Inequality and Fiscal Policy
Chapter

Chapter 19. Growing (Un)equal: Fiscal Policy and Income Inequality in China and BRIC+

Author(s):
Benedict Clements, Ruud Mooij, Sanjeev Gupta, and Michael Keen
Published Date:
September 2015
Share
  • ShareShare
Show Summary Details
Author(s)
Serhan Cevik and Carolina Correa-Caro 

Introduction

China provides an important case for analyzing the long-term evolution of household income inequality. Since the first wave of economic liberalization in the late 1970s, China has grown at an astonishing rate of almost 10 percent per year, raising 660 million people out of poverty. Per capita income increased from $320 in 1980 to about $5,500 in 2012, and the number of people living on less than $1.25 a day declined from 85 percent of the population in 1980 to 11 percent by 2012, according to the World Bank’s World Development Indicators database. But the fruits of the transition from a system of centrally planned socialism to a market-oriented economy are not being widely shared across the society. Income inequality—as measured by the Gini coefficient for pretax market income—has exhibited an increasing trend from 0.28 in 1980 to 0.44 in 2000 and 0.52 by 2013. Income inequality also varies significantly within China at the regional level. This widening in the gap between rich and poor demonstrates China’s transition from a relatively egalitarian society to one of the most unequal countries in the world.1

The sharp increase in income inequality since 1985 appears to be a result of China’s investment-and export-led development model.2 The growth incidence curve—the extent to which each quintile of households benefits from growth in real terms—shows a cumulative increase of 331 percent for the lowest income quintile between 1980 and 2012, but 1,042 percent for the highest income quintile. As a result, the top quintile now captures 47 percent of total income (up from 38 percent in 1980), whereas the lowest quintile accounts for only 4.7 percent (down from 8.7 percent).3 In other words, China’s widening income inequality is largely a reflection of faster income growth among the rich rather than stagnant living standards among the poor. Reforms to the Chinese economy (gaige kaifang) started in 1978 by decollectivizing agricultural land and allocating it to individual households and then expanded in the early 1980s to industrial development in coastal urban areas with greater openness to international trade and finance. This reform strategy ultimately aimed for high aggregate growth rates at the expense of an increase in income inequality.4 However, mounting evidence suggests that income disparities become detrimental to economic growth in the long term, with significant social consequences, especially in a country like China, which is aiming to move beyond its “middle-income” status. Berg and Ostry (2011), among others, show that income inequality is a key determinant of the pace and sustainability of economic growth, even after taking into account other economic and institutional factors.

Income distribution depends on a complex array of factors, including the design of fiscal policy instruments. According to the Standardized World Income Inequality Database (SWIID), the market-income Gini coefficient (before taxes and transfers) in China increased by 82 percent between 1980 and 2013, whereas the net Gini coefficient (after taxes and transfers) increased by 90 percent (Figure 19.1). This difference indicates that the redistributive impact of fiscal policy is eroding. Fiscal redistribution—defined as the difference between market-income and net Gini coefficients—amounted to an average of 1.7 Gini-index points in the 1990s, but it turned negative during the period 2000–13, averaging -1.1 Gini-index points. Fiscal policy, however, can matter not just for redistribution, but also for market-income inequality. Redistribution is, of course, a function of the level and composition of taxation and spending as well as their distribution across income groups. The distribution of taxes in China, for example, remains regressive, with taxes accounting for 10.8 percent of annual income among the bottom decile of households (and 13.3 percent among the bottom 5 percent) compared with 8.7 percent among the top decile of households.

Figure 19.1Income Inequality and Fiscal Redistribution in China

Sources: Standardized World Income Inequality Database (panel 1); OECD Economic Surveys, China, 2013 (panel 2); CEIC China database (panel 3); World Bank, World Development Indicators database (panel 4); National Bureau of Statistics (panel 5); Standardized World Income Inequality Database and authors’ calculations (panel 6).

The aim of this chapter is to isolate the distributional effects of fiscal policy in China during the period 1980–2013. The empirical objective is to identify the proximate determinants of income inequality, with a focus on the distributional effects of fiscal policy. From an econometric point of view, there are two significant challenges. First, potential endogeneity (or reverse causality) between inequality and growth is a problem. Second, time-series analysis based on a small sample may degrade the quality of the estimations and result in misleading conclusions. To deal with the potential endogeneity of economic growth to income inequality, the analysis uses the number of international tourist arrivals as an instrumental variable (IV) for real GDP per capita and uses the IV estimation via the two-stage least squares and the generalized method of moments (GMM) estimators. To overcome the data constraint of a time-series approach in a single-country case, this investigation undertakes a panel data analysis of income inequality for BRIC+ countries (Brazil, Russia, India, and China, plus 30 other emerging market economies) using the IV-GMM estimator. This approach allows for a more vigorous empirical analysis, including a broader set of explanatory variables.

The empirical results support the hypothesis of an inverted U-shaped relationship between income inequality and growth. For China, government spending and taxation are found to have opposing effects on income inequality in the short term as well as in the long term. Whereas government spending appears to be associated with worsening income inequality, taxation improves the distribution of household income. The results of the panel data analysis of the BRIC+ countries are broadly consistent with the China-specific findings, with one important exception. Both government spending and taxation have the desired redistributive effect, but it is statistically insignificant. Although the redistributive impact of fiscal policy in China appears to be stronger, the combined effect is still not sufficient to compensate for the adverse impact of other influential factors identified in the analysis.

