VI Fiscal Federalism

Eswar Prasad
Published Date:
June 2004
  • ShareShare
Show Summary Details
Annalisa Fedelino and Raju Jan Singh 

The increased decentralization of intergovernmental fiscal relations since the early 1980s has played an important role in supporting the process of transition to a market economy and promoting growth in China. Some scholars have suggested that the devolution of authority from the center to local governments has contributed to China’s spectacular growth performance over the last two decades (Box 6.1).

At the same time, however, China’s impressive growth performance has not benefited all provinces equally.1 Large income disparities remain across provinces; for example, in 2002 the highest provincial per capita income level was more than ten times greater than the lowest (Table 6.1). Among the top ten richest provinces, nine were in the east, while the ten poorest were in the central and western regions. Furthermore, the provision of public services is skewed in favor of richer provinces: for example, annual per capita healthcare expenditure varies from a maximum of about RMB 200 in Beijing and Shanghai to less than RMB 20 in the central provinces of Henan and Hunan. The role of center-local relations in perpetuating these disparities has come under increasing scrutiny.

Table 6.1.Per Capita Income Levels of Selected Provinces, 2002
In RenminbiIn U.S. Dollars1As a Ratio to Maximum
Top two provinces
Bottom two provinces
Memorandum item:
National average9,2551,11527.8
Sources: China Finance Yearbook, 2003; and IMF staff calculations.

At the exchange rate of US$1 = RMB 8.3.

Sources: China Finance Yearbook, 2003; and IMF staff calculations.

At the exchange rate of US$1 = RMB 8.3.

Local governments are largely responsible for public service delivery and implementation of social policies. However, a widening gap is emerging between local governments’ expenditure mandates (accounting for about 70 percent of total budgetary expenditure), which are assigned by the center, and resources immediately available to local governments (about 45 percent of total budgetary revenue, before transfers from the center). This imbalance implies that local governments, which are not allowed to borrow directly and whose transfers from the center are inadequate, experience difficulties in financing basic services, and more so in poorer provinces where revenue is lower (Ahmad and others, 2004). This largely reflects the design as well as the implementation of the system of intergovernmental fiscal relations introduced in 1994. This system has not kept pace with the challenges posed by the process of transition to a market economy and growing regional disparities.

Against this background, this section looks at the main elements of the 1994 reform, focusing on revenue assignments, expenditure mandates, and the transfer system, mainly at the provincial level. The main conclusion is that, ten years into the reform, there is a pressing need to reexamine some aspects of intergovernmental fiscal relations, in particular the design and clarity of expenditure mandates and the transparency and efficiency of the transfer system.

The 1994 Fiscal Reform

Recentralization of Revenue

The core element of the 1994 reform was to ensure higher revenue as a ratio to GDP while also boosting the share of the central government in total revenue—both had declined dramatically since the early 1980s, mainly due to weak tax administration capacity and lax control over local governments’ collections (Figure 6.1) (Wong, 1997; and Ahmad, 1997). A new tax sharing system (TSS) was introduced, which shifted revenue collection and distribution away from a negotiated basis to a mix of tax assignments and tax sharing. A new value-added tax (VAT) was adopted, to be shared between the center and the local governments. Revenue assigned to local governments included 25 percent of the new VAT, the business tax (a turnover tax), the personal income tax, and the enterprise income tax (EIT) levied on local state-owned enterprises (SOEs) and foreign-financed enterprises. The central government received 75 percent of the VAT, consumption taxes (excises) and trade-related taxes, and EIT from central SOEs (see also OECD, 2002a; and Ahmad and others, 2002 and 2004). The State Tax Administration was made responsible for the collection of central and shared taxes, while local government agencies were in charge of collecting local taxes. However, tax authority remained solely vested with the central government, as local governments had only limited powers to set rates for a few local taxes. These arrangements are still broadly applied, as no major changes have been introduced since 1994.2

Figure 6.1.Revenue Developments1

Sources: Ministry of Finance; China Finance Yearbook; and IMF staff estimates.

1 For comparability with older data, the official definition of revenue is used.

Box 6.1.Intergovernmental Fiscal Relations and Market Reforms in China: A Review of the Literature

The market-preserving fiscal federalism school is based on the idea that fiscal decentralization has provided local governments with significant incentives to promote growth (see, among others, Qian and Weingast, 1997; and Blanchard and Shleifer, 2001). More recently, Jin, Qian, and Weingast (2003a and 2003b) have tested the relationship between provincial governments’ fiscal incentives and regional development over the period 1970–92. There are two results relevant to the analysis in this section. First, fiscal incentives (defined as the share of retained revenue) faced by provincial governments have grown increasingly stronger as a result of the decentralization started in the 1980s: the correlation between ex post (realized) provincial budgetary revenue collection and expenditure after decentralization reforms is about four times as high as before reforms. Second, higher ex ante fiscal incentives (measured by the contractual revenue retention rate of provincial governments) are found to have a positive effect on provincial growth performance. In other words, while fiscal decentralization per se would not lead to stronger growth, stronger fiscal incentives would do so. However, further empirical work would be needed to test the impact of the latest round of intergovernmental fiscal reforms implemented in 1994.

