored to tilt the transfers more to the poorer states by including an equity criteria into the formula for allocating shared taxes, the different objectives of tax sharing and grants have become blurred. The multiple and overlapping criteria now used make it difficult to orient the transfers to any clear central objective. The third problem commented on by Bagchi concerns the overlapping tax jurisdic- tions of the center and the states. This is a particularly serious problem in relation to do- mestic trade taxes such as the Union excise and the sales tax. The problems are also acute in terms of the overlapping and cumulation involving the Central Sales Tax (CST). Be- cause the CST is levied on the basis of the origin of the goods, it has enabled the rela- tively advanced states in terms of industry to export the burden of their own taxes to the consuming states, which are often less developed. The result is a considerable concentra- tion of CST revenues: 40 percent of total collections go to just four states, which account for only 20 percent of India's total population. The system is a considerable incentive to "competitive federalism" and also to distortions such as the camouflaging of interstate sales as transfers on consignment to branches or depots of the same company. Bagchi noted that efforts to better harmonize the main domestic trade taxes to avoid these and related problems have been relatively unsuccessful. In particular, the substitu- tion of the sales tax on three major commodities (textiles, tobacco, and sugar) by a cen- trally levied additional excise duty has come under pressure with the states attempting to circumvent it in various ways. However, the main issue still to be resolved is that the ex- cessive complexity and competition associated with the separation of powers to levy domestic trade taxes have been inimical to the development of a more effective common market in India. Since such a market is surely one of the main justifications for a federal form of government in a large country with distinctive regional characteristics, this is a very serious comment on present arrangements. 9. This compares with a per capita income differential of 3.1 between the richest and the poorest states. 28 Macroeconomic Management and Fiscal Decentralization Bagchi concluded by noting a number of lessons that can be drawn from the Indian experience, some of which are also pertinent to the discussions now taking place in China. First, some vertical gaps at the subnational level are likely to be unavoidable since taxes with the broadest bases are collected most efficiently at the national level. To keep these gaps to a minimum, expenditure responsibilities should be defined in such a way that only those of clear national concern are dealt with at the center. Subnational gov- emments should be assigned as many taxes as is consistent with the ease of administra- tion of those taxes. However, this assignment should leave the center in a strong enough revenue position to macro-manage the economy effectively and also to deal with spil- lovers, externalities, and a degree of equalization in the provision of public services across different regions. Second, with the main revenue taxes being assigned to the center, the transfer of funds from the center to the subnational governments is unavoidable. However, the de- sign of the system of transfers should endeavor (a) to link the expenditure and revenue- raising decisions at least at the margin; (b) to avoid any adverse incentive effects for revenue raising and excessive expenditures at subnational levels; and (c) to respect as far as possible the preferences of local populations regarding the public services to be pro- vided to them. Among other things, this second point argues for some conditionality in the transfers linked, for example, to tax effort, as well as for some monitoring by the cen- ter of the purposes for which transfers are used. The third lesson is that the correction of horizontal imbalances should not rely only on specified selected taxes. This reduces the incentive of the center to exploit the tax capacity of the shared taxes to the full. More certainty and less distortion can be generated by a system that specifies a share of central taxes as a whole to belong to the divisible pool to be transferred. Finally, the experiences of India show that even a long-lived system of fiscal relations is subject to a complex web of ever-changing pressures. Although the broad structure of the system might remain unchanged, the details of its implementation need to be kept under constant and objective review. The Finance Commission in India is supposed to play this role, but its efficacy is undermined by the temporary character of the appointees who serve on it as well as by its lack of research-underpinning on a sustained basis. Bag- chi was strongly of the view that a permanent body could more effectively discharge the tasks assigned to the Finance Commission. Further Examples from Europe In the final presentation relating to the experiences of other countries, Berndt Spahn compared the arrangements in the two neighboring European countries of Germany and Switzerland. He noted that the two countries have adopted radically different ap- proaches to fiscal relations. Switzerland provides its lower levels of government with high degrees of independence, while the German system emphasizes centralization, standardization, and equalization. Switzerland Spahn explained that in this country, which is the oldest federation in the world, having been established in the thirteenth century, there is an explicit federal system. It treats the Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 29 twenty-three cantons as sovereign states for most purposes. The lower levels of govern- ment enjoy a high level of independence and in that sense have much in common with the provinces of China. "Direct" democracy also applies in that the electorates are re- quired to vote on changes to the Constitution and on all other major pieces of specific legislation. This imposes considerable complexity on the processes of government at the higher levels. For example, the attempt to introduce a modern value-aded tax has in- volved ten years of wrangling and even now is accepted only up to the year 2006. There- after, further voter approval needs to be sought. In spite of this lower level independence in most areas of government, in the area of expenditure assignment, the Constitution of 1848 attributes so-called "principal" responsi- bilities to each layer of government. Thus, the cantons have exclusive responsibility for public order, public welfare, health care, schools, regional and local land planning, high- ways, and water and other resources. The Confederation government has the responsi- bility for defense, citizenship, asylum, civil and penal law, social protection, economic and monetary management, energy policy, national transportation, and telecommunica- tions. However, the complex network of payments, subsidies, incentives, joint financing, and delegation that has evolved over the years tends to blur these apparently clear-cut lines of responsibility. As regards tax assignment, the original vertical separation gave the center the respon- sibility for all the indirect taxes, while the cantons and municipal governments were as- cribed the direct taxes on income and wealth. Over time, the center has acquired powers in the area of income taxation, mainly as a result of the evolving vertical imbalances in- herent in the previous system and the diverging revenue needs of different regions. There are now twenty-six separate income tax regimes at the lower levels of government in addition to the federal income tax and some municipal direct taxes. This is in sharp contrast with the uniformity now emerging in China. In spite of this evolution of the initial design, each tier of government retains either full or partial authority for a number of taxes and not only one. The subnational govern- ments, for example, have a right to levy, in addition to income tax, user charges and fees, and they have an exclusive right to tax motor vehicles (cantons). The Confederation now obtains about 41 percent of its total revenue from direct taxes, originally assigned to the lower levels of government, and 52 percent from indirect taxes. The cantons derive 50 percent of their total revenue from direct taxes and only about 8 percent from indirect taxes and licenses and other fees. Both federal and lower level go v- ernments are able to discharge their responsibilities without being financially too de- pendent on each other. Thus, the grant element of cantonal revenues is equal to only 21 percent of the total, which is considerably lower than in the more centralized case of Australia. Spahn noted, however, that the competing tax powers at the different levels of gov- ernment and the great diversity in fiscal federal arrangements create enormous problems of tax coordination, competition, and harmonization. In this regard, the Swiss case par- allels that of the United States, which also has a varied and complex system of subna- tional taxes. As regards equalization issues, another objective firmly embedded in Swiss law is that of providing for a similar level of public service in each canton without the need for taxes in some to be significantly more onerous than in others. This objective is achieved 30 Macroeconomic Management and Fiscal Decentralization through three types of vertical financial adjustment: some reimbursement of federal taxes; tax sharing; and specific purpose and normally conditional grants from the center. The federal government reimburses part of its collections of custom duty revenues to lower level governments. This is because certain responsibilities, including road building, have been passed down to the cantons for administration on behalf of the central gov- ernment. Tax sharing is done partly to compensate the cantons for the dilution of their tax sov- ereignty associated with the transfer of some greater tax powers to the center as ex- plained above. The tax-share revenues can be viewed as the main source of unconditional general revenue grants of the lower level governments. The tax shares are distributed in accordance with regional revenue collections, population, the level of private incomes, and by reference to a measure of fiscal needs including a limited set of identified expen- diture needs-but far fewer than in the Australian case. The last of these criteria is ac- cepted by the Swiss as a simple and fair way of compensating the special needs of par- ticular regions. In 1992, the cantons received 30 percent of the federal direct taxes on income and profits, 10 percent of the withholding tax, and 20 percent of the tax on exemptions from military service, which together constituted a distributable pool. The distribution criteria applied to this pool have similarities with the systems in both Canada and Germany. Above all, the fiscal revenue of the cantons is standardized so that distribution shares cannot reflect the differential tax effort of different cantons. In Germany, where tax rates are more standardized, this step is less important. Vertical intergovernmental transfers have been dominated by conditional grants-in- aid allocated in accordance with the policy priorities of the donor government. Many have strings attached in the form of matching requirements, and most have conditions requiring their pass through to lower levels of government such as the municipalities. While some equalization is involved in all three aspects of this system (that is, poorer cantons are certainly subsidized), the general judgment reported by Spahn is that equali- zation in grant programs is a matter of subsidiary interest only in the Swiss case. This is a significant contrast with, for example, Canada and also Germany. Germany The German Constitution established after World War II created the states (or Lander) as the first organs of government. The states continue to have a very important role in the legislative process, and they have provided some sort of model for the political structure of the present-day European Union. These constitutional arrangements were used from an early stage, Spahn explained, to establish the regional shares of the tumover tax as an important mechanism to rectify vertical imbalances as they occurred. In 1969 shared taxes were extended to include the income tax as well, and these are shared on a rigid one-to-one formula. But the vertical imbalances are still addressed only by way of the turnover tax or more correctly by its modern equivalent, the value-added tax. Overall shared taxes are very important in the German system, accounting as they do for about two-thirds of all tax revenues. Spahn noted that China, like Germany, has a uniform tax law, and it separates certain taxes to individual levels of government. Recently, China also has followed the German example of tax sharing. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 31 In its early years the German system resulted in large fiscal imbalances as the federal government expanded relative to the states, particularly when Germany was re- militarized within NATO. But later that situation reversed, with the expenditure shares of the states increasing relative to the federal government. Many labor-intensive services, such as education and health, expanded more rapidly in cost of provision than most fed- erally provided services. A few of the shared taxes are assigned individually to the federal level, and this pro- vides some fiscal basis for alleviating vertical imbalances. Unlike Australia, where most revenues first accrue to the center for distribution to the states, for the most part, Ger- many avoids major vertical fiscal imbalances, and this has major consequences for fiscal equalization. The way this problem is solved is via special agreements among the states. Germany is unique in that it has fully horizontal redistribution of resources without cen- tral governmental involvement. The conclusion for China is clear. If you want the gov- ernment to do the job of equalization, you must leave the central government with some surplus of fiscal means for redistribution. If you do not do this, you must rely, as in Ger- many, on explicit cooperation among the provinces to arrange the process of redistribu- tion and regional equity. Spahn explained that there are three levels of redistribution in the German system. The first involves the tax assignments, including the tax-sharing arrangements already described. Personal income taxes are shared among the federation, states, and munici- palities. The criteria for horizontal distribution is the residence principle-significantly different from the derivation principle used in China for the value-added tax sharing. The corporate income tax is shared using a "modified residence principle" as the crite- rion. It is difficult to apportion the income accruing to any one firm, and so the taxes it pays, to the different states in which that firm operates. The modified residence principle provides one reasonable solution to that problem through the use of a formula including local turnover and the local wage bill paid out in the different locations of the firm. This issue proved particularly interesting to the Chinese participants, especially because of the task of allocating the corporate profits and taxes of multinationals. It was noted that there is no unique theoretical solution to this problem: normally a formula approach as in Germany is the way to solve the problem. As regards the value-added tax, the German and the Chinese arrangements differ sig- nificantly. Presently, the federal government receives about 56 percent of the total com- pared with 44 percent for the Lander. The second element of the redistribution system is a definition of fiscal capacity for each state, which is roughly the sum of state tax revenues corrected for special burdens, population density, and urbanization. An equalization yardstick is then derived as the average per capita fiscal capacity calculated across all states and multiplied by each state's population. The final step is the equalization of revenues in accordance with that yardstick but using graduated rates. There is an underlying guarantee that fiscal capacity including the equalization transfers should reach at least 95 percent of the average across all states. Spahn noted that over time this German system resulted in a very even regional dis- tribution of infrastructure and a reasonably even regional distribution of income. This is in sharp contrast with concentration of development on certain "growth centers." He suggested that as China's development proceeds, more attention would need to be paid to regional equalization and "cohesion," if only for political reasons. But how much you 32 Macroeconomic Management and Fiscal Decentralization should equalize is a political matter depending on the willingness of the richer provinces to sacrifice some resources. The question of how to organize an equalization scheme is more straightforward, even though there are wide variations in international experiences as evidenced by the contrasts between Canada and Australia. Spahn argued that the Australian system, although it may seem to provide a rigorous and scientifically based set of criteria, includes many criteria that are arbitrary and politi- cally influenced as is the size of the pool to be distributed. Spahn advised against the adoption of such a scheme in China mainly because of the formidable informational re- quirements that its application would require (for example, regionalized data on female fertility and by age group) and the likelihood of a proliferation of informational require- ments. It was suggested that if China wishes to give the provinces general revenue to cover their basic expenditure needs, there are simpler methods available than those implicit in the Australian approach. Revenue equalization along Canadian lines would be prefer- able, for example, provided that it could ensure some elements of basic needs satisfac- tion. Finally, Spahn spoke briefly about investment grants. In the area of capital formation, the systems used in Germany, and in the European Union more generally, involved transfers requiring matching contributions. This system encourages good coordination between regional and center governments and might also be a useful model for China. Reflections on Remaining Problems in China The final substantive session of the seminar provided an opportunity for the three work- ing groups to present and discuss their findings. As was noted earlier, the topics given to the working groups were expenditure responsibility assignments; taxation responsibility assignments; and the design of an intergovernmental transfer system for vertical and horizontal equality. Each of the three working groups considered some or all of these topics. Group One This group reported first their conclusion that the problem of assigning expenditure re- sponsibilities was a crucial one still needing resolution. It was noted that this was a con- clusion that had surfaced at several earlier stages of the seminar proceedings. Next the group focused on a variety of issues pertaining to the Chinese system of in- tergovernmental transfers. The present arrangements, founded as they are on adjust- ments to a base-year allocation, lack flexibility and adaptability to changing circum- stances. Moreover, it was recognized that there is presently no transparency in the sys- tem and no scientific base against which to assess the levels of the needs of different re- gions. Thus, some scientific formula ought to be established to provide this, when adopting a more robust system. The issue was addressed of whether, within a new system of transfers, a formal Fi- nance or Grants Commission was needed along Australian or Indian lines. It was felt that such an arrangement was not necessarily required in China and that much of the work could be done within the Ministry of Finance itself. The more important thing, the group argued, was to establish a clear and transparent mechanism for deciding on the allocation Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 33 of grants. Above all, this should reduce the amount of discretion left to individuals or to the political process more generally. As regards taxation and the assignment of taxation responsibilities, the group ex- pressed concern about the relatively low level of the revenue:GDP ratio in China. Al- though that there was plenty of scope for raising it, such an improvement would require stronger laws, especially in relation to joint ventures and other foreign enterprises. It was also felt that there was scope for an improved revenue from the personal income tax and that higher receipts from nonwage income would need some significant redesign of the arrangements for reporting such incomes and for checking on evasion. It was not entirely clear from the group report what the implications of its findings were for the level of government at which the personal income tax should be assigned. However, it became clear in the ensuing discussions that there was much disagreement on this issue among the Chinese experts. Because of the relatively limited revenues to be expected from the personal income tax, some felt that it could be assigned safely to the local government rather than to the central level of government. Others, however, saw a considerable potential for revenues from this tax in some provinces, and they felt it needed to be assigned to the central government. Group Two This group also focused initially on expenditure responsibilities. It argued that the main issue for the near future in China is to establish an appropriate standard for public serv- ice provision in each province. Population size, the degree of urbanization, the size of the land area, cultural and ethnic differences, and the fiscal capacity of each province should be taken into account. The emerging gaps between provinces due to reforms were not primarily attributable to weaknesses in the scope of expenditure responsibilities between the center and local governments, the group argued. However, it did recognize some problems associated with the fact that policies and prescriptions were made at the central government level, while the expenditures themselves were generally more localized. As regards the assignment of tax responsibilities, the group broadly supported the status quo and saw no inherent problems with the 75:25 sharing of VAT revenues already agreed. It felt that without this level of revenue from one of the main taxes, the central government would find it extremely difficult to meet its expenditure responsibilities. However, there was room for an overall increase in the total of revenue collected, and the group agreed with Group One that the critical issue was the corporate income tax. Group Two was worried in particular about the taxation of large foreign investment companies that spread their activities over several provinces. On balance, the problems of competi- tion for the tax base and excessive incentives for interprovincial protection would be re- solved best if the corporate income tax was assigned to the central government. How- ever, there was a good case for sharing some of the revenues with the local governments. As regards transfers, it was explained that the Chinese system had three elements: the surrender of any local govemment surpluses of revenues over expenditures to the central government (this is equivalent to 72 percent of revenues in the case of Dalian); some earmarked grants; and some special grants to the poor regions. These three arrangements together result in some equalization, but the basis on which this occurs is an arbitrary one. The group agreed on the need for a more scientifically based system capable of con- tributing to the delivery of some minimum standards of public service across the country 34 Macroeconomic Management and Fiscal Decentralization as a whole. However, it anticipated that any such system would and should leave some discrepancies favoring the more economically successful provinces. Group Three This group began with the issue of taxation and argued that a modem international sys- tem of tax sharing was required in China. The reforms of January 1994 were likely to re- sult in an increase in the aggregate ratio of revenue to GDP, but at the same time they could reduce the revenues received by some of the local levels of government. Some ad- ditional ingredient seemed to be necessary to ensure that the richer provinces were asked to make a larger sacrifice of revenues as a response to the rapid changes in the balance of the economy. It was not entirely clear what this implied about the suitability of the VAT sharing system recently introduced. However, the group did note that the 75 percent share for the center would be greatly affected by the rules about incremental shares and that these would mean the receipt of far more than a 25 percent share in some localities. As regards the allocation of the revenues from corporate income tax, it was argued that the state enterprises had emerged from the Chinese system of planning and that those local authorities that had made investments under this system should receive the benefits in terms of the tax revenues. The group argued that any decisions regarding revenue sharing should be linked more directly with the issue of the expenditure responsibilities assigned to the local tiers of government. In this regard it was also important to ensure that the quality as well as the quantity of provision of public service were properly catered for and monitored. As regards the system of transfers, the group supported the need for a more scientific basis for making the allocations. However, it pointed to the enormous variation in cond i- tions in the different provinces of China. The needs of high altitude, mountainous, cold, inland, remote provinces such as Shanghai were radically different from the needs of the more accessible coastal provinces. These differences argued for some blending of condi- tional and unconditional grants and for the inclusion of both real and nominal criteria in the allocation formulas. Chinghai, for example, has a high average wage but generates little in the way of other incomes. Hence, a wage-based formula for allocating transfers would be quite misleading. The Views of the Foreign Experts We can conclude this report by contrasting these Chinese judgments about the emerging system of intergovernmental fiscal relationships with the views presented by foreign ex- perts during earlier sessions of the seminar. These opinions are arranged by topic rather than speaker, and it should be noted that some of the points were made by more than one speaker. EXPENDITURE ASSIGNMENTS. There was widespread agreement that expenditure as- signments is one of the most important areas still needing the attention of the Chinese authorities. Until this assignment is well defined, it is very hard to do what recent Chi- nese reforms have attempted-namely, to establish a sound basis for revenue assignment and tax sharing. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 35 In this area of system design, there was strong support for relying on the benefits principle as explained earlier by Ved Gandhi. However, where functions clearly overlap, it was also thought to be useful to establish a centralized framework, albeit one with various degrees of administration at the subnational levels as in Germany. Some of the subnational governments might not be able to administer properly everything that ide- ally is assigned to them. Hence, some residual central government powers (for example, to ensure reasonable minimum standards of provision) were called for. This could apply, for example, to some aspects of university administration. Given these exceptions, there was a great deal to be said for assigning expenditure re- sponsibilities to the lowest possible levels of government consistent with the coherence of the nation as a whole and the economic efficiency of service delivery (the "subsidiarity" principle of the European Union). This approach would be likely to improve incentives for the deliverers of services and also keep them closely in touch with the needs of the beneficiaries. It was noted that present arrangements in China tend to make the provin- cial governments the agents of the center and not equal partners in decisionmaking and delivery as they are, for example, in the Canadian system. The subsidiary issues raised in relation to expenditure assignments were as follows. First, it was pointed out that many of the services already provided by, or prospectively provided by, local governments in China involve zero or extremely low user charges. Examples included water, power, gas, and the Beijing subway services. Many felt these arrangements should be reviewed and more realistic user charges introduced in some cases. Second, there was thought to be a strong cases for removing the responsibility for as many as possible of the social services presently provided by SOEs and reassigning these explicitly as responsibilities of local governments. Among other things, this could greatly ease the tasks of enterprise restructuring and reform. Conversely, there were strong ar- guments for removing the local governments from direct involvement in the ownership and management of joint ventures and other commercial activities in their areas. This reassignment would be greatly eased if local governments were given more revenue autonomy in the conventional sense (see the next paragraphs). TAX ASSIGNMENTS AND TAX SHARING. The general opinion was that the decisions already taken in China on this subject conform reasonably well to theoretical principles and also with international experience. In particular, the treatment of customs duties and the VAT as central taxes seems to be correct. The main exception to this was the present assignment of the business income tax and the sharing of the revenues according to the location of the enterprises being taxed. It might be preferable, it was argued, to assign this tax as a central tax but then provide revenue to subnational governments through a formula-sharing approach as in Germany, or by allowing local surcharges on top of the central rate of tax as in Canada. There was general agreement that the sharing of the VAT revenues on the basis of a 1993 allocation and the derivation principle would be an insuf- ficiently flexible arrangement to meet the needs of a rapidly changing economy. It was argued also that there was much to be said for unifying the control over tax bases so as to avoid the competition for tax bases that has been experienced in India. Chinese arrangements so far have provided subnational governments with only lim- ited autonomy on the revenue side of the budget. Unless tax assignments and present tax 36 Macroeconomic Management and Fiscal Decentralization sharing arrangements are amended, the central government will collect considerably more revenue than it needs for its own purposes, and the subnational governments will become very dependent on transfers from the center to meet their needs. This situation is not necessarily unmanageable, as the Australian experience illustrates, but it is likely to cause difficulties. Thus, the vertical imbalances might be reduced somewhat by intro- ducing additional sources of tax revenue of relevance to local governments such as prop- erty taxes or a strengthened personal income tax. For the reasons already given, the verti- cal imbalances should not be reduced by limiting the scope of the expenditure responsi- bilities assigned to the local level. TRANSFERS TO ADDRESS VERTICAL AND HORIZONTAL IMBALANCES. A general view was expressed that a close matching of expenditure and tax assignments was desir- able in order to limit the scale of transfer-dependence of the subnational governments. However, given that some vertical imbalances would be likely in the Chinese system, the advice offered was to avoid a gap-filling approach of the type used in India, especially one that merely provides money to cover the unmet costs of certain local activities. Such an arrangement almost certainly would result in wasteful spending. It was suggested that a gap-filling approach should be considered only where a genuine central govern- ment function was transferred explicitly to the local governments. In these cases transfers should normally be conditional on the delivery of some adequate level of service. Special purpose transfer programs should be avoided if possible since they tend to proliferate in both nLmber and total cost. Their use for certain special purposes such as disaster relief and help to ethnic minorities was probably unavoidable, but attention should always be given to the negative disincentive effects. It is preferable to provide most transfers as general revenue grants using an approach like Canada's that avoids possible revenue- effort and other disincentive effects. However, objective rules for the allocations as well as monitoring of those parts of the transfers that are conditional are also called for. Some but not all of the speakers advocated for China a special Finance Commission or Grants Commission along Australian or Indian lines. The issue of horizontal equalization was acknowledged to be an inherently political one. China essentially had to decide whether it wished to pursue such equalization as a high-level policy goal (like Germany has) or accept a significant degree of ongoing inter- provincial inequality. More specifically, the tradeoff in the Chinese case seemed to be between, on the one hand, some curtailment of growth by requiring larger transfers from the fast-growing Eastern and Southern provinces and, on the other, some risks to na- tional political cohesion if the inequalities became too great and the fast-growing prov- inces were asked to bear only a modest additional fiscal burden. In the event that horizontal equalization arrangements were adopted in China, several reservations were expressed about the suitability of the Australian system based on ap- parently scientific formulas and objective principles. The informational requirements of such a system were once again argued to be prohibitive for the Chinese circumstances, and the residual scope for political interference, overriding scientific judgments, was ar- gued to be considerable. Instead there was considerable support among the foreign ex- perts for a simpler system, possibly along the lines of that presently operated in Canada. Intergovernmental Fiscal Relations in China: Report of a Senior Policy Seminar 37 Annex: List of Main Terms 1. Intergovernmental fiscal relations. Relations between different levels of government involved in government financial activities such as taxation, expenditure, borrowing, subsidies, transfers, and grants. 