Appropriate fiscal redistribution can bring about balanced and sustainable growth by reducing net income inequality.5 The empirical findings presented in this chapter have important policy implications. Fiscal policy in China and in BRIC+ countries can be redesigned to foster inclusive growth and to reverse the pattern of widening income inequality without undermining fiscal sustainability and causing distortions and efficiency losses. First, the tax base needs to be broadened and the tax system made more progressive with a shift from indirect to direct taxation to help narrow income inequality. Second, there is great scope for improving progressivity through well-targeted spending programs that champion greater access for the poor to education, health care, and other social services, particularly in rural areas, which account for more than 95 percent of poor households in China.

The remainder of this chapter is organized as follows: First, China’s experience is put in a comparative perspective. A brief overview of the literature on income inequality is then provided, followed by an outline of China’s fiscal policy space. The data sources and the salient features of this chapter’s empirical strategy are then described. The econometric results are presented, followed by concluding remarks that focus on broad fiscal policy implications.

China’s Experience in the Global Context

How does China’s experience look from a comparative perspective? Across the world, intracountry income inequality has widened since the 1980s to levels unprecedented in the postwar period, with some exceptions in Latin America and sub-Saharan Africa (Figure 19.2). The gap between rich and poor households within countries has widened despite a significant degree of convergence in per capita income levels between countries. The unweighted world average market Gini coefficient increased from 0.41 in the 1980s to 0.43 by 2007 and, after the recent global crisis, to 0.45 in 2013. Even after accounting for taxes and social transfers, income inequality followed a similar widening trend, with the unweighted “world average” net Gini coefficient increasing from 0.36 in the 1980s to 0.38 by 2013.

Figure 19.2China and the Rest of the World

Sources: Standardized World Income Inequality Database and authors’ calculations.

1Or latest data point available between 2009 and 2013.

These averages, however, mask significant differences in inequality across countries and over time. They reflect country-specific demographic, institutional, and economic characteristics and varying degrees of progressivity in taxation and expenditure policies. At one end of the spectrum are countries such as Bulgaria and Belgium with traditionally low levels of income inequality. At the other end are countries such as Kenya and Indonesia with income inequality significantly above the average. Within this global context, although having achieved an exceptionally high rate of economic growth and remarkable progress in poverty reduction, China has recorded a significant deterioration in income inequality, with the market and net Gini coefficients rising from an average of 0.30 and 0.29 in the 1980s, respectively, to 0.52 and 0.53 by 2013. Similarly, the income share of the top 1 percent of households in China increased from 2.8 percent in 1980 to 4.9 percent in 2009. This ratio is relatively low compared with that in many other countries (such as 7.2 percent in Korea and 19 percent in the United States), but it is still significantly higher than the total income share of the lowest quintile of households in China.

Fiscal policy has contributed to changes in income inequality in most countries. On average, the decrease in income inequality brought about by tax and transfer policies was greater in economies with higher inequality of pretax income. The redistributive effect of fiscal policy, as measured by the difference between market and net Gini coefficients, diminished from an unweighted world average of 7.1 Gini-index points in the 1980s to 6.3 Gini-index points in the 1990s, but recovered to an average of 6.9 Gini-index points in the 2000s. Against this global background, fiscal policy in China has been much less effective at decreasing income inequality. The extent of fiscal redistribution declined from an average of 1 Gini-index point in the 1980s and 1.7 Gini-index points in the 1990s to -1.1 Gini-index points during the period 2000–13.

An Overview of the Literature

There is a vast literature on income inequality, but most of these studies focus on the relationship between income inequality and economic development. In a seminal paper, Kuznets (1955) conjectured that a country’s income distribution becomes less egalitarian as its level of economic development increases, and that growth brings about more equality only after the level of income reaches a threshold. In other words, the income distribution evolves along an inverted U-shaped curve: growth results in relatively more inequality in the initial stages of economic development and greater equality at advanced stages. Greenwood and Jovanovic (1990), Banerjee and Newman (1993), Galor and Zeira (1993), Perotti (1993), and Barro (2000) find a positive correlation between growth and income inequality in a cross-section of international data. This hypothesis, however, has been challenged by other studies. Adelman and Robinson (1989), Anand and Kanbur (1993), and Ravallion (1995), among others, show that there is no empirical support for Kuznets’ conjecture.

A strand of the literature has looked beyond Kuznets’ hypothesis, aiming to identify the fundamental determinants of income inequality. Extensive evidence suggests that high inflation tends to depress income growth for the poor and lead to greater income inequality (Datt and Ravallion 1998; Ferreira, Leite, and Litchfield 2007). However, one of the most debated issues is the role of globalization—the increased openness to foreign trade and investment. From a theoretical point of view, the impact of trade openness on income inequality depends on factor endowments—countries with higher (lower) levels of human capital experience increases (decreases) in inequality. In the empirical literature, however, some scholars, such as Dollar and Kraay (2004), argue that globalization benefits the poor, while others, such as Barro (2000) and Milanovic (2005) show that greater openness leads to an increase in inequality, especially in countries with higher income levels. Similarly, the relationship between foreign direct investment (FDI) and income inequality has been extensively investigated and found to be positive. While Evans and Timberlake (1980) argue that dependence on FDI tends to exacerbate income inequality by altering the occupational structure of developing economies and producing both a highly paid elite and large groups of marginalized workers, Alderson and Nielson (1999) show an inverted U-shaped relationship between income inequality and the stock of FDI per capita.

