Examining the Impact of Fiscal Resources on Anti-Poverty Expenditure: Evidence from China

Abstract

In developing countries, anti-poverty programs are often implemented by local governments. However, due to the limitation of fiscal resources, the amount of anti-poor expenditure by the local government is generally less than what is needed for the poor. In this paper, we investigate whether an increase in the fiscal resources of local government will lead to an increase in anti-poor fiscal expenditure using county-level Chinese data. Using the fixed effect model, we show that local governments will put more fiscal resources into the minimum living standard guarantee (MLSG) system if they receive more intergovernmental transfers from high-level governments, but this effect only exists in urban areas. Moreover, the off-budget fiscal revenue does not affect the anti-poverty expenditure, both in rural and urban areas.

1. Introduction

Economic growth depends critically on local administration. The literature has devoted a lot of attention to the distribution of public finances by local authorities in developing nations [1]. In contrast to wealthy nations, local authorities in developing nations often lack sufficient budgetary resources, forcing them to depend on other sources of funding, such as intergovernmental subsidies from higher levels of government and the unofficial tax system [2,3]. Because of limited resources, local governments in developing countries face trade-off choices when deciding where to use their fiscal revenues. Therefore, knowing which fiscal expenditure category receives more fiscal funds from the local government when they obtain additional fiscal resources is important. In this paper, we make a small step toward answer this question by studying the impact of intergovernmental transfer and off-budget fiscal revenue on the anti-poor expenditure of the local government in China. There are five layers in the Chinese government: central-level government, provincial-level government, prefecture-level (city-level) government, county-level government and township-level government. We refer to the last four layers as the local government.

China is the largest developing country in the world with many people still living in poverty in the period 2008–2009 that we examine in this paper. According to the data from the World Bank, the poverty headcount ratio at USD 1.90 a day (2011 PPP) was 14.9% in China in the year 2008. To guarantee the basic needs of the poor, China launched the national minimum living standard guarantee (MLSG) system (the MLSG system is also known as the Dibao system) from the earlier 1990s. The MLSG system provides a cash transfer to the poor who have an income lower than a certain level. The payments of the MLSG are undertaken by prefecture-level and county-level governments which decide the eligibility thresholds based on the living standard, the income of the household as well as the fiscal resources [4]. However, the eligibility thresholds decided by the local government tend to be lower than the required level due to the lack of fiscal resources of local governments in China [5,6].

Many Chinese citizens were living in poverty in 1949 when the People’s Republic of China (PRC) came into being. Ever since, several actions have been performed to reduce poverty by governments at various levels. The number of poor people in China has dropped substantially from ”730.4 million in 1981 to 18.5 million in 2014” [7]. Intending to eradicate poverty by 2020, the Chinese authorities and the communist party initiated a national effort to combat poverty in 2016. The Chinese government announced success in the war against poverty at the end of 2020, pulling all people out of extreme poverty and so embracing China’s “moderately prosperous society.”

Local government in China relies on the intergovernmental transfer from the central government and they also obtain a large amount of off-budget fiscal revenue from land selling. In 1994, the Chinese government conducted a reform of the fiscal revenue-sharing system that aimed at increasing the fiscal revenue share of the central government. After the reform, the largest two parts of the tax, namely, the value-add tax (VAT) and corporate income tax (CIT), were shared between the local and central governments [8]. The fiscal revenue share of the local government decreased from 77.98% to 44.30% immediately after the reform. However, the fiscal burden was not transferred to the central government accordingly, so the fiscal expenditure of the local government relies on intergovernmental transfer. The fiscal gap of local governments in China remains at a very high level after the 1994 reform. Furthermore, local governments in China obtain large amounts of off-budget revenue that do not show in their budget. The land-selling revenue is the main part of the off-budget revenue [9]. The ratio of land-selling revenue to the local budget revenue reached 43% in the year 2009.

The intergovernmental transfer is an essential part of the local budget revenue in China. There are two different intergovernmental transfers in China, namely, general transfer and special transfer, which correspond to unconditional grants and conditional grants that are usually used in the literature [10]. Local government can use the general transfer income for whatever project they want, but it can only use the special transfer income for the earmarked corresponding project. In this paper, we use the ratio of general transfer to total fiscal revenue as the independent variable because the general transfer does not limit the usage. It represents the fiscal resources that come from the higher-level government. In addition to the intergovernmental grants, land revenue became an indispensable part of fiscal revenue for the local government in China [11]. More importantly, the land revenue is the off-budget revenue of the local government that does not show in the local fiscal budget. Therefore, it may have a very different influence on the behavior of the local government in China since it receives much less monitoring from the central government [12]. So, we use the ratio of land-selling revenue to the total fiscal revenue of the local government as another independent variable to show the effect of the off-budget revenue of the local government.

In this paper, we investigate the problem of whether the increase in the intergovernmental transfer and off-budget revenue of local government in China will lead to an increase in the fiscal expenditure on the MLSG. Figure 1 plots the correlation between the general transfer ratio and MLSG expenditure ratio. It shows a positive correlation between these two variables, both in urban and rural areas. The positive correlation still exists when we look at different regions, namely the east, middle and west areas.

