Policy Research WORKING PAPERS Public Economics Country Economics Department The World Bank February 1992 WPS 863 Intergovernmental Fiscal Relations in China Roy Bahl and Christine Wallich Is there a "best" way to divide fiscal responsibilities between China's central and local governments in China's three-tiered fiscal system? Policy Research Worcing Papers disseminate the findings of work in progrss and encourage the exchange of ideas among Bank staff and all others interested in developrnentissues.Tbesepapers.disLributed by thc Research Advisory Staff,carry thenamnesoftheauthors,reflect onlytheirviews,andshouldbeused and cited accordingly. befindings, interpretations, andconclusions aretheauthors'own.Tneyshould not be attributed to the Wodd Bank, its Board of Directors, its management, or any of its member countries. Policy Research - Public Economics WPS 863 This paper-a product of the Public Economics Division, Country Economics Department -is part of a larger effort in the department to explorc interg(vemrnmcntal fiscal rclations in developing and formerly socialist economies. Please contact Ann Bhalla, room NIO-053, extension 37699 (February 1992, 58 pages). The choice of the "right" fiscal relationship Centralizing the fiscal system nevertheless between central, provincial, and local govem- reduces the potcntial for vesting more budgetary ments depends on how a govemment weighs the decisionmaking powers in iocal govemments benefits of decentralized economic development and can crode local and provincial governments' policies against the costs of having less effective incentives for raising revenues, another goal of central fiscal management. system reform. Moreover, there are major problems with introducing fiscal centralization in Three strong forces justify more fiscal a country with a heterogeneous population of I centralization in China's highly decentralized billion and relatively little tnidition of central fiscal system at the present time: govemment fiscal administration. * Bouts of inflation and recurrent fiscal Bahl and Wallich conclude that a reformed deficits can be seen as calling for more central system ol intergovernmental finance must meet control over the budget. the center's necds for stabilization and the provinces' necds lbr revenue and equalized - Reform of an economic system relies spending capacity. They argue that such equal- heavily on the use of tax policy as an allocative ization should be based on objective indicators instrument to influence economic decisions. of need and that a formula-based grant system Local control of the implementation of the tax best meets this latter objectivc. A rcformed system can and probably has compromised some system must also underpin price and enterprise objectives of the central government's tax reform - and should be designed so as not to policy. Gaining tighter control over the revenue require major rccalibration or adjustments while system will probably require reducing if not such rcforms are taking place. eliminating local govemment discretion in providing special tax concessions. Bahl and Wallich also conclude that reformn of the rclationishlip of central and local govem- * If the center wants to movc ahead with price ments should be supplemcnted by an improved reform and to encourage entcrprise reform, it system of Financing local capital expenditures needs a more centrally controlled revenuc through borrowing, a system of benefit charges, sharing or assignment system that reduces the and improved financial planning and tax admin- dislocating effects of such reforms. istration. Thc Policy Research Working Papcr Series disseminates the findings of work inder way in the Bank. Anobjective of the series is to get these fuidings out quickly, even if presenations are less than fully polished. The findings, interpretations, and conclusions in tlJ.se papers do not necessarily represent official Bank policy. Produced by the Policy Research t)isseniination Center CHINA CENTRAL-LOCAL FISCAL RELATIONS Table of Contents Page No. Introduction.. .1 China's Fiscal Reforme a n d Revenue Trends 2 A. THE PRESENT SYSTEM OF SUBNATIONAL FINANCE IN CHINA . . . . . . . . . 4 Provincial Government Finances . . . . . . . . . . . . . . . . . 5 Budgeting and Financial Planning. 8 Tax Administration . . .. 8 Provincial and Local Revenues .9 Intergovernmental Fiscal Relations . . . . . . . . . . . . . . 11 B. PROBLEMS AND ISSUES IN CENTER/LOCAL FINANCES IN CHINA . . . . . . . . 16 Compatibility with System Reforms. . . . . . . . . . . . . . . . 16 The Macroeconomic Perspective . . . . . . . . . . . . . . . . . . 18 Revenue Adequacy . . . . . . . . . . . . . . . . . . . . . . . . 19 Incentives for Tax Avoidance ... . . . . . . . . . . . . . . . 23 Provincial Tax Effort . . . . . . . . . . . . . . . . . . . . . . 23 Incentives for Infrastructure Development . . . . . . . . . . . . 23 Equalization Properties. . . . . . . . . . . . . . . . . . . . . 25 C. OPTIONS FOR REFORM. . . . . . . . . . . . . . . . . . . . . . . . . . 27 Some Alternativesn.ative .. . . 27 Principles of Tax Tax Sharing and Tax Assignment . . . . . . . . 30 Considerations in Weighing the Options . . . . . . . . . . . . . 32 Options: 33 (i) Tax Reassignment: A Centralized Version . . 33 (ii) Tax Reassignment: Decentralized Version ... . . . . 35 (iii) The "Local Surcharge" or Tax Base Sharing Option . . . 36 (iv) Higher Provincial Retention Rates . . . . . . . . . 38 (v) Lower Provincial Retention Rate. . . . . . . . . . . . 38 (vi) Reforming the Present System . . . . . . . . . . . . . 38 Other Essentials for Center-Local Fiscal Reform ... . 40 A System of Transfers . . . . . . . . . . . . . . . 40 Borrowing Powers . . . . . . . . . . . . . . . . . . . . . . . 43 Benefit Charges . . . . . . . . . . . . . . . . . . . . . . . 43 Tax Administration . . . . . . . . . . . . . . . . . . . . 44 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Valuable comments from Peter Harrold, Ram Agarwala, Raja Chelliah, and Anwar Shah are gratefully acknowledged. Errors and omissions remain those of the authors. TABLES IN TEXT 1. China: Developments in Government Revenue - Consolidated Government 1978-89 . . . . . . . . . . . . . . . . . . . . . . . . 4 2. Tax Collection and Expenditure of Central and Local Governments, 1980-89 ... . 6 3a. China: Revenue-Sharing System Between the Central and Provincial Governments, 1985-1987 . . . . . . . . . . . . . . . . . . . . 13 3b. China: Revenue-sharing System Between the Central and Provincial Governments, 1988-1990 . . . . . . . . . . . . . . . . . . . . 14 4. Chinat Price Ccntrolled Products: Shares of GVIO of the Largest Producing Provinces in 1989. . . . . . . . . . . . . . . 17 5. Tax Collection and Expenditure ................ . 21 BOXES IN TEXT 1. Principles of Revenue Allocation and Tax Sharing 1985 Reforms 10 2. Provincial Tax Effort . . . . . . . . . . . . . . . . . . . . . . 24 3. Equalization Properties of China's Provincial Contracting . . .. 26 4. Fiscal Federalisms and Centralized Systems . . . . . . . . . . . 28 5. Principles of Tax Assignment . . . . . . . . . . . . . . . . . . 32 6. Profit and Sales Tax Sharing Under a Base-Sharing Model . . . . . 36 7. Weighing the Six Options in China ............... . 41 8. Formula Grants ........................ . 43 CHARTS IN TEXT 1. Government Structure in China .... . . . . . ...5... . . S 2. China: Ratio of Revenue to GNP .... . . . ....... . . 20a 3. China: Ratio of Revenue to GNP (Adjusted) . . . . . . . . . . 20a 4. China: Ratio of Expenditure to GNP . . . . . . . . . . . . . . 20b 5. China: Budget Balance: Central and Local . . . . . . . . . . . 20b ANNEXES 1. Procyclical Effects of Contracting with Provinces . . . . . . . . 46 2. Counterequalizing Properties of Provincial Tax Contracting . . . 48 STATISTICAL ANNEX Table 1: Tax Collection and Expenditure Table 2a: Sharing Arrangements in 1987 by Type of Tax Table 2b: Sharing Arrangements in 1988 by Type of Tax Table 3: Revenue-Sharing System Between the Central and Local Governments, 1985-87 Table 4: Revenue and Expenditure Disparities Among Provinces Table 5: Budgetary Expenditure by Province Table 6: Percent Increase in Revenues and Expenditures, by Province for 1983-86 Table 7: Changes in the Ratio of Expenditures to Collections: for 1983-86 by Province Table 8: OLS Regression Results for Chinese Revenues and Expenditures Against Selected Independent Variables: By Province BIBLIOGRAPHY CHINA CENTRAL-LOCAL FISCAL RELATIONS Introductlon 1. The purpoee of this paper is to analyze options for reform to the system of subnational finance in China--the central-local revenue-sharing system-- with a focus, inter alia, on the support to system reform, macrostabilization, equalization and allocative effects. 2. Both the government and outside observers have come to realize that China's system of public financing requireo change in order to better support the system reform. Under the previous system of strict central planning, the objectives of fiscal policy were limited: its allocative function was to administratively redistribute resources--derived mostly from enterprise surpluses--through the budget to enterprises and government units; its role in stabilization consisted mostly of expenditure or investment cutbacks to reduce aggregate demand pressures. Under the system reform, tax policy plays a role in both mobilizing resources for government spending purposes and for indirectly controlling the economic decisions of enterprises.l/ 3. Matters such as central government tax reform are usually a concern of the national treasury. This is not the case in China. Perhaps more than in any other country in the world, China's national tax structure and system of intergovernmental fiscal relations are linked. The success of any central government tax reform will be determined by the policy reactions of local governments, and by local implementation of central policy. While uniform national tax laws are laid down in Beijing, subnational governments are responsible for tax administration, share in revenue collections, and may make policy by giving tax incentives to enterprises. 4. By the same token, any reforri in the system of intergovernmental relations must be evaluated in a context of the objectives of the system reform. Enterprise and price reforms affect the tax base available for sharing in each province, revenues mobilized for each level of government must reach some prescribed target, and fiscal disparities among provinces must stay within some prescribed range. Finally, and underneath much of the clamor for a reform of the intergovernmental system, is to move government fiscal decisions "closer to the people." There is no question but that this goal of the systen, reform argues for a stronger local government sector. S. The choice the "right" central-local fiscal relationship will depend on how the govf ..ient weighs the benefits of decentralized economic develop- ment policies a ainst the costs of having less effective central fiscal management. There are three strong forces justifying more fiscal centralization in China at the present time: (a) inflation and recurrent deficits are interpreted by some government policymakers as calling for more central control over the budget; (b) if the center wants to move quickly with price reforms and to encourage enterprise reform, it will need a centrally controlled revenue system which reduces the dislocating effects; (c) the economic svstem reform relies heavily on the use of tax policy as an allocative instrument to influence economic decisions, and local control over the implementation of the tax system can (and probably has) compromised some of the objectives of central government tax policy. In order to gain tighter control over the revenue system, it will be necessary to reduce if not eliminate local government discretion in providing special tax concessions. To centralize the fiscal system, however, sacrifices possibilities of vestilug 1/ See Revenue Mobilization and Tax Policy, World Bank Red Cover Report, 1989, for a fuller treatment of tax reforms in China, and Finance and Investment, World Bank, Red Cover Report, 1988, for a description of China's fiscal system in the 1979-83 period. more budgetary decision-making powers in local governments and can erode the revenue-raising incentives of provincial and local governments, which are also goals of system reform. Moreover, there are major problems with Łntroducing fiscal centralization in a cous;try with a heterogeneous population of one billion and relatively little tradition of central government fiscal administration. 6. This research is addressed to the question of fiscal decentralization in China, with a view to better understanding its present features, their consequences and possible reform options. The paper has three parts. The first outlines the present system of subnational finance in Chir.A; the second describes the impacts of this system on resource allocation and income distribution; and the third develops options and recommendations for reform. China plans such a reform following the ninth Plan period. As necessary background, the following paragraphs briefly review recent fiscal developments and the tax structure in China. China's Fiscal Reforms and Revenue Trends 7. Fiscal policy in China has evolved in the context of overall system reforms undertaken since 1979. These emphasized decentralization of economic management, with greater autonomy given to provinces and state enterprises, and allowed the development )f the non-state sector. Associated reforms were initiated in prices, the financial sector, and the foreign trade and payments system. Corresponding to the priorities implicit in enterprise and price reform, fiscal reforms focused on three main areas: (i) the reform of direct (enterprise) taxation; (ii) the reform of indirect taxes; and (iii) reform of the system of center-local fiscal relations. 8. Under the reforms of direct taxes, full profit remittance was replaced by a "profits tax," with enterprises allowed to retain a portion of their profits. The objective was to provide enterprises with incentives. Profit taxation was introduced in 1984 first as a flat rate tax (55 percent), supplemented by an "adjustment tax" designed to compensate for the impact on profitability of factors external to the enterprise, such as administered prices and distorted costs. Tax rates, therefore, varied significantly between enterprises, and gave rise to significant lack of uniformity when com- pared to tax systems in market economies. 9. The discretionary element soon took on increased importance under the "responsibility system," in which enterprises' tax liability was negotiated on a case-by-case basis through a "contract" designed to stimulate enterprise production and investment. Contracts called for a tax quota to be paid, with any profits above the quota amount taxed at a lower (sometimes zero) rate. Thus, as output and profits expand, the effective tax rate falls, and enter- prises retain a larger share of total profits earned. At end-1989, some 90 percent of state-owned enterprises were subject to contracting, according to the Chinese Ministry of Finance sources. Since the introduction of contracting, the elasticity of enterprise income taxes has been below unity and the overall tax-GDP ratio has fallen. In the longer run, however, this practice may expand the tax base and tax receipts if it is successful in promoting growth and efficiency. 10. Contracts designed with the low marginal tax rate also introduce a strong procyclical aggregate demand element into tax policy. As production, profits and the economy grow, relatively more resources are kept by the enterprises which further increase demand. (These demand pressures exist whether the profits are reinvested or paid out as wages.) In contrast with other countries where the tax system is structured so as to restrain aggregate demand when the economy grows excessively rapidly, the stabilization effects of contracting in China could be said to be perverse. This would be less ser- ious if the authorities had other effective macroeconomic instruments to con- -3- trol aggregate demand. The contracting system in its widely applied contract management responsibility system (CMRS) version has also reduced the central government's flexibility to introduce new direct tax policies in a timely manner. This occurs because tax contracts--fixed for us to five years--aet a nominal revenue level. Even if the contract itself were fixed for a shorter period (this could be undesirable in that it would increase uncertainty for enterprises), the fact is that most contracts are negotiated by the provincial and local governments which have no macroeconomic responsibilities, and which therefore pursue developmental, not stabilization, objectives. 11. Indirect tax reforms have increasingly replaced a cascading sales tax (product tax) with a VAT accompanied by excises and a gross receipts tax on businesses (business tax). The restructuring of taxes on production was a necessary consequence of the elimination of profit remittance. Under profit remittance, the fact that prices did not reflect costs of production was not very important. However, under profit taxation, since enterprises kept a certain proportion of their profits, and the pricing system generated consid- erable profit differentials between enterprises, this had consequences for their investment and growth prospects. This inequality could have been reduced by price reform, but instead differential product taxes were used to offset the inequities arising from distorted prices. As 'he price of a good to its purchaser was fixed, an indirect tax could be used to reduce the 'orice received by the producer and thus reduce excess profits. The main consideration in setting product tax rates in 1984 was to equalize profits (as a proportion of sales value) between products, with some modifications to encourage or discourage the production of particular goods. This approach led to a very large (over 260) number of different indirect tax rates. Ir. China, therefore, indirect taxes also carry a discretionary element, since they are used to equalize intersectoral profitability. 12. Central-local fiscal relations also underwent major reform, pari passu with the regional decentralization which has been such an i .portant aspect of overall reform. In China, local governments are responsible for collecting virtually all major taxes, with revenues "shared upward" to central government. The sharing proportions differ between provinces based on the outcomes of their negotiations with the central government. Over time the trend has been to allow provinces a larger share of the revenues they collect. The system in place from 1981-86 required a percentage (which differed by tax) of each tax to be shared with the center. More recently, contracts have been introduced which call for provinces to deliver a "tax quota" to the central government and allow above-quota collections to be retained at a higher rate, or fully retained. 13. These fiscal reforms have, in fact, had a major impact on revenue structure and trends. Direct taxes make up some 19 percent of total revenues (and 3.7 percent of GDP) while indirect taxes comprise 4S percent, or 9.3 percent of GDP. (Other tax and non-tax revenues, including trade taxes contribute 35 percent of the total--7 percent of GNP.) A notable result of the tax reforms and the decentralization of the fiscal regime has been the decline, continuous since the beginning of the reforms in 1979, in the ratio of total revenue in relation to GNP, from 34 percent in 1978 and 32 percent in 1979 to about 19.8 percent in 1989 (see Table 1). To some extent, this has been the intended consequence of policies to increase enterprise profit retention and enhance provincial autonomy. However, the declining trend also reflects the unintended consequences of the design of the reforms themselves, and the system of both provincial and enterprise contracting. 4 Table I CHINA: DEVELOPMENTS IN GOVERNMENT REVENUE - CONSOLIDATED GOVERNMENT 1978-89 a/ (In % of GOP) 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 19'89 Total revenue 31.6 29.4 29.0 27.2 27.4 26.4 26.3 24.8 22.2 20.4 19.8 Direct Taxes d/ 18.4 16.9 16.2 13.9 12.9 11.6 8.1 7.1 5.9 4.9 3.7 Profit tax (1.1) (1.0) (0.9) (0.9) (1.0) (1.3) (7.6) (6.7) (5.0) (4.5) (3.4) Profit Remittance (17.2) (15.9) (15.1) (12.9) (11.5) (9.9) (0.5) (0.4) (0.4) (0.4) (0.3) Indirect Taxes 11.4 10.8 10.9 11.1 10.2 10.1 11.2 10.8 9.6 9.2 9.2 Product tax (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (6.9) (5.6) (4.7) (3.5) (3.3) VAT (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (1.7) (2.4) (2.2) (2.8) (2.7, Business tax (0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (2.5) (2.7) (2.?) (2.9) (3.1) AgricuLturat taS (0.3) (0.2) (0.2) (0.2) (0.2) (0.1) (0.1) (0.1) (0.0) (0.1) (0.1) Taxes on int. traie 0.7 0.8 1.1 0.9 0.9 1.5 2.4 1.6 1.3 1.1 1.2 Other taxes b/ 0.3 0.2 0.2 0.5 2.4 2.0 3.4 3.3 3.2 3.1 3.3 Nontax revenue 0.9 0.8 0.7 0.8 1.0 1.1 1.1 2.0 2.2 2.1 2.5 Memo items: Extrabud. receipts: - of government 0.0 0.0 0.0 0.0 1.6 1.8 1.7 1.6 1.5 1.3 1.3 - of pubtic agencies S/ 0.0 0.0 0.0 0.0 0.0 0.1 0.3 0.2 0.3 0.2 0.2 Source: Ministry of Finance. (Note: Adjusted according to the IMF's Govermrent Finance Statistics (GFS) format, which differs from Chinese budget). aJ ConsoLidated central, provincial anm tocal govermnents b/ Includes taxes on extrabudgetary receipts, on extrabudgetary construction, and wage bonus tax. p/ Excludes enterprise retained earnings and depreciation funds. d/ Includes profit remittance 1978-84. A. THE-PRESENT SYSTEM OF SUBNATIONAL FINANCE IN CHIN 14. Despite its size and diversity, China has chosen a unitary system 'f government in which the Constitution does not expressly delineate the powers and responsibilities of the various levels of government. Yet the central, provincial and local levels of government have distinct powers and responsibilities, and in many respects the Chinese system functions as a federalism. Fiscal behavior varies enough across provinces to contravene national uniformity in China. 