Financial development has been shown to affect the distribution of income through multiple channels. A plethora of studies show that financial development affects income equality by enhancing human capital accumulation, improving the access to capital for entrepreneurial activity, and changing the sectoral composition of employment (Beck, Demirgüç-Kunt, and Levine 2007; Demirgüç-Kunt and Levine 2009). Most of the empirical literature reaches the conclusion that financial development lowers income inequality in the long term (Galor and Zeira 1993; Banerjee and Newman 1993; Clarke, Xu, and Zou 2006), except at the very early stages of development (Greenwood and Jovanovic 1990). However, because the distribution of capital income is significantly more unequal than the distribution of labor income, the concentration of wealth could become one of the root causes of income inequality over time (McKenzie and Woodruff 2006; Rajan 2010).6

The literature has also focused on the relationship between demographic and social characteristics and income inequality. Population growth is found to be critical, mainly through its effect on the demographic composition. First, while an increase in the supply of unskilled young workers may depress income growth (Alderson and Nielsen 1999), an increase in the share of the population older than 65 years tends to worsen income inequality (Deaton and Paxson 1997). Second, as pointed out by Kuznets (1955), the urbanization process becomes decisive, especially in the initial stage of economic development, because the evolution from an agrarian economy to industrialization leads to significant income disparities between and within rural and urban areas. Third, education forms a vital link between the pace and quality of growth and income distribution, although the relationship is not straightforward. Although cross-country studies indicate that a higher level of educational attainment brings about greater equality in the distribution of income, the type, quality, and distribution of education result in an intricate effect on income inequality, particularly in connection with skill-biased technological change (Barro 2000; Checci 2000).

Another critical dimension of income inequality is related to the distributional effects of fiscal policy. Although fiscal policy is traditionally assigned a limited role that focuses on the provision of public goods and services and long-term fiscal sustainability without directly taking into account distributional considerations, Musgrave (1959), among others, shows that fiscal policy can have an activist role in achieving an equitable distribution of income among households. The large variation in net income inequality across countries indicates that fiscal policy can influence the distribution of income (Feenberg and Poterba 1993; Auten and Carroll 1999; Benabou 2000; Muinelo-Gallo and Roca-Sagales 2011). The key consideration is the level and progressivity of taxation and expenditure policies. Well-targeted public spending can improve income distribution by providing greater equality of access to education and health care, thereby redistributing ownership of the factors of production.7 Taxation plays an important role in attaining greater equity in the distribution of income through the progressivity of the tax system and by generating sufficient revenues to fund public spending on social programs. Although taxation, especially of the top earning bracket, is presented as an obstacle to growth and an ineffective tool for fiscal redistribution (Bird and Zolt 2005), Bastagli, Coady, and Gupta (2012) show that direct income taxes and cash transfer schemes reduced the average Gini coefficient by about one-third in Organisation for Economic Co-operation and Development (OECD) countries during the period 1985–2005.

In China, uneven educational attainment and large geographical disparities are shown to be the primary causes of inequality.8Ping (1997), Knight and Song (1999), Sicular and others (2006), and Whyte (2010) identify the rural-urban income gap—driven by a secular decline in agricultural prices and rapid urbanization—as the key determinant of income inequality. In particular, the hukou (household registration) system contributed to rural-urban income inequality by restricting internal migration (Herd, Koen, and Reutersward 2010). Walsh and Yu (2012) find that inflation exacerbates income inequality; Xu and Zou (2000) show that the increase in income inequality is associated with, besides high inflation, the decline in the share of state-owned enterprises and, to a lesser extent, trade openness. With regard to the relationship between financial development and income distribution in China, the empirical evidence is not conclusive. Although most studies identify an adverse effect, Deng and Su (2012) find that financial deepening contributes to income growth among the poor and thereby to reducing income inequality. Taken together, the findings in the literature suggest that income inequality in China appears to be intimately linked to the state-engineered, export-led development model that has reshaped the opportunity landscape and the rates of return on human, financial, and physical capital.

A Synopsis of China’s Fiscal Policy

China has reformed its fiscal policy framework to improve revenue collection and enhance the effectiveness of public spending. One of the key components of fiscal reform is the devolution of authority for fiscal policymaking from the central government to provincial governments, which now account for 53 percent of general government revenues and 85 percent of expenditures. Another strategic development has been the enactment of a range of tax policy and administration reforms, including the introduction of the value-added tax (VAT) regime that replaced the wholesale turnover tax. Although the VAT still suffers from a narrow tax base that is confined to goods and a few services, the system has helped remove tax-induced distortions and provided a significant and stable source of revenue.

China’s tax-to-GDP ratio doubled from less than 10 percent in the early 1990s to 19 percent in 2013. It is, however, still significantly below the average of about 35 percent in the OECD countries. Furthermore, China collects more than half of its revenues from indirect taxes. Personal income taxes amount to 6 percent of total tax revenues (and 1.1 percent of GDP), whereas indirect taxes on goods and services account for more than 50 percent of total tax revenues (and about 10 percent of GDP). Even though China has a progressive personal income tax rate schedule with a top rate of 45 percent, its broad tax brackets and generous allowance schedule result in a very low ratio of personal income taxes to indirect taxes, which is a rough measure of the overall progressivity of taxation.9

China’s government spending has grown steadily from 18 percent of GDP in 1990 to 29 percent in 2013, but it remains below the OECD average of 45 percent. Off-budget spending by local governments, however, is substantial and amounts to about 15 percent of GDP.10 This increase in government spending is largely due to higher outlays for infrastructure investment and public administration, while social spending accounts for about 6 percent of GDP. The government has recently expanded the minimum subsistence allowance (dibao) system and introduced a new pension scheme in rural areas, but these programs have limited coverage and provide a low level of income compared with what urban workers earn.11 An additional complication is the existing system of fiscal relations between the central government and subnational governments. Although subnational governments are responsible for more than half of total spending, they have limited revenue-raising capacity and experience substantial differences in per capita allocations for basic public services (Figure 19.3).

Figure 19.3Evolution of Fiscal Policy in China

Sources: IMF, World Economic Outlook (panels 1, 2, and 6); National Bureau of Statistics and authors’ calculations (panel 3); National Bureau of Statistics (panels 4 and 5).

