POLICY RESEARCH WORKING PAPER 1608 Fiscal Decentralization, Fifeenyears of efforts to promote flscal Public Spending, and decentralization in China have tfailed to promote economic Economic Growth in China grhinChinasprovinces Ths finding is surprising in the Tao Zhang light of arguments that fiscal Heng-fu Zou decentralzation usually promotes provincial or local economic growth. The World Bank Policy Research Department Public Economics Division May 1996 I POLICY RESEARCH WORKING PAPER 1608 Summary findings Zhang and Zou use data on China to demonstrate how reforms begun in China in the early 1980s have probably the allocation of fiscal revenue and expenditures between failed to promote the country's economic growth. central and local governments has affected economic This result is consistently significant and robust in their growth since reforms that began in the late 1970s. empirical examinations. It is also surprising, in the light They find a higher degree of fiscal decentralization of the argument that fiscal decentralization usually associated with lower provincial economic growth over contributes positively to provincial or local economic the past 15 years in China. This implies that fiscal growth. This paper - a product of the Public Economics Division, Policy Research Department - is part of a larger effort in the department to study fiscal decentralization and economic growth. The study was funded by the Bank's Research Support Budget under the research project "Fiscal Decentralization and Economic Growth" (RPO 680-02). Copies of this paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Cynthia Bernardo, room N1O-053, telephone 202-473-7699, fax 202-522-1154, Internet address prdpe@worldbank.org. May 1996. (46 pages) The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be used and cited accordingly. The findings, interpretations, and conclusions are the authors' own and should not be attributed to the World Bank, its Executive Board of Directors, or any of its member countries. Produced by the Policy Research Dissemination Center Fiscal Decentralization, Public Spending, and Economic Growth in China* Tao Zhang and Heng-fu Zou Policy Research Department, World Bank 'We thank Hamid Davoodi, Shantayanan Devarajan, Gang Fan, Bert Hofman, John H. Jia, Hongyi Li, Stephen McGurk, Richard Newfarmer, Yingyi Qian, Terry Sicular, Christine Wong, and Danyang Xie for their comments and suggestions. The authors also acknowledge financial assistances from the Washington Center of China Studies, Ford Foundation, and the World Bank (Research grant RPO #680-02). 1 Introduction Many developing countries and transitional economies have a mandate to de- centralize some aspects of their public finance. In addition, many developed economies such as the United States, the United Kingdom, and Canada are reviving their policy debates on devolution. Decentralization of expenditure and revenue decisions of the central government is seen as part of a package to improve the efficiency of the public sector, cut the budget deficit, and pro- mote economic growth (Bird and Wallich, 1993; Bahl and Linn, 1992; Rivlin, 1992; Gramlich, 1993; Oates, 1993; and Bird, 1993). The argument is that decentralization will increase economic efficiency since local governments are better positioned to deliver public services that match local preferences and needs than the national government (Oates, 1972). Over time, efficiency gains will lead to fast local as well as national economic growth. This wisdom is shared by numerous studies on intergovernmental fiscal relations in China (Bahl and Wallich, 1992; World Bank, 1990, 1992, 1995). Many proposals favor assigning more revenue and expenditure responsibil- ities to localities from the center. However, a concern has emerged that decentralization in China has been implemented too fast and has gone too far, and that this is threatening macroeconomic control and stability. Fur- thermore, in this process, national priorities in public spending have often been crowded out by local public projects. Despite the foregoing policy concerns, there has been no empirical at- 3 tempt to explore the relationship between fiscal decentralization and eco- nomic growth in developing countries in general and China in particular. In this paper, we use the Chinese case tS demonstrate how the allocation of fiscal revenues and expenditures between the central government and local governments has affected economic growth since reforms of the late 1970s. The outline of this paper is as follows. A theoretical model of fiscal decen- tralization and growth is laid out in the next section. Section 3 summarizes the trends in fiscal allocations between the center and provincial govern- ments. Section 4 contains the empirical application of the theoretical model to the Chinese economy. Section 5 concludes the paper. 2 A Growth Model with Different Levels of Government Spending Following Barro (1990)1, we begin with an endogenous growth model consist- ing of a production function with two inputs: production capital and public spending where the function exhibits constant returns to scale in the two in- puts. We depart from the Barro model by assuming that public spending is carried out by two levels of government: central and provincial governments. Let k be the capital stock, g the total government spending, f central gov- ernment spending and s local government spending: f +s=g (1) 'Some related models are presented in Devarajan, Swaroop and Zou (1993), and Davoodi, Xie, and Zou (1995). 4 The production function is CES: y = [ak4 + OfI + 7ysl] /l,) -oo < < 1 (2) where a, ,B, and -y are all in (0,1) and a + /3 + y = 1. The CES production functions include the Cobb-Douglas specification as a special case (o = 0). The introduction of public spending by different levels of government creates a potential link between fiscal decentralization (i.e. differential effects of spending by the two levels of government) and growth. As in Barro (1990), when specifying the production function we abstract from human capital and labor, but we allow for these additional inputs in the empirical work. The consolidated government spending g is financed by a flat output tax at rate T: g = 'ry (3) To derive the long-run growth rate of the economy, we first analyze the decisions made by the production-consumption sector. We consider a long- lived producer-consumer unit which maximizes its discounted utility, max i [C l] e-Ptdt (4) where c is the consumption of a single good produced in this economy; a is the inverse of the intertemporal elasticity of substitution; and p is the rate of time preference. The dynamic budget constraint for the producer-consumer unit is: = (1- r) [aek' +±f 0 + as+] - c, ko given. (5) 5 The producer-consumer unit takes as given the government's announce- ment of the fixed tax rate r,the spending at different levels of governments, f and s. It then chooses optimally the consumption path {c(t): t > 0} and the path of the capital stock {k(t): t > O}. To characterize the producer- consumer unit's optimal allocation of resources, we write down the Hamilto- nian: H 1a _ Iy + A {(I -T) [akO + #f + s]/c} (6) The first order conditions are given by, A x(7) A = pA - Aa(l - r) [akO + Of f + ysO] I1 k)/4 " (8) The transversality condition is kAe-Pt -- 0 as t approaches infinity. Equations (5), (7), (8) with the initial transversality conditions determine the producer-consumer unit's optimal responses. One immediate result from these equations is that the growth rate of consumption is given by, c= r(x)- p(9 c al where x denotes the vector (k, f, s, r); r(x) has the interpretation of the real interest rate and is defined by r(x) = a(1 - r) [akO + Of + tys kO-. (10) Let us define the spending shares for the central and local governments as Vof and W., respectively: (tOf = Ws,fa = 9 (11) g g 6 Then, we can solve the long-run growth rate, G, of the economy explicitly as spending shares, income tax, and other exogenous factors: G T) -T 1P (12) Thus, the allocation of public spending among different levels of govern- ment can affect the rate of economic growth as seen from equation (12). To examine how the long-run growth rate responds to various spending shares and income taxation, we assume that the government's objective is to max- imize the growth rate in (12) by choosing r, pf and s. This is the same as maximizing the producer-consumer unit's consumption growth (which co- incides with the rate of growth of output and capital) in (9) subject to the government budget constraint of (3). Hence the problem can be formulated as follows: r(x)-p (13) f,s,r 1J subject to: f + S < r [akX + /3f + ±ys'l 0] (14) The growth-maximizing tax rate is given in the following equation: = + ') (15) where II = 31/(10-) ± 5,l/('-). The growth-maximizing shares of the central and local government spend- ing are given by, OV- (16) 71/(1-0) 5°s= II (17) 7 From the results above, we can obtain the real interest rate r(x) as follows: r(x) = (1 - -)(y/k)'- = C(1 - r) {IIX - vI} (18) where T is the solution from equation (22). The maximum rate of growth of consumption can then be calculated using equation (9). Here we can interpret 31/(0-) and 7l/(l-) as the measures of the individual productivity of spending by the central and local government, respectively. In the same light, rI = 1/(l-0) +± l/(0-) represents the aggregate productivity of both levels of government spending . From equations (16) and (17), it is not difficult to see that the growth-maximizing spending shares are equal to the ratios of individual productivity to the aggregate productivity. If the actual spending shares do not correspond to these growth-maximizing shares, some reallocation of resources among the two levels of government will be growth- enhancing. This point can be made most clearly in the case of the Cobb-Douglas production function. With the Cobb-Douglas technology, q = 0. Then, the growth-maximizing tax rate given by equation (22) is very simple: -7* =0 +±Y, (19) which is the same as the formula in Barro (1990) after the notation is made consistent. II is simply equal to (oB + y). Thus the growth-maximizing shares of the central and local government speeding are also very simple: *= + (20) - (21) 8 To test the impact on economic growth of spending by different levels of government, we use provincial panel data on China over the period of 1978 to 1992. But before our formal statistical analysis, we take a brief look at the data to see the patterns of revenue and expenditure allocations between the center and provinces and among provinces in China. 3 Trends of Intergovernmental Fiscal Alloca- tions in China: 1978-1992 Since the late 1970s, several rounds of fiscal reform have been conducted in China in an effort to decentralize the fiscal system and fiscal management (World Bank 1990; Wong et al 1993; Zhou and Yang 1992). Can we say that the resulting Chinese fiscal system is decentralized? The following exami- nation of the Chinese central-local fiscal status suggests that the question should be answered very carefully.2 3.1 Trend in overall fiscal status Following the public finance literature, we measure the overall fiscal status in China as the spending-to-GDP ratio of all governments, including the central and local governments.3 Government spending is measured in three differ- ent ways: budgetary spending, extra-budgetary spending, and consolidated 2Local governments include all sub-national governments in this paper. 3The data used in our calculation is described in detail in the Data Appendix. 9 spending, which is the sum of budgetary and extra-budgetary spending. The time trends of government spending are shown in Figure 1. The spending-to-GDP ratio in budgetary finance was 18.27 percent in 1992 compared to 30.77 percent in 1978. Although there were insignificant rises from 1978 to 1979, from 1985 to 1986, and from 1988 to 1989, the budgetary spending-to-GDP ratio declined continuously since the beginning of the reform in 1978. On the extra-budget side, changes in spending-to-GDP ratio were merely arithmetic, although the ratio rose from 14.17 percent in 1982 to 15.20 percent in 1992.4 The consolidated budget exhibited an inverted U-shape with the measure of spending-to-GDP ratio first up during 1982-1986 and then down during 1986-1992, from 36.43 percent in 1982 to 40.35 percent in 1986, and to 33.47 percent in 1992. The above results show that the overall fiscal status of consolidated gov- ernment weakened during the reform period, especially for budgetary finance. This trend can also be confirmed by the measure of revenue-to-GDP ratio of consolidated government revenue, as shown in Figure 2. 3.2 Relative status of the central-local budget Fiscal decentralization is mainly measured by the relative size of the local budget with respect to the central budget. Despite China's unitary sys- tem, in which tax laws and tax policies are set by the central government and applied uniformly nationwide, there are no delineations of powers and 4The central and provincial-aggregated data of extra-budgetary spending became avail- able in 1982. 10 responsibilities between the central and local governments in practice. In many respects, as pointed out in a World Bank study, "the Chinese system functions as a federalism" (World Bank 1990). Therefore, both central and local governments have contributed significantly to the allocation and uti- lization of public resources. On the spending side, the central government had a of 46.89 percent share of the total government budgetary spending in 1978. This share became 42.60 percent in 1992, indicating small progress in budgetary decentralization. But a closer examination reveals that the central spending share first kept rising until 1981 when it reached at 54.01 percent, then went down as far as 36.35 percent in 1989, and subsequently rose again to almost return to its original level. Although the total spending share of the central government rose and fell, the central budgetary spending share declined over most of the past decade, as shown in Figure 3. 3.3 Central-local enterprise relations Turning to extra-budgetary revenue and spending, we see quite a different picture. In China, government budgets include not only budgetary funds but also extra-budgetary funds (EBF), about 80 percent of which come from the income and profits generated by state-owned enterprises (SOEs) (World Bank 1990).5 Given the nature of EBF, fiscal decentralization can be achieved if 5Extra-budgetary funds are created to supplement budgetary resources among govern- ments, public agencies, and SOEs. They are usually managed and allocated with a certain amount of revenue assigned by the central government. Local governments can also create such assignments to their "owned" SOEs under the delegation of the central government. For details, see World Bank (1990), Shirk (1993), and Wong et al (1993). 11 the central government shifts public resources to local governments through assigning more SOEs to local governments. To measure fiscal decentralization this way, we use the central shares of extra-budgetary revenue (EBR) and of extra-budgetary spending (EBS). That is, a higher degree of decentralization is indicated by a smaller central share of the EBR (and/or EBS). As shown in Figure 3 and Figure 4 respectively, the central shares in EBR and EBS showed similar movements over time. The central EBR share rose from 33.72 percent in 1982 to 41.57 percent in 1985, and again to 44.30 percent in 1992. The central EBS share also increased: from 30.91 percent in 1982 to 40.87 percent in 1985, and again, to 43.64 percent in 1992. Unlike central budgetary spending, figures in central EBR and EBS manifested a trend of fiscal decentralization over the post-reform period. 3.4 Fiscal decentralization in provincial perspective We now further explore fiscal decentralization by looking into government spending in twenty-eight provinces.6 In addition to the complications in the central-local (aggregate) fiscal relations described earlier, more variations can be found in fiscal decentralization on the basis of provincial comparisons. First, there is significant cross-province heterogeneity in the degree of fis- cal decentralization. The ratio of budgetary spending to provincial income on average (1980-1992) ranged from 8.96 percent in Jiangsu (a coastal province) 60f the total thirty provincial areas in China (mainland), two provincial areas, Tibet and Hainan, are excluded due to their special status. For a complete list of the twenty- eight provincial areas used in this study, see the Data Appendix. 