Soft-Budget Constraint on Local Governments in China
Jing Jin and Heng-fu Zou
A significant feature of China‘s economic reform since 1978 is the devolution of the central government‘s control over the economy to subnational governments. The fiscal system is
decentralized among five levels of government: national, provincial, municipal, county, and township governments, which are broadly categorized into center, provincial, and local governments (all sub-provincial governments). This paper mainly focuses on soft-budget constraints in the relationship between the central government and the provinces. The term ―local‖ and ―subnational‖ refers to the provincial level or aggregated subnational governments,
unless otherwise specified.
China‘s subsidy, taxation, credit and administrative pricing systems are all subject to soft budget constraints. Prior to 1994 under the Chinese fiscal regime, the collection of all taxes and profits followed the pre-reform pattern: local government collections were remitted to the center and then transferred back to the provinces according to expenditure needs approved by the center. Policymakers in the central government decided what type of revenues should be collected and how these revenues were to be reallocated for national and local public good provisions. Most expenditures at subnational levels were financed by central transfers and complemented by a few self-retained local tax receipts. The prereform fiscal system resulted in a fundamental lack of incentives and efficiency, which became the major concern of the central authorities. In the 1980s, a series of reforms were implemented to revamp the fiscal relations between the central
and subnational governments. Although incentives to spur tax collection efforts by local
1 they also reduced the share of revenues passed governments were successful to a certain extent,
on to the central government. Before the 1994 tax system reform, the central government‘s share of total revenue declined from 44 percent in 1978 to 23 percent in 1993, while the total subnational revenue share increased from roughly 56 percent to 77 percent during same period. At the same time, the consolidated government revenue share in GDP also shrank from 47 percent in 1978 to 13 percent in 1993. Despite the fact that fiscal decentralization in the 1980s shifted more resources to local governments in terms of its increased share in total revenues, the shrinking pie also considerably reduced the budgetary resources allocated at the provincial level (Table 1).
1 Unlike the previous system, reform in the 1980s allowed provincial authorities to retain all or a proportion of the
tax collected after sharing with the center.
Central and Provincial Government Revenue Shares in Total Revenue and GDP Year Tax Revenue Revenue Total Central Tax Total Subnational Share of Share of GDP Total
( Billion from Revenue Revenue Central Revenue Total Central Subnational (Billion Government
Yuan) Enterprises (Billion (Billion Yuan) Revenue (Billion Revenue in Revenue in Yuan) Revenue as a
[without SOE (Profit Yuan) (Billion Yuan) Total Total Percentage in
remittance] Remittance, Yuan) Revenue Revenue GDP
Billion (%) (%) (%)
1978 113.2 57.2 170.4 17.6 74.8 95.6 43.9 56.1 362.4 47.0 1979 114.6 49.5 164.1 23.1 72.6 91.5 44.2 55.8 403.8 40.6 1980 116 43.5 159.4 28.4 71.9 87.5 45.1 54.9 451.8 35.3 1981 117.6 35.4 153 31.1 66.5 86.5 43.5 56.5 486.2 31.5 1982 121.2 29.6 150.8 34.7 64.3 86.5 42.6 57.4 529.5 28.5 1983 136.7 24.1 160.8 49 73.1 87.7 45.5 54.5 593.5 27.1 1984 164.3 27.7 191.9 66.5 94.2 97.7 49.1 50.9 717.1 26.8 1985 200.5 4.4 204.9 77 81.4 123.5 39.7 60.3 896.4 22.9 1986 212.2 4.2 216.4 77.8 82 134.4 37.9 62.1 1020.2 21.2 1987 219.9 4.3 224.2 73.6 77.9 146.3 34.7 65.3 1196.3 18.7 1988 235.7 5.1 240.7 77.4 82.5 158.2 34.3 65.7 1492.8 16.1 1989 266.5 6.4 272.9 82.3 88.7 184.2 32.5 67.5 1690.9 16.1 1990 293.7 7.8 301.5 99.2 107 194.5 35.5 64.5 1854.8 16.3 1991 314.9 7.5 322.4 93.8 101.3 221.1 31.4 68.6 2161.8 14.9 1992 348.3 6 354.4 98 104 250.4 29.3 70.7 2663.8 13.3 1993 434.9 4.9 439.8 95.8 100.7 339.1 22.9 77.1 3463.4 12.7 1994/b 521.8 521.8 290.7 290.7 231.1 55.7 44.3 4675.9 11.2 1995 624.2 624.3 325.7 325.7 298.6 52.2 47.8 5847.8 10.7 1996 740.8 740.7 366.1 366.1 374.6 49.4 50.6 6788.4 10.9 1997 865.1 865.1 422.7 422.7 442.4 48.9 51.1 7446.3 11.6 1998 987.6 987.6 489.2 489.2 498.4 49.5 50.5 7939.6 12.4 Note: a/ Before 1984, a considerable amount of central revenue came from SOEs' Profit Remittance, which exclusively went to the central treasury. Since