Data Description

The empirical analysis is based on annual data spanning 1980–2013, covering a panel of 33 countries along with China. The dependent variable is income inequality as measured by the net Gini coefficients for China and the panel of BRIC+ countries, which are drawn from the SWIID, constructed by Solt (2009) using the Luxembourg Income Study as the harmonized benchmark for comparable estimates. The SWIID provides two definitions of the Gini coefficient—the first based on market income and the second net of taxes and transfers—on an annual basis, using a custom missing-data multiple-imputation algorithm to standardize observations collected from various sources.12 The SWIID is the preferred source of data on income inequality for this analysis because it provides comparable figures across countries and for a longer span of time. Nevertheless, although these series allow for better identification of the sources of income inequality and the redistributive impact of fiscal policy, they are still subject to measurement uncertainty as depicted in Figure 19.4. This is why the analysis is limited to the net Gini coefficient, which describes income distribution across size-adjusted households after taxes and transfers are taken into account.

Figure 19.4Confidence Intervals of Gini Coefficients in China

Source: Standardized World Income Inequality Database.

Drawing on the literature and facing data constraints, the analysis focuses on a list of key explanatory variables. In China-specific regressions, the exercise includes real GDP per capita (instrumented by the number of international tourist arrivals), tax revenues as a share of GDP, and government spending as a share of GDP as the variables of interest. In BRIC+ panel data estimations, the list of explanatory variables is broadened to include real GDP per capita (instrumented by the number of international tourist arrivals), tax revenues as a share of GDP, government spending as a share of GDP, trade openness, financial development, an index of human capital, urbanization, and old-age dependency. Economic and financial statistics are compiled from the IMF’s Government Finance Statistics, International Financial Statistics, and World Economic Outlook databases; the World Bank’s World Development Indicators database; and the National Bureau of Statistics of China. Descriptive statistics for the key variables of interest are presented in Table 19.1 for China and in Table 19.2 for the panel data set for BRIC+ countries.

Table 19.1Descriptive Statistics: China
VariableObservationsMeanMedianStandard DeviationMinimumMaximum
Net Gini Coefficient3440.337.99.527.353.6
Real GDP per Capita (yuan)344,440.33,229.03,640.9806.613,164.4
International Tourist Arrivals (millions)3427.823.318.25.757.7
Government Spending (percent of GDP)3419.418.45.210.729.1
Tax Revenue (percent GDP)3414.614.63.29.322.2
Source: Authors’ calculations.
Table 19.2Descriptive Statistics: BRIC+ Countries
VariableObservationsMeanMedianStandard DeviationMinimumMaximum
Net Gini Coefficient1,01240.040.89.219.262.6
Real GDP per Capita (U.S. dollars)1,0733,492.62,867.42,443.4221.711,533.8
International Tourist Arrivals (millions)7926.83.38.80.057.7
Government Spending (percent of GDP)66528.126.69.70.055.4
Tax Revenue (percent of GDP)56318.517.97.93.950.0
Trade Openness (percent of GDP)96362.153.437.811.5321.1
Domestic Credit to Private Sector (percent of GDP)1,03139.828.032.31.2167.5
Index of Human Capital per Person9202.32.40.51.33.3
Urban Population (percent of total)1,15657.557.618.718.395.0
Source: Authors’ calculations.Note: A country is classified as BRIC+ if it belongs to the IMF’s “Emerging Market and Middle-Income Economies” country group classification, which includes Algeria, Angola, Argentina, Azerbaijan, Belarus, Brazil, Chile, China, Colombia, Croatia, the Dominican Republic, Ecuador, Egypt, Hungary, India, Indonesia, Iran, Kazakhstan, Malaysia, Mexico, Morocco, Pakistan, Peru, the Philippines, Poland, Romania, Russia, South Africa, Sri Lanka, Thailand, Turkey, Ukraine, Uruguay, and Venezuela.

Before proceeding with the estimations, it is important to analyze the time-series properties of the data to avoid spurious results. The Augmented Dickey-Fuller (ADF) test is commonly used in the literature to investigate the integration order, but it may suffer from size distortions and fail to differentiate between a highly persistent stationary series and a nonstationary process. Accordingly, the Ng-Perron (Ng-P) test is also performed to ensure the robustness of the empirical results.13 The unit-root results, available upon request, indicate that the variables used in the analysis are stationary after logarithmic transformation.

Empirical Strategy and Results

The analysis investigates the determinants of income inequality in China, with a particular focus on the redistributive contribution of fiscal policy. In view of data constraints and the need to have sufficient degrees of freedom, only a limited number of explanatory variables can be considered in the specification. Accordingly, using the IV regression via the two-stage least squares and GMM estimators, the following equation is estimated:

in which GINIt is the net Gini coefficient at time t; β0 is the intercept term; GDPPCt is real GDP per capita; EXPt and TAXt are government spending and tax revenues as a share of GDP, respectively; and εt is the error term.

To control for potential reverse causality, international tourist arrivals are used as an instrument for real GDP per capita. The relationship between economic growth and income inequality may exhibit contemporaneous reverse causation, given that income inequality influences the pace of growth. The challenge is to find a robust, time-varying IV, which needs to be correlated with real GDP per capita and needs to be exogenous with respect to real GDP per capita, but will have no effect on income inequality, except through its effect on per capita income. Although several empirical studies have used variations in rainfall and international commodity prices as IVs for economic growth, these may not be plausible for China and for most of the sample of BRIC+ countries. First, these countries are no longer highly dependent on the agriculture sector. Second, economic developments in these countries may likely influence the behavior of international commodity prices. As an alternative, the number of international tourist arrivals is introduced as the IV for real GDP per capita. According to the test statistics, this is a robust IV for per capita income with no direct effect on income inequality for China. However, although the number of international tourist arrivals is also a plausible IV for BRIC+ countries as a group, it should be noted that the strict exogeneity assumption may not hold in some countries where tourism plays a more significant role in economic activity.