2. Vertical imbalances. Imbalances between the fiscal positions of various levels of gov- ernment: chronic deficits for some levels of govemment (for example, the central government) and surpluses for other levels (for example, the provincial govern- ments). 3. Horizontal imbalances. Imbalances between the fiscal positions of different govern- ments at the same level such as provincial, local, county, or prefecture levels. 4. Resource allocation efficiency. The allocation of economic resources (labor, capital, land) is efficient when it takes place in response to the returns, rewards, and economic sig- nals given by market conditions, unaffected by government intervention. 5. Fiscal equity. Equity between people, regions, industries, and economic sectors that results from the operation of the fiscal system and so from taxes, expenditures, bor- rowing, subsidies, and transfers. 6. Tax assignment. The formal assignment of sources of tax revenue to the different lev- els of government. 7. Tax exporting. The shifting of the burden of taxes imposed by one government onto the populations or industries of other provinces, regions, or localities. 8. "Piggy-backing" of taxes. This occurs when tax revenues are raised not by levying a new tax but by superimposing an additional rate on a tax that already exists and which is possibly imposed by some other authority. 9. Revenue sharing according to the derivation principle. The sharing of revenues according to where they have been derived or where they have originated. 1 0. Expenditure responsibilities. The formal assignment of spending responsibilities to different levels of government. 11. Benefit spillover. This occurs whenever the benefits of expenditures incurred by one government level accrue to the population or industries of other provinces, regions, localities. (This concept is the expenditure counterpart of tax exporting.) 12. Conditional grants. Grants provided by one level of government to another level of government if it meets certain conditions. 13. Open-ended grants. Grants without any cap or limit of Yuan amounts and which will be provided in the amounts needed to meet the objectives of the grant. 14. Fiscal accountability. Expenditures that are incurred must reflect the tastes and prefer- ences of the public at large, and public services should be provided at minimal cost in terms of the real resources used. 15. Macreconomic stability. Control of inflation of domestic prices as well as the volatility of the rate of change of the value of foreign currency. 2 Intergovernmental Fiscal Relations and Economic Performance Ved P. Gandhi It is well known that the public-private sector split of an economy (that is, the level of real resources absorbed by the public sector and the manner in which these resources are mo- bilized and utilized) can affect its performance significantly. What is less well known, however, is that the split of a given level of public resources between different levels of government can also affect an economy's performance for lack of a good fiscal policy. For example, it can make the establishment of an efficient and modern tax system extremely difficult, if more important taxes essentiaUy belong to lower level governments. As an- other example, it can create difficulties for the implementation of an optimal and (macroeconomically) sound public debt policy, if lower level governments have complete authority to borrow from whatever source they may wish. Tlhis chapter attempts to suggest the ways in which intergovernmental fiscal relations can influence (i) the efficiency with which public resources can be mobilized and utilized; (ii) the degree to which macroeconomic stabilization can be attained; (iii) the extent to which fiscal equity can be achieved; and (iv) the feasibility that a high level of overall performance of the economy can be continued. Before the chapter concludes, a few gen- eral principles, based primarily on economic considerations, are set out. These principles can guide the prospective reform of intergovernmental fiscal relations in China. The chapter is divided into four sections. The first identifies the major parameters of intergovernmental fiscal relations. The second section highlights the effects that each pa- rameter of intergovernmental fiscal structure could have on the determinants of a coun- try's economic performance. In the third section the economic principles for designing a good intergovernmental fiscal structure are outlined. The fourth section offers general observations concerning reform of intergovernmental fiscal relations. The views expressed in the paper are strictly personal and should in no case be attributed to the organization to which the author belongs. The author is grateful to Angelo Faria for his comments on an earlier draft. 39 40 Macroeconomic Management and Fiscal Decentralization Major Parameters of Intergovernmental Fiscal Relations and the Diversity of Country Practices Major Parameters Intergovernmental fiscal arrangements are a necessary requirement of decentralized gov- ernment administration, particularly in a large country.1 At least, five, generally interre- lated parameters define these arrangements. EXPENDITURE RESPONSIBILITIES. What level of government (central, provincial, or local) will be responsible for which public expenditures and which public services? The answer is often the result of historical, constitutional, and social developments in a country. TAX ASSIGNMENTS. What level of government (central, provincial, or local) will have the right to levy, collect, and retain which categories of taxes, and what will the methods of their collections and administration be? This, too, is frequently the consequence of historical, constitutional, and other factors. FISCAL IMBALANCES, IF ANY. Given their established expenditure responsibilities and tax assignments, and varying cost structures of public services (due to demographic factors and urbanization) and taxable capacities (due to resource endowments and de- gree of industrial development), different levels of government can end up with either fiscal surpluses or fiscal deficits, implying "vertical" and/or "horizontal" imbalances of varying degrees.2 How serious these imbalances are, can define the relative stability of intergovernmental fiscal arrangements. AVAILABILITY OF FINANCING ALTERNATIVES. A level of government facing fiscal deficits may truly have financing alternatives. Among others, these can include transfers from other (usually higher) levels of government in the form of revenue sharing (fixed proportions of revenue yields of certain taxes), grants (specific purpose or general grants, tied or untied grants, matching or nonmatching grants, conditional or unconditional grants, open-ended or closed grants), or subsidies. All these alternatives may be subject to criteria that the levels of government having fiscal deficits must meet. Financing their deficits through borrowings from banks, other financial institutions, the public at large, 1. There is some evidence that decentralization increases as a country moves away from central planning (for example, the Russian Federation). There is also evidence that, at times, autonomous governments may be willing to give up some of their national sovereignty, and form a confedera- tion, in the interest of reaping the benefits of a more unified economic space and larger markets (for example, the European Community). 2. Fiscal surpluses for one level of government, while other levels of government face chronic fiscal deficits, imply "vertical fiscal imbalances." Similarly, fiscal surpluses for wealthy and eco- nomically advanced provinces or localities, while less wealthy and economically backward prov- inces or localities face serious fiscal deficits, imply "horizontal fiscal imbalances." Horizontal fiscal imbalances also exist whenever the governments of richer and economically advanced provinces or localities are able to provide very high levels of public services, while the governments of poorer and economically backward provinces or localities are unable to provide even the most basic public services. Intergovernmental Fiscal Relations and Economic Performance 41 and even higher level goverrunents are other alternatives, and these, too, may be subject to preconditions. Where none of these financing alternatives is available, the levels of government with fiscal deficits must either curtail their public expenditures and lower the levels of public services or raise additional revenues from their "own" tax bases and other nontax sources. INSTITUTIONAL PARAMETERS. Finally, intergovernmental fiscal arrangements eve- rywhere contain some elements that are strictly institutional. Examples of these include the degree of autonomy and independence that governments enjoy in planning for and administering the public expenditure responsibilities given to them3 or levying or raising taxes assigned to them without interference from or support of other levels of govern- ment; freedom to grant tax reliefs or expenditure subsidies with "own" resources to en- tice enterprises and businesses from other jurisdictions; rules of contracting and pur- chasing for the provision of public services from "own" revenues, and so on. Effects of Intergovernmental Fiscal Structure on Economic Performance This section illustrates how each of the five major parameters of intergovernmental fiscal structure identified earlier can affect the workings and the performance of an economy. Expenditure Responsibilities Economic efficiency requires that expenditure responsibilities be allocated between levels of government in a way that provides public services according to citizens' preferences and at lowest possible resource cost. However, expenditure responsibilities will not meet these criteria in the following situations. * If public services whose needs can best be gauged by lower levels of governments (for example, needs for police and fire protection, water and sanitation, urban transport and road network) are not assigned to them, they will not be provided in adequate quantities or will not be provided most efficiently. * If public expenditures that can be quickly accelerated or slowed down (for exam- ple, public works program) or which automatically adjust with natural ups and downs of business activity in the economy (for example, unemployment benefit program) are not allotted to the central government, the capacity of the govern- ment to moderate fluctuations in private business activities, and to achieve macroeconomic stabilization, will be seriously eroded. 3. Perhaps the autonomy and independence of the lower levels of government will always be limited, because economic policies designed by the central government often have serious implica- tions for the budgets of provincial and local governments. For example, policy declarations by the central government relating to universal primary education and free health services may directly affect the expenditure levels of lower level governments. Other examples of central government policies bearing on the budgets of lower level governments are: interest rate policy on the cost of their borrowings; incomes policy on their civil service wage bills; pricing policy on the budget sub- sidies; and excise duties on energy products on the costs of urban transport and utilities. Trade and exchange rate policies, too, can affect the economies of provinces and localities, depending upon the composition of output and natural resources of the region. The intergovernmental fiscal struc- ture must be able to deal with the budgetary consequences of central government policies for the provincial and local governments and vice versa. 42 Macroeconomic Management and Fiscal Decentralization * If central governments were allotted public services whose overhead costs tended to grow with expansion or which suffered from "diseconomies of scale" (for ex- ample, provision of law and order or maintenance of smaller roads and urban streets), such public services would not be provided in optimal quantities. Similar problems would arise if lower levels of governments were allocated public ex- penditures whose benefits accrued widely and far beyond the frontiers of provin- cial or local areas (for example, provision of university education, financing of major roads and highways, or specialized hospitals). * If provincial and local governments incurred widely different amounts of public expenditures per capita, or provided public services at widely different levels, this could easily encourage unnecessary migration of labor and capital, unless high levels of public expenditures per capita in those areas were financed through high levels of taxes per capita as well. • If lower level governments are responsible for expenditures with large potential benefits to other regions or "spillover effects" on other jurisdictions (for example, on education, road maintenance, or environmental improvement), they are un- likely to be adequately provided for with local fiscal resources alone, and this could have serious consequences for the performance of the economy. Tax Assignment Economic efficiency requires the assignment of taxes to different levels of government in a way that the built-in elasticity, equity, and administrative ease of different taxes is fully exploited and the incidence of taxes is borne by those for whom it is intended and who benefit from the use of revenues raised from them. The way taxes are assigned should also support such desirable goals as economic growth, macroeconomic stability, and dis- tributive justice. However, in reality, this may not be so: - If taxes were assigned to provincial and/or local governments whose burden could easily be shifted or exported to consumers of other provinces and localities (for example, natural resource tax, manufacturing level sales tax or excise duty, import and export duties at the ports, airports or border areas, taxes on tourism), this not only could be grossly unjust, but also could encourage lower level gov- ernments to pitch those taxes at excessively high levels to the detriment of the growth of the national economy. * If taxes were assigned to central governments whose administration required de- tailed information (for example, property tax, land tax), and which the central government could gather but only at high costs, such taxes would be poorly ad- ministered and easily evaded, and the cost of their collection would be undesira- bly high. * If taxes that were primarily redistributive in character (for example, a progressive personal income tax, a progressive wealth tax, or a progressive inheritance tax) were not levied by the central government, and on a global basis, the overall eq- uity orientation of the tax system could be seriously jeopardized. * If taxes that normally have large built-in elasticity (for example, personal income tax, enterprise profits tax, sales tax, or value-added tax), and whose revenues Intergovernmental Fiscal Relations and Economic Performance 43 grow automatically in boom periods and decline in recessionary periods, were not assigned to the central government, the ability of the government to achieve macroeconomic stabilization would be greatly curtailed. Fiscal Imbalances Maintaining macroeconomic stability and providing for economic growth require that overall fiscal imbalance for all levels of government, net of transfers and grants, is such so as not to cause inflation or create shortages of investible funds for either the private or productive sectors of the economy. Nevertheless, this may not always be the case: * If expenditure responsibilities of central governments were mismatched with taxes assigned to them (that is, public expenditures were far more income elastic than tax revenues), central government budgets would chronically be in deficit, with serious implications for macroeconomic stabilization. * If any level of government often had budgetary surpluses because of given ex- penditure responsibilities and tax assignments, it would have an incentive either to overspend or to lower tax burdens on its citizens. The former would become a source of inefficient spending, while the latter could become a source of tax com- petition with governments of other jurisdictions. * If there were serious vertical imbalances, and appropriate mechanisms for the transfers of resources from the surplus governments to deficit governments did not exist, governments with budgetary deficits would be forced to scale back their public expenditures, including expenditures on social and economic infrastruc- ture, or to increase the ratio of taxes and user charges, including those with seri- ous disincentive effects. Either way, the performance of the economy would suf- fer, and fiscal equity would be seriously eroded. * If serious horizontal imbalances existed and certain governments were unable to provide even minimum levels of public services while others were able to offer levels of services far above the average, it would encourage undue movement of labor and capital, which in itself may be undesirable. With increasing population density, such movement of labor and capital could raise the "average cost" of providing public services in well-off areas and result in their inefficient provision. Financing Alternatives Macroeconomic stability requires that financing alternatives for meeting fiscal imbalances should be noninflationary. Besides, neither the curtailment of essential public services nor the raising of distortionary and disincentive causing income and profit taxes should be used as financing alternatives, since they can seriously curtail the pace of economic growth. In reality, however, inappropriate financing alternatives often tend to be used to "finance" fiscal imbalances: * If the only financing alternative for governments with fiscal deficits were to cut their public expenditures (including those on essential public services and the op- eration and maintenance of infrastructure) or to raise their taxes to excessively high levels, there could be serious equity and efficiency consequences. In other 44 Macroeconomic Management and Fisci. Decentralization words, such decisions could seriously affect overall fiscal equity and/or cause undesirable mobility of labor and capital. They could also affect the performance of the national economy, if the province or the locality were an important element of the national economic space for the transportation of a country's imports or ex- ports or the supply of raw materials and natural resources to industries in other areas of the country. * If the central government shared revenues from its most buoyant taxes with the lower levels of government, it could discourage local tax effort and encourage wasteful spending. * If the lower levels of government received subsidies, which were not earmarked to pay for local expenditures with positive externality or "spillover" effects or which did not require that those goods be provided at minimum possible costs, much wasteful use of limited fiscal resources may take place. * If the lower levels of governments had fiscal deficits and could easily borrow from banks and other financial institutions (as in Brazil and India), it could seriously erode the capacity of the central governments to formulate and implement a monetary policy consistent with macroeconomic stability, crowd out availability of credit to the private sector, encourage undesirable interprovincial competition for borrowed funds, and raise the interest cost of productive investment. e If the lower levels of governments with fiscal deficits depended heavily on central government budgets for revenue sharing, transfers, and subsidies (all with rigid formulas), the flexibility of the central govermnents to manage their fiscal policies consistent with macroeconomic stability would be seriously curtailed. - If the budget of the higher levels of a government was in deficit while the budgets of the lower levels of government were in surplus, or the budgets of certain pro- vincial or local governments were in deficit while those of others were in surplus, and if no mechanisms existed for the transfer of fiscal resources between govern- ments, allocation of fiscal resources nationwide will be grossly inefficient. There could be serious implications for macroeconomic stability as well as overall per- formance of the economy, depending upon how governments with deficits fi- nanced their deficits (for example, through monetary expansion, curtailing their public expenditures, or increasing revenues from disincentive causing and distor- tionary taxes). Institutional Aspects Resource allocation in the economy and the costs at which public services are provided can also be seriously affected by the institutional aspects of intergovernmental fiscal rela- tions, as follows: Providing a large degree of independence and autonomy to lower levels of gov- ernment in carrying out their expenditure responsibilities or formulating their policies with respect to assigned taxes can help ensure their accountability (matching local preferences for public services) and allow efficient allocation of their fiscal resources. Intergovernmental Fiscal Relations and Economic Performance 45 * If lower level govemments are given the responsibility for administering central government taxes, and if they enjoy a large degree of autonomy in doing so, they can erode central govemrnent revenues through generous tax incentives, and they can seriously affect macroeconomic stability. The loss of revenues to the central government, if those revenues are large, can curtail its capacity to meet its public expenditures, including providing grants to local govemments in poor regions to meet the most basic needs of the population. At the same time, the local govern- ments that brought about this loss of revenue will enjoy prosperity without pay- ing for it. This is a clear case of undesirable inequity. Allowing lower levels of govemment complete independence with respect to their tax policies can result in competition among them for tax reliefs and concessions to entice businesses and enterprises (with serious implications for the efficiency of allocation of resources nationwide). It can also mean unnecessary duplication of tax administration efforts, and an inefficient use of limited tax administration re- sources available nationally, if the relevant information gathered by one taxing ju- risdiction about taxpayers and their economic characteristics goes unused by other taxing jurisdictions. * Permitting provinces and localities to limit contracting and purchasing within their own borders (and not nationwide) could result in the provision of public services at higher costs that will be inefficient. Economic Principles for Designing a "Good" Intergovernmental Fiscal Structure Clearly, the major parameters of an intergovernmental fiscal structure can affect the effi- ciency of allocation of resources in an economy as well as other determinants of its per- formance.4 Nowhere are intergovernmental fiscal structures designed from scratch (or de novo), and economic considerations are not the only factors in their design. Nevertheless, certain economic principles have been developed that can help guide the reform of inter- governmental fiscal structure in any federal country, provided its policymakers are con- cerned about the efficiency of resource allocation and the performance of the economy. I would now like to highlight a few of these economic principles. Expenditure assignments to different levels of govemment should be guided by at least three economic principles. First, public services that benefit everyone (for example, de- fense) and whose supply benefits from economies of scale (for example, electricity gen- eration) should be provided by the central government. Second, public services whose benefits are localized (for example, law and order) and for which different people can have different preferences (for example, local roads) should be provided by local gov- ernments. Third, public services whose benefits spill over to jurisdictions other than those where they are actually provided (for example, public education and public health) 4. The intergovernmental fiscal structure can affect the performance of the economy, and the economy's performance, in turn, can affect intergovernmental fiscal outcomes. For example, the buoyancy of taxes assigned to the central government can affect the size of the revenue sharing pool, and the buoyancy of taxes assigned to lower level governments can reduce their needs for grants from higher level governments. 