Lee Kuan Yew School of Public Policy. Accepted Paper Series

Transcription

1 Lee Kuan Yew School of Public Policy Accepted Paper Series Intra-Provincial Fiscal Disparities And The Role Of Provincial Fiscal Transfer System In China: A Case Study Of Henan Province Li Hui Lee Kuan Yew School of Public Policy National University of Singapore spplih@nus.edu.sg September 25, 2017 Published Paper No.: LKYSPP Electronic copy available at:

2 ABSTRACT This study seeks to explore intra-provincial fiscal disparities at the county level in Henan province and the extent to which the undesirable fiscal inequalities can be traced to the provincial fiscal transfer system. It tests the equalization effects of provincial-local transfers by decomposing the Gini coefficient by revenue sources. The primary findings are that only general-purpose transfers (including both equalization transfers and other general-purpose transfers) have an equalization effect, while both tax rebates and specific-purpose transfers have anti-equalization effects. In addition, the size of the general transfers equalization effect is too limited to mitigate the overall anti-equalization effect of total transfers. It also found that so-called equalization transfers have not shown stronger equalization effects than other general-purpose transfers; tax rebates anti-equalization effect is limited, but specific-purpose transfers anti-equalization effect is significant due to its rapid growth during the study period of Keywords: intra-provincial fiscal disparity, fiscal equalization, fiscal decentralization, fiscal transfer 1 Electronic copy available at:

3 1. INTRODUCTION While the efforts over the last two decades to reform China s fiscal system through decentralization have driven tremendous economic growth, there are growing concerns about the fiscal decentralization reform s negative effects on fiscal disparities throughout the country. The issue of fiscal disparities, referred to as differences across subnational governments in their ability to raise revenue to meet public expenditure needs (Otter, 2008), has been widely studied, but this strand of literature has mainly addressed fiscal disparities between coastal and inland regions, rural and urban areas, and across provinces. The focus has been the role of central equalization policies in mitigating the growing fiscal disparities resulting from uneven regional economic development. Intra-provincial disparities and the role of provincial equalization policies have not gained much attention in large part due to the lack of detailed data at the subprovincial level in the early years of this research. Yet provincial-level data might disguise intra-province fiscal disparities, and an improvement in a province s overall fiscal capacity as a result of central fiscal equalization efforts does not necessarily translate into reduced intraprovincial fiscal disparities. Based on sub-provincial information and relevant decomposition techniques, a few recent studies have demonstrated that intra-provincial fiscal disparities are actually more profound than inter-provincial fiscal disparities and account for the majority of the overall fiscal disparity in the country, although the relative contributions of the two types of disparities vary in different studies (Liu et al 2014; Wang and Herd, 2013: Zhang, 2013). These studies have also found that the extent of intra-provincial disparities differs across provinces (Wang and Herd, 2013; Zhang, 2013; Zhou and Yan, 2010), which implies that the individual provincial fiscal transfer systems play a varying role in equalization. A few scholarly efforts have been made to examine the effects of the fiscal transfer system in reducing overall national fiscal disparities at the sub-provincial level (primarily the county level), but the results are inconsistent (Liu et al, 2014; Tsui, 2005; Wang and Herd, 2013; Yin, 2008). Those using conventional equalization measurement methods find some equalization effects of the fiscal transfer system (e.g. Liu et al, 2014; Wang and Herd, 2013), while those using the techniques of decomposition by revenue sources reveal opposite results due to the decomposition method s much stricter criterion of equalization effects (e.g. Tsui, 2005; Yin, 2008). Nevertheless, a shortcoming of this body of literature is that the approach of pooling more than 2,000 county-level governments together ignores the variation of provincial fiscal transfer systems and their different effects on intra-provincial fiscal disparities within individual provinces. These aggregate analyses can shed some light on the redistributive outcome of provincial transfer systems at the aggregate national level, but they do not present the fiscal equalization performance of individual provincial fiscal transfer systems. Zhou and Yan (2010) presented the level of intra-provincial fiscal disparities and the contributions by different types of transfers in each individual province. Their findings demonstrate the varying extent of fiscal equalization across provinces and underpin the importance and urgency of single-province study to look into the details of individual provincial fiscal systems and their equalization effects on intra-provincial disparities. This study differs from previous studies as it is the first single-province analysis to investigate the level of intra-provincial fiscal disparities and explicitly explore the role of fiscal equalization policies pursued by a provincial-level government. Through a case study of Henan Province, it examines the extent to which the undesirable fiscal inequalities at the county level 2 Electronic copy available at:

4 can be traced to the provincial fiscal transfer system. The study of the provincial government s role in fiscal equalization policies is vital because in a multi-level government structure, the effectiveness of national equalization efforts greatly depends on the further degree of vertical decentralization and the equalization efforts of intermediate-level governments, particularly the provincial governments (Martinez-Vazquez et al, 2008). The main features of China s current sub-provincial fiscal relations date back to the Tax Sharing System (TSS) Reform of 1994, which determined fiscal relations between the central and provincial governments. This reform gave provincial governments the discretion to stipulate fiscal arrangements between the provincial and local levels. As a result, provinces vary in their assignment of revenue powers and expenditure responsibilities between the provincial and local levels and also in the design of the provincial fiscal transfer system. The provincial fiscal transfer system can potentially align with or nullify the equalization efforts of central transfers, depending on how the fiscal transfer system has been designed in individual provinces. Provincial governments as economic agents have their own preferences, incentives, development strategies and availability of fiscal resources, and their efforts in implementing the national equalization policy might therefore be different (Qiao et al, 2008). The key puzzle to be unraveled is how sub-provincial fiscal relations in individual provinces have shaped within-province fiscal capacity and contributed to intra-provincial fiscal disparities at the local level. This research is designed as a case study of Henan Province to understand the equalization effects of its provincial transfer system. Single-province analysis is much-needed because the provincial transfer system itself and the context in which the provincial transfer system develops varies across provinces. Variations include sub-provincial fiscal arrangements such as revenue power and spending responsibility assignment, the major policy concerns when the provincial transfer system was designed, as well as redistributive results the system is expected to achieve or economic and fiscal conditions in that province. Thus, it is important to investigate the research questions with a contextual perspective. By concentrating on the fiscal equalization efforts in one province, this study affords the opportunity to capture the provincial transfer system in the context that produced and sustained it. The paper is organized as follows: the first section reviews the research literature, followed by an overview of the transfer system in Henan Province and county-level disparities, then an empirical analysis of the equalization effects of different components of the provincial-local transfers. The last section summarizes the findings and concludes the paper. 2. LITERATURE REVIEW Intergovernmental transfers as a policy instrument to address different policy objectives have been well-documented in the public finance literature. Several objectives justify some fiscal transfers to lower levels of government: (i) to deal with the fiscal gap, resulting from insufficient own-source revenues to meet the expenditure responsibilities assigned to subnational governments (vertical imbalance); (ii) to achieve equalization among subnational governments so that they can provide a comparable level of service by levying comparable tax rates (horizontal imbalance); (iii) to correct for inter-jurisdictional spillovers (externalities); and (iv) to achieve national objectives (Shah, 1994). 3

5 Broadly, unconditional transfers (revenue sharing, general-purpose transfers and block grants) are more appropriate for achieving objectives (i) and (ii), while conditional or specificpurpose transfers (earmarked transfers) are better intended for achieving objectives (iii) and (iv). The general purpose of unconditional transfers is either to improve the vertical fiscal balance by providing general-purpose funding to the subnational level or to improve the horizontal fiscal balance by compensating for fiscal disparities across jurisdictions (Otter, 2008). Addressing horizontal fiscal imbalance, i.e. objective (ii), is a key concern for the design of the intergovernmental transfer system. Fiscal equalization transfers are conceptually justified on equity, efficiency and stability grounds, but the primary objective is to ensure horizontal equality among residents across jurisdictions, i.e. ensuring that, subject to local preferences, all persons or firms in a country can obtain comparable public services at comparable tax rates (Blöchliger, and Charbit, 2008). It can be seen as the natural companion to fiscal decentralization to potentially ameliorate horizontal fiscal disparities resulting from regional autonomy (Blöchliger, and Charbit, 2008). Fiscal equalization is above all evaluated on how it reduces fiscal disparities, 1 measured either by fiscal capacity (revenue-raising capacity) or fiscal needs (differentiated expenditure needs). Equalization based on fiscal capacity tends to equalize the difference between standard local revenue and effective local revenue at standardized tax rates, using the so-called Representative Tax Systems (RTS), while equalization based on fiscal need tends to cover the difference between a standardized local need expenditure (measured on the basis of the socalled Representative Expenditure System) and some benchmark (Dafflon and Mischler 2008; Otter, 2008; Shah, 2012). Fiscal need equalization is more comprehensive and accurate but also more complicated to implement because expenditure needs are more difficult to define and derive than fiscal capacity (Shah, 2012). In practice, while some countries only have revenue equalization arrangements, most countries including China equalize both revenue-raising capacities and expenditure needs (Blöchliger, 2014). Nevertheless, implementing such a comprehensive system of equalization transfers presents a difficult challenge, especially given the difficulty of objectively measuring expenditure needs. In practice, comprehensive equalization has failed to achieve the intended fiscal equalization effects in many countries (Shah, 2012). In some cases, these equalization systems can even exacerbate fiscal disparities, if the arrangements take into account fiscal effort by giving the equalization formula a matching character, or if the equalization formulae leave out substantial revenue sources in their fiscal capacity definition, or if the true fiscal needs are not assessed adequately (Blöchliger and Charbit, 2008). Furthermore, some countries clearly separate equalization from other unconditional transfers while others use compound arrangements where funding and equalization are combined (Blöchliger and Charbit, 2008). The latter arrangement might either compromise the intended equalization effects or render difficult the evaluation of equalization. The equalization effect of China s intergovernmental system has received significant attention both from scholars in China and those in international organizations, such as the 1. Fiscal equalization policies are specifically devoted to promote horizontal equity but they also have impacts on efficiency and stability, for instance, efficiency implications in terms of the level of tax rates and public expenditure. There is a large body of literature in this respect, for example, Dahlby (2011), Kelders and Koethenbuergher (2010), Kotsogiannis (2010), Petretto (2013), Smart (1998), and Smart (2007). This body of literature is not discussed here because the scope of this paper deals only with the equity implications. 4

