Regional Decentralization and Fiscal Incentives federalism chinese style

1 We would like to thank Alberto Alesina, Bill Evans, Roger Gordon, John McMillan, Peter Murrell, Barry Naughton, Gerard Roland, Seth Sanders, Robert Schwab, Andrei Shleifer, Shang-Jin Wei, David Wildasin, Alwyn Young, Heng-fu Zou, and the participants at the Fifth Nobel Symposium in Economics held in Stockholm in 1999 for helpful comments and discussions and the Center for Research on Economic Development and Policy Reform at Stanford for financial support. 2 Corresponding author. Email: yqian@econ.berkeley.edu Regional Decentralization and Fiscal Incentives: Federalism, Chinese Style1 Hehui Jin Department of Economics Stanford University Yingyi Qian2 Department of Economics University of California, Berkeley and Barry R. Weingast Hoover Institution and Department of Political Science Stanford University This Version: December 2004 2 Abstract Aligning the interests of local governments with market development is an important issue for developing and transition economies. Using a panel data set from China, we investigate the relationship between provincial government’s fiscal incentives and provincial market development. We report three empirical findings. First, we find that during the period of “fiscal contracting system” the discrepancy between ex ante contracts and ex post implementation was relatively small, suggesting that the fiscal contracts were credible. Second, we find a much higher correlation, about four times, between the provincial government’s budgetary revenue and its expenditure during 1980s and 1990s as compared to 1970s, demonstrating that provincial governments faced much stronger ex post fiscal incentives after reform. Third, we find that stronger ex ante fiscal incentives, measured by the contractual marginal retention rate of the provincial government in its budgetary revenue, are associated with faster development of the non-state sector as well as more reforms in the state sector in the provincial economy. This holds even when we control for the conventional measure of fiscal decentralization. Finally, we compare federalism, Chinese style, to federalism, Russian style. JEL Classification: P35, P51, H77 1 I. Introduction Reforming the government is a crucial component of both the transition from a planned to a market economy and economic development. Creating thriving markets in these economies typically requires transforming a highly centralized and interventionist government into one that supports the market and fosters decentralized economic activities. Democracy, separation of powers, and the rule of law are among the important institutions that allow citizens to hold the government accountable for its economic actions and to secure markets from arbitrary state intrusion. By devolving power from the central to local levels, federalism is another institution that helps implement a limited yet effective government conducive to market development. Economic theories of federalism have traditionally emphasized allocative benefits of decentralization in the provision of public goods and services, such as education and health care. There are two related ideas. First, Hayek (1945) discussed the use of knowledge in society, emphasizing that local governments have better access to local information, which allows them to provide public goods and services that better match local preferences than the national government. Second, Tiebout (1956) introduced the inter-jurisdictional competition dimension and argued that such a competition provides a sorting mechanism to better match public goods and services with consumers' preferences. Drawing on these ideas, Musgrave (1959) and Oates (1972) built a theory of fiscal federalism, stressing among other things the appropriate assignment of taxes and expenditures to the various levels of government to improve welfare. Our main concern in this paper is the relationship between fiscal incentives facing local governments and local government promotion of market development in the local economy. Recent 2 3 The recent theory of “market-preserving federalism” studies the general question of how federalism can be structured to promote market development (e.g., McKinnon, 1997; Qian and Roland, 1998; Qian and Weingast, 1996; Weingast, 1995; Wildasin, 1997; and Zhuravskaya 2000). experiences of transition and developing economies have shown that a central barrier to economic development in these countries is from the governments, especially local governments, as their policies are often hostile to local business development. Local government policies, such as business regulation and levies, may have either favorable or adverse effects on the entry and expansion of local business enterprises. This leads to two types of government role that have been identified in the literature (Shleifer and Vishny 1998): The government either plays the role of the “grabbing hand” by restricting and preying on productive enterprises and protecting unproductive ones, or it plays the role of the “helping hand” by supporting productive enterprises and disciplining unproductive enterprises. Our study of federalism centers around the question of how the central-local governmental relationship affects the local government's behavior toward business enterprises and market development. A crucial issue is what kind of federalism better aligns local government incentives with promoting markets and productive enterprises.3 Inter-jurisdictional competition can serve as an important incentive device, as emphasized by Tiebout and Brennan and Buchanan (1980): competition rewards local governments friendly to markets as factors of production move to their regions, while it punishes heavily interventionist local governments as they lose valuable factors of production. But this mechanism is not perfect, competition may result in the phenomenon known as “race to the bottom.” Another mechanism, the focus of this paper, concerns the fiscal incentives of local government. This mechanism works when pro-business local government policy promotes local 3 business development, which rewards local governments by increasing the local tax revenue base. A critical aspect of this incentive, however concerns whether the local government is able to keep a significant portion of the increased tax revenue that results from their policy decisions. If so, they have strong fiscal incentives to support market development. On the other hand, if a local government’s fiscal reward is unrelated