46 Macroeconomic Management and Fiscal Decentralization either should be provided centrally or should be subsidized by the central government, depending on the estimated value of the spill-over benefit. Tax assignments to different levels of government could be guided by at least three ba- sic principles. First, taxes that are levied on (a) relatively mobile factors of production (for example, labor and capital), (b) at progressive rates (for example, personal incomes), (c) cyclically sensitive tax bases (for example, personal incomes and corporate profits), and (d) tax bases that are unevenly distributed among jurisdictions (for example, natural re- sources) should all be assigned to the central government. Second, taxes that are levied on immobile tax bases (for example, land and property), and whose burden is not ex- portable to other jurisdictions, should be assigned to local governments. (Neither taxes on the consumption of goods and services, nor the corporate income tax, meet this crite- rion.) Third, large revenue yielding taxes that are inherently buoyant (for example, per- sonal income tax, enterprise profits tax, broad-based consumption tax) can be assigned to the central government or to local governments. If they are assigned to the latter, their rates should vary within a limited range, and their base should be rnore or less uniform, if tax competition among jurisdictions is to be avoided. Expenditure assignments should match tax assignments, as far as possible, in order to avoid largefiscal imbalances at different levels of government. If they are unmatched, and lower level governments are likely to face chronic fiscal deficits, "piggybacking" of the more elastic taxes of central governments (for example, of income taxes and/or value- added tax), or of overlapping assigned taxes, should be allowed. If, despite this strategy, fiscal imbalances continue for lower levels of government, alternative financing methods may have to be considered. Revenue sharing should relate to any central government tax that (a) is progressive and/or has a large built-in elasticity (for example, personal income tax or enterprise in- come tax) and (b) is less distortionary and is a large revenue source (for example, value- added tax or excise duties). No economic principles exist that can help decide the exact proportions of the revenues of these taxes to be transferred to the revenue sharing pool. Whatever the size of that pool, its distribution among different lower level governments, as I mentioned earlier, can be guided by either the "needs" principle or the "derivation" principle. Depending upon whether the equity or the efficiency objective is considered paramount, the former or the latter will become the main basis for distribution. "Tax ef- fort" made by the lower level government could be added as yet another factor in de- signing the revenue distribution formula. Grants are of different kinds, and each one has its own raison d'etre and meets one or more special needs. Generally, however, matching or conditional grants are preferable to nonmatching or unconditional grants (since they can encourage local tax effort and dis- courage wasteful spending); closed grants are preferable to open-ended grants (since they will imply constrained and quantifiable costs to the central budget); and earmarked grants are preferable to nonearmarked grants (since they will ensure that grant objectives will be achieved, such as meeting otherwise unmet basic needs of the local population). Whatever forms grants take, at least two principles must guide their design. First, they must make a measurable contribution to a clearly defined goal. Second, they must be in- versely related to the grantee government's "fiscal capacity" and directly related to its "fiscal needs." "Fiscal capacity" and "fiscal needs" of each jurisdiction must be estimated with a clear, quantifiable, and transparent formula, as far as possible. Intergovernmental Fiscal Relations and Economic Performance 47 Subsidies should be limited to finance spill-over benefits, if any, from the public ex- penditures of lower level governments, and they should be transparent as far as possible. Hidden subsidies, in the form of tax reliefs (for example, to holders of local government bonds), should be avoided. Borrowings by lower level governments from the public to finance their deficits are le- gitimate, if they are not so excessive as to "crowd out" the private sector, and if the funds are raised at market interest rates, without any implicit or explicit subsidy. Borrowings by them from the banking system should be limited and from the central bank prohib- ited, if the central government is not to lose its control over the country's monetary policy and if local government finances are not to become a major cause of macroeconomic in- stability and inflation in the economy. The foregoing principles are primarily derived from economic considerations. How- ever, intergovernmental fiscal relations in no country of the world are based upon eco- nomic considerations alone. It is, therefore, quite reasonable to expect that historical, po- litical, social, and other noneconomic considerations will continue to play an important role in the design of the intergovernmental fiscal structure in your country. Final Words Before concluding the paper, I must reiterate that intergovernmental fiscal relations in every country of the world will continue to remain prisoners of history, traditions, insti- tutions, and politics. What can be expected, at best, is only an incremental change from the present reality rather than a very profound and fundamental change. Whatever the extent and form of change, one hopes that it will make the structure transparent, stable, and predictable. Changes along these lines alone will be deserving of being called reform. Would the establishment of an autonomous intergovernmental fiscal commission on a permanent basis, with a secretariat of its own, or on an ad hoc basis every five years (as in India) help? It would certainly make intergovernmental fiscal relations more transpa r- ent. WVhether or not the structure will become more stable and predictable is less clear. This will depend upon how autonomous and apolitical a body that commission will be. A review of experiences of intergovernmental fiscal relations in different countries of the world can help one see what the alternative solutions to common problems are, and it can help one decide if any of the alternative solutions adopted elsewhere are applicable to a given country. It is hoped that this chapter has provided a framework to understand and debate the various alternatives that a country may have in the reform of its intergov- ernmental fiscal structure. 3 Intergovernmental Fiscal Relations in Australia Paul Bernd Spahn and Anwar Shah General Characteristics of the Australian Federation Australia was established as a federation at the beginning of this century when six sepa- rate British colonies formed the Commonwealth.1 The aim of this political confederation was twofold: defense, and a more intensive economic cooperation and exchange on the Australian continent.2 The Commonwealth obtained relatively few exclusive powers, but its responsibilities to be exerted concurrently with the states were quite extensive. In these areas of concurrent responsibilities, Commonwealth legislation is paramount. Thus concurrency provisions have remarkably strengthened the center over the years. Never- theless, the Australian states retain substantial power in terms of spending functions. 3 As to government revenue, the states had expected their financial independence to be pre- served under the Constitution. However, legal constraints as well as historical events have led to centralization of revenue functions and an ever-increasing financial depend- ence of the states on the central government. Australia has a two-tiered highly centralized federation of six states (New South Wales, Victoria, Queensland, Western Australia, Tasmania) and two federal territories (Australian Capital Territory and the Northern Territory) with a total of 1994 end-year population of 17.9 million. The center emphasizes uniformity of public services across the nation and uses conditional grants to achieve that purpose. Tax administration and collection are central, representing 80 percent of revenues. Local governments are exten- sions of states but are given reasonable autonomy in local service delivery. The national government has sole responsibility for defense, trade, immigration, ex- ternal affairs, social security, and employment. States are responsible for education, health and social services, transport, railways, electricity, and water. (See Figures 3-1 and 3-2 for composition of public spending by various levels of government.) 1. In addition to the six founding states, two territories have more recently achieved self-rule, the Northern Territory (in 1978) and the Australian Capital Territory (in 1989). 2. At that time, the states' overseas trade connections were generally much more developed than interstate trading within Australia. 3. The following account is largely based on Mathews and Jay (1972), Hunter (1977), Mathews (1982), Galligan and Walsh (1990), Walsh (1990), Groenewegen (1991). 49 50 Macroeconomic Management and Fiscal Decentralization Figure 3-1. Own-Purpose Outlays by Category, 1992-93 Federal Commonwealth 16% 3% 5%% 8% 4% 13% State and Territory Social Security & Welfare O Health 29% 26% ~~~~~*Defense _ Transport & Communications * General Public Services [MlI Education 5% 6% EB~~~~~~ Other 7%3 '> > _ O Public Order & Safety 10% 7% M Housing & Community Amenities U Fuel & Energy D Recreation & Culture Local 3% 16%ll1 66% 17% Source: The Treasurer of the Commonwealth of Australia (1994). Intergovernmental Fiscal Relations in Australia 51 Figure 3-2. Commonwealth and State/Local Outlays by Major Function, 1992-93 Billions of dollars 45 40 - E Direct Commonwealth Outlays 355 - Transfers from the Commonwealth 300- State/local Outlays 25 - 20- 0- Social Security Health Education Transport Housing and and Welfare and Community Communications Amenities Source: The Treasurer of the Commonwealth of Australia (1994). The federal govemment nonetheless exercises stiong influence in areas of state re- sponsibility through conditional transfers. In tax assignment, customs and excises are reserved for the center, and concurrent responsibilities are assigned in all other areas. One half of customs proceeds are mandated for states. The Uniform Taxation Act of 1942 eliminated the role of states for income taxes, and subsequent court rulings closed sales and excise taxation fields to states. State-local governments are responsible for 50 percent of the total outlay of the public sector but raise less than 20 percent of revenues (see Fig- ure 3-3). Vertical Tax Assignment, Federal and State Taxation, and Grants Tax assignment in Australia may be seen as strictly following the layer-cake approach whereby each level of govemment is reserved its own tax base. The strict separation of taxable sources avoids tax competition among governments; only fees and fines are em- ployed by all levels of govemment. A separation of tax sources is also found in other fed- erations, like the United States, Canada, or Switzerland, but, contrary to the latter,the prescript has led, under the conditions prevailing in Australia, to the federal government controlling virtually all lucrative taxes in the end. It may be useful to review some of the major historical events that have led to this consequence. The historic perspective illus- trates the importance of idiosyncratic constitutional and political constraints that have shaped intergovemmental fiscal relations in Australia and continue to bear on modem arrangements. 52 Macroeconomic Management and Fiscal Decentralization Figure 3-3. Revenue by Source, 1992-93 Commonwealth 4% 2 g ~~~~47% 23% 15% State and Territory El Income taxes levied on individuals El Income taxes levied on enterprises 10% * Taxes on provision of goods and services 1 Other taxes, fees and fines 8% [1111 Commonwealth payments EB Payroll taxes 10% N~~~~~~~~~~ Taxes on property Taxes on uses of goods and 1% 9% perfomance of activities 0 Net operating surplus of PTEs ZI Fees and fines Local ElI Payments from the States M Payments direct from the 10% Commonwealth 6% Source: The Treasurer of the Commonwealth of Australia (1994). Intergovernmental Fiscal Relations in Australia 53 A Brief Historic Outline The following historical events have contributed to the steady erosion of state tax com- petences in Australia: * At the inception of the federation, the Australian states transferred, via the Con- stitution, exclusive powers on customs and excise revenues to the Common- wealth. However, starting in 1908, the High Court has consistently interpreted these powers more and more generously in favor of the central government by extending the definition of excises.4 In particular, state sales taxes, as they exist in the United States, or a value-added tax operating at the state level, as in the mem- ber states of the European Union, could not be reconciled with the Constitution in Australia. * To compensate them for the loss of customs duties and excise taxes, which then were relatively more important than today, the states were originally handed back three-quarters of "surplus" revenue collected (Section 87 of the Constitution). Yet on the expiry of the clause, the central government did not continue this form of tax sharing. It accorded per capita financial assistance instead (1911-27), a lump- surn grant that was rapidly eroded in real terms by inflation.5 A special grant was accorded to Western Australia in view of its high contribution of customs duties. - An offer was made in the early 1920s by the Bruce-Page government to vacate the area of personal income taxation in favor of the states. This failed on the rejection of New South Wales (all other states had accepted). A similar move to hand over smaller taxes and to reduce income taxation was also discarded at a Premiers' Conference in 1926. The states had then no desire to administer these unpopular taxes. * A change in the provision of financial assistance through the Financial Agreement of 1927 led to the Commonwealth intruding on state sovereignty: the (unconditional) per capita grants were transformed into annual contributions to- ward interest charges on the public debt of states. The Agreement also introduced greater Commonwealth control of state borrowing through the establishment of a Loan Council. * In the early 1930s, discontent with the fiscal federal arrangements was at a peak. Western Australia passed a referendum in favor of secession. A system of fiscal equalization for the financially weaker states, involving special grants from the Commonwealth, was introduced in 1933. An independent Commonwealth Grants Commission was established to review states' claims for special assistance and make recommendations. * The most important fiscal development during World War II was the taking-over, in 1942, of all income tax by the Commonwealth as a wartime emergency act. The federal government, "having tasted the delights of full control of this great reve- nue machine, wanted to retain that control" (Bird 1986, p. 110)-and was success- 4. The important rulings in 1908 were the Wire-netting Case and the Steel Rails Case. 5. "The real burden of grants (to the Commonwealth) had been at least halved by the increase in prices during the war, and of course was far less than 75 per cent of customs and excise revenue would have been" (Mathews and Jay 1972, p. 118). 54 Macroeconomic Management and Fiscal Decentralization ful in doing so.6 At the same time, income tax rates that had varied across states became uniform across Australia. Uniformity of the income tax has been pre- served until now despite various attempts to give the states more "flexibility" in revenue raising, for instance through a surcharge on the Commonwealth income tax. * Financial compensation for the loss in income tax was given through "tax reim- bursement grants," which became-with amendments-the cornerstone of today's intergovernmental fiscal arrangements in Australia. This issue will be discussed in more detail later in this chapter. * The only (minor) reversal of this centripetal trend came in 1971 with the handing over of payroll taxation to the states.7 * Since World War II, there has been a massive increase in specific-purpose pay- ments to the states in the main functional fields for which they have constitutional responsibility (such as education and health).8 The states resent this as another form of Commonwealth guardianship. The Present Tax System The important income taxes on individuals, enterprises, and nonresidents are now as- signed to the Commonwealth. Furthermore, it collects the sales tax, excise taxes, and taxes on international trade. The states' most important own revenue sources are the pay- roll tax, taxes on financial and capital transactions, and taxes on gambling, on insurance, and on motor vehicles. Moreover, the states employ "franchise fees," taxes on excisable goods converted into service taxes for constitutional reasons.9 Local governments, finally, are given taxes on immovable property, the municipal rates. The assignment of taxes has lead to a significant "vertical fiscal imbalance" in Aus- tralia, a mismatch between own resources available and expenditure needs at the vari- ous levels of government. The Commonwealth's own revenues greatly exceed its outlay responsibilities.10 However, the states-which retain control over almost all major gov- errunent functions (except defense, foreign affairs, interstate matters, social welfare, postal and communications, and banking and insurance) are denied access to the two 6. Four of the states challenged the validity of the legislation that transferred income taxes to the Commonwealth government. Yet to the dismay of the states, the High Court declared the essential parts of this legislation valid, not merely under defense powers, but also under normal conditions of peace (Mathews and Jay 1972, pp. 171-174). 7. Arguably, legislation under the Fraser government allowed the states to impose surcharges (or give rebates) on federal income taxation. However, this provision was never applied. The Commonwealth scrapped the legislation in 1988 because it feared the provision might be em- ployed. 8. In 1982-92, 44 percent of Commonwealth transfers to the states were specific purpose pay- ments (tied grants). 9. These taxes have emerged in response to the states being denied access to excises by the Con- stitution. The states have argued that a franchise is a legal entitlement, accorded to the suppliers of goods, to deliver certain commercial services (the distribution of goods). But such license taxes have emerged in typical areas of excise taxation (on alcoholic beverages, on tobacco, and on petro- leum products), and the tax base is often indistinguishable from the excise tax base itself. It is there- fore a puzzle that constitutional lawyers and even the Courts have accepted state franchise fees. 10. In 1992-93, Commonwealth revenue was 67 percent of total public sector revenue, while fi- nal demand for Commonwealth services was only 33 percent (Rye and Searl 1994). Intergovernmental Fiscal Relations in Australia 55 major sources of revenue in Australia: income taxation and taxes on the production and distribution of goods.11 Commonwealth taxes were about two-thirds of all public sector revenue in fiscal 1992-93, while public sector final demand corresponded to only one- third of total government spending.'2 The states, however, met about 60 percent of public sector final demand, and collected only 28 percent of all revenue, leaving a sig- nificant "fiscal gap" to be closed. Local governments collected about 5 percent of reve- nue while satisfying 8 percent of total public sector demand (see Figure 3-4). Figure 3-4. Public Sector Revenue and Demand in Australia, Fiscal 1992-93 Public Sector Revenue Local (4.81 %) State (27.80%)- Commonwealth (67.39%)- Public Sector Final Demand Local (8.33%)- State (59.25%)- Commonwealth (32.42%) Source: The Treasuer of the Commonwealth of Australia (1994,1994-95 Budget Paper No. 1). 11. The High Court has ruled the states out of taxes on goods but not services; this is the reason why the states imposed "franchise fees" on liquor and tobacco and petroleum, which are calculated on the value of past period sales (see also footnote 9). 12. Final public demand is measured by public final consumption expenditure (which excludes transfer payments) plus public gross fixed capital expenditure. In terms of total government out- lays (that include transfer payments for social welfare, health and other purposes), the Common- wealth's share is higher. It was around 43 percent in 1992-93 (1994-95 Budget Paper No. 1, State- ment 6, p. 6.5). 56 Macroeconomic Management and Fiscal Decentralization Australian states thus depend largely on revenue to be transferred from the Com- monwealth. The financing of state budgets is predominantly effected through uncondi- tional general revenue assistance as well as specific-purpose payments or categorical grants (according to Section 96 of the Constitution). In 1992-93, the states relied for more than 40 percent of their revenue needs on Commonwealth grants.13 This fact made some authors speak of the states' dependence "on federal largesse" (Bird 1986, p. 125) or of the Commonwealth's "financial domination over the states" (Mathews and Jay 1972, p. 291). Apart from vertical fiscal imbalance, the Australian tax system also relies heavily on the income tax. Income taxes alone account for about 60 percent of total revenue. The next single tax, sales tax, constitutes only one-sixth of the magnitude of income taxes. There is wide agreement among scholars of federalism that both the lopsided reliance on the income tax and the vertical fiscal imbalance among the tiers of government con- stitute the chief deficiencies of the modern Australian tax system. In order to remedy the excessive reliance on income tax, the opposition parties at the Commonwealth level proposed, in their latest electoral platform "Fightback!", the intro- duction of a broader based consumption tax; their defeat in the elections is widely as- cribed to this particular reform proposal. For political reasons, a broad-based goods and services tax (GST) seems to be out of the question for some time, all the more as the pres- ent Commonwealth government is strongly opposed to such a tax. As to the vertical as- signment of taxes, a Premiers' Conference in 1990 essentially affirmed the present system,14 although the position is now being reviewed as part of a general process of reforming intergovernmental relations. Tax sharing had been tried earlier (under the Fraser government in the 1970s) but was later abandoned. Since a reassignment of taxes seems to be out of the question, tax sharing is likely to come back on the political agenda. Vertical fiscal imbalance is now essentially dealt with through an extensive system of grants. The contribution of taxes and grants to total revenue at each level of government is depicted in Table 3-1 (for 1989-90). Financial Assistance and Revenue Sharing By far the most important single revenue source of subcentral governments in Australia is grant money. These grants are mainly paid in the form of general purpose financial assistance, which is explained by their historic roots as "income tax reimbursements" fol- lowing the implementation of uniform income taxation in 1942. The Tax Reimbursement Act provided the legal basis for grants to any state that abstained from levying income tax.15 In addition, more than half of Commonwealth assistance is in the form of tied or conditioned grants (for health, education, roads, and public housing). In 1992-93, the total of transfers from the Commonwealth to the states was $31.7 billion in gross terms, whereas state expenditures amounted to $51.8 billion.The various categories of grants are shown in Figure 3-5. 13.1994-95 Budget Paper No. 1, p. 6.7. 14. At a Special Premiers' Conference in July 1991, there were signs that the Commonwealth was prepared to accept greater revenue freedom for the states; for a summary of results see Walsh (1991, pp. 18-20). 15. The states were thus not prohibited from exploiting this tax base; yet they were encouraged to refrain from doing so (Mathews and Jay 1972, p. 175). Intergovernmental Fiscal Relations in Australia 57 Table 3-1. Sources of Revenue for Each Level of Government in Australia, 1991-92 (Percentage of total revenues) Revenue source Commonwealth State Local Own-source revenue Taxes, fees and fines: Personal income tax 48 Company income tax 16 Excises 11 Sales tax 9 Customs duty 3 - Payroll tax 1 9 Stamp duties b - 6 Motor vehicle taxes - 4 Franchise taxes - 4 Municipal rates - - 51 Other 3 11 3 Total 91 34 54 Net operating surplus' 3 10 7 Other 6 14 17 Total 100 58 78 Payments from higher levels of government From Commonwealthd na 42 14 From state na na 8 Total revenue 100 100 100 a. Includes fringe benefits tax. b. Includes financial institutions duties. c. Net operating surplus is included as revenue in the ABS Government Financial Estimates. d. Payments from the Commonwealth to states for on-passing to local government recorder under local. Source: Fletcher and Walsh (1993). 58 Macroeconomic Management and Fiscal Decentralization The more recent history of general revenue assistance to the Australian states falls into three phases: from 1959 to 1976, when general revenue funds were given to the states under the label of "financial assistance grants"; from 1976 to 1985, when such revenue was transferred to the states through revenue sharing; and from 1986 onward, when financial assistance grants were reintroduced. Financial assistance grants were increased annually (from 1959 to 1976) by a formula stressing, for each state, three factors: population changes, average wage increases, and a so-called "betterment factor" designed to allow the states to expand their relative level of services.16 On their reintroduction, these grants have been determined "by fiscal equali- zation factors based on revenue/expenditure disabilities for different states while the size of the pool is determined by a percentage growth rate set to reflect specified real (inflation-adjusted) changes in assistance which can be negative when fiscal restraint is Figure 3-5. Commonwealth Transfers by Category, Fiscal 1992-93 General revenue funds for Current purposes (44.37%) . Capital purpose (1.57%) - ' ' ' Specific purpose funds for Current purposes (40.22%) Capital purposes (I 3.84%) Source: The Treasurer of the Commonwealth of Australia (1994, 1994-95 Budget Paper No. 3, Table 17). 16. This betterment factor was 1.2 percent in the beginning; it was then raised to 1.8 percent in 1971 and 3 percent in 1975. Intergovernmental Fiscal Relations in Australia 59 called for" (Groenewegen 1991). Since 1990 the Commonwealth has agreed to maintain the real value of such grants in order to provide greater certainty in the funding of state budgets. The Premiers' Conference of March 1994 extended this commitment in per cap- ita terms for the following three years.17 The revenue-sharing interlude had been initiated by Fraser's "New Federalism" poli- cies. Sharing was first adopted for the income tax alone. The pool was formed by ap- plying first a fixed proportion and later an increasing proportion of the tax proceeds collected in the states. This led to the states' revenues being determined by factors simi- lar to those used for financial assistance grants before. Later, revenue sharing was ex- tended to federal tax receipts as a whole, which somewhat lowered the rate of growth of the states' revenue pool (since revenue elasticities of other Commonwealth taxes are typically lower than for the income tax). 18 As noted before, the revenue-sharing experiment was abandoned largely because of uncertainties about the Commonwealth's planned tax policies and for political rea- sons-since the states had found themselves to be "more vulnerable to unilateral federal decisions on tax policy" (Bird 1986, p. 116). General revenue assistance is given for recurrent and for capital purposes in Austra- lia. The latter reflect the fact that the Commonwealth used to borrow on behalf of the states and distribute the funds through the Loan Council. This program, however, has become an anachronism, and the Premiers' Conference decided to abolish this type of grant in 1994. Specific purpose grants in the form of recurrent and capital grants are given for spe- cific state functions like social services (health, education), social security and welfare, economic services (roads, transport, industry assistance, water resources), and other services (like housing and urban renewal, regional development, disaster relief and debt charges). Three types can be distinguished: (i) payments "to" states for funding direct state outlays, (ii) payments "through" states to be passed on by the states to other bodies or individuals,19 and (iii) direct payments to local governments. The conditions attached to these grants vary widely. One category simply requires the payment to be made for a specified activity, with varying degrees of budgetary discretion available to the states; another entails program stipulations (for instance, for a hospital funding grant, that free public hospital treatment to Medicare patients is provided). Or grants may hinge on the observance of principles of service provision and program delivery mechanisms as es- tablished by Commonwealth state agreements (1994-95 Budget Paper No. 3, p. 31). From a political point of view, specific-purpose grants are a way to "signal" the grantor government's concern for specific public functions that lie outside its own realm of competence. This may pay off in national elections, yet it may also be resented by state legislators and governments (as well as "informed" voters). From an economic 17. Adjustments are made not only for inflation, as has been recent practice, but also for total population growth in Australia. 18. During the period of revenue sharing (1975-1985), the proceeds from income taxes were ex- periencing high income elasticities through the fiscal drag induced by high inflation rates (and a failure to index taxes consistently). 19. The main payments relate to institutions of higher education, nongovernment schools, and local government general-purpose payments. 60 Macroeconomic Management and Fiscal Decentralization point of view, however, the budgetary effect of specific-purpose grants is questionable. Specific-purpose grants for recurrent purposes may not differ much from general pur- pose grants to the extent that they release resources in the tied area for general pur- poses. This is true whenever the specific function would have been financed at the state level anyway. In this case, the "tied" grant money is "fungible" (that is, it cannot be at- tributed to the public function performed).20 If, on the other hand, the amount of the specific-purpose grant exceeds the amount that would reflect public demand for that service at the state level, and the central government insists on it being spent in the spe- cific area, then waste and inefficiencies would result since the grant does not reflect the regional pattern of demand for the public service. Some specific-purpose grants in Australia are provided with matching requirements and/or they depend on project approval, information requirements, and conditions for operation. To this extent, the Commonwealth is able to impose some standards, and one may even speak of "joint decisionmaking" although not in the institutionalized sense of the German arrangements (see Chapter 6). Moreover, there is evidence that these grants are associated with substantially increased state-local taxation.21 T'his indicates con- forming additional increases in state-local expenditure (or saving) through expenditure requirements and relative price effects. In any case, the specific nature of the grants af- fects the horizontal regional distribution of these resources, subject however to the Grants Commission taking grant differentials for most of the recurrent specific purpose grants into account in assessing general revenue grant relativities. Over the past twenty years the percentage of specific purpose grants increased dra- matically, from about one-quarter in the early seventies to more than one-half today. This occurred in two leaps during the second half of the seventies and over the past five years. It is very much resented by the states as Commonwealth intrusion into their sov- ereignties. The states have typically accepted this form of grant money since they real- ized the high degree of discretion remaining through their "fungibility." Recently, how- ever, the major growth in tied grants has been in areas where the states' own expendi- tures are limited and so free up very little state money. The Council of Australian Gov- ernments has recently established a working group whose role is to improve the effi- ciency of specific purpose payments as well as to reduce their interference with state re- sponsibilities. Equalization and the Role of the Commonwealth Grants Commnission General Remarks The horizontal allocation of government resources among the Australian states depends on the distribution of the states' own tax revenue; on the distribution of specific purpose 20. The 1994-95 Budget Paper No. 3 recognizes that "for some of the large state expenditure items that are funded through SPPs (for example, hospital funding grants and government schools), a relatively large percentage of these funds would be directed to the same purpose re- gardless of the form of the funding" (p. 32). 