The panel data analysis in the context of BRIC+ countries allows the analysis to overcome data constraints and include a broader set of control variables. The cross-country analysis for a panel of BRIC+ countries for the period 1980–2013 is conducted with the following specification:

in which GINIi,t is the net Gini coefficient in country i at time t; GDPPCi,t is real GDP per capita instrumented by the number of international tourist arrivals; and EXPi,t and TAXi,t are government expenditures and tax revenues as a share of GDP, respectively. The term Xi,t is a vector of control variables including trade openness, financial development, human capital accumulation, and urbanization. The ηi and Vt coefficients denote country- and time-specific effects, and εi,t is an idiosyncratic error term that satisfies the standard assumptions of zero mean and constant variance.

Because panel data tend to have complex error structures, standard estimation techniques are likely to yield inefficient estimates with biased standard errors. The Wooldridge-Drukker test is performed and indeed detects the presence of first-order serial correlation in the panel data used in this analysis.14 To account for the persistence of income inequality, the IV-GMM estimator is applied in a dynamic model that includes lagged values of the dependent variable as a regressor. The GMM approach takes into account unobserved country effects and possible endogeneity of the explanatory variables, providing more robust and consistent parameter estimates. This method also allows a dynamic specification with the lagged dependent variable as an explanatory variable to be used, thereby taking into account the persistence of income inequality over time.

Instrumental Variable Models—China

Table 19.3 presents the IV estimation results for China, relating income inequality to the principal explanatory variables. Below each coefficient are reported the robust standard errors that account for heteroscedasticity and first-order autocorrelation in the error terms.15 Model (1) regresses the net Gini coefficient on real GDP per capita—instrumented by the number of international tourist arrivals—and its square term. Model (2) incorporates government spending and taxation as a share of GDP.16 The results indicate that an increase in per capita income leads to a worsening of income inequality, while its square term lowers inequality. The coefficients on both real GDP per capita and its square term are highly statistically significant, and support the hypothesis of the existence of a Kuznets curve—an inverted U-shaped relationship between income inequality and economic development.

Table 19.3China: Instrumental Variables Estimation
Log (net Gini coefficient)
Variable(1)(2)
Log (real GDP per capita)0.682**1.396***
(0.208)(0.461)
Log (real GDP per capita) (squared)−0.025*−0.069**
(0.013)(0.028)
Log (government spending as percent of GDP)0.152**
(0.070)
Log (tax revenue as percent of GDP)−0.085*
(0.043)
Adjusted R20.9590.963
HAC score Chi25.058*6.197**
HAC regression F27.084***39.810***
Source: Authors’ calculations.Note: Real GDP per capita is instrumented using the number of international tourist arrivals. The sample period is 1980–2013. Heteroscedasticity and autocorrelation consistent (HAC) standard errors are reported in parentheses.*** p < 0.01; ** p < 0.05; * p < 0.1.

Government spending is found to be a statistically significant factor with a worsening effect on the distribution of household income. This outcome reflects the fact that government spending in China is low and dominated by infrastructure investment and public administration. Conversely, taxation appears to have the desired negative coefficient and comes out to be statistically significant. This suggests that taxation has a redistributive effect in China, where the tax-to-GDP ratio almost doubled during the past two decades, even though it still remains significantly below the OECD average. The adjusted R2 of this model is slightly higher, and the coefficients on per capita income and its square term are larger, than when the impact of economic development is considered separately.

Instrumental Variable Panel Data Analysis—BRIC+

The results of the BRIC+ panel estimations, presented in Table 19.4, are broadly consistent with the findings of the time-series analysis of China. The IV-GMM estimations indicate a high degree of persistence in income inequality. The coefficient on the lagged net Gini coefficient is positive and statistically significant across all specifications, although it becomes marginally smaller with the inclusion of other explanatory variables. The exercise finds that the level of per capita income, instrumented by the number of international tourist arrivals, widens income inequality, while its square term has a narrowing effect. Both of the estimated coefficients are statistically significant across all specifications, confirming the existence of the Kuznets curve in the BRIC+ panel. With regard to the impact of fiscal policy on inequality, the estimation finds that the coefficient on government spending has the desired negative sign, indicating that higher government spending lowers the net Gini coefficient. The magnitude of this effect, however, is small and statistically insignificant. Similarly, taxation appears to have a redistributive impact but comes out to be a statistically insignificant factor.

Table 19.4BRIC+ Panel: IV-GMM Estimation
VariableLog (net Gini coefficient)
(1)(2)(3)
Log (net Gini coefficient)t-10.919***0.874***0.894***
(0.010)(0.050)(0.031)
Log (real GDP per capita)0.194***0.322*0.210***
(0.027)(0.177)(0.070)
Log (real GDP per capita) (squared)−0.013**−0.020*−0.014***
(0.002)(0.010)(0.005)
Log (government spending as percent of GDP)−0.009−0.001
(0.014)(0.015)
Log (tax revenue as percent of GDP)−0.004−0.013
(0.016)(0.018)
Log (trade openness as percent of GDP)−0.003
(0.008)
Log (domestic credit to private sector as percent of GDP)0.005
(0.005)
Index of Human Capital per Person−0.034
(0.036)
Log (urban population as percent of total)0.064
(0.040)
Wald χ210,352.73***3,429.81***3,408.46***
Source: Authors’ calculations.Note: Real GDP per capita is instrumented using the number of international tourist arrivals. The sample period is 1980–2013. Heteroscedasticity and autocorrelation consistent standard errors are reported in parentheses.*** p < 0.01; ** p < 0.05; * p < 0.1.