1984, the Profit Remittance had been increasingly replaced by Enterprises Income Tax (Li gai sui). Profit Remittance from SOEs remained as a
residual category until 1993 before it terminated.
b/ Tax assignment system reform introduced. Data after 1994 are not compatible with those before 1994. Sources: China Statistical Yearbook (1999)
China Government Finance Yearbook (Various issues).
Throughout the 1980s, the central government‘s inability to cut spending to stay within declining revenue created persistent budget deficits, which contributed to mounting inflationary pressures. At the same time, subnational governments faced greatly expanded expenditure responsibilities stemming from obligations imposed by national policy (Wong, 1991). As the central government responded to fiscal pressure by attempting to devolve expenditure responsibilities to lower levels of government, it left provincial governments starved for revenues. Apart from the intensified bargaining between central and local governments over the sharing schemes, fiscal pressures created by the contract system of the 1980s led to undesirable responses by subnational governments. Examples include the diversion of resources from budgetary to extrabudgetary channels, the duplication of industries to capture revenues that formerly flowed to the national treasury, generous tax concessions to local state-owned enterprises (SOEs) under their own jurisdictions, and expanded local bank lending to these SOEs. All these measures circumvented the central government‘s efforts to impose hard budget
constraints and weakened overall financial discipline.
As the country moved toward economic federalism with the fiscal
decentralization marked by a continuous decline of government revenue as a percentage of national income, the unitary political system was also transformed and decentralized. Even though the central bureaucratic hierarchy continued to select, assign, and promote top provincial cadres (Huang, 1996), since 1983, bureau level officials (e.g., the heads of provincial fiscal bureaus and the managers of provincial branches of national banks) have been selected by provincial governments and appointed by the corresponding level of the People‘s Congress. No central approval is required. Driven by common economic
interests and the pressure to seek growth, which is the most important measure of their political performance, the directors of fiscal and banking agencies tend to ―stand where they sit‖ rather than delegate their central line administrators. As a result, the former
hierarchical management has been considerably weakened and increasingly transformed into horizontal administration featured by a highly fragmented economy. The central authority‘s attempt to strengthen the hierarchical management-- by strengthening
personnel management at the level of provincial party secretaries and governors-- thus may not necessarily be able to penetrate the horizontal coalitions increasingly shaped by common interests and contiguity at the subnational level. According to Yang (1997), the heads of faster-growing provinces now tend to be promoted more quickly than otherwise would be. Bo (1996) also finds that provincial leaders of more populous and richer provinces are more likely to be promoted than those in less populous and less developed ones.
This paper outlines some major economic and administrative mechanisms that undermine the central government‘s endeavor to harden the budget constraint on provincial governments. Section II describes briefly the evolution of China‘s
intergovernmental fiscal relations in the post-reform period. Section III presents the major channels of soft budget constraints on provincial governments. Section IV concludes.