to, or even worse, negatively related to its policy effort, it has no fiscal incentives to support local business. Studies on China's transition to markets have long noticed the general local government support for local business development, especially in the non-state sector (e.g., Montinola, Qian, and Weingast, 1995). What are the reasons for the local governments in China to play the “helping hand” for local business development? Using a provincial panel data set, we conduct an empirical study on the Chinese style of federalism with a focus on provincial government’s fiscal incentives. We report three empirical findings. Our first two findings both concern the change in fiscal incentives facing provincial governments after the reform. First, in assessing the "fiscal contracting system" operating between the central and provincial governments from 1980-93, we find that the discrepancy between ex ante contracts and ex post implementation declined over time and was relatively small on average. These small differences imply that the fiscal contracts between center and province were credible. We also find that cases of extra subsidies were much more common than cases of extra revenue remittance. This suggests that, as far as the central-provincial relations in China, the “soft budget constraint” problem (Kornai, 1996) was a greater problem than the “predation” problem. Second, we find a strong correlation between the current provincial budgetary revenue and its expenditure for the period of 1982-91 when the “fiscal contracting system” was implemented, 4 4 The conventional, more readily available data of the ex post ratios of revenue retention over collection measures only the average realized revenue retention and is thus less suitable for the study of the effects of fiscal incentives on other economic variables. about four times as large in the magnitude as the before reform period of 1970-79, and such a strong correlation remained in the post-1994 period when the “fiscal contracting system” was replaced by the “separating tax system.” The finding provides the evidence that provincial governments in China faced much stronger ex post fiscal incentives after the reform. Our third finding concerns the effects of fiscal incentives on provincial economic development and reform. We use the ex ante marginal revenue retention rate of provincial governments, as specified in the fiscal contracts between the central and provincial governments in the period between 1982 and 1992, as the measure of fiscal incentives faced by provincial government.4 We find that stronger fiscal incentives are associated with faster development of non- state enterprises in terms of the employment growth rates in rural enterprises and in all non-state enterprises, even controlling for the conventional measurement of fiscal decentralization. Similarly, stronger fiscal incentives are also associated with greater reform in state-owned enterprises, as measured by the increased shares of contract workers in the total state employment and bonuses in total employee wages. With these results in mind, we compare federalism, Chinese style, with federalism, Russian style. Studies of Russia's transition stress the problematic role of the government in reform. Shleifer (1997) and Frye and Shleifer (1997), for example, provide evidence that local governments in Russia have been playing the role of “grabbing hands” that retard private business development. Zhuravskaya (2000) finds that the existing revenue sharing schemes between the Russian regional and local governments provide the latter with no fiscal incentives to increase their tax base: increases 5 in local government revenues were almost entirely exacted by the regional government. The lack of fiscal incentives in part explains why local governments in Russia prey on private businesses. Our perspective suggests that the distorted incentives faced by local governments in part explains the disappointing performances of the Russian reforms. Interestingly, Russia has done more than China in terms of privatization of state-owned enterprises and liberalization of markets (Shleifer, 1997; Frye and Shleifer, 1997; de Figueiredo and Weingast 2001; Lavrov, Litwack, and Sutherland 2000; and OECD 2000). But apparently it has failed to provide local governments with appropriate fiscal incentives to pursue local prosperity. Liberalization and privatization without altering government fiscal incentives are insufficient to produce meaningful economic reform. We emphasize the critical importance of government fiscal incentives for successful reform. In addition to fiscal incentives, political incentives facing local governments also matter in comparing the Chinese and Russian federalism (Blanchard and Shleifer, 2000). A complementary approach studying the political incentives facing local governments in China has recently began. For example, Maskin, Qian and Xu (2000) documented an empirical correlation between the provincial economic performance and the provincial representation in the Party Central Committee. Li and Zhou (2004) found evidence that the central government uses personnel control over promotion and dismissal of provincial top leaders to induce provincial economic growth. The remainder of the paper is organized as follows. Section II describes the changing fiscal system in China during the reform. Section III develops our theoretical perspective. Section IV describes the data and the construction of variables. Section V assesses the credibility of the “fiscal contracting system” for the period of 1982-92. Section VI examines the ex post link between provincial revenue collection and its expenditure before and after reform. Section VII estimates the 6 5 Below the township level, the village is an informal level of government. A municipality can be one of the levels of a province, prefecture, or county; most municipalities are at the prefecture level. effects of fiscal incentives on provincial development and reform. Our conclusions follow. II. Fiscal Relations Between the Central and Provincial Governments in China China's fiscal system has five hierarchical levels of government: (1) central; (2) provincial; (3) prefecture; (4) county; and (5) township.5 In this paper, we will focus on the central-provincial fiscal relations. The central-provincial fiscal relations have evolved over time in three distinct phases: the pre-reform phase prior to 1979, the transitional phase of 1980-93, and the post-1994 phase. Prior to the reform of 1979, the fiscal relations