12 to 40.52 percent in Ningxia (an inland minority province), indicating a gen- eral tendency for provincial government to participate less in developed ar- eas than in less developed areas. Further complications came from the three metropolitan cities, Beijing, Tianjin, and Shanghai, which held the highest ranks in per-capita income and above average levels of the ratio of budgetary spending to provincial income. Table 1 shows the provincial variations in provincial per-capita income and the size of provincial budget in the period 1980-1992. In measures of fiscal decentralization, cross-province disparities are also evident. As shown in Table 2(a), 2(b), and 2(c), on average in the period of 1978-1992, the ratio of provincial budgetary spending to central budgetary spending ranged from 1.08 percent in Ningxia to 8.91 percent in Guangdong (a leading province in economic reform). For extra-budgetary funds, the ratio of provincial consolidated budgetary spending to central consolidated budgetary spending varied from 0.91 percent in Ningxia to 10.28 percent in Liaoning, one of China's heavy industrial centers. Measured by the ratio of per-capita provincial budgetary spending to per capita central budgetary spending, fiscal decentralization was as low as 70.57 percent in Guizhou (a mountainous minority province) and as high as 436.18 percent in Beijing (the nation's capital). Second, the intertemporal picture in fiscal decentralization is also diver- sified among provinces. Guangdong, a coastal province favored by central government policies and among the first to start economic reforms in 1978, experienced the greatest increase in fiscal decentralization. Measured by the ratio of provincial budgetary spending to central budgetary spending, there 13 was a 112.97 percent increase in Guangdong from 1978 to 1992.7 At the other extreme, Qinghai, one of the eight minority provincial areas, hardly saw any changes in fiscal decentralization, with its provincial-to-central bud- getary ratio narrowing from 1.31 percent in 1978 to 1.20 percent in 1991, and to 0.99 percent in 1992. Between Guangdong and Qinghai stand mostly in- land provincial areas. Sichuan, the most populous province in China, started with its provincial-to-central spending share at 6.86 percent in 1978 (higher than Guangdong's 5.51 percent in 1978) and ended with only 8.54 percent in 1992 (lower than Guangdong's 11.73 percent in 1992). Henan, a long-time center of ancient China, also saw changes in a much narrowed range, from 6.47 percent in 1978 down to 5.56 percent in 1992 in its provincial-to-central budgetary spending ratio. Comparisons of fiscal decentralization in these four provinces are presented in Figure 5. 3.5 Summary The above discussion suggests that fiscal reform in China does not yield a clear pattern of decentralization. (1) Budgetary spending became more de- centralized since 1978. (2) Extra-budgetary spending, however, showed an increasing central share through the entire reform period. (3) The consol- idated (budgetary and extra-budgetary) central spending share fluctuated, starting with 42.09 percent in 1982, 43.44 percent in 1985, 37.54 percent in 1987, and ending at 43.09 percent in 1992. (4) On the revenue side, man- 7Due to data availablity, intertemporal comparisons of provincial fiscal decentralization employ only the ratio of provincial budgetary spending to central budgetary spending. 14 agement and collection became even morje centralized. Combining budgetary and extra-budgetary funds, the central revenue share increased from 27.57 percent in 1982 to 41.92 percent in 1992, indicating a strengthened central control in governmental revenue. (5) Fiscal decentralization was found to be highly divergent across provinces and over time. In the following section, we will quantify the impact of various decentralization measures on provincial economic growth. 4 Empirical Estimations with Provincial-Level Data 4.1 Variables and estimation equation Our empirical estimations are based on the annual data over the period from 1980 to 1992 for 28 provinces. The dependent variable is the provincial income growth rate in real terms.8 The explanatory variables fall into three categories: (1)production inputs; (2) measures of fiscal decentralization; and (3) others, such as tax rates, foreign trade, and inflation rates. We use the following data in our estimations: Y = real growth rate of provincial income, measured in the percentage change; L = growth rate of the provincial labor force, measured by the percentage change in the total number of the labor force; 8Provincial income is defined as provincial equivalenct of national income (Guomin Shouru), which measures net provincial output according Chinese statistics. 1 15 I = provincial investment rate, measured by the ratio of investment (ac- cumulation in fixed asset and circulating funds) to provincial income; F = degree of openness of provincial economy, measured by the share of total volume of foreign trade (exports and imports) in provincial income; T = degree of distortion in provincial economy, measured by the ratio of provincial revenue collection to provincial income; R = inflation rate, measured by the overall social retail price index in each province; DC = degree of fiscal decentralization, measured by the following six different indicators: DCc.e. = the ratio of total provincial spending to total central spending; DCcse2 = the ratio of per-capita provincial spending to per-capita central spending;g DCbel = the ratio of provincial budgetary spending to central budgetary spending; DCbe2 = the ratio of per-capita provincial budgetary spending to per- capita central budgetary spending; DC~eib= the ratio of provincial extra-budgetary spending to central extra- budgetary spending; DCebe2 = the ratio of [provincial extra-budgetary spending/provincial in- come] to [central extra-budgetary spending/national income]; 9The central per capita spending is the central spending divided by the total population of China. 16 To these provincial-level data we fit our growth model as follows: 28 YDt = 13m Mt + 3nN8st + /3dODC8t ± oSzD8 + u8t (22) s=1 where s and t indicate province and year, respectively, Mst is a set of variables always included in the regression, DCst denotes variables of interest, Nt is a subset of variables identified by the literature as potentially important explanatory variables of growth, Ds denotes provincial dummy variable, and finally, ust denotes error term. Numbers of variables included in the M- and N-sets are denoted by m and n respectively. The M-variables consist of the growth rate in the total labor force (L) and the tax rate (T). The labor growth rate is not explicitly considered in our theoretical model for analytical simplicity. The tax variable is our aggregate measure of distortion introduced by governments to finance their spending. Other variables identified as potentially important explanatory variables of growth in many studies on economic growth are included as the N-variables. They are degree of openness (F), the inflation rate (R), and finally, the in- vestment rate (I). The usual argument for including the degree of openness as a determinant of growth states that more exports lead to more efficient resource allocation as a result of external competition in the world market and more imports are the means to import advanced technology from devel- oped economies (see, Feder 1983). Inflation can generate a positive effect on growth because higher inflation leads people to invest more in physical capi- tal and cut their real-balance holdings (the Tobin portfolio shift effect), but at the same time, inflation raises the transaction cost of economic activities (consumption and investment) and may reduce the rate of economic growth. 17 The investment rate appears as a "must-include" variable in conventional specifications of growth estimation. But in our analysis, it is endogenous. In order to make sure that our results are robust across different specifications of regression equations, we also include the investment rate as one explanatory variable in our sensitivity analysis. From our theoretical model in equation (12), our central concern is the third set of variables DC8t in (22): the six different indicators of fiscal decentralization. 