Sociology Mind
2012. Vol.2, No.4, 373-381
Published Online October 2012 in SciRes (http://www.SciRP.org/journal/sm) http://dx.doi.org/10.4236/sm.2012.24049
Income Inequality in Transitional Urban China: The Effect of
Market versus State
Qiong Wu, Barry Goetz, David Hartmann, Yuan-Kang Wang
Department of Sociology, Western Michigan University, Kalamazoo, USA
Received July 2nd, 2012; revised August 4th, 2012; accepted August 13th, 2012
The rise of inequality in China is one of the most serious social problems in the reform era in China. Pre-
vious studies have debated the relative importance of human capital, political capital, and other factors in
determining personal income. Using a new dataset from 2006 China General Social Survey (CGSS, 2006),
the first author replicates earlier tests to measure whether the market or state has more impact on incomes
as a way to the competing hypotheses related to human versus political capital. The results of the ordinary
least squares regression analysis show no significance in party membership, state ownership, and work
experience, while the first author does find high returns to education, which supports Nee’s market transi-
tion theory. Moreover, the findings indicate that market sectors, including domestic private enterprises
and foreign enterprises have remarkable advantages in earnings, and there is a great income gap between
different regions, sectors, and within the sectors. To summarize, the market and state play a dual role in
determining income in transitional urban China.
Keywords: Income Inequality; Urban China; Market Effect; State Effect
Introduction
China’s Communist Revolution was founded upon the idea
of equality of wealth. In pre-reform China, the society was
relatively equal in income distribution and resource allocation.
Since 1978, China has been carrying out a transformation from
a socialist planned economy to market economy, along with a
great social change from relative social egalitarianism to a new
era of individualism and competition under the market mecha-
nism.
Sir Arthur Lewis said, “development must be inegalitarian
because it does not start in every part of the economy at the
same time” (Lewis, 1976: p. 26). In terms of China, the gov-
ernment has started a policy to allow and encourage some peo-
ple to get rich first and some regions to develop quickly, and
coastal and urban areas obtained the priority to develop first
and faster. As a result, the income gap between the rich and
poor, between urban and rural areas, and between different
regions has become larger.
Compared to the pre-reform era, though inequalities have in-
creased dramatically between workers and professionals, east-
ern-coastal regions and western regions, “under a market sys-
tem, everyone ostensibly has an opportunity to try for better
jobs and income” (Tang & Parish, 2000: p. 51). Chinese soci-
ety has become more diverse. Specialization helps build a more
organic society, in which an individual’s needs are served by
markets, rather than by the state.
However, according to the survey results from the national
China Household Income Project 2002, 81.5% of people think
that the current situation on income distribution is not fair, and
the 2006 China General Social Survey also indicated that over
50% of the respondents feel unfair about the income distribu-
tion. People’s attitude towards the unfairness of income distri-
bution, to some extent, reflects income inequality in China that
ordinary people feel the widen gap between the rich and the
poor, the urban-rural divide, between different social classes,
and different regions. The income gap has become the most
serious social problem in current China, far ahead of crime and
corruption, which rank in second and third place based on a
survey in 2004 (Xinhua, 2004).
In studies of social change and problems in the societal
transformation in the state socialism, there are three contradic-
tory theories