between the central and provincial governments are best described as the one of “unified revenue collection and unified spending” (tongshou tongzhi). Basically, the provincial governments collected most of revenue generated from within the province, on average over 80 percent, which included taxes and (mostly) profits from state-owned enterprises. Then the central government made a plan of spending for each province. This system earned a nickname “eating from one big pot” (chi daguofan), which captured its essence. Starting 1980, the central-provincial fiscal relations altered in a dramatic way. Between 1980 and 1993, the institution governing the central and provincial fiscal relations is the so called “fiscal contracting system” (caizheng chengbao zhi), also known by its nickname “eating from separate kitchens” (fenzao chifan). Under the fiscal contracting system, provincial governments entered into relatively long-term fiscal contracts (typically five years) with the central government. Because of 7 their experimental nature, the contractual arrangements varied across provinces and over time. The fiscal contracting system worked as follows (Wong, 1997; Bahl, 1999): First, “central fixed revenue” was defined to include custom's duties, direct tax or profit remittance from the central government supervised state-owned enterprises (SOEs), and some other taxes. All other revenue falls under the heading “local revenue.” On average, the local revenue accounted for about 66% of total government budgetary revenue over these years. Second, the local revenue was then divided between the central and provincial governments according to pre-determined sharing schemes. For example, between 1980 and 1987, Guangdong province agreed to remit a fixed amount per year, and between 1988 and 1993, it agreed to remit an amount that increased by a fixed 9 percent per year. Guizhou province agreed to receive subsidies that increased by a fixed 10 percent per year. On the other hand, Jiangsu province agreed to remit a fixed share of revenue to the central government. Over time, many provincial governments retained 100 percent of the total local revenue at the margin, which effectively made them residual claimants over the local revenue. The actual (ex post) expenditure of local government did not necessarily match that from the sharing scheme for several reasons. After the division of local revenue according to the sharing scheme, some extra remittance and transfer payments took place between the central government and the provinces. For example, the central government sometimes “borrowed” funds from the provinces. On the other hand, the central government also made additional transfer payments (not specified in the sharing schemes) to provinces, which generally fell into two categories: earmarked subsidies (zhuanxiang butie), such as price subsidies for urban residents compensating them for food price increases, and matching grants (peitao buokuan), such as funds for highway building. Clearly, the larger this type of ex post redistribution, the less important the pre-determined revenue sharing 8 schemes. Starting 1994, the “fiscal contracting system” was replaced by “separating tax system.” Under the new system, “local revenue” has been redefined as revenues from local taxes and the local portion of the shared taxes (Bahl, 1999). The major local taxes are now the income taxes from all enterprises other than central government enterprises, business tax from the sales of services, and personal income tax. The most important shared tax is the value added tax (VAT), of which 25% belongs to the provincial government, uniform across provinces. The post-1994 phase has eliminated the variations of the revenue sharing rules from the 1980-93 phase. In addition to the budgetary revenue, another category of revenue exists called “extra- budgetary revenue,” which consists of tax surcharges and user fees levied by central and local government's agencies, as well as some earnings from SOEs. The extra-budgetary revenue emerged in the 1950s but only became institutionalized after the reform. Unlike the budgetary local revenues, the extra-budgetary local revenues are not subject to sharing with the central government. In 1978, total extra-budgetary revenue was about 10 percent of the GDP while total budgetary revenue was about 31 percent. In 1993, the extra-budgetary revenue was up to 16 percent of the GDP and the budgetary revenue was down to 16 percent of the GDP (Statistical Yearbook of China, 1995). While about three-quarters of the extra-budgetary funds are earnings retained by SOEs and by their supervisory government agencies at the central and local levels, about 30 percent of the extra- budgetary funds are used for government expenditures to supplement the budgetary funds (Fan, 1996). 9 III. Fiscal Incentives and Economic Development: A Theoretical Perspective How do fiscal incentives of local government contribute to the local economic development? As Hayek stressed, decentralization of authority has the benefits of more efficient use of dispersed local knowledge possessed by the local government. In contrast, centralization of government authority is costly because information transmission from local to central government is often distorted and incomplete. But decentralization of authority is meaningless if the central government takes away all revenue generated in the local economy as a result of local government’s action. This suggests a link between fiscal incentives of local government and local development which is a function of local government’s policy. Consider the following very simple model for the purpose of illustration. Let Y(e) be the value created by the local business development, which is a function of local government’s “effort” e. This effort is related to local government’s policies concerning local productive enterprises, such as the non-state enterprises. These policies could reduce excessive regulation and controls over business entry, speed up of approval of projects and permits, eliminate of onerous fees imposed on firms, or fight against unfavorable ideology toward the development of non-state firms. This effort is also related to reform efforts in non-productive state enterprises in order to reduce their losses. Because effort here is interpreted as policies rather than public goods provisions such as those in education and health care, its effect on the local economy is immediate. Higher local government effort means more favorable local business environm