Tax Sharing and Fiscal Imbalances: Reconstruction of the VAT Sharing System in China

2.1 Institutional Background

Over a long period of time, business tax and VAT have been important factors in determining the level of local government’s fiscal capacity, and are also the main reasons for regional fiscal imbalances. Since the tax sharing reform, China’s local government fiscal autonomy has been declining, especially since the income tax sharing reform in 2002, local governments’ fiscal power has been further weakened, followed by the deepening of vertical fiscal imbalances. Since then, China’s central-local fiscal relationships have basically stabilized, the central and local governments share fiscal revenues in important tax sources, while local governments gradually develop a financial pattern with business tax as their main source of fiscal revenue. However, due to the differences in industrial structure and economic scale, the regional distribution of business tax and VAT is very uneven, and the scale of the two taxes in the eastern region is much larger than that in the central and western regions. Table 1, by breaking down the variation coefficient of provincial government revenue according to the composition of budget revenue, presents that two tax revenues, including business tax and VAT, are the main causes of the imbalance of local governments’ fiscal capacity, and the contribution of the two tax revenues to fiscal revenue imbalance is as high as 32%.

After the reform of “replacing business tax with VAT”, in the absence of the main taxes of local governments, tax sharing has become an important link in the fiscal relationship of China’s central government and local governments. Since May 1, 2016, China has launched a nationwide pilot of “replacing business tax with VAT”, with all business tax taxpayers paying VAT. Prior to the reform, business tax was the main source of local government revenue, and VAT revenue was shared by the central government and local governments according to the ratio of 75:25. In order to alleviate the impact of “replacing business tax with VAT” on local government revenue, China has adjusted the proportion of central and local VAT sharing, in an attempt to seek the relative stability of central and local fiscal capacity. The State Council has determined that the central return and local surrender base should be determined with the year of 2014 as a base number, and that the central and local governments share 50% of the VAT revenue respectively, and that the VAT paid by enterprises in all industries is included in the scope of central and local sharing. During the transitional period, revenue from VAT sharing becomes the main source of local revenue.^[1] In fact, since the reform of 50:50 sharing, there have been obvious differences in the growth of VAT revenue in the eastern, central and western regions of China, with the VAT revenue in the eastern region growing faster and increasing at a faster rate after the reform.^[1] However, this trend of VAT may contain two effects, one is the increase of local VAT revenue caused by “replacing bussiness tax with VAT”, and the other is the rapid increase of VAT brought about by the increase of local VAT share, which is the focus of study in this paper.

2.2 Theoretical Analysis

In theory, when residents of different jurisdictions within a state get the same level of utility from government public services, there would be no inefficient flow of laborers among regions due to their preference for public services, thus enhancing the overall well-being of society, which is also the government’s goal. By subsidizing local governments to ensure the smooth realization of their policy objectives, if the central government funds local governments according to the source of tax revenue, which, despite reducing the pressure on local government expenditure, and easing the vertical fiscal imbalances of local governments, also widens the fiscal gap among different regions, leading to horizontal fiscal differentiation. As has been recognized in existing studies, the change in China’s vertical fiscal imbalances stems mainly from the adjustment of fiscal revenues (Lu and Li, 2019), but few studies have noted that such adjustments in intergovernmental fiscal revenues also lead to horizontal fiscal imbalances. Intuitively, the 50:50 VAT sharing scheme is fair to the provinces, but it is easy to ignore that the impact of “replacing business tax with VAT” on local fiscal capacity varies significantly among provinces, and the scale of business tax and the composition of the tax base will affect local governments’ tax revenues after the reform of “replacing business tax with VAT”. In fact, this “one size fits all” sharing method is essentially the redistribution of fiscal revenue according to the source of revenue, with significantly different impact on the fiscal revenues of different provinces. The level of fiscal capacity of a province may be differentiated at a faster pace in the process of sharing, and the degree of vertical and horizontal imbalances of the provinces with scarce fiscal capacity may be deepened.

Horizontal fiscal imbalances are caused by the imbalance of fiscal revenues that can be drawn from different jurisdictions, and the deeper cause is the difference in economic scale and industrial structure across different regions, which determines the total amount and structure of local tax revenue. A number of studies have found that difference in regional fiscal capacity is an important cause for fiscal imbalances (Boadway and Tremblay, 2006; Bird and Villancourt, 2006). In theory, a liquid base tax should not be taxed by local governments, and its allocation to local governments would create a shortage of local governments to provide public goods and induce local governments to compete for fiscal capacity, which could further exacerbate the relative fiscal gap between local governments. The conclusions of empirical studies are in line with theoretical expectations. Some studies have found that the fiscal incentive effect of tax sharing will change the behavior of local governments. The study conducted by Han and Kung (2015) shows that China’s income tax sharing reform has changed the development strategy of local governments, which devote more efforts to develop industries conducive to local tax, so as to grab more business tax revenue, and these actions in the long run can exacerbate horizontal fiscal imbalances.