There are three advantages to using MLSG expenditure rather than other fiscal expenditures of the local government as the dependent variable. First, though we can obtain data on the broader categories of local government spending (from the dataset of “Fiscal Statistics for Prefectures, Municipalities and Counties”; see more details in Section 4), they contain too much information that cannot reflect the preference of the local government. For example, the category of expenditure on insurance and unemployment not only includes the welfare of the poor but also the subsidy of administrators that have retired, who are generally not the poor. Jia et al. [13], using the government expenditure on insurance and unemployment data, found no significant effect of fiscal decentralization and fiscal vertical imbalance on the expenditure on insurance and unemployment. They defined vertical imbalance as quite similar with the general transfer ratio used in our analysis, but the results are different as we use different dependent variables and consider rural and urban areas separately. Instead, the MLSG expenditure is a special case of the government expenditure that only contains the funds that transfer to the extremely poor. Though the MLSG program is designed to provide for the basic needs of the poor, it does not mean that all the funds are actually going to the right people. However, in this paper, we do not consider such a mismatch problem; we treat the MLSG spending as the expenditure that the local government wants to transfer to the poor. Therefore, it can provide more accurate information on anti-poor expenditure. Second, the MLSG expenditure data contain both urban and rural areas, so we can investigate whether there is an urban bias of local government or not. Finally, there is evidence showing that the MLSG expenditure does rely on the local fiscal resource due to the limitation of the local government’s fiscal capacity.

We use county-level panel data of local governments in the period 2008–2009 to investigate the effect. The advantage of using panel data is to eliminate the unobserved county fixed effect that may influence both the dependent variables and the independent variables. Our result shows that there is a significant positive effect of intergovernmental transfer on the anti-poor expenditure of the local government in urban areas. However, such an effect is not significant in rural areas. The off-budget revenue has no significant effect on the anti-poor expenditure, both in rural and urban areas.

This paper mainly contributes to three parts of the literature. Firstly, previous studies have investigated the incentives of the local government to provide public goods under the intergovernmental transfer system [12]. Some literature has documented that local governments in China prefer to increase fiscal expenditure that can boost economic growth, rather than improve public services [14]. Our paper contributes to the literature by studying a very specific area of local government expenditures, i.e., the payment for the extremely poor. By using the county-level MSLG data, our study suggests that local governments in China do respond to the anti-poor target of the central government if they obtain more intergovernmental transfers, at least for the urban areas. Secondly, the existing literature has been devoted to understanding the effectiveness of the MLSG system. Most of the studies are trying to figure out whether the MLSG funds are successfully going to the poor people or not [13,15]. In contrast to this approach, we emphasize the fiscal incentive of the local government, i.e., whether the local government has the incentive to increase the total expenditure of the MLSG when they obtain more fiscal resources. Finally, little attention has been paid to the role of off-budget revenue in developing countries, especially its effect on government expenditure. By employing the data of land-selling revenue, which is the main part of the off-budget revenue of local governments in China, our results demonstrate that the local government does not increase its anti-poor expenditure when it collects more off-budget revenue.

The following structure is arranged as follows: the second part reviews the existing literature; the third part describes the policy background of China’s intergovernmental transfer system and the national minimum living standard; the fourth part describes the dataset used in this paper as well as the estimation strategy used in this paper; the fifth part gives the empirical results of our estimation; the last part provides the conclusions and policy recommendations.

2. Literature Review

Social justice and equality are connected to the fight against poverty. According to Walzer [16], having little money prevents a person from fully participating in society. According to Rawls [17], social justice necessitates a system of institutions that maximize the well-being of the least fortunate people in society. According to Arneson [18], “the sufficientarian principle of justice says that institutions and practices should be arranged and actions chosen so that of those people who will ever live, as many as possible reach the sufficient level.” In addition, egalitarians often think it is unjust if one individual is worse off than someone else without their own mistake or decision [19]. This is because poverty is not the choice of poor people, but rather, it has been imposed on them due to the following factors such as war, a lack of economic opportunity and inadequate infrastructure. Consequently, the government’s initiative to combat poverty has serious moral consequences for maintaining a fair and equitable society [20].

It is long known that local governments will increase their public expenditure when they obtain grants from the higher-level government, and this is called the flypaper effect [21]. Developing countries have different institutional arrangements compared with developed countries, the power of taxation of the local government is often restricted by the central government and the ability to collect tax revenue is limited [1]. Given the limited tax power and fiscal resources, it is not surprising that the flypaper effect is also found in developing countries [22,23].

However, since the local government in developing countries generally lacks fiscal resources [24], which of the fiscal expenditure categories does the local government prefer more when it receives fiscal resources? Do local governments want to spend more on poverty alleviation? The answer to this question is inconclusive. One study [25] investigates the effect of intergovernmental fiscal arrangements on poverty alleviation in Viet Nam and points out that both general and specific transfers play an important role in poverty alleviation. The result is consistent with the research of Nursini and Tawakkal [26]. Granado et al. [27] investigate the effect of decentralization on the composition of public expenditures using cross-country data. Their result shows that decentralization increases the share of education and health expenditures. In contrast, Sepulveda and Martinez-Vazquez [28] find that fiscal decentralization has a negative effect on poverty alleviation using cross-country data. Kappeler and Välilä [29] focus on the public investment of European countries, and they find that fiscal federalism increases the fiscal expenditure that is economically productive, such as infrastructure investment. Boret et al. [30] and Adu et al. [31] note that fiscal resources have an asymmetric effect on poverty.