15. China's system of intergovernmental transfers is an important component of tota'. public financing, perhaps more so than in any other country. China's shared tax system based on sales and profits taxes makes local government revenue in principle responsive to growth in income and prices. The distribution of tax shares among provinces is changing from a broad formula- based approach with an equalizing intent to an ad hoc approach charactsrized in recent years by substantial negotiation and a variety of special purpose discretionary grants. The Chinese system does not make wide use of categorical or cost reimbursement grants to stimulate spending for particular purposes, nor does it use formula-based grants formula-based grants which take account of the population characteristics of provinces. -5- Provincial Government Finances 16. In the Chinese system, the central Syvernment has direct relationa with provincial governments. This system is rzughly described in Chart 1. All governnments within a province report direv.tly (or indirectly) to the provincial government, and carry out their duties subject to provincial zegulations. This system of vertical relationships creates a setting within the province which in principle would allow a very substantial degree of fiscal decentralization to the local government level. For example, in Zhejiang Province at the end of 1985, there were 8 provincial cities (municipalities), 66 counties, 3 country-level cities, 3 prefectures and 508 towns. Chart 1: GOVERNMENT STRUCTURE IN CHINA Central Government Province (Urban) (Rural) Municipalities F|Prefectures /4 City Urban Nonurban Proper /a Counties /a l Counties (Districts) Towns/ Towns Villages La Indicates administrative, not governmental, status. 17. In some ways, the Chinese fiscal system is as decentralized as its gov- ernmental structure. Municipal and county governments' tax bureaus directly assess and collect about 65% of all taxes. Expenditure responsibility is less decentralized in that provincial and local governments account, on average, for over 60% of total direct expenditures. Only a few countries in the world can claim as great a degree of expenditure or revenue decentralization and none can claim this degree of decentralization in tax administration.2/ 2/ The comparable ratios for the United States--which is a decentralized fiscal system by world standards--are 43% of taxes collected and 42% of expenditures made by state and local governments. The ratio of subnatio- nal to central government expenditure exceeds 75% in Denmark, Australia, Switzerland, Ital and Canada, but subnational government revenue autono- my is more limited. Roy Bahl, "The Design of Intergovernmental Transfers in Industrialized Countries," Public Budgeting and Finance, Winter 1986, -6- Table 2 shows recent trends in the share in revenues and expenditures of central ar.d subnational governments. 18. The central government speiads more than it collects. Central government expenditures, until recently, amounted to half or more of total gavernment outlays (seq Table 2). The central government'a major areas of expenditure reuponnibility are defense, foreign affai.rs and foreign aid, national universities and research, central ministries, general administration, and large investment pLojects. In addition, the central government provides net revenue transfers to poor provinces. Table 2. TAX COLLECTION AND EXPENDITURE OF CENTRAL AND LOCAL GOVERNMENTS, 1980-89 Revenue Unified system Sharing Contracting 1980 1981 1982 1983 19i 1985 1986 1987 1988 1989 Total revenues 100o,: 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 X center 19.3 20.6 22.9 29.7 34.9 37.0 39.3 35.3 36.4 37.5 X tocal 80.7 79.4 77.1 70.2 65.0 62.9 60.7 64.7 63.6 63.5 Total expenditures 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 X center 53.6 54.0 49.9 50.0 47.5 44.0 41.2 39.5 35.9 n.a. % local 46.3 46.0 50 1 50.0 52.5 56.0 58.7 60.5 64.1 n.a. Ratio of Cuttection to GNP 25.0 22.6 21.6 21.4 21.4 21.7 23.1 20.1 18.9 18.1 Subnational 19.4 18.0 16.6 15.1 14.0 13.? 13.8 13.4 12.0 11.3 Central 4.6 4.6 5.0 6.3 7.4 8.0 9.3 7.3 6.7 (6.8) Ratio of Expenditures to GNP 27.9 23.1 22.1 22.4 22.7 22.4 23.8 23.3 18.7 16.4 Subnational 14.9 12.4 11.5 11.2 11.9 12.5 139 12.9 12.0 n.a. Central 12.9 10.6 11.4 11.2 10.7 9.8 9.8 8.4 6.7 n.a. Central Goverrment Share of: Collectiors 19.3 20.6 23.0 29.8 34.9 37.0 39.3 35.3 36.1 37.5 Expenditures 53.6 54.0 49.9 49.6 47.8 43.3 41.3 39.5 35.9 n.a. Expenditure-Collection Ratio Center 309.5 267.6 222.8 172.5 141.0 115.7 104.9 115.0 99.5 n.a. Subnational 64.9 50 3 66.7 74.1 82.7 88.9 101.8 96.1 99.6 n.a. Sources: See Table 5. Notes: On Chinese definition of revenues and expenditures, which do not correspond to GFS format data reported elsewhere in this report and in tables focusing on national level taxes. 19. The responsibilities of the provincial and lower-level authorities include their own investment projects, and most, but not all, public expenditures on education and health, local administration (and tax collection), culture, science and agricultural support, including irrigation, agricultufal research, extension activities and other rural expenditures. Since the inception of the economic reforms, there has been a gradual trend towards decentralizing expenditures, with the central government's share in total outlays decliniiig and a corresponding increase in the expenditure share of local governments. Consequently the center's share of total expenditures has bren declining over the decade from about 53% of tota. expenditures to 35% Volume 6, Number 4, pp. 3-22. (Table 2). Little is known, however, about the types of expdnditure shifts which have taken place. 20. Decrees of Autonomy: Revenues. Revenue collection and exper.diture disbursement are not the only dimensions of fiscal decentralizatior, and on'i can find many areao where fiscal decisions are subject to substantial central control and direction. By comparison with most count-'es in the world, subnat4onal governments in China have little formal, or legal independence in matters of structuring thei.r 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 thosa whose rate or base the subnational govarnment can unilaterally fix--at the subnational level.3/ Moreover, the central government determires, for each province, a share of taxes to be t ned 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. 21. Even with this degree of centralization in the rules, however, subnationial governmentn have ani i4niportant 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. In particular, provinces determine the share of tax collections that will be retained by each local government. The allocation of loans to local enterprises and the distribution of grants to local governments are also determined by the provincial government. Moreover, because provinces can set the tax shar ng rates for each local government, they may also indirectly affect the rela_ive rate of tax collection or tax effort the local administration makes. Local governments have a substantial degree of autonomy to affect the level and composition of taxation, rablic service delivery and capital investment. This autonomy arises from the fact that they control tax collection and assessment with apparently a minimum of direct central or even provincial supervision. Local governments also have a subBtantial degree of freedom in awarding tax contracts to their enterpr.ins. Responsibility for implementation of the tax system is a very powerful p-iicy instrument in the hands of local government and indications are that they use it. 22. Expenditure Autonomy. Autonomy on the expeneiture 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. 23. 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 expenditures within the prqvince. There appears to be great variation in the system of province-local :elations across provinces, suggesting the provincial governments have significant room to adjust fiscal decisions to accommodate local needs and preferences within the parameter' set by the central goveriiment. 24. Within the system of "vertical" responsibility, each province must account to the central government for its activities. In this process of vertical accountability, the following principles restrain, or guide, budgetary choices of provincial governments: a) there cannot be a deficit; 3/i Local oovernments are entitled to set surtax rates on a small number of taxes. Local governments also collect a set of extrabudgetary fees a.ad charges. -8- (b) current expenditures to maintain infrastructure have the highest priority among urban construction-relaLed expenditures; (c) the provision of social everhead facilities such as education, scientific research institutes and hospitals take a high priority; expenditures on culture and education are mandated to increase by at least the same rate as tota; oxpenditures; (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, i.e., for public utilities %nd public facilities. Budoetina and Financial Plannjfg 25. In theory, China has a unified system of budgeting--covering all accounts--in which all the financial plans and accounts of the central and subnational governments are j3intly presented.4; Eaclh provincial government in principle also has a consolidated budget which includes the budgets of all lower level local governments. However, in the provinces visited for this work, the budgets were neit}her fully unified nor fully consolidated.5/ 26. Because government budgets do not fully include all revenue and expenditure accounts, it is difficult to construct an estimate of the total amount of revenues raised or expenditures made by the subnational governments in particular local areas. For example, extrabudgetary revenues and nonplan expenditures are reported in the budget, along with budgetary receipts and outlays. However, departmental revenues and grants received are generally not reported in the accounts. Tl.e tranpfers between, and overlaps among, the budgets of the general government, the SOEs and public utilities are not apparent. The example of transfers from the general government to the SOEs can illustrate the nature of the problem. The provincial government budget shows transfers to SOEs under the "technical transformation" heading, but it does not distinguish between grants and loans. Other Lransfere to cover SOE losses are shown as an expenditure in the general provincial budget and not as a transfer, unless the loss is "unplanned," in which case it is shown as negative profits' tax revenue (i.e., it is subtracted from revenues--not added to expenditures). A further complication is that grants to provincial and local governments do not appear to be shown separately in the budget 27. With regard to the distinction made in some countries between capital and current budgets, each local (urban) government has a regular budget and a construction budget, but these do not correspond to a division of current and capital expenditures, and there is no separate reporting of capital financing. There does not appear to be an operative concept of capital budgeting. Tax Administration 28. Provincial and city governments cannot vary the nominal rates of tax, nor may they redefine the leqal tax base. However, they have almost complete autonomy in assessing and collecting taxes, and along w.th the lower level county government can and do give tax relief without having to seek approval from the Center. One could fairly say that subnational governments can substantially alter the level and pattern of effective tax rates paid by enterprises. 4/ "Unified" budgeting is used here to mean a budget which incorporates all accounts of any particular government unit. "Consolidated" uses the Chinese terminology and reflects the joint presentation of the budget of all levels of government. 5/ The provinces were Anhui, Jiangsu, Zhejiang, Shandong, Shanghai and Beijing (which have the status of a province). -9- 29. The organization of tax administration in China centers around the activities of two organizations: the Tay Bureau and the Finance Department. There are separate Tax Bureaus and Finance Departments at the city, county and province levels. The functions of the Tax Bureau and the Finance Department seem clearly distinguishable in principle. The Finance Department conducts tax policy allowed by law at the provincial and local levels, and manages the expenditure side of the budget. The Tax Bureau is responsible for implementing central tax law and collecting taxes. In theory, both are Et the same time organs of the central and local governments. In practice, the division of responsibility is not so clear, and the directives given to the Tax Bureau by the central government and ito subnational government are not always consistent. Many of the Tax Bureaus' actions in the provinces visited for this work suggest substantial tension as a result of this system of "dual leadership." The assessment based on this field work is that the tax and finance bureaus are more likely to act as agents of the provincial or local governments than of the central government. The Finance Department plays the dominant policy leadership role within the provincial and local governments. 30. Provincial and local goxernments have a surprising amount of discretion in granting tax relief. Their activities in this regard are referred to as the policy of "stimulating enterprises through tax expenditures." Three types of tax Bxperliture, or methods of granting preferential tax treatments, are used. First, if the provincial government wants to promote a new product or a pioneer industry, it may authorize a reduced tax rate or a tax holiday for a number of years (usually not to exceed five years). Second, the Finance Bureau may enter into a contract arrangement with an enterprise for payment of a negotiated (as distinct from schedular) amount of taxes. Third, the Tax Bureau may grant ad hoc tax relief to enterprises on a case-by-case basis depending on the needs of the enterprise. There is every indication that provincial and local governments use this discretion to promote economic development in the local area, even though the preferences granted sometimes do not conform with the objectives of the central government and may seriously Impair its revenue position. Provincial and Local Revenues 31. Provincial governments have four revenue sources: own taxes and shared taxes; extrabudgetary funds; user fees; and capital grants. China's revenue sharing system is primarily a division of sales anc profit taxes among the central, provincial, and local governments. Whereas in most countries the taxes are collected by the central government and then allocated to the lower level subnational governments, in China they are collected by the local governm'.nts and "shared-up" to the higher levels. The amount of shared tax revenue finally ending up at each level of government depends on the tax base and rate, tax administration, the sharing formulae between the province and its municipalities, and the sharing formula between center and province. To understand the revenue-sharing system in China, one must understand all of these dimensions. 32. Tax Revenues. By law, there are three categories of revenues--"fixed central government revenues," "fixed local government revenues," and "shared revenues." Box 1 shows the principal taxes in each of these categories prior to the 1988 proposed changes. In all three cases, however, rate determination and base definition are not under local control. Revenues collected from local taxes are, in principle, assigned fully to the local government and are referred to as "local fixed revenues." The actual practice of revenue sharing in China has not matched this scheme. Most "local fixed" and "shared" taxes have neen subject to sharing, apparently because adherence to these categories caused a revenue shortfall to the central government. Since 1988, other minor changes have taken place in the allocation of these taxes to different levels - 10 - Box 1: PRINCIPLES OF REVENUE ALLOCATION AND TAX SHARING 1985 REFORMS 1. "Fixed Central Goverrnment Revenues": 1. Income and adjustment tax of all central goverrinent enterprises. 2. Business tax from railroads, bank and insurance conpany headquarters. 3. Profit remittances by atl enterprises producing arms. 4. Price subsidies paid to producers of grain, cotton and oiL (treated as a negative revenue of the central goverrment). 5. Fuel oil special tax. 6. Income taxes, sales taxes and royalties from offshore oil activities of foreign companies and joint ventures. 7. Treasury bond income. 8. 70X of the three sales taxes collected from enterprises owned by the Ministry of Industry, the linistry of Power, SINOPEC, and the China Nonferrous Metals Company. 9. All customs duty and all VAT and product taxes collected at customs. 10. Tobacco Tax and Business Tax on Tobacco. II. Product tax on Liquor and tobacco. II. In 1985-87. the "local fixed revenues" were as follows: 1. Income tax and adjustment tax of locally-owned enterprises. 2. Income tax from coltectively owned enterprises (ICIT). - AgricuLture tax. Rural market trading tax levied on private sector traders. 5. Local goverrament grain trading Loss (a negative tax). 6. Fines for delinquent taxes. 7. The Urban Maintenance and Construction Tax (UMCT).1/ 8. Housing tax.g/ 9. Vehicle utilization tax. 10. 30X of the sales tax reveiiue collected from enterprises owned by the Ministry of Power, SINOPEC, and the China Nonferrous Met..s Conpany. 11. Individual income tax. 12. Wage bonus tax. 13. Self-empLoyed Entrepreneurs Tax 14. Slaughter Tax. 15. CattLe Trading Tax. 16. Contract Tax. III. Taxes shared between the central and local governments: 1. All sales taxes (valued-added, business, and product) revenues from all enterprises, except those expressly excluded as described above under 1:6, 9 and 10. 2. Natural resource taxes. 3. Construction tax. 4. Salt tax. 5. Industrial and commercial tax, and income tax, Levied on foreign and joint venture enterprises. 6. Energy and transportation fund tax. I/ The Urban Maintenance and Construction Tax (UMCT) is set at 7% of total sales tax Liability for for municipaLities (5% for towns and 1% everywhere elsel. of government. More generally, as central revenue needs grow, the shift of new taxes into the "central fixed" category is made. 33. Extrabudoetary Funds. other sources of revenue for provincial and local governments, earmarked for capital purposes, are extrabudgetary funds. The amount of revenue involved is relatively small compared to other provincial and local revenue sources, and accounts for only 3% of total extrabudgetary funds (enterprise retained earnings are about 80% and extrabudgetary revenues of government agencies are about 17%) and less than 1.6% of government budgetary revenue. - 11 - 34. User Charges. Though the public utility enterprises attempt to recoup a portion of costs through user charges, there has been no strong sentiment to raise rates to efficient (marginal cost) levels. In China, cost recovery is a much bigger matter than simply raising the level of the user ch>-ge. Water, sewerage and gas (LPG) charges, bus fares and housing cannot be adjusted independent of national wage and price policy and enterprise and tax reform policy. Perhaps as important, but less widely recognized, is the relationship between increasing the rate of user charge and the sharing of revenues among the three levels of government. An increased user charge--paid by enterprises or by individuals and compensated by an increase in wages--will lower profits and therefore the tax liability of enterprises. The result will be a shift in revenue power (a) from the central to the local level because the whole of the user charge "stays at home," and (b) from the general government to the public utility enterprises. Unfortunately, no data are available for making a good estimate of the percentage of total costs recouped by user charges, but it is probably quite low. Residential user charges have changed little since Liberation, though there has been some increase in the rates charged to enterprises. Within limits set by the central government, local governments can increase user fees, and have done so for commercial and industrial users. 35. Capital Grants and Borrowing. China has no regular, formula grant program to support capital projects; all grants are on an ad hoc basis. There is no mechanism or formal program for lending to local governments, and there is no formal mechanism that guides local governments in developing beneficiary financing schemes. Capital financing is done from some combination of current revenues, planned loans or grants, special exceptions to the restrictions on borrowing, and creative, ad hoc approaches to benefit financing. 36. Provincial and local governments in China cannot borrow. However, there appear to be ways to avoid these restrictions. Short-term borrowing (less than one year) and even some longer term credit financing does occur. In some cases, municipally owned enterprises borrow for infrastructure projects and in some cases the municipal government has pledged its general revenues to secure loans to its enterprises. Nevertheless, credit financing is quite limited. As a result, the "price" of capital construction is high because it must be financed from current revenues rather than loans, i.e., by current rather than future beneficiaries and by the general public rather than by direct beneficiaries. Intergovernmental Fiscal Relations 37. There are three important dimensions to the system of intergovernmental fiscal relations in China. the first is the system of tax sharing, outlined above, which determines the amount of resources that will be allocated to the subnational government sector. The second is the distribution of these funds among provinces, and includes both the tax sharing formulae and the flow of grants and subsidies. The third is the system of horizontal fiscal relations within the province, the method by which the provinci&l government allocates fiscal resources among its counties and municipalitiea. 38. Central-Provincial Transfers. The central government determines the tax sharing arrangements with the provinces. The system in China is essentially a sharing of revenues from a specitied set of taxes (Box 1), almost all of which are central government taxes that are collected by the local governments. The total size of the distributable pool is thus determined primarily by the amounts collected. The distribution of central revenues by province is then determined by a combination of (a) origin of collections, (b) formula, and (c) negotiation and ad hoc decisions. The exact sharing rate, which are in the - 12 - last analysis the result of a central-provincial negotiation, are presented in Table 3.6/ 39. The basic ehring formula, first applied in 1985, and t..Jified later to use a 1987 exp iditure base, is: E(83) RATIO =- R(83) where Ratio = The share of collections to be retained by the province E(83) = Actual amount of "Allowable" local government expenditures in 1983 R(83) = Actual amount of local fixed plus shared revenues collected in 1983 40. The results of applying this procedure are described in Table 3a for the 1985-87 period and in Table 3b for 1988-90. For example, allowable expenditures in Beijing were equivalent to 48.2% of total shared and local fixed revenues in 1985, hence Beijing in 1985 would be able to retain 48.2% of all it collected from those two categories of tax. Of course, all fixed central government revenues would have to be turned over to the center. As may be seen from Table 3a, 15 provinces were in such a surplus position in 1985, i.e., shared plus local fixed revenues exceeded allowable expenditures and the tax sharing ratio was less than unity. 