In addition, the transfer of tax revenue caused by the principle of productive tax collection will also have an impact on governments’ fiscal situation. China’s tax system has typical productive characteristics in that the scale of local production determines the level of local fiscal capacity. But consumption ends up in the nationwide market, with goods flowing to other regions at tax-inclusive prices, which causes residents of other jurisdictions to bear the tax burden of this jurisdiction. Tax revenue based on the place of production can result in a tax burden export problem, transferring some tax burdens of some jurisdictions to other jurisdictions, and compromising the fiscal capacity of the tax revenue outflowing places. Consumption-based taxation, which can effectively avoid this defect, is conducive to improving the supply and consumption environment of public goods in a jurisdiction. This may become an important part of future tax reform.

The two characteristics of VAT in China determine that VAT sharing have an impact on governments’ horizontal fiscal pattern. On the one hand, China’s VAT sharing allocates tax revenue according to the source of revenue. The central government returns local VAT revenue according to the size of the VAT collected, with regions of larger VAT size acquiring a larger amount of VAT revenue in sharing. On the other hand, as China’s VAT is levied according to the principle of the place of production, the regional transfer of VAT revenue affects a place’s fiscal revenue. For the inflowing place, the degree of vertical fiscal imbalance may decrease; for outflowing places, fiscal imbalances are likely to deepen.

3.1 Measurement of Fiscal Imbalances

Fiscal imbalances include vertical and horizontal fiscal imbalances. Vertical fiscal imbalances (VFI) refer to the relative imbalance of fiscal capacity between central and local governments, also known as “fiscal gap” or “transfer payment dependency”, which occur when local governments’ own revenue cannot meet the demands of their expenditure responsibilities. Based on the common practice of the existing literature, this paper uses the proportion of a jurisdiction’s own revenue in fiscal expenditure to construct a measurement index of vertical fiscal imbalance (Boadway and Tremblay, 2006).

In this model, p represents a province, t represents time, S^own represents fiscal expenditure, R^own represents its own fiscal revenue, and VFI_pt represents the degree of vertical fiscal imbalance of the p province in the t year, which measures the portion of the province’s expenditure responsibilities that cannot be met by the province’s own income.

Horizontal fiscal imbalances (HFI) refer to financial imbalances between governments at the same level, and the available literature generally uses locally owned fiscal revenue to calculate the Gini coefficient and the Theil index (Tsui, 2005). However, this approach has resulted in fiscal capacity and fiscal effort being included in fiscal revenue, and the policy schemes proposed therefrom are often prone to factor distortions and even reverse incentives to government behaviors (Bird and Villancourt, 2006). To avoid these problems, Liddo et al. (2014) have proposed a measure based on fiscal capacity, arguing that the essence of fiscal imbalances across different jurisdictions lies in the imbalance of fiscal capacity, using imbalance of fiscal capacity between administrative units as a measure of horizontal fiscal imbalances. Fiscal capacity measures the ability of jurisdictions to raise fiscal revenue, which is generally considered to be the part of fiscal revenue excluding fiscal effort, depending on the size of a region’s economy and the degree of its industrial development. A growing number of studies have introduced regression methods to measure fiscal capacity, using socio-economic variables to regress fiscal revenues and predict the size of a region’s fiscal capacity (Bird et al., 2008; Liddo et al., 2014). Fiscal effort reflects the extent to which the government taps into its existing tax base and tax capacity. This paper takes fiscal revenue as an explanatory variable, and takes industrial structure, population, consumption and other variables as control variables, to estimate the fiscal capacity and fiscal effort of the provinces on the basis of controlling the two-way fixed effect.^[1]