Contrasting these views is the market-hampering school, claiming that competition among local governments has resulted in antimarket behavior (see, for example, Findlay, Wu, and Watson, 1995; and Young, 2000a). Under the centrally planned system, prices were skewed to concentrate value added, and hence profits, in industry. The gradual pace of the reform process, with few segments of the economy freed of control at any time, has created incentives to seek opportunities in the remaining distortions, thereby generating new distortions, mainly in the form of interregional barriers to trade. On this basis, this strand of literature concludes that the transition to a market economy has resulted in the fragmentation of the domestic market, possibly hampering the reform process down the road and preventing China from reaping the benefits of a large unified internal market.

From the perspective of raising revenue and increasing the share of revenue accruing to the center, the 1994 reform has been successful: revenues have quickly recovered from the trough in the mid-1990s, and the center’s share has surged to above 55 percent—more than twice the level registered just before the reform (see Figure 6.1). However, important differences in revenue-generating capacity across provinces remain, and these have increased over time. For example, not only did Beijing raise almost four times the amount of revenue, as a share of GDP, than Henan in 2002, but the latter’s revenue had remained virtually flat since 1995, while the former’s had grown by more than 8 percent of GDP (Table 6.2).

Table 6.2.Differences in Revenue Performance by Province, 1995 and 20021(In percent of provincial GDP)
19952002Change from 1995 to 2002
Top two provinces
Bottom two provinces
Unweighted average210.211.81.6
Source: China Finance Yearbook, 2003.

Province refers to the first tier of subnational governments.

Includes all provinces.

Source: China Finance Yearbook, 2003.

Province refers to the first tier of subnational governments.

Includes all provinces.

The reason for this disparity lies mainly in the structure of the taxes assigned to local governments, which tends to favor richer provinces. As explained above, local tax revenues derive mainly from shared VAT, business taxes, and EIT (together accounting for about 70 percent of total local governments’ revenue). The bases for these taxes typically cover the secondary (manufacturing) and tertiary (services) sectors; hence, coastal regions—where the share of GDP of secondary and tertiary industries is relatively high—benefit from the current fiscal system more than central and western provinces, whose economic structure relies heavily on the primary sector (agriculture-related taxes account for only 6 percent of local government taxes). A similar pattern applies to the personal income tax, whose yields are higher the larger the average household income—again, richer coastal provinces are favored.

Broadened Expenditure Mandates

The “recentralization” of revenue has not been accompanied by a reduction in expenditure assignments. To the contrary, expenditure pressures on subnational governments have intensified, especially when measured relative to their own resources. While local governments’ share of total expenditure has remained quite stable at around 70 percent during the last decade, local governments’ expenditure has nonetheless become increasingly burdensome relative to local governments’ own resources. For example, the ratio of local governments’ expenditure to their own revenue has surged from about 103 percent in 1991 to about 180 percent in 2002. With an appropriate system of central transfers, there would be no financing gap, but the transfer system suffers from a number of drawbacks (see below). In addition, local governments’ expenditure shares are above 90 percent in a number of areas (Table 6.3).

Table 6.3.Shares of Central and Local Governments in Selected Budgetary Expenditure, 2001
Shares of Total Expenditure (In percent)Share of Center (In percent of item)Share of Local Government (In percent of item)
Of which:
Culture, education, science, and health17.810.889.2
Capital construction13.434.365.7
National defense7.799.20.8
Technical updates, transformation, and science5.225.075.0
Tax administration5.035.164.9
Policy subsidies3.940.359.7
Urban maintenance and construction3.40.0100.0
Pension for retired employees3.29.190.9
Armed police1.292.37.7
Sources: Ministry of Finance data. Local expenditures include earmarked transfers from the central government.
Sources: Ministry of Finance data. Local expenditures include earmarked transfers from the central government.

Various factors explain the proliferation of local governments’ expenditure responsibilities. The industrial restructuring process has transferred previous spending responsibilities of SOEs—especially in social areas, such as education and health—to subnational governments as SOEs are reformed; increased and fast urbanization has created a need for subnational governments to provide basic infrastructure services (such as electricity and transportation); and the administration of pensions remains largely decentralized, in most cases all the way down to the county level, where there are signs of difficulties in paying pensions. In addition, minimum service standards set by the center create challenges for poorer counties, while richer subnational governments do not seem to experience similar difficulties. For example, the provision of healthcare services in rural areas—to be partly covered by specially-designed central subsidies—requires matching funds from receiving counties, further stretching their service provision capacity.