6 OECD and World Bank. Most of the literature about China s fiscal equalization measures the disparity-reducing effect by the dispersion of fiscal capacity before and after equalization. Scholars apply the revenue-raising approach and measure local fiscal capacity by per capita revenue or per capita expenditure and interpret the dispersion of the measure as fiscal disparities (Zhao, 2009). However, empirical research on this topic has produced mixed results. Furthermore, most of the studies are provincial-level analysis with a focus on the central transfer system. There are some studies on aggregate sub-provincial fiscal disparities focusing on the overall fiscal transfer system, but very few examine intra-provincial disparities at the sub-provincial level and they tend to pay inadequate attention to the role played by provincial transfer systems. The following sections will review the research on the intergovernmental transfer system s equalization effects on provincial- and sub-provincial disparities published by both Chinese and Western journals and try to explain the mixed results in current empirical studies through a closer review of the research methods employed in those studies. 2.1 INTERGOVERNMENTAL TRANSFER SYSTEM S EQUALIZATION EFFECTS ON PROVINCIAL-LEVEL DISPARITIES IN CHINA Scholars have found that China s intergovernmental system has helped to reduce provincial-level fiscal disparities (Qiao et al, 2008), although some of them found the equalization effects very limited (Dabla-Norris, 2005; Hu and Zhang, 2007; Ou-yang, 2007; Zeng, 2000). Most studies suggest, however, that the system has not reduced but instead has further expanded provincial disparities (Liu, 2002; Wang, 2005). The impacts of different types of transfers have also been studied. Tax rebates are generally considered to have an antiequalization effect, but the effects of general-purpose and specific-purpose transfers are more debated. Gao (2010) found that general-purpose transfers have demonstrated a certain degree of anti-equalization, while Liu (2007), Hu and Zhang (2007), Huang and Chen (2012), Shah and Shen (2006) found them equalizing. For specific-purpose transfers, Wang (2005), Liu (2008), and Shah and Shen (2006) demonstrated their equalization effects, while Zhou and Yan (2010) and Huang and Chen (2012) found them anti-equalizing. The above mixed results can be explained by the different time periods examined in different studies. For example, Knight and Li (1999) analyzed the inter-provincial Gini coefficient changes, finding that central transfer s equalization effect decreased during the period of Zeng (2000) found that the transfer system during the period of did not have a significant equalization effect. Zhao (2009) demonstrated that it was equalizing in the tax-sharing period ( ), but anti-equalizing in the contracted fiscal system (cai zheng bao gan zhi) period ( ), with an improved anti-equalization effect after the TSS reform ( ). Huang and Chen (2012) focused on the post-1994 period but their data analysis covers only the period of Wang and Herd (2013) compared the effect of transfers in reducing inter-provincial disparities in 2000 and 2010 and found that China s transfer system became markedly more redistributive in It is not surprising that recent data show different equalization effects because China s transfer policies have changed rapidly. However, Wang and Herd (2013) only compared two years and it is not clear how the trend changed between the two years observed. More empirical evidence is needed to update our knowledge and understanding of China s fiscal transfer reforms. 5

7 2.2 INTERGOVERNMENTAL TRANSFER SYSTEM S EQUALIZATION EFFECTS ON SUB-PROVINCIAL DISPARITIES IN CHINA Although the issue of China s intergovernmental transfers equalization effects has been widely studied, most of the research has focused on provincial-level disparities. In China, subprovincial governments, especially at the county level, have the primary responsibility of delivering basic public services to residents, such as primary education, public health and social welfare. The fiscal capacity of sub-provincial governments will obviously affect the quantity and quality of public service delivery, which makes the issue of sub-provincial fiscal disparities extremely important. Empirical studies have produced contrasting findings on the overall equalization effects of the intergovernmental transfers at the sub-provincial level. While early research efforts report that China s intergovernmental transfer system has widened sub-provincial disparities (Tsui, 2005; Yin, 2008), more recent studies demonstrate some equalization effects (Liu et al, 2014; Wang and Herd; 2013). This is partly because of the different time periods examined in those studies, but also because of the different research methods used, as discussed in the section 2.3. One limitation of this body of research is the aggregate level analysis of the studies. As Zhang (2013) and Zhou and Yan (2010) have pointed out, the approach of pooling more than 2,000 local governments together did not take into account the variation in provincial-local fiscal arrangements across provinces. Zhang (2013) sought to examine the extent of intra-provincial disparities at the subprovincial level by calculating sub-provincial fiscal disparities by province. Using 2005 data, Zhang (2013) found that although the transfers expanded rather than narrowed inter-provincial disparities, they did reduce intra-provincial disparities, though the equalization effect varied from province to province. Wang and Herd (2013), using 2009 data, differentiated withinprovince from between-province inequalities and also found a reduction in within-province inequalities at both the prefecture and county level due to transfers, with the extent of fiscal equalization varying across provinces. In both studies, the reduction in inequality is measured using the conventional technique, i.e. as the change in inequality indices (the coefficient of variation or Gini coefficient) ex-anti and ex-post total transfers, and there is no breakdown of different types of transfers. Zhou and Yan (2010) used the Gini coefficient decomposition method to decompose the contribution of different types of transfers to county-level fiscal disparities within each of China s 27 provinces during They found that the equalization effect of the formulabased equalization transfers is not significant, and that the effects vary across different provinces. The varying extent of fiscal equalization across provinces implies different roles played by different provincial transfer systems, thus pointing to the importance and urgency of individual province study, which should take into account the practical effect of sub-provincial transfers to better measure the micro policy effect of China s intergovernmental transfer system. 6