3.2 VAT Sharing Reform and Fiscal Imbalances

3.2.1 VAT Sharing Reform and Vertical Fiscal Imbalances

The adjustment of the rules of financial distribution between different levels of finance can affect the vertical fiscal structure of the government. To analyze the impact of VAT sharing reform on vertical fiscal imbalances, we must consider the impact of comprehensively “replacing business tax with VAT”. In 2016, China implemented two important fiscal and tax reforms, “replacing business tax with VAT” and VAT sharing. Although both reforms affect vertical fiscal imbalances, they exert impact in the opposite direction on such imbalances. “Replacing business tax with VAT” have an impact on local government fiscal capacity, resulting in the lack of local main taxes, and worsening vertical fiscal imbalances; whereas the VAT sharing reform enhances the ratio of local governments’ VAT sharing and reduces the negative impact of “replacing business tax with VAT”, thus alleviating vertical fiscal imbalances. Therefore, this paper first calculates the degree of fiscal imbalances before “replacing business tax with VAT” as a benchmark, then calculates the degree of vertical imbalances after “replacing business tax with VAT” and VAT sharing reform, as a comprehensive response to the reform effect of “replacing business tax with VAT” and VAT sharing, and finally restores the VAT share ratio to the distribution ratio before “replacing business tax with VAT” to simulate the degree of fiscal vertical imbalances in the counter-factual state of the VAT sharing reform.

Figure 1 presents the results of measuring vertical fiscal imbalances. On the whole, the degree of vertical fiscal imbalances in the eastern provinces is lower than that in the central and western regions, and the degree of vertical fiscal imbalances in the western provinces is the highest. Comparing the degree of fiscal imbalances after the reform of “replacing business tax with VAT” and VAT sharing, it is found that the reform of VAT has alleviated the vertical imbalance of local finance to a certain extent. Compared with the situation of only “replacing business tax with VAT”, the vertical fiscal imbalance of each province has decreased to varying degrees, but the relative fiscal situation of the provinces has not been restored to the level before the reform of “replacing business tax with VAT”. This means that the VAT sharing reform may have a negative impact on the horizontal fiscal structure of different regions and provinces.

Figure 1

The Degree of Vertical Fiscal Imbalance in Different Provinces and Regions

3.2.2 VAT Sharing Reform and Regional Horizontal Fiscal Imbalance

The adjustment of the rules of vertical fiscal allocation not only affects the financial structure governments at different levels, but also changes the relative fiscal situation of governments at the same level. Figure 2 directly calculates the extent of the decline of vertical fiscal imbalances in provinces and regions after the sharing reform, from which it can be found that the distribution of the effect of the VAT sharing reform varies greatly, with Beijing, Guangdong, Shanghai, Jiangsu and Zhejiang having the largest decline in vertical fiscal imbalances, which means that these provinces have benefited the most from the VAT sharing reform. On the whole, the eastern provinces have seen the decline of their fiscal vertical imbalances more than those in the central and western provinces after the sharing reforms, which may mean that the fiscal gap between different regions may widen further after the sharing reform.

Figure 2

VAT Shares the Extent to Which Vertical Fiscal Imbalances Have Declined Since the Reform

The result of this change in vertical fiscal imbalances in Figure 2 tells us that although the reform of VAT 50:50 sharing has led to a decline in the vertical fiscal imbalances of the provinces, the degree of decline is different, and the fiscal capacity of the provinces may be differentiated at a faster pace in the process of sharing. To further investigate the differences in the impact of the VAT sharing reform on different regions, the following measurement model has been constructed.

ΔVFIpt=β0+β1Regionp× Reformt+β2Xpt+λp+δt+εptΔVFIpt=VFIptVATsharing+replacingbusinesstaxwithVAT−VFIptbeforetherefrom−VFIptreplacingbusinesstaxwithVAT−VFIptbeforetherefrom

In this model, ΔVFI_pt means the difference between the VFI value after the sharing reform and the VFI value assuming that the sharing reform is not implemented, which measures the vertical fiscal imbalance caused by the VAT sharing reform.^[1] This practice excludes the impact of the reform of “replacing business tax with VAT” on vertical fiscal imbalance and directly measures the impact of the sharing reform on vertical imbalance. Region_p means that the provinces of China are divided into eastern, middle and western regions, with the eastern regions taken as the base group, assigned to 1, while the central and western regions are assigned to 2 and 3 respectively. Reform_t is the virtual variable of the VAT sharing reform and is assigned to 1 after 2015. X_pt represents a series of control variables, including per capita GDP, industrial structure, tax effort, urbanization level and other variables. λ_p and δ_t represent the provincial fixed effect and the time fixed effect respectively. ε_pt represents residuals. In this model, the impact of the VAT sharing reform on the fiscal capacity of different regions is captured by β₁. If β₁ is negative, it shows that since the sharing reform, the vertical fiscal imbalance between the central and western regions has deepened, which may lead to a widening of the regional fiscal gap. If β₁ is positive, it indicates that the fiscal gap between the central, western and eastern regions is narrowing.