21. The propensities for taxes to increase in response to this type of grant money were about 35 percent for grants for recurrent purposes and nearly two-thirds for capital grants (Spahn 1977, pp. 140-42). Intergovernmental Fiscal Relations in Australia 61 payments; on the distribution of the general revenue grants (respectively, tax-sharing funds); and on the distribution of loan money. The distribution of the states' loan monies is centrally controlled in Australia through the Loan Council, discussed later in this chapter. As to the distribution of own revenue, this may vary across regions because of an inequitable dispersal of tax bases and because of variations in tax rates. For reasons of equalization, own revenue of states is standardized, however. Equalization formulae are always based on a standard tax effort that renders explicit equalization independent from regional tax policy and strategies. Own revenue is, of course, unconditional as are the means obtained from general revenue grants. These grants are apportioned on the rec- ommendation of the Commonwealth Grants Commission which uses a sophisticated model that is described below.22 Since the distribution formula chosen for allocating these funds leads to varying weights to be attached to different states, a gradual change in the relative provision of state revenue may result. The distribution of specific purpose payments hinges on several factors, including the historical distribution, which was often arbitrary or reflected state expenditure policies and the Commonwealth's policy priorities (for example, road construction, university education). Specific purpose payments are related to specific programs but are not al- ways granted on the basis of relative economic advantage or needs. The Grants Commis- sion should take most of these grants into account when assessing the transfers needed to achieve horizontal fiscal balance. The Commonwealth Grants Commission's methodol- ogy takes four different approaches to dealing with specific purpose payments: - Inclusion is used for those grants that are considered to meet in part state expen- ditures needs (for example, government schools). In these cases, special purpose funding is a full substitute for general revenue assistance. * Absorption is a variant of the inclusion method. Specific purpose payments are added to the pool of financial assistance grants, and per capita relativities recom- mended by the Commission are determined with regard to that pool (for example, for unquarantined hospital funding grants). * Deduction is used when specific purpose grants are considered to finance expen- diture in addition to what the states would otherwise have undertaken. In these cases, only the state-funded portion of expenditure is included in the Commis- sion's assessment for general revenue grants (for example, a number of health grants). * Exclusion is used where the Commonwealth has largely accepted financial respon- sibility (most grants "through" the states). In this case, all expenditure in the par- ticular area is excluded from the assessment. Intergovernmental distribution of funds in Australia may be typified as "asymmetrical vertical grants"-except for states' own revenue. In designing these grants, relative costs of providing standard services at the state level, relative specific purpose grant differentials, and relative revenue-raising capacities are all taken into ac- count. These relativities are assessed by the Commonwealth Grants Commission, 22. General revenue grants are the main instrument for achieving horizontal fiscal balance through the formula designed by the Grants Commission. One major specific grant (for hospital funding) is also distributed on the basis of the Commission's advice. 62 Macroeconomic Management and Fiscal Decentralization following detailed submissions from and exhaustive discussions with the states and the Commonwealth Treasury. Relative per capita revenue capacities of the Australian states are depicted in Figure 3-6. Compared with most other countries, notably China, the regional variations of fiscal capacity seem to be small. The "richest" states (Western Australia and New South Wales) exceed the "poorest" state (Tasmania) by only about 3 percent. The standard deviation for fiscal capacity of all states is only 11 percent. Apart from standard revenue capacities, the methodology also examines relative standardized expenditure needs for all states. The problems of defining standard needs are further discussed later in this chapter. It is obvious that standardization requires specification of a basket of state functions, which may vary across regions in accordance with different locational criteria-whether a service must be provided extensively over vast geographical areas or intensively in urban agglomerations; whether it must be pro- vided at different quality levels among regions (for example, basic versus highly spe- cialized medical service), or at different quantitative levels in accordance with differ- ences in local demand patterns (for example, a different proportion of school-aged chil- dren). Furthermore, the differences in relative costs must be taken into account (for ex- ample, different wage levels, transportation costs). Figure 3-6. Relative Revenue Capacities of Australian States and Territories, Average 1988-89 to 1992-93 Australian Capital Territory *_ Northern Territory Tasmania South Australia * Western Australia _ Queensland * _ Victoria New South Wales _ 0 20 40 60 80 100 120 Percent Source: Rye and Searl (1994). Intergovernmental Fiscal Relations in Australia 63 As to the relative costs of public services, the discrepancies in the Australian federa- tion are significant (see Figure 3-7). The cost of providing services in the Northern Ter- ritory, for instance, is almost three times as high as in Victoria. However, the Northern Territory is clearly an exception, and it could be dealt with separately without the need to consider relative cost elements for all states. The standard deviation for the relative cost of service is only 7 percent for all states if the Northern Territory is excluded. The Commonwealth Grants Commission The Commonwealth Grants Commission was set up in 1933 and charged with the task "of inquiring into and reporting on applications by the states for financial assistance under Section 96 of the Constitution" (Mathews and Jay 1972, p. 154). The Commission defines its own role as "an independent, impartial and authoritative arbiter in relation to dis- tributional aspects of fiscal federalism in Australia" (Commonwealth Grants Commission 1983, p. 160). Over the years the Commission developed and refined criteria for horizontal fiscal equalization. The guiding principle for supplying special grants was initially that of re- gional "financial needs." A few points explaining the history of horizontal equalization in Australia are worth mentioning. For many years, the system was restricted to "claimant" states, the needs of which were compared with the two "standard" states (New South Wales and Victoria). The horizontal review was hence only partial. With some states ceasing to claim special Figure 3-7. Relative Costs of Public Services in Australian States and Territories, Average 1988-89 to 1992-93 Australian Capital Territory Northern Territory I Out of range 273percem Tasmania South Australia Western Australia l Queensland Victoria New South Wales I I I I l __ 0 20 40 60 80 100 120 Percent Source: Rye and Searl (1994). 64 Macroeconomic Management and Fiscal Decentralization grants the scope for equalization through the Grants Commission was further reduced. The special grants recommended by the Commission (its recommendations were always accepted by the Commonwealth) were relatively small because they reflected the distri- bution of the financial assistance grants and because the financially weaker states were also the states with the smallest populations. The special grants were financed from the Commonwealth budget and did not affect the total amount of the financial assistance grants. Both "revenue needs" and "expenditure needs" were assessed comprehensively for all recurrent revenue/expenditure categories. This was different from the formula govern- ing German Finanzausgleich, for instance, which essentially emphasizes aggregate fiscal capacity, leaving expenditure functions aside.23 However, the needs criteria developed by the Commission had (and have) some more general bearing on the philosophy governing horizontal equalization in Australia-even for general financial assistance. This point was ostensibly made clear when the Commis- sion was asked to review the initial allocation of tax-sharing funds introduced in 1976 through the New Federalism Policy. The Grants Commission's detailed reports on tax sharing led to a great deal of controversy in Australia. Initial reactions were described "as cold, if not actively hostile" (Bird 1986, p. 138). There were two main reasons for the con- troversy. First, a full review of state relativities was to define losers and winners of the proposed redistribution exercise-which always meets political resistance. Second, the report was so specific in many ways-for instance addressing outlays on shark and crocodile protection (Commonwealth Grants Commission (1985, vol. 2, p. 66)-that peo- ple started to wonder what it had to do with the transfer of general revenue to states- although it has to be acknowledged that a systematic assessment would have to include all state revenue and expenditures. Despite some criticism, the Commission has been "the cynosure of federal finance specialists throughout the world" (Bird 1986, p. 142) for over fifty years and its achieve- ments are out of question. What may have been overlooked, however, was the fact that the recommendations initially affected only a small pool of redistributive means-the special grants for the claimant states-and that these grants were far from being compre- hensive equalization payments, both quantitatively and as to their coverage of regions. When the recommendations started addressing a broader range of issues, the obvious emerged: that horizontal equalization is essentially a political process. This process may follow the advice of "independent experts"-yet only as long as it suits existing political interests. Another lesson to be learned from the Grants Commission's tax-sharing exercise is that too specific horizontal redistribution formulae for a substantive amount of general revenue are likely to meet political resistance. In addition, overly specific formulae are likely to jeopardize any move aiming at establishing a greater degree of cohesion. 24 Nevertheless, the basic methodology developed by the Commission with the assis- tance of the states and the Commonwealth Treasury-the equalization model for assess 23. The Australian model also defines relative "tax effort" (own "tax effort" relative to standard "tax effort"). Such a distinction would make little sense in Germany, where legislation governing state taxation is uniform. 24. A similar conclusion may be drawn from the German horizontal Finanzausgleich. Intergovernmental Fiscal Relations in Australia 65 ing general revenue grant relativities-has now been accepted by all parties. The main issue between them is the extent to which the equalization process should be carried-a political question-and the details of revenue and expenditure assessments, which in- volve technical and data considerations. While the political question-the extent of equalization-needs to be determined by governments, it is generally accepted in Aus- tralia that an independent body is more likely to achieve an equitable distribution of grants than is a process of political bargaining by governments of different political com- plexions and different fiscal strengths. The present Australian system developed because of the perceived unfairness of the previous arrangements. Assessing Standard Revenue and Expenditure Needs: The Distribution Formula The distribution model that has been used in Australia since 1981, replacing the special grants arrangements, embraces all states and territories and determines the distribution of the total amount of Commonwealth general revenue grants. It involves the calculation of grant relativities by reference to the relative per capita revenue-raising capacities for all recurrent own-source revenues, the relative per capita expenditure needs (costs of providing standard services) for all recurrent expenditures, and the differential per capita amounts of most recurrent specific purpose grants (those for the expenditures that have been equalized). In effect, each state's or territory's share of the total Commonwealth general revenue grants depends on its standardized deficit, which is the product of its population and its per capita grant relativity. The latter is assessed as its per capita stan- dardized expenditure minus its per capita standardized own-source revenues, plus (or minus) its differential per capita specific purpose grants. Separate assessments are made for nineteen revenue categories, forty-one expenditure categories, and several grant cate- gories. The grant relativities are assessed following a major review every five years. Methodological issues are also reviewed between the major inquiries. Of all federal countries, Australia is best noted for its balanced emphasis on expend i- ture need and revenue means factors in determining state relativities for the distribution of unconditional equalization transfers. Section 13(3) of the States (Personal Income Tax Sharing) Act of 1970 contains general guidelines concerning equalization as follows: ...respective payments to which the states are entitled under this act should enable each state to provide, without imposing taxes and charges at levels appreciably different from the levels of the taxes and charges imposed by other states, government services at stan- dards not appreciably different from the standards of government services provided by the other states. Total equalization pool is determined by 39.87 percent of personal income tax collec- tions for the previous year. The Commonwealth Grants Commission is entrusted with the task of developing state relativities based upon the above principle for use in grant allocation. These relativities are defended in open adversary proceedings by the Com- mission, and a final report is presented to the Commonwealth cabinet for review. The cabinet occasionally revises the recommended relativities based on its own view of rela- tive fiscal needs. Following this review, a final determination is made in the annual Pre- miers' Conference. The Commission's approach to fiscal equalization is summarized here briefly. 66 Macroeconomic Management and Fiscal Decentralization Tax sharing entitlement of state i: Gi = Bj / 2:(Bijf where G is 40 percent of personal income tax collection in the preceding year, Gi is the state's entitlement, and Bi / I B1 is state i's proportion of total entitlement. Assessed grant: (Bi) = Revenue + expenditure needs - assessed needs met by other federal transfers. Revenue: = Pi(R., / P, )O -- Pi (R, /Ps)/(Ys / P - Yj /Pj) where Pi is population of state i, RS/PS is per capita standard tax revenue, Oi is relative revenue capacity of state i, Ys/Ps is per capita standard tax base, and Yi/Pi is per capita tax base of state i. The Commission measures each state's expenditure needs for a service or category of expenditure by calculating the differential cost, for the state whose needs are being assessed, of providing services at a standard level, range, and quality. Thus, per capita expenditure need is per capita differential cost-or per capita standardized expen- diture minus per capita standard expenditure. Standardized expenditure is the amount the state would need to spend to provide a standard level and range of public services and operate at standard efficiency, and standard expenditure is the population weighted average expenditure of all six states. Standard expenditure: = Total of six states' expenditure on category Total population of six states Mathematically, Expenditure need = Pi (EsAPs)Vi = Pi (ES/Ps) [(UiSidi)ei) -1] where Pi is populations of state i, ES/PS is per capita standard expenditure, Vi is expen- diture need of i relative to standard, Ui is eligible population, Si is scale factor, di is dis- persion factor, and ei is social, cultural, and environmental factors. Intergovernmental Fiscal Relations in Australia 67 Expenditure need factors can be categorized as follows: * Scale factors * Population and related factors - Dispersion - Urbanization - Social composition - Age structure Environment factors - Physical - Economic An example: * Expenditure category Health * Subexpenditure General medical services * Units of use Total population Other adjustment factors * Age and sex composition factor based on hospital bed use data. * Index of health status based on standardized mortality rates adjusted by female fertility rates. * Social composition factor based on aboriginal populations. A state's expenditure as measured by these procedures is either positive or negative. Sociodemographic composition, population density, urbanization, and physical envi- ronment factors figure predominantly in assessing differential costs. Table 3-2 provides details on the 1994-95 distribution of these transfers and Table 3-3 using 1991-92 data compares the allocation of transfers under the existing system with alternate distributions using equal per capita and income tax payments criteria. The Australian approach to fiscal equalization is more comprehensive than that found in Canada and Germany. An attempt is made to equalize both the revenue raising capacity and the ability to provide a given level of services. Municipal revenues are, however, not included in revenue equalization. The Australian procedure for the assessment of expen- diture needs has some elements of subjectivity. Determinants of expenditure need are sometimes arrived at using broad judgment rather than hard quantitative analysis. These calculations are done every five years and projected for interim years. The data used for some of the need calculations can be quite dated. Past allocations influence current enti- tlements, and any major change in fiscal position is accommodated with a significant de- lay. The fixed nature of overall allocations negated open-ended commitment. While the Australian Grants Commission's philosophical attachment to a comprehensive system of fiscal equalization is commendable, actual procedures used by the Commission may be open to further refinement. Table 3-2. Commonwealth Payments to the States and Direct to Local Government Authorities, 1994-95 (Estimates) (dollars per capita) New South Western South Australia Capital Northern Payment Wales Victoria Queensland Australia Australia Tasmania Territory Territory Total General revenue assistance 701.0 691.3 863.9 911.6 1,005.8 1,325.7 1,073.7 4,974.8 836.4 General purpose capital assistance 17.7 14.3 8.1 12.9 15.1 12.7 18.6 30.6 14.5 Total general purpose payrnents 718.7 705.5 872.0 924.5 1,020.9 1,338.4 1,092.4 5,005.4 850.9 Specific purpose payments 'to' the states 515.3 512.5 587.3 632.1 801.9 695.4 561.1 1,245.8 586.9 Specific purpose payments 'through' the states 357.3 392.4 346.7 366.1 363.8 364.0 200.9 407.4 363.5 Total specific purpose Payments 908.5 904.9 933.9 998.3 1,165.7 1,059.4 762.0 1,653.2 950.5 Total payment to states and local government authorities (net) 1,530.8 1,526.3 1,739.1 1,829.1 2,038.6 2,113.8 1,796.7 6,269.3 1,702.3 Population, December 1, 1994 (in thousands) 6,079 4,490 3,228 1,711 1,474 474 305 170 17,931 Source: The Treasurer of the Commonwealth of Australia (1994). Intergovernmental Fiscal Relations in Australia 69 Table 3-3. Impact of Horizontal Fiscal Equalization on the Distribution of the Pool of Fi- nancial Assistance Grants and Hospital Funding Grants, 1991-92 ($ million) Distribution Distribution Distribution of the pool State or territory of pooling of the pool on Difference in on the basis Difference in using rela- an equal per distribution of income distribution tivities capita basis (1) - (2) tax paid (1) - (4) (1) (2) (3) (4) (5) New South Wales 5,251 6,002 -751 6,501 -1,250 Victoria 3,709 4,433 -724 4,589 -880 Queensland 3,325 3,187 138 2,641 684 Western Australia 1,830 1,690 140 1,630 200 South Australia 1,772 1,456 317 1,316 457 Tasmania 709 468 241 402 308 Australia Capital Territory 270 301 -31 465 -195 Northern Territory 837 168 669 159 678 Total 17,704 17,704 " 17,704 Source: The Treasurer of the Commonwealth of Australia (1994). Constraints on Deficit Financing A full discussion of revenue instruments would have to include loan monies to be raised at vertical levels of government as well. Loan money-if not controlled-may help soften budget constraints at all levels of government; loan control is thus most delicate for mul- tilayer government finance. Moreover, government borrowing is intertwined with monetary policy and macroeconomic demand management (through its impact on capi- tal markets). Effective control of public sector borrowing is hence a major concern for all federations. State borrowing (general government as well as state business enter- prises) is determined by the Loan Council, which determines not only the total amount, but also the allocation among Australian states. The Loan Council was set up by the Fi- nancial Agreement of 1927. It provides essentially what may be called a "joint- decisionmaking" machinery, yet the Loan Council's decisions throughout its history have predominantly reflected the interests of the Commonwealth. 25 In the early 1980s, it appeared that the Australian Loan Council was relaxing control. The 1980s brought about an increase in "off-program-borrowing" activities at all levels of government. Furthermore, the states had increasingly developed devices to circumvent Loan Council control. State budget constraints had thus been softened quite considera- bly. And the constraints on regional loan financing appeared to be softer than it seems at 25. The federal government must convince only two states in order to have a motion passed, but in any case it is able to control the Council's decisions because of its dominating fiscal strength and grant powers. 70 Macroeconomic Management and Fiscal Decentralization first sight. Nevertheless, the Commonwealth retains control over the total borrowing program, and it has introduced a new global limits formula, which was used restric- tively. 26 Recently, the Loan Council has reformed its proceedings. Under the new arrange- ments, each jurisdiction nominates an amount to be allocated under the proceedings of the Loan Council in accordance with "its estimated general government deficit/surplus ..., public trading enterprise sector ... net financing requirement and certain memoran- dum items. These nominations are considered by Loan Council having regard to each jurisdiction's fiscal position and reasonable infrastructure requirements, as well as to the macroeconomic implications of the aggregate figure" (1994-95 Budget Paper No. 2, p. 42). Summary and Outlook The lack of flexible tax resources under state control is often cited as a severe problem for multilevel government finance in Australia. Yet it is questionable whether the states have really been "at the mercy" of the Commonwealth. A number of factors suggest that the states have been able to preserve-and even extend-their responsibilities: • General revenue grants based on quasiconstitutional rules and institutional iner- tia ("no one should lose a benefit once received") provide a secure revenue source covering half of state resources on average. This is prone to foster a stable envi- ronment for state policy decisions and for long-term planning of public sector developments. In recent years, however, general revenue grants have been cut substantially in real terms. * If the states had continued to impose their own income taxes, they would proba- bly have received less revenue than they obtained through general revenue as- sistance. This was certainly true for the postwar period until the mid-sixties (Mathews and Jay 1972, p. 317); and it is probably true for the initial tax-sharing phase in the 1970s. The fact that the Commonwealth was able on several occa- sions to reverse these trends is not an argument for Commonwealth dominance; it is rather an indication that intergovernmental financial control may work even without the often cited tax/outlay link. * Specific purpose payments have almost certainly released the states' own funds in certain policy areas that could be used for general purposes. This is true in spite of matching requirements in some instances. Matching requirements sometimes may have introduced a bias in the structure of state government outlays-creating, for instance, an incentive to increase expenditure on university education relative to primary and secondary schooling. Deploring this type of inefficiency is justified; but it calls for a revision of the set of matching requirements associated with spe- cific purpose grants-not for a revision of the whole system of federal finance, and this has generally been achieved in Australia. * Special grants and, later, the general revenue equalization arrangements have helped to ease the budgetary constraints of the poorer states. 26. This was facilitated politically because high real borrowing costs have reduced the desire of states to borrow. Intergovernmental Fiscal Relations in Australia 71 * The states have recently started to exploit more heavily some own resources - notably user charges, resource taxation, and business franchise tax revenue. * The consistent reduction in the Loan Council's role in controlling state borrowing has further contributed to easing budget constraints at state levels. However, important flaws in the Australian federal arrangements remain. In addition to the vertical fiscal imbalance and the lopsided dependency of the fiscal system on in- come taxes, recent reform proposals have focused on how the special grants system has interfered with equity and the definition of relative needs. Proposals also were made to strengthen general revenue grants at the expense of specific purpose payments. A Premiers' Conference in 1992 acknowledged the need for further reforms, and a working party was set up to examine the adequacy of the current fiscal system. The working party found the system of equalization to be highly complex by international standards, and it considered simplification of the scheme (for instance, the distribution of a part of the grant on a straightforward per capita formula or the reduction of the for- mula to fiscal capacity equalization only). The working party also looked into some of the possible implications for economic efficiency arising from the practical operation of the methodology. Whatever the prospects for reform in Australia, the fiscal federal machinery is an in- teresting object of study for the scholar of federalism. Despite the highly centralized revenue functions and the importance of Commonwealth interference in state affairs through its specific purpose payments, the significance of the lower tier of government for the provision of public goods cannot be denied. There is no doubt that the states are resilient and play an important role in Australia. The Commonwealth cannot ignore the states and cannot keep cutting back grants to them. It is forced to reach broad agreements with the states (Fletcher and Walsh 1991) in order to preserve political stability. An out- side observer of Australian federalism may well come to the conclusion that" overall, the pattern of Australian fiscal federalism seems to be fairly stable" (United States Advisory Commission 1981, p. 78). Despite this positive interpretation of the effects of multilevel finance on state sovereignty in Australia, a number of writers continue to underline the inefficiency of these provisions-notably by stressing the missing tax/expenditure link. The importance of this link may be debatable-in particular for smaller vertical fiscal im- balances (Spahn 1993). The question is whether the vertical inbalance in Australia is se- vere enough to warrant such concern. It should not be overlooked that the political process may generate a corrective re- sponse-a response different from the traditional median voter model, yet one that leads to similar results. The Commonwealth government may thus come under direct pressure from its own electorate to constrain the overall level of taxation; the Commonwealth may then pass this pressure on to the states by demanding a revision of the revenue-sharing formula. Indeed, such revisions have been taking place in Australia; the mechanism may in fact be fully appropriate for a polity in which "neither the Senate nor the Cabinet pro- vides adequate regional representation at the national level, because they are primarily dominated by party rather than by regional considerations" (Bird 1986, p. 130). Federal financial arrangements in Australia have preserved and secured state independence. The lower tier of government is far from being at the mercy of the Commonwealth. 72 Macroeconomic Management and Fiscal Decentralization References Bird, Richard. 1986. Federal Finance in Comparative Perspective. Toronto: Canadian Tax Foundation. Commonwealth of Australia, The Treasurer. 1994. Commonwealth Financial Relations with Other Lev- els of Government, 1994-95. 1994-95 Budget Papers Nos. 1, 3. Canberra: Australian Government Publishing Service. Commonwealth Grants Commission. 1983. Equality in Diversity: Fifty Years of the Commonwealth Grants Commission. Canberra: Australian Government Publishing Service. . 1985. Report on Tax Sharing Relativities, vols. 1 and 2. Canberra: Australian Government Publishing Service. Fletcher, Christine, and Cliff Walsh. 1991. "Intergovernmental Relations in Australia: A Manageri- alist Revolution?" Canberra: Australian National University, Federalism Research Center. Mimeo. - . 1993. State Taxation in Australia: Lessons for South Africa. Centre for Human Rights, Univer- sity of Pretoria, South Africa. November. Galligan, Brian, and Cliff Walsh. 1990. "Australian Federalism: Developments and Prospects." Pub- lius: The Journal of Federalism, vol. 20, no. 4. Groenewegen, Peter. 1991. "Financing the States in the Australian Federation." In Financing Federal and State Governments, edited by the OECD, Centre for Cooperation with European Economies in Transition, Committee on Fiscal Affairs, Paris. Unpublished. Hunter, J. S. H. 1977. Federalism and Fiscal Balance. Canberra: Australian National University, Centre for Research on Federal Fiscal Relations. Mathews, Russell L. 1982. The Commonwealth-State Contract. Reprint No. 46. Canberra: Australian National University, Centre for Research on Federal Fiscal Relations. Mathews, Russell L., and W. R. C. Jay. 1972. Federal Finance: Intergovernmental Financial Relations in Australia since Federation. Melbourne. Rye, C. R., and R. J. Searl. 1994. "The Fiscal Transfer System in Australia." Paper presented at the International Conference on Fiscal Transfer Systems in Quingdao, China. July. Spahn, Paul Bernd. 