8 2.3 THE RESEARCH METHOD ISSUE The mixed evidence on the equalization effects of China s intergovernmental transfers suggests that research findings are very sensitive to data periods and research methods. Most studies use a simple comparison approach, measuring equalization effects by calculating the decrease of selected inequality indices after adding intergovernmental transfers as a source of revenue. The studies employ various inequality indices, including the standard deviation coefficient, coefficient of variation, convergence coefficient, Gini coefficient, Theil Index, and the way those indicators are calculated can influence the results. Moreover, as Yin (2008) has noted, a simple before-after comparison of inequality indices may show misleading results, as it merely considers individual income distribution without concerning the extent to which the transfer distribution inclines toward the poor. Given that the fundamental policy goal of China s transfer system is to compensate for regional disparities and support poorer regions, the measurement of equalization effect should focus on whether and to what extent the transfers are inclined to poorer regions instead of whether they have made inequality indicators fall. In general, the overall inequality index after adding a revenue source is not only determined by the distribution of this revenue and original income, but also relies on the relationship between them (Yin, 2008, pp.119). In other words, the decline in inequality indices after introducing one source of income (including different components of transfers) can only prove that the inequality of income distribution after the introduction of this new income is lower than before. It cannot explain whether this source of income has an equalization effect if the strong get more and the weak receive less ; only a distribution plan where the strong get less and the weak receive more can achieve the goal of equalization. To better illustrate this point, Yin (2008) gives a hypothetical example: suppose that the pre-transfer income distribution for three regions is [100, 60, 1], where their coefficient of variation, Gini coefficient and Theil coefficient are 0.93, 0.41, and 0.47 respectively. If there is a transfer scheme [30, 20, 10] for the three regions, the overall income distribution would be [130, 80, 11], with the three inequality indicators reduced to 0.81, 0.36, and 0.18 respectively (decreased by 12.9%, 13.89% and 61.70% respectively). This transfer scheme is clearly contrary to the policy goal of equalization although it reduced the three inequality indices. More recent studies have overcome this methodological challenge by employing the decomposition method used in income inequality studies. Zhao (2009), Tsui (2005) and Zhou and Yan (2010) decomposed overall provincial- or county-level fiscal disparities by individual income sources. Gao (2010) and Yin (2008) also applied the Gini coefficient decomposition method but went further to decompose each income source to analyze the specific way each component of transfers contributes to overall local fiscal disparities. Gao (2010) and Yin (2008) s extended two-step decomposition method is a breakthrough and partly explains their inconsistent findings with other studies. However, both of them pool all local government units without controlling for province-specific effects, which may also produce misleading results. The variation of province-specific fiscal transfer systems warrants case studies of individual provinces to better understand the micro policy effect of China s intergovernmental transfer system. This research is the very first single-province analysis focusing on province-specific transfer system s role in addressing intra-provincial disparities. It uses Henan Province as a case study and longitudinal data for a more recent period ( ). It employs the extended 7

9 two-step decomposition techniques by Gao (2010) and Yin (2008) to analyze the equalization effects of different components of transfers and their specific effects on intra-provincial disparities at the county level. 3. PROVINCIAL TRANSFER SYSTEM AND COUNTY-LEVEL GOVERNMENT FINANCE IN HENAN PROVINCE Henan Province is in the middle region of China and is a major agricultural province with a large rural population. Its budgetary revenues rank 10 th and budgetary expenditure ranks 5 th, however, its own-source revenues only cover 31.9% of budgetary expenditures. As a deficit province, it is a typical recipient of central government transfers. Despite its big budget size due to its large population, Henan Province ranks at the bottom in terms of budgetary revenues per capita (only 2670 RMB), about 60% of the national average. It has 17 prefecture-level government units and 159 county-level government units including 88 counties, 21 countylevel cities and 50 urban districts. 2 Following the central-provincial transfer system established by the central government, Henan Province established its own transfer system after the 1994 TSS reform. Table 1 shows the composition of the central-provincial transfers to Henan Province and provincial-local transfers in Jiyuan City s status was changed from county-level to sub-prefecture level (between county-level and prefecture-level) city in