4.2 Regression results 4.2.1 Base case As our base case, we first choose the M-variables and one of the six indicators of fiscal decentralization, DCbe2, the ratio of per-capita provincial spending to per-capita central spending, while ignoring the potentially important N- variables. The LSDV (Least Squares Dummy Variables) regression results over the 1986-1992 period are: Yst = 0.274L,t -0.407Tt -0.05DCbe2., (0.878) (-2.628) (-3.541) (Adjusted R2 = 0.328, number of observation = 196; values of t-statistics appear in parentheses). The M-variables have signs as predicted by our model but are not significantly different from zero at the 5 percent signifi- cance level. Our primary concern is the sign and magnitude of the coefficient for decentralization, which is -0.05 and is significantly different from zero at the 5 percent significance level. The regression result implies that fiscal 18 decentralization in China did not promote provincial economic growth, indi- cating an inappropriate level of decentralization or too much decentralization in government budgetary spending. To see whether this conclusion is robust to changes in the conditioning information set, we conducted sensitivity tests against the base equation. Table 3 presents the sensitivity results for each of the M-variables and the indicators of fiscal decentralization. Eight estimations are conducted along with different selections of the three N-variables. The labor coefficients are positive but not significantly different from zero at the 5 percent confidence level, and the non-significance result is consistent between the lower bound and the upper bound of the labor coefficient. Similar results are observed with the coefficients of the tax rates. Our major variable of interest is fis- cal decentralization, DCbe2. The decentralization coefficient is negative and robust.'0At the upper bound, the decentralization coefficient is -0.047 with the t-statistic of -3.67, while it becomes -0.069 with the t-statistic of -4.821 at the lower bound. This result implies that economic growth falls by as little as 4.7 percent and by as much as 6.9 percent for each additional unit of decen- tralization measured by DCbe2 the ratio of per-capita provincial budgetary spending to per-capita central budgetary spending. 4.2.2 Structural changes To further investigate our result of a negative effect of fiscal decentralization on Chinese provincial economic growth, structural changes are introduced in lFollowing Levine and Renelt (1992), we say the result is robust if the regression co- effiecient remains significant and of the same sign at the extreme (lower and upper) bounds. 19 the following two ways with respect to the base case: (1) different sample periods, (2) different selections of decentralization indicators, and (3) cross- province estimations based on provincial average values in the period 1986- 1992. Table 4 contains results of structure changes introduced in the base case. First, the negative correlation between fiscal decentralization and growth maintains for all the three sample periods: 1986-1992, 1980-1992, and 1985- 1989. The first sample period is our base sample period for which we have consistent data from the State Bureau of Statistics (SBS). The second one extends the base sample period to the beginning of reforms and contains data from provincial bureaus of statistics in addition to those from SBS. The third sample is selected to specify the period during which extensive fiscal decentralization was observed. As shown in Columns (1) - (3), the decentral- ization coefficients are consistently negative and significantly different from zero at the 5 percent level of confidence. The magnitude of the negative effect of decentralization on growth, however, changes across different samples: in the sample covering the entire reform period, the negative effect of decentral- ization on growth appears weaker than in the base case, while in the sample of extensive decentralization, the negative effect becomes stronger. Second, structural changes are also introduced by testing the growth ef- fect of four other indicators of decentralization. Instead of measuring de- centralization only with budgetary spending, we select DCcse2, measuring decentralization with per-capita total (budgetary and extra-budgetary) gov- ernment spending, to conduct our estimation. The result is shown in Column (5). Following international conventions, we also choose DCcsei, the ratio 20 of total provincial spending to total central spending, as the decentraliza- tion measure and the result is shown in Column (1). We further test the growth effect of decentralization by dividing total spending into budgetary and extra-budgetary spending. Column (6) shows the results with decentral- ization measures in both budgetary and extra-budgetary spending. Column (7) contains the results for the same regression when budgetary spending and extra-budgetary spending are adjusted by provincial population and income size, respectively. The negative and significant effect of fiscal decentralization on growth is observed across all these different decentralization indicators. If the ratio of total provincial spending to total central spending rises by I percent, provincial economic growth will fall by 3.12 percent; if the ra- tio of provincial budgetary spending to central budgetary spending rises by 1 percent, the provincial growth rate will be lowered by 2.15 percent; for extra-budgetary spending, a rise in the relative provincial share by 1 percent will reduce the growth rate by 0.96 percent. To introduce the second structural change with respect to the base case, we estimate the base equation and the augmented base equation ( including all the M- and N-variables) with average provincial data over the period 1986- 1992. Estimation of the base equation based on provincial average values is shown in Column (8): the growth impact of decentralization is still negative, but not significant. Estimation of the augmented base equation is reported in Column (9), labeled as the non-base equation, which presents a negative and significant coefficient for fiscal decentralization. In this column, we also note that the inflation rate has a positive and nonsignificant effect on growth, the effect of provincial openness is positive and statistically very significant, and 21 the rate of investment has the conventional sign (positive), and is significant. 4.2.3 Alternative specifications and their sensitivities Following the results of negative and significant signs of coefficients on vari- ous indicators of fiscal decentralization, we now further pursue the robustness of the results. To do so, sensitivity analyses are conducted with respect to each selection of four additional decentralization indicators introduced ear- lier. The robustness is then further examined with random-effect estimations, as shown in Table 5. Similar to Table 3, sensitivity tests are conducted against all the decen- tralization indicators where the M-variables are labor (L) and the tax rate (T) and the N-variables include openness (F), inflation (R), and investment (I). Eight regressions are estimated along with different selections of the three N-variables. The results are quite robust: the negative association between decentralization and growth prevails among the 32 estimations, although the result of negative significance for DCb62 appears relatively weak in two esti- mations at the 5 percent significance level. Moreover, the coefficients on the measure of openness and on the investment rate are positive and significant in most of those estimations, some of which are included in Table 6. Table 7 shows the results of random-effect estimates with the generalized least square (GLS) regression, which follows the specification used in the base case. Negative and significant coefficients on fiscal decentralization are found for three indicators as shown in Column (1) - (3). Column (4) gives a negative sign for DCcbe2, but it is not significant. The coefficient for DCbe2, the 22 indicator with respect to budgetary funds adjusted by provincial population, turns out to be positive but insignificant, as shown in Column (5). However, overall results from the random-effect estimations show their consistency with those in our previous (fixed-effect) estimations. 