41. The remainder of the provinces (with the exception of Guangdong) were in a deficit position. In these cases, the province was allowed to retain all of its fixed and shared revenue collections and the central government paid a subsidy equivalent to the size of the 1983 deficit. Eight of these deficit provinces--the autonomous regions, the provincos with heavy minority populatione and those which are least developed--were singled out for special treatment. They were to receive the deficit subsidy, but this amount was to be increased by 10% per year. Apparently, 10% was taken as a number that would roughly approximate needed revenue growth. In fact, the 10% increment was given in 1986 but was reduced to 5% in 1987 and 1988 with the intention to eliminate it in 1989. 42. This approach has now undergone major changes, with a growing number of provinces contracting with the center for a delivery of a fixed tax quota, following the example of Guangdong, which retained all fixed local and shared collections, but turned over a fixed annual amount of Rmb 778 million to the center. This was a first step toward the proposed 1988 reforms, under the name of "provincial contracting." 43. Since 1988, most provinces have "bargained down" the center in contracts which reduce their obligations to share and transfer funds. The purpose of this program is to give these provinces greater incentives to collect more taxes. Since 1988, contracting has been applied in most of China's prosperous (and high-yield) provinces and cities including Jiangsu, Zhejiang, Hebei, 6/ This formula has been adapted a number of times since 1984 to adjust for the transfer of state-owned enterprises from provincial to central owner- ship. - 13 - Table 3a. CHINA: REVENUE-SHARING SYSTEM BETWEEN THE CENTRAL AND PROVINCIAL GOVERNMENTS, 1985-1988 Percentage of total Province Provinces cotlections receives subsidy Fixed or contract and regions retained b Province aereed from the Center a/ delivery to the Center 1985 1986 1987 1985 1986 1987 1985 1986 1987 X)I ------------tin millions of yuan) ------------------ North China Beijing 48.2 49.55 49.55 Tianjin 39.5 39.45 39.55 Hebel 69.0 72.00 72.00 Shanxi 97.5 97.50 97.50 Inner Mongolia 1783* 1961.74 2059.83 Northeast China Liaoning 51.1 52.66 52.66 Jilin 397 396.62 396.62 Heilongiiang 96.0 142.70 142.70 East China Shanghai 26.0 23.54 23.54 Jiangsu 39.0 41.00 41.00 An Hui 80.1 80.10 80.10 Fujian 235 234.86 234.86 -- Jiangxi 239 239.46 239.46 Shandong 59.0 77.47 75.0 Central/South China Henan 81.0 81.00 87.1) Hubei 66.5 100.00 100.09 Hunan 88.0 88.00 88.0( Guangdong 772.08 778.08 778.00 Guangxi 716* 788.03 827.43 Southwest China Sichuan 89.0 100.00 100.00 Guizhou 743* 817.57 858.45 Yunnan 637* 925.88 972.17 Tibet 750* 825.32 866.59 Northwest China Shaanxi 270 270.26 270.26 Gansu 246 245.60 245.60 Qinghai 611* 671.88 705.47 Ningxis 494* 543.14 570.30 Xinjiang 1450* 1594.85 1674.59 Source: Data supplied by MOF. a/ Asterisk indicates subsidies were to increase by 10% per year after 1985. Two asterisks indicate quota contract delivery, also known as "provincial" contracting. No informa- tion available on amounts. Information not available for the province's arrangements with the Center for 1988. Beijing, Tianjin, Guangdong and Shanghai (see Table 3b). Under this system, a basic amount (quota) of shared revenues must be transferred to the central government while revenues collected over and above this quota may be kept in full by the province or city (see belowl. Table 3b: China: Central-Local Revenue-Sharing Contracts 198P-90 (i) (ii) (iii) (iv) (v) (vi) Basic Sharing with Growth Incremental Sharing Fixed Quota with Growth Basic Contracted Basic Fixed Initial Contracted Payment Basic Retention Rate of Retention Marginal Quota Amount Amnuat Rate to Deficit Sharing Rate Increase Rate Retention to State to State of Increase Provinces Rate X ()() Y 100 mn.) (7 100 on.) (X) (t 100 mn.) Beijing * 50.0 4.0 - Hebei - 70.0 4.5 - Liaoning - 58.3 .5 - Shenyang - 30.3 4.0 - Harbin - 45.0 5.0 - Jiangsu - 41.0 5.0 - Zhejiang - 61.5 6.5 - Ningbo 27.9 5.3 Henan - 80.0 5.0 Chongqfng/a - 33.5 4.0 Tianjin 46.5 - - Shanxi 87.6 Anhui 77.5 - - - - Dalian - - - 27.7 27.3 Qindiao - 16.0 34.0 Uuhan/a - - 17.0 25.0 - - Guangdong tinc. Guangzhou) - - - - - - 14.1 9.0 - Huian - - - - - - 8.0 7.0 - Shanghai - - - - 105.0 - - - Shandrg - - - - - 2.9 Hei Ictiang Imm 3.0 Jitin - - - - - - 1.1 Jianxi - 0.5 Shaanxi (inc. Xian) - - - - - - - 1.2 Gansu - - - - - - - - 1.3 Fujian (beginning 1989) - - - - * - . 0.5 Inmer Mongolia - - - - - - - 18.4 Guangxi - - - - - - - - 6.1 Tibet 9.0 Ningxia 5.3 Xinjiang 15.3 Guizhou 7.4 Yurnan - - - - - - - - 6.7 Qinghai - - - - - - - - 6.6 Hainan - - - - - - - 1.4 Source: Ministry of Finance /e After the cities of Wuhan and Chongqing were separated from Hubei and Sichuan provinces, the provinces changed from net providers to the state to net recipients of suibsidies from the state. Data are not avaLtable on the other "independent cities", such as Nanjing, Ningpo, etc. - 15 - 44. Another important change in the past two years has been to create "special cities" which hava a direct relationship with the central government and whose revenue-sharing arrangement is independent of the arrangement which exists for their province. As of 1991, these include Harbin, Changzou, Shenyang, Dalian, Quingdao, Xian, Chengdu, Chongqing, Wuhan, Nianjing, Ningpo, Xiamen, Guangzhou, and Shenzhen. 45. There are now 6 basic variants of the "quota" and sharing arrangements in which provinces contract with the central government. Table 3b outlines them for the 1988-90 period, both for the provinces and for the cities with independent status. These six main types of contracts under which transfers are effected from the provinces to the central government can be characterized as:7/ li) "Basic Sharing" - a fixed proportion of all revenue is remitted to the center; (ii) "Basic Sharing with Growth Adiustment" - based on 1987 revenue, the localities retain a specified proportion ranging from 28 percent to 80 percent of any revenue that is within a certain percentage growth (ranging from 3.5 to 6.5 percent) from the previous year (revenue in excess of this growth rate is retained entirely by the province); (iii) "Incremental Sharing" - a certain proportion is retained up to a quota, and then a different (usually higher) proportion of revenue is retained in excess of the quota; (iv) "Fixed Ouota Delivery" - a specific nominal amount is contracted to be transferred to the center with no annual adjustments; (v) "Fixed ouota with Growth Adiustment" - a specified nominal amount is contracted to go to the center in the initial year; in subsequent years, the amount was to increase at a contracted rate (7 percent or 9 percent). (vi) Deficit provinces continue to receive fixed subsidies. 46. As the sharing proportions in Table 3b show, the provinces in which significant revenue is collected (Shanghai, Liaoning, Beijing) retain a significant proportion of what they collect, and that incremental sharing and growth rates are quite low, the latter well below GDP growth. 47. While the provincial contracting arrangements can provide increased incentives to collect revenue, they do not eliminate the potential tonflict of interest between local and central governments, nor do they alleviate a crucial problem associated with decentralization, that is, the loss of control and flexibility by the Center over the use of tax policy for microeconomic purposes. 48. Moreover, the provincial contracting system, if more widely applied, can lead to negative revenue consequences for the center. The fact that most important contracts are fixed in nominal terms has left an increased share of fiscal resources in the hands of local governments. This reduces the growth and, potentially, the real volume of resources in the hands of central government. Finally, the provincial contract arrangement introduces a pro- cyclical bias to the fiscal system. This occurs because the revenue received 7/ There is, in fact, an infinite variety of these sharing and quota/incremental contracts and it is to some extent forced to categorize them in this fashion. Essentially, these are bargains, not a revenue- sharing "system." - 16 - by the central government under a "fixed rominal" contract remains relatively constant regardless of the underlying growth of economic activity. B. PROBLEMS AND ISSUES IN CENTER/LOCAL FINANCE IN CHINA 49. China's experience with fiscal decentralization is marked by four patterns. First, as decentralization has proceeded, the share of taxes in GNP has declined. Second, the "fiscai surplus" available to the central government--the excess of its total direct expenditures over its direct revenue collections--has fallen (see Table 2). Third, local governments have taken a great deal of discretion in the implementation of tax policy, with important results: (a) the application of the "national" tax system varies widely from province to province, and (b) use of the tax system to promote provincial or local goals has not always been in keeping with central guide- lines. This reduced central leverage over fiscal revenues has led to problems of macroeconomic management, an erosion of the revenue base, allocative distortions, and generally to a more ad hoc tax system. Fourth, revisions to the revenue sharing system have had the effect of improving the relative fiscal position of higher income provinces. This pattern of change raises some important problems which must be faced in restructuring central-local relations in China. Moreover, the current system may represent an important impediment to price reforms and reforms addressing enterprise restructuring, both of which are important system reforms in the next Plan period. Compatibilitv with System Reforms 50. China's reform program in the Eighth Plan has the implementation of price reform and enterprise reform as its twin linchpins. The present system of central-local fiscal relations cannot support these reforms. 51. Consider first the area of price reforms. Since localities receive revenues from their locally owned enterprises, and also have an entitlement to all sales and excise taxes generated in their locality, local revenues can be markedly affected by price reform. Increasing market prices for intermediate goods, for example, would greatly improve the collection potential of profits and sales tax in provinces producing such goods. Provinces such as Shaanxi and other raw materials producers would benefit both by the increase in the tax base, and by the fact that their fixed nominal tax contract to the center was tailored to profits and value added tax yields estimated on the basis of lower pre-reform prices and profit levels.8/ Conversely, in the absence of retail price liberalization, provinces using basic raw materials would see their enterprises' profitability decline (and their tax yield correspondingly lower),9/ and would find it correspondingly harder to meet the terms of the contract. To illustrate the magnitude of these potential effects, the data in Table 4 show, for major sectors in China's input/output table, the provinces which are the largest producers of commodities subject to price controls. Column 4 shows an indication of the proportion of the total output nationwide, subject to such controls on the assumption that retail prices in general are 8/ The interprovincial impact could be quantified using an input-output framework. A corresponding sharing or transfer system could also be devised which would modify tax contracts in line with the shifts in tax yields implied by price reform. However, since prices will continue to undergo adjustment, such transfer system would require ongoing adjustments following each price change. 9/ If both intermediate and final goods prices are liberalized, provinces which are predominantly intermediates-producing would gain, while "proce- ssing provinces" could come out about equal, gaining from the increase in final goods prices, but losing from the higher input prices. - 17 - Table 4: CHINA: PRICE CONTROLLED PRODUCTS: SHARES OF GROSS VALUE OF INDUSTRIAL OUTPUT (GVIO) OF THE LARGEST PRODUCING PROVINCES IN 1989 (percentage shares into total sectoral GVIO) (1) (2) (3) Total Share Sales of Top 3 to Final Total Province Provinces Consumers GVIO by sector (X) CX) (Y billion) Smelting and pressing Liaoning (17.1) of ferrous metals Shangha (12.7) 38.0 1 114.0 Sichuan (8.3) Smelting and pressin Liaoning (9.9) of nonferrous metals Gansu (7.5) 24.2 1 47.15 Shanghai (6.8) Metal products Jianssu (12.1) Shanghai t10.0) 31.5 1 49.42 Guangdong (9.4) Electrical equipment Guangdong (14.6) and machinery Jiangsu (11.7) 36.9 9 84.98 Sha ghai (10.6) Electronic & telecom. equip. Jiangsu (19.4) Shanghai (12.9) 40.7 - 55.12 Guangdong (8.3) Coal mining & dressing Shanxi (19.3) Shandong (10.7) 38.6 26 40.81 Henan (8.6) Petroleum processing Liaoning (19.7) Shandong (10.7) 43.1 5 45.73 Heilongjiang (11.3) Petroleum and natural Heilongjiang (37.2j gas extraction Shandong (19.1) 65.4 5 36.24 Liaoning (9.1) Chemicals and allied products Jiangsu (13.2) Li oning (7.1) 27.2 2 137.53 Shanghai (6.9) Machine building industry Jiangsu (12.7) Shanghai (10.9) 32.0 8 172.66 Liaoning (8.4) Food manufacturing Jiangsu (9.9) Shandong (9.4) 27.1 - 120.30 Hunan (7.9) Textile industry Jiangsu (18.7) Shandong (10.7) 39.9 74 210.96 Zhejiang (10.5) Paper making and paper prod. Shandong (8.7) Guangdong (7.5) 23.5 39 37.21 Liaoning (7.4) Other industry Guangdong (18.1) Jiangsu (14.1) 39.3 66 7.24 Zhejiang (7.1) Source: China: Sta*istical Yearbook 1990, pp. 405-18. largely liberalized; intermediate sales are price controlled to varying degrees. The magnitude and distribution of the positive impact of price liberalization is a function of the wedge between market and controlled prices, and the share of price-controlled output in provincial GVIO. [Data on the largest users (who would be negatively impacted) are not available.]52.It is because tax contracts and quotas are fixed in nominal terms and based on the present provincial tax bases and estimaterd yields, that price reform implies shifts in the effective burden of fixed nominal contracts. If contracts were instead defined as a share of total collections (as they were - 18 - earlier), the tax yield and the contracted amount due would change in the same proportion, and price reform would not give rise to these effects. The strong vested interests of some provinces in the existing status quo, could form a powerful impediment to price reform unless the fiscal system can be redesigned to compensate their losses. 53. The present system of intergovernmental contracting is similarly unsupportive of the enterprise reforms needed to strengthen industrial policy. As pointed out by many researchers on China, a major objective of China's industrial policy is to integrate its markets and reap the benefits of industrial specialization and economies of scale. The present fiscal system, by contrast, is a prime contributor to the widely-noted tendency for local governments to set up identical industries with inefficient plant scale and lack of regional specialization, since localities reap the taxes from enterprises they own. The preference for setting up processing industries-- which yield higher value added under the present distorted centrally-guided price structure--is further encouraged by the design of the fiscal sharing system under which provinces retain incremental tax yields above a nominal tax quota.l0/ 54. The rationalization of industrial structure, and esp-cially the improvements in scope and scale economies which can be achieved through a well-executed strategy of horizontal integration or cooperation between enterprises, enterprise mergers, and enterprise industrial groups, is difficult to implement in the current system. This is because the revenue- sharing system, under which revenues accrue to the level of government (provincial, local, etc.) owning the enterprise, interferes with mergers across localities, since this would require agreement on the sourcing and subsequent division of profits.l1/ A system in which ownership plays no role in the accrual of fiscal revenue to localities would be preferable. The Macroeconomic Perspective 55. Hitherto in China and other socialist economies, the stabilization aspect of tax policy was not a major concern, since most sources of macroeconomic instability--unforeseen (or unwanted) changes in investment, consumption and savings--were "controlled" by planners. The present reforms decentralize decision-making, and make it necessary for planners to give more consideration to the stabilization aspects of fiscal policy. This in turn requires improved management of the major tax instruments. 56. From a macroeconomic ,gerspective, the present fiscal system has three shortcomings. First, provincial contracting has been an important factor contributing to the lower income elasticity of government revenues. Second, by locking in a relatively fixed revenue level for a multiyear period, contracts deprive the central and local governments of a flexible revenue source and, in fact, introduce a procyclical bias to the fiscal system. Reve- nue received by the central government under the prevailing "fixed tax quota" system remains relatively constant regardless of the underlying growth of economic activity. (Even where the center collects a small share of above- quota collections, its revenues still grow more slowly than GNP because the above quota sharing rate tends to be low.) By contrast, rapid economic growth 10/ See China: Between Plan and Market, Red Cover Report 1990, World Bank for further discussion, and Tidrick and Chen, China's Industrial Reforms, Oxford University Press, 1988. 11/ Even in countries as advanced as the United States, the determination of the base for state-level corporate income taxes is a difficult proposition. Complicated formulae are used to prorate the base for firms with multi-state locations. - 19 - increases the revenues realized by local governments (which feel the full fiscal effect of booms). 57. Since local governments do not have demand management responsibilities, they may increase their spending when a booming economy produces higher revenues (producing the multiplier effect of a "balanced budget"), but would not, as might the central government, try to dampen agg-egate demand. Alternatively, local governments could respond to economic booms by targeting collections to some fixed revenue level. In this case, those resources which would have been taxed are retained by enterprises to further fuel aggregate demand through the multiplier effect of a "tax cut." Either local response will be expansionary and procyclical. Similarly, but in reverse, in an economic downturn, provincial contracts which are fixed in nominal terms would produce a drain on local fiscal resources, reducing consumption and procyclically intensifying the downswing, unless, at the central level, this is offset by a compensating budgetary stance. 58. The third shortcoming of the present quota-based system derives from its lack of flexibility. The provincial quotas' three- to five-year terms lock the central government into a relatively fixed revenue level which is not fully responsive to growth in either output or prices. Thus, the central government is deprived of a flexible fiscal tool. The result is that the center has less discretion over net spending, and a less effective fiscal instrument for fine-tuning the economy. Local governments have been delegated the responsibility for tax administration, absent effective central monitoring and oversight, and the authority to grant tax relief within their juris- dictions. This effectively severs the link between revenue policy (and tax policy) set at the central level and collections at the local level, and has put local governments in the position of determining effective tax rates and collection levels. 59. The reduced flexibility over tax policy emerging as a result of this specific form of decentralized fiscal arrangement represents a potentially worrisome development. Because this decentralized fiscal arrangement gives local governments discretion in determining the total revenue take, the center cannot run a fiscal policy suitable for and responsive to differing economic conditions. The Chinese government must weigh the benefits of this form of fiscal decentralization against the costs of reduced capacity for central management and direction over tax policy. 60. To improve central r.tanagement of revenues, expenditures, and stabilization policy, it may be necessary to reduce local governments' discretion over tax relief and to monitor their tax administration. The status quo with regard to central and local finances appe&rs untenable in the long run, and mechanisms need to be developed--some combination of new taxes for local governments, a reassignment of existing taxes or a different sharing of tax bases, a different assignment of expenditure responsibilities--to serve better both central and local interests. Revenue Adeouacv 61. Fiscal decentralization has shifted economic resources from the general government sector to the enterprise sector. Specif-cally, while the yield of the Chinese tax system has kept pace with prices and population (i.e., there has been positive, but low, real revenue growth in the 1980s), it has not kept pace with growth in total real incomes or output. For the local government sector specifically, the ratio of taxes collected to GDP fell from 19.4 percent in 1980 to 10.8 percent in 1989, while for the center, it rose from 4.6 percent to 9.3 percent in 1986, before falling to 6.5 percent in 1989 -20 - (see Table 5 and Chart 1).12/ Sinc" the inception of provincial contract- ing there has also been a decline in the share of revenues accruing to the central government. This shift in resources between central and local govern- ment does not appear to have been commensurate with any change in their respective expenditure responsibilities. 62. Judging overall revenue adequacy is difficult in the absence of concreta projections of expenditure needs. However, one can safely say that public service levels are deficient in all parts of China, and the infrasturcture gap is likely to be an especially critical problem in the future. Moreover, many reforms--especially those which attempt to replace enterprise-based services with privately or government-provided equivalents such as housing reform, socia) security reform and education financing--almost certainly will not be expenditure-neutral and may well incroase the financial burden on all levels of government.l3/ 63 Judging whether the "balance" is right, i.e., whether the system generates sufficient revenues for the respective levels of government, is also difficult in the absence of estimates of expenditure needs for each level of government. However, the data do tell us that the balance is changing. The central government's share of revenue collections has increased (see Table 5) over the period 1980-86; however, this was not because of a strengthening of central government revenue-raising capacity. It is primarily because provincial and local government revenue collections as a percent of GDP fell by nearly half. This decline largely reflects a shifting of the financial resources to the enterprise sector. Since 1986, the balance of financial resources seems to have shifted back toward provincial and local governments as central government revenues collected also have begun to decline as a share of GDP. Real revenue collections by the central government declined by 4 billion yuan between 1986 and 1989. 