10 Table 1. Henan Province Central-Provincial and Provincial-Prefecture Transfer Composition in 2011 (Unit: 10,000 YUAN) Transfers from Central Government to Henan 25,089,960 Transfers from Henan Province to Its 20,476,190 Province Prefectures/Counties Tax Rebates 2,434,039 Tax Rebates 1,232,738 VAT and Excise Tax Rebates 1,428,090 Personal and Enterprise Income 290,949 Tax Rebates The rebates for the reform of the prices of 715,000 and taxes and fees on refined oil products General-Purpose Transfers 13,107,751 General-Purpose Transfers 11,004,284 Equalization Transfers 4,626,500 Equalization Transfers 4,460,849 Total Other General-Purpose 8,450,853 Total Other General-Purpose 6,543,435 Transfers Transfers Transfers for Minority Regions 13,905 Transfers for Minority Regions 13,905 Transfers for Increasing Wages of 1,724,493 Transfers for Increasing Wages of 1,572,538 Civil servants Civil Servants Transfers for Rural Tax Reform 777,234 Transfers for Rural Tax Reform 796,291 Awards and Subsidies for 645,684 Implementing the Mechanism of Ensuring Basic Funding to County-Level Governments Awards for Major Grain (Oil)- 262,109 Awards for Major Grain (Oil)- 244,430 Producing Counties Producing Counties Transfers for Year-end Settlement 29,266 Transfers for Year-end Settlement 73,958 Transfers for Resource-Exhausted 67,600 Transfers for Resource-Exhausted 67,600 Cities Cities Transfers for Changes of Enterprises 186,897 and Public Institutions Subordinate Relationship Transfers for the Reform of Prices and 240,400 Transfers for the Reform of Prices 32,978 Taxes and Fees on Refined Oil Products of and Taxes and Fees on Refined Oil Products Transfers for Removal of Two Fees 33,274 by Industry & Commerce Department Transfers for Public Health 987,327 Transfers for Public Health 987,223 Transfers for Public Safety 203,875 Transfers for Public Safety 324,367 Transfers for Education 1,044,681 Transfers for Education 1,246,403 Transfers for Social Security and 2,160,434 Transfers for Social Security and 1,027,833 Employment Employment Transfers for Agriculture, Forestry and 73,674 Transfers for Agriculture, Forestry 154,453 Water Conservancy and Water Conservancy Transfers for General Public 273 Services Others 30,398 Others 1,183 Specific-Purpose Transfers 9,548,170 Specific-Purpose Transfers 8,239, THE CENTRAL-PROVINCIAL TRANSFERS TO HENAN PROVINCE China s current central-provincial transfer system was introduced after the 1994 TSS reform and the categories and allocation methods have been reformed several times since then. Since 2009, China s Ministry of Finance has adopted three broad categories of transfers: (a) tax rebate transfers, (b) general-purpose transfers and (c) specific-purpose transfers (or earmarked transfers). Table 1 provides the composition of central-provincial transfers to Henan Province. 9

11 3.1.1 TAX REBATES The tax rebate transfers are origin-based revenue sharing transfers between the central and sub-national governments, consisting of VAT and excise tax rebates, personal income tax (PIT) and enterprise income tax (EIT) rebates, and rebates for the reform of the prices of and taxes and fees on refined oil products. The design of those tax rebates is based on the harmless principle to win support for the fiscal reform from sub-national governments. VAT and Excise Tax Rebates. With the 1994 TSS reform, VAT and excise taxes were brought under central administration and sub-national governments share 25% of the proceeds of VAT (and zero % of excise taxes) collected in their provinces. VAT and excise tax rebates were introduced to assure the provinces a minimum level of VAT and excise tax revenues the VAT and excise tax revenues they retained in 1993 as well as 30% of the growth in VAT and excise tax revenues each year. Personal income tax (PIT) and enterprise income tax (EIT) rebates. The income tax revenue-sharing scheme was introduced in Under this scheme, the central and subnational governments shared the personal and enterprise income taxes on a basis for 2002 and on a basis since A tax rebate program similar to the VAT and excise tax rebates was instituted. The provincial vested interests were respected under this scheme by the provision that the amount of income taxes received by provinces would not be less than the amount they received in Rebates for the reform of the prices of and taxes and fees on refined oil products. This rebate was introduced in 2008 to compensate for sub-national governments revenue losses due to the reform of the prices of refined oil products and taxes and fees on them. The rebate uses provinces proceeds from road maintenance fees and five other fees retained during the year 2007 as the base. The above three types of tax rebates represent about 10% of total central transfers to Henan Province in GENERAL-PURPOSE TRANSFERS According to the MOF s classification, general-purpose transfers (called fiscal capacity balanced transfers before 2009) refer to the subsidies from the higher to lower level governments to close the lower level government s fiscal gap in accordance with standardized allocation approaches. These transfers have no specific spending purposes attached and subnational governments have discretion in fund allocation and use. The primary category of general-purpose transfers is equalization transfers, originally called transitory period transfers after the 1994 TSS reform and later renamed general-purpose transfers after 2001 until the adoption of the new classification method in The equalization transfer was China s first comprehensive formula-based transfer scheme. It took into account certain objective factors such as the gap between standard revenues and expenditures and the size of available central funds, with a policy objective of addressing regional fiscal disparity by equalizing both fiscal capacity and expenditure needs. Equalization transfers to Henan Province amount to 35% of total general-purpose transfers and 18% of total transfers in