5 Conclusions The negative effect of fiscal decentralization on provincial economic growth has been found to be consistently significant and robust in China. It sug- gests that fiscal reforms began in China since the early 1980s seem to have failed to promote the country's economic growth. This finding is surprising in light of the conventional wisdom that fiscal decentralization usually makes a positive contribution to provincial or local economic growth. The result seems to suggest that provincial spending has failed to deliver fast economic growth as widely expected. Perhaps this is understandable because, in the current stage of the Chinese economic development, the central government is constantly constrained by the limited resources for public investment in national priorities and nation-wide externalities such as highways, railways, power stations, telecommunications, and energy. These key projects may have far more significant impact on growth across provinces than their coun- terparts in each province. This finding also has some implications for those developing countries and transitional economies pursuing fiscal decentralization. The merits of fiscal decentralization have to be measured relative to existing revenue and expenditure assignments and to the stage of economic development. In the 23 early stage of economic development, the central government may be in a much better position to undertake public investment with nation-wide exter- nalities. More importantly, if local shares in total fiscal revenue and expendi- ture are already high, further decentralization may result in slower economic growth. In this connection, the dangers of decentralization put forward by Prud'homme (1995) seem to be empirically relevant. Our empirical assessment of China is still quite preliminary. For future research, we need to consider the composition of central and local govern- ment spending and identify the contribution of public spending to provincial growth by both functional forms and levels of government when the data be- come available. Furthermore, the role of intergovernmental transfers should be considered explicitly in the process of fiscal decentralization. While some analytical work has been done in this respect (Zou 1994, 1995), the collection of provincial data on both conditional and non-conditional grants from the center to localities in China should be a priority. Finally, instead of measur- ing distortions by a simple output tax, we should pay more attention to the revenue side of fiscal decentralization and quantify the impacts of local taxes and central taxes on economic growth. 24 Data Appendix Our empirical estimations are based on annual data for 28 provinces. Data sources are all official publications in China. Although over 100 volumes of statistical publications are involved, major data sources include China Statistical Yearbook and provincial statistical yearbooks for various years."1 Variables used for estimations are listed below with their data sources. Names of provincial areas included in our estimations are also listed. Y = real growth rate of provincial income, measured by the percentage rate Derived from index of provincial income measured at constant price level (Sources: for 1980-1985: China National Income Statistics 1949-1985 (Guomin Shouru Tongji Ziliao Huibian 1949-1985); for 1985-1992: China Statistical Yearbook (Zhongguo Tongji Nianjian), various issues L= growth rate of the provincial labor force, measured by the percentage change in the total number of the labor force Derived from the total number of the labor force in the whole society (Sources: for 1980-1985, various volumes of provincial statistical yearbooks; for 1986-1992: China Statistical Yearbook (Zhongguo Tongji Nianjian), var- ious issues I = provincial investment rate, measured by the rate of accumulation in fixed asset and circulating funds "Provincial statistical yearbooks cover 28 provinces for various years up to 1994. 25 Derived from index of total accumulations in provincial income mea- sured at constant price level (Sources: for 1980-1985: China National Income Statistics 1949-1985 (Guomin Shouru Tongji Ziliao Huibian 1949-1985); for 1985-1992: China Statistical Yearbook (Zhongguo Tongji Nianjian), various issues) F = degree of openness of provincial economy, measured by the share of total volume of foreign trade (exports and imports) in provincial income Derived from the total volume of provincial exports and imports (divided by provincial income) (Sources: Almanac of China's Foreign Economic Re- lations and Trade (Zhongguo Duiwai Jingji Maoyi Nianjian), various issues in 1984-1994/95.) T = degree of distortion in provincial economy, measured by the ratio of provincial revenue collection in provincial income Derived from provincial budgetary revenue collection (divided by provin- cial income) (Sources: various volumes of provincial statistical yearbooks) R = inflation rate, measured by the overall social retail price index in each province (Sources: China Statistical Yearbook (Zhongguo Tongji Nianjian), various issues) DCcsel = decentralization measured by the ratio of total provincial spend- ing to total central spending Total (central or provincial ) spending is the sum of budgetary spending and extra-budgetary spending (Sources: see sources of budgetary spending and of extra-budgetary spending) DCrse2 = decentralization measured by the ratio of per-capita provincial spending to per-capita central spending 26 Per-capita (central or provincial) spending is derived from spending di- vided by population. (Sources: for provincial population: various volumes of provincial statistical yearbooks; for the central government, national pop- ulation is used, China Statistical Yearbook (Zhongguo Tongji Nianjian), various issues) DCbeI = decentralization measured by the ratio of provincial budgetary spending to central budgetary spending (Sources: for province: various vol- umes of provincial statistical yearbooks; for the central government: China Statistical Yearbook (Zhongguo Tongji Nianjian), various issues) DCbe2 = decentralization measured by the ratio of per-capita provincial budgetary spending to per-capita central budgetary spending (Sources: see sources of budgetary spending and of population) DCebel= decentralization measured by the ratio of provincial extra-budgetary spending to central extra-budgetary spending (Sources: China Government Finance Statistics (Zhongguo Caizheng Tongji, 1950-1991), Yearbook of China Government Finance (Zhongguo Caizheng Nianjian, 1993) DCebe2 = decentralization measured by the ratio of [provincial extra- budgetary spending/provincial income] to [central extra-budgetary spend- ing/national income] (Sources: see sources of provincial income and of extra- budgetary spending) List of provincial areas: Beijing, Tianjin, Hebei, Shanxi, Neimenggu (Inner Mongolia), Liaoning, Jilin, Heilongjiang, Shanghai, Jiangsu, Zhejiang, Anhui, Fujian, Jiangxi, Shandong, Henan, Hubei, Hunan, Guangdong, Guangzi, Sichuan, Guizhou, Yunnan, Shaanxi, Gansu, Qinghai, Ningzia, and Xinjiang. 27 Figurel: Government Spending Share 1978-1992 45.00% 40.00% _ 35.00%-------_...__..__.. ........____..._.. _ 0 a30.00%- 20.00% - - .-.--. . -. ...... . . . . _. ..... .... 15.00%.- .------------------------------------ . ........ ......... .............. 10.00% I I 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 Year -8-- Consolidated Spending -- Budgetary Spending -O- Extra-budgetary Spending 28 Figure 2: Government Revenue Share 1978-1992 45.00%- 2 3 5. 0 0 % _ . _ _ _ __ _ _ .. ... _ .- --- - - - - - -.-.- -.-- - - - . .. -.-. .. -..-.- . ...... .. .............. ..... .. - ------ .=_ Ct30.00%t -- =--- - ----- .--. --. -- --- ---. W 305.00% .- . ............. 20.00% _ ... -. . . -_ _ ._ _ .... ... ...... ... . ...... . .... _ . ........ ........... ... ... . . . _ 15.00% I I 1 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 Year -i+ Consolidated Revenue -F- Budgetary Revenue -_- Extra-budgetary Revenue 29 Figure 3: Central Government Spending 1978-1992 Consolidated Spending Budgetary Spending 3Xt-l- -l----- t - -l Is l - Iwo IoIo q l- l -19 I 1 M M m I 1001 - --- Til| - I 1101 - - i - 737 179" 111 oW 111 11 111 Ill 1_ 1W- 11 111 111 101 1062 117 1171 114 711 1062 111 111 11_1 1_ 1111 11_I 131 1111 1332 Yer Year Ezdrt-budgetary Spending - - -- -- -- - -- - - -- - - - -.- 10.00% -.o4- ................. .... ..... .. ..| ....... .., l ............. ...,Z , 1070 1171 10 1161 ¶1 12 lo low low 117 loa i 110 1.111 1fa 2 Year 30 I£ J10A,~~~~~~~ WOI - 0 w uI - -. - .... . .- - - - - - -~~~~~~~~~~~~~~~~~~~~........... . - - O00 *nu*n. k6 ApnqUmU3 ML ml Am MLz "el _[i azL cu1_ t i';e ML ua uI. OW O ML IU i"; l ML _lg aMt "t "a ML ML % . WuO 11-1--1--l-l-~~~~~~~ ................ .-- ...... _%OO'O-1- to' '- 1 -'-'-''----'--- -0 ..... . .._,__,. .......... _% ___OW ......... . . ....... ............_ ..._. ..