The empirical findings of the baseline IV-GMM model remain robust to the inclusion of various control variables. In line with the literature, the analysis finds that trade openness and human capital accumulation improve the distribution of household income in the BRIC+ panel, but the estimated coefficients do not reach the threshold of statistical significance. However, financial development has a worsening effect on income inequality, as expected, but it is statistically insignificant. The results suggest that an increase in the share of urban population worsens income distribution, but this effect is not statistically significant in the panel of BRIC+ countries.

Concluding Remarks and Policy Issues

China has made remarkable progress in reducing poverty, but this achievement has been accompanied by widening income disparities across the society. Although this situation reflects an intricate array of developments, including the country’s investment- and export-led growth model as well as socioeconomic, financial, and institutional undercurrents, fiscal policy appears to have played an important role through the impact of taxes and transfers on income distribution. The market and net Gini coefficients increased by 82 percent and 90 percent, respectively, during the period 1974–2013. In other words, the egalitarian effects of fiscal policy fell from an average of 1.7 Gini-index points in the 1990s to -1.1 Gini-index points during the period 2000–13.

China’s tax-to-GDP ratio has almost doubled since 1995 to 19 percent, but it remains significantly below the OECD average of about 35 percent. This low level of tax support effectively sets a limit on public expenditures, including redistributive measures. Furthermore, China’s system of taxation distributes the tax burden in a regressive manner across income groups, largely because China collects more than half of its revenues from indirect taxes. Personal income taxes amount to 6 percent of total tax revenues, whereas indirect taxes on goods and services account for more than 50 percent of total tax revenues. Although China has a progressive personal income tax rate schedule with a top rate of 45 percent, its broad tax brackets and generous allowance schedule diminish the effective progressivity of the tax regime, resulting in a very low ratio of personal income taxes to indirect taxes.

Although government spending has grown from 18 percent of GDP in 1990 to 29 percent in 2013, it is still significantly below the OECD average of 45 percent. This increase in government spending is largely due to higher outlays for infrastructure investment and public administration, while social protection and health care account for only about 6 percent of GDP (compared with an average of 15 percent in OECD countries and 9 percent in upper-middle-income countries). In other words, excluding social protection and health care, China’s nonredistributive government spending is comparable to that in OECD countries. Furthermore, the incidence of benefits from public services and transfers is shown to favor high-income groups in urban areas.17 For example, the top quartile of households receives about 80 percent of pension spending, compared with only 2 percent for the bottom quartile. As a part of the “harmonious society” strategy, the government has expanded the minimum subsistence allowance system and introduced a new pension scheme in rural areas, but these programs have limited coverage and provide a low level of income compared with what urban workers receive. An additional complication is the existing system of fiscal relations between the central government and subnational governments. Although subnational governments are responsible for more than half of total spending, they have limited revenue-raising capacity and experience substantial differences in per capita allocations for basic public services.

The empirical findings presented in this chapter are consistent with the stylized facts about China. First, the analysis shows that an increase in real GDP per capita—instrumented by the number of international tourist arrivals—leads to an increase in the net Gini coefficient, while its square term lowers income inequality. This confirms the existence of an inverted U-shaped relationship between income inequality and economic growth. Second, the exercise shows that government spending and taxation have opposing effects on income inequality. Whereas government spending appears to worsen inequality, taxation improves the distribution of household income. The results of the panel data analysis are broadly consistent with the findings of the time-series analysis for China, with one important exception—both government spending and taxation have the desired redistributive effect, although it is statistically insignificant. Altogether the redistributive impact of fiscal policy in China appears to be stronger than what is identified in the BRIC+ panel, the “net” effect is still not sufficient to compensate for the adverse impact of other influential factors identified in the analysis.

Fiscal policy can be redesigned to have a greater redistributive effect, especially in the long term. On the taxation front, the system needs to be broadened in a more progressive way to help narrow income inequality. The effective number of personal income tax payers is less than 3 percent of the working population, indicating a high degree of informality and tax avoidance. Strengthening tax administration and broadening the personal income tax, including capital gains—thus increasing effective taxation of the rich—and imposing VAT on services, which tend to be consumed more by the rich, would make the tax regime more progressive and also create additional fiscal space. In particular, China has room to lower high labor taxation that hurts the low- and middle-income brackets more than the rich, while increasing direct taxes on capital and wealth, especially through more effective land and property taxation. The planned extension of a recurrent property tax from pilot implementation to the rest of the country is a step in the right direction to generate additional revenues and improve the progressivity of the tax regime.

On the expenditure side, there is scope for making public spending a more effective tool by improving progressivity through well-targeted programs that champion greater access for the poor, particularly in rural areas, where more than 95 percent of poor households in China reside. To this end, given the decentralized nature of China’s fiscal system, expenditure assignments need to be realigned with revenue sources across all layers of government. Expanding the social safety net, including means-tested income support to the poor and unemployment insurance, is critical, and untargeted energy subsidies, which tend to benefit the rich more than the poor, should be reduced.18 Reforming the pension system, including redesigning the eligibility criteria for the basic retirement pension, could have a positive redistributive impact, but it needs to be accompanied by structural and parametric changes (that is, pooling of provincial-level pension funds and adjusting the retirement age and replacement rates) to ensure long-term sustainability, especially in view of China’s rising old-age dependency ratio.19

The distributional effects of fiscal policy should be taken into account, but redistributive measures need to be consistent with the objective of fiscal sustainability. In particular, the expansion of social assistance programs needs to take into account the fiscal cost of rapid population aging. Ultimately, fiscal policy is only one aspect of an inclusive growth strategy that requires a comprehensive range of structural reforms aiming to sustain economic growth as well as to provide every segment of the society with greater access to emerging opportunities.

References

    Adelman, I., and S.Robinson.1989. “Income Distribution and Development.” In Handbook of Development Economics, edited by H.Chenery and T.Srinivasan. (ed)Amsterdam: North-Holland.