12 In addition to this comprehensive equalization transfer scheme, the general-purpose transfers also include transfers for minority regions; transfers for increasing wages of civil servants; transfers for rural tax reform; transfers for year-end settlement; transfers for resourceexhausted cities; transfers for changes of enterprises and public institutions affiliation; transfers for the reform of prices of and taxes and fees on refined oil products; transfers for removal of two fees by the industry & commerce department; transfers for local public security organs, procuratorial organs and people's courts; transfers for rural compulsory education; transfers for basic pension and subsistence allowances; transfers for the new rural co-operative medical care system; and others general-purpose transfers. According to the MOF, those transfers are used to help finance basic public service provision of sub-national governments. Their allocations are also based on objective factors, which are the characteristics of general-purpose transfers, therefore they are classified as general-purpose transfers. Indeed those transfers in practice are closer to general-purpose than specific-purpose transfers and the author believes that they are more suitably classified as block grants (other than general grants or categorical grants) following the classic classification in the public finance literature. For the purpose of this study, the author follows MOF s classification and labels them as total other general-purpose transfers (other than equalization transfers). They represent 65% of general-purpose transfers and 34% of total central transfers to Henan Province. 3 Together with the equalization transfers, they represent 52% of total central transfers to Henan Province in SPECIFIC-PURPOSE TRANFERS (EARMARKED TRANFERS) According to MOF, the specific-purpose transfers (earmarked transfers) refer to the subsidies from the higher to lower level governments to achieve specific economic or social development goals, and must be spent by recipient governments according to the specified purposes. Those transfers are policy tools to carry out high-priority policy objectives of the central government, primarily in the field of agriculture, forestry and water conservancy, education, health, social security and employment, transportation, energy saving and environmental protection. There were approximately 220 specific-purpose transfer programs in 2013; the number declined to about 150 in Our data source provides no details of individual specific-purpose transfers so Table 1 only reports the total amount of all specificpurpose transfers which represent about 38% of total central transfers in THE EXISTING SYSTEM OF PROVINCIAL-LOCAL TRANFERS IN HENAN PROVINCE All central transfers intended for prefectures and counties have to be disbursed by provincial governments, and the sub-provincial fiscal transfer system (as well as the subprovincial revenue and expenditure assignments) may vary across provinces. It therefore matters how provinces design their fiscal transfer system and further distribute central transfers 3. Note that in Table 1, Henan Province listed the awards and subsidies for implementing the mechanism of ensuring basic funding to county-level governments and the awards for major grain (oil)-producing counties under general-purpose transfers other than equalization transfers. Those two transfers were included in the equalization transfers by MOF. 11

13 down to sub-provincial governments, especially for provinces, like Henan Province, that rely heavily on central transfers to balance their budgets. This section will briefly discuss the existing system of the sub-provincial fiscal transfer system in Henan Province REVENUE SHARING AND TAX REBATES Following the central tax rebate system, Henan Province s tax rebate system is also based on the harmless principle and has experienced three major adjustments since 1994 (in 1996, 2004 and 2009). Starting in 1994, for the 25% of the proceeds of the VAT and excise taxes and later the 40% of the proceeds of the PIT and EIT) retained by provinces, the ratio of provincial and sub-provincial shares was 2:8. For the tax rebates returned by the central government, subprovincial governments only retained half of the 30% growth in those revenues retained in the province (for a province-sub-province ratio of 1:0.15). The 1996 adjustment let sub-provincial governments retain all of the 30% growth in VAT returned by the central government (for a ratio of 1:0.3 for the growth in the VAT). The 2004 adjustment increased the sub-provincial governments share of the growth in business tax, PIT and EIT retained in the province from 50% to 80%, and the 2009 adjustment changed sub-provincial share of the growth in the VAT back to 80% and EIT to 85%. Generally speaking, the sub-provincial tax sharing arrangement and tax rebate program favor sub-provincial governments, with the provincial government having a gradually smaller share in the growth of revenues. The implication might be that the provincial government slowly plays a less redistributive role. In 2011 it distributed approximately 51% of central-provincial tax rebates to sub-provincial governments, representing about 6% of total provincial-local transfers GENERAL-PURPOSE TRANSFERS As the central government, Henan Province s general-purpose transfer system includes both equalization transfers and other general-purpose transfers. The equalization transfer scheme underwent a major reform in 2007, which introduced two types of equalization transfers: transfers for ensuring minimum basic operation expenditures, and transfers for basic public service equalization. The transfers for ensuring minimum basic operation expenditures are designed to ensure salary payment to public employees, normal operation of government institutions, and the funding of necessary programs, calculated as the gap between the recipient government s minimum basic operation expenditure needs and its standard fiscal capacity. The former are calculated as the sum of personnel expenditures, operation expenditures and program expenditures, and the latter is estimated using the previous year s actual fiscal capacity measured as the weighted sum of tax revenues, tax rebates and selected other general-purpose transfers such as transfers for increasing wages of civil servants, transfers for rural tax reform, and awards for major grain-producing counties 4. Personnel expenditures mainly consider the amount of fiscal dependents and provincial average salary levels for different government departments. Operation expenditures are 4. The specific components of other general-purpose transfers included for calculation and the assigned weight to each component are subject to change each year. 12