In the case of China, the intergovernmental grants system is part of the fiscal decentralization that draws considerable attention in the literature [32,33]. The effect of fiscal decentralization on the expenditure structure of local government in China is shown in the research of Jia et al. [13], which finds that decentralization increases the expenditure of the local government in general. More specifically, Sanogo [34] and Hussain et al. [35] report that fiscal decentralization tends to increase the local government spending in capital construction, and reduce education expenditure spending and administrative expenditure spending, but does not affect the expenditure on insurance and unemployment subsidies.

The MLSG program in China has drawn a lot of attention from researchers. Most of the research focuses on the effectiveness of the program. Using the China Household Income Project (CHIP) data, Gao et al. [5] estimate MLSG eligibility for the urban population and document that only half of the families that are eligible for MLSG receive insurance from it. Kakwani et al. [20] focus on the targeting accuracy of rural MLSG; they show that most of the poor are not included in the MLSG. Ravallion [6] points out that there is geographic inequality if the anti-poverty program was delegated to the local government due to the budget constraint of the local government. He finds that local governments in China with less fiscal capacity tend to set a lower poverty line. He uses survey data that contain the 35 largest cities in China. The purpose of MLSG is not reached mainly because of the lack of fiscal resources of the local government.

2.1. The Intergovernmental Transfer System in China

China has an institutional arrangement that is different from the developed countries in many aspects. Xu [36] refers to China’s institution as the “regionally decentralized authoritarian regime”. The local government in China is economically decentralized in terms of their fiscal burden, and most of the fiscal expenditure is conducted by local government. Meanwhile, the power of personnel appointment is highly centralized in the higher-level government. Therefore, the local government is more likely to respond to the higher-level government rather than the residents. Moreover, local governments are highly competitive with each other, which has been taken as a key reason that explains the economic miracle China created since 1978 [37,38].

The intergovernmental transfer system was established after the reform of tax-sharing systems in 1994 [39]. In 1994, the central government took 75% of the total value-added tax (VAT) revenue. In 2002, the corporate income tax was equally shared between the local and central government. The value-added tax and corporate income tax are the two most important taxes in China. One year later, the share of corporate income tax that belonged to the central government increased to 60%. However, the devolution of fiscal revenue was not matched by the fiscal expenditure, and local governments still had the responsibility for a large part of fiscal spending, including the infrastructure, public health and education [40]. More importantly, the local government in China did not have the power to decide the tax rate and tax base, which meant it could not lower the tax when receiving the intergovernmental transfer. This led to a consequence of lacking fiscal resources for local governments to support their fiscal expenditure. To meet the need of local governments, the central government chose to build an intergovernmental transfer system, i.e., local governments obtained fiscal grants from the central government to fulfill their fiscal expenditure.

There are two kinds of intergovernmental transfer: one is called “general transfer”, and another is “special transfer”. There are two important differences between general transfer and special transfer. The first one is the allocation of intergovernmental grants. The general transfer is decided by the Ministry of Finance based on some formulas which capture the potential fiscal resources and the fiscal burden of local government. The central government decides the amount of general transfer payment to go to the provincial government, which in turn, decides the payment of the prefecture government and so on. The special transfer depends on the particular project that is delegated to the local government. Such projects are controlled by different ministries of the central government which have the power to decide the assignment of the project across local governments. The second difference is the usage of the transfer income. There is no limitation on how the local government uses the general transfer revenue, but as for the special transfer, the local government must use it for the corresponding project.

In the year 2008, the general transfer accounted for 50.1% of the total transfer [41]. The largest part of the general transfer is called the balance transfer. (Before 2009, the general transfer was called fiscal transfer and the balance was called general transfer. The data we use are the transfer income of the year 2008–2009; however, we use the new name after 2009). It was designed to equalize the fiscal resources across the regions. The amount of balance transfer is decided by some fixed rules that capture the fiscal capacity of the local government. Typically, local governments with a low fiscal income can obtain more of a balance transfer income. Other parts of general transfers are transferred for ethnic minorities’ areas, for ecological function areas, for the basic fiscal capacity and so on. However, some researchers point out that the general transfer does not achieve its equalization purpose [42].

2.2. The Land-Selling Revenue

In addition to the general public budget revenue, the local government in China also obtains government funds and collects revenue from state-owned firms [43]. The land-selling revenue was the main part of off-budget government funds for local governments in the period 2008–2009 which is considered in this paper.

In China, while the land in the urban area is state-owned, the land in the rural area is collective-owned. In 1998, a new land management law was passed which set a restriction on the usage of collective-owned rural lands. It ruled that collective-owned rural lands cannot be used for non-agricultural construction [44]. Only the local government can transfer the collective-owned rural lands into the state-owned urban lands [45]. According to the law, the compensation for land acquisition was based on the agricultural output of the land, but the sales revenue of land depended on the value of industrial and commercial output, which was far more than the value of agricultural goods. The land management law that first passed in 1987 ruled that the compensation fee was set to no more than twenty times of the average output in the previous three years. In 1998, the compensation fee was changed by thirty times of the average output in the previous three years. Therefore, the local government can convert the agricultural-used rural land into non-agricultural construction-used land at a very low cost and acquire a large land conveyance fee.