64. Provincial and local collections over the 1980s have grown in real terms, but have not kept pace with GDP.14/ The reasons may include "tax preferences" given to enterprises, lax administration; central revenue-sharing policies which reduce local collection incentives (tax effort), or central policies which reduce enterprise profits (such as wage or price changes) and thereby the provincial tax base. 65. On the exnenditure side, the center's direct share of expenditures (Table 5 and Chart 3) has fallen from 53 percent in 1980 to 36 perc3nt in 12/ The center's major collection vehicles include customs duties, revenues from centrally owned enterprises and joint-venture firms. Generally speaking, revenues collected by the center accrue to the center. Reve- nuee collected by localities are in fact "central revenues" but subject to sharing. 13/ One cannot safely say that the overall decline in the tax-GDP ratio means a reduction in government financia'l capacity. Against the 1980-89 decline of 4 percent of GDP, one must set off the reduced claim by enterprises on general government revenues. 14/ Data note: Revenue data (unadjusted) presented in Table 5 includa--as per the standard Chinese budgetary format--central government borrowing, both foreign and domestic in the revenue base. This substantially overstates (especially since 1982 when the borrowing became important) the true amount of central revenue. Obtaining a closer picture of the central revenue trend requires adjusting the central collections by the amount of central borrowing. Local data require no such adjustment, as local government revenues do not include borrowing. The results of this tabulation are shown in Table 4 under "central revenue (adjusted)." - 20a - CHINA: RATIO OF REVENUE TO GNP % OF GNP 36 20 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 * Total {- Central - Local CHINA: RATIO OF REVENUE TO GNP (Adjusted) . OF GNP 30. 25' 20 * 10 i 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 IN TOTAL 8 CENTRAL(ADJUSTED) -9LOCAL - 20b - CHINA: RATIO OF EXPENDITURE TO GNP % OF GNP 30_.- 25 20 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 Total Central Local CHINA: BUDGET BALANCE: CENTRAL & LOCAL of GNP 25 20 _ 15 5 0 d____________________.. . jfi 1980 1981 1982 1983 1984 1986 1986 1987 1988 1989 ~' CENTRAL RevenusiGNP -a- LOCAL Revenue/GNP CENTRAL Expand.IGNP -4- LOCAL Exp*nd./GNP Table 5: TAX COLLECTION AND EXPENDITURE By Level of Government 1980-90 billions of R 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 Total revenues collected 108.5 108.9 112.4 124.9 150.1 186.6 231.0 236.0 262.8 285.9 Central government 21.0 22.5 25.8 37.2 52.4 69.0 91.7 83.3 94.5 107.2 of which: borrowing -- 7.3 8.4 8.0 7.7 10.0 13.9 16.9 23.1 22.0 Local goverrnent 87.5 86.5 86.6 87.7 97.7 117.6 134.3 152.7 168.3 178.7 of which: own 11.8 11.6 12.5 15.3 15.9 19.0 -- -- -- -- Total expenditure 121.2 111.5 115.3 128.4 153.6 186.5 233.0 242.5 263.5 295.0 Centrat 65.0 60.2 57.5 64.2 72.8 81.9 96.2 95.8 94.6 -- Local 56.2 51.3 57.8 65.8 80.8 104.6 136.8 146.7 168.9 -- Central collections deficit -44.0 -37.7 -31.7 -27.0 -20.4 -12.9 -4.5 -12.5 -0.1 Funded by: Local collections surplus: 31.3 35.2 28.8 21.9 16.9 13.0 -2.5 6.0 -0.6 -- Central borrowing - 7.3 8.4 8.0 7.7 10.0 13.9 16.9 23.1 22.0 Total revenue collections (adjusted) 108,5 101.7 104.0 116.9 142.4 176.6 212.1 219.1 239.7 263.9 Central 21.0 15.2 17.4 29.2 44.7 59.0 77.8 66.4 71.4 85.2 Local 87.5 86.5 86.6 87.7 97.7 117.6 134.3 152.7 168.3 178.7 (in percent) Total expenditure 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Centre L 53.63% 53.99X 49.87X 50.00X 47.40X 43.91Z 41.29Z 39.51X 35.90X - Local 46.37X 46.01X 50.13% 51,251 52.601 56.09% 58.711 60.491 64.101 - Total revenues collected 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Central 19.35X 20.66% 22.95% 29.781 34.91X 36.98% 39.701 35.301 35.96% 37.501 LocaL 80.651 79.43% 77.05% 70.22% 65.09% 63.02% 58.14X 64.701 64.041 62.50X Total revenues coltected (adjusted) Central 19.351 14.951 16.73X 24.98% 31.39X 33.411 36.68X 30.31X 29.79X 32.28X Local 80.65% 85.05% 83.27% 75.02X 68.61X 66.591 63.32X 69.69X 70.211 67.72% (in X of GNP) Revenues collected/GNP 24.27X 22.811 21.67X 21.59% 21.681 21.851 23.77X 20.781 18.96% 18.17X Central 4.70% 4.71X 4.971 6.43% 7.571 8.081 9.431 7.331 6.821 6.811 Local 19.57X 18.121 16.70% 15.161 14.11% 13.77% 13.821 13.451 12.14% 11.36% Central adjusted revenues 4.701 3.18% 3.36% 5.05% 6.461 6.911 8.001 5.851 5.15X 5.42% Expenditures/ANP 27.111 2,.35% 22.231 22.201 22.181 21.841 23.971 21.351 19.011 - Central 14.541 12.611 11.091 11.10% 10.51% 9.59X 9.90X 8.441 6.831 - Local 12.577 10.741 11.15% 11.381 11.67X 12.25% 14.071 12.92X 12.191 Memo Items: Real rewenues collected 108.5 107.0 110.1 120.4 137.2 156.4 183.8 177.6 180.2 179.5 Central 21.0 ,22.1 25.3 35.9 47.9 57.8 73.0 62.7 64.8 67.3 Local 87.5 85.0 84.8 84.6 89.3 98.6 106.8 114.9 115.4 112.2 Real revenue collection growth Central 5.25% 14.331 41.96% 33.52% 20.751 26.131 -14.081 3.41X 3.831 - Local -2.891 -0.18X -0.291 5.60% 10.38% 8.39% 7.54% 0.46X -2.82X - GNP 447.1 477.5 518.6 578.4 692.4 854.0 972.0 1135.7 138'5.8 1573.1 Deflator 100.0 101.8 102.1 103.7 109.4 119.3 125.7 132.9 145.8 159.3 Source: MOF - 22 - 1988. As a percentage of GDP, the center's expenditures have fallen from about 14.0 percent in 1980 to 6.8 percent in 1988, with a corresponding, but substantially smaller decline in the provincial share, from 13.0 percent to 12 percent. Again, these declines are indicative of the shift in resources to the enterprise sector. 66. To get a true picture of what has happened to the balance of fiscal power in the Chinese intergovernmental syster.m in the 1980s, the revenue and expenditure sides must be considered together. While the provinces' revenue and expenditures have both fallen as a percentage of GNP, their revenue share has fallen faster. As a result, the provincial "surplus" available for transfer to the center has shrunk in nominal terms (see Chart 4). In 1980, the center financed almost 50 percent of its total expenditures from local transfers, and provinces spent only two-thirds of what they collected and transferrea the remainder to the center. By 1986, the provinces' direct expenditures were 1 percent greater than their collections and the center tranbferred net, to them, Y 2.5 billion (4 percent of its expenditures). By 1988, there was a transfer to the provinces of Y 0.6 billion. Thus, while in the past the ceiter could spend more than it collected--in 1980, some 50 percer.t more--this discretionary expenditure authority has been eroded since 1986 (see Table 2).15/ There are many reasons for this. Provincial contracting has reduced the net transfer to the center, as has enterprise contracting which has reduced the revenue share being "passed up." 67. one might be tempted to interpret this as "greater balance," i.e., each level of government is now spending about what it raises. This would not be correct: any discussion of revenue adequacy needs to be put in the context of the expenditure responsibilities of each level of government. There are also important implications of this "self-sufficiency" of the national government sector for stabilization and equalization capabilities. This shift means that the central government's ability to use discretionary policy to redistribute fiscal resources among provinces, or to manage national budget finances (i.e., to run a surplus or deficit), is much more limited in 1989 than in 1980. 68. These trends also show that the central government has borrowed in order to maintain its expenditure share, i.e., to replace revenues provided earlier by the provinces. (First differences of provincial collections as a share of GNP have declined since 1986 approximately in parallel with increased central borrowing, and in 1989, by more than the center's borrowing.) Continuation of these trends could complicate macroeconomic management, as central borrowing to make up for reduced local transfers adds to demand pressures. And if these funds are not used for capital purposes, this practice shifts the payment for current services to future years. Ircentives for Tax Avoidance 69. The tax-sharing system in China presents provincial and local governments with mixed incentives to improve their tax administration efforts and may dampen revenue mobilization efforts. Simply put, high sharing rates may encourage tax avoidance: if provinces cannot keep all that they collect, they may not put much effort into collection. Many avenues are open for "tax avoidance" including concessions to help enterprises over hardship periods, tax holidays to encourage new activities, and enterprise contracts with reduced tax liabilities to generate a "supply-side" response--all within the spirit and the letter of the law. Local governments can also allow evasion by becoming more lax in assessment and collection efforts and permitting enterprises to underdeclare tax liability. 70. Local interests in stimulating local economic activities give rise to strong forces aaainst collecting tax revenues, particularly those that must be shared with the center at a high rate. Local tax exemptionei allow resources to remain under local jurisdiction, to be tapped through "voluntary" IS/ Expenditure data for 1989 and 1990 are not available. - 23 - contributions that are not shared with the center. Thus, while the effective tax and nontax burden on enterprises may not be reduced, budgetary revenue of the central government is eroded (see Annex 1 for a formal treatment). Whether the retention rates in China are high enough to encourage good tax administration is an oven question. Higher retention rates may be needed to convince local governmente to tax the marginal dollar into the public budget rather than leave it in enterprises' retained earnings.16/ Alternatively, it may not be feasible for the present shared tax system to coexist with local responsibility for tax collection and discretion in administration. While the quota arrangement can in principle provide incentives to collect revenue, the conflict of interest between local and central governments remains. There is also no guarantee that the "supply-side', response--assuming it is achieved through these contracts--will materialize in a way that can be easily predicted and coordinated with expenditure plans or other macro requirements. Provincial Tax Effort ,I. Collections data show that, on average, the richer provinces collected more revenue per capita than poorer provinces over the 1980-86 period (see Box Table 1 and Annex Table 4). However, revenue growth in six of the ten richest provinces was below the national average, while nine of the ten poorest provinces had above-average growth in revenue collections (See Annex Table 6). The extent to which provinces are exploiting their tax base thus clearly needs to be addressed (see Box 2 for a discussion of the measurement of provincial tax effort.) For policy purposes, the Chinese government wants to know not only which provinces have a greater capacity to collect revenues, but how extensively they use this capacitv. Otherwise, there is the risk of subnational governments letting central transfers substitute for what other- wise would have been increased local government revenue mobilization. 72. A first approximation of revenue effort is the "tax ratio," i.e., the rates of revenues to gross output (see column (1) of Box table]. The "tax ratio" in Shanghai was 18 percent compared to 9.8 percent in Gansu and 8.7 percent on average for all provinces. This is only a partial measure of tax effort, because much of the interprovincial variation would be expected due to differences in taxable capacity relating to levels of income and urbanization. The tax ratio does not therefore show that higher-income Shanghai exerts more than twice the revenue effort of Gansu. In fact, comparing tax collections to estimated "taxable capacity" suggests that Gansu actually makes a greater revenue effort than Shanghai. In general, it would appear that many of the higher income provinces make a lower level of revenue effort: Jiangsu, Zhejiang, Shandong and Liaoning all make below average efforts and Shanghai is just about average. While tax collections are lower in provinces where per capita income is lower, many low-income provinces raise more revenue than might be expected given their economic base (column 3). Incentives for Infrastructure Development 73. While infrastructure bottlenecks are a major problem facing almost all Chinese localities, the fiscal system accentuates the capital financing gap 16/ Finance Minister Wang Bingqian stated that some people "evaded taxes, retained a larger share of profits than they were entitled to for their own enterprises, falsified accounts about losses to secure subsidies, and diverted state funds to uses other than those prescribed..." (1987 Budget speeches). Many of the provinces "hold back" on the proper level of tax remittance, authorizing preferential tax treatment of certain firms, and thereby lowering the total taxable base. In this way, the after-tax profits of locally owned firms may be increased and more resources may be "kept at home" in the province. - 24 - Box 2: PROVINCIAL TAX EFFORT Taxable capacity of a province is measured as a function of its income level and its degree of urbaniza- tion. The higher the level of per capita income--proxied here by the per capita gross value of output--the greater the capacity to raise profit and sales taxes. Urbanization may also contribute to taxable capacity because urban economic activities are more easily reached by the ackninistrative system (provide better tax handles) than do ruraL activities.l/ The results show that taxable capacity is significantly higher if per capita output is higher, and is positively (though not significantly) retated to the level of urbanization. Using this estimated relationship, one can predict the tax ratio, based on a provinces's per capita income and urbanization. To take Beijing as an example, one could say that, based on the average practice and its own Level of per capita output and urbanization, we would "expect" Beijing's revenue ratio, or taxable capacity, to be 12.78 percent.?1 Tax effort is the extent to which a province uses this capacity, and may be measured as the ratio of the actual rate at which output is taxed (Box Table 1, column 1) to estimated taxable capacity. To follow the example above, we expect Beijing to raise revenues equivalent to 12.78 percent of output (coLumn 2), the province actually raised 15.24 percent in 1986 (column 1), hence there is an above average effort--specifically an effort which is 19 percent above average as is shown by the effort index of 1.19 in column 3 of the Table below. Shanghai, by contrast, raises questions about what would be expected and makes an average tax effort with an index of 1.01. j/ The method used here follows the approach developed in Roy Bahl, "A Regression Approach to Tax Effort and Tax Ratio Analysis," International Monetary Fund Staff Papers, Vol. 18, No. 3, November 1971. 2/ The average relationship between the share of output raised and these two indicators of taxable capacity is determined from a linear OLS estimate as described in the following equation: T/Y 0.059 + 1.514 (E-05) Y/P + 0.0002 u (5.168) (0.600) where T 2 tax colLections Y = gross vaLue of industrial output P = population u = percent of population living .n urban areas l 2 0.63, and t-values shown in parenthesis Box Table 1: COMPARISONS OF TAXABLE CAPACITY AND TAX EFFORT BY PROVINCE Ratio of budgetary "Taxable Index of Tax Percent Ranking: Per coLlection to capacity" tax effort effort of total capita national Province gross output, 1986 1986 1.0 /a ranking population income: 1985 Beiiing 1S.24 12.78 1.19 5 0.92 2 Tianjin 14.42 13.30 1.08 8 0.78 3 Hebei 7.52 8.90 0.94 15 5.33 14 Shanxi 8.93 8.48 1.05 9 2.52 12 Inner Mongolia 7.87 7.92 0.99 13 1.93 16 Liaonirz 9.75 10.57 0.92 15 3.54 4 Jilin 7.75 9.09 0.85 18 - 8 HeiLongjiang 7.82 9.09 0.86 17 3.18 5 Shanghai 18.06 17.95 1.01 11 1.17 1 Jiangsu 6.30 9.49 0.66 19 5.97 6 Zhejiang 7.93 9.18 0.86 17 3.87 7 Anhui 6.49 7.55 0.86 17 4.95 21 Fujian 9.35 8.05 1.16 6 2.61 17 Jiangxi 7.10 7.59 0.94 15 3.32 22 Shandong 5.44 8.70 0.63 20 7.39 10 Henan 7.45 7.42 1.00 12 7.41 25 Hubei 7.63 8.49 0.90 16 4.74 11 Hunan 8.06 7.73 1.04 10 5.40 19 Guangdong 8.51 8.33 1.02 11 6.01 9 Guangxi 8.90 7.43 1.20 4 3.72 28 Sichuan 7.18 7.52 0.96 14 9.79 26 Guizhou 8.79 7.18 1.23 3 2.85 29 Yunnan 12.35 7.28 1.70 1 3.27 27 Tibet 6.02 6.97 0.86 17 0.19 18 Shaanxi 7.86 7.84 1.00 12 2.88 24 Gansu 9.86 7.76 1.27 2 1.96 23 Qinghai 7.85 7.69 1.02 11 0.39 15 Ningxia 8.61 7.74 1.11 7 0.40 20 Xinjiang 6.25 8.14 7.68 - 1.31 13 /a Tax effort tax collections as a percentage of taxable capacity. - 25 - and discourages local governments from spending a greater share of their budgets for capital purposes. Only two taxes are specifically earmarked for capital construction and maintenance: the UCMT and the public utility surcharge. In theory and rhetoric, these taxes are intended to nncourage spending to improve the capital stock, but they remain a very small share of total local resources and of total capital construction financing. 74. A second constraint is the absence of a formal mechanism for local governments to borrow for capital construction, even if repayment potential is not in question. Long-lived and expensive projects must be financed from current general revenues, ad-hoc grants from higher-level governments, accumulated savings from current revenue, or enterprise "contributions." The inability to borrow shifts the full burden of financing a project with future benefits onto current taxpayers and the general public, rather than on future users of the facility and specific beneficiaries. This raises the price of infrastructure investments and enhances the relative attractiveness of expenditures with current benefits.17/ 75. Second, there are no formal programs of "benefit" or "user" charges to finance capital projects. Even where project beneficiaries exhibit a strong willingness to pay, there is no mechanism to tap this willingness, so that desired projects may go unfinanced. Instead, localities have turned to ad-hoc levies and "fiscal predation" of enterprises, both of which contribute to the erratic and nontransparent nature of the revenue system. Eaualization Properties 76. Income distribution issues also arise because of the design of the present system of intergovrnmental relations. Income inequalities appear to have increased between rich and poor provinces in China, and may have been accentuated by fiscal decentralization which benefits better-off provinces through increased local tax retention powers. In many countries, the fiscal system is pointed toward reducing income differences among provinces, the goal being to raise the poorer provinces' fiscal capacities and per capita expen- ditures, so that the inhabitants of rich and poor provinces receive a more comparable level of services. In many countries, equity is promoted by grants or transfers specifically targeted to poor areas, as distinct from transfers directed at achieving vertical balance. 77. China's revenue-sharing system does not appear to go very far towards equalizing per capita expenditures. A major reason is the inherent conflict between equity and incentives for revenue collection which have led to low contract quotas, the benefits of which accrue to better-off provinces. The objective of equity would be served if better-off provinces remitted more in order to transfer greater amounts to poor provinces. The government has not yet faced up to this inherent conflict in designing the current system. In fact, one could make the argument that the current system is intended to enable provinces to maintain a historical, benchmark level of expenditures, which were themselves unequally distributed in the first place. The result is that greater emphasis has been given to stimulating tax effort and to supporting historical levels of expenditure than to promoting interprovincial fiscal equalization. 