Correction of China's income inequality for missing top incomes

• Published date: 01 August 2023
• Author: Haiyuan Wan,Yangcheng Yu
• Date: 01 August 2023
• DOI: http://doi.org/10.1111/rode.13010

REGULAR ARTICLE

Correction of China's income inequality for

missing top incomes

Haiyuan Wan

1

| Yangcheng Yu

2

1

Business School, Beijing Normal

University, China

2

School of Economics, Shanghai

University of Finance and Economics,

Shanghai, China

Correspondence

Yangcheng Yu, School of Economics,

Shanghai University of Finance and

Economics, Shanghai, 200433, China.

Email: yyc961220@sina.com

Abstract

After accounting for top incomes missing from the

Chinese Household Income Project, this paper exam-

ines the income inequality trend in China during the

21st century. Our analysis, which involves fitting the

upper tail of the income distribution to a power-law

model, reveals that the authoritative income survey

data miss 0.68% of the population and 6.19% of aggre-

gate income in 2018. Despite the most recent survey

data providing better coverage of top incomes, our cor-

rection is still crucial. The raw survey data indicate a

consistently increasing Gini coefficient between 2002

and 2018, but the corrected index starts to decline from

2013. Meanwhile, the revised top 1% income share

increases from 7.01% in 2002 to 7.89% in 2013 and then

slightly decreases to 7.64% in 2018, while the revised

top 10% income share stabilises at around 33% through-

out the period. Notably, China's revised top 10% and

top 50% income shares in 2018 are close to those of the

United Kingdom but are considerably lower than those

of the United States.

KEYWORDS

household survey, income inequality, missing top incomes,

power-law model

JEL CLASSIFICATION

C81, D31, E01

Received: 19 January 2021 Revised: 9 May 2023 Accepted: 12 May 2023

DOI: 10.1111/rode.13010

Rev Dev Econ. 2023;27:1769–1791. wileyonlinelibrary.com/journal/rode © 2023 John Wiley & Sons Ltd. 1769

1|INTRODUCTION

China has experienced impressive economic growth in the last 20 years, becoming the world's

second-largest economy. Many studies have focused on the unprecedented scale of poverty reduc-

tion in China in the last two decades, but relatively little is known about the change in income

distribution at the top tail over this period. Although researchers have agreed that the survey cov-

erage of top incomes is often insufficient, national statistical agencies do not typically incorporate

top-income adjustments when they report Gini coefficients (Jenkins, 2022). An exception is the

National Bureau of Statistics (NBS) of China, which historically adjusted Gini coefficients by com-

bining official survey data with data from other sources, such as tax data, before disclosing them

to the public (Luo, 2019). Unfortunately, the Chinese government last released a national Gini

coefficient in 2000 but the index has not been officially updatedsince then.

Researchers now commonly take household survey data as the primary source for empirical

analysis of inequality. However, the income Gini coefficients published by several well-known

nationally representative surveys in China are inconsistent. For example, the 2010 Chinese

Household Finance Survey (CHFS) yielded a Gini coefficient of 0.677, whereas the 2010 Chi-

nese Family Panel Survey (CFPS) reported a Gini coefficient of 0.554. In 2013, the Chinese

Household Income Project (CHIP) yielded an even lower Gini coefficient of 0.451. The samples

used in the CHIP survey are drawn from the official government income survey and therefore

provide an appropriate data foundation to evaluate the national income inequality level in

China (Zhang et al., 2014).

In their study of China's income distribution, Piketty et al. (2019) combined survey data

with tax records to correct for missing top incomes. The Chinese government published income

tax records for high-income individuals between 2006 and 2010, but the data released were

restricted to a limited number of provinces from 2011. Combining survey data and income tax

data was no longer feasible. Instead, in a pioneering study, Li et al. (2020) collected top-income

samples from various industries to construct the first Top Incomes in China (TIC) micro data-

base. Li et al. (2021) combined CHIP data with TIC data to estimate China's level of income

inequality in 2016. However, the data on top-income earners were obtained from limited indus-

tries and self-collected, making their representativeness questionable. In addition, Li et al.

(2021) assumed that the top tail of the survey income distribution followed a right truncated

Pareto distribution, but they did not empirically test this hypothesis.

The current paper corrects for top incomes missing from the CHIP to document China's

income inequality trend in the 21st century. The main contributions are as follows. First, our

correction for missing top incomes requires a parametric assumption only at the top tail of the

survey income distribution. We neither restrict the distribution for other parts of the survey

data nor include any external top-income data. Second, most empirical studies have not

attempted to test the power-law hypothesis quantitatively. We generate a pvalue to test the

goodness of fit of our parametric model and compare our power-law model with alternative dis-

tributions using likelihood ratio tests. Third, the estimates of the model parameters are obtained

by maximum likelihood estimation (MLE) and the Kolmogorov–Smirnov (KS) statistic rather

than by ordinary least squares (OLS) regression or simple visualisation. These estimates are rel-

atively objective and are found to be asymptotically unbiased in our model. Finally, the house-

hold data used in this paper are subsamples of the government's official household income

survey and thus provide a more authoritative measure of income inequality than other survey

data. We use three waves of CHIP data to document income inequality trends between 2002

and 2018, against the backdrop of growing public concern over income inequality in China.

1770 WAN and YU