14 calculated based on the operation costs of four sectors of government: general administration, public safety (including public security organs, procuratorial organs and people s courts), education, and all other departments. Algebraically, the amount of operation expenditures for a county or city is given by: OEi = (GAi+ PSi + EDi + AOi) * Ri (1) where OEi = the amount of operation expenditures for county or city i; GAi = the operation costs of general administration for county or city i; PSi = the operation costs of public safety for county or city i; EDi = the operation costs of education for county or city i; AOi = the operation costs of all other departments for county or city i. For the sectors of general administration, public safety, and all other departments, each sector s operation costs are determined by the specified annual per capita operation spending standard for the sector multiplied by the sector s authorized staffing number for county or city i. For the education sector, the operation costs are calculated as the product of annual per student education spending standard and student population. Ri = cost variance ratio for county or city i, determined by the weighted average of fiscal dependent population ratio and area size ratio for county or city i. Program expenditures include statutory expenditures in education, science and technology, and agriculture programs, and priority expenditures in cultural, health, and environmental protection programs and other necessary programs funded by county-level governments (i.e. family planning, infrastructure, and earmarked programs requiring matching funds). The transfers for basic public service equalization are based on the difference between the recipient government s standard expenditures and standard revenues. Standard expenditures are calculated based on different per capita spending for different categories of outlays for adjusted cost factors such as population size and population density. The categories of spending are based on classification of government expenditure by function, including general public services, public safety, education, culture and sports and media, medical and health care, urban and rural community service, agriculture, forestry, transportation, and social security and employment. Standard revenues are composed of general budgetary revenues of the previous budget year, plus the current budget year s estimated tax rebates, transfers for ensuring minimum basic operation expenditures, other general-purpose transfers (excluding transfers for resource-exhausted cities, transfers for key provincial eco-functional zones, transfers for minority regions, awards for major grain (oil)-producing counties), and specificpurpose transfers. The transfer amount is also adjusted for coefficients that take into account the size of the standard gap and also the total equalization grant available in the budget year. Algebraically, it is calculated as: TRi = (SEi SRi) * σi (2) where TRi = the amount of the transfers for basic public service equalization for county or city i; SEi = total standard expenditure for county or city i, calculated as the sum of total standard 13

15 spending for 10 sectors 5. For each sector, the average per capita standard expenditure for the recipient government is multiplied by its population and adjusted by the same cost variance ratio Ri used in the calculation of the amount of operation expenditures for transfers for ensuring minimum basic operation expenditures ; SRi = total standard revenue for county or city i; σi = the transfer coefficient for county or city i, determined by the total equalization grant available in the budget year and the gap between total standard expenditure SEi and total standard revenue SRi. It is clearly stated in Department of Finance s official document that the first type of equalization transfer is given priority over the second, accordingly, the equalization transfer system in Henan Province focuses more on ensuring the basic operation of government than on ensuring basic public services. As shown in Table 1, Henan Province distributed 44.6 billion RMB to sub-provincial governments, 96% of the equalization transfers it received from the central government. Regarding other general-purpose transfers, Henan Province reduced the number of categories by pooling some of the general-purpose transfers by functional area. For example, transfers for local public security organs, procuratorial organs and people s courts, transfers for rural compulsory education, transfers for basic pension and subsistence allowances, and transfers for the new rural co-operative medical care system are grouped into transfers for public safety, education, social security and employment, and public health, respectively. Total other general-purpose transfers amount to 65 billion RMB, 77% of total other general-purpose transfers it received from the central government. Equalization and other general-purpose transfers amount to 54% of total provincial-local transfers (22% and 32% of total provincial-local transfers respectively) SPECIFIC-PURPOSE TRANSFERS (OR EARMARKED TRANSFERS) Specific-purpose transfers amount to 82.4 billion RMB in 2011, approximately 86% of central-provincial specific-purpose transfers. They contribute to 40% of total provincial-local transfers in Itemized information is not available. Two observations can be made based on the comparison between the central-provincial and provincial-local transfers in Henan Province. First, the composition of the two transfer systems is similar, with the equalization transfers making up a slightly higher portion in the provinciallocal transfer system. Second, the provincial government did not pass all central transfers to sub-provincial governments. According to Table 1, the provincial government retained 46 billion RMB (18%) central transfers, including 12 billion RMB tax rebates (49%), 1.6 billion RMB equalization transfers (4%), 19.4 billion RMB other general-purpose transfers (23%) and 13 billion RMB specific-purpose transfers (14%). The possible explanation for equalization transfers high pass on percentage is that the central government s equalization transfer scheme is more transparent than other transfer schemes and its calculation is directly based on the 5. Some categories of spending are based on actual spending, such as defense, science and technology, water conservancy, environmental protection, financial regulation, land and resources, public housing, etc., considering the difficulties posed by the objective measurement of cost factors and also high degree of policy differences across jurisdictions. 14