1977. "Federal Grant Policy and State-local Taxation." In State and Local Taxa- tion, edited by Russell L. Mathews, pp. 125-142. Canberra: Australian National University Press. . 1993. The Community Budget for an Economic and Monetary Union. Houndsmill, Basingstoke, Hampshire: Macmillan. United States Advisory Commission on Intergovernmental Relations. 1981. Studies in Comparative Federalism: Australia. Washington, D. C. Walsh, Cliff, ed. 1990. "State Taxation and Vertical Fiscal Imbalance: The Radical Reform Options." In Issues in State Taxation, pp. 1-22. Canberra: Australian National University, Centre for Re- search on Federal Financial Relations. . 1991. Reform of Commonwealth-State Relations 'No Representation Without Taxation'. Discus- sion Paper No. 2. Canberra: Australian National University, Federalism Research Centre. August. 4 Comparative Federalism: Trends and Issues in the United States, China, and Russia Roy Bahl The United States, China, and Russia share some features: large populations, large land areas, and historical pressures for decentralization in governance. The United States and Russia are formally federal countries, and China in many ways behaves as though it is a fiscal federalism. 1 In the past few years, all three countries have undergone substantial changes in their intergovernmental systems, and central governments in all three coun- tries are under fire from lower level governments. Many observers would see intergovernmental fiscal relations in the United States vs. Russia and China, as being so different as to be incomparable. Likewise, it might be ar- gued that the same theoretical model cannot explain changes in central-local relations in a capitalist transition and socialist economy. The objective of this paper is to compare the systems of intergovernmental fiscal relations in these three countries and to ask whether the theory of fiscal federalism can explain the differences and similarities in the struc- tures of tax and expenditure assignment, and the recent trends that have occurred. The next section deals with the theoretical model, that is, the reasons why one would expect more or less decentralization in government finance from countries such as these. I then turn to a description of each of the fiscal systems, and recent trends in each country, and finally to some attempt at explaining the impacts and potential reasons for these changes. Theory and Fiscal Decentralization Fiscal decentralization has to do with the amount of fiscal autonomy and responsibility given to subnational governments. It is a subject on the policy agenda in many develop- ing, transition, and industrialized countries. There always has been a cry for more de- centralization of government, resulting from a combination of people wanting to get The author is a professor of economics and the director of the Policy Research Center at Georgia State University. A version of this paper was originally presented at a conference on decentraliza- tion in Washington, DC, in August 1994 sponsored by the National Academy of Public Admini- stration (Washington, D. C.) and the National Institute of Research Advancement (Tokyo). 1. In fact, China is a unitary state. It is part of this comparison because its fiscal system shares many of the features of fiscal systems in federal countries. 73 74 Macroeconomic Management and Fiscal Decentralization more involved in the process of government and the inability of central governments to "get the job done." Many countries around the world have moved toward more decentralized structures in the past two decades, but quite different paths have been taken. Many other countries have remained highly centralized and have been loathe to relinquish any taxing and spending powers to lower level governments. China and Russia are examples of coun- tries that have been decentralizing the flow of resources, that is, subnational govern- ments have been claiming an increasing share. Current policy, however, seems to be in the direction of finding a way to reclaim a larger share for the center. In the United States, policy and practice have been in the direction of cutting the dependence of the state and local sectors on the federal government. Unlike the United States, neither China nor Rus- sia has granted taxing autonomy to their subnational units of government. Empirical research has shown that countries that have given greater fiscal powers to their state and local units tend to be higher income, larger in population and land area, and to have a more heterogeneous mix in their populations. It also has shown that coun- tries at war or threatened by war tend to be more centralized (Bahl and Nath 1986). By these criteria, Russia, China, and the United States are all countries that would seem to be candidates for a strong degree of fiscal decentralization. Another way to examine the determinants of fiscal decentralization is to consider the a priori reasons why a country might choose decentralization of its fiscal structure, that is, the theory of fiscal federalism (Oates 1972). These reasons might be broken down, as be- low, into the advantages and disadvantages of fiscal decentralization.2 Advantages of Decentralization Economists would invoke efficiency criteria in arguing for smaller local government, that is, in arguing for fiscal decentralization. If preferences for public services do differ across subgroups of the population, and if externalities are not present, then national welfare is maximized if local communities vote their preferences and provide the level and mix of public services that they want. Noneconomists might take the same view but couch it in different terms: getting government closer to the people will lead to more participation in government, will likely provide an outcome that is closest to the preferences of the me- dian voter, and will allow the political process to guarantee a more efficient operation of local government. Either way, the results should be the same: * The mix of services provided will match the demands of the local population. * Government officials will become more accountable to voters for the quality of services they provide. * Local populations will be more willing to pay for public services, since their preferences will be honored. A second argument for fiscal decentralization, not often made, is that it can enhance revenue mobilization. Some taxes are suited to local government in that their assessment and collection require familiarity with the local economy and population, and because they are perceived as quasi-benefit charges that finance local area services. The property 2. The arguments concerning the advantages of decentralization are developed more fully in Bahl and Linn (1992, 385-427). Comparative Federalism: Trends and Issues in the United States, China, and Russia 75 tax and other land based taxes are usually thought of as local government taxes. It is also true that central government value added and income taxes often do not reach far down in terms of the amount of the tax base that they pick up. Typically, small firms and work- ers outside the larger formal sector firms are left out of the tax net for administrative rea- sons. Local governments, it is argued, might be able to capture this untapped fiscal ca- pacity because of their greater familiarity with the local tax base. The United States, Rus- sia, and China all have capitalized on this advantage of fiscal decentralization but in dif- ferent ways. Only the United States among this group, however, makes extensive use of property taxation. Are these arguments really valid? Can local governments actually respond to citizens' preferences for many or few local services, or to a willingness to pay more tax to receive local services? In fact, the efficiency case for fiscal decentralization is much stronger in industrial than in developing or transition countries. Consider first the notion that mov- ing service provision closer to the people can lead to gains in the welfare of consumer- voters. Because the theory of fiscal assignment was developed in industrialized countries, it was heavily influenced by democratic processes of budgetmaking (for example, the median voter theories of public expenditure determination). In this model, the level of tax effort and the expenditure mix in local areas are responsive to changes in relative prices and income, and the potential losses in efficiency caused by interference from a higher level of government can be substantial (as can the potential efficiency gains from the greater fiscal autonomy of local government). Although the model is based on a number of questionable assumptions, empirical research has shown that the behavior of U.S. state and local governments more or less squares with it (Bocherdering and Deacon 1972). The model does not so easily fit transition countries such as China and Russia, how- ever, and the efficiency gains from decentralization therefore may not be so great. This is partly because voter preferences are not as readily translated into budget outcomes in transition countries as in industrial countries. Local councils are often not elected, chief officials are often not locally appointed, and adjustments in the allocation of local re- sources are often severely constrained by central government controls. These controls include approval of the budget, central appointment of chief local government officers, central government regulations of tax administration, mandates as to salary levels of lo- cal government employees, and the general absence of a mechanism by which local vot- ers can reveal their preferences for a larger or smaller government. In this setting-where the devolution of revenue authority and expenditure responsibility is not accompanied by a relaxation of central government control over local fiscal decisionmaking-there is less to be gained from decentralization of taxes and expenditure than would be the case in industrial countries. Given this state of affairs, a transition country that could derive maximum gains from a more decentralized local government structure would have the following characteris- tics: (a) enough skilled labor, access to materials, and capital to expand public service de- livery when desired, (b) an efficient tax administration, (c) taxing power able to capture significant portions of community income increments, (d) an income-elastic demand for public services, (e) popularly elected local officials, and (f) some local discretion in shap- ing the budget and setting the tax rate. This list suggests that the setting for decentraliza- tion is clearly present in the United States, but that the case for efficiency gains in China and Russia is much less easy to make. 76 Macroeconomic Management and Fiscal Decentralization Advantages of Centralization The arguments for fiscal centralization are stronger in transition countries than in indus- trial countries. As recent history has shown, stabilization is especially important in such countries. This argues for central government control of the main tax and borrowing in- struments. It has also become a concern in the United States as the federal government tries to cope with a substantial budget deficit. In transition countries that are undergoing privatization and building a public and industrial infrastructure, the need for a coherent growth policy is also an argument for fiscal centralization, because investment capital is scarce and must be controlled by the central government to maximize returns. If local governments are given access to major tax bases, they may compete with the central government and therefore limit the amount available for the central tax. As a corollary, centralization allows the national government to allocate fiscal resources to goods and services with national benefits, whereas local autonomy would inevitably lead to greater expenditures on those services that have more local benefits. The struggle over financing health care in the United States is a good example of this issue. Several arguments for income distribution also support fiscal centralization. The most important is that regional (and rural-urban) disparities in income and wealth may be ac- centuated by fiscal decentralization because wealthier urban governments will benefit most from greater taxing powers. Centralization allows the national government more discretion in shaping regional differences in levels of public service and taxation, which is an especially important consideration for governments that intend to use tax and sub- sidy policy to shape the spatial distribution of economic development. China, Russia, and the United States are all characterized by significant fiscal disparities. China and Russia in particular have faced difficult choices as regards equalization. China was forced to choose between funneling more resources to the lower income provinces or to leave the retained revenues higher in the coastal growth provinces. Russia has faced the difficult decision of choosing among equalization, central government fiscal solvency, and ap- peasing the potential breakaway provinces. In both cases, the central governments re- tained control over the fiscal resources and were in a position to make the decision. The U.S. federal government would have been in much less of a position to affect a regional redistribution of resources. The U.S. Federal System The system of fiscal federalism in the United States gives substantial autonomy to subna- tional governments on both the tax and expenditure sides of the budget and relies heav- ily on these governments for revenue mobilization and the provision of social and infra- structure services. The subnational government sector is composed of fifty states and, within each, numerous local governments (counties, cities, towns, school districts, and other special districts). Each state may decide on the powers and responsibilities that it will give to its local governments. Tax and Expenditure Structure Subnational governments in the United States (states and local governments) spend an amount equivalent to 13 percent of GNP (1993), less than the federal government share Comparative Federalism: Trends and Issues in the United States, China, and Russia 77 when defense and international assistance are excluded (18.4 percent). State and local governments raised about 41 percent of all tax revenues in 1993. For every one federal government employee, there are 5.5 state and local government employees (1993). Clearly, the state and local government sector is an important part of the American econ- omy. The federal government relies almost exclusively on income taxation, with almost two-thirds of tax revenues coming from the individual income tax. None of the revenue is directly shared with state and local governments, and collections are made by a federal agency, the Internal Revenue Service. Many subnational governments have chosen to adopt the same base as the federal income tax for their own individual and corporate taxes, and there is a system of information sharing between the Internal Revenue Service and the states' tax collection agencies. State and local governments have chosen a wide variety of tax structures and depend, to varying degrees, on taxes on individuals and businesses (see Table 4-1). There is a relatively even split among income, consumption, wealth taxes, user charges, and inter- governmental assistance. The U.S. fiscal system is relatively balanced. State governments tend to rely heavily on income and sales taxation, though six states impose no income tax, and four do not impose a retail sales tax-the preferred form of indirect tax in the United States. Local governments tend to rely heavily on the property tax, user charges, and state government grants. Table 4-1. The United States: State and Local Government Fiscal Structure, 1991 Category State and local governments Local governments Total revenue (millions) $902,177 $541,791 Income tax 14.6% 2.2% Sales tax 13.9 4.1 Property 18.6 29.8 Federal (and state) grants 17.1 37.3 User charges 13.9 14.4 Other 10.8 8.7 Total expenditures (millions) $908,470 $542,045 Education 32.7% 39.9% Highways 7.1 4.8 Health 8.9 7.9 Police and fire 5.2 Public welfare 14.0 5.0 Interest on debt 5.7 5.3 Other 31.2 31.1 Total 99.6% 99.2% Note: In 1993, state and local government expenditures as a share of GNP = 13.9%; state and local government taxes as a percentage of personal income = 11.8%; and federal grants as a percentage of state and local government revenues = 21.0%. Source: U.S. Department of Commerce (1994). 78 Macroeconomic Management and Fiscal Decentralization Fiscal Autonomy State and local governments have substantial fiscal autonomy. They are free to choose their own tax structures, so long as they do not violate the federal Constitution. The pri- mary issues of concem here are that states not restrict interstate commerce and that they not discriminate against any subgroup of the population. Likewise, they may deliver ex- penditures in any manner they desire, except that the federal Constitution requires that they provide citizens "equal protection." In recent years, the courts have been hearing challenges against the method of financing public schools in U.S. states. In some cases, they have ordered states to change the method of financing to provide more equal serv- ices to all students in the state. Tax rates, tax bases, and user charges may be set by states without approval by the federal governrnent. States may borrow from whatever source they choose, subject only to the limitations on general financial practices (for example, disclosure) laid down by federal agencies. Finally, states have independent tax collection agencies and are only loosely tied to the federal government (information sharing). On the expenditure side, the states may select whatever number of employees, and whatever compensation rates for those employees, it chooses. Likewise, it is free to de- liver services at whatever level it chooses. There are, however, some restrictions on ex- penditure autonomy. States have restricted themselves, through state constitutions, to disallow deficit financing. In other words, states must balance their recurrent budgets every year. Borrowing is only for capital financing purposes. Moreover, the federal gov- emment has laid down some mandates that restrict state and local government expen- diture decisions: environmental and health regulations; conditionality on the receipt of federal grants, particularly for health and welfare services, and so on. Likewise, state governments have mandated certain actions by their local governments. Various analysts have estimated the costs of such mandates at substantial levels. The fiscal autonomy of local governments is more limited and is determined by each state. The taxes that local govemments may levy are prescribed, and though there is us u- ally some freedom in choosing tax rates, the state usually provides for a maximum levy. The rules under which local governments must seek voter approval for fiscal actions (tax rates, annexations, new borrowing, and so on) are carefully prescribed. The fiscal impor- tance of local governments in the United States varies widely, from 40 percent of total state and local government spending in Vermont to 68 percent in Nevada. State govern- ments also give grants to local governments, and sometimes share the proceeds of tax revenues on a derivation basis, but there is no single pattern that best describes the prac- tice. Recent Trends Over the past fifteen years, the federal govemment has reduced the level of its explicit and implicit transfers to state and local govemments. This has had important effects on the activities of the state and local government sectors. Three important issues stand out: the decline in the federal subsidy, the pattem of disparities and interstate competition that has accompanied this "fend-for-yourself" federalism, and the offsetting pattern from federal mandates and court cases on school finance. Comparative Federalism: Trends and Issues in the United States, China, and Russia 79 Fiscal Dependence The federal government has taken a series of actions to increase the fiscal independence of state and local governments, that is, to make them more independent. It has reduced the rate of increase of federal assistance, particularly the level of federal grants to state and local governments. It also has eliminated the only program of general assistance to state and local governments (general revenue sharing), and it has all but eliminated the direct flow of aid to local governments. Between the late 1960s and the late 1970s, federal grants increased dramatically in real terms, as a share of the federal budget and as a share of state and local government ex- penditures (see Table 4-2). The 1980s saw a decline in the reliance on federal grants by state and local governments, and a declining emphasis on grants in the federal budget. The real level of grants was lower in 1990 than in 1980. The first three years of this decade saw a resurgence in federal grants to state and local governments, by all indicators, ex- cept that the increase was concentrated in grants that are passed to individuals rather than grants to the governmental units. The federal government also reduced an implicit subsidy that operated through the fed- eral income tax. In the past, all state and local government general purpose taxes were allowed as deductions on the federal individual income tax. This preferential treatment was available to those who itemized deductions (approximately the one-third highest income taxpayers). At a 50 percent marginal income tax rate, this meant that itemizers could shift, at the margin, about half of their taxes on to the federal government. This re- duced the resistance to higher state taxes and, therefore, acted as a subsidy to state and local governments. The 1986 federal tax reform eliminated the deductibility provision for general sales taxes3 and reduced the top marginal income tax rate to 33 percent, thereby reducing the value of the remaining deductibility provision for income and property taxes.4 Table 4-2. The United States: Federal Grants to State and Local Governments, Selected Years 1970-1993 Percentage of Percentage of Grantsfor payments to Real amount state and local ex- federal government individuals Year (billions) penditures expenditures (% of total) 1970 73.6 19.0 11.5 36.3 1975 105.4 22.6 13.7 33.7 1980 127.6 25.8 15.2 35.7 1985 113.0 20.9 10.8 46.6 1990 119.7 19.4 10.7 57.0 1993 163.2 23.0 13.6 62.0 Source: U.S. Department of Commerce, Survey of Current Business, National Income and Product Accounts, Tables 3.2 and 3.3, selected years; and Advisory Commission on Intergovernmental Re- lations (1993). 3. The gasoline tax deduction had been disallowed earlier. 4. The Clinton Tax Reform of 1993 increased the top marginal tax rate to 39 percent thereby re- storing some of the value to deductibility. 80 Macroeconomic Management and Fiscal Decentralization The net effect of these changes is to increase the price that a state resident must pay to buy a dollar's worth of public expenditures. This means that states must be more hesitant to propose budget increases, and there is a kind of dampening effect introduced into state and local government fiscal decisionmaking. On the other hand, state government officials must be more accountable to their constituencies. Even so, the state and local government tax share of personal income rose from $19.90 per $100 of personal income in 1986 to $22.40 in 1991. Disparities and Interstate Competition The United States has always been characterized by wide fiscal disparities. There has been no major trend of lessening these disparities, even though personal income dispari- ties among the states are converging. Per capita expenditures vary from highs of $5,482 and $5,064 in New York and Wyoming to lows of $2,664 in Missouri and $2,440 in Ar- kansas, a range equivalent to approximately 77 percent of the mean. The federal grant system is not systematically related to the level of personal income and therefore does not equalize these fiscal disparities. Interstate competition for industry is a significant part of state tax policy in the United States. To some extent, this competition is regulated by interstate compact, under which states share the proceeds of corporate income taxes according to a three-factor formula based on the share of a multistate company's activities in each state. States compete ag- gressively with one another to attract industry by giving preferential tax treatment- most notably income and property taxes-and direct subsidies for infrastructure devel- opment and worker training. The Russian Federal System The Russian Federation is a three-tiered federal state, consisting of ninety-one provinces or states directly subordinate to the federal government. With varying degrees of ad- ministrative autonomy, the ninety-one states "directly subordinate" to the federal gov- ernment comprise (1) the oblasts, okrugs, and krais, (2) metropolitan cities with "oblast status" (Moscow and St. Petersburg), (3) republics, which until mid-1992 were called "autonomous republics," (4) autonomous regions, and (5) national regions. Some repub- lics have their own governments (Soviets) with defined degrees of autonomy, and some have declared independence (although these declarations have not been recognized by the Supreme Soviet of the Russian Federation or by any other country). Within each oblast are local governments, cities, and smaller units called rayons.5 Tax and Expenditure Structure Most subnational government revenues are derived from four shared taxes whose rates and bases are set by the federal government. The system is not well balanced, and about two-thirds of all tax revenues come from two taxes-the enterprise income tax and the value added tax. All taxes are shared on a derivation basis (that is, by point of collection), 5. In this section on Russia, I draw heavily from Bahl and Wallich (1995); Bahl and Wallace (1994); and Bahl (1994b, 129-180). Comparative Federalism: Trends and Issues in the United States, China, and Russia 81 and the sharing rates are set by the government and the Parliament. The basic sharing system is summarized in Table 4-3. The individual income tax is fully retained by local governments (rayons) on the basis of place of employment. Coverage of the self-employed and small firms is limited, and revenue yield is sensitive to the present high rates of inflation and to the resulting changes in the tax brackets, personal exemption levels, and wage levels. The tax is largely administered by the enterprises, which collect it on a withholding basis and keep all em- ployee records. Most workers are subject to a 12 percent rate. The enterprise income tax is levied at a 35 percent rate on company profits. Oblast level governments may retain an amount equivalent to that derived from a 22 percent tax, and they must turn the remainder over to the federal government. Oblasts may in- crease the tax rate to 38 percent or reduce it to 33 percent, but the federal claim must re- main at 13 points. Revenues from the enterprise income tax are sensitive to federal deci- sions about input and output prices, capital allocations (past and present), and redefini- tions of the tax base. In 1992, about two-thirds of the enterprise income tax was retained by local govermnents (Table 4-4). Table 4-3. Russia: Revenue Sharing by Revenue Source Amount allocated to subnational Method of distribution Revenue source governments among oblasts Comments Individual income tax 100 percent Derivation, by place Fully allocated to the of employment rayon level Company income tax Tax rate is 32 percent; Derivation Oblast may reduce rate 22 percent rate belongs by 2 percent to the oblast Value added tax 20 to 50 percent of Derivation, with an ad collections, depending hoc determination of on the oblast the percentage for each oblast Excise on vodka 50 percent of collections Derivation Other excises 100 percent of collec- Derivation Excise on motor vehicles tions to center Subventions and Ad hoc determination Ad hoc determination Distribution largely transfers to autonomous based on "approved" regions deficits, and special projects Source: Bahl and Wallace (1994). 