78. A hypothetical example (see Box 3) illustrates that, although the reve- nue-sharing system in China makes net transfers to poorer provinces, these do not go far towards equalizing fiscal capacities between richer and poorer 17/ There is a similarly low capital expenditure share in the extrabudgetary accounts, suggesting a similar effect at work. See also Remy Prud'homme: "Urban Finances in Shanghai," IBPD mimeo, for a discussion of urban and local financing reforms. - 26 - Box 3: EQUALIZATION PROPERTIES OF CHINA'S PROVINCIAL CONTRACTING China's present revenue-sharing system involves Central-Provincial contracts of three types: (i) for high- and middle-income provinces, the center requires a "quota', delivery and applies a zero or near-zero above-quota sharing rate on above-quota collections, to provide coLlection incentives where taxabte capacity is greatest; (ii) for some, mostly middle-income, provinces, there is "aroPortional sharjnM," in which the same sharing rate is applied both to the quota collections, and above-quota collections. This has been mostly phased out now and replaced with a simple "fixed auota delivery as in the high income provinces;n and (iii) a lump-sum transfer is given to most poor provinces. An important question is whether this approach helps to equalize expenditure capacities in the various provinces. The example below explains this in a hypothetical case. The simulation tracks the growth in expenditure capacity (i.e., the amount that the local governments are able to spend) under this system over a 3-year period in which revenues are assumed to grow by 20 percent annually. Provin- cial expenditure capacity is defined as collections, plus or minus the transfers to/from the central government. The transfer to deficit provinces is assumed to be Y 50, fixed for five years. The simulations show that: (i) in the deficit provinces that retain all revenues collected and receive a fixed transfer, expenditure capacity grows the slowest; (ii) in provinces where proportional (25 percent) contracting applies, expenditure capacity grows at the same rate as revenues (20 percent); and (iii) under the fixed-quota contract, expenditure capacity grows faster than collections. By implication, the richer (high tax yield) provinces' expenditure capacity grows far faster than the deficit provinces that receive direct transfers. At the end of year 3, the expenditure capacity under the fixed contract system has grown by nearly 60 percent; that under proportional sharing by 44 percent; and that of the transfer recipients, by nearly 30 percent. Box Table 1: HYPOTHETICAL EFFECTS OF REVENUE SHARING ON EXPENDITURE CAPACITY (Yuan) index of Transfer expenditure to/from Expenditure capacity Collections center capacity (En + E.) Exam le 1: Deficit Province (Receives Fixed Nominal Transfer), Year 1 100 +50 150 100 Year 2 120 +50 170 113 Year 3 144 +50 194 129 Examtle 2: Prooortional Contract (25X sharing) Year 1 200 -50 150 100 Year 2 240 -60 180 120 Year 3 288 -72 216 144 Examnle 3: Contract System (Pays 25X of Base Year Collections. Fixed In Nominal Terms) Year 1 200 -50 150 100 Year 2 240 -50 190 128 Year 3 288 -50 238 158 provinces. If, in addition, the poorer provinces cannot or do not make the same tax effort as better-off prov4.nces, the disparity in budgetary revenues available would be even wider. (See Annex 2 for a formal treatment of the counterequalizing effects of the system.) China's actual experience shows that expenditures in China's richer provinces grew faster than in the poorer ones, and per capita expenditures in better-off provinces are also significantly higher (Annex Tables 4 and 6). - 27 - 79. For a tax-sharing system to be effective in equalization, it must first identify what is to be equalized, and then it must find some objective measures to allocate resources to achieve these objectives. Indicators of interregional differences and/or economic status must be made explicit in the allocation criteria of the transfers. For example, in Brazil and India (two countries with wide regional disparities) some 50 percent of the funds allocated among the states are based on "indicators of need." Indonesia and Malaysia use other even more detailed indicators of need, such as an index of physical infrae_ructure (road and transport network, etc.) and social infrastructure (schools, hospitals, etc.), per-capita income and population.18/ C. OPTIONS FOR REFGRM 80. The economic system reform requires a change in the roles of central and local governments in China. China needs an intergovernmental system which: o is consistent with maintaining an appropriate income elasticity of revenues O provides incentives for public infrastructure development; o which supports central government macroeconomic management; O will influence resource allocation in line with national goals and priorities; O will promote income distribution goals; and, most importantly, o supports and is consiatent with system reforms such as price and enterprise reforms. Some Alternatives 81. The foundations are now bein7 laid for a reform of the intergovernmental fiscal system, and a number of spetcific measures have recently been considered. While a decision was made in the December 1990 Plenum of the Communist Party to retain the present system for another five years (i.e., through 1995), the present system is not fully adequate for either level of government and fiscal reform remains a very live discussion. 82. Several different fiscal reform proposals either are under discussion in government or have been considered in the past. These are: 0 Strengthening the local tax base by designating more taxes as strictly local and not subject to sharing; 0 The reassignment of certain taxes (responsibility for both policy and administration) to the central government, and others to the local government; O A local surcharge on the national government profits or sales taxes; O Continued revenue sharing with a reduction of provincial sharing rates to increase local tax effort, supported by an equalizing transfer eystem. 18/ A survey of formula tranqfers is included in Roy Bahl and Johannes Linn Urban Public Finance in Developing Countries, Oxford University Press, forthcoming. - 28 - O Continued revenue sharing with an increase in the provincial sharing rates to increase central revenues, accompanied by a transfer system. O A more piecemeal reform of the oresent system. 83. There is no easy choice among these models. Certainly one could not look to the experience elsewhere in the world to find one system that works best. Aiternative structures of intergovernmental finance differ widely among countries, and the "right" choice depends on political considerations, regional differences, constitutional requirements and administrative c&pabilities. The most basic issues that differentiate systems are the command of each level of government over fiscal resources, and the resultant independence in spending these resources. 84. Whatever the tax assignment between levels of government, fiscal transfers are usually required to get a better correspondence between revenue- raising capacity and expenditure needs. "Vertical balance" of revenues and expenditures--the amount raised and the amount spent at each level of government--is seldom perfect, and it ought not to be, for several reasons: first, central governments (particularly in developing countries) have a decided advantage in revenue raising. However, many expenditures are better planned and delivered by the local level, where local preferences can be considered and the service can be better "fitted" to the local situation. For this reason alone, intergovernmental transfers are needed. A second justification for intergovernmental transfer is to finance services with "externalities," such as those which contribute to meeting national health or education objectives, and those where purely local financing could lead to underprovision from a regional or national perspective. Third is the equalization issue. with wide variations in resource endowments, intergov- ernmental compensating transfers may have a redistributive role to play among regions. There also is a political motive. Intergovernmental transfers allow the central government .o maintain control over rG'venues--depending on how the grant system is designed--while delegating expenditure delivery responsibility to local governments. For all of these reasons, one would expect to find intergovernmental transfers as part of the fiscal system in all but the smallest and most centralized countries.19/ Box 4 outlines some international experience with different systems. Box 4: FISCAL FEDERALISMS AND CENTRALIZED SYSTEMS More countries are organized as unitary than as federal systems. Under a unitery system, there is not a legal statement of the powers and responsibilities of subnational governments, and provincial government fiscal powers are given through direct central regulation. The important issue here is "fiscaL autonomy," the control over sufficient resources to plan and manage the provision of local public services without continuous interference and control by higher level authorities. The size of local government (expenditure) and the revenues of local govermnent are important indicators. However, the degree of autonomy may also be defined by: (a) whether there is latitude in revising tax rates and bases; (b) whether borrowing powers are circumscribed; (c) whether the local bidget is subject to higher level approval and monitoring; and (d) whether the local officials are appointed by a higher level gov- errment. Evidence suggests that there is considerable divergence on the exoenditure side from country to country. On the whole, local governments in most OECD countries and responsible for the delivery of most direct services to citizens, such as primary and secondary education, health, social welfare, housing, and the provision of local services such as street repair, refuse removal, and the like. Local goverrnments in many developing countries have similar responsibilities but more often the central government tekes a Larger role, with overlapping functions in some cases. 19/ This is also true in the U.S. where federal grants to state and local government expenditures account for 16 percent of total state and local government revenues and 10 percent of total federal government expenditures. - 29 - Box 4: F.-CAL FEDERALISMS AND CENTRALIZED SYSTEMS (Continued) The size and pattern of local government revenues also varies greatly from country to country. One common element, however, is that in no country do local taxes come close to financing local expenditures. In nine (European) countries for which information is availabLe, Local taxes accounted on average for 41 percent of local revenue in 1988. Non-tax revenue (such as user charges) for .O percent and grants for 39 percent. There is also considerable variation from country to country in the role of grants: in most countries grants include both general grants, often with an explicit equalization element, and specific grants of many varieties. Local goverriment authority to adjust tax rates and to enact new taxes is limited in most countries, but is more limited in developing countries. In general, the national or state law prescribes the tax bases availabLe (or unavaiLable) to local goverrvnents and sets maximum rates within which local goverrunents must operate. These restrictions usually hold even for the largest cities. When the rate ceilings are binding, Local governments have little revenue discretion and are dependent on the higher level government for approval of every revenue proposal. A similbr arrangement holds for the adju:.ting of user charges for most major services, e.g., water rates, bus fares, rents. The issue then becomes whether or not the approving central or state goverrnent will permit the requested increases in rates and charges. Experience varies but some countries have consistently refused requests for local rate increases, e.g., cities in Bangladesh have been held at 1960 property tax rates despite repeated requests for increments. All countries are not subject to sJch stringent controls. Brazil and Venezuela are among the exceptions in that municipal laws are not subject to approval by higher level goverrinents, though some tax changes do require approval by a central agency. A number of conclusions are suggested by th^ range of outcomes sketched above. First, national governments clearly exercise considerable discreticn in deciding how large a role local governments play, the extent to which local activities are financed from local revenues, and the types of taxes levied by local governments. Chart 1 below shows, for a variety of countries, the relative importance of the subnationat fiscal sector in overal national finance, and the degree of fiscal autonomy which the sector has, as measured by the importance of its "own revenues" in total spending. The size of the subnational sector in China, at 45X, is not insignificant. With respect to the degree of revenue autonomy, formally defined, China's subnational governments are retatively revenue dependent, with own-source revenues only 11 percent of their total revenues. This compsres with the subnational sector in countries such as Korea, and Indonesia, which are also large in the overall financial picture but are not self-financing. De facto, China's localities are, conparatively, much more revenue independent than these most. At the other extreme are countries such as ChiLe, Kenya and Sri Lanka, whose subnational governments, inportance in the overall fiscal picture is small, but which are self-financing. Chart 1: Subnational Finance: Revenue Independence and Scale of Subnational Sector Subnational government spending (percentage of total government spending) 80 Yugoslavia . 60 _ India * China* 40- * Republic of Korea Braz * n Colombi * Argentina * Colombial Indonesia Romania The Gambia 20 - Mexico Thailand Philippines * Costa Rica Malawi Chile \ Kenva Tunisia * * ' \ 0 I SriLanka I 0 1 30 50 70 90 Own-source revenue (percentage of gross revenue) Source: R. Baht and J. Linn, op. cit, and R. Bird and C. Wallich; "Local Goverrwnent Finance in Hungary" World Bank Working Paper, forthcoming; Chart source: 1988 WDR, World Bank. - 30 - Principles of Tax Sharina and Tax Assignment 85. There is no ideal assignment of taxes between central and local governments. However, there is general agreement that responsibility for setting the rate and base of certain tax sources is more appropriats for local governments and while for other taxes, the national government should have such responsibility. These principles of "tax assignment" relate to the respective responsibilities of central and local governments in resource allocation, income redistribution, and macroeconomic stabilization and growth. In developing countries there is the issue of the administrative capabilities of local governments; in large and diverse countries there is an issue of tax harmonization; and in China, there is the added dimension of compatibility of the state-local system with China's economic reforms. 86. The stabilization-growth objective of the government fiscal system clearly calls for central control over the money supply and most debt management. Taxes on international trade and a substantial share of income and general sales taxes should also reside with central governments. The central government must have the tax instruments to deal with a central deficit or with inflation. This does not mean that local governments should not have access to sales or income taxes. Rather it means that the assignment of these taxes between levels of government should recognize the need for the central government to use tax policy in controlling the macroeconomy. Where local governments do have taxes with stabilization properties, stabilization policy may be made more difficult if local government expenditure policy conflicts with the center's stabilization objectives. The converse of this is that central governments can tive with more unstable sources of revenue because they can borrow to maiize up the shortfall. Local governments, by contrast, require relatively etable sources of revenue. 87. There are "special" taxes that ought to be assigned to the central government, if they are at the center of national policy. The tax on motor fuels, and perhaps that on major natural resource-based exports, are examples. 88. If the assignment of taxing powers were to be made solely on a basis of macroeconomic control considerations, local governments would not have access to any of the broad-based taxes. But government also has an allocative function, i.e., it decides on how much of national (local) resources shall be spent on which government-provided goods. Here local governments should play a greater role, it can be argued, because they are closer to the people and better able to take the pulse of the local population for public services. The efficiency argument for local government revenue raising powers would be based on the types of local services delivered: O general purpose local services whose benefits do not spill across local boundaries should be financed by general purpose local taxes; O services that are locally delivered and can be "priced," e.g., public utilities, should be financed with user chzrges; and O services with benefit spillovers should be financed either directly by the central government or via revenue shariny. The Chinese system does not even approximate these norms, indicating that efficiency considerations play little role in the assignment of functions to local governments. 89. In China, there are other important efficiency considerations that pull the assignment of taxing power toward centralization. It is important that the fiscal system accommodate the enterprise restructuring through mergers, conglomerates, joint ventures or joint stock companies. This suggests, a priori, that enterprise income taxes should be central government taxes. To - 31 - The Chinese system does not even approximate these norms, indicating that efficiency considerations play little role in the assignment of functions to local governments. 89. In China, there are other important efficiency considerations that pull the assignment of taxing power toward centralization. It is important that the fiscal system accommodate the enterprise restructuring through mergers, conglomerates, joint ventures or joint stock companies. This suggests, a priori, that enterprise income taxes should be central government taxes. To the extent that price reforms generate major shifts in value added across provinces, a case can be made for centralizing a portion of the VAT, and using its proceeds to offset these effects, at least until the transition is largely complete. 90. There are also equalization norms that can be used to guide the decision about which taxes should be local and which should be central. The distributive function of government would seem to be an argument for a centralized taxing system in China. If there are wide disparities in income and wealth across provinces, as there are in China, then local taxing powers can exacerbate these disparities. Central taxing powers coupled with an equalizing distribution of transfers can partially offset these disparities, as can the subsidization of "priceable" public services. The present situation in China is one where such equalization has been an important part of the system, but where the present trend is in the direction of less equalization through the fiscal system. 91. The capacity to administer taxes is always an important constraint to the assignment of taxing powers in low income countries. If local governments cannot effectively aseess and collect the taxes, then the system will tend toward centralization. This is not an issue in China because the tax system is already administered by the local government sector. In fact, the administrative question in China is whether the central government is able to effectively control and monitor the local tax administration syetem. The present situation is one of de facto local government taxing powers because the central government is unable to control the local administration (or because it has chosen not to). 92. A final consideration is tax harmonization. There are certain taxes that simply ouaht to be central because it is too difficult to try and harmonize them acrolss provinces. The two most important examples are customs duties and company income taxes. In the former case, the revenue would accrue to the entry province, or every province would levy an .ntry tax on commerce.20/ Neither are workable alternatives. In the case of company income taxes, the problem with allowing provinces to levy separate rates is how to assign the tax base and therefore revenues to each province.211 In China there is the special problem of how to compensate for firms whose profits are a result of national subsidy. Box 5 summarizes one poss.ble approach to tax assignment. 93. The conclusion from this discussion is that there is no "best" way to divide taxation responsibility between the central and local governments. The 20/ Indian cities levy such a tax, known as "octroi." It is charged against all products entering or passing through the corporate boundaries of a city. It is revenue productive but widely criticized for being disruptive to interstate trade and corrupt in its administration. 21/ U.S. states levy separate company income taxes, and divide the taxable base for multistate firms according to a complicated three factor formula. In Switzerland, local jurisdictions (the cantons) are allowed to levy corporate income taxes as well as personal income taxes and natural resource taxes. - 32 - Box 5: PCINCIPLES OF TAX ASSIGNMENT The table below outlines one possible arrangement of tax assignments that derives from the principles Just outLined. As shown in the table, taxes usuatly tevied at the national level inctude resource taxes and personal income taxes. Customs duties are also almost aLways national-tevel taxes because of their strategic importance, as are resource and income taxes because of their important role in stabitization and distribution policy. In theory, customs duties and income taxes coutd be Levied at uniform rates in all localities with the same resutt, but this woutd be harder to administer. Sales taxes are often thought to be best levied at the provincial level, If the region is big enough to avoid revenue loss from consumers shopping across the border in provinces with tower taxes. The revenue growth from sales taxes is usually stable. Excises and surcharges are also appropriate as provinciat taxes. Box Table 1: TAX ASSIGNMENT Centrat level taxes Provincial and Local taxes Personat income tax Surcharges /a Corporate income tax Sates taxes Natural resource taxes Excises Excises Property Taxes Customs duties User fees charges Surcharges La /a Preferabty at rates standardized across localities to avoid inequities across regions and distortion of prices. The actual practice of tax assigrvnent does not aLways parallet the above principles. For exanple, in Switzerland and the United States, tocaL jurisdictions (the cantons and states) are atlowed to levy corporate income taxes ap well as personal income taxes and natural resource taxes. case for centralization is usually built around macroeconomic considerations and equalization, and the case for local government taxing powers around efficiency considerations. The "best" way to do thingr depends on how the goverinment weights these considerations. In China, the weighting has come out, de jure, in favor of a very centralized system, i.e., provincial and local governments have very little tax policy autonomy. It might be argued that the present arrangement is too centralized for China at its present stage of development and the time is now right to grant more local autonomy. 94. As noted above, there are several problems with the current Chinese version of fiscal centralization. First, local government revenues are procyclically unstable, while those of the center are too fixed. Second, local governments' free hand in the implementation of tax policy e..ables them to effect major changes in the incentive structure facing economic agents. Third, most local financing is from shared taxes, and the present sharing formulae are moving away from equalization of interprovincial expenditure capacity. Finally, the present system complicates price reforms, because of its revenue impact on different provinces (there would be major gainers and losers), and interferes with enterprise reforms in that new organizational modes such as joint ventures or mergers, especially across taxing jurisdictions, are effectively discouraged. Considerations in Weiahing the Options 95. The major questions to be faced up to in evaluating the choices for fiscal decentralization are whether local governments will take on more responsibility for financing local services (in particular, whether local - 33 - governments will be given some degree of revenue-raising autonomy); whether tax administration can remain a local government responsibili.ty; and whether the intergovernmental transfer system should include a formula-based transfer program. 