___ soo Z66 L-8L6t anUGA9N JUGWUJGAOD IDJIUG D :t sJn6Bj Figure 5: Fiscal Decentralization (Selected Provinces: 1978-1992) 0.14 - 0. 13 - 0.12 -_ - - ....-.-...-.--._ _. . . -..._..- - O0 . 1 1 - __ .,,, _ ___.---.-...-.. - - ------- - --..- .. ......-... 0LO.7 -_.;; ........._._ ... .... . . . . _ _ . ._ ... .z. ..... ...... ... ............ . _.... _. .. ... _.........._ v0.0- ............. ......... CD 0.0 O.07 -... . .. . ............ ... .. .... _ ..... .. ?%0. 06 - ........... -------....... 0 0 0 .. . . ..... .... .. ........ . .. ............... . .... ......._ tm ~004 .............-....................................................... ........ ...........-........... ...... . . .... ...... 0.03 ---------. ................................................... . ...... . .. .................................-. .........- --....... ... 0.02 . ....-...-.-.... .... - ..... .... . ........... ............ _ ........ . 0.01 .--... 78 80 81 82 83 84 85 86 87 88 89 90 91 92 Year |+-- Guangdong -e- Henan -i- Qinghai -0- Sichuan | 32 Table 1: Size of Provincial Budget and Income Size of Budgetary Spending I[ Per capita Income Provincial Areas Average (1980-1992) Initial Level (1980) Current Level (1992) Average (1980-1992) Initial Level (1980) Current Level (1992) Three Metropolitan Cities 14.57% 12.01% 1515% 2823% 16.51% 42.16% Beijing 17.06% 13.47% 16.72% 24 67% 12.46% 39.12% rianjin 17.09% 15.76% 18 37% 21 89% 12.43% 33.23% Shanghai 10.23% 6.79% 12.35% 38.14% 24.83% 54.14% Costal Areas 11.71% 1317% 11.67% 10.60% 3.82% 18 32% Jiangsu 8.96% 10.69% 8.72% 1170% 4.56% 18 53% Zhejiang 10.84% 1095% 9.84% 11.72% 414% 20.09% Fujian 17.25% 20.65% 16.99% 8.49% 2.89% 1513% Shandong 10.24% 11.93% 9.60% 930% 3.46% 16.13% Guangdong 13.37% 1162% 13.22% 13.09% 4.03% 21.74% Inland Areas 15.04% 16.47% 15.66% 7.98% 3.72% 12.54% Hebei 13.01% 12.67% 1150% 7.85% 364% 12 74% Shanxi 20.34% 2309% 18.17% 738% 3.43% 11.35% Liaoning 13.76% 13.66% 17.42% 1429% 7.16% 2208% Jilin 20.18% 20.38% 22.14% 9.34% 3.85% 14 53% Heilongjiang 15.93% 13.74% 17 82% 1110% 585% 17.59% Anhui 12.59% 1303% 1575% 626% 2.61% 8.82% Jiangxi 15.93% 16.18% 1545% 620% 303% 1011% Henan 13.71% 16 42% 10 93% 5 95% 2 66% 9 54% Hubei 13.45% 15 26% 14.09% 8 48% 3.71% 12 97% Hunan 14.25% 17 88% 12.36% 6.69% 316% 10 56% Sichuan 13.55% 12.63% 15 28% 600% 2.70% 9.78% Shannxi 19.51% 22.74% 18.96% 6.20% 2.84% 10.38% Minority Areas 26.13% 28.16% 24.58% 6.52% 287% 1095% Neimeng 29.65% 34.83% 25 30% 7 20% 2.81% 12 06% Guangxi 13.93% 15.01% 14 70% 5 33% 2.32% 8 86% Guizhou 14.52% 12.81% 19.05% 4.51% 1.90% 7 36% Yuannan 25.92% 23.12% 29.03% 5.70% 2 38% 10 09% Gansu 22.70% 20.27% 22 72% 6 25% 316% 10 00% Oinghai 40.46% 42.28% 3448% 7.41% 389% 11.63% Ningxia 40.52% 48.80% 30.58% 6.77% 3.15% 11 07% Xinpang 25.74% 31.33% 20.81% 8 96% 3.59% 16 52% Minimum 8.96% 6.79% 8.72% 4.51% 1.90% 7.36% Maxmum 40.52% 48.80% 34 48% 38.14% 24.83% 54 14% Standard Deviation 0.08 0.09 0.06 007 0.04 010 Note: (1) Measured by the ratio of provincial budgetray spending to provincial income (2) Intial year forXinjiang is 1985 Sources: See Data Appendix 33 Table 2 (a) Descriptive Statistics of Provincial Decentralization in Budgetary Finance (1) Provincial Areas Average Minimum Maximum Standard Deviation Initial Level Current Level Growth Rate | _____________________________ (1978-1992) (1978-1992) (1978.1992) (1978.1992) (1978) (1992) (19781992) Three Metropolitan Cities 3 94% 2.50% 5.19% 0.0078 3 90% 3.80% -2.57% Beijing 3.93% 2.29% 538% 0.0095 3.91% 3.83% -201% Tianjin 3.01% 2.25% 3.72% 0.0045 2.79% 2.49% -10 89% Shanghai 4.86% 2 95% 6 63% 0 0116 4 99% 5.07% 1.64% Costal Areas 6.21% 3.76% 896% 0.0162 5.29% 7.30% 57.15% Jiangsu 6.05% 3.95% 8.13% 0.0127 5.45% 6.72% 23.43% Liaoning 7.16% 4.43% 10.34% 00186 6.01% 7.93% 3204% Zhejiang 4.53% 2.66% 6.77% 0.0124 3.35% 5.09% 52.19% Fujian M 3.79% 2.31% 5 47% 0.0106 2.31% 4.51% 95.17% Shandong 6.83% 4.24% 10.29% 0.0183 612% 7.78% 27.10% Guangdong 8.91% 4.16% 12.77% 0.0292 5.51% 11.73% 11297% Inland Areas 4.96% 3.62% 8.50% 0.0092 4.91% 4.91% 2.06% Hebei 5.60% 3.67% 6.99% 0.0088 6.23% - 5.41% -13.18% Shanixi 3.87% 290% 4.60% 0.0046 4.05% 3.43% -1525% Jilin 4.19% 2.64% 6.07% 00119 3.14% 4.27% 3621% Heilongjiang 5.80% 3.98% 7.72% 0.0112 8.05% 5.47% -s.59% Anhui 3.90% 2.58% 5.26% 0.0082 3.48% 396% 13.73% Jiangai 3.36% 2.33% 4.41% 0.0062 3.12% 3.s5% 18.98% Henan 5.95% 4.90% 7.33% 0.0058 6.47% 5.56% -14 16% Hubei 541% 3.92% 7.15% 0.01 5.75% 5.29% -8.12% Hunan 5 09% 4.45% 8.23% 0.0044 5.37% 4.96% -7.71% Sichuan 7.92% 4.96% 11.51% 0.0211 6.86% 8.54% 24.55% Shaanxi 3.48% 2.72% 41.0% 0.0053 3.52% 3.49% -085% MinorityAreas 2.77% 1.72% 3.87% 0.006 2.32% 2.98% 20.72% Innr Mongolia 4.02% 2.72% 5.08% 0.0o67 3.59% 385% 7.33% Guangoi 2.73% 1.82% 3.75% 0.0088 2.75% 3.27% 18.91% Guizhou 1.92% 1.03% 3.01% 0.0086 1.20% 2.53% 110.35% Yunnan 4.89% 2.81% 7.41% 0.016 3.51% 6.50% 85.08% Gansu 2.84% 1.88% 3.73% o.o055 2.75% 2.86% 3.77% ONhai 1.18% 0.90% 1.42% 0.0015 1.31% 0.99% -23.80% Ngna 1.08% 0.77% 1.28% 0.0015 1.11% 0.85% -23.55% Xinjang (2) 3.47% 3.00% 3.80% 0.0024 3 42% 3.00% -12.36% Minimum 1 06% 0.0015 1.11% 0.85% -23.80% Mawimum 8.91% 0.0292 8.86% 11.73% 112.97% Standard Deviation 0.0181 0.0162 0.0224 0.1950 Note: (1) Fiscal Decnalizaton is measured by t ratio princda budgetay spendibg cental budgetray spending (2) Iniial yea are 190 and 1985 fo Fujian and for Xiny respectey. Source: see DC Append Table 2 (b) Descriptive Statistics of Provincial Decentralization in Consolidated Finance (1) Provincial Areas Average Minimum Maximum Standard Deviation Initial Level Current Level Growth Rate (1986-1992) (1986-1992) (1986-1992) (19B661992) (196) (1992) (1966-1992) Three Metropolitan Cities 563% 4.27% 6.23% 0.0082 5.92% 4.27% -27.78% Betang 5 81% 5.13% 6.22% 0 0035 5.87% 5.13% .9.56% rin 3.56% 269M 4.18% 0.0048 4.18% 2.09% .35.71% Shanghai 7.54% 5.00% 8.63% 0.0112 7.90% 5.00% -36.s7% Costal Areas 7.92% 7.02% 88.9% 0.0057 7.02% 7.43% 5.82% Jienpu 8.45% 7.90% 9.35% 0.0048 8.06% 7.90% -1.91% Lsa nmg 10.28% 856% 11.59% 0.0109 9.00% 8.56% -4.93% Zhua 8.11% 5.45% 8.82% 00D47 5.45% 5.89% 8.12% Ft#on 418% 3.55% 4.77% 0.0044 3.55% 4.0s% 15.10% Sh -aon 8863% 7.41% 10.12% 0o.09 7.41% 7.8% 6.12% Guadong 9s91% &8% 1087% 0.0o 86.6% 1028% 18.71% Irland Areas 5.49% 4.78% 6.09%6 0.0043 5.37% 4.78% .10.ss% Heb~ 6.35% 5.s3% 7.23% 0.0058 5.85% 5.75% -1.85% Shad 4.16% 3.54% 4.63% 0.0032 4.20% 3.54% .17.28% Ain 5.27% 4.30% 580% 0.0047 5.26% 4.30% -18.21% Hh N9 8.30% 5.15% 8.96% 0.0052 6.52% 5.15% -20.99% A9di 4.40% s87% 4.72% 0.0026 4.52% 3.87% .14.42% tJI J g 36% l39% .95% 0.0017 3.59% 3.39% -5.73% Huian 5.75% 5.26% 6.36% 0.0037 5.26% 5.29% 0.18% Fl 5.91% 4.s9 1.47% 0.0044 5.95% 4.99% .16.13% aFo- 5.39% 4.89% 5.91% 0.0029 5.16% 4.9% .5.62% Si". 9.48% 6.17% 10.79% 0.0079 9.16% &17% .10.82% a 369% 3.S1% 4.15% 0.0027 3.45% 331% -4.19% Miwrt Ares 2.79s 25ss% 3065% 0S0 2.61% 2.61% -0.07 h.p &.ola 4.00% 3.51% 4.36% o0025 4.01% 351% .12.53% Guine 3.50% 2Z02% 3S.% oCoSS 292% 3.72% 2.30% Gouahou 2.14% 1.61% 2ss% 0Q002 1.81% 2.06% 1527% YU, s5.19% 4.34% 6.9% 0.0056 4.34% 5.16% 1&04% G . w 2.80% 243% .03% 0O.18 2.84% 2.43% .14.32% af% w 1.02% 0.61% 1.12% 0.001 1.07% 0.81% -24.22% Nb,da 0.91% 0.72% 0.99% 00o0 0.99% 0.72% -27.93s Xk;eg 2.74% 2.42% 289% 0.0015 269% 2.42% -16.13% M inum 0.91% 0.000D9 0.Q9C 0.72% -36.67% MmndmLun 10.26% 0.0112 9.16% 10.26% 27.30% Standard Devtiaon 0.0242 0.021 0.022S 0.1554 Note: (1) Fa Oen- Iaon is n,umed bhy tm fto d piw*aM crmaids sd spending b c C011c10d een rs Source: see Dab Appenwd rable 2 (c) Descriptive Statistics of Provincial Decentralization in Consolidated Finance (1) Provincial Areas Average Minimum Maximum Standard Deviation Initial Level Current Level Growth Rate (198-1992) (1980-1992) (1980-1992) (1960.1992) (1980 (1992) (1980-1992) Three Metropoltan Cites 420.83% 28.52% .549.88% 0839 268.52% 407.41% 51.73% Beillg 438.18% 254.e9% 594.24% 1.088 254.80% 429.79% 55.75% rtrion 3s8.84% 297.11% 487.71% 0.t05 297.11% 331.31% 11.51% Shanghai 427.4u% 253.76% 555.68% 1.094 253.70% 48115% 81.73% Costal Areas 140.43% 82.98% 202.70% 0.3614 85.87% 163.25% 90.12% Jianpu 103.64% e5.7% 140.22% 0.232 73.95% 11399% 54.14% U oning 209.58% 115.52% - 300.72% 0.5741 146.34% 234.s5% 58.32% Zhejian 12.85% 86.74% 181.10% 0.3479 e8.74% 130.20% 102.51% Fujian 14U.95% 9068% 213.47% 0.4029 90.e8% 172.41% 90.21% Shianrdng 94.20% 57.38% 141WG% 0.2611 62.52% 106.29% 70.eo% Guangdong 115.54% 70.90% 238.92% 0.5309 70.9s% 212.69% 199.72% Inland Areas 120.01% 85.s8% 158.09% 0.244e s6.8s% 118.42% 