82 Macroeconomic Management and Fiscal Decentralization Table 4-4. Russia: Revenue Structure and Revenue Sharing in 1992 (in billions of rubles) Percentage re- tained by subna- Budgeted average Subnational tional govern- retention rate for Tax/revenue Total collections amount ments 1993 Individual income tax 31.3 431.3 100.0 100.0 Enterprise income tax 1,566.8 920.9 58.8 66.7 Value added tax 1,998.9 498.7 24.9 30.4 Excises 211.5 110.8 52.3 61.5 Foreign trade taxes 467.4 8.0 1.7 n.a. All other taxes 573.8 374.3 65.2 n.a. Total revenue 5,249.7 2,344.0 44.7 43.4a a. This total refers only to the income and value added taxes and excises. Source: Bahl and Wallace (1994). The value added tax is shared among subnational governments on a derivation basis, but the percentage of collections that may be retained varies by oblast. These retention rates have been established in an ad hoc way and have been changed frequently in the past two years. In 1993, most oblasts retained half of VAT collections, but in 1992, only about 30 percent was locally retained. Until November 1991, tax administration was the responsibility of the Ministry of Fi- nance, and the State Tax Service (STS) was a department within that ministry. Highly decentralized oblast and rayon (local) offices were supervised by their respective oblast and rayon level Departments of Finance and by finance officers of the federal ministry. No single authority (such as a "revenue commissioner") was in charge of all tax admini- stration activities. This "dual leadership" of the tax officers was a source of major conflict of interest, often resolved by giving precedence to interests of the lower level. Since 1991, the STS has been an autonomous agency with ministerial ranking, and it is now in charge of administering all taxes in the Russian Federation. Its organizational structure has three levels: central, oblast, and rayon. The assignment of public expenditure responsibility to different levels of government in Russia is based on the principle of geographical dimension of benefits (Martinez- Vasquez 1994). This economic efficiency rule, rather than an equalization rule, was in- herited from the former Soviet Union. Public service activities whose "benefit area" is the entire nation are provided by the central government. Those with a regional dimension (for example, universities and tertiary and psychiatric hospitals) are provided by the oblast and autonomous republic level of government; and those with a local dimension (elementary schools and parks) are provided by rayon and city governments. The federal budget includes large and important enterprises, pipelines, electric power, marine transport, and national (but not local) environmental problems. The federal gov- ernment also is responsible for international trade activities (export and trade subsidies) and fundamental science. In the social sector, the federal budget accounts for a small Comparative Federalism: Trends and Issues in the United States, China, and Russia 83 share of financing universities, higher learning institutions, specialized health-care facili- ties, and culture and museums. This is in keeping with the principle of minimizing fed- eral involvement in cases where the 'benefit zone" is local or regional. The oblast level is responsible for facilities of an interjurisdictional nature: river trans- port, oblast roads, environment at the oblast level, the preservation of forests, oblast-run vocational schools, health care for oblast hospitals, and specialized clinics. The oblasts and republics are also responsible for small and medium enterprises-local light industry and consumer goods. Oblasts are increasingly transferring such enterprises down to the rayon level. The expenditure responsibilities of rayons and townships are concentrated in the social services area. Rayon budgets account for almost 100 percent of total expendi- tures on basic education, 85 percent of total expenditures on health, 60 percent of kinder- garten services, 60 percent of housing expenditures, and 80 percent of public utility ex- penditures. While this basic assignment fits the accepted norms, the actual practice does not. There is a lack of clarity in the assignment of specific service responsibilities to specific levels of government. This has led in some cases to provision by more than one level of government and in others to provision at the "wrong" level. To make matters more com- plicated, the assignment of responsibility has included not only the traditional public service functions, but also ownership of certain commercial and industrial enterprises and the expenditure responsibilities related to that ownership. Fiscal Autonomy Subnational governments have relatively little autonomy on the revenue side of the budget. All tax rates are set at the federal level, and taxation is uniform across the coun- try. All major user charge rates are nationally set. Subnational governments must simply adjust to the level of taxation chosen by the federal government. Technically, subnational governments may not borrow, although they do manage some credit financing by bor- rowing through their enterprises. The level of expenditures is largely controlled by the federal government in four ways. First, the sharing rates for all major taxes are determined by the federal govern- ment. Second, the oblasts are required to gain federal approval for their budgets each year. Third, the federal government provides a set of ad hoc grants to supplement shared taxes. Fourth, the assignment of expenditure responsibilities is determined by the federal government. This is not to say that the subnational government sector has no voice in the determination of the budget, because these decisions are all negotiated and subnational governments are an important force in Russia. Subnational governments have much more discretion when it comes to the composi- tion of government expenditures. Basic pay rates are centrally set, but otherwise local governments have a good deal of flexibility in choosing how to allocate their budgets among competing expenditure needs. While the unit cost "expenditure norms" of the past have been rendered obsolete by high rates of inflation, the federal government still im- poses some mandates on the subnational governments. Each oblast may decide how it will treat its constituent local government units. Oblasts determine the tax sharing rates, give grants to the local governments under ad hoc distributions, and approve the budgets of local units. Cities and rayons have rela- 84 Macroeconomic Management and Fiscal Decentralization tively little discretion when it comes to determining the overall size of the budget but considerable discretion when it comes to determining the mix of expenditures used. Recent Trends The Russian Federation is too new to be characterized by fiscal trends. All changes in fis- cal structure now are driven by the transition to a market economy and by the extremely unstable macro economy. There have been major annual changes in the tax structure, and frequent changes in both the system of revenue sharing and the assignment of expendi- ture responsibilities to lower level governments. The underlying theme of these policy changes seems to be a mix between finding the right balance of fiscal powers between the levels of government and holding the federation together by quieting the calls for sepa- ration with revenue sharing devices. The constraint, however, is the pressure on the cen- tral government to find resources to resolve its economic problems. SHIFTING CENTRAL FUNCTIONS DOWNWARD. Recent changes in the assignment of expenditure responsibility have violated these principles in very fundamental ways. For example, in early 1992, the central government shifted responsibility to the oblast and rayon governments for most of the price subsidy and income maintenance programs. These programs had previously been financed with transfers from the central govern- ment (equivalent to about 5 percent of GDP). While the burden of price subsidies will cease after prices are freed, the underlying problem of financing social protection for those most hurt by economic change will not. The government has neither generated es- timates of the cost of financing this social protection nor developed a plan as to how this cost might be matched with revenues available in each oblast. Until recently, the federal level was responsible for approving, financing, and imple- menting all subnational capital investment. In mid-1992, many investment responsibili- ties-both those with national significance (such as highways, military housing, and air- ports) and those with local significance-were shifted to subnational budgets. This policy decision took place after the second quarter (1992) revenue-sharing agreement yielded apparent budgetary surpluses for many oblasts. Shifting these responsibilities down to lower-level governments may have been a response to short-run budget pressures, but it is inconsistent with expenditure assignment principles.6 Moreover, it was a badly timed action. The assignment of fixed and uniform sharing rates to oblasts from the second quarter of 1992 made oblasts much less willing to accept additional expenditure responsibilities. The main rationale for the reassignment of ex- penditure responsibilities appears to have been to balance the budget of the central gov- ernment and to "claw back" the apparent surpluses of subnational governments, which were thought to have been created by the introduction of fixed sharing rates. In fact, the shift in responsibilities appears to have placed serious budget pressures on some subna- tional governments, creating the risk that important social expenditures (health care and education) would be crowded out. The more general issue here is the absence of concreteness in assigning responsibili- ties. Both subnational and central governments reap advantages from the continued 6. This discussion, and later references to intergovernmental fiscal relations in Russia, draw from Wallich, ed., Russia and the Challenge of Fiscal Federalism (Washington, D.C.: World Bank, 1994), Chapter 1, "Russia's Dilemma," pp. 9-10. -Christine Wallich. Comparative Federalism: Trends and Issues in the United States, China, and Russia 85 murkiness. Subnational governments use their broader responsibilities to bargain for a larger share of revenue, and the federal government has an additional instrument- jettisoning expenditure responsibilities-to balance its own budget. This lack of defini- tion cannot be perpetuated much longer if the system of intergovernmental relations in the Russian Federation is to move away from a "bargaining" mode of budget determina- tion toward greater certainty and predictability. More important, if this trend persists, what important expenditure functions will the central government perform to justify its existence to skeptical regional governments? By eschewing the concrete assignment of responsibilities, the federal government may inad- vertently be contributing to its worst fear - the disintegration of the Russian Federation (Martinez-Vazquez 1994). An important part of social expenditures and most capital ex- penditures have been delegated to the subnational governments; however, revenues pro- vided to subnational governments may not have taken these new responsibilities into account. REVENUE SHARING. The Russian system of revenue sharing has two distinct features. First, unlike most systems of intergovernmental finance, in which the center collects and shares most national revenue with lower levels of government, revenue in the Russian system is "shared upward" from rayons, to oblasts, and then to the federal budget. Sec- ond, the system is not a "system," but a collection of ad hoc, bargained, nontransparent agreements whose effects are not well understood. The bargaining inherent in this sys- tem makes subnational governments highly dependent on the center and creates consid- erable uncertainty about their fiscal autonomy and responsibilities. The sharing rates for the enterprise income tax and the VAT changed significantly during 1992 and early 1993. The direction appears to favor subnational governments, though it is not clear whether the increased claim on local revenues has been adequate to offset the increased expenditure responsibilities assigned (Martinez-Vazquez 1994). The retention rate on the enterprise income tax increased from 15 to 22 points (on the 32 per- cent rate) between early 1992 and early 1993. The increase in the value added tax sharing rate is more difficult to estimate because the sharing rate varies by oblast. However, the median VAT retention rate doubled during this period (Bahl 1994b). REVENUE DETERMINATION. Is there some rough justice in the present system of revenue sharing? In other words, are revenues implicitly distributed among the oblasts in some systematic way? To try and answer this question, Bahl and Wallace (1994) carried out a multiple regression analysis on actual 1992 data for eighty-eight oblasts with per capita retained revenues (excluding subventions) as the dependent variable. Three results from this analysis stand out. First, oblasts with a larger population retain more revenue on a per capita basis. Second, oblasts with a higher average wage, and with a higher growth rate in the average wage, retain significantly more on a per capita basis. Third, there does not appear to be a strong association between indicators of need in the population and per capita retained revenues. Neither hospital beds, infant mortality rates, nor highway density turned out to be significant determinants of the share of re- tained revenues. Taken together, these results suggest that in 1992, revenue distributions were driven in significant part by the strength of the economic base (wage levels, wage growth, and population size). This finding is consistent with the results of a similar analysis carried out by the World Bank on data for the first half of 1992 (Bahl 1994b). 86 Macroeconomic Management and Fiscal Decentralization Equalization The assignment of the individual and corporate taxes to the subnational level on a deri- vation basis necessarily means that higher income territories will derive more revenue. Since this system is likely to allocate more revenues to higher income oblasts, some form of subvention must be provided to protect the budgetary position of territories whose economic base is not strong enough to support an adequate level of public services. In fact, the federal government undertakes three types of discretionary action to equalize the distribution of fiscal resources. First, it approves the final budget expenditures of each oblast. Second, it determines the level of subventions that will flow to each oblast, and this is partly determined by the approved budget level. Third, it determines the VAT retention rates. How does one measure the "success" of the Russian system in equalization? First, in an equalizing system, oblasts with a higher level of expenditure needs and a lower level of fiscal capacity would have a greater tax retention rate. Bahl and Wallace attempted to measure this effect by regressing the ratio of taxes retained to taxes collected, against general indicators of need and fiscal capacity. The multiple regression results showed that relatively little of the variation in the level of the retention rate across oblasts could be explained, suggesting that much of the process is ad hoc. However, there were some systematic relationships. The tax retention rate was significantly higher where the rate of urbanization was lower and where there was a smaller concentration of elderly population. The results also indicated that, other things being equal, the autonomous republics have a higher retention rate than do other oblasts. Perhaps the most significant finding here is that the tax-sharing system does not redistribute resources toward oblasts with a lower per capita income. One may not call this tax-sharing system an equalizing scheme. If the system in operation in 1992 had an equalizing element, it would have to be sub- ventions (grants). The allocation of the subvention among oblasts is not done according to a strict formula but appears to take the form of an ex-post, deficit-filling grant. The (unweighted) mean level of subvention among the eighty-eight oblasts studied was about 25 percent of total revenues, and it ranged from a low of zero to a high of over 90 percent. Were the subventions distributed on an equalizing basis? Bahl and Wallace (1994) estimated that per capita subventions are significantly higher in oblasts with a greater fiscal capacity (measured by enterprise profits), suggesting a counterequalizing pattern. The Chinese "Federal" System China is not a federalism. It is a unitary state. But in many ways, the Chinese public fi- nance system has characteristics in common with a fiscal federalism.7 There is a distinct middle level of government, the Province, which is free to determine the fiscal responsi- bilities of its subordinate local governments. There is a type of formula system for allo- cating revenues among provinces, and provincial governments have some discretion in forming their budgets. 7. This section on the Chinese "federal" system draws on Bahl and Wallich (1992) and Bahl (1994a). Comparative Federalism: Trends and Issues in the United States, China, and Russia 87 Other features make China less of a federal structure. The central government and only the central government sets the tax rates and the tax bases, and it even determines most of the rates of user charges. Provincial and local governments cannot borrow. Tax and Expenditure Structure The Chinese tax structure is dominated by the income tax on enterprises and by a family of indirect taxes (the product tax, the business tax, and the value added tax). The enter- prise income tax is levied on all companies with a maximum rate of 33 percent. The big- gest departure of the Chinese company tax from traditional company income taxes is that all the wages are not deductible, and depreciation rates are low. The indirect taxes are dominated by the value added tax, which is now levied at two rates. Services and small traders are taxed under a "business tax," which is a turnover tax on gross sales, and cer- tain manufacturers are taxed under the product tax, which is a gross receipts tax. All taxes are collected by the State Tax Administration, a central government agency, but the collection machinery is decentralized to the lowest level. As in Russia, taxes are collected at the bottom and "shared up." Revenues are shared between the central and provincial governments according to a negotiated formula. Provinces then determine sharing rates in any way they choose. Up until 1994, the tax sharing system involved the determination of tax retention rates by central government negotiation with each province. In addition, there is a system of ear- marked grants distributed on an ad hoc basis. Local Autonomy By comparison with most countries in the world, subnational governments in China have little formal, or legal, independence in matters of structuring their tax system or deciding on the level and composition of expenditures. All tax rates and bases are set centrally, and so there are no truly local taxes-defined as those whose rate or base the subnational government can unilaterally fix-at the subnational level. Moreover, the central govern- ment determines, for each province, a share of taxes to be turned over to the center. In effect, subnational tax collections in China are central government taxes whose revenues are allocated among provinces, municipalities, and the central government. Even with this degree of centralization in the rules, however, subnational govern- ments have an important impact on spending levels and on the amount of revenues raised within their provincial jurisdiction. This follows because provinces design and implement the system of intergovernmental relations between the province and local governments. Moreover, a substantial amount of autonomy arises because local govern- ments exert some central control over tax collection and assessment. Local governments also have considerable freedom in awarding tax contracts to their enterprises. 8 Responsi- bility for implementation of the tax system is a very powerful policy instrument in the hands of local governments, and indications are that they use it. 8. A 1994 change in the tax law eliminated this feature. 88 Macroeconomic Management and Fiscal Decentralization Expenditure Autonomy Autonomy on the expenditure side of the budget is limited for provinces. Subnational government budgets are determined as part of a consolidated central, provincial, and local budget and as such must satisfy the (negotiated) fiscal targets laid down by higher level government. The budgetary choices of provincial governments are further limited by expenditure rules, mandates, and monitoring by higher level government. Within the province, there is more room for discretion. At the local level, provincial governments are responsible for approving the budgets and financial plans of municipal and county governments. This means they can control the spatial distribution of expen- ditures within the province. There appears to be great variation in the system of provin- cial-local relations across provinces, suggesting that the provincial governments have significant room to adjust fiscal decisions to accommodate local needs and preferences within the parameters set by the central government. Within the system of "vertical" responsibility, each province must account to the cen- tral government for its activities. In this process of vertical accountability, the following principles restrain, or guide, budgetary choices of provincial governments: (a) there can- not be a deficit; (b) current expenditures to maintain infrastructure have the highest pri- ority among urban construction-related expenditures; (c) the provision of social overhead facilities (such as education, scientific research institutes, and hospitals) take a high pri- ority; expenditures on culture and education are mandated to increase by at least the same rate as total expenditures; (d) employment levels and wage rates are fixed by the central and provincial governments; (e) all revenues from the urban maintenance and construction tax must be spent for urban maintenance and construction, that is, for public utilities and public facilities. Recent Trends The intergovernmental fiscal arrangement has remained a point of debate in China dur- ing much of the past decade. With rapid economic growth has come a demand for more resources to cover central government needs, and a demand for a greater share of reve- nues at the lower levels to support local projects. Three important issues have arisen: the shifting balance between the central and subnational level, continuing fiscal disparities among the provinces, and equalization policies. THE DIVISION OF REVENUES: VERTICAL BALANCE. Between 1983 and 1992, real expe n- ditures of provincial and local governments increased by 35 percent. While many coun- tries in the world could not boast such an increase, it is in some respects modest for China. The data in Table 4-5 show that, between 1983 and 1992, budgetary revenues col- lected by local governments have not grown as fast as total output, or as fast as local government budgetary expenditures. In fact, at the time of the income tax reform in 1983, the local government sector was spending an amount equivalent to 73 percent of what it collected and turning a net amount of 27 percent over to the central government for na- tional purposes. By 1992, the local government sector was spending an amount equiva- lent to their total collections and was receiving an additional subsidy from the center equivalent to about 6 percent of expenditures (See Table 4-5). Comparative Federalism: Trends and Issues in the United States, China, and Russia 89 Table 4-5. China: Revenue Collections as a Percentage of GNP Local Local government Central collec tions as a expenditure as a Total budget- Local govern- government percentage of percentage of local ary revenues ment buidget- budgetary total gozverniment Year collected ary collections collec tionisb collections collections 1983 27.4 15.1 12.3 55.1 83 1988 19.9 11.2 8.7 56.3 106 1992 16.6 10.1 6.5 60.8 106 Income elasticity (I 983-1992) a 0.57* 0.76* 0.47* * Statistically significant at the .05 level. a. Estimated from InR=a+bln GNP where R = revenues (with 0 adjustment for discretionary changes) b= revenue-income elasticity b. Estimated as the difference between total revenue for the consolidated government sector (IMF definition) and local government budgetary revenue (Chinese definition). Source: World Bank estimates and Bahl (1994b). The problem that has arisen is that the central government share of total revenue is not sufficient to cover central government expenditures. Historically (through the mid- 1980s), central fixed revenues were considerably less than central government expendi- tures, and the difference was made up with net transfers from the local government sec- tor. For example, in 1984, the local governments ran a "collections surplus" of about 17 billion yuan, while the central government ran a "collections deficit" of about 21 billion yuan (see Table 4-6). The net transfer was from local to center. However, by 1989, the situation was reversed: the central government was collecting more than it spent, and the local government sector was running a collection deficit. In fact, in 1990, there was a net transfer from the center to the local sector (excluding any earmarked grants) equivalent to about 10 percent of central collections. It is this change in the division of revenues, and the consequent growth in extrabudgetary revenues, that has prompted a proposed re- form of the revenue sharing system. FISCAL DISPARITIES. Revenue collections are roughly proportional to income level. The five highest income provinces in China account for 18 percent of local government reve- nue collections and 19 percent of national income. However, these provinces account for only 10 percent of the national population. This suggests that there are significant dis- parities in per capita revenue collections among the provinces. In fact, this disparity is from 1,180 yuan in Shanghai to 96 yuan in Anhui and 8 yuan in Tibet-a substantial range (see Table 4-7). 90 Macroeconomic Management and Fiscal Decentralization Table 4-6. China: Tax Collection and Expenditure of Central and Local Governments, 1984, 1989, 1990 (in billions of yuan) Category 1984 1989 1990 Central government Revenue collections 52.4 110.6 136.8 Expenditures 72.8 110.5 137.3 Collections deficit/surplus -21.4 0.1 -1.5 Local government Revenue collectionsa 97.7 184.2 194.5 Expenditures 80.8 193.5 207.9 Collections deficit/surplus 16.9 -9.3 -13.4 Note: The difference between the local governments' surplus and the central government's deficit is the central government's foreign borrowing and domestic budget deficit. a. Central and local government collections are before transfer from local goverrunents to the cen- tral government, and before earmarked grant distribution to the local governments. Source: World Bank estimates and Bahl (1994b). What are the determinants of this variation in per capita revenue collections across provinces? The revenue collection statistics and income rankings shown in Table 4-7 sug- gest that per capita income is a major determinant. Provinces with higher per capita in- comes, other things being equal, collect more revenue. One might also argue that the rate of urbanization (or conversely the agricultural share of GDP) would be major determi- nants because the tax system covers urban industrial and commercial activity, and urban business activity, to a much greater extent than it does rural and agricultural activity. Fi- nally, because some of the less populous provinces tend to have greater rates of urbani- zation and less of an agricultural sector, one would expect them to raise more on a per capita basis. We have tested these hypotheses by estimating an ordinary least squares (OLS) re- gression of per capita budgetary collections against per capita income, the rate of urbani- zation, and population size. The results of this OLS regression, presented in Table 4-8, more or less confirm these hypotheses. There is a strong significant relationship be- tween per capita revenue collections and per capita income. The relationship holds over the 1987 and 1990 period, and it was also reported for 1985 (Bahl and Wallich 1992). The income elasticities are high. In 1990, a 10 percent difference in per capita income tended to be associated with a 16 percent difference in per capita revenue collections.9 Popula- tion size did show the expected negative relationship with revenues but was not signifi- cant in either the 1987 or the 1990 equations.10 9. Hofman (1993) has pointed out that virtually all of the variation in per capita collections can be explained by a squared per capita income term. 10. Urbanization was not included in the final equations because of collinearity with income. Comparative Federalism: Trends and Issues in the United States, China, and Russia 91 Table 4-7. China: Collection and Expenditure Disparities Among Provinces Collections Budgetarv expenditures Per capita Per capita Percentage Per capita amount Percentage amount Percentage of total income Province (in yuan) of total (in yuan) of total population rank Beijing 702 3.85 633 3.22 0.95 2 Tianjin 508 2.27 455 1.88 0.78 3 Hebei 132 4.11 142 4.09 5.40 20 Shanxi 185 2.72 196 2.67 2.54 12 Inner Mongolia 151 1.66 282 2.86 1.90 16 Liaoning 326 6.54 308 5.73 3.48 4 Jilin 204 2.56 289 3.36 2.18 9 Heilongjiang 216 3.87 262 4.35 3.11 6 Shanghai 1180 7.98 566 3.55 1.17 1 Jiangsu 201 6.89 149 4.73 5.94 10 Zhejiang 244 5.14 192 3.76 3.66 5 Anhui 96 2.76 108 2.89 4.98 29 Fujian 187 2.87 225 3.21 2.66 11 Jiangxi 107 2.05 133 2.38 3.34 27 Shandong 129 5.55 146 5.81 7.45 21 Henan 99 4.35 106 4.31 7.59 28 Hubei 143 3.94 156 3.98 4.77 18 Hunan 118 3.66 134 3.86 5.38 22 Guangdong 206 6.63 237 7.06 5.57 8 Guangxi 112 2.41 154 3.08 3.74 24 Hainan 111 0.37 263 0.82 0.58 26 Sichuan 111 6.06 132 6.69 9.48 25 Guizhou 113 1.87 149 2.28 2.87 23 Yunan 208 3.92 243 4.26 3.27 7 Tibet 8 0.01 582 0.61 0.19 30 Sha'anxi 134 2.25 172 2.68 2.91 19 Gansu 152 1.73 204 2.15 1.98 15 Qinghai 163 0.37 382 0.80 0.39 14 Ningxia 143 0.34 318 0.70 0.41 17 Xingjiang 165 1.28 313 2.24 1.34 13 Source: Computed from data provided by Ministry of Finance. 92 Macroeconomic Management and Fiscal Decentralization Table 4-8. China: OLS Regression Results for Local Government Revenues and Expenditures Against Selected Independent Variables, 1987 and 1990 1990 1987 1990 1987 1990 1987 Dependent Per capita Per capita Per capita Per capita Per capita Per capita variable collections collections expenditure expenditure collections collections Constant -113.70 -156.330 244.245 217.856 80.987 28.46 (1.82) (2.65) (3.23) (3.61) (2.15) (0.86) Per capita 0.250 0.362 0.088 0.07 income (17.63) (21.73) (5.14) (4.13) [1.605] [1.933] [0.487] [0.377] Per capita 0.052 0.131 income ... ... ... ... (31.36) (45.74) squared Population -0.454 -0.478 -2.974 -2.707 (1.01) (1.05) (5.493) (5.79) R2 0.93 0.95 0.74 0.73 0.97 0.99 N 30 30 30 30 30 30 Note: The t-statistics are shown in parenthesis below the regression coefficients. Elasticities are shown in brackets. There appears to be some weakening in the relationship between collections and in- come. In 1985, the four provinces with the highest levels of per capita income (Shanghai, Beijing, Tianjin, and Liaoning) raised 31.7 percent of revenues and accounted for only 6.4 percent of the national population. In 1990, these same provinces raised only 20.6 percent of locally raised revenues. In fact, virtually all of the higher income provinces had growth rates in budgetary collections that were below the 1987-92 average (see Table 4-9). Disparities are much less pronounced on the expenditure side, with a per capita variation from 633 yuan in Beijing to 106 yuan in Henan and 108 yuan in Anhui (Table 4- 7) 11 The five highest income provinces, with 10 percent of the population, account for 17 percent of the expenditures. The pattem of variation across provinces might be explained in the following way. Higher income provinces spend more because of the greater de- mand for public services by their citizens and their enterprises, their ability to raise more "local fixed" revenues, their ability to attract more grants, and very importantly, their ability to slow the flow of revenues from the center. More urbanized provinces spend 11. The range in 1991 per capita state and local government expenditures in the United States (excluding Alaska) was from $6,525 in New York to $2,715 in Arkansas. Comparative Federalism: Trends and Issues in the United States, China, and Russia 93 Table 4-9. China: Percentage Increase in Revenues and Expenditures, by Provinces, 1987-92 Budgetary collections Budgetar,l expenditures Per capita incote Province Percentage Relative Percentage Relative rank increase (average = 100) Rank increase (average = 100) Rank (1990) Beijing 28.36 44.12 27 46.57 59.01 27 2 Tianjin 12.85 19.99 28 49.29 62.47 26 3 Hebei 72.87 113.37 16 87.60 111.00 6 20 Shanxi 72.51 112.80 17 54.58 69.16 23 12 Inner Mongolia 101.13 157.33 8 58.19 73.73 21 16 Liaoning 40.34 62.76 26 84.78 107.44 9 4 Jilin 51.89 80.73 22 53.21 67.42 24 9 Heilongjiang 57.27 89.10 20 55.24 70.00 22 6 Shanghai 6.59 10.25 29 70.55 89.40 16 1 Jiangsu 42.12 65.53 24 85.09 107.32 8 10 Zhejiang 55.00 85.57 21 86.01 108.99 7 5 Anhui 41.97 65.29 25 66.43 84.18 19 29 Fujian 130.25 202.63 4 111.30 141.04 3 11 Jiangxi 75.08 116.80 15 80.95 102.58 12 27 Shandong 91.40 142.19 10 94.10 119.24 4 21 Henan 64.73 100.71 19 78.50 99.48 13 28 Hubei 44.06 68.54 23 62.30 78.95 20 18 Hunan 70.61 109.85 18 77.19 97.81 14 22 Guangdong 132.21 205.67 3 127.36 161.40 1 8 Guangxi 100.39 156.18 9 15.92 20.18 29 24 Hainan n.a. n.a. n.a. n.a. n.a. n.a. 25 Sichuan 86.34 134.32 12 82.16 104.11 10 26 Guizhou 119.45 185.83 6 91.87 116.41 5 23 Yunan 191.60 298.07 2 125.70 159.28 2 7 Tibet 2,825.00 4,394.86 1 81.93 103.82 11 30 Sha'anxi 80.80 125.70 13 72.63 92.03 15 19 Gansu 75.95 118.16 14 70.54 89.38 17 15 Qinghai 103.49 161.00 7 51.88 65.74 25 14 Ningxia 88.08 137.02 11 33.95 43.03 28 17 Xingjiang 125.13 194.66 5 66.49 84.25 18 13 94 Macrtlectoomic Management and Fiscal Decentralization more, arguably because urbanization reflects a greater capacity to raise and retain taxes. Again, less populous provinces would be expected to spend more on a per capita basis, if the effects on income level are controlled. The regression results, presented in the middle two panels of Table 4-8, confirm these expectations. About three-fourths of the inter- province variations in per capita expenditures can be explained by variations in per cap- ita income and in population size. The income elasticities for expenditure variations are much less strong than in the case of collections, suggesting some degree of equalization in the system. In 1990, a 10 percent difference in per capita income was, on average, asso- ciated with a 4.8 percent difference in per capita expenditures. One might also ask whether the share of expenditures accruing to each province has changed significantly during recent years. The answer is that it has, as is shown in Table 4-10 for the 1987-92 period. There is no apparent pattern to this change. 