96. The six models under discussion each have differing implications for consistency with and support to reforms, stabilization policy, equalization potential, resource allocation, tax administration, and the extent and nature of the autonomy given to provinces. Their pros and cons are evaluated in turn in the following section and summarized in Box 7. Whatever model is chosen, or even if the present system is essentially retained, thero must be provision for (i) the institution of new methods of infrastructure financing, (ii) improved fiscal planning, and (iii) a strengthening of the local tax administration. Options (i) Option 1: Tax Reassiqnment: A Centralized Version 97. Reassignment of taxing authority--giving local governments the power to introduce their own taxes and to set tax rates--is one reform possibility that would fit the objectives outlined above: it could raise more revenue and, if tax rates can be set by localities rather than by central mandate, it could make local officials more accountable to their populations for the quality of services delivered. Successfully implemented, it could stimulate the tax collection of local governments. This approach, applied in its extreme form, would call for fully separate central and local taxes, and an end to the shared tax system. The three questions to resolve would be: which taxes to give to each level of government, how to administer the center's taxes (since they are currently collected by local governments), and how to design an appropriate transfer system. 98. One possibility, call it "Version A", is a very centralized approach under which the enterprise income tax and the product, businese, and value- added taxes would become fully central revenues. Provincial governments would be given minor taxes (e.g., the 13 "fixed local taxes" assigned to them in 1985) and allowed to introduce new local taxes and user charges. There would be separate central and subnational tax administrations because local governments would be unlikely to aggressively collect central taxes in which they do not receive a share. 99. If provincial and local governments were given only the minor taxes, they would be unable to finance all services for which they are now responsible. One response to this mismatch would be for the central government to assume direct responsibility for provision of certain services. This option is not in keeping with the tenets of system reform and increased decentralization; the adju;stment costs would be very great; and, moreover, China is simply too large and diverse a country for a centralized public expenditure system to be managed efficiently. A more likely solution is to allow local governments some additional taxes and to create a regular program of central grants to rake up the revenue shortfall. The grant could be distributed on a formula basis, with the elements of the formula chosen to reflect provincial income, or need (see below). For example the grant distribution among provinces might be based on some combination of per capita income level, population size, some indicator of infrastructure needs, urbanization, etc. 100. A very great advantage of this centralized version of tax assignment is the flexibility that it gives to the central government to influence interregional equalization and the sectoral composition of investment. It also puts the government in a better position to use the tax system for macrostabilization goals. And, because it merges responsibility for tax structure and tax administration, the tax system can be used to achieve - 34 - allocative goals. The latter point is especially important: the tax "levers" that the central government designed would be more effective. This option-- since the center would receive all profits and sales taxes--would also reduce the fiscal impact of price reforms and make changes in enterprise organizational forms irrelevant to provincial government finances. The biggest disadvantages to this approach are the loss of provincial and local control over revenues, the reduced incentive for revenue mobilization by local governments, and the increased expense associated with two separate tax administration machineries and with the maintenance and operation of a grant system. 101. This approach is used by most low-income countries that have unitary forms of government, but it may not be feasible for China. There are few local taxes currently in place, and together they account for 19 percent of total taxes collected, 9 percent of provincial and local expenditures, and about 1.5 percent of GNP in 1985 (see Table 5).22/ To increase the local tax share significantly beyond this would imply allowing local governments to introduce new taxes with significant revenue potential (see below) or making a formal reassignment of one or more major taxes to the local governments [discussed as Option (ii) below], qlternatives which do not seem likely. 102. Of the various options for increasing local revenue-raising powers, the land use tax is prominently mentioned. The rationale for levying this tax is straightforward. Since land owned by the state has a location value, it is appropriate to charge for its use, and to charge accordins, to its location so aa to avoid the overuse of land relative to other factors of production. The problem with a larj-use tax lies less with its justification in a socialist economy than with its imple.nentation. Valuing property and location value is difficult when there is no formal market in which land is bought and sold. Similar'.y, taxes may not induce a better use of arban land when enterprise location mobility is very limited. The adequacy of the existing cadastre and the problems of recordkeeping are additional issues. Finally, there is the question of the land tax's revenue potential. Unless levied at a high rate, it is unlikely to have any significant revenue impact or to influence decisions; however, a high rate may not be feasible due to constraints or lags in other system (e.g., price) reforms.23/ 103. User charges can generate more local resources, cover the cost of services, and charge users for benefits received. Transport, water supply, gas and housing are each frequently cited as important services where present charges are nominal. While at first glance, increased user charges might seem to be a way of financing certain services, the situation in China is more complicated, and it is not clear how large a revenue boost there would be. Increased user charges would lower the profitability of consuming enterprises and thereby profits tax revenue, since services are tax-deductible costs. This would be partially, but not completely, offset by an increase in sales and profits tax revenues from the public utility enterprise, since public utilities are typicalXy taxed at preferential rates. Localities may well receive a net reduction in revenues if public utility user charges are increased. However, it also could be that, where government subsidies to the utilities have been necessary, increased user charges will relieve some pres- 22/ The "13 own taxes" are the Urban Maintenance and Construction Tax (UMCT), vehicle utilization tax, the profits tax on collectives, land occupancy tax, vehicle tax, house tax, slaughter tax, animal trading tax, free- market transaction tax, salt tax, agriculture tax and local vehicle utilization tax. 23/ For a detailed discussion of the land tax use in China, see Roy Bahl and Jun Zhang, mimeo, World Bank, 1988, "Land Use Taxes in China." - 35 - sure on the general local budget. In short, the net impact of increased user charges in the local government budget is not at all clear. (ii) Tax ReassiLnment: Decentralized Version 104. Another approach (call it "Version B") would reassign to local governments one or more if the productive tax bases, such as the sales or profits tax. This would provide sufficient or nearly sufficient revenues for the higher income provinces, and only an equalizing supplementary grant scheme would be necessary. This solution would give provincial governments a considerable amount of discretion in determining the level of revenues and expenditures, and would increase the economic decision making role of local governments. There is precedent for this more decentralized approach. The assignment of broad provincial and local taxing powers can be found in large countries such as the United States, Brazil, Colombia and Nigeria, where state and local governments raise 25 to 50 percent of all taxes. 105. The major issue is the choice or taxes to hand over to subnational governments. In practice, the central government is unlikely to give up either of the two major taxes, and it is not at all clear that it should. The sales tax is the major revenue producer in China and will almost certainly not be decentralized, although it would be a good candicate for this. Most central governments levy a general sales tax (the United States is the most notable exception), though many share the proceeds with subnational gov- ernments.24/ The profits tax would be a good choice in that it is revenue- productive, but, as a purely local government revenue, it would raise a number of problems. Local budgets would be at the mercy of the typically cyclical behavior of profits and the impact of central government macroeconomic, price and wage policy on the tax base. Local governments whose enterprises continue to be affected by central price controls would be sorely disadvantaged. Moreover, the assignment of the profits or product taxes to provinces would lead them to be very unequally affected by system reform. Price reforms that affected, say, intermediate goods-producing provinces would benefit them vis- a-vis user-provinces. Full local retention of the profits tax could continue to discourage new corporate forms, until rules for the sourcing of profits across business units are devised. 106. The second major issue is determinirg the required revenue base. This requires, first, clarifying the expenditure responsibility of the central and local governments, respectively. In China, these responsibilities are not clearly separated; and the issue of revenue adequacy cannot be discussed separate from expenditure responsibility. The assignment of the profit or sales tax to local government would almost certainly require separate central and local administrations, requiring duplicate audits, reporting requirements, and the loss of economies of scale in tax administration. 107. Finally, this option would be counterequalizing, in that the highest income provinces would generate the greatest amounts of revenue. For disadvantaged communities, a program of compensating grants will be required. The challenge is to design a compensation program that will equalize, but not dampen the incentive to mobilize local revenues. As the experience in many countries will i test, designing and implementing such a system is difficult. 24/ Note that if the VAT were made into a provincial tax, rates would have to be uniform in all provinces to avoid special administrative problems. - 36 - (iii) The "Local Surcharge" or Tax Base Sharinq Option 108. As an alternative to reassigning taxes or allowing localities to levy new taxes, localities could be given the option of "surtaxing" the present national taxes at a prescribed higher (local) rate. This model involves a sharing of the tax base, and is fundamentally different from a reassigmnent of taxing power. Provincial governments would be permitted to levy (withLn a range) a surtax rate on the enterprise, product, business, and/or value-added taxes. In the case of the enterprise tax, a basic rate (say 35 percent instead of the present 55 percent) would belong to the central government; provincial governments could levy an additional rate of up to, say, a maximum of 20 percent. Alternatively, the subnational tax might be levied as a percentage of the central tax liability, in much the same way as the present UCMT is now calculated as a percent of sales tax liability. (Box 6 illustrates how a local surcharge might work). The UCMT offers an interesting precedent for this approach, differing from the proposal here only in that the surtax rate is fixed by the central government. 109. Whether revenues under this option would be adequate to meet provincial expenditure needs depends on how the surtax limits are set. If, on average, rates were set to reflect expenditure needs (again, these would have to be defined), higher-income provinces could generate adequate revenues but the lower-income provinces could not. Equalizing formula grants would be required to compensate for the low fiscal capacity of some provinces. If local tax administration is retained under a ba"e-sharing program, provincial and local governments coula not be permitted to .'ngage in any tax relief policies that would affect the ba-a or rate of the ce'tral government tax. If provincial governments chose z3 provide tax prefereices, this could be done only by reducing their own tax base or surcharge -ate. The effect of this principle is illustrated in the two cases in Box 6 balow. Box 6: PROFIT AND SALES TAX SHARING UIDER A BASE-SHARING MODEL The simple numerical exanpte here describes one version of how an enterprise would be treated under such a system. The central -ales and profits tax rates are assumed to be 10 and 20 percent, respectively, and, to make matters b;mple, the tax bases are gross sales and gross profits, respec- tively. In case A, the provincial government chooses rates of 5 and 15 percent, with no preferential treatments, with the result that it collects Rmb 950 from this enterprise by comparison with Rmb 1,600 for the centraL goverrment. Case A Case B Gross Sales (Rmb) 10,000 10.000 Central sales tax at 10X 1,000 1,000 Provincial sales tax at 5X 500 Provincial sales tax at 3X - 300 Less Production Expenses 5,500 5,500 Eouals Gross Profits (Rmb) 30 3.000 Central profits tax at 20X a/ 600 600 Provincial profit tax at 15X 450 - Provincial profit tax at 10X - 320 Total Central Revenue (Rnb) 1.600 1.600 Total Local Revenue (Rmb) 950 620 In case B, the provincial govermnent offers a preferential treatment to this enterprise and limits the sales and profits tax rates to 3 and 10 percent, respectively. Central revenues would not be affected, Rmb 1,600 is still raised in Case B. but local revenues would be reduced to Rmb 620 because of the tax abatement. YWe make the assumption that provincial and local government sales taxes would be deductible for provincial and local income taxes but not for central income taxes. - 37 - 110. A choice would have to be made as to which taxes to surcharge. The sales and profits taxes would be the candidates in China. One could make strong arguments for the sales tax: (a) a major problem with an income tax surcharge would be prorating profits between provinces in the case of multiplant firms; (b) a sales tax would cover more firms (those making profits as well as those not making profits), hence the financing of local services would be more broad-based; (c) more efficient enterprises would not be penalized with a higher rate of profits tax; and (d) local accounting practices are such that total sales value can probably be estimated more accurately than total profits.ZS/ The principal argument in favor of the profits tax is that it is a better ability-to-pay indicator. 111. Compared to the local tax option and more balanced tax reassignment models discussed above, this approach implies less central governance of the tax base and considerably more local fiscal self-determination. There are disadvantages to this shared-base approach. Provinces with a stronger economic base would have an advantage and the supplementary grant program would have to be created and maintained. The local tax administration system would have to be carefully monitored by a better management information system (MIS) than presently exists to ensure proper collection of the center's tac share. The central government would also give up some control over the tax system's stabilization and allocative properties at the margin, especially if the enterprise tax were surtaxed, and therefore would have less leeway in using tax policy to pursue stabilization and macroeconomic go&ls or to address allocative objectives. A major consideration in structuring a surcharge would be how to recalibrate the system (through grants or other mechanisms) to offset the effects of price reforms on fiscal disparities. To the extent that provinces continue to generate a major share of revenue from profits tax surcharges, this could impede enterprise reforms. 112. On the other hand, the surcharge has great advantages over the alternative of designing and implementing a fully new tax. The administrative structure to assess and collect the surcharge is already in place and at the right rate it can be a substantial revenue producer, and would give local governments some revenue autonomy. Compared to a low-yield land tax, a profits or sales tax surcharge are clearly superior as revenue-raising measures. Thus, despite the disadvantages, one might be able to make a good case that this could be an appropriate approach for China. It would allow provincial governments to continue to use taxes to promote local industrial policy, while allowing the central government to retain some control over the level of taxation and complete control over the definition of the tax base. This system would give the local governments a significant incentive to improve the efficiency of their tax administration. It might also be argued that, by comparison with the present system, central government macroeconomic control would not be compromised markedly. 25/ In the case of turnover taxes, the addition of a local surcharge poses few conceptual or practical difficulties. An invoice-based VAT could raise some difficulties. One problem is that provinces specializing in intermediate goods would almost certainly prefer to "export" taxes to other provinces rather than zero-rate their own "exports." A true destination VAT would require some kind of adjustment for taxes on "imported" goods, as will now be the case in EEC countries. However, it is doubtful that local governments would wish to give credit for "foreign" taxes on inputs purchased from other provinces. While the experience of VAT harmonization in the European Community offers some parallels here, it appears unlikely that a regionally rate- differentiated VAT could operate satisfactorily in China without central control. See Tait, Alan, VAT: International Experiences and Problems, 1986. - 38 - (iv) Higher Provincial Retention Rates 113. Another way to increase subnational revenues is simply to change the sharing formula, i.e., to allow local governments to retain a greater percentage of what they collect. The central government is already moving in this directicn by increasing the marginal retention rates for provinces with a greater taxable capacity, and by entering into special contracting arrangements with certain provinces (e.g., Shanghai). The sharing rate is in fact used as a policy instrument within provinces, where retention rates for cities vary widely. (Low-income cities/counties usually are given a higher retention, with lower retentions applied to the richer urban areas.) 114. This approach has the strong supports many of China's provincial govern- ments. Its effects, however, are uncertain and require further analysis. On the positive side, substantial revenues would remain at the local level and resources available for allocation to local projects would increase. A greater local tax share would create a greater sense of local autonomy, and local officials would have a greater incentive to improve tax administration. System reform which gives more autonomy to enterprises combined with an increased sharing rate would reduce the gains from collusion between local governments and enterprises and would increase overall resource mobilization. In all of these cases the issue is whether the magnitude of the response would be significant. 115. Increased local retention also has serious drawbacks. The central government could view an increased provincial share as a further drain on its total revenues. This would exacerbate the center's budget problem and reduce its ability to reallocate revenues among provinces. This proposal would, therefore, involve a trade-off between encouraging more revenue mobilization by local governments on the one hand, and lessening the redistribution of fiscal resources from richer to poorer areas, on the other. Therefore, it could well turn out to be a counterequalizing measure. Finally, inasmuch as price reforms will affect provinces tax bases, according to whether they are producing or consuming goods which are liberalized or continue to be subject to controls, higher local retention will need to be linked with some method of recalibrating the provincial quotas to offset the impact of price reform on the revenue base. (v) Lower Provincial Retention Rates 116. The obverse approach is the choice of the center. A lower provincial retention rate would, it is hoped, give the center more revenue, assuming localities do not respond by reducing their tax collection efforts. The greater revenue base could improve the center's ability to reallocate revenues among provinces and equalize resources through a grant system. It could provide the center with more resources for nationally important investment projects, or projects spanning several provinces, and would strengthen its ability to use fiscal policy for stabilization purposes. The major question here is whether central revenues would rise, or whether local government reductions in tax effort would be the dominant impact. Provincial and local resistance to this option would be strong. (vi) i&eformina the Present System 117. All of the schemes reviewed above are "big" changes. They call for eliminating the existing tax-sharing system, giving local governments some rate-setting autonomy, creating a grant system, and changing the nature of responsibility for tax administration. Another possibility is that reform will go more slowly, and a decision will be made to eliminate some of the most objectionable features of the present system while developing its strengths. Among the advantages of the current system are that its local orientation fits some of the goals of decentralization. It recognizes that provincial - 39 - officials are in a good position to determine how tax relief can best stimulate the local economy, leaving it to them to decide between tax contracts to stimulate production, a tax holiday to protect a pioneer industry, a one-year abatement to help an enterprise through a cash flow problem, etc. There is an efficiency case for the ad-hoc approach since each local government is able to trade between preferential tax treatments and revenues for general government services. 