    Alderson, A., and F.Nielsen.1999. “Income Inequality, Development, and Dependence.American Sociological Review64 (4): 60631.

    Anand, S., and S.Kanbur.1993. “Inequality and Development: A Critique.Journal of Development Economics41 (1): 1943.

    Auten, G., and R.Carroll.1999. “The Effects of Income Taxes on Household Income.Review of Economics and Statistics81 (4): 68193.

    Banerjee, A., and A.Newman.1993. “Occupational Choice and the Process of Development.Journal of Political Economy101 (2): 27498.

    Barro, R.2000. “Inequality and Growth in a Panel of Countries.Journal of Economic Growth5 (1): 532.

    Bastagli, F., D.Coady, and S.Gupta.2012. “Income Inequality and Fiscal Policy.IMF Staff Discussion Note No. 12/08, International Monetary Fund, Washington.

    Beck, T., A.Demirgüç-Kunt, and R.Levine.2007. “Finance, Inequality and the Poor.Journal of Economic Growth12 (1): 2749.

    Benabou, R.2000. “Unequal Societies: Income Distribution and the Social Contract.American Economic Review90 (1): 96129.

    Berg, A., and J.Ostry.2011. “Inequality and Unsustainable Growth: Two Sides of the Same Coin?IMF Staff Discussion Note No. 11/08, International Monetary Fund, Washington.

    Bird, R., and E.Zolt.2005. “Redistribution via Taxation: The Limited Role of the Personal Income Tax in Developing Countries.UCLA Law Review52: 171.

    Checci, D.2000. “Does Educational Achievement Help to Explain Income Inequality?UNU-WIDER Research Paper No. 2000/208, United Nations University–World Institute for Development Economics Research, Helsinki.

    Clarke, G., L.Xu, and H.Zou.2006. “Finance and Income Inequality: What Do the Data Tell Us?Southern Economic Journal72 (3): 57896.

    Datt, G., and M.Ravallion.1998. “Farm Productivity and Rural Poverty in India.Journal of Development Studies34 (4): 6285.

    Deaton, A., and C.Paxson.1997. “The Effects of Economic and Population Growth on National Saving and Inequality.Demography34 (1): 97114.

    Demirgüç-Kunt, A., and R.Levine.2009. “Finance and Inequality: Theory and Evidence.World Bank Policy Research Working Paper No. 4967, World Bank, Washington.

    Deng, H., and J.Su.2012. “Influence of Financial Development on the Income Distribution in China.Social Science Letters1 (1): 7379.

    Dollar, D., and A.Kraay.2004. “Trade, Growth, and Poverty.Economic Journal114 (493): F22F49.

    Drukker, D.2003. “Testing for Serial Correlation in Linear Panel-Data Models.Stata Journal3 (3): 16877.

    Dunaway, S., and V.Arora.2007. “Pension Reform in China: The Need for a New Approach.IMF Working Paper No. 07/109, International Monetary Fund, Washington.

    Evans, Peter B., and MichaelTimberlake.1980. “Dependence, Inequality, and the Growth of the Tertiary: A Comparative Analysis of Less Developed Countries.American Sociological Review45 (4): 53152.

    Feenberg, D., and J.Poterba.1993. “Income Inequality and the Incomes of Very High Income Taxpayers.Tax Policy and the Economy7: 14577.

    Ferreira, F., P.Leite, and J.Litchfield.2007. “The Rise and Fall of Brazilian Inequality: 1981–2004.Macroeconomic Dynamics12 (September): 199250.

    Galor, O., and J.Zeira.1993. “Income Distribution and Macroeconomics.Review of Economic Studies60 (1): 3552.

    Greenwood, J., and B.Jovanovic.1990. “Financial Development, Growth, and the Distribution of Income.Journal of Political Economy98 (5): 1076107.

    Herd, R., V.Koen, and A.Reutersward.2010. “China’s Labor Market in Transition: Job Creation, Migration and Regulation.OECD Working Paper No. 749, Organisation for Economic Co-operation and Development, Paris.

    Knight, J.2014. “Inequality in China: An Overview.World Bank Research Observer29: 119.

    Knight, J., and L.Song.1999. The Rural-Urban Divide: Economic Disparities and Interactions in China. New York: Oxford University Press.

    Kuznets, S.1955. “Economic Growth and Income Inequality.American Economic Review45 (1): 128.

    Lee, I., M.Syed, and X.Wang.2013. “Two Sides of the Same Coin? Rebalancing and Inclusive Growth in China.IMF Working Paper No. 13/185, International Monetary Fund, Washington.

    Li, S., and T.Sicular.2014. “The Distribution of Household Income in China: Inequality, Poverty and Policies.China Quarterly217 (March): 141.

    McKenzie, D., and C.Woodruff.2006. “Do Entry Costs Provide an Empirical Basis for Poverty Traps? Evidence from Mexican Microenterprises.Economic Development and Cultural Change55 (1): 342.

    Milanovic, B.2005. Worlds Apart: Measuring International and Global Inequality, 1950–2000. Princeton, New Jersey: Princeton University Press.

    Muinelo-Gallo, L., and O.Roca-Sagales.2011. “Economic Growth and Inequality: The Role of Fiscal Policies.Australian Economic Papers50 (2–3): 7497.

    Musgrave, R.1959. The Theory of Public Finance. New York: McGraw Hill.

    Ostry, J., A.Berg, and C.Tsangarides.2014. “Redistribution, Inequality, and Growth.IMF Staff Discussion Note No. 14/02, International Monetary Fund, Washington.

    Perotti, R.1993. “Political Equilibrium, Income Distribution and Growth.Review of Economic Studies60 (4): 75576.