Different from the valued-add tax, the land conveyance fee is fully retained by the local government and does not show in the general public budget of the government (the land revenue is part of the “government funds”, with no detailed public information on income and expenditure), which means the use of the land conveyance fee is a relative lack of supervision compared to the revenue from tax or higher-level government. Little is known about the detailed information on the expenditure structure of the land conveyance fee. Local government has a strong incentive to increase the land conveyance fee since it does not need to share with the central government [46,47]. With the acceleration of urbanization, the land conveyance fee keeps rising and has become an indispensable part of fiscal resources for local government. In 2009, land conveyance fees reached a level equal to 43% of the local within-budget fiscal revenue [48].

2.3. The National Minimum Living Standard Guarantee System of China

The minimum living standard guarantee (MLSG) system is an unconditional targeted cash transfer program aimed at providing basic needs for the poor. The Chinese government began to build the MLSG in the earlier 1990s in some cities. In 1997, the State Council officially announced that the MLSG would be conducted nationwide in the urban area. The original purpose of the MLSG was to provide subsistence for the increasing amount of urban poor who lost their jobs and social security during the economic reform in the 1990s [5]. During that time, China’s government decided to let many small state-owned firms (SOEs) go broke or be privatized due to the overall inefficiency of the SOEs [49]. This led to a rapid increase in laid-off workers in urban areas.

The eligibility of the urban MLSG is decided by the household register system, known as the hukou system, which sets barriers to interjurisdictional mobility. The transfer of rural hukou to urban hukou is also restricted. People are supposed to live and die in the same locality. Only people with urban hukou can apply for the urban MLSG. Poor families who have a per capita income less than the minimum living standard line are eligible to apply for the program. Families who successfully acquire the qualification will receive money equal to the gap between the income and minimum living standard line. The person who wants to apply for the MLSG must go to their registered residence, so it will not attract the poor from other counties if local governments increase their payment.

Before 2007, some cities in China began to establish the rural MLSG to provide insurance for the rural poor. In 2007, China officially decided to expand the MLSG to whole rural areas. After several years of efforts, now the MLSG serves as a last resort for providing basic social security for all poor people. The central government delegates the power of setting the minimum living standard line to local governments (prefecture and county level) which have information advantages over the central government. The minimum living standard line is required to satisfy the basic needs of the poor and should be adjusted yearly according to the local price of necessities and the living standard of the residents. However, in reality, the minimum living standard line may not satisfy the requirement due to the lack of fiscal capacity of the local government.

3. Data and Empirical Strategy

3.1. Data and Summary Statistics

The data used in this study come from four different sources. The first is the Fiscal Statistics for Prefectures, Municipalities and Counties which was released by the Ministry of Finance. The data are available for the years 1993 to 2009 and provide detailed public finance data for all the prefecture-level and county-level governments, including fiscal revenues, expenditures and intergovernmental transfers. We obtained the data of general transfer, special transfer, fiscal revenues and fiscal expenditures from this dataset for the year 2008–2009. Unfortunately, the 2008–2009 data do not contain the detailed composition of fiscal expenditures. Though the dataset provides the main categories of fiscal expenditure before 2007, as we mentioned in Section 1, the information it contains is not clear. For example, the expenditure on insurance and unemployment not only includes the welfare of the poor but also the subsidy of administrators that have retired. The total MLSG payment only accounted for 11.86% of the total fiscal expenditure on insurance and unemployment in 2008 (total fiscal expenditure on insurance and unemployment data comes from the China Statistical Yearbook; the total MLSG payment data come from the China Civil Affairs’ Statistical Yearbooks released by the Ministry of Civil Affairs). To deal with this problem, we obtained the MLSG expenditure data from the Ministry of Civil Affairs. The data provide the county-level seasonal data on the number of eligible people, the expenditure of the MLSG and the minimum living standard line for urban from 2007 to 2016. For the rural area, we could obtain the county-level seasonal data of the number of eligible people and the expenditure of the MLSG for the year 2008–2016, but we could not obtain the data of the minimum living standard line (the prefecture-level data of the minimum living standard line were available for both urban and rural areas from 2008 to 2019). The third data source is the China Yearbook of Land Resource in which we could find the land conveyance data. These data are prefecture-level data that document the revenue of land sales for the years 2004–2016. Other county-level data such as GDP, population, banking credit and secondary industrial output were obtained from the Statistical Yearbooks of China.

We chose the sample period of 2008–2009 because this is the period that had both intergovernmental transfer and the MLSG data. After combining data from different sources and deleting the missing variables, we acquired a balanced panel dataset containing 1224 counties. There are two advantages to using the county-level data: the first is the data provide more examples for our analysis; the second is that county-level government is the lowest level of government that has the authority to decide the allocation of MLSG funds. However, because the land conveyance data were only available at the prefecture level, we also used the prefecture-level data when we considered the land revenue. The summary statistics are shown in

Table 1. Summary statistics of variables (2008–2009).

	(1)	(2)	(3)	(4)	(5)
Variables	N	Mean	sd	Min	Max
General transfer ratio (% total fiscal revenue)	2448	0.362	0.129	−0.0049	0.731
Special transfer ratio (% of total fiscal revenue)	2448	0.310	0.114	0.0397	0.869
Urban MLSG expenditure ratio (% of total fiscal expenditure)	2448	0.0127	0.0078	0.0001	0.0910
Rural MLSG expenditure ratio (% of total fiscal expenditure)	2448	0.0137	0.0104	0.0004	0.314
Local fiscal revenue	2448	24,719	27,735	167	283,823
GDP per capita	2448	14,812	11,056	1881	166,683
Urbanization rate	2448	0.183	0.111	0.0113	0.941
Credit ratio (% of GDP)	2448	0.416	0.265	0.0016	3.707
Average salary	2448	22,428	5406	1410	52,792

.