12 EQUALIZATION. Evaluation of the equalization features of the present system is no straightforward matter. A central question is whether the system presents a significant potential for redistribution. That is, does the central government take back enough in transfers and give enough in earmarked grants to significantly alter the disparities in fis- cal capacity among the provinces? There are three ways in which redistribution might take place: * earmarked grants to the local governments, which augments their resources; * an increased tax sharing rate for local governments, which augments their re- sources; and • transfers to the center, which reduces the resources of each province. The total of transfers and grants was equivalent in amount to about 30 percent of local government collections in 1990. This suggests a significant potential for redistribution. Whether or not redistribution occurs, however, depends on how the transfers are ex- tracted and how the grants are allocated. Bahl's (1994b) empirical analysis gives mixed results. A simple correlation analysis shows that per capita transfers to the center are significantly higher in provinces with higher per capita incomes. There is some equalization in the distribution of transfers. However, the same is not true of earmarked grants. On a per capita basis, higher income provinces receive significantly more in earmarked grants. The pattern is counter- equalizing. These results also show that less populous provinces receive significantly less in per capita grants. Extrabudgetary Revenue Another important source of revenue for local governments in China is extrabudgetary revenues. These are of two types. The first are the "fiscal extrabudgetary funds" of the government. They are earmarked for capital purposes and include a set of taxes and charges that are controlled by the local government finance department. The most im- 12. I correlated the changing share of total expenditures with per capita income and with population but could find no significant relationship. Comparative Federalism: Trends and Issues in the United States, China, and Russia 95 Table 4-10. China: Changes in Expenditure Shares, 1987-92 Percentage shares of total budgetary expenditures Province 1987 1992 Change Beijing 3.49 2.83 -0.66 Tianjin 2.19 1.81 -0.38 Hebei 3.73 3.88 0.15 Shanxi 2.92 2.50 -0.42 Inner Mongolia 3.20 2.80 -0.40 Liaoning 5.65 5.77 0.12 Jilin 3.67 3.11 -0.56 Heilongjiang 4.64 3.98 -0.66 Shanghai 3.52 3.33 -0.19 Jiangsu 4.78 4.89 0.11 Zhejiang 3.60 3.71 0.11 Anhui 3.13 2.88 -0.25 Fujian 2.81 3.29 0.48 Jiangxi 2.65 2.66 0.01 Shandong 5.28 5.68 0.40 Henan 4.58 4.53 -0.05 Hubei 4.28 3.85 -0.43 Hunan 3.93 3.85 -0.08 Guangdong 6.79 8.54 1.75 Guangxi 4.76 3.05 -1.71 Hainan 0.00 0.99 n.c. Sichuan 6.17 6.22 0.05 Guizhou 2.22 2.36 0.14 Yunan 3.78 4.73 0.95 Tibet 0.64 0.65 0.01 Sha'anxi 2.66 2.54 -0.12 Gansu 2.20 2.08 -0.12 Qinghai 0.86 0.72 -0.14 Ningxia 0.83 0.62 -021 Xinjiang 2.37 2.18 -0.19 Source: Computed from data provided by Ministry of Finance. portant are the public utility surcharge-a 10 percent tax on the utility bills of consum- ers-and the urban construction and maintenance tax. There also are some minor reve- nues received from public housing and public property, and some institutional income that accrues to the various city enterprises. The latter include such items as fees and charges from hospitals, road maintenance charges, advertisement fees, and so on. Fiscal extrabudgetary revenues are relatively small compared with other revenue sources. The World Bank (1986 p. 259) estimated that in 1986, these extrabudgetary rev e- 96 Macroeconomic Management and Fiscal Decentralization nues of local governments accounted for less than 3 percent of total state budgetary reve- nue. The other type of extrabudgetary revenues is the retained earnings and depreciation funds of locally owned enterprises. In principle, these funds should not be classified as part of the government budget because they are not resources over which the local gov- ernments have complete control. On the other hand, they are funds that may be used for social purposes and to support the economic development goals of the local government administration, and their growth may be heavily influenced by local taxing and con- tracting practices. Extrabudgetary funds of the second type expanded very rapidly in the 1980s because of the objective of the reform program to give SOEs more flexibility in managing social funds and in revitalizing their operations (World Bank 1990, pp. 17-19). The enterprise contracting system that grew out of this economic reform allowed the local governments to shift funds from the budgetary to the extrabudgetary side. Extrabudgetary funds are by their very nature sensitive to the business cycle, as is shown by the time series data in Table 4-11. In 1992, extrabudgetary revenues of local governments were equivalent to about 89 percent of local government budgetary collec- tions. This share appears to have trended upward after 1986 when enterprise contracting began to grow significantly. The responsiveness of extrabudgetary revenues to GNP would appear to be greater than the responsiveness of budgetary revenues. I found a positive and significant relationship between the ratio of local government extrabudge- tary revenues to budgetary collections, and GNP, over the 1983-92 period.13 Table 4-11. China: Extrabudgetary Revenues of Local Governments, 1984-92 Percentage of central government Total Per capita Percentage of extrabudgetary Year (in billions of yuan) (in yuan) budgetary revenue revenues 1984 71.8 69 74 151 1985 89.4 86 76 141 1986 102.0 97 77 143 1987 120.1 111 82 146 1988 145.3 133 92 161 1989 157.5 142 84 148 1990 163.5 143 83 152 1991 186.2 161 76 n.a. 1992 214.7 183 89 n.a. 13. The ratio of extrabudgetary to budgetary revenue increased with GNP over the 1983-92 pe- riod, as indicated by the following regression results: EXB/B = 69.8 + 0.00093 GNP R2 = 0.43 (2.45) Where: EXB = extrabudgetary revenue; B = budgetary revenue; and the figure in parenthesis is the t-statistic Comparative Federalism: Trends and Issues in the United States, China, and Russia 97 Finally, it can be seen from Table 4-11 that local government extrabudgetary revenues are about 1.5 times greater than central government extrabudgetary revenues. This spread has remained approximately constant over the past decade. Tax Sharing: 1988-92 The tax sharing system was changed beginning in 1988 to a much more negotiated ap- proach. The basic idea is still to allow the local governments to retain an amount of reve- nues that will enable them to cover a basic level of expenditure needs. The base year cho- sen to define the expenditure amount was 1987. Local governments may retain this amount plus a share of any increase in revenues, according to a pre-determined negotia- tion. This agreement, originally planned to be in force for three years, led to the following six arrangements for tax sharing.14 1. Revenue growth up to a defined limit is divided between center and province according to a formula. Both the limit and the sharing ratio are negotiated. The province may retain all collections above the limit. Ten areas use this method of tax sharing. 2. The ratio of provincial revenue collections to provincial approved expenditures defines the local retention rate. In effect, the local government retains a fixed percent of total collections no matter what the increment (or lack of an incre- ment). The share is determined from approved base figures for the previous two years. Three areas use this method. 3. The retained share of revenues is based on the ratio of revenues to approved expenditures in 1987. A retained share of all incremental revenues is negotiated separately. Three areas use this method. 4. The amount to be paid to the central government in the first year is determined by base year revenues and expenditures. This amount then increases at an agreed rate in later years. Two areas use this method. 5. Provincial governments retain an amount equivalent to (a) approved expendi- tures in the base year, and (b) a fixed amount of revenues raised that are in ex- cess of this level. Thus, the central government receives a fixed payment each year. Three areas use this method. 6. The central government provides a fixed subsidy, of an amount equal to the dif- ference between actual collections and expenditure needs in the base year. This method is used by sixteen areas. Earnarked Grants Local governments receive earmarked grants from the central government for a variety of purposes. These include appropriations for capital construction projects, price subsidies for urban grain consumption, social relief funds, and special subsidies for health and education of the poor, minority, and border provinces (Hofman 1993). There is no set formula to determine the amount of earmarked grants to distribute in any given year, 14. This description is elaborated in Qiang (1993). 98 Macroeconomic Management and Fiscal Decentralization and the distribution among the provinces appears to be ad hoc rather than formula- based. Earmarked grants have grown significantly, as may be seen in Table 4-12, and they are now greater in magnitude than tax sharing transfers to the central government. In 1990, earmarked grants were equivalent in amount to 14.4 percent of local government budgetary expenditures, slightly less than in 1985. Some would argue that a primary purpose of earmarked grants is equalization. As may be seen from the simple correlations in Table 4-13, however, per capita earmarked grants are distributed in significantly heavier amounts to higher income provinces. These are project rather than entitlement grants, and it is not unusual for more developed re- gions to absorb these because of their greater capability at project preparation and their greater ability to "buy in" on a matching basis. Conclusions There are great differences in the systems of intergovernmental fiscal relations in the United States, Russia, and China. The major difference is the degree of autonomy af- forded to subnational governments in the United States, where taxing powers and ex- penditure discretion reside at the subnational level. In both Russia and China, all tax rates are set at the central level, and they are applied uniformly across the country. Second, the tax administration machinery in the United States is three-leveled with independent federal, state, and local administrations. Each level collects its own revenue. In Russia and China, all revenues are collected by a central administration, but collection is at the bottom and shared up to the top. Third, in the United States revenue sharing is by the assignment of taxes to the sub- national level and a system of grants in aid to local governments. In both Russia and China, there is a combination of derivation tax sharing and grants to local governments. Table 4-12. China: Estimated Net Redistribution, 1985-90 (in billions of yuan) Local government Contracted transfers Earmarked grants Net transfer Year budgetary revenue (local to central) (central to local) to central 1985 1176.55 310.52 182.98 127.54 1986 1325.59 303.67 244.26 59.41 1987 1465.69 171.38 268.24 -96.86 1988 1582.48 254.41 263.71 -9.30 1989 1881.31 262.02 277.73 -15.71 1990 1976.83 267.83 285.13 -17.30 Source: Computed from data provided by Ministry of Finance. Comparative Federalism: Trends and Issues in the United States, China, and Russia 99 Table 4-13. China: The Ratio of Expenditures to Collections, 1987 and 1992, by Province Ratio Province 1987 1992 Change 1987-92 Beijing 0.78 0.89 0.11 Tianjin 0.56 0.74 0.18 Hebei 0.92 1.00 0.08 Shanxi 1.24 l.ll -0.13 Inner Mongolia 2.34 1.84 -0.50 Liaoning 0.74 0.98 0.24 Jilin 1.39 1.40 0.01 Heilongjiang 1.23 1.21 -0.02 Shanghai 0.30 0.49 0.19 Jiangsu 0.63 0.83 0.20 Zhejiang 0.67 0.81 0.14 Anhui 1.15 1.34 0.19 Fujian 1.21 1.11 -0.10 Jiangxi 1.34 1.38 0.04 Shandong 1.03 1.05 0.02 Henan 1.03 1.12 0.09 Hubei 0.93 1.05 0.12 Hunan 1.03 1.07 0.04 Guangdong 1.01 0.99 -0.02 Guangxi 2.22 1.28 -0.94 Hainan n.c. 1.69 n.c. Sichuan 1.16 1.13 -0.03 Guizhou 1.47 1.28 -0.19 Yunan 1.44 1.11 -0.33 Tibet n.c. n.c. n.c. Sha'anxi 1.34 1.28 -0.06 Gansu 1.39 1.35 -0.04 Qinghai 3.06 2.28 -0.78 Ningxia 2.88 2.05 -0.83 Xinjiang 2.91 2.15 -0.76 Mean 0.97 1.06 0.09 Source: Computed from data provided by Ministry of Finance. Because of these differences, it is not possible in Russia or China to separate the issues of tax policy, tax administration, and intergovernmental fiscal relations. In the United States, it is possible to treat reform in each of these areas independently of the other two. There are also similarities in the systems of the three countries. In the United States and in Russia, about 40 percent of expenditures are made by subnational governments, 100 Macroeconomic Management and Fiscal Decentralization while in China, the share of local governments is over half. At least on the expenditure side, all are highly decentralized systems. All three countries have intergovernmental transfer systems that to some extent are not transparent. Part of the U. S. grant system is distributed on an ad hoc basis, while all of the Chinese and Russian systems of grants are distributed on an ad hoc basis. In all three countries there are large disparities among subnational units in fiscal capacity, and in none of the three countries is there any evidence that the intergovernmental transfer system is equalizing. In all three countries it is left to the middle level of government to decide how reve- nue and expenditure responsibility should be distributed among its local governments. In all three countries the middle level governments have made a variety of choices. Finally, on the issue of trends, both China and Russia are still on a path toward in- creased fiscal decentralization, at least in terms of how much of their total resources is allocated to the subnational government level, while the United States apparently is re- centralizing. This raises an interesting question in light of the conventional theory of fis- cal federalism, which suggests that the efficiency objectives of decentralization should be much weaker in China and Russia. References Advisory Commission on Intergovernmental Relations. 1992. Intergovernmental Perspective. Vol. 18, No. 4 (Fall). Washington, DC. . 1993. Significant Featuires of Fiscal Federalism. Vol. 2 (September). Washington, DC. Bahl, Roy. 1994a. "China: Tax Structure and Intergovernmental Fiscal Relations." Paper prepared for the World Bank (March). _ - 1994b. "Revenues and Revenue Assignment: Intergovernmental Fiscal Relations in the Rus- sian Federation." In Russia and the Challenge of Fiscal Federalism, ed. Christine Wallich. Wash- ington, DC: The World Bank. Bahl, Roy, and Johanne:s F. Linn. 1992. Urban Public Finance in Developing Countries, 385-427. New York: Oxford University Press. Bahl, Roy, and Shyam Nath. 1986. "Public Expenditure Decentralization in Developing Countries." Government and Policy 4: 405-418. Bahl, Roy, and Sally Wallace. 1994 "Revenue Sharing in Russia." Environiment and Planning: Govern- ment and Policy 12 (3): 293-307. Bahl, Roy, and Christine Wallich. 1992. "Intergovernmental Fiscal Relations in China," PRE Work- ing Paper 863 (February). Washington, DC: The World Bank, Policy Research Department. _____. 1995 "Intergovernmental Fiscal Relations in the Russian Federation." In The Decentralization of the Socialist State: Intergovernmental Finance in Transition Economies, ed. Richard M. Bird, Rob- ert D. Ebel, and Christine 1. Wallich. Washington, DC: The World Bank. Bocherdering, T.E., and R.T. Deacon. 1972. "The Demand for the Services of Non-Federal Govern- ments." American Economic Review 62 (5): 891-901. Hofman, Bert. 1993. "China: Budgetary Policy and Intergovernmental Fiscal Relations," World Bank paper (July), 72-78. Martinez-Vazquez, Jorge. 1994. "Expenditures and Expenditure Assignment." Russia and the Chal- lenge of Fiscal Federalism, ed. Christine Wallich. Washington, DC: The World Bank. Oates, W.E. 1972. Fiscal Federalism . New York: Harcourt Brace Jovanovich. Comparative Federalism: Trends and Issues in the United States, China, and Russia 101 Quiang, Gao. 1993. "Problems in Chinese Inter-governmental Fiscal Relations and the Main Diffi- culties in Future Reform." Paper presented at Shanghai conference sponsored by the Ministry of Finance on October 25, 1993. U.S. Department of Commerce. 1994. Survey of Current Business. Vol. 74, No. 6 June), Economics and Statistics Administration, Bureau of Economic Analysis. Washington, DC. U.S. Department of Labor. 1994. Employment and Earnings. Vol. 41, No. 6 Uune), Bureau of Labor Statistics. Washington, DC. World Bank. 1986. China: Revenue Mobilization and Tax Policy. Washington, D.C. .1990. Budgetary Policy. Washington, D.C. . 1992. "Russia: Intergovernmental Fiscal Relations in the Russian Federation," Report 11302- RU. Washington, D.C. 5 Russia's Dilemma of Fiscal Federalism Christine 1. Wallich This paper, which explores the complex and conflictive questions of fiscal federalism in Russia, was prepared for a Chinese audience. It begins by comparing the two countries' intergovernmental systems before describing the Russian system in greater detail, with the challenges it faces. Introduction Fiscal Federalism in China and Russia: An Abstract of Common Issues China's intergovernmental system shares many of the same features as Russia's. First, both are structured as three-tiered administrations. In each, fiscal decentralization has been an important dimnension of reforms and has proceeded quickly, with far-reaching consequences. And, in both countries, regional interests are important, and provincial administrations are increasingly powerful, with all that this implies for the design of a system of fiscal federalism. China's and Russia's fiscal federalisms have many features in common. Indeed, China's system had Russian influence, owing to the postliberation technical cooperation programs. The Chinese and Russian systems thus have much in common, but they have evolved in different ways based on political exigencies, differing objectives, and changing circumstances. These differences are interesting to explore and contrast. TAX ADMINISTRATION. Until recently, both Russia and China had the "bottom up" system of tax administration, common to many former socialist economies, in which lower level tax offices remitted tax revenues collected at the local level upward to the provincial and central government tax offices. This system of "bottom up" tax collection confers substantial vulnerability on the central government, which must rely on the com- pliance of provincial (or oblast) governments to remit the agreed amounts of revenue. In The author is Lead Economist in the Central Europe Department of the World Bank. This paper draws on the following chapters in Wallich (1994): Roy Bahl, "Revenues and Revenue Assignment: Intergovernmental Fiscal Relations in the Russian Federation"; Jennie Litvack, "Regional Demands and Fiscal Federalism"; Charles McLure, Jr., "The Sharing of Taxes on Natural Resources and the Future of the Russian Federation"; as well as several chapters by the author, in the same volume. 103 104 Macroeconomic Management and Fiscal Decentralization Russia, the refusal of a number of oblasts to remit their revenues has led to revenue shortfalls for the federal budget and complicated macroeconomic management. China has recently taken the vital step of reforming the State Tax Service, and it has established a unified tax administration responsible for collecting central government revenues. While it will take time for the State Tax Service to become effective, this is an important first step in regularizing the revenue flow to the central budget and modernizing Chinas tax administration. CONTRACTING AND NEGOTIATED REVENUE SYSTEMS. In addition, both China and Russia have in common an intergovernmental revenue sharing system that has not been transparent. Negotiation between the provinces and the central government has deter- mined the provincial shares of central taxes as well as the level of transfers to the prov- inces (if any). Tax shares were differentiated by tax and different for each province. For- mula-based transfers have not been used in either country. This is changing now in China with the formal assignment of certain taxes to the central and the local levels and the plan to introduce a formula-based transfer system with equalization features. This reform now under way is an important step in the direction of making China's system more transparent, ensuring revenue adequacy and improving the equity dimensions of the system. EQUALIZATION. In neither country has equalization been an explicit feature of the intergovernmental system, although both countries exhibit significant inter-regional di s- parities. In China, the richest province (Shanghai) has a per capita income more than seven times that of the poorest (Guizhou). In Russia, the ratio is 10:1. Is there a role for greater equalization, and if so, how should it be introduced? Russia had no growth from 1991 to 1994, and indeed, it has experienced a severe output decline. One could argue that it would be better not to push for too much equalization (taking away from the rich provinces to give to the poor provinces) but rather to leave the resources in the richer provinces, which could take advantage of their stronger fiscal base and invest in growth- promoting public services and investments. This is essentially what the contracting sys- tem has done in China. Others argue that with a decade and a half of rapid growth since 1979, but increasing regional disparities, China can now afford for the fiscal system to be more equalizing. A formula-based transfer system is the way to begin. UNCODIFIED AND MURKY SPENDING ASSIGNMENTS. Both countries have rather murky expenditure assignments based on tradition and inertia; neither has codified ex- plicit spending responsibilities in a law. In both, there has been growing decentralization of expenditures to the provincial level, with the subnational governments responsible for a growing share of total spending. This murkiness (and inability to quantify spending responsibilities) make it difficult to design an intergovernmental fiscal system and to en- sure it provides for revenue adequacy at the central and local levels. In addition, both countries have put the cart before the horse by focusing first on the design of tax and transfer systems rather than on expenditure responsibilities. Thus, revenue availability is driving spending, rather than the other way around. Russia's Dilemma of Fiscal Federalism 105 THE SOCIAL SAFETY NET. Both Russia and China put responsibility for the social safety net "downstairs" at the subnational level. In Russia, until recently, price subsidies were administered by oblast governments, but they were financed by central transfers. Then in 1992 responsibility for the social safety net was transferred to the oblasts, an outlay equivalent to some 6 percent of GDP. China and Russia may need to reconsider the financing of the safety net and whether it should be the responsibility of local gov- ernments alone or whether the national government should share in its financing. A strong safety net is in the national interest over the rocky reform period ahead. UNIFIED NATIONAL TAX SYSTEMS AND LACK OF SUBNATIONAL REVENUE DISCRETION. The overall tax system in Russia and China is "unified," meaning that the central government sets the tax rates and defines the tax bases for all national-level and local-level taxes. In both countries, local governments have been assigned only minor, "nuisance" taxes of the sort that do not generate significant revenues. Local governments have no authority to determine tax rates, even of the taxes that are notionally assigned to them. This lack of fiscal discretion makes it difficult to modify local budgets to local needs and means that both the efficiency gains and the fiscal accountability that poten- tially should come from decentralization are being missed. SCOPE OF THE INTERGOVERNMENTAL SYSTEM. Both countries must decide on the appropriate scope of the intergovernmental fiscal system: should the provinces be re- sponsible for fiscal matters relating to towns and cities within their boundaries, or should the central government be responsible also for intraprovincial revenue sharing and trans- fers? Russia has yet to decide; the oblasts argue for full responsibility for "local affairs." The federal government, concerned that oblasts may not implement central policies, would prefer a system under which the center also allocates fiscal resources to the lowest level administrations. Arguably, in countries as large as Russia and China, assigning and monitoring spending responsibilities and designing and monitoring a revenue sharing and transfer system for up to 10,000 local communities would be very difficult. A federal structure, perhaps with some framework agreement, would work best. ENTERPRISES AND THEIR SOCIAL ASSETS. In both countries, state-owned enterprises have been important providers of public services and infrastructure. As privatization oc- curs in Russia, and as commercialization and harder budget constraints are introduced in China, enterprises will no longer be able to afford these outlays, and many will have to be put onto the budget. And since many of the enterprise functions are in areas of tradi- tionally local spending responsibility, it is the subnational budgets that will have to take them on. Both Russia and China will have to quantify these additional spending respon- sibilities and accommodate the additional budgetary outlays via the revenue sharing system. INTERLINKAGE OF TAX, REFORM, REVENUE SHARING, AND ADMINISTRATION. To begin, policymakers and legislators must change their approach toward public finance reform. In Russia, certain fiscal issues and reforms have traditionally been viewed in iso- lation-expenditure assignment and spending mandates, tax sharing, subventions and norms, and tax policy and deficit-reduction macropolicies. They must now be considered as a whole and the effects of each incorporated into the intergovernmental system. More 106 Macroeconomic Management and Fiscal Decentralization generally, fiscal policy, tax administration, and intergovernmental fiscal relations are so interconnected in Russia that all must be reformed simultaneously. CENTRALIZATION VS. DECENTRALIZATION. China's and Russia's intergovernmental systems have undergone substantial and rapid decentralization in recent years. Some argue that decentralization has been too rapid and has gone too far. There is concern that the emergence of such a major vertical imbalance will deprive the center of sufficient re- sources for macrostabilization and equalization. China's recent attempt to re-centralize and to increase the center's revenue share has been more successful than Russia's. China's premier, in his 1993 tour of the rich southern provinces, argued that "strengthening the center strengthens the locals." In Russia, the oblast/center relationship was likened to "a river with tributaries": stronger tributaries are needed to strengthen the center. As a re- sult, Russia has had little success in re-centralizing the resource flow, and the central budget remains highly vulnerable. Russia's Systemi: An Overview Russia's moment of truth is fast approaching. It is in the midst of an economic and politi- cal transition never attempted anywhere. Russia is trying not only to restructure its entire economic svstem but, at the same time, to protect the well-being of all citizens, stabilize prices and its external balance, and provide public services. It is also trying to establish a system of governance acceptable to far-flung regions whose cultural identity, natural re- source endowments, and degree of economic development differ widely. It is a herculean task for the new Russian Federation. At the heart of all these challenges is one issue: intergovemmental finance; more spe- cifically, the division of expenditure responsibilities and the assignment of revenues to different levels of government (see Figure 5-1). In a nutshell, it is about how the national revenue cake should be divided and which government (federal, oblast, rayon, or okrug) should be responsible for which spending. How Russia determines the division of ex- penditure responsibilities between the federal level, and how the national revenue pie is divided among Russia's three tiers-federal, oblast, rayon-will be key to Russia's eco- nomic growth, regional disparities, macroeconomic stabilization, and privatization. What are some of the challenges to be addressed? First, in the present system there is no "correspondence" (or matching) of responsibilities and resources. Most subnational governments do not have enough revenue to meet their spending responsibilities. So, Russia must also design a system of transfers that will both meet the shortfall and sup- port more efficient and equitable provision of services. These fiscal changes will deter- mine the efficiency with which the economy performs and its future direction. The new intergovernmental fiscal system now evolving gives subnational govern- ments new spending responsibilities (especially in investment and the social sectors), as well as new budgetary rights and new financial resources. Many of the important changes proposed under laws passed since December 1991 have not yet been fully im- plemented, however, and subnational governments are operating under transitional rules. Some oblast governments also want special treatment or "channels" outside the laws. Russia's Dilemma of Fiscal Federalism 107 Figure 5-1: Russia's Administrative and Federal Structure The Russian Federation is the largest and one of the most diverse countries in the world, with regions whose cultures, politics, and resource endowments vary widely. This diversity repre- sents a challenge to effective administration, to budgetary management, and to stabilization and structural policies. Russia is organized as a three-tiered federal state consisting of eighty nine provinces or states directly subordinate to the federal government. The eighty-nine states directly subordinate to the federal government comprise (1) the oblasts, okrugs, and krais, (2) metropolitan cities with "oblast" status (Moscow and St. Petersburg), (3) republics that until mid-1992 were called "autonomous republics," (4) autonomous regions, and (5) na- tional regions. (Map 1). Below them are municipalities and rayons, subordinate to the oblast government. Each oblast supervises the rural and urban areas within its jurisdiction, and from a budgetary perspective, each has a so-called independent (that is separate and free standing) budgetary and administrative status. Although the oblast-level finance depart- ments are officially autonomous according to the law, oblast finance officers are paid by the central government budget. Oblast finance departments therefore still have some allegiance to the federal government in Moscow and may respond to competing realms of authority or, in Russian parlance, be under "dual leadership." Russian Federation Central Government Oblasts (Urban) (Rural) Municipalities Rayons City Proper Urban Districts (District) or Rayons Urban 1 Urban Rural Soviets | Soviets Soviets Note: In some oblasts, and in some special cases, urban rayons may be directly subordinate to the oblast government. IBRD 27718 RUSSIAN FEDERATION Republics, krais, oblosts, and okrugs Obl-t, 1- -~~~~~~~~~~~~~~~~~~bk b-d- ~ ~ ~ ~ ~ o," - Chiplo. km.e, or repobd.c boordorre C ff Arcnc Cceoor Autonoooros oblost obrjg, or 0551 COOrl.bOrnu e --Inte,oot,onol boordones 2; _e Illpln )# 't' tRA9~~ol T- ;3t : 4 (5Russ- a n Feder a'tionjt~ Sea W-Ji tt~~0 ;o M , Ni< ; 9 ',1 .0 S =_ X 3S G \ SS A ~ ~ ~ ~ ~ ~ N~ ere . a>-' SI So| dmn n-n 5 r;r . ..efn 6Vo logd Ob>40 217 Vld 7rbls 34.0ooOls 54\ OrnugO0s 15orios'rg R oTkutnm>O 8. Ln ng ndOl,v 24 Reub f C -sos 40 Uloos9Ols 6.K/'em5-o.'ht bos 2JwshAtnmosOls 9 Novgorod Oblost 25 K,rov Oblost 41 Volgogrpd Oblast Autonomous Okrug 70 Agin Buryof 83 , oqodon Oblost~~~~~~~~~~~~~~~~.