Inadequate Transfers

In addition to a clear redefinition of the tax assignments, the 1994 reform also redesigned the transfer system, moving away from ad hoc negotiated transfers toward a more rules-based and transparent mechanism. The transfer system is based on four main pillars.

  • Revenue returned provides each province with 30 percent of the increase in VAT receipts and excise tax collection over the 1993 base (the year before the reform).
  • Specific-purpose grants are earmarked transfers allocated on an ad hoc basis.
  • Subsidy transfers or general-purpose grants help ensure that each province has adequate resources. They are rules-based and depend on variables such as provincial GDP, student-teacher ratios, number of civil servants, and population density.
  • Fixed subsidies ensure that every province has total revenues no lower than subsidies in 1993.

Ahmad and others (2004) provide a snapshot of these transfers (Table 6.4). Central transfers are sizable, representing about 44 percent of total local government’s resources and financing 48 percent of local governments’ expenditure in 2002 (World Bank, 2002 and 2003). Despite their large size, transfers have nonetheless proved inadequate to provide sufficient financial support to the provision of essential services such as rural education and rural public health. The predominance of “transfers” under the revenue returned principle makes the current system implicitly regressive, as richer provinces receive most transfers. Although declining, these transfers still represent about 40 percent of total transfers from the center to the provinces. In contrast, general-purpose transfers (including transfers for income equalization), although growing, represent only a quarter of the total. The system’s ability to redistribute fiscal revenues across provinces therefore remains limited.

Table 6.4.Transfers from the Central to Local Governments
(In percent of GDP)
Revenue returned2.
Specific-purpose transfers0.
General-purpose transfers10.
Fixed subsidies under old system0.
(In percent of total transfers)
Revenue returned70.562.753.048.938.9
Specific-purpose transfers18.126.433.830.634.6
General-purpose transfers17.57.510.217.924.5
Fixed subsidies under old system3.
Sources: Chinese authorities; IMF staff estimates; and Ahmad and others (2004).

Include transfers for income equalization.

Sources: Chinese authorities; IMF staff estimates; and Ahmad and others (2004).

Include transfers for income equalization.

Specific-purpose grants comprise hundreds of different earmarked grants, allocated on an ad hoc negotiated basis. Their increasing share reflects the proactive regional policy that the center is carrying out. However, by their nature these grants make the transfer system less transparent and more difficult to monitor, as the center lacks the ability to track how the related funds are effectively spent; they also undermine the rules-based character of the transfer system that the 1994 reform aimed to introduce. Ongoing reforms to create a treasury system should assist in improving the transparency and control of expenditure at all levels of government. Finally, fixed subsidies, introduced to guarantee that every province maintained revenue no lower than the 1993 level, are progressively becoming less relevant and are gradually being phased out.

Uncertainty about the size and timing of central transfers further complicates budgetary formulation and execution at the local government level. Where local governments do not have sufficient resources of

Some Implications of the Current System

The imbalance between increased expenditure by subnational governments and their budgetary revenue has widened over time, resulting in a sharp deterioration of their fiscal position (before transfers) since 1994 (Figure 6.2). Such vertical imbalances are not uncommon in large federations (for example, Australia and India), as they allow the center to play a bigger role in macroeconomic stabilization and equalization. In China, however, the large share of revenue returned in total transfers limits the amount of unencumbered resources available to the center for redistribution and works against the principle of equalization (Ahmad and others, 2004).

Figure 6.2.Central and Local Governments’ Budgetary Positions1

(In billions of yuan)

Source: China Finance Year book.

1 Based on official data for budgetary revenue and expenditure (for local governments, revenue is before transfers).

Faced with intensifying financial pressures—resulting from limited tax setting powers, an inadequate transfer system, and a legal prohibition to borrow—local governments have continued to raise revenue outside the budget system, mainly in the form of fees and charges that accrue to locally managed extrabudgetary funds. While reported extrabudgetary funds managed by the central government declined significantly after the 1994 reform, those controlled by subnational governments have continued to increase (Figure 6.3).3

Figure 6.3.Extrabudgetary Funds

(In billions of yuan)

Source: China Finance Year book.

Box 6.2.Fiscal Risks of Local Governments: The 2002 Audit Report1

The Chinese authorities are aware of possible fiscal risks generated at the local government level. This box summarizes some key findings of the 2002 Audit Report, which illustrates some cases of such fiscal risks and identifies the underlying causes for their buildup.

Misuse of Government Loans

Some local projects financed by loans from the central government are plagued by loss, waste, and ineffective utilization of funds. For example, 37 sewage treatment projects costed at about RMB 6 billion received about RMB 2 billion in government loans. But, owing to inadequate planning, failure to provide counterpart funding, and insufficient operating funds, construction was not finished (or, in some cases, not even begun) on 15 projects; seven of 16 completed projects failed to meet design specifications, and the completion of supporting facilities frequently lagged behind the main projects, with the result that equipment lay idle and the programs failed to produce their planned overall effect.

The audit of projects involving 18 major airports and 38 secondary-route airports revealed that many of the airports had suffered huge losses and that their operations were in financial difficulty. Nine of 12 major completed airports were losing money, with the combined loss amounting to RMB 1.4 billion. Thirty-seven of the 38 secondary-route airports had also lost money, with a combined loss of RMB 1.5 billion in 2000–2001. In 2001, the passenger turnaround capacity of 38 secondary-route airports was only a quarter of that estimated in feasibility studies, and the Mianzhou airport achieved only 3 percent of its design capacity.

Insufficient Funding from the Center and Increasing Arrears

At the end of 2001, 49 cities and counties had a total of RMB 1.6 billion in debt, equivalent to 2.1 times their disposable financial resources for that year. In some cases, accumulated arrears kept on growing. As of September 2002, 42 counties and townships had RMB 1.8 billion in wage arrears—more than three times the arrears in 1998.

Underlying Reasons

The Audit Report also identifies some causes of these problems:

  • Lack of stable revenue sources at the subnational level. Currently, counties and townships rely on transfers from higher government levels for 60 percent of their expenditures. However, many earmarked subsidies require counterpart funds, which increases the fiscal pressure on these local governments.
  • “Improper attitudes toward fiscal management,” which induce local governments to “blindly set up projects and even create image projects.”
  • Excessively rapid growth in the numbers of persons supported by public funds (unfunded mandates). Since 1994, while the total population of counties and townships has increased by 4.6 percent, the number of people supported by public funds has increased by 22 percent.
1 The Audit Report is produced annually by the National Audit Office of the People’s Republic of China (CNAO), the official institution in charge of auditing government finances. The CNAO is required by the Constitution to report its findings to the State Council.

Financing constraints have also induced subnational governments to seek ways to circumvent their legal funding limits by creating channels for raising indirect financing and shifting some public functions to seemingly “nongovernment” entities. Large-scale infrastructure projects offer an example: in most provinces, there has been intense activity to build highways, airports, and urban ring roads. Most of these projects have been financed by bank loans, on the expectation that the sales of appreciated land leases in the areas where these projects are implemented will make the latter financially viable. However, these projects may also represent significant fiscal risks in case local governments, and ultimately the central government, are called upon to shoulder the associated fiscal costs. The shift of public functions to nongovernment entities to overcome legal borrowing constraints might also generate opportunities for waste and corruption, without necessarily improving effective service delivery. The authorities are aware of these problems, as candidly described in the 2002 Audit Report (Box 6.2).


The 1994 reform of the structure of intergovernmental fiscal relations has been successful in a number of areas: it has streamlined the tax system and enhanced tax administration, boosted revenue, and increased budgetary resources available to the center. However, there are growing signs that the transfer system, whose aim was to compensate local governments for revenue lost to the center and promote equalization across regions, is in need of reform; in particular, it has proven inadequate to address the large regional income disparities. Local governments’ expenditure mandates remain unclear and, in some cases, largely unfunded; pension costs are a case in point, where local governments are illequipped to shoulder the related costs and higher-level pooling is called for. Finally, indirect means of local government financing and creation of implicit liabilities at the local level may represent significant fiscal risks, underscoring the need for a comprehensive and centralized monitoring of subnational operations. These challenges will need to be addressed early on, not only to reduce regional disparities but, most importantly, to prevent erosion of social cohesion and weakening of public support for future reforms.

1China’s government structure has four subnational (local) levels: the first tier consists of provincial level authorities (22 provinces, five autonomous regions, and four municipalities—Beijing, Shanghai, Tienjin, and Chongqing). The second tier comprises prefecture level authorities (some 330 prefectures and cities at prefecture level). The third tier encompasses more than 2,100 counties/cities at the county level, while the fourth tier includes about 48,000 villages/townships—overall, some 50,000 entities.
2The 2002 budget introduced a new sharing formula for personal and corporate income taxes. According to the new formula, all proceeds from these taxes above the 2001 level are shared between the center and the provinces on the basis of a 50:50 ratio (changed to 60:40 in 2003). The share of these increased revenues received by the center is to be used for transfers to poorer provinces.
3A provincial pilot project, launched in 2001, aims to replace the numerous fees with a surcharge on the local agricultural tax. This so-called tax-for-fees reform is to be gradually extended to all provinces.

    Other Resources Citing This Publication