    Ping, Z.1997. “Income Distribution during the Transition in China.UNU-WIDER Research Paper No. 1997/138, United Nations University-World Institute for Development Economics Research, Helsinki.

    Rajan, R.2010. Fault Lines: How Hidden Fractures Still Threaten the World Economy. Princeton, New Jersey: Princeton University Press.

    Ravallion, M.1995. “Growth and Poverty: Evidence for Developing Countries in the 1980s.Economics Letters48 (3–4): 41117.

    Shawki, A.1997. “China: From Mao to Deng.International Socialist Review1 (Summer).

    Shen, K., and S.-H.Lee.2014. “Benefit Incidence of Public Transfers: Evidence from the People’s Republic of China.ADB Working Paper No. 413, Asian Development Bank, Manila.

    Sicular, T., X.Yue, B.Gustafsson, and S.Li.2006. “The Urban-Rural Income Gap and Inequality in China.UNU-WIDER Research Paper No. 2006/135, United Nations University-World Institute for Development Economics Research, Helsinki.

    Solt, F.2009. “Standardizing the World Income Inequality Database.Social Science Quarterly90 (2): 23142.

    Walsh, J., and J.Yu.2012. “Inflation and Income Inequality: Is Food Inflation Different?IMF Working Paper No. 12/147, International Monetary Fund, Washington.

    Whyte, M.2010. One Country, Two Societies: Rural Urban Inequality in Contemporary China. Cambridge, Massachusetts: Harvard University Press.

    Wooldridge, J.2002. Econometric Analysis of Cross section and Panel Data. Cambridge, Massachusetts: MIT Press.

    Xu, L., and H.Zou.2000. “Explaining the Changes of Income Distribution in China.China Economic Review11 (2): 14970.

    Zhang, Y., and S.Barnett.2014. “Fiscal Vulnerabilities and Risks from Local Government Finance in China.IMF Working Paper No. 14/4, International Monetary Fund, Washington.

This paper benefited from helpful comments and suggestions by Santiago Acosta-Ormaechea, Andrew Berg, Benedict Clements, Mark De Broeck, Maura Francese, Sanjeev Gupta, Takuji Komatsuzaki, Bin Grace Li, Leandro Medina, Philippe Wingender, and participants at a seminar at the Fiscal Affairs Department of the International Monetary Fund.

There are various methods of estimating the Gini coefficient, resulting in significant differences. For example, the China Household Finance Survey conducted by Texas A&M University and Southwestern University of Finance and Economics in Chengdu estimated that the overall Gini coefficient was 0.61 in 2010.

Lee, Syed, and Wang (2013) provide a detailed account of China’s development model and its potential effects on income distribution.

According to the Chinese Family Panel Studies, conducted by Peking University and covering 14,960 households in five province-level areas, the top 5 percent bracket earned 23 percent of total household income, whereas the households in the lowest 5 percent bracket accounted for just 0.1 percent of total income.

At the beginning of gaige kaifang, Deng Xiaoping’s popular slogan was “We should let some people get rich first, both in the countryside and in the urban areas” (Shawki 1997).

Ostry, Berg, and Tsangarides (2014) find that lower net income inequality is robustly correlated with faster and more durable growth.

This chapter does not explore wealth inequality because of data constraints. Cross-country data, however, indicate that wealth inequality tends to be worse than income inequality and has a significant bearing on income inequality through nonlabor income accrued to the asset-rich bracket of households. For China, available data based on the China Household Income Project surveys show that the share of asset income increased from 8 percent of total household income in 2002 to 15 percent in 2007, contributing 13–19 percent of income inequality (Li and Sicular 2014).

In general, social spending such as pensions and unemployment benefits transfer resources to groups of the population who do not earn labor income and whose risk of falling into poverty would otherwise be large.

Knight (2014) provides a comprehensive overview of the literature on income inequality in China.

The effective number of personal income tax payers in China is less than 3 percent of the working population.

Zhang and Barnett (2014) provide an extensive analysis of off-budget activity in China.

According to the China Household Finance Survey in 2010, retirement insurance coverage was 34.5 percent in rural areas, compared with 87 percent in urban areas; annual pension income was 12,000 yuan for rural households, compared with 33,000 yuan for urban households.

The latest version (5.0) of the SWIID database is available at http://thedata.harvard.edu/dvn/dv/fsolt.

Regarding the optimal lag order selection for the ADF test, we use the modified Schwarz information criteria. For the Ng-P test, we use the Quadratic Spectral kernel–Andrews bandwidth combination to take into account the sample characteristics such as size and possible existence of structural breaks.

Implementing an idea originally proposed by Wooldridge (2002), Drukker (2003) developed an easy-to-use test for serial correlation in panel data based on the ordinary least squares residuals of the first-differenced model.

The results presented in Table 19.3 are based on the two-stage least squares approach. The IV estimator using the GMM yields similar findings, which are available upon request.

Although collinearity between government spending and tax revenue is a potential problem that may lead to unstable parameter estimates, collinearity diagnostics yield a variance inflation factor of 1.24, which is significantly less than the critical threshold of 10.

Using the framework of the National Transfer Accounts Project and household-level data, Shen and Lee (2014) provide an analysis of the benefit incidence of public spending across socioeconomic groups in 2009 and conclude that (1) education spending was equally distributed at the primary and secondary level, but favored high-income urban households at the tertiary level; (2) health care spending skewed toward high-income urban households; and (3) pension spending was far more favorable to high-income urban households.

As of 2011, posttax subsidies for petroleum products, electricity, natural gas, and coal accounted for more than 3.8 percent of GDP in China—almost four times the amount of health care spending.

Dunaway and Arora (2007) provide an approach to strengthening the pension system and dealing with the “legacy costs” associated with the relatively more generous benefits provided under the old system.

    Other Resources Citing This Publication