37.91% Hebei 108.03% 70.04% 134.04% 0.169 70.04% 100.94% 44.11% Shanxi 152.55% 115.69% 181.77% 0.1854 12D.11% 135.07% 12.45% Jilin 198.27% 118.21% 285.83% 0.58 11884% 202.46% 70.36% Heilongliang 18e.17% 121.96% 252.87% 0.4117 121.96% 181.88% 49.13% Anhw 80.21% 51.65% 105.96% 0.1752 51.18% 79.75% 54.38% Jiangxi 101.78% 70.56% 132.63% 0.1891 74.18% 109.39% 47.50% Henan 79.10% Ws.34% 97.24% 0.074 s8.34% 73.49% 10.78% Hubei 114.99% 82.73% 154.35% 0.2279 85.91% 112.38% 30.79% Hunan 95.15% 8.45% 116.77% 0.0844 85.77% 93.55% 9.08% Sichuan 83.49% 50.02% 121.21% 0.2388 51.76% 91.44% 76.84% Shaanxi 122.32% 95.07% 161.95% 0.196 97.95% 122.30% 24.88% Minority Areas 187.39% 122.27% 233.73% 0.3275 126.92 178.02% 3s.e8% Inner Mongolia 213.61% 142.79% 268.19% 0.3697 148.51% 204.45% 37.67% Guangxi 74.44% 50.O0% 101 .7% 0.1721 52.73% 87.88% e6.65% Guizhou 70.57% 38.84% 106.99% 0.2312 36.84% 90.05% 144.45% Yunnan 154.04% 81.13% 228.91% 0.4931 8Z79% 196.84% 139.92 Gansu 147.20% 95.83% 193.83% 0.29se 97.29% 146.30% 50.37% Qinghai 300.67% 236.81% 32s94% a3698 23.81% 252.81% 6.75% Ningia 271.91% 199.84% 316.50% 0.3895 233.49% 205.90% -11.82% xin %.ng ( 266.71% 222.13% 294.64% 0.212 265.91% 222.13% -16.46% Minimum 70.57% 0.074 36.84% 73.49% *1e.4% Maximum 438.18% 1.094 297.11% 481.15% 199.72% Standard Deviation 1.1281 0.7277 0.9279 0.4613 Note: (1) Fiscal Dcentralz ation is measured by the rabo of per capita paAncial budgetary spending t per capita central budgtary spending (2) Initial year for Xinjiang Is 1985 Source: se De" At w Table 3: -SENSITIVITY RESULTS FOR BASIC VARIABLES: Base Case Variable: Real Growth Rate Co, 181 - 1992 Variafe Coefficient Standard Error t R-Sqtr AdW;d R-Sqmu S.E. Of Regrson ther Variables L (Labor) high 0.44 0.312 1.41 0.373 0.247 0.043 I, R base 0.274 0.312 0.878 0.328 0.203 0.044 ks 0.143 0.307 0.466 0.367 0.244 0.043 F T (Tax Re) high -0.115 0.154 -0.749 0.444 0.32 0.041 I. R, F base -0.4069 0.155 -2.628 0.328 0.203 0.044 low -0.4069 0.155 -2.628 0.328 0.203 0.044 DCbS2 high -0.047 0.0139 -3.413 0.367 0.244 0.043 F base -0.05 0.014 -3.541 0.328 0.203 0.044 _________________ k -0.069 0.014 -4.821 0.444 0.328 0.041 I, R. F Notes: 1. Al esavna honav conC red pmWciai Abed ofc*s but Uio risai ae not mppde hew. 2. NuiberofobewvaJs 196. 3. DCbe2= decetdkalon measured by te rabo of per-cph p'r**wd budgutw * e to per-capb VW"b uAdaty tpene. Souas: See Ws deb awxk. wJ Table 4: -Estimates of Structural Changes Dependent Variable: Real Growth Rate Variable (1 )=Base (2) (3) (4) (5) (6) (7) (8)= Base, averap (9)z Non-base, a8erage (19661S2 (I1992) (19651961 (196192) (19661992) (19619W2) (198tl19-2 (1986199 (1961-W2 Corndt 0.104 0.01 (4.13) 0.1945 L (Labor) 0.27 0.06 0.003 0.23 0.31 0.22 0.27 .0.23 053 p77) p.191) (0009 pD79) (0991 ID.77) P0.4 (-.40A (1.30 T (Tax Rab) 4.41 .021 -0.18 4.36 .31 0.36 4.35 .006 -004 (-2.62W (.1.532) (-90m) (-25791 (.1.75) (-240 (2345) (4M52 (4037) DCcsel -3.12 DCcs2 (053 (.2.591 DCbel .215 (4.715) DCebel 4.98 DCbe2 .05 -.01 -.09 40.02 4.003 4.02 (-3.54) (437) (4.4A (-.6) (-.51 (2.85) DCebe2 .0.10 (-4.25) R (Infbtion Rate) 0°S4. (0.324) F (Openness) 1525 (5.21) I (Investrent) 0.14 ____________________________ (~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~2.33) Number of Obervdions e19 308 137 196 196 196 196 26 28 R-Square 0o328 0.134 0.344 0.068 0.62 Adjusted R-Square 0202 004 0.156 0.257 0.175 0.254 0.279 -0.051 0.512 S.E. Refession 0.04 009 004 0.0428 0.0451 042 0.0421 0.022 0.015 Notes: 1. Estmnaton equations (1) - (7) have considered pmvincdal Axed efflcts, but the resu*s are not reported her. 2. Vahues of t-stadtiscs appear in pamntheses. 3. DCcse 1= decentrakation measured by the ratio of total provincial spendng to total central spending. 4. DCcse2= decentrazabon measured by the ratio of per-capita proincdal spendng to per-capNta contra spendhg. 5. DCbM = decentralzation measured by the ratio of promndal budgetary spendng to central budgetary spendng. 6. DCebet= decentralization measured by the ra6io otprowindal extra-budgetary spendng to cenfra exta-budgetary spendog 7 DCbe2= decentiraztion measured by the ra6o of per-capita pmrvndal budgetary spendng to per-capite central budgetary spendig 8 DCebe2= decentrazation measured by the raio of [provinh l exina-budgetary spending#omidalhwome to [centa oexra-budgetary spendhat nalh come). Sources: Tabe 6: -SENSITIVITY RESULTS FOR FISCAL DECENTRALIZATION INDICATORS Variabe: Real Growth Rate m. 198 - 1m Variable Coemci i Standard Effor I R wa A4doid R rquwe S.E. Rei onm OiwiorVi L (Labor) high 0 409 0 29 0137 0 415 0.299 0.041 . R base 0235 0o299 076 0373 0257 0043 blw 0112 0.26 0369 0.425 0.31i 0.041 F T(Tax Rate) high -0051 0.142 -0361 0.513 0.411 0.0Q8 l,R,F bow -0 3S1 0.14 -0.2579 0.373 0.257 0.043 ow -00361 0146 -02579 0.373 0.257 0.043 DCcsl hih 3.124 0.605 s.163 0.373 0.267 0.043 be - 3124 o.e05 4.163 0.373 0.257 0.043 b.v -4039 0 569 4657 0.464 0.361 O.9 R, F L (Labt) high Q4s 0.31S 1342 0.51 022 0O0436 R base O.314 0.317 0.989 0304 0175 0.0451 km 0.106 0.313 0.e 0.33 0Q211 0.0441 F T (Tax Rate) high - O6 012 -0473 0Q364 0.230 0.043 R, F be" .0313 0.176 .1759 0.304 0.175 0.0451 iw4 S 0313 0.176 -1.759 0304 0.175 0.0451 DCc*.2 high o.304 0.142 -2.134 0.33s 0.211 0.0441 F base -0.372 0.144 -2.50 0.304 0.175 0.0451 l -0S51 0.15 4.371 0.351 0.22 0.0436 1 R L (Labor) high 0.404 0.3 1.347 0.418 0.20a 0.0415 1, R bsa 0.216 0.3 0727 0.375 0.254 0.0427 low 0.061 0.29 0.261 0.434 0321 0.040 F T (Ta Rle) hih o.026 0.142 -0190 0.531 0.429 0.0374 l, F bao -.0377 0.1514 -2.466 0.375 0.254 0.0427 kaw -0377 0.1514 -24a6 0.375 0254 0.0427 DCbe1 high .21i3 0.564 4791 0406 0.2 0.0417 1 be -2147 0.576 4.715 0.375 0.254 0.0427 low .3.42 0.56e -6.99 0.507 0.403 0.0382 R. F DCabel hi9h 40.775 0.463 -1604 05057 0Q403 0.02 R. F ba -0.96 0 539 -1.7a 0.375 0.254 0.0427 ki -1.065 0.527 -2022 0.406 0Q2 0.0417 1 L (Labor) hih 0.402 0.303 0.133 041 0291 0.0418 R be" 0 265 0.297 0.694 0.3 0.279 0.042 m 014S 0.202 049 04 0.313 0.041 F T (Ta Rae) high -0.0724 0.145 -0406 0Q507 0.4 0.0365 , R. F baef -0.347 0.146 -2346 0.36 0.279 0.042 low -0347 0145 -2.346 0.396 0.279 0.0422 DCbe2 high 40021 0015 *145 0429 0313 0.041 F bas -0023 0.015 -1.516 0396 0.279 0.0422 l 0.0419 0.014a -2.36 0.507 0.4 0.03 i, R. F DCebe2 high -0t09 00211 -4206 0.47 0.3a 0.03o7 R,F bae -0005 0.0224 -425 0.396 0.279 0.0422 kw -9 0.0217 4525 0.439 0.326 0.0406 Akfbs Formcpbnoms of vwmbbds mid dab bouafs.am Teh 2 Table 6: -Estimates with Openess and investment Dependent Variable: Real Growth Rate Variable (1) (2) (3) (4) (1986-1992) (198-1992) (19-1992) (1986-199) L (Labor) 0.39 0.39 0.37 0.38 (1.41) (1.28) (1.38) (1.35) T (Tax Rate) -0.05 -003 -0.03 -0.07 (-0.36) (-0.20) (-0.20) (4.498) DCcsel -4.01 (-8.99) DCcse2 -0.49 (4.37) DCbel -3.32 (-5.99 DCebei -0.88 (-1.83) DCbe2 -0.04 (-2.84) DCebe2 -0.09 (.4.47) R (Inflation Rate) 0.18 0.11 0.21 0.15 (3.59) (2.04) (4.19) (2.98) F (Openness) 0.29 0.23 0.32 0.24 (5.62) (3.79) (8.14) (4.45) I (Investment) 0.16 0.20 0.15 0.18 (3.06) (3.30) (2.8M (3.41) Number of Obervations 198 196 198 198 Adjusted R-Square 0.411 0.281 0.429 0.400 S.E. Regression 0.038 0.042 0.037 0.038 Notes: 1. Values oft-statistics appear in parentheses. 2. DCcsel= decentralization measured by the ratio of total pro vncmiaspendNng to total centralsp.nding. 3. DCcse2= decentralization measured by the ratio of per-capita provincial spending to per-capita central spencding 4. DCbel= decentralization mneasuredby the ratdo of proncim calbudgetary spending to central budgetary spending. 5. DCebe 1 = decentralization measured by the ratio of provincial extra-budgetary spending to central extra-budgeter 6. DCbe2= decentralization measured by the ratio of per-capita provnctal budgetary spending to per-capita central 9. DCebe2= decentralization measured by the ratio of [ provincial extra-budgetary spendin gprovincial income) to (central extra-budgetary spending/national income]. Sources: See the Data Appendix. 40 Table 7: -Estimates with Random Effects (GLS) Dependent Variable: Real Growth Rate Variable (1) (2) (3) (4) (5) _______________________ (19W1992) (198-1962) (1988-1962) (1986-1992) (1986-1992) Constant o.168 0.229 0.233 0144 (8.09) (6.769) (6.885) (8 057) L (Labor) 0.17 0.20 018 0.22 0.20 (0.581) (0.885) (0.8349) (0.725) (705) T (Tax Rate) -0.38 -0.41 -0.43 -0 36 -0 32 (-2.6) (-3.09) (--3.176) (-2 26) (-2 36) DCcsel -1.78 (-3.87) DCcse2 -0.08 (-0.9se) DCbel -0.02 -1.54 (-1.79) (-3.009) DCebel -0.24 (-0.53) DCbe2 0.01 (0.4986) DCebe2 -0.09 DCebe2 (-4 40) Number of Obervations 196 19s 196 198 196 S.E. Regression 0.043 0.041 0.043 0.041 0041 Notes: 1. Values of t-statstcs appear in parentheses. 