3.2. Estimation Strategy

We examined the effect of the intergovernmental transfer income of local government on the anti-poverty expenditure by regressing the MLSG expenditure ratio on the general transfer ratio and land-selling revenue ratio. The MLSG expenditure ratio is the ratio of MLSG expenditure to total fiscal spending. A higher ratio means local government allocates relatively more fiscal resources to the anti-poverty program. The general transfer ratio is the ratio of general transfer to the local fiscal total revenue. It is the measure of whether the local government acquires relatively more fiscal resources from the central government. The land-selling revenue ratio is the ratio of land-selling revenue to the local fiscal total revenue. It represents the relative size of the off-budget revenue of the local government.

There are two reasons why we used general transfer instead of special transfer. The first reason is the lack of detailed data on special transfers. As mentioned above, the special transfer consists of many specific projects, and the corresponding funds can only be used for this particular project. Surely some projects provide funds for poverty alleviation, but we could not obtain detailed data on such projects; we only had the total special transfer data. One may ascertain that if most of the expenditure of MLSG comes from the special transfer, then it may not be convenient to use general transfer as the independent variable.

The other reason why we used the general transfer ratio is that the general transfer can be freely used by the local government. The special transfer also helps the local government to fulfill the fiscal gap, but the government can only use such transfer income for special projects. Since our interest is whether local government will choose to increase the spending on the anti-poverty program, it was better to use unconditional grants, i.e., general transfer as an independent variable.

Another question regards whether the local government lacks fiscal resources to fulfill the MLSG expenditure. Though we do not have direct evidence, we do have some information that can give us some insight into this question. The local government of all-levels should increase the expenditure for MLSG; local governments should not reduce the expenditure of MLSG given that the central government increases the input in MLSG spending.

From the document, we can see that the central government commands local governments to expand their fiscal expenditure for MLSG, even though the central government increases the special transfer to the local governments. Local governments do face the trade-off of allocating their fiscal resources between MLSG and other areas.

We used the fixed effect model to estimate the effect of the general transfer ratio on the MLSG expenditure ratio. We specified our regression equation as:

$$\begin{array}{l}MLS{G}_{u,it}=\alpha +\beta Transfe{r}_{it}+\gamma X_{it}+{\tau }_t+{\delta }_i+{ϵ}_{it}\\ MLS{G}_{r,it}=\alpha +\beta Transfe{r}_{it}+\gamma X_{it}+{\tau }_t+{\delta }_i+{ϵ}_{it}\end{array}$$

where $MLS{G}_{u,it},MLS{G}_{r,it}$ denote the MLSG expenditure ratio for urban and rural of the city i in time t, respectively. $Transfe{r}_{it}$ is the key independent variable which means the ratio of general transfer to the local total fiscal revenue of city i in time t. $X_{it}$ is the vector of control variables including the special transfer ratio (the ratio of special transfer to total fiscal revenue), GDP per capita, urbanization rate, the average level of salary, the local fiscal revenue which is the share of fiscal revenue that directly goes to local government and the credit GDP ratio which represents the financial deepening. In the following regression, we included the GDP per capita and the average level of salary in log forms for convenience. ${\tau }_t,{\delta }_j$ represent the time effect and the fixed effect. Using a fixed effect model and controlling the time effect, we could eliminate the influence of the variables that are different between counties but do not change with time and control for the variables that have the same effect on all counties but change with time. In the following regression, standard errors were clustered at the county level.

4. Empirical Results

4.1. Basic Result

The result of regression for urban and rural MLSG is shown in

Table 2. General transfer ratio and urban MLSG payment ratio.

Dependent Variable: Urban MLSG Ratio
Variables	OLS	FE
	(1)	(2)	(3)	(4)
General Transfer Ratio	0.0173 ***	0.0117 ***	0.0071 ***	0.0071 ***
	(0.002)	(0.002)	(0.003)	(0.003)
Special Transfer Ratio	0.0085 ***		−0.0040	−0.0036
	(0.001)		(0.003)	(0.003)
GDP Per Capita (log)	−0.0014 ***		0.0001	0.0002
	(0.000)		(0.001)	(0.001)
Urbanization Rate	0.0154 ***		0.0036 *	0.0036 *
	(0.002)		(0.002)	(0.002)
Credit GDP Ratio	−0.0014 **		0.0004	0.0005
	(0.001)		(0.001)	(0.001)
Average Salary (log)	−0.0038 ***		−0.0014	−0.0012
	(0.001)		(0.001)	(0.001)
Constant	0.0519 ***	0.0085 ***	0.0240 ***	0.0204
	(0.008)	(0.001)	(0.009)	(0.015)
Observations	2448	2448	2448	2448
R-squared	0.214	0.033	0.040	0.040
County FE	No	YES	YES	YES
Year FE	No	No	No	YES

Note: Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

and

Table 3. General transfer ratio and rural MLSG payment ratio.