< .A h s >~~ ~ 135* W-A. ~ ~ ~~ ~~~~~~~~~~~~~~~~~~~N -S ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~kn 11 Btcyonskt Obloot N7 Oodov 0009'i 432 Vororef Repbro58Udut 49 Reotpol k 71o oo TooS Obloot 75 Nrroon,k -tonoc>Cti 12 Nvonovo Autonot 0 28 SNoonOloNovont Ols 54 ArohobrnoObclohteubi 5O KRoKrocoo-Ceos Sgc 65 ls-ryanOboott 86 SokaenOhA.ao,'.'s.- 13 Relugo.oo o hr0 4 St Polorhoorg City 23 toroslool 0131091 29 RepubLic of Totorston 55 Perot Osfoet 68 Sopaohl c of Sutyot SI NNooho,o,nh <"to 8.Le,ogooOroro2 Spohoco Coos 4 Uyrtocf Olot56 om-Ptio o.ChtoOlo'.3 ecsi Aonooo blh 9 ogoo bos 2 to hno 1Vo59o6Clo uoronnu org7 8mBoyl13 'o1o30.9 10Pko hle2 SpblrfMoo E 2Roolo fDohstn5 Sedloh ho AtooasO`23 4CuhtAo'"'ros2E, 11StosoOloi27M'doSaolo 3 045 e~bro58Umof Spblc7fIbo2 Obst876ooom 12loo,noObos 2 NnioyNogoodOboo 4 Koooro-olorReoblo59AlotKto 2 stOryr Sarot SoSofil 66lo NJKooo hon 9 elaodCbo 5 lrnodt(to60 RpbloofGto, lo,Atoooah hg 7Seoo o EN26 Io 14fletno bot 0hosEOlor 6 eooocofAygto6 KrotooObos 7 rooooskI( 3. Retail Sales Tax (general rate) 5-12% (9 provinces) 7-12"', (9 provinces) 4. Fuel Taxes Gasoline (cents/litre) 7.5-13.35 (8 provinces) 9.0-15.7 5.Tobacco Tax 1.484.8 ¢/cigarette 6.52-11 c/cigarette 6. Corporate Capital Tax General Rate (%) 0.2-0.45 (5 provinces) 0.3-0.6 (5 provinces) 7. Payroll Tax 1.5-3%. (2 provinces) 1 95-3.75% (4 provinces) 8. Health Care Premiums Individual ($/yr) 168-357 (3 provinces) 360432 (2 provinces) Federal transfers 1984-85 1993-94 1. To Persons Family Allowance Universal and taxable Replaced by income-tested, nontax- able child benefit. Child Tax Credit Income-tested, nonitaxable Replaced by income-tested, nontax- able child benefit. Sales Tax Credit Small amiount. income tested Enriched, income tested. OAS Universal Universal, subject to clawback. Unemployment Insurance Restricthons for those quitting jobs. 2. To Provinces Annual growth rate limited to 5% CAP for Ontario, B.C., Alberta. EPF Cash payments frozen until 1995-96. Equalization Growth rate limited to growth of CDP over 5-year period. Source: Ruggeri, Howard, Robertson, and Wart (1493). Figure 10.3. Federal Provincial Fiscal Arrangements The Federal-Provincial Fiscal Arrangements Tax Harmonization Fiscal Transfers Fiscal and Economic Policy - Tax collection agreements - Established programs financing - First Ministers' Conferences - Allocation rules - Canadian fiscal equalization program - Meetings of Finance Minsiters - Reciprocal taxation - Canada Assistance Plan - Committees of Officials - Other Intergovernimental Fiscal Relations in Canada: An Overview 243 Figure 10.4 Federal Transfers 4a. Major Federal Transfers, 1983-84 to 1994-95 Billions of dollars 40 35 30 25 - l '5 10 5 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 4b. Major Federal Transfers, 1994-95 Dollars per capita 3,000 2,500- 2,000 I .000 500 Nfld. PE.I. N.B. N.S. Man. Que. Sask. B.C. Alta. Ont. | R Equalization Canada Assistance Plan Established Programs Financing 244 Macroeconomic Mantagem?ient and Fiscal Decenitralizationi Figure 10.4 (continued) 4c. Major Federal Transfers as a Share of Provincial Revenues, 1994-95 Percent 50 ______-_________ 40 30 20 10 0 Nfld. PE.I. N.B. N.S. Man. Que. Sask. B.C. Alta. Ont. 4d. Total Federal Transfers, 1994-95 Other ($4.7 billion) Equalization ($7.7 billion) Established Programs Financing ($21.3 billion) Canada Assistance Plan ($8.2 billion) Note: Total federal transfers $41 9 billiwn. Equalization associated with tax transfers under EPF is included under ElF and excludlei from equalization. In tergovern ticnltal Fiscal Relationis in Canada: An Overview 245 Figure 10-5. How Established Programs Financing Works, 1994-95 Dollars per capita 735 - _ _ - - - - - -- - - - - - - - - - 600 400 200 0 Nfld. PE.I. N.S. N.B. Que. Ont. Man. Sask. Alta. B.C. * Tax transfer RJID Quebec abatement U Associated equalization El Cash transfer Note: The calculation starts with each province's total per capita entitlement, which is the same for all provinces. It equals the national average per capita federal contribution to shared-cost programs in 1975 plus $20 per capita for Extended Health Care Services (starting in 1977), escalated to the current year bv the growth in the Canadian et m*omv, as measured by GNP per capita. Beginning in 1986, the rate of escalation was reduced to twvo percentage points below the GNP escalator. The 1989 federal budget reduced the rate of escalation to three percentage points below the GNP esca- lator. However, this was superseded by the Expenditure Control Plan. As part of the Expenditure Control Plan, from 1900-91 to 1994-95, the per capita entitlement is frozen at its 1989-90 level. Next, the current per capita values to provinces of tihe tax transfer (13.5 points of personal income tax and one point of corporate income tax) is calculated, along with the equalization associated with it. This amount i. paidci to provinces under- the equalization program. Finally, the equalized tax transfer for each prov ince is subtracted from the total entitlement per capita and the remainder (times population) is paid to each province in cash. Under an arrangement offered to all provinces in 1964, Quebec chose to receive an additional tax transfer (8.5 points (if personal income tax) in respect of hospital insurance. As shown above, the current value of this amount is deducted froin Quebec's total entitlement to determine its cash payment. Source: Canada (1994). Table 10-5. Relevant Data on Equalization by Revenue Source, 1994-95 Preliminary Payout ratio Equalization entitlement (S000'000) (total Revenues equaliz- sulbject to tion per equali- Total $100 of zation receiving reve- R~evenue source ($000'000) Nfld. P.El. N.S. N.B. Que. Ont. Man. Sask. Alta. B.C. provinces ries) 1. Personalincornetaxes 39,811 349.0 65.1 297.2 327.9 1,603.3 -2,053.0 361.4 427.9 -46.5 -339.5 3,431.7 8.62 2. Businessincomnetaxes 5,517 65.7 11.2 103.4 72.7 -105.4 -80.6 83.8 45.2 -206.7 57.1 278.6 5.05 3. Capital tax revenues 3,211 32.1 9.6 39.3 28.6 102.6 -211.0 23.6 30.3 -133.6 54.6 266.0 8.28 4. Generalandnuscellaneo-ilestaxes 21,027 96.4 17.4 57.8 79.3 437.6 -273.8 107.2 104.3 -423.7 -375.4 900.1 4.28 5. Tobacco taxes 1,696 11.5 -0.2 -18.8 -0.7 -65.6 -12.8 12.3 8.3 -8.9 57.8 -53.2 -3.14 6. Gasoline taxes 4,506 4.7 -3.7 -14.2 15.9 128.3 -81.8 -12.1 -22.7 -138.2 -11.7 64.5 1.43 7. Dieselfueltaxes 1,371 10.6 2.8 6.3 -3.8 69.1 1.3 -13.6 -46.8 -72.0 -10.2 24.8 1.81 8. NoncomnerciaCvehicle heenses 1,687 7.3 -0.3 -0.1 3.3 33.5 0.1 -5.0 2.6 -32.9 -31.3 41.5 2.46 9. Commercialvehiclelises 712 2.2 0.8 2.9 -2.2 30.0 -17.1 -1.3 -2.7 -34.4 -9.0 29.8 4.18 10. Sales ofspirits, wine andbeer 3,065 -4.1 0.8 -1.0 19.0 118.7 -40.5 -0.3 7.6 -36.0 -85.6 140.8 4.59 11. Health inurance premniums 1,268 2.8 -1.5 -0.6 0 2.9 -0.6 -2.7 2.7 -1.1 -2.3 3.6 0.28 12. Race track taxes 138 2.9 0.2 3.8 3.3 15.0 -25.0 2.3 4.3 -3.3 3.4 31.8 23.01 13. Forestry revenues 1,216 15.1 5.6 34.6 -4.7 126.7 308.9 36.9 17.7 5.1 -490.3 232.1 19.08 14. Newoilrevenues 441 2.0 0.5 3.2 2.6 24.8 35.0 1.6 -57.3 -350.2 -4.1 -22.7 -5.14 15. Oldoilrevenues 278 0.9 0.2 1.5 1.2 11.5 16.3 0 -22.4 -235.8 -5.3 -7.1 -2.57 16. Heavy oil revenues 57 0.9 0.2 1.5 1.2 11.6 17.4 1.8 -36.6 -14.4 5.7 -19.3 -33.98 17. Mined oil revenues 51 0 0 0 0 0 0 0 0 -50.6 0 0 - 18. Natural gas revenues - domestic sales 770 4.2 1.0 6.7 5.5 52.6 75.1 8.1 -57.9 -577.7 -77.9 20.1 2.61 19. Naturalgasrevenues-exportsales 769 1.9 0.4 3.0 2.4 23.5 35.4 3.6 3.2 -683.1 -65.8 38.2 4.96 20. Sale of mineral rights on oil and gas lands 343 1.5 0.3 2.4 2.0 18.9 27.4 2.6 -25.0 -273.2 -23.9 2.7 -0.80 21. Otheroilandgasrevenues 253 1.1 0.2 1.7 1.4 13.1 19.1 1.3 -20.5 -204.8 -13.0 -1.7 -0.66 22. Mineralsotherthanpotash 255 -8.9 1.1 3.8 -5.2 16.5 21.9 -5.7 -14.4 3.8 -18.3 -13.3 -5.24 23. Potash 44 0.9 0.2 1.4 -5.3 11.3 17.0 1.7 -35.6 4.2 5.6 -25.3 -58.04 24. Water power rentals 484 -9.4 2.5 16.3 9.7 -114.9 145.6 -19.6 14.3 49.1 -25.4 -101.1 -20.89 25. Insurancepremium revenues 949 8.0 1.1 7.4 2.5 20.9 -27.3 6.8 5.5 1.2 -6.0 52.3 5.51 26. Payroll taxes 6,077 42.3 8.3 35.7 36.0 157.9 -253.3 35.6 54.3 -26.1 5.5 370.1 6.09 27. Property taxes 29,629 225.9 40.4 229.6 248.1 823.7 -832.5 218.7 182.4 -431.1 -392.3 1,968.9 6.65 28. Lottery revenues 2,035 -4.9 3.5 14.6 17.9 -84.7 113.0 19.9 20.8 35.8 -69.1 -12.9 -0.63 29. Miscellaneous own-service revenues 14,088 97.7 17.9 88.0 93.9 377.5 -436.7 95.9 94.8 -178.8 -131.5 865.7 6.14 30. Federalsharedrevenues 127 -0.5 0.2 2.2 2.5 3.7 -12.4 -0.9 2.2 -1.6 7.3 9.3 7.36 Total 141,874 959.9 185.9 931.0 923.3 3,864.9 -3,524.9 964.0 686.0 -4,065.7 -1,990.7 8,515.8 6.00 Note: Totals may not add because of rounding. The above revenue sources include subsources in the case of spirits, wine and beer (line 10), minerals other than potash (line 22) and federal shared revenues (line 30). Where there are subsources, separate tax bases are used for each subsource. Negative payout ratios occur for a revenue source where the equalization receiving provinces as a whole have a higher per capita tax base than the five provinces making up the program standard. Source: Canada, Department of Finance (1994). Intergouernmental Fiscal Relations in Canada: An Overview 247 * Portability: Persons must remain covered while temporarily absent from their province and for three months by original province while relocating to a new province in Canada. * Public insurance: Health plans must be insured by a nonprofit public agency. The act requires the federal government to withold funds if any of the above condi- tions are not met. Canada Assistance Plan Provincial governments in Canada provide social assistance to the needy as a "last resort" when benefits from other programs such as Unemployment Insurance, Old Age Security, and public and private pensions are either exhausted or not available. Through the Can- ada Assistance Plan established in 1966, the Government of Canada shares on a 50:50 ba- sis all eligible provincial expenses for social welfare. Figure 10-6 shows the historical growth of these transfers as well as 1994-95 per capita transfers by province. The primary objectives of this program are to provide federal support for the provision by provinces of (a) adequate rates of social assistance and institutional care for persons in need; and (b) welfare services that have as their object the lessening, removal, or prevention of the causes and effects of poverty, child neglect, or dependence on public assistance. Federal-provincial agreements under the CAP bind the provinces to provide assis- tance to needy Canadians on the basis of the fact of need rather than its cause. The prov- inces also agreed not to restrict the access to CAP assistance by nonresidents. The CAP beneficiaries include the disabled, single mothers, children, the aged, and the unemployed. In addition, child care, rehabilitation, and home-making services are provided. The provinces enjoy complete autonomy in design and administration of these cost-shared programs within their jurisdiction. The federal government simply reim- burses its share of the eligible expenses. Major Unconditional Transfers Equalization transfers to have-not provinces and formula financing of territorial govem- ments represent two important federal unconditional transfer programs to subnational governents. The Canadian Fiscal Equalization Program Fiscal equalization is a federal program of annual, unconditional payments to provinces with a substandard capacity to finance public services for their citizens. This program was implemented in 1957 upon the recommendations of the Rowell-Sirois Commission of 1941. Successive governments have stated that the purpose of equalization is to make it possible for all provinces to provide reasonably comparable levels of public services at reasonably comparable levels of taxation. A commitment to a federal program with this goal is set out in the Canadian Constitution. Specifically, subsection 36(2) of the Consti- tution Act of 1982 states: 248 Macroeconomic Management and Fiscal Decenttralizntioni Figure 10-6. Transfers Under the Canada Assistance Plan a. CAP Transfers, 1983-84 to 1994-95 Billions of dollars 10 _ ___ 9 9 ____ 8 7 6 5 3 2 2 0 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 b. CAP transfers, 1994-95 Dollars per capita 400 - __ __ _ _ _ 350 300 250 200 150 100 50 0 Que. Nfld. PElI. N.S. N.B. Man. B.C. Alta. Sask. Ont. Intergovernmental Fiscal Relations in Canada: An Overview 249 Parliament and the Government of Canada are committed to the principle of making equalization payments to ensure that provincial governments have sufficient revenues to provide reasonably comparable levels of public services at reasonably comparable levels of taxaticn. The present system of fiscal equalization in Canada has the following characteristics: * Use of the representative tax system approach to the measurement of fiscal ca- pacity. * Comprehensive coverge of provincial-local revenues from own sources extend- ing to virtually the whole range of provincial-local revenues from own sources plus federal revenues from taxation shared with provinces on a point-of-origin basis. * Use of a "five-province standard" consisting of British Columbia, Saskatchewan, Ontario, Manitoba, and Quebec. * Application of national average rates of taxation to revenue bases. * Use of total provincial population as a means for comparing provinces of differ- ent size, and as a proxy for the relative needs for and costs of public services in each province. The Equalization Program also has two special features. First, it has a sliding scale floor to protect individual provinces against large year-over-year decreases in their equaliza- tion entitlements. Second, the program has a ceiling provisior whereby total equalization entitlements may not increase more rapidly from a base year than GNP increases from the same base year. Mathematically, E a = ((R1/B10 x B5/P5) - (R10/B10 x Ba/Pa)) x Pa where Ea = Equalization entitlement of Province a from revenue source I (i =1 to 30). Rlo = Total revenues obtained by ten provinces Blo = Tax base in ten provinces B5 = Tax base in five provinces (standard) P5 = Population of five provinces Ba = Tax base in Province a Pa = Population in Province a Equalization per capita for revenue source (i): = (National average tax rate)i [(Per capita base in 5 provinces)i - (Province's own base per capita)i I = (Per capita potential revenue in 5 province s) - (Per capita standardized revenue in Province a) A province may have equalization entitlements that are positive from some revenue sources and negative from other revenue sources. These are added from all sources, and if the overall sum is positive, the province receives a transfer from the federal govern- ment of the equivalent amount. A negative sum is simply ignored. Table 10-6 and Figure 10-7 illustrate how this program worked for the fiscal year 1994-95. Figure 10-8 shows 250 Macroeconomic Management and Fiscal Decentralization Figure 10-7. Equalization Standard 1994-95 Dollars per capita 7,000 - 6,000 - Five-province standard $4,857 . 5,000 4,000 3,000 2,000 Fica capacity per capitas 2,000 - _ l 11 l 1 | (prov~incial ability to rais e revenues) E 1,000 M 1l 0 Nfld. PE.I. N.B. N.S. Man. Sask. Que. Ont. B.C. Alta. Note: By levying average rates of tax, any province, together with its local governments, vwill be able to derive revenues of at least $4,857 per capita (equal to $19,428 for a family of four) with which to finance public services. Table 10-6. How Equalization Works: Summary of Calculation, 1994-95 (1) (2) (3) (4) (5) Estimated revenues of prozv- Prozvinice's per Proz'ince's Provintce's per intcefrom applying the Rep- capita sihortfall equalization capita revenue- resentative Tax from formula entitlement raising capacity SYSteni standard of (Col 3, if after equalization (a) (b) $4,874 positive, X (Col. 2(b)+ Col 3 Fotal Per capita Col. 1) if positive) Prozinice PWItion ($00,00 _ ($) ___ (S) _ $000,000)___ Newfoundland 582,600 1,870 3,209 1,647 960 4,857 Prince Edward 132,200 456 .k,451 1.406 186 4,857 New Brunswick 755,500 2,746 3,635 1,222 923 4,857 Nova Scotia 928,200 3,577 3,8514 1,003 931 4,857 Manitoba 1,119,300 4,472 3,996 861 964 4,857 Saskatchewan 1,001,900 4,179 4,171 685 687 4,857 Quebec 7,262,500 31,408 4,325 532 3,865 4,857 Ontario 10,907,700 56,503 5,180 (-323) 0 5,18() British Columbia 3,594,100 19.447 5,411 (-554) 0 5,411 Alberta 2,707,400 17,215 6,359 (-1,502) 0 6,359 Five-province standard 4,857 8,516 Note: Provinces are listed in ascending order of per capita revenue-raising capacity as shown in col. 2(b). Source: Canada, Department of Finance (1994). Intergovernmental Fiscal Relations in Canada: An Overview 251 Table 10-7. Calculation of Equalization Entitlement of Newfoundland for General Sales Tax Revenue Source, 1993-94 Step 1: Revenues to be equalized 1. Total revenues of 10 provinces from general sales taxes, 1993-94 $19,880 million Step 2: Newfoundland standardized revenues from general sales taxes at national average tax rate 2. Newfoundland standardized tax base for general sales tax revenue source (sales of goods and services in province which are typically subject to general sales tax, 1993-94) $3,643.3 million 3. National average tax rate for general sales tax = total revenues of all provinces from general sales tax total standardized tax base of all provinces for general sales taxes 8.4175 percent $19,880 million $236,175 million 4. Newfoundland standardized revenues from applying national average tax rate to its standardized tax base line 2 x line 3 =$3,643.3 million x 0.084175 $306,677,000 Step 3: Comparison of Newfoundland's standardized revenues from the general sales tax base with revenues it would derive if it had a tax base as productive as the average for the 5 standard provinces 5. Newfoundland population 581,100 6. Newfoundland standardized sales tax revenues per capita = line 4 = $306,677,000 line 5 = 581,100 = $527.75 7. Standardized tax base of the 5 standard provinces for general sales tax = $192,976 million 8. Standardized revenues of the 5 standard provinces for general sales tax = line 7 x line 3 = $192,976 million x 0.084175 = $16,244 million 9. Population of the 5 standard provinces 23,583,700 10. Standardized sales tax revenues per capita of 5 standard provinces = line 8 = $16,244 million line 9 = 23,583,700 = $688.77 11. Newfoundland's equalization for general sales tax = Newfoundland population x per capita shortfall from 5 province sta ndard = line 5 x (line 10 - line 6) = 581,100 x ($688.77 - 527.75) = 581,100 x $161.02 = $93,568,000 Note: The general sales tax revenue source in equalization also includes miscellaneous sales taxes (levied by some provinces on hotel accommodation, meals, cable TV, amusements, and telecommunications). These are reflected in both the revenues and the tax base and, because they tend to be levied at higher rates than general sales taxes, have the effect of slightly raising the n a- tional average tax rate shown on line 3 from what it would be in respect of general sales taxes only. This table does not show a ceiling impact (although a ceiling did apply in 1993-94) as this potentially final step in the determination of equalization enti- tlements is applied in respect of overall entitlements and not on a tax-by-tax basis. However, in any year when the ceiling ap- plies, it could be pro-rated over the various taxes subject to equalization. Thus, if general sales taxes accounted for, say, 11 percent of total pre-ceiling equalization as of 1993-94, then 11 percent of any ceilmg impact in that year could be apportioned to the general sales tax. Source: Canada, Department of Finance (1994). 252 Macroeconomic Management and Fiscal Decentralization Figure 10-8. Equalization Transfers, by Province in 1994 and over time 8a. Equalization Transfers, 1994-95 Dollars per capita 2,000 1,500 1,000 500 0 Nfld. PE.I. N.B. N.S. Man. Sask. Que. 8b. Provincial Shares of Equalization, 1994-95 cPrince Edward Island Saskatchewan 2. 3% Manitoba Quebec ... ..... 46.2% Nova Scotia New Brunswick Newfoundland 1 1.4% !itcr,g)o?ernmnc,,tal Fiscal Relationts in Canada: Ant Overview 253 Figure 10-8. (continlued) 8c. Equalization Transfers, 1983-84 to 1994-95 Billions of dollars 10 8 6 4 2 0 1983 1984 198S 986 1987 988 1989 1990 1991 1992 993 1994 Table 10-8. Other Federal-Provincial Transfers, 1994-95 Tranisfe'rs Millions of diollars Territorial finan-icing 1 143 Infrastructure 687 Gross revenue inIsurance plani 487 Grants in lieu of property taxes 435 Other- 351 Miscellaneous hiealth and welfare 310 Transportation 260 Official languages in education 255 Public utilities Income tax transfer 237 YoLung offendlers 158 Preferred share dinidend taxes 150 justice 93 Statutorv subsidties 40 Total 4,656 Source: Canada, Department of Finance (5994). 254 Macroeconomic Managemenit and Fiscal Decentralization equalization payments by province as well as the historical record of such transfers. Table 10-7 provides an illustrative example of steps involved in calculating equalization pay- ments for a specific revenue category. The Canadian Fiscal Equalization Program has endured and is often referred to as the glue that holds the federation together" (Courchene 1984). It is a comprehensive per capita fiscal capacity equalization program, and it considers all provincial-local revenues. Since the program uses fiscal capacity as a criterion, it allows the federal government to monitor fiscal positions of subnational governments on a timely basis. The program is largely free of grantsmanship and strategy, and both the size of the pool and allocations are determined objectivelv. The program, however, implicitly assumes that per capita expenditure needs are the same across all provinces. It is a federal program, and there- fore it constrains the federal government's fiscal position. The program also neglects other federal transfers in calculating equalization transfers. Finally, it separates taxing and spending decisions in a major way in Atlantic Canada resulting in reduced account- ability for provincial-local governments in that region. See Shah (1994a) for a critique of this program and the effect on provincial allocations under a comprehensive equalization program. See Shah (1994b) for a discussion of the design of equalization programs; see also Boadway and Hobson (1993). Formula Financing of Territorial Governments The governments of the Northwest Territories and Yukon are not eligible to receive equalization payments. Instead, they receive formula-based federal assistance. This grant program was initiated in 1985-86. The federal grant is based on a measure of the differ- ence between expenditure needs of the territories and their revenue means-a fiscal-gap- filling approach. Expenditure needs are based on 1982-83 expenditures escalated annu- ally, and reveneue means are based upon fiscal capacity and other federal transfers re- ceived. In 1994-95, $1.2 billion in federal assistance was provided under this program (see Table 10-8). Conclusion In this chapter, I have provided a brief overview of the Canadian federal system to serve as a study guide for comparative federal systems. While I have not presented an analyti- cal review of this system, it is quite obvious that the Canadian system would receive plaudits for its simplicity and design, objectivity of the criteria used, and consistency of the design with objectives sought. Thus, a familiarity with this system should be of bene- fit to any one engaged in serious deliberations on institutions and mechanisms to secure constitutional objectives under a federal system of government. References Boadway, Robin, and Paul Hobson. 1993. Intergovernmental Fiscal Relations in Canada. Canadian Tax Paper No. 96. Toronto, Canada: Canadian Tax Foundation. Canada, Department of Finance. 1994. Federal Transfers to the Provinces. Ottawa, Canada. Courchene, Thomas. 1984. Equialization Paymenets. Toronto, Canada: Ontario Economic Council. Intergovernmental Fiscal Relations in Canada: An Overview 255 Ruggeri, G.C., R. Howard, G.K. Robertson, and D. Van Wart. 1993. Imbalances in the Canadian Fed- eral System. Presented at the Canadian Economic Association Meetings, Atlanta, Georgia, June 1993. Shah, Anwar. 1994a. A Fiscal Need Approachi to Equalization Transfers in a Decentralized Federation. Policy Research Working Paper 1289. Washington, DC: The World Bank. April. . 1994b. The Reform of Inter',overnmental Fiscal Relations in Developing and Emerging Market Economies. World Bank Policy Research Series 23. Washington, D.C.: The World Bank. June. Distributors of World Bank Publications Pnces and credit terms vary from country to country Consult your local distributor before placing an order ALBANIA CYPRUS IRELAND PAKISTAN LUbnrr Intemrnaonal AEDOS Adnon Lid Conf of Appked Researt Gowernment Suppies Agency MirZa Book Agency Consel de Cent 391 Pedat Renhepi Str Cyprus College Oig9 nt iSolithair o5. Shahah-e-Quaide-Azam 08009 Barelona Pall. Shk 1, Ap. 4 6. Drognes Stnot Engori 4-S Hwcourt Roac P.O. Box No. 729 Tel: (31488-3009 Tirana PO Box2006 Dubin 2 Lahore 54000 Fax: (3) 487-7659 Tel (42) 274 19 22172 Nicosia Tfel (1)481-3111 Tel: 142) 7353601 Fax (42)274 19 Tel: 244-1730 Fax: (1)475-2670 Fax: (42)7585283 Fax 248-2051 SRI LANKA, THE MALDIVES ARGENTINA ISRAEL Oxford Uninersiy Press Lake House Bookshop CarlosHrsch, SRL CZECHREPUBLIC YeeostaLiraitu. Ltd 5BangabreTwpn P.O .ox244 Galena Guemes Nabonal Infomiation Centar P.O. Bo 58055 Shirar Fayssl 100, Sir Chigampelam A Florida 165, 4th Floor poxensa Konmrkitka S Tel Avie 81560 PO Boo 13033 GamdinarMawvatha oic 4531485 CS -113 57 Pragu 1 Tel: (3) 5285-397 Karaohi-75350 COlOmfo 2 1333 Buenos Aires Tel: (2)2422-9433 Fan: (3)5285-397 Tel (21)446307 Tel. (1)32105 Tel (1)331-2391 Fax (2) 2422-1484 Fax (21)454-7840 Fax (1)432104 Fax (1) 331-1787 R.O Y lnermabonal DENMARK PO Box 13058 PERU SWEDEN Oticina del Ubro Intemanonal SamnundsULteratur Tal Aniv 61130 EdtonI Deseanollo SA Fntzes Customer Service Al Cordoba 1877 Rosenoems Alli 11 Tel (3)5481423 Apartado 3824 Regangsgaton 12 1120 Buenos Aires DK-1970 Fedenksberg C Fax (3) 54814442 Lima 1 S-106 47 Stockholm Tel (1)815-8354 Tel: (31)-351942 Tel (14)285380 Tel: (8)6909090 Fax- (1)815-8156 Fax (310357822 PaleatiniomAutrlontyM,ddie East Fax (14)288628 Fax (8)21 4777 lode ntolomahon Sernices AUSTRALIA, FIJI. PAPUA ECUADOR POB. 19502 Jerusalem PHILIPPINES errnergren-Wiliams AS NEW GUINEA, SOLOMON Feoultad LAstoamencana 4de Cienas Tel: (2) 271219 Intemabonal Booksoune Carer Inc P 0. Boo 1305 ISLANDS. VANUATU. AND Socieiles Sunie 720. Ctyland 10 S- 17125 Solna WESTERN SAMOA FLASCO-SEDE Ecoador ItALY Condorinium Toner 2 Tel: (8)705-97-50 GA Inlormat,onServices Crall Ulpiano Paz I18yAv. PaDna LiUws Commissiona SansoniaSPA H.V doaI Costa, omar Valem St Fax: (8)27-00-71 648 Whitehorse Road Quato. Ecueaor Vla Dica Di Calabna, 1/1 Makab. Mario Manila M,tcham3132 Tel: (2)542714 542718: 528200 CaselaPostaieSS2 Tel (2)817-9678 SWITZERLAND V,ctona Fax: (2) 56 139 50125 Firoze Fax (2) 817-1741 Libreine Payot Tl. (3)398734411 Tel (55)845-415 Senvice lnsltitionnel F.u (3) 873 5679 EGYPT, ARAB REPUBLIC OF Fax. (55)841-257 POLAND Cliies-as-Montbenon 30 Al Ahntm Intemnalonal Publishing Service 1002 Lausanne AUSTRIA AlGaLeaStreet JAMAICA Ul Piekna3137 Tel (021)320C2511 Gepid and C. Cailm ln Randle Publishers Ltd 00-577Warzawa Fax (0211-320-2514 Graben 31 Tel. (2) 578-6083 206 Old Hope Road Tel (216286-089 A-1011 Wen Fax, (2)578-832 Kingstons Fax (2) 821-7255 Van Diemen Ediltions Techniques Tel (1) 533-50-14-0 Tel BD0-927-2085 Ch. de Lacuez 41 Fan (1) 512-47-31-29 The Middle East Orseamer Fax: 809-77-0243 PORTUGAL CH1807 Blonay 41. Sherif Street Li,r,naPonugal Tel: (02119432873 BANGLADESH Caro JAPAN Rue Do Camro 70-74 Fax (021) 943 3605 Mamo Industnes Development Tel: (2)393-9732 Eastem Book Service 1200 Lisbon Assislance Soccety (MIDAS) Fax (2) 393-9732 Hongo 3-Chome BunkyP-ku 113 Tel (1) 347-4982 TANZANIA House S. Road 16 Tokyo Fax (1) 347-0264 Oxford Uni-rsiy Press Dnanmondi R/Area FINLAND Tel (03) 3818-0881 Maktaba Sireet Dnaka1209 Akateemmen Kirakauppa Fax (03138180864 ROMANIA PO Box 5299 Te (2)326427 PO Box23 Compmni De Libam Bucurest SA. Dares Salaam Fax 12)811188 FIN-00371 Heisminn KENYA Sto L pscani no 26. sector3 Tel: (51) 29209 Tel (0) 12141 AlncaBookSeIn,ce(E.A.)Ltd. Bucharest Faa(51t46822 BELGIUM Fax (0)1214441 OuaranHouse,MlangarrS9neet Tel (1)813 9645 Jean D. Lannoy PD Box45245 Fax (1)3124000 THAILAND A, dJ Roi 202 FRANCE Naimobi Central Books DstribuLon 1060 Brussels World Bank Publicaxons Tel (2) 23641 RUSSIAN FEDERATION 306 Silom Road Te (2) 538-5169 66, anesue d lena Fax: (21330272 (sOatelsIvo