16 measurement of county-level units in each province although a lump sum is given to each province. 3.3 COUNTY-LEVEL GOVERNMENT FINANCE IN HENAN PROVINCE As mentioned above, Henan Province consists of 17 prefecture-level and 159 county-level governments, including 88 counties, 21 county-level cities and 50 urban districts. For the purpose of this study, subsequent analysis of county-level government units only concerns counties and county-level cities. 6 As shown in Table 2, during the study period of , county-level government budgetary revenues grew from 10.4 billion RMB to 53 billion RMB at an annual rate of 26.68%. Despite the rapid growth in amount, the share of budgetary revenues in total revenues (budgetary revenues plus net transfer revenues 7 ) decreased, from half in 2001 to less than a third in Over the same period, the growth of net transfer revenues outpaced budgetary revenues; net transfer revenues grew from 10.3 billion RMB in 2001 to billion RMB in 2011, nearly twice the size of local budgetary revenues. Two thirds of local government expenditures are now financed by transfers, which makes transfer redistribution a critical issue for local finance. 6. The analysis in this paper only concerns counties and county-level cities due to the absence of urban district population data. 7. Net transfer revenues = tax rebates + general-purpose transfers + specific-purpose transfers general-purpose transfer remittances to upper governments specific-purpose transfer remittances to upper governments. General-purpose transfer remittances to upper level governments include original-system remittances and export rebate remittances to upper level governments. 15

17 Year Budgetary Revenues (Own- Source Revenues) ,041,501 (50.33%) ,144,329 (42.54%) ,275,694 (41.30%) ,587,730 (39.92%) ,014,191 (39.71%) ,460,543 (39.00%) ,052,413 (37.04%) ,575,624 (35.61%) ,836,277 (31.73%) ,381,864 (33.00%) ,309,151 (32.69%) Table 2. County-Level Government Finance in Henan Province ( , 2001 Constant Price, Unit: 10,000 RMB) Total (Net) Transfers Total 1,027,641 (49.67%) 1,545,831 (57.46%) 1,813,366 (58.70%) 2,389,660 (60.08%) 3,057,890 (60.29%) 3,848,320 (61.00%) 5,188,444 (62.96%) 6,464,042 (64.39%) 8,252,638 (68.27%) 8,895,045 (67.00%) 10,934,130 (67.31%) Tax Rebates 236,339 (11.42%) 346,361 (12.88%) 370,025 (11.98%) 383,996 (9.65%) 382,229 (7.54%) 376,803 (5.97%) 372,952 (4.53%) 368,395 (3.67%) 378,856 (3.13%) 380,873 (2.87%) 380,838 (2.34%) Sub-Total 550,841 (26.62%) 892,351 (33.17%) 1,034,763 (33.50%) 1,510,915 (37.99%) 2,046,785 (40.35%) 2,764,305 (43.82%) 3,477,359 (42.20%) 4,253,980 (42.37%) 4,609,113 (38.13%) 4,920,719 (37.06%) 6,956,513 (42.83%) General-Purpose Transfers Equalization Transfers 45,463 (2.20%) 120,525 (4.48%) 214,759 (6.95%) 481,438 (12.10%) 647,545 (12.77%) 963,449 (15.27%) 1,309,602 (15.89%) 1,984,800 (19.77%) 1,921,035 (15.89%) 2,361,906 (17.79%) 2,952,853 (18.18%) Other General Purpose Transfers 505,378 (24.42%) 771,826 (28.69%) 820,004 (26.55%) 1,029,478 (25.88%) 1,399,239 (27.59%) 1,800,857 (28.54%) 2,167,757 (26.30%) 2,269,180 (22.60%) 2,688,078 (22.24%) 2,558,813 (19.27%) 4,003,660 (24.65%) Specific- Purpose Transfers 212,522 (10.27%) 265,347 (9.86%) 359,149 (11.63%) 427,029 (10.74%) 543,039 (10.71%) 707,212 (11.21%) 1,338,133 (16.24%) 1,841,666 (918.34%) 3,264,670 (27.01%) 3,593,454 (27.07%) 3,596,780 (22.14%) Revenues (After Transfers) 2,069,142 2,690,160 3,089,060 3,977,390 5,072,081 6,308,863 8,240,857 10,039,666 12,088,915 13,276,909 16,243,281 The size of tax rebates has grown relatively slowly compared with other types of transfers. Since 2003, tax rebates have stabilized at 3.8 billion (2001 constant price), consistent with the policy objective of the TSS reform to limit the overall size of tax rebates to sub-national governments. General-purpose transfers, the largest portion of total transfers in Henan Province, grew steadily at 27.28% annually over the ten years. The share of total revenues made up by general-purpose transfers experienced fast growth until 2006, increasing from 26.62% to 43.82%. Their share of total transfers declined during the period due to the drastic increase of specific-purpose transfers since 2006, but bounced back to 42.83% in The increase of general transfers was mainly driven by rapid growth in equalization transfers, which jumped from 2.2% of total revenues in 2001 to in The size of specific-purpose transfers grew most rapidly among the three transfer components, with an average annual growth rate of 51.8% over the 10 years. The growth rate peaked during (130%) and (143.97%) due to the substantial increase in transfers related to the central government s Rise of Central China (zhongbu jueqi) plan. As a result, the share of total revenues contributed by specific-purpose transfers expanded from 10.27%in 2001 to 27.07% in 2010 (but slightly declined to 22.14% in 2011). Table 3 presents descriptive statistics of per capita budgetary revenues and expenditures of county-level governments for the period The discrepancy between the maximum and minimum per capita budgetary revenues widened dramatically over the years. In 2002, per capita budgetary revenues were 450 RMB in the richest county (city) but 70 RMB in the poorest county (city), a difference of In 2008, per capita budgetary revenues for the richest county 16