Columns (1) and (2) of Table 2 report the regression results of model (1), present the changes in vertical fiscal imbalances after the sharing reform in the eastern, central and western regions respectively. The first column only controls the time and provincial fixed effect, while the second column adds per capita GDP, industrial structure, tax effort, urbanization level and other variables. It is not difficult to find that after the sharing reform, the vertical fiscal imbalance in the central and western regions has decreased less than that in the eastern region, while the relative vertical fiscal imbalance has deepened. The adjustment of tax sharing ratio will not only affect local governments’ fiscal revenue, but also their fiscal expenditure, although some research has found that China’s fiscal imbalance mainly stems from the adjustment of income (Lu and Li, 2019), but we still want to prove that revenue adjustment has an important impact on the adjustment of local government’s fiscal situation.

Table 2

Changes in the Relative Level of Fiscal Capacity in the Eastern, Central and Western Regions after the Sharing Reform

	(1)	(2)	(3)	(4)
Variables	ΔVFI	ΔVFI	ΔVFI_adj	ΔVFI_adj
Policy Reform	−0.0178***	−0.0146***	−0.000404*	−0.00105***
	(0.000906)	(0.000985)	(0.000222)	(0.000244)
	1.92e−09	0.0671**	3.00e−09	−0.0100
Constant	(0.000992)	(0.0283)	(0.000243)	(0.00701)
Control	NO	YES	NO	YES
Year FE	ANDIT’S	ANDIT’S	ANDIT’S	ANDIT’S
Pro FE	YES	YES	YES	YES
Observations	341	341	341	341
R2	0.915	0.931	0.971	0.976

1. Note: Standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1.

To this end, this paper proves the importance of revenue adjustment to the relative level of finance from three aspects: Firstly, we map the trend of changes in per capita fiscal expenditure in different regions, and find no significant change in per capita fiscal expenditure before and after the sharing reform.^[1] Secondly, on the basis of controlling the primary and secondary items of the time trend, the per capita fiscal expenditure is used to regress the year of VAT reform, and the regression coefficient is not significantly different.^[2] Finally, to directly eliminate the impact of the sharing reform on expenditure, we deduct the increase in revenue brought by the value-added tax sharing reform from the expenditure data, recalculated ΔVFI_adj for regression, and obtained the regression results in columns (3) and (4) in Table 2. The regression results are consistent with the results in column (1).

3.3 The Effect of Sharing Reform Measured on the Basis of the Input-Output Method

In order to completely separate the impact of “replacing business tax with VAT” and the VAT sharing reform on fiscal imbalance, the key lies in accurately estimating the change of local government tax revenue following the two reforms, which requires accurately defining the tax base of VAT. Although we have tried to prove that the VAT sharing reform will affect the government’s fiscal situation, we are still worried that the “replacing business tax with VAT” may cause a deviation in the measurement. To solve this problem, we use the data of the 2012 provincial input and output table to directly simulate the relative fiscal situation of local governments under the impact of the two policies of “replacing business tax with VAT” and the VAT sharing reform. Using the input and output data for 2012 is not affected by the reform of “replacing business tax with VAT”, and is thus conducive to our relatively clean simulation of the effects of the VAT sharing reform.^[1]

In the literature, there are three ways to estimate the VAT base: the Aggregate National Accounting Account Method,^[2] the Industry Account Method^[3] and the Input and Output Method^[4] (Li and Fang, 2013; Minth, 2007). These three methods theoretically produce the same results, but each of the three approaches has its advantages and disadvantages. The advantage of the input-output method is that it can make full use of input-output and consumption data for goods and services in the economy. To this end, this paper uses China’s input and output data in 2012 to simulate the effects of the reform of “replacing business tax with VAT” and the VAT sharing reform.^[5] First of all, the proportion of the two tax revenues of the provinces in the national two tax revenues before the reform of “replacing business tax with VAT” is calculated, and secondly, the proportion of the VAT revenue of the provinces in the national VAT revenue after the reform of “replacing business tax with VAT” is calculated. The result represents the impact of the reform of “replacing business tax with VAT” on the fiscal capacity of the provinces. Numerical simulation results presented in Figure 3 show that it is not difficult to find that the reform of “replacing business tax with VAT” can cause a rapid rise of the proportion of the VAT revenue of regions with a larger scale of the consumption-based service industry in the national VAT revenue; while the proportion of VAT revenue of regions with a larger scale of production-based service industry in the national VAT revenue can decrease in the course of reform. As the eastern region has a larger scale of the consumption-based service industry, the proportion of VAT revenue in the region such as Guangdong, Zhejiang, Jiangsu and Shanghai has significantly risen.