Dependent Variable: Rural MLSG Ratio
Variables	OLS	FE
	(1)	(2)	(3)	(4)
General Transfer Ratio	0.0120 ***	−0.0213 ***	0.0074	0.0092
	(0.002)	(0.004)	(0.006)	(0.006)
Special Transfer Ratio	0.0012		0.0130 **	−0.0012
	(0.002)		(0.006)	(0.006)
GDP Per Capita (log)	−0.0049 ***		0.0040 ***	0.0001
	(0.001)		(0.001)	(0.001)
Urbanization Rate	−0.0112 ***		0.0051	0.0048
	(0.002)		(0.005)	(0.005)
Credit GDP Ratio	−0.0016 ***		0.0045	−0.0004
	(0.001)		(0.003)	(0.002)
Average Salary (log)	0.0048 ***		0.0077 ***	−0.0037
	(0.001)		(0.003)	(0.003)
Constant	0.0089	0.0214 ***	−0.1102 ***	0.0443 *
	(0.008)	(0.001)	(0.022)	(0.023)
Observations	2448	2448	2448	2448
R-squared	0.176	0.015	0.055	0.091
County FE	No	YES	YES	YES
Year FE	No	No	No	YES

Note: Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

. Column (1) in Table 2 gives the result of basic OLS regression. As a comparison, column (3) shows the result of the fixed effect model. We can see that the effect under both models is significant at the 1% level, but once we control for the fixed effect, the effect of general transfer on MLSG is 0.0071 which is less than the case in OLS regression. In the fixed effect model, only the urbanization rate has a positive effect on the urban MLSG expenditure, and other control variables do not have a significant effect, which is quite different from the OLS model. In column (4), we also control the time effect, and it does not change much of our results.

However, when it comes to the rural MLSG ratio, the results are different from the case of urban areas as can be seen in Table 3. Though the estimator from OLS is significantly positive, the fixed effect model gives a positive but not significant estimation of the effect of the general transfer on MLSG expenditure. Thus, based on our result, we find that the anti-poor expenditure of the local government in China is urban-biased. Local governments will spend more on urban MLSG when they obtain more general transfers from the high-level government. This reflects an urban-biased policy that has been observed in the literature [50].

4.2. Consider Different Regions

Since GDP per capita is only the proxy of the poverty level, we also use another way to check our results. We include the interaction term of general transfer ratio and a dummy variable that represent whether one county is located in the east, central or west regions to see if there will be a difference between different regions. Traditionally, China divides all of its 34 provinces into three parts: the east coastal areas, the western areas and the middle areas. The east areas are the richest provinces, and the west are the most underdeveloped provinces. Therefore, if the rich counties already have enough fiscal revenue to finance their anti-poverty expenditure, then there should be a different effect between rich and poor regions. Furthermore, the allocation of the general transfer is highly influenced by the region in which the county is located. Therefore, it is worth seeing whether the effect of the general transfer on MLSG expenditure depends on the location in different regions.

Columns (1) and (3) of

Table 4. General transfer ratio and MLSG payment ratio considering different regions.

Fixed Effect Model
Variables	Urban		Rural
	(1)	(2)	(3)	(4)
General Transfer Ratio	0.0089 **	0.0253 ***	0.0080 *	0.0101
	(0.004)	(0.008)	(0.005)	(0.011)
General Transfer Ratio × East	−0.0078 **		0.0155 **
	(0.003)		(0.007)
General Transfer Ratio × West	0.0032		−0.0134 **
	(0.004)		(0.007)
Special Transfer Ratio	−0.0029	−0.0030	−0.0032	−0.0012
	(0.003)	(0.003)	(0.006)	(0.005)
GDP Per Capita (log)	0.0003	0.0003	−0.0003	0.0001
	(0.001)	(0.001)	(0.001)	(0.001)
Urbanization Rate	0.0034 *	0.0035 *	0.0055	0.0048
	(0.002)	(0.002)	(0.005)	(0.005)
Credit GDP Ratio	0.0006	0.0005	−0.0005	−0.0004
	(0.001)	(0.001)	(0.002)	(0.002)
Average Salary (log)	−0.0012	−0.0012	−0.0037	−0.0037
	(0.001)	(0.001)	(0.003)	(0.003)
Constant	0.0185	0.0196	0.0490 **	0.0443 *
	(0.015)	(0.015)	(0.023)	(0.023)
Observations	2448	2448	2448	2448
R-squared	0.045	0.045	0.096	0.091
County FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES

Note: Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

show the result of including the region dummy variables. The middle region is chosen as a benchmark. The effect of general transfer on urban MLSG expenditure has no difference between the middle and west region, but is smaller in the rich east area: only 0.0021 compared with 0.0089 for the middle and west areas. This makes sense because the counties in the east area are generally richer than their counterparts in the middle and west, which means they have more fiscal resources and fewer poor people and do not rely much on the general transfer. Nevertheless, after controlling for the difference between regions, the effect is still significant in urban areas. Interestingly, when it comes to rural MLSG expenditure, adding the interaction term makes the estimator significant for all three regions, though the effect is not significant as a whole. Different from the urban case, the effect of the general transfer on rural MLSG expenditure is highest in the east region and lowest in the western region. This seems to imply that with limited fiscal resources, local government will consider the urban MLSG expenditure more of a priority than the rural MLSG, so rich areas tend to use more fiscal resources to finance the MLSG expenditure of rural areas.

4.3. The Effect of Off-Budget Revenue

As mentioned above, we only have data on land-selling revenue at the city level. The regression in

Table 5. Land revenue ratio and MLSG payment ratio.