2. DCcset= decentralIzatIon measured by the raeo of total provindal spend/ng to total central spending. 3. DCcse2= decentralization measured by the rato of per-capita proWincial spending to per-capita central spending. 4. DCbel= decentralzaton measured by the rato of provincial budgetaly spending to central budgetary spending. 5. DCebe 1 = decentralizaton measured by the rato of provincial extra-budgetary spending to central extra-budgetary spending. 6. DCbe2= decentralizaton measured by the rato of per-capita provincial budgetary spending to per-capita central budgetary spending. 9. DCebe2= decentralization measured by the reao of [provincial extra-budgetary spending/provincial income] to [central extra-budgetary spending/national income). Sources: See the Data Appendix. 41 References [1] Bahl, Roy W. and Johannes Linn (1992) Urban Public Finance in De- veloping Countries. New York: Oxford University Press. [2] Bahl, Roy W. and Christine Wallich (1992) "Intergovernmental Fis- cal Relations in China." Working Papers, Policy Research Department, World Bank, Washington D.C.. [3] Barro, Robert, and Sala-i-Martin (1992) "Public Finance in Models of Economic Growth" Review of Economic Studies, 59: 645-661. (4] Barro, Robert (1990) "Government Spending in a Simple Model of En- dogenous Growth." Journal of Political Economy, 98: S103-S125. [5] Bird, Richard M., and Christine Wallich (1993) "Fiscal Decentraliza- tion and Intergovernmental Relations in Transition Economies: Towards a Systematic Framework of Analysis" Country Economics Department Working Paper, World Bank, Washington D.C.. [6] Bird, Richard M. (1993) "Threading the Fiscal Labyrinth: Some Issues in Fiscal Decentralization," National Tax Journal, XLVI, no. 2: 207-227. (7] Davoodi, H., Danyang Xie, and Heng-fu Zou (1995) "Fiscal Decentral- ization and Economic Growth in the United States." Policy Research Department, World Bank, Washington, D.C. 42 [8] Devarajan, Shantayanan, Vinaya Swaroop, and Heng-Fu Zou (1996) "The Composition of Public Spending and Economic Growth." Jour- nal of Monetary Economics, April. [9] Feder, Gershon (1983) "On Exports and Economic Growth." Journal of Development Economics 12: 59-73. [10] Gramlich, Edward, and H. Galper (1973) "State and Local Fiscal Behav- ior and Federal Grant Policy, Brookings Papers on Economic Activity 1: 15-65. [11] Gramlich, Edward M. (1993) "A Policy maker's Guide to Fiscal Decen- tralization" National Tax Journal, Vol XLVI, No. 2: 229-235. [12] Jia, Hao and Zhimin Lin, (Eds.) (1994) " Changing Central-local Rela- tions in China." Boulder: Westview Press. [13] Levine, Ross, and David Renelt (1992) "A Sensitivity Analysis of Cross- Country Growth Regressions." American Economic Review, 82, No.4: 942-963. [14] Oates, Wallace E. (1972) Fiscal Federalism. New York: Harcourt Brace Jvanovich. [15] Oates, Wallace E. (1993) "Fiscal Decentralization and Economic Devel- opment," National Tax Journal, XLVI, no. 2: 237-243. [16] Peltzman, Sam (1980) "The Growth of Government" The Journal of Law and Economics, 23: 209-287. 43 [17] Prud'homme, Remy (1995) " The Dangers of Decentralization." The World Bank Research Observer, Washington D.C.. [18] Romer, Paul (1986) "Increasing Returns and Long-Run Growth." Jour- nal of Political Economy, 94: 1002-1037. [19] Rosen, Harvey (Ed.) 1988. Fiscal federalism: Quantitative studies. The University of Chicago Press. [20] Shirk, Susan L (1993) "The Political Logic of Economic Reform in China." University of California Press. [21] Tiebout, Charles (1956) "A Pure Theory of Local Public Expenditure." Journal of Political Economy, 64: 416-424. [22] Wong, Christine P.W., Christopher Heady, and Wing Thye Woo (1993) "Economic Reform and Fiscal Management in China: Report of Technical Assistance Project." Asian Development Bank and Ministry of Finance of China. [23] Wong, Christine P.W.(1991) "Central-local Relations in an Era of Fiscal Decline: The Paradox of Fiscal Decentralization in Post-Mao China." China Quarterly, December. [24] World Bank (1995) "Macroeconomic Stability in a Decentralized Econ- omy." China and Mongolia Department, Washington D.C.. [25] World Bank(1992) "China: Reforming Intergovernmental Fiscal Rela- tions." World Bank Discussion Papers, China and Mongolia Depart- ment, Washington D.C.. 44 [26] World Bank(1990) "Revenue Mobilization and Tax Policy." A World Bank Country Study, Washington D.C.. [27] Zhou, Xiao-chuan and Zhigang Yang (1992) "Fiscal Management and Tax System in China: Problems and Solutions (Zhongguo Caishui Tizhi de Wenti yu Chulu)." Tijin People's Press. [28] Zou, Heng-fu (1994) "Dynamic Effect of Federal Grants on Local Spend- ing." Journal of Urban Economics 36: 98-115, 1994. [29] Zou, Heng-fu (1995) " Taxes, Federal Grants, Local Public Spending and Growth." (forthcoming) Journal of Urban Economics. 45 Policy Research Working Paper Series Contact Title Author Date for paper WPS1585 Public Finances and Economic Luca Barbone March 1996 C. Pelegrin Transition Hana Polackova 85087 WPS1586 Did External Barriers Cause the Azita Amjadi March 1996 S. Lipscomb Marginalization of Sub-Saharan Ulrich Reincke 33718 Africa in World Trade? Alexander Yeats WPS1587 Payments and Finance Problems Constantine April 1996 M. de la Puente in the Commonwealth of Michalopoulos 31206 Independent States WPS1588 Social Insurance in the Transition Deborah Mabbett April 1996 L. Biely to a Market Economy: Theoretical 36280 Issues with Application to Moldova WPS1589 The Analysis of Emeiging Policy Sudarshan Gooptu April 1996 R. Vo Issues in Development Finance 31047 WPS1590 The Latvian Banking Crisis Alex Fleming April 1996 S. Coffey Lessons Learned Samuel Talley 32535 WPS1591 Competition in Network Industries Michael Klein April 1996 S. Vivas 82809 WPS1592 The Stockmarkets as a Source of Cherian Samuel April 1996 C. Samuel Finance: A Comparison of U.S. and 30802 Indian Firms WPS1593 Rural Finance for Growth and Poverty Saeed Qureshi April 1996 C. Anbiah Alleviation Ijaz Nabi 81275 Rashid Faruqee WPS1594 Why Have Some Indian States Done Gaurav Datt April 1996 P. Sader Better Than Others at Reducing Martin Ravallion 33902 Rural Poverty? WPS1595 In Search of Owners: Lessons of Cheryl W. Gray April 1996 M. Geller Experience with Privatization and 31393 Corporate Governance in Transition Economies WPS1596 Agrarian Structure in Poland: The Johan van ZVl April 1996 M. Williams Myth of Large-Farm Superiority Bill R. Miller 87297 Andrew Parker WPS1597 Egypt and the Uruguay Round Bernard Hoekman May 1996 F. Hatab Arvind Subramanian 35835 WPS1598 Logistical Constraints on Francois Amiot May 1996 L. Lewis International Trade in the Maghreb Ovadia Salama 37369 Policy Research Working Paper Series Contact Title Author Date for paper WPS 1599 Economic Implications for Turkey Glenn W. Harrison May 1996 M. Patena of a Customs Union with the Thomas F. Rutherford 39515 European Union David G. Tarr WPS1600 Is Commodity-Dependence Nanae Yabuki May 1996 G. Ilogon Pessimism Justified? Critical Takamasa Akiyama 33732 Factors and Government Policies that Characterize Dynamic Commodity Sectors WPS1601 The Domestic Benefits of Tropical Kenneth M. Chomitz May 1996 PRDEI Forests: A Critical Review Kanta Kumari 80513 Emphasizing Hydrological Functions WPS1602 Program-Based Pollution Control Shakeb Afsah May 1996 A. Williams Management: The Indonesian Benoit Laplante 37176 PROKASIH Program Nabiel Makarim WPS1603 Infrastructure Bottlenecks, Private Alex Anas May 1996 S. Ward Provision, and Industrial Productivity: Kyu Sik Lee 31707 A Study of Indonesian and Thai Michael Murray Cities WPS1604 Costs of Infrastructure Deficiencies Kyu Sik Lee May 1996 S. Ward in Manufacturing in Indonesia, Alex Anas 31707 Nigeria, and Thailand Gi-Taik Oh WPS1605 Why Manufacturing Firms Produce Kyu Sik Lee May 1996 S. Ward Some Electricity Internally Alex Anas 31707 Satyendra Verma Michael Murray WPS1606 The Benefits of Alternative Power Alex Anas May 1996 S. Ward Tariffs for Nigeria and Indonesia Kyu Sik Lee 31707 WPS1607 Population Aging and Pension F. Desmond McCarthy May 1996 M. Divino Systems: Reform Options for China Kangbin Zheng 33739 WPS1608 Fiscal Decentralization, Public Tao Zhang May 1996 C. Bernardo Spending, and Economic Growth Heng-fu Zou 37699 in China