II. The Evolution of Intergovernmental Fiscal Relations from the 1980s through the
(1) 1980-1993: Fiscal Contract System (Tax-Sharing System)
In 1980, the centralized fiscal regime was replaced with the fiscal contract system whereby each level of government contracted with the next level up to meet certain revenue and expenditure targets. Central and subnational governments shared the revenue proportionately or in the form of a fixed quota plus a percentage share. At the same time, subnational governments were required to finance their own expenditures through self-generated and shared revenues—a step in the direction of hardening the budget constraint on local governments.
Unlike other countries where taxes are collected by the central government and then allocated to subnational governments, in China local authorities collected all tax revenues and remitted a portion to the higher levels. The amount submitted to central coffer depended on provincial receipts, and the sharing formula between the center and provinces. Given such a highly decentralized revenue collection system, the center had to resort to various instruments to ensure revenue remittance from local authorities. These instruments, in turn, led to perverse reactions from the provinces, which always found ways to retain more revenues through their relaxed revenue collection for and the negotiations with the center regarding shared revenue.
From 1980 through the early 1990s, four revenue-sharing systems were employed, with many variations. Until the tax system reform in 1994, six different contract types were in use between the central government and provinces, with many more at the subprovincial level (Table 2) (also see World Bank, 1993, and Bahl and Wallich, 1992).
Revenue-Sharing System between the Central and Provincial Governments, 1988-1992
Type A Type B Type C Type D Type E Type F
Incremental Contract Basic Proportional Sharing and Remittance Incremental Fixed-Fixed-Proportional Incremental Sharing Contract Remittance Subsidy Sharing
Contracted growth Retention Proportion Incremental Remittance Incremental (100 million Yuan) rate(%) rate (%) sharing (100 million) Contract (%) Beijing 4.00 50.00 Hebei 4.50 70.00 Liaoning 3.50 58.30 Shenyang 4.00 30.30 Haerbin 5.00 45.00 Jiangsu 5.00 41.00 Zhejiang (exl. Ningpo) 6.50 61.50 Ningpo 5.30 27.90 Henan 5.00 80.00 Chongqing a/ 4.00 33.50 Tianjin 46.50 Shanxi 87.60 An Hui 77.50 Da Lian 27.70 27.30 Qingdao 16.00 34.00 Wuhan a/ 17.00 25.00 Guangdong 14.10 9.00 Hunan 8.00 7.00 Shanghai 105.00 Heilongjiang 2.90 Shangdong (exl. Qingdao) 4.90 Hubei (exl.Wuhan) 1.22 Ji Lin 1.07 Sichuan (exl. Chongqing) 1.79 Jiangxi 0.50 Sha'anxi 1.20 Gansu 1.30 Fujian 0.50 Inner Mongolia 18.40 Guangxi 6.10 Tibet 9.00 Ningxia 5.30 Xinjiang 15.30 Guizhou 7.40 Yunan 6.70 Qinghai 6.60 Hainan 1.40
Source: Ministry of Finance, P.R. China
Also see Bahl and Wallich, 1992, and World Bank, 1993.
a/: After the cities of Wuhan and Chongqing were treated differently from Hubei and Sichuan provinces, the provinces
changed from net providers to the state to net recipients of subsidies from the state.
Type A: Incremental contract – Based on 1987 revenues, the provincial retention
rate of all tax revenues ranged from 28 percent to 80 percent while local remittance to the center needed was to increase from 3.5 percent to 6.5 percent (contracted growth rate) on an annual basis. Tax revenues in excess of the stipulated growth rates was retained entirely by provinces.
Type B: Basic Proportional Sharing – A fixed proportion of all revenues was
remitted to the center.
Type C: Proportional Sharing and Incremental Sharing – A certain proportion of
the actual revenue collection of the previous year was retained, and then a different (usually higher) proportion of revenues was retained for the incremental amount in excess of the total revenues for the previous year.
Type D: Remittance Incremental Contract – A specific nominal amount was
transferred to the center in the initial year; in subsequent years, the remitted amount increased at a contracted rate (9 percent for Guangdong province and 7 percent for Hunan province).
Type E: Fixed Remittance – A specific nominal amount was transferred to the
center with no annual adjustments.
2 Type F: Fixed Subsidy – Deficit provinces received fixed subsidies.
Two crucial features survived every change in revenue-sharing systems. First, central fixed revenues were not subject to the revenue sharing, so that whatever was designated as central revenues left the pool of revenues to which revenue-sharing
2 Bahl and Wallich (1992).
3 Second, enterprise income, both remitted profits and direct tax formulas were applied.
revenues (after 1984), were still divided among governments according to their administrative subordination, i.e., state-owned enterprises subordinated to the central, provincial, and local governments, respectively.
Problems of the Fiscal Contract System
Central revenue declined as a percentage of total revenue. The decentralized nature
of tax collection meant that the central government lacked effective supervision of tax collections and remittances by provincial governments. Consequently, local governments avoided sharing revenues with the center through various means. Before the tax system reform in 1994, around 10 provinces and municipalities (include Beijing, Hebei, Liaoning, Shengyang, Haerbin, Jiangsu, Zhejiang, Nigpo, Henan and Chongqing) were under the ―incremental contract‖ scheme (Table 2). Take Jiangsu for example: if the total revenue
collected by Jiangsu provincial government was within the total amount of the previous year (taking 1987 as the base year) plus a 5 percent increase, Jiangsu provincial government could retain 41 percent of the total revenue collected. Any amount exceeding the total of increased revenue can be retained by the provincial government. Frequently tax revenues stagnated for years, limiting the amount for sharing with the center. This phenomenon was prevalent among the provinces under contracts A, B, and C. Assuming some growth in taxes accruing to the provinces by either transferring budgetary revenue to extrabudgetary items, or allowing generous tax recessions to local enterprises so that
3 E.g. income taxes from railways, coal mining, petroleum and airline industries as well as income taxes of
banks and insurance companies, etc.
benefits could be accrued within the enterprises under the jurisdiction of subnational governments—the center‘s share would decrease.
In other cases, tax remitted to the center was fixed in nominal terms for many years and growth was retained by the province. Guangdong, one of the fastest growing economies in China, is a case at point—the remittance from Guangdong was fixed at 1.4
billion Yuan for many years. Not until 1988 had the ―remittance incremental contract‖ been implemented (Table 2), under which the remittance of Guangdong was set at an annual increase of 9 percent with 1987 as the base year. By 1993, its remittance increased to 2.4 billion Yuan, barely 7 percent of its total 34.7 billion revenue. Hunan, a relatively
4 Jilin (excluding Changchun), and poor province, was under the same contract. Shanghai,
Shandong (exclude Qingdao) were categorized into ―fixed remittance‖ schemes, representing 10.5 billion, 0.29 billion, and 0.49 billion, or 46 percent, 4 percent and 3 percent of their respective total revenue in 1993.
With the power of tax collection provincial governments acted strategically to escape sharing their revenues with the center, which resulted in a decline of central revenue share vis-a-vis that of the local in total revenue (Table 1). Tax generation in such a fiscal system tends to be inelastic with respect to GDP and procyclical. In a rapidly growing economy with fiscal contracts containing a large fixed component, the rate of increase in tax revenues would be less than that of income growth. Tax policy thus becomes a procyclical mechanism that exacerbated economic fluctuations instead of
4 Before the tax system reform in 1994, Shanghai was far more heavily taxed with nearly half of its revenue
remitted to the center under the fixed-remittance term. However, it might be ―compensated‖ with a
credit allocation beyond the local deposit base (see World Bank, 1993).