118. On the other hand, there are arguments against the ad-hoc, decentralized approach. First, because local governments can grant tax relief, the present system probably reduces, rather than increases, general government revenues, at least in the short run. Second, its ad-hoc nature produces a very uneven pattern of effective tax rates across enterprises. Whether this horizontal inequity is acceptable depends on the objectives of the local economic plan. Third, the ad hoc approach interferes with efficient tax administration. Proliferation of special treatments for enterprises makes the tax system more complicated and more difficult for local officials to administer. Moreover, there is almost certainly some tendency to be lax in assessments when there are so many possibilities for special treatment. 119. But perhaps the most severe criticism of the ad-hoc approach is that it destroys the notion of a system of taxation. The central government increasingly discusses the pcssibilities of using the tax system as a "lever" to influence economic activity. However, if local governments can change the pattern of effective tax rates without agreeing such changes with higher level governments, then the intent of using the tax system to influence economic activity can be defeated. There is a direct trade-off between the central government's objectives to indirectly 'nfluence the allocation of resources in the economy, and decentralization in the form of power to grant tax relief. 120. The central policy question becomes whether it is possible to allow local governments some revenue powers as an incentive to increase revenue mobilization and, at the same time, allow them to retain the power to grant abatements, holidays, and relief through contracts. To the extent the government's objective is to create more uniformity in its system, the proper strategy iB to move toward eliminating the ad-hoc power of local government to influence the distribution of effective tax rates. To the extent decentralization is more important than the use of taxes as an economic lever, there might be some argument for a continuation of the present practice. The granting of increased taxing powers and a continuation of the present ad-hoc practices, however, are not compatible. 121. These considerations suggest that, if the current system is to be "patched up," the areas where reform is most needed are: O rationalizing the rates of central-provincial revenue sharing; O rationalizing the system of provincial grants; O bringing the implementation of central tax policy under control; and O modernizing the systam of tax administration and financial management. 122. The following changes should be made in the system of central-provincial sharing rates. First, the trend to fixed nominal quotas should be eliminated, and a sharing ratio should be determined on some objective basis rather than in an ad-hoc way. If one province's retention rate is to be lower than another's, the difference should be based on objective indicators. A formula should be substituted for the present negotiated and judgmental approach. For example, the formula might compensate for lower fiscal capacity or greater fiscal need, or reward greater revenue mobilization (see next section). - 40 - Second, the tax-sharing ratios should remain fixed for a number of years to provide local governments with some certainty, and to discourage the sharing ratio from being a subject of annual negotiation. Third, the provincial and local governments should know the sharing ratio before the beginning of the fiscal year, to ensure better fiscal planning and the realization of any incentive effects. 123. Additionally, the grants system should be rationalized,i.e., it should be made more regular, should have a known distribution formula, and should be coordinated with the shared tax system (see below). one possibility would be to set up a "Grants Commission" along the lines of those in India and Australia, to study and recommend a fixed, five-year program of revenue sharing. In the interim, and before price reforms are completed, the grants system must incorporate some objective compensating mechanisms for the impact of price reforms on provinces revenue base. Finally, the enterprise tax should not be allowed to accrue to the level of government owning the enterprises. This interferes with merger, exploitation of scale economies and other interjurisdictional joint venture. 124. The central government must also rationalize its delegation of administrative powers to provincial governments through a better monitoring and tax information system. Before the central government can evaluate the costs and benefits of tax relief, it must understand the pattern of taxation that is being applied and the existing distribution of effective rates. At present, the central government has no information system that can produce this information. It would be fair to say that the central government does not have a good idea about the impacts of its overall system. 125. Box 7 summarizes some pros and cons of the various options developed and ranks them according to their effectiveness in reaching the selected objectives of stabilization, equity, and incentives for revenue mobilization. Some form of tax sharing or base sharing rates higher on most indicators than does reforming the present system, or simply changing the present retention rates. Other Essentials for Center-Local Fiscal Reform 126. A System of Transfers. Whatever the choice of a new intergovernmental fiscal system, it is not likely that the new revenue authority will match expenditure responsibility for every province. There will be a need for some kind of intergovernmental transfer system. In some ways this is a happy compromise between centralization and decentralization. A system of transfers permits central governments to retain control over the taxes on the more pro- ductive bases but it guarantees state or local governments a flow of revenues and can have a significant impact on interprovincial equity. On the other hand, transfers can make local governments less accountable for their fiscal decisions (they can increase spending without increasing taxes); hence, there will be less incentive to improve local government operations or be innovative in service delivery. In the case of China, special attention also needs to be paid to the impact of price reforms on the revenue base of the different provinces and how the transfer system can help offset the impact the price changes on fiscal capacity. 127. With respect to allocative effects, a transfer system should be structured so as to encourage efficient management and fiscal planning by local governments. From this perspective, shared taxes with or without a "formula" distribution offer the best possibilities for designing a transfer system which enhances local fiscal planning. If the transfer system is intended to influence budget priorities, other types of transfers may be better. Partial cost reimbursement grants which lower the relative price of one government service relative to others offer the best possibility for stimulating spending (teacher salary grants and public works grants are - 41 - Box 7: WEIGHING THE SIX OPTIONS IN CHINA The table below summarizes how these approaches rate against t.e criteria that are generally used to evaluate an intergovernmental fiscal system. Box Table 1: ECONOMIC EFFECTS OF ALTERNATIVE APPROACHES Alto- Reform Revenue Macro- cative com- mobili- stabili- effi- pati- Equity zation zation ciency bility (i) Tax reassignment - "Al N /a L H L/N H (fi) Tax reassignment - 'B" L N/L L/h. 4 H N/L (iii) Local surcharge on central taxes L/N H N/L/H N/H N (iv) Increased local retention rate L H/N L H N/L (v) Decreased local retention rate H La L/N H L N/H (vi) Reforming present system L L L L L /a Equity impact depends on the accompanying transfer system. Definitions: H = Positive economic effects or improvement with respect to present system. N = Neutral L = Negative economic effects Reform option (i), assigning certain taxes to localities, has many attractions: St has a neutral impact on the center's ability to conduct stabilization policy since local governments will be assigned only minor taxes. Compared to the present system, the central government's fiscaL management position would be stronger. It is Likely to encourage local government tax effort since localities will retain aLl their revenues, but, the local adninistration would now have less incentive to collect central taxes, and would likely have to be replaced by a central government administration. Since poorer local goverrnents will have a smaller tax base, the local tax option is likely to be counterequalizing. Howe- ver, a transfer program could offset this. It is highly compatible with price and enterprise reforms, assuming an appropriate transfer system is in place. The decentralized version of the reassignment of taxes, option (ii), could be positive or negative from the point of view of macroeconomic management, depending on which taxes are given to the local governments. In China, it is unlikely that the provinces will be given access to a major revenue source, hence the macroeconomic management function may not be harmed. On the other hand, if important revenue sources are not passed down, it is difficult to see why provinces should improve tax effort. Similarly, the increased self-reliance of provinces suggests low equalizing potential, unless the central government simultaneously introduces increased transfers. This option ranks lower than Ci) on compatibility with reforms. Option (iii) is neutral to low on macroeconomic grounds depending on which tax bases are surtaxed and how the enterprise income tax is administered. It must even improve things to the extent its introduction is coincident with a sharing out of local tax concessions including contracting. From an equalizing point of view, however, its properties are poor (in the absence of transfers), in that the tax base in poorer provinces is likely to be smaller than in better-off provinces. From a revenue point of view, the credentials of the surcharge option are strong, depending on the latitude of the local governments in rate determination. Its allocative properties are positive because local governments would have more flexibility to choose a level of taxes and a level of expenditures. It is not highly compatible with reforms, unless contracting can also be eliminated Option (iv) of an increased local tax share has appeal from the perspective of its impact on revenue mobilization. It rates low on equity. On the stabilization side, it rates low, inasmuch as provinces obtain a (guaranteed) larger share of major taxes. From an allocative perspective, the more local control, the closer will the local budget match local preferences, hence it rates high. It also rates Low in terms of compatibility with reforms. Option (v), reduced local retention, would have the opposite effects. Lastly, there is the reform of the present system; its disadvantages from the perspective of revenue mobilization and tax effort have been described in some detail. We have also noted that its properties, from the perspective of the central government, falls short of what is needed for sta- bilization and reduces scope for discretionary central expenditure policy and equalization. - 42 - examples of such programs). Conditional, earmarked formula grants are another approach. Their effectiveness depends on the income elasticity of demand for the expenditure in question, whether the expenditure would have been made in any case, and whether local revenues are "fungible."26/ 128. The impact of t-ansfers on tax effort, and the possibility that grants may affect local revenae mobilization, needs attention in transfer design. Ideally, one would make the transfers stimulative, so that the transfer has the net effect of increasing total local government expenditures. One way to do this is to build tax effort directly into the allocation formula, so that governments which exert a relatively greater tax effort, receive a larger grant. This has been tried in a number of countries, but without great success. 129. An important feature of grants is the extent to which they equalize fiscal capacities and public service levels. Typically the objective is to equalize the expenditure capacity of local governments to finance needed services. Usually, full expenditure capacity equalization is not realistic, since disparities in fiscal capacity are simply too large to be fully offset. So, most grant systems try and allocate relatively more funds to those areas that are thought to have the least capacity to finance services, or to those that have the greatest expenditure needs. With respect to the former objective, allocation is usually by formula and relies on some measure of financial capacity. In this respect, per capita income figures are an important ingredient in such revenue-sharing formulae. Expenditure "needs" also can guide the distribution of grants, but need is a subjective concept and most governments have chosen objective proxy measures rather thansophisticated indicators. Population is often used as a crude indicator of needs (i.e., equal per capita allocations of central government assistance are thought to address variations in needs). Land area, to reflect the costs of covering a more dispersed population, and equal shares, to reflect the fixed costs of government, are common measures. Some countries try to allocate special shares to provinces with especially heavier concentrations of the poor. It is not clear, however, that any of these approaches has led to more equalization in the distribution of resources across provinces in developing countries. In the last analysis, it has not been possible in most countries to find satisfactory indicators of capacity and need for inclusion in the formula. 130. An important purpose of the grant system is to ensure adequacy of revenues for local gov3rnments. "Adequacy" is defined in light of two considerations: (i) that transfers be large enough to redress the imbalance between the revenue bases and expenditure responsibility assigned to local governments, and (ii) that transfer revenues should grow at least in proportion to the growth in local population and prices, i.e., that transfers allow local governments to hold real per capita expenditures constant. This, in turn, depends on: (a) how the growth in the pool of funds available for transfers is determined; (b) how the distribution among local governments is made; and (c) whether the central government actually makes the full monetary distributions called for by the tranpfer system. Box 8 outlines some international experiences with transfers and formula grants. 131. Finally, until price reforms are completed, the transfer system will need to address the impact of price reforms on the revenue base of each province. In principle, there should be positive and negative offsets, with some provinces (e.g., intermediates-using, processing provinces) ceribus paribus receiving higher transfers than other (e.g., intermediates-producing) provinces. Empirically, this will require estimates of the revenue base 26/ See Bahl and Linn, Chapter 13, forthcoming; and Shah, A., "Fiscal Federalism in International Perspective," World Bank Working Paper. - 43 - Box 8: FORMULA GRANTS Four basic approaches to atlocating central revenue resources among tocaL governments are: (i) tax sharing, (ii) ad-hoc distributions, (iii) reimbursement of the costs of specified undertakings, and (iv) b9 formuta. Formula grants are popular because they are objective and easily understood, and because they give the centrat government the opportunity to target the distribution of funds among local governments in ways consistent with national policy. An added advantage is that as local governments outgrow their needs for revenue sharing, or as their needs change relative to other local governments, the grant dis- tribution is automatically adjusted by the formula and no discretionary government action is catted for. Formuta grants are usually pointed towards either a recognition of variations in expenditure needs or of differences in fiscal capacity, i.e., in the ability to raise revenue while making an average effort. Expenditure needs are proxied in many ways including per capita income, population, infrastructure adequacy, and the like. The distribution of education grants in Colombia is based pri- marlty on the population size of each province. Philippine general-purpose grants are distributed according to population and land area. Transportation grants to Brazilian state and local governments are allocated according to population, land area, and the consumption of imported fuels. The distri- bution of India's excise tax grant has made use of a poverty index. Grants to coipensate some provinces for low-income and low-fiscal capacity are often distributed partially by the reciprocal of per capita income. This is done in Brazil and India. Yet other formula grant systems have explicitly included a tax effort measure to induce provinces to increase their rate cf revenue mobilization (examples are the United States, Nigeria, India). The disadvantages to the formula grant are that the choice ot the grant elements can be influ- enced by politics. Moreover, the choice of a formula may be limited to what data are available, and as such may reflect neither true expenditure need for fiscal capacity. Moreover, detailed and timely data are almost never available at the local government leveL. Source: Roy Bahl and Johbnnes Linn; Urban Public Finance in Developing Countries; forthcoming. exante, and the impact of the price change on enterprise profitability and retail sales. 132. Borrowing Powers. Reform of China's center-local system should also address the need to improve financing sources for capital outlays, which involve investment in long-lasting infrastructure, e.g., public utilities and road infrastructure. Local governments in China make no use of borrowing to finance capital projects. In most countries, local governments borrow to finance infrastructure development. This is done in a variety of, albeit restricted, ways--bond finance, from a central government loan fund, from a development bank capitalized by the central government, etc. Usually, the central government sets the terms of the loan, defines the acceptable uses of debt finance and controls the flow of loan funds. In the United States, local governments have greater autonomy in deciding how much to borrow, under what terms and from whom. 133. There is potential for local governments to carry debt in China, at least from the point of view of adequacy of repayment potential for long- lived projects. However, borrowing powers need to be accompanied by a substantially reformed system of local financing--taxes that were more responsive to income growth--and price reform to ensure benefit charges are able to cover costs (see below). 134. Benefit Charges. The other financing possibility is self-financing, i.e., services that are to be finan:ed in part by beneficiary charges. There are many forms of benefit charge, e.g., road and bridge tolls, pollution charges, full cost recovery from public utility users. Indeed, while there is no formal benefit charge program in China, there is evidence of many - 44 - innovative self-financing schemes. Many local governments have taken advantage of particular opportunities to finance capital projects with charges to beneficiaries, but these seem to have been developed on a case-by-case basis and there is no common practice. The time may be right for China to make increasing use of benefit charges. Infrastructure needs are acute and enterprises are willing to pay for capital improvements. What is needed now is central government guidance in establishing such programs and encouraging their use. 135. Tax administration. The present system of tax administration will not support the objectives of China's economic reform. China's tax structure has very recently been modernized (a profits tax and a value-added tax have been introduced since 1984) but its tax administration has not kept pace. Tax administration will become more difficult with the growing number of small private firms and collectives--which are the hard-to-tax sectors in China.27/ 136. Without a survey of administrative problems, it is difficult to be specific about the elements of a reform program. There are, however, four general areas for improvement. The first is personnel policies and training programs and the development of an adequate number of qualified tax administrators. The second relates to the need to adapt the administrative system to accommodate the changing economic system. The growing private and collective sector implies a greater need to identify new tax-paying enterprises and track their activities, hence the need for a tax-paying numbering system, more information on transactions, more government attention to assessment and audit, and computerization. Third is a reexamination of the tax system to see whether its complications block effective administration. It may be that the first step toward better administration is a simplification of the tax structure. In particular, it is difficult to imagine that the present income tax system with its contracting feature could ever be efficiently administered. 137. The fourth area is the most difficult to address: should the government create for itself a separat.e central tax administration similar to that which exists in most countries? The arguments in favor of this are strong. Under the present system, there are inadequate incentives for aggressive assessment and collection efforts. A centralized system would eliminate this problem by taking local governments out of the business of collecting central government taxes. Another advantage is that procedures could be standardized across the country and the processes of manual preparation, monitoring, and gathering and reporting statistics would all realize economies of scale. The central government is more able to bring specialized technical assistance (including the development of a computerized tax information system) to the whole system of tax assessment and collection, and is in the best position to modernize the tax administration to keep it in step with the modernization of the tax structure. It is important that China's new tax structure be implemented so that it achieves the intended economic impacts, and that it be implemented in a uniform way across the country. Finally, there are advantages that relate to the staffing of an efficient tax administration service. Among these are central organization of the training programs, the ease of transferability of personnel within a cen- tralized system, and the greater possibility of promotion and advancement within a central revenue service. 