Variables	Urban		Rural
	(1)	(2)	(3)	(4)
Land Revenue Ratio	0.0006	0.0015	−0.0004	0.0001
	(0.001)	(0.003)	(0.001)	(0.002)
Land Revenue Ratio × East		−0.0007		−0.0012
		(0.003)		(0.002)
Land Revenue Ratio × West		−0.0046		0.0080
		(0.006)		(0.006)
Special Transfer Ratio	−0.0004	−0.0009	0.0042	0.0061
	(0.006)	(0.006)	(0.005)	(0.006)
GDP Per Capita (log)	−0.0053 **	−0.0051 **	−0.0019	−0.0024
	(0.002)	(0.003)	(0.003)	(0.003)
Secondary Industry Ratio	−0.0096	−0.0094	0.0005	0.0001
	(0.008)	(0.008)	(0.007)	(0.007)
Credit GDP Ratio	−0.0018	−0.0017	−0.0061 ***	−0.0061 ***
	(0.001)	(0.001)	(0.002)	(0.002)
Average Salary (log)	0.0046	0.0047 *	0.0041	0.0040
	(0.003)	(0.003)	(0.003)	(0.003)
Constant	0.0242	0.0212	−0.0127	−0.0074
	(0.033)	(0.031)	(0.034)	(0.036)
Observations	494	494	494	494
R-squared	0.129	0.134	0.257	0.275
City FE	YES	YES	YES	YES
Year FE	YES	YES	YES	YES

Note: Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1.

uses city-level panel data containing 247 cities in the year 2008–2009. The result shows that the off-budget revenue of the local government has no significant effect on the MLSG payment, both in urban and rural areas. The effect also has no differences between the three different regions. The finding is inconsistent with Tang et al. [51], who noted that the land revenue of local government increases the expenditure on capital construction projects while decreasing the public services’ expenditure.

4.4. Results Discussion

In general, our findings suggest that the increased transfers of financial resources to local government can increase anti-poverty expenditures in China. Local governments can play a significant role in developing a fair and equitable society. The provision of financial resources to local administration can help to locate the underprivileged class of society more easily because the local leaders are normally aware of their surroundings and the condition of the poor people. As a result, they will divert the financial sources to the most needed faction of society with greater accuracy and efficiency, which will significantly contribute to uplifting the status of poor people. Our findings are consistent with Digdowiseiso [52], Hussain et al. [35] and Siburian [23], which conclude that fiscal decentralization helps in eradicating poverty.

Since restructuring and openness, China has significantly achieved results in poverty reduction levels. China’s poverty reduction phase involves overcoming obstacles and hurdles [5]. The communist party of China recognized poverty reduction as one of the three key fights that must be waged to finish the construction of a prosperous society in all aspects during its 19th National Conference. It has made significant preparations and offered a plan of action for winning the struggle against poverty. The process of eradicating poverty in 2018–2020 serves as the set of actions for winning the anti-poverty war in China. The amount of special financial assistance allotted by the centralized administration for eradicating poverty in 2018 was CNY 106.09 billion, a CNY 20 billion increase over 2017. As a result, local governments will be in a better position to expedite the execution of their budgets and wage a successful battle against poverty. Studies of the effects of fiscal input on anti-poverty are thus helpful in hastening the speed of eradicating poverty and prompting recognition of the interplay between the authorities, business and society, in addition to the links between special, industrial and social poverty reduction.

Our results are also supported by other earlier investigations. For instance, an analysis of cross-province data in China by Cui et al. [53] reveals that financial regulations may raise the level of wealth of the underprivileged by energizing the local economy and transferring earnings. Nevertheless, the impact of reducing poverty may be countered by financial instability. Some claim that financial regulations could hinder the fight against poverty. Although the impact is not immediately apparent, Huang et al. [54] demonstrate that financial growth in rural regions supports poverty alleviation in the near-term. Long-term financial progress may prevent the alleviation of poverty (Tang et al. (2019)).

China has lifted more than 850 million people out of poverty, which accounts for more than 70% of global poverty reduction. In terms of fiscal resources, the Chinese government has made significant investments in anti-poverty programs [55]. In 2020, China’s central government allocated CNY 140 billion (USD 21.5 billion) to poverty alleviation efforts, an increase of 40.4% from the previous year. This funding has been used to support a range of programs, including poverty relief, education, healthcare and employment training. One of China’s most significant anti-poverty policies is its targeted poverty alleviation program, which identifies impoverished areas and households and provides targeted assistance to lift them out of poverty. The program includes measures such as improving infrastructure, providing financial assistance and offering vocational training.

The European Union (EU) also has a range of policies aimed at reducing poverty. The EU’s main instrument for poverty reduction is the European Social Fund (ESF), which provides funding for education, training and employment programs [56]. In addition to the ESF, the EU has implemented a range of other policies aimed at reducing poverty and promoting social inclusion. These include the European Regional Development Fund, which supports the development of disadvantaged regions, and the European Fund for Aid to the Most Deprived, which provides food and basic assistance to people in need. One of the EU’s most significant anti-poverty policies is the Europe 2020 Strategy, which sets out a range of targets for reducing poverty and social exclusion across the EU [57]. The strategy includes targets such as reducing the number of people at risk of poverty or social exclusion by at least 20 million by 2020.

5. Conclusions and Policy Implications

Most of the anti-poor programs in developing countries are conducted by local governments which typically have a larger fiscal gap than developed countries. Whether the local government will increase the anti-poor expenditure when they obtain additional fiscal resources remains a question in the literature. This paper uses the 2008–2009 panel data of local governments in China to investigate the effect of intergovernmental transfer on anti-poor expenditure.