Figure 3

Simulation of the Effects of Reform with the Input-Output Table 2012

4.1 Comparison of the Equalization Effects of Different Policy Schemes

We compare the equalization effects of different policy schemes from two perspectives: one is to directly compare the changes in the proportion of VAT revenue of different provinces in the national total, to observe the changes in the proportion of different provinces; and the other is to calculate the standard deviation of the proportion of VAT revenue of different provinces in the national total under different sharing schemes, the smaller the standard deviation, the stronger the degree of equalization.

Based on a given central and local share ratio of 50:50, this paper measures the equalization effect of different sharing methods based on the 2017 data, and the numerical results show that considering other VAT distribution principles will help promote the balance of fiscal capacity in the a jurisdiction.^[1]Figure 4 shows the standard deviation of the VAT proportion in the country under different sharing schemes, and it can be found that the introduction of other criteria for the distribution of VAT is conducive to the equalization of fiscal capacity in different jurisdictions and the enhancement of their capacity to supply public services. From the measured results, it is not difficult to find that the degree of VAT revenue inequality is the highest under the principle of the place of production, which also means that compared with other distribution methods, the degree of horizontal fiscal imbalances under this method is the most serious. Distribution according to the principle of the place of consumption can alleviate horizontal fiscal imbalances, and compared with the principle of the place of production, the standard deviation under this principle falls by 0.04. The distribution of VAT revenue according to population size and fiscal capacity can further highlight the equalization effect of distribution. The method of distribution according to fiscal capacity or in line with the “population + consumption + fiscal capacity” multi-principle method can best highlight the equalization effect and improve the current unfair distribution of VAT revenue in China.

Figure 4

Fiscal Equalization Effects of Different Sharing Schemes

4.2 The Choice of the Best Sharing Scheme for VAT

In theory, the residents of a country can avoid the inefficient flow of labor and maximize social welfare when they enjoy equal public services. However, the decentralization system may cause different scales and standards of government public services in different jurisdictions, which could lead to the different public services received by residents. The key to address this problem is to improve the public services supply capacity and realize fiscal equalization in areas with backward fiscal capacity. As far as VAT per se is concerned, if it is levied according to the principle of place of production, local governments would be induced to compete for fiscal capacity and distort the allocation of resources. In practice, how should the optimal allocation method be chosen? In choosing the optimal allocation method of VAT, there are two dimensions to consider. The first is whether the equalization effect of a sharing scheme is conducive to narrowing the regional fiscal gap. The second is an incentive mechanism of government behaviors. Under the principle of production place, the government would try to expand the scale of local productive tax base to maximize budget constraints, and the most direct reactions of the government are to attract investment, increase the scale of revenue by promoting enterprises in the jurisdiction to expand their production. Under the principle of place of consumption and population, to increase fiscal revenue, the government would focus on creating a good consumption environment and be encouraged to be responsible to the residents in its jurisdiction. The principle of fiscal capacity has a great distortion effect on the incentive of government behavior, because places with less fiscal revenue would get larger subsidies from the central government, which may cause the government to race to the bottom.

The VAT sharing reform aims to establish a VAT distribution mechanism combining equalization and government incentive. From the experience of other countries, Germany’s VAT is mainly distributed on the basis of population and fiscal capacity, and Canada has established a system of Harmonized Sales Tax (HST). China’s VAT sharing scheme, on the basis of the principle of tax collection from consumption places, should gradually establish a VAT sharing system that takes into account various factors of consumption, population and fiscal capacity. Figure 5 compares the equalization effect of different VAT sharing schemes, with the highest degree of imbalance under the principle of place of production. From the point of view of equalization effect, the effect of fiscal capacity and “consumption + population + fiscal capacity” is similar, but the distribution of VAT revenue according to the principle of fiscal capacity alone could have a huge distorting effect on local government behavior. It can be found from the figure that the areas with higher VAT revenue get lower share of revenue in this distribution scheme, which makes local governments race to the bottom. The advantage of collecting VAT according to the principle of place of consumption is that it can improve the degree of responsibility of the local government to the residents of the jurisdiction and improve the quality and level of public service supply. Meanwhile, on the basis of the principle of consumption place, the inclusion of comprehensive factors such as population and fiscal capacity can not only promote the equalization of fiscal capacity, but also establish incentives in fiscal revenue distribution from many dimensions. Therefore, the gradual establishment of a “consumption + population + fiscal capacity” combined collection method on the basis of the principle of consumption place is an ideal policy scheme for future VAT sharing in China.

Figure 5

A Comparison of the Equalization Effects of Different Sharing Schemes