138. There are also disadvantages. If coupled with increased local fiscal autonomy, there would have to be some local tax administration for local taxes. Separate central and local systems would involve duplication of effort 27/ For a discussion of tax administration in China, see Revenue Mobi- lization and Tax Policy, Red Cover Report, 1990, World Bank. - 45 - and inevitably a weaker, "second-class" local administration. Another problem is that a central revenue system in China would be an enormous bureaucracy, and inevitably would have to decentralize to accommodate the great diversity in tax administration needs within the country. The biggest disadvantage to centralization is that intimate familiarity with the local economy and its tax-paying base can be lost. For example, most central sales and income tax systems in low income countries do not have a very broad coverage of firms, whereas locally administered systems seem more able to identify and assess smaller firms. 139. Perhaps a better course would be to reform the tax administration to capture the best features of centralization and decentralization. A few principles that might be considered are: (a) general procedures for taxpayer identification, recordkeeping, and assessment should be centralized, and all related manuals should be centrally prepared and updated; (b) a major staff training program should be centrally designed and implemented; (c) a "statistics of taxation" series should be organized centrally and regularly produced to help in monitoring the performance of the tax system and the administrative efforts of each decentralized local unit; (d) assessment, collection and audit responsibilities should remain at the local level, but procedures should be established by the Central Government and regulated through the provincial level; (e) a computerized tax information 3ystem should be developed, and (f) tax sharing is a good scheme to stimulate local tax effort but the local shares should be the same for all taxes so as not to encourage different levels of administrative effort for different taxes. Conclusions 140. This paper argues that a reformed revenue-sharing system must meet the center's needs for stabilization and the provinces' needs for revenue and equalization of expenditure capacity. It has also argued that equalization should be based on objectivE indicators of need and that a formula-based grant system best meets this objective. A reformed system must also underpin price and enterprise reforms, and not require major recalibration or adjustments, because such reforms are taking place. These three elements ar3 the foundations of an improved system of local finance. The central-local reforms would be supplemented by local borrowing powers, a system of benefit charges, and improved financial planning and tax administration. ANNEX I Page 46 CHINA CENTER-LOCAL FISCAL RELATIONS Procvclical Effects of Contracting with Provinces for Revenue Co-wction 1. The provincial contracting arrangement can be formalized to illus- trate the type of problems discussed in the text and the factors affecting revenue elasticity. The tax remittance/transfer to the center by a province in year i (PT,, is PT; ab(C) + am (C-C;) ab - provinces' quota contracted sharing rate am = marginal, above-quota sharing rate (which may be zero) (amO, expenditure capacity grows at a rate related to (Q-g). For a given collection rate, in those provinces where g=O, but a transfer from the central government is received, expenditure capacity does not grow relative to GNP inasmuch as the growth rate in the lump-sum transfer is zero. Table 1: TAX COLLECTION AND EXPENDITURE Wy Level of Goverrmient, 1980-90 1930 1?81 1982 1983 1984 1985 1986 1987 1988 1989 1990B Total revenues 108.5 108.9 112.4 124.9 150.1 186.6 231.0 236.0 262.8 285.9 Central goverrmnt 21.0 22.5 25.8 37.2 52.4 69.0 91.7 83.3 94.5 107.2 of which: borrowing -- 7.3 8.4 8.0 7.7 10.0 13.9 16.9 23.1 22.0 Local goverment 87.5 66.5 86.6 87.7 97.7 117.6 134.3 152.7 168.3 178.7 of which: own 11.8 11.6 12.5 15.3 15.9 19.0 -- -- -- -- Total expenditure 121.2 111.5 115.3 128.4 153.6 186.5 233.0 242.5 263.5 0.0 Central 65.0 60.2 57.5 64.2 72.8 81.9 96.2 95.8 94.6 -- Local 56.2 51.3 57.8 65.8 80.8 104.6 136.8 146.7 168.9 -- Central collections deficit -44.0 -37.7 -31.7 -27.0 -20.4 -12.9 -4.5 -12.5 -0.1 107.2 Funded by: Local collections surplus: 31.3 35.2 28.8 21.9 16.9 13.0 -2.5 6.0 -0.6 178.7 Central borrowing -- 7.3 8.4 8.0 7.7 10.0 13.9 16.9 23.1 22.0 Total revenues (adjusted) 108.5 101.7 104.0 116.9 142.4 176.6 212.1 219.1 239.7 263.9 Central 21.0 15.2 17.4 29.2 44.7 59.0 77.8 66.4 71.4 85.2 Local 87.5 86.5 86.6 87.7 97.7 117.6 134.3 152.7 168.3 178.7 (in percent) Total expenditure 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Central 53.63% 53.99X 49.87% 50.002 47.402 43.912 41.292 39.512 35.90% ERR Loce; 46.37X 46.01X 50.13X 51.252 52.602 56.09X 58.71X 60.492 64.10X ERR Total revenues 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Central 19.35X 20.66X 22.95f 29.78X 34.91X 36.98X 39.70X 35.30X 35.96X 37.50X Local 80.652 79.432 77.05% 70.222 65.092 63.022 58.142 64.702 64.042 62.502 Total revenues (adjusted) Central 19.352 14.952 16.73% 24.982 31.39X 33.412 36.682 30.312 29.792 32.232 Local 80.652 85.052 83.27X 75.02X 68.61% 66.592 63.322 69.692 70.212 67.722 Reverues/GNP 24.27% 22.812 21.672 21.59X 21.682 21.852 23.772 20.782 18.962 18.17X Central 4.702 4.712 4.97X 6.432 7.57X 8.08% 9.43X 7.33X 6.82X 6.812 Local 19.572 18.122 16.702 15.162 14.112 13.77% 13.822 13.452 12.142 11.362 Central adjusted revenues 4.70X 3.18X 3.362 5.052 6.462 6.912 8.002 5.852 5.15X 5.422 Expenditures/GNP 27.112 23.352 22.232 22.202 22.18X 21.842 23.972 21.35% 19.01% 0.00% Central 14.542 12.61211.092 11.102 10.512 9.59% 9.90X 8.44Z 6.83% 0.00O Lcal 12.5" 10.742 11.152 11.382 11.67 12.25X 14.07X 12.92X 12.192 0.002 Memo Items: Real revenues 108.5 107.0 110.1 120.4 137.2 156.4 183.8 177.6 180.2 179.5 Central 2'.0 22.1 25.3 35.9 47.9 57.8 73.0 62.7 64.8 67.3 Local 87.5 85.0 84.8 84.6 89.3 98.6 106.8 114.9 115.4 112.2 Real revenue growth Central 5.252 14.332 41.962 33.522 20.752 26.132 -14.08X 3.41X 3.832 Err Local -2.892 -0.18 -0.29X 5.602 10.382 8.39X 7.542 0.462 -2.822 Err GNP 447.1 477.5 518.6 578.4 692.4 854.0 972.0 1135.7 1385.8 1573.1 Deftator 100.0 101.8 102.1 103.7 109.4 119.3 125.7 132.9 145.8 159.3 Table 2a: SHARING ARRANGEMENTS IN 1987 BY TYPE OF TAX (in X) "Fixed "Fixed Tax category "Shared" Central" Local" Industrial-Co*..ercial Taxes 1. Product and VAT- a. Enterprioes owned by four Ministries 30 70 0 b. Tobacco products produced by centrally owned enterprises 0 100 0 c. Other general taxes 100 0 0 d. Product tax and VAT on imported goods 0 100 0 e. Refund of product tax and VAT to central goverrment foreign trade company 0 100 0 f. Refund of product tax and VAT on export goods, to industrial enterprises and local foreign trade enterprises 100 0 0 2. Business Tax a. Enterprises owned by four Ministries 30 70 0 b. Railway, central tobacco enterprises, bank headquarters 0 100 0 c. Genera I business tax 100 0 0 d. Self_employed urban and rural households in industry and commerce 100 0 0 3. Consolidated Industrv and Cofferce Tax a. Offshore OIL enterprises 0 100 0 b. Other enrprises 100 0 0 c. Imported products 0 100 0 Other Taxes '.SpTeial Adjustment Tax 0 100 0 5. Collective Enterprises Income Tax 100 0 0 6. Self-Employed Households in Industry and Commerce 100 0 0 7. Individual Income Tax 100 0 0 8. Individual Income Adjusted Tax 100 0 0 9. Joint Venture Income Tax a. Offshore oiL 0 100 0 b. All other 100 0 0 10. Fore%in Ente ristr Inconme Tax aOfsore oL 0 100 0 b. All other 100 0 0 11. Urban Construction and Maintenance Tax 0 0 100 12. Vehicle Utilization Tax 100 0 0 13. Local Vehicle Utilization Tax 0 0 100 14. House Tax 100 0 0 15. Slaughter Tax 100 0 0 16. Animal Trading Tax (livestock transactions) 100 0 0 17. Free Market Transaction Tax 100 0 0 18. Natural Resource Tax 100 0 0 19. Central Resource Tax 0 100 0 20. SOE Bonus Tax 100 0 0 21 SOE Wage Adjustment Tax 100 0 0 22. Institutions Bonus Tax 100 0 0 23. Collectives Bonus Tax 0 0 100 24. Construction Tax 100 0 0 25 Special Fuel Using Tax (crude oil burning tax) 0 100 0 26. Deduction and Refund of Fuel Tax 100 0 0 27. Salt Tax 100 0 0 28. Revenue from Penalties and Fines a. Customs duty categories 0 100 0 29. Agricutture Taxes a. Animal husbandry 100 0 0 b. Forestry and special products 100 0 0 c. Central land occupation tax 0 100 0 d. Local land occupation tax 0 0 100 30. Income Tax a. SOEs income tax 100 0 0 b. SOEs adjustment tax 100 0 0 c. SOEs profit remittance 100 0 0 d. Subsidies for planned losses 100 0 0 31. Contribution for Enery Transportation Projects a. Paid by central SOEs 0 100 0 b. Paid by local SoEs 0 100 0 32. Interest Income n.a. n.a. n.a. 33. Earmarked Revenue n.a. n.a. n.s. 34. Revenue from Loan Repayment for Capital Construction n.a. n.a. n.a. 35. Revenue from Other Sources a. Other revenue from joTnt ventures 100 0 0 Table 2b: SHARING ARRANGEMENTS IN 1988 BY TYPE OF TAX (in X) "Fixed "Fixed Tax category "Shared" Central" Local" Industria t-Coa iercial Taxes 1. Produt and VAT a. Entter-p-rises owrned by four Ministries 30 70 0 b. Tobacco products produced by centrally owned enterprises 0 100 0 c. Other general taxes 100 0 0 d. Product tax and VAT on imported goods 0 100 0 e. Refund of product tax and VAT to central govermient foreign trade company 0 100 0 f. Refund of product tax and VAT on export goods, to industrial enterprises end local foreign trade enterprises 100 0 0 2. Business Tax a. EnterprFses owned by four Ministries 30 70 0 b. Rail: a, central tobacco enterprises, bank headquarters 10 0 0 c. General business tax 100 0 0 d. Self-employed urban and rural households in industry and commerce 100 0 0 3. Consolidated Industry and Commerce Tax a. Offshore oIl enterprises 0 100 0 b. Other enterprises 100 0 0 c. Imported products 0 100 0 Other Taxes 4. Special Adjustment Tax 0 100 0 5. Collective Enterprises Income Tax 100 0 0 6. Self-Employed Households in Industry and Commerce 100 0 0 7. Individual Income Tax 100 0 0 8. Individual Income Adjusted Tax 100 0 0 9. Joint Venture Income Tax a. Offshore oil 0 100 0 b. ALl other 100 0 0 10. Foreion Enterprise Income Tax aD. Offshore oil 0 100 0 b. All other 100 0 0 11. Urban Construction and Maintenance Tax 0 0 100 12. Vehicle Utilization Tax 100 0 0 13. Local Vehicle Utilization Tax 0 0 100 14. House Tax 100 0 0 15. Slaughter Tax 100 0 0 16. Animal Trading Tax (livestock transactions) 100 0 0 17. Free Market Transaction Tax 100 0 0 18. Natural Resource Tax 100 0 0 19. Central Resource Tax 0 100 0 20. SOE Bonus Tax 100 0 0 21. SOE Wage Adjustment Tax 100 0 0 22. Institutions Bonus Tax 100 0 0 23. Collectives Bonus Tax 0 0 100 24. Construction Tax 100 0 0 25. Special FueL Using Tax (crude oil burning tax) 0 100 0 26. Deduction and Refund of Fuel Tax 100 0 0 27. Salt Tax 100 0 0 28. Revenue from Penalties and Fines a. Customs duty categories 0 100 0 29. Aariculture Taxes aAnima husbandry 100 0 0 b. Forestry and special products 100 0 0 c. Central land occupation tax 0 100 0 d. Local land occupation tax 0 0 100 30. Income Tax a. SOEs income tax 100 0 0 b. SOEs adjustment tax 100 0 0 c. SOEs profit remittance 100 0 0 d. Subsidies for planned losses 100 0 0 31. Contribution for Energy Transportation Projects a. Paid by central SOFF 0 100 0 b. Paid by local SOEs 0 100 0 32. Interest Income n.a. n.a. n.a. 33. Earmarked Revenue n.a. n.a. n.a. 34. Revenue from Loan Repayment for Capital Construction n.e. n.a. n.a. 35. Revenue from Other Sources a. Other revenue fraro joint ventures 100 0 0 Table 3: REVENUE-SHARING SYSTEM BETWEEN THE CENTRAL AND LOCAL GOVERNMENTS, 1985-87 La Fixed percentage Province retains Province retains own of total revenue all own revenue and receives revenue and pays fixed reained bv,:rovvnce fixed amount from the center amount to the center Provinces & regions 1985 1986 1987 1985 1986 1987 -------- t ........ ------------------- (Yuan mitlion) ----------------- North China Beijing 48.20 49.55 49.55 Tianjin 39.50, 39.45 39.45 Hebei 6;.00 72.00 72.00 Shanxi 97.50 97.50 97.50 Inner Mongolia 1,783 1,961 2,059 Northeast China Liaoning 51.10 52.66 52.66 Jilin 397 396 396 Heitlongjiang 96.00 142 142 East China Shanghai 26.00 23.54 23.54 Jiangsu 39.00 41.00 41.00 Zhejiang 55.00 55.00 60.81 Anhui 80.10 80.10 80.10 Fujian 235 234 234 Jiangxi 239 239 239 Shandong 59.00 77.47 75.00 Central/South China Henan 81.00 81.00 87.71 Hubei 66.50 100.00 100.00 Hiunan 88.00 88.00 88.00 Cuangdong 772 778 778 Guangxi 716 788 827 Southwest China Sichuan 89.00 100.00 100.00 Guizhou 743 817 858 Yunnan 637 925 972 Tibet 750 825 866 Northwest China Shaanxi 270 270 270 Gansu 246 245 245 Qinghai 611 671 705 Ningxia 494 543 57r Xinjiang 1,450 1.594 1,674 /a Subsidies were to increase by 10 percent per year after 1985. Source: Data supplied by Ministry of Finance. Table 4: REVENUE AND EXPENDITURE DISPARITIES ANOWG PROVINCES (1985) Budgetary Ranking: revenue cotlections Budoetarv exoenditures Percent Per Per capita Percent Per capita Percent of total capita mount of amount of popu- national (Rlb/person) total (Rmb/person) total lation income BeiJing 609 4.97 344 3.15 0.92 2 Lg Tianjin 597 4.10 334 2.58 0.78 3 L Hebei 81 3.84 75 3.98 5.33 14 Shanxi 95 2.12 135 3.40 2.52 12 Inner Mongolia 72 1.22 170 3.27 1.93 16 Liaoning 231 7.24 154 5.44 3.54 4 L Heilongjiang 98 2.76 135 4.27 3.18 5 Li Shanghai 1,492 15.43 346 4.02 1.17 1 La Jiangsu 135 7.14. 81 4.83 5.97 6 La Zhejiang 145 4.95 93 3.57 3.87 7 /a Anhui 58 2.56 66 3.24 4.95 21 Lb Fujian 92 2.13 113 2.93 2.61 17 Jiangxi 56 1.66 81 2.68 3.32 22 Lb Shandong 88 5.74 67 4.90 7.39 10 /a Henan 63 4.16 64 4.73 7.41 25 /b Hubei 102 4.27 88 4.17 4.74 11 Hunan 70 3.33 71 3.83 5.40 19 Guangdong 112 5.92 107 6.41 6.01 9 La Guangxi 52 1.71 77 2.84 3.72 28 Li Sichuan 57 4.99 63 6.13 9.79 26 /b Guizhou 49 1.23 80 2.28 2.85 29 lb Yunnan 80 2.33 108 3.51 3.27 27 /b Tibet 40 0.07 517 0.98 0.19 18 Shaanxi 68 1.72 92 2.63 2.88 24 Li Gansu 81 1.40 118 2.29 1.96 23 Lb Qinghai 60 0.20 248 0.96 0.39 15 Ningxia 70 0.25 237 0.94 0.40 20 /b Xinjiang 62 0.72 210 2.73 1.31 13 Total 100.00 /a Among 10 richest provinces. ,b Among 10 poorest provinces. Source: Data supplied by MOF. Table 5: BUDGETARY EXPENDITURE BY PROVINCE (in bitlion yuan) 1983 1986 Percent increase Per capita income ranking: 1985 BeIjing 19.61 44.27 125 2 L Tianjin 20.49 34.85 70 3 La Hebei 2-.27 53.82 90 14 Shanxi 24.01 41.17 71 12 inmer Mongolia 22.83 43.89 92 16 Liaoning 34.17 75.51 120 4 L Heitongjiang 30.71 61.47 100 5 a Jiangsu 32.29 66.16 104 6 Lh Shanghai 19.03 56.95 199 1 /a Zhejiang 21.94 50.95 132 7 a Anhui 20.38 46.18 126 21 /b Shtndong 32.41 67.94 109 10 Henwn 30.06 69.20 130 25 Lb Hubei 28.32 58.02 104 11 Hunan 25.31 54.29 114 19 Sichuan 36.64 87.74 139 26 Lb Jilin 19.41 50.12 158 8 a Jiangxi 17.27 36.63 112 22 /b Shaanxi 18.81 35.59 89 23 gb Gansu 15.53 30.01 93 15 FuJiat 17.55 37.62 114 17 Guangxi 18.84 42.22 124 28 Lb Yunnan 24.03 47.31 97 27 /b Tibet 5.88 8.97 61 18 Qinghai 7.39 12.22 65 15 Ningxia 6.95 12.02 73 20 Lb Xinjiang 18.61 35.12 89 13 Guangdong 37.65 89.55 138 9 La Guizhou 15.55 30.39 95 29 L Total 649.00 1.380.00 112 Source: Comuted from data provided by Ministry of Finance. Table 6: PERCENT INCREASE IN REVENUES AND EXPENDITURES, BY PROVINCE FOR 1983-86 Budgetarv Expenditures_ Budgetary Collections Ranking: Percent Retetive Percent Relative Per capita Province increase (average100) Rank Increase (average=100) Rank national income (1988) Beijing 125.75 111.92 8 61.05 127.30 18 2 La Tianjin 70.08 62.38 27 40.68 84.83 22 3 a Hebei 90.38 80.44 22 40.62 84.69 23 14 Shanxi 71.47 63.61 26 18.59 38.77 26 12 Inner Mongolia 92.25 82.10 21 129.18 269.38 2 16 Liaoning 120.98 107.68 10 12.49 26.05 28 4 L Heilongjiang 100.16 89.15 17 109.07 227.43 5 5 a Jiangsu 104.89 93.36 15 119.67 249.53 4 5 La Shanghai 199.26 177.35 1 15.86 33.06 27 1 /a Zhejiang 132.22 117.68 5 32.40 67.55 24 7 LA Anhui 126.59 112.67 7 64.18 133.83 15 21 Lb Shandong 109.63 97.57 14 58.33 121.63 19 ^0 Henan 130.21 115.89 16 135.57 202.70 1 25 /b Hubef 104.87 93.34 16 77.75 162.14 12 11 Hunan 114.50 101.91 11 21.15 44.10 25 19 Sichuan 139.47 124.13 3 50.71 105.75 20 26 Lb Jitin 158.22 140.82 2 42.38 88.38 21 8 /a Jiangxi 112.10 99.77 13 62.79 130.94 16 22 Lb Shaanxi 89.21 79.40 23 125.74 262.19 3 23 Lb Gansu 93.24 82.99 20 81.64 170.24 9 15 Fujian 114.36 101.78 12 62.74 130.82 17 17 Guangxi 124.10 110.45 9 88.88 105.33 a 28 Lb Yurnan 96.88 86.23 18 74.96 156.30 13 27 /b Tibet 52.55 46.77 29 -85.42 -178.12 29 18 Qinghai 65.36 58.17 28 65.68 136.96 14 15 Ningxia 72.95 64.93 25 81.28 169.50 10 20 /b Xinjiang 88.72 78.96 24 108.44 226.13 6 13 Guangdong 137.85 122.69 4 105.62 220.24 7 9 /a Guizhou 95.43 84.94 19 81.17 169.27 11 29 /b La Among 10 richest provinces. Lb Among 10 poorest provinces. Table 7: CHANGES IN THE RATIO OF EXPENDITURES TO COLLECTIONS: FOR 1983-86 BY PROVINCE Ranking: Ratio Change Per capita Province 1983 1986 1983-86 national income (1988) BeijIng 0.52 0.73 0.21 2 L TianJin 0.53 0.64 0.11 3 _Q Hebei 0.78 1.05 0.27 14 Shanxi 0.99 1.44 0.45 12 Inner Mongolia 3.27 2.74 -0.53 16 Liaoting 0.39 0.76 0.37 5 L Jilin 1.37 1.70 0.33 a L Neilonga]ang 1.42 1.29 -0.13 5 L! Shanghai 0.12 0.32 0.20 1 La Jiangsu 0.43 0.67 0.24 5 LA Zhejiang 0.53 0.74 0.21 7 La Anhui 0.91 1.30 0.39 21 Lb Fujian 1.42 1.29 -0.13 17 Jiangxi 1.28 1.52 0.24 22 Lb Shandong 0.63 1.09 0.46 10 Henan 0.82 1.26 0.44 25 L_ Hubei 0.70 1.01 0.31 11 Hunan 0.86 1.14 0.28 19 Guangdong 1.04 1.09 0.05 9/a Guangxi 1.36 1.67 0.31 28 Lb Sichuan 0.89 1.30 0.41 26 Lb Guizhou 1.78 1.85 0.07 29 Lb Yunnan 1.40 1.58 0.18 27 Lb Tibet 12.25 128.14 115.89 18 Shaanxi 1.29 1.48 0.19 23 Lb Gansu 1.42 1.52 0.10 15 4inghaf 4.80 3.81 -0.99 15 Ningxia 3.90 3.28 -0.62 20 Lb Xinjiang 3.31 3.44 0.13 13 Table 8: OLS REGRESSION RESULTS FOR CHINESE REVENUES AND EXPENDITURES AGAINST SELECTED INDEPENDENT VARIABLES: BY PROVINCE FOR 1985 REGRESSION COEFFICIENTS LA Per capita Percent of popu- Equa- Logarithms (L) output lation living _ tion Dependent variable or Linear (N) Constant (100 RMB) in urban areas Population 12 N 1. Per Capita Revenue L -5.049 1.357 0.333 ... 0.91 29 (8.53) (0.114) (0.129) 2. Per Capita Revenue L -5.608 1.391 0.314 0.053 0.91 29 (7.06) (11.790) (2.241) (1.052) 3. Per Capita Expenditures L 5.433 0.436 0.106 -0.443 0.93 29 (11.38) (6.154) (1.362) (-14.783) 4. Ratio of Expenditures L 6.739 -0.558 -0.168 -0.308 0.82 29 to Revenues Raised (11.322) (6.315) (1.728) (8.242) 5. Ratio of Revenues to N 0.059 1.514 (E-05) 0.0002 ... 0.63 29 Total Output Value (6.454) (5.168) (0.600) / T-statistic shown in parentheses below the regression coefficient. CHINA CENTRAL-LOCAL FISCAL RELATIONS Bibliographv Bahl, Roy and Jun Zhang, mimeo, World Bank, 1988, "Land Use Taxes in China." Bahl, Roy and Johannes Linn, Urban Public Finance in DeveloxDing Countries, Oxford Univ. Press, 1992. Bahl, Roy, "A Regression Approach to Tax Effort and Tax Ratio Analysis," International Monetary Fund Staff Papers, Vol. 18, No. 3, November 1971. Bahl, Roy, "The Design of Intergovernmental Transfers in Industrialized Countries," Public Budaeting and Finance, Winter 1986, Volume 6, Number 4, pp. 3-22. Bird, Richard and Christine Wallich, "Local Government Finance in Hungary" World Bank Working Paper, 1991, forthcoming. Blejer, Mario, and George Szapary, "The Evolving Role of Fiscal Policy in Centrally Planned Economies: The Case of China," 1989, IMF, Washington, D.C. Prud'homme, Remy, "Urban Finances in Shanghai," IBRD mimeo, 1989. Shah, A., "Fiscal Federalism in International Perspective," World Bank Working Paper Series No. 557. State Statistical Bureau, China: Statistical Yearbook 1990, Beijing, China. Tait, Alan, VAT: International ExDeriences and Problems, 1988, IMF, Washington, D.C. Tidrick and Chen, China's Industrial Reforms, Oxford University Press, London, 1988. World Bank, Finance and Investment, Red Cover Report, Washington, D.C., 1988. World Bank, Revenue Mobilization and Tax Policy, Red Cover Report, Washington, D.C., 1989. World Bank, China: Between Plan and Market, Red Cover Report, Washington, D.C., 1990. Policy Research Working Paper Series Contact Title Author Date for paper WPS 842 Capital Flows to South Asia and Ishrat Husain January 1992 S. King-Watson ASEAN Countries: Trends, Kwang W. Jun 31047 Determinants, and Policy Implications WPS843 How Financial Markets Affect Long- Ejaz Ghani January 1992 A. Nokhostin Run Growth: A Cross-Country Study 34150 WPS844 Heterogeneity, Distribution, and Ravi Kanbur January 1992 WDR Office Cooperation in Common Property 31393 Resource Management WPS845 Inflation Stabilization in Turkey: Luc Everaert .J?.1uary i992 B. Mondestin An Application of the RMSM-X Model 36071 WPS846 Incorporating Cost and Cost- Larry Forgy January 1992 0. Nadora Effectiveness Analysis into the Diana M. Measharn 31091 Development of Safe Motherhood ^-:.ne G. Tinker Programs WPS847 Coping with the Legacies of Silvia B. Sagari January 1992 M. Guirbo Subsidized Mortgage Credit in Loic Chiquier 35015 Hungary WPS848 How EC 1992 and Reforms of the Merlinda D. lngco February 1992 P. Kokila Common Agricultural Policy Would Donald 0. Mitchell 33716 Affect Developing Countries' Grain Trade WPS849 Financial Structures and Economic Ross Levine February 1992 W. Pitayatona'zarn Development 37666 WPS850 Fiscal Adjustment and the Real Kazi M. Matin February 1992 D. Ballantyne Exchange Rafe: The Case of Bangladesh 38004 WPS851 Sources of Wor!d Bank Estimates Eduard Bos February 1992 0. Nadora of Current Mortality Rates My T. Vu 31091 Patience W. Stephens WPS852 How Health Insurance Affects the Joseph Kutzin February 1992 0. Nadora Delivery of Health Care in Developing Howard Barnum 31091 Countries WPS853 Policy Uncertainty, Information Gerard Caprio February 1992 W. Pitayatonakarn Asymmetries, and Financial 37664 Intermediation WPS854 Is 1'here a Case for an Optimal Export Takamasa Akiyama February 1992 G. Ilogon -fax on Perennial Crops? 33732 WPS855 Sovereign Debt: A Primer Jonathan Eaton February 1992 S. King-Watson 31047 Policy Research Workirng Paper Series Contact Title Author Date for paper WPS856 Latin American Women's Earnings George Psacharopoulos February 1992 L. Longo and Participation in the Labor Force Zafiris Tzannatos 39244 WPS857 The Life Insurance Industry in the Kenneth M. Wright February 1992 W. Pitayatonakarn United States: An Analysis of Economic 37664 and Regulatory Issues WPS858 Contractual Savings and Emerging Dimitri Vittas February 1992 W. Pitayatonakarn Securities Markets 37664 WPS859 Macroeconomic Management and Ibrahim A. Elbadawi February 1992 V. Barthelmes the Black Market for Foreign Exchange 39175 in Sudan WPS860 The Restrictiveness of the Multi- Refik Erzan February 1992 G. llogon Fibre Arrangement on Eastern Christopher Holmes 33732 European Trade WPS861 Private Saving in Mexico, 1980-90 Patricio Arrau Fobruary 1992 S. King-Watson Daniel Oks 31047 WPS862 Higher Education in Egypt Alan Richards WPS863 Intergovernmental Fiscal Relations Roy Bahl February 1992 A. Bhalla in China Christine Wallich 37699