Our results show that local governments in China tend to spend more on the MLSG expenditure when they acquire additional intergovernmental transfer income from the higher-level government, but this effect is significant only for urban areas. The effect does not exist in rural areas, which reflects the urban-biased policy that is well documented in the literature. We also show that the off-budget fiscal revenue of local government does not affect the anti-poor expenditure. The result suggested that it is important for the policymaker to notice the incentive of local government and distinguish the different fiscal sources of government revenue, since they have different effects on the usage of fiscal expenditure. In the case of China, based on the results of this paper, how to increase the incentives of local government to take more consideration of rural areas and how to transfer the off-budget fiscal revenue into the formal fiscal budget are important aspects for future reform.

This article formulates the following policy proposals in conjunction with the aforementioned study results. First, we should increase spending on infrastructure development and agriculture welfare services by transferring funds to the local governments because the local leaders are much more aware of the actual situation of the deprived people and areas, and they can spend the funds with more accuracy and can expedite the fund transfer to address the attributes of the poor, inadequate transportation and distribution systems and subpar circumstances for living. There is a need to increase the social and economic welfare of rural areas, and the government should adopt appropriate policies for land revenue. Second, in order to address the lack of funding and challenges associated with poverty reduction in impoverished remote regions, we will energetically establish unique poverty reduction loan programs and work-relief initiatives. This can be accomplished by encouraging the transfer of resources to local governments, which would be crucial in raising the income levels of the poor. The government should increase financial support to rural areas with the objective to combat poverty. In this regard, more investment is required for rural sector infrastructure development. Government investment is required in network infrastructure in rural areas, which includes the construction of roads, and installation of telephone and Internet systems. Roads will connect farmers to the markets, and mobile networks and Internet will accelerate the application of ICT in rural areas. The government should increase expenditures on rural health insurance that enable poor residents to attain better health services. Lastly, we should prioritize the fiscal policy of transferring the money to the local governments of deprived regions to reduce poverty when developing financial policies to combat high poverty levels in underdeveloped regions. While continuing to execute fiscal poverty reduction measures, we should concentrate on boosting financial help to local governments to alleviate poverty in disadvantaged regions. When poverty is reduced to the lowest, i.e., when poverty is complete, we must increase financial assistance to eradicate poverty despite retaining the financial reducing poverty strategy.

The policy of ”targeted poverty alleviation” has been one of the main cornerstones of the attempts to reduce poverty under Xi’s presidency. China has developed a nationwide registration process for low-income families for this reason, allowing local authorities to collect information from each individual, home and town. In addition, President Xi has emphasized the need to adhere to a five-batch plan centered on industrialization, resettlement, eco-compensation, education and social protection in this focused plan to reduce poverty. The fight against poverty has also benefited from fiscal measures such as tax reductions and corporate incentives. To promote the growth of small enterprises, the authorities have waived income taxes on them and offered tax advantages for companies who invest in underdeveloped regions.

However, the distribution and administration of finances still have a lot to be improved to increase the efficacy of China’s anti-poverty initiatives. Although local governments have the authority to approve the distribution of financial help, their choices are often arbitrary. As a consequence, some low-income families are left without financial assistance, while those with more means are given more financial aid. In certain places, attempts to reduce poverty are hampered by excessively rigid bureaucratic procedures.

On the other hand, in the European Union, social security initiatives and redistributionist tax reforms work together to fund anti-poverty efforts. Through minimum wage laws and collective bargaining agreements, the EU is very focused on eliminating income inequality. Additionally, the EU has an integration policy framework that tries to lessen social and economic inequalities across the various member states. Nevertheless, the EU’s budgetary structure may make it difficult for member countries to implement comprehensive anti-poverty measures. Since the “Stability and Growth Pact” restricts budget deficits at 3% of GDP, it may be challenging for governments to boost expenditures on social safety nets during recessions. Furthermore, EU member states may be less able to adapt the economic strategy to their unique circumstances due to the EU’s “One Size Fits All” approach.

In terms of fiscal resources, China appears to have a larger budget for anti-poverty programs than the EU. However, it is worth noting that the EU’s budget is spread across a range of member states, each with their own poverty reduction policies. Both China and the EU have implemented targeted poverty alleviation programs, but there are differences in their approaches. China’s targeted program is focused on identifying impoverished areas and households and providing targeted assistance, while the EU’s approach is more focused on education, training and employment. Overall, while there are some differences in approach, both China and the EU have made significant efforts to reduce poverty through their policies and programs.

Despite the significant contribution of the study, this study has various limitations. The data used in the analysis are old, and future studies must include updated data, i.e., up to 2020 or 2021. Moreover, future studies must also update the range of data and include the latest data, i.e., from 2012 to 2020, because using too-old data does not provide a clear picture in the context of said relationship. The current analysis gathered data from 2008 to 2009, which did not include the impacts of COVID-19 while investigating said relationship. Hence, future studies must update the data and capture the impact of the COVID-19 pandemic. Further, in order to make the analysis more interesting, future studies should try to gather data from the past century and compare them with the most recent data. Last but not least, the structure of the local administration is different in each province, and their impact on anti-poverty campaigns could be different. Therefore, future studies should perform an analysis by gathering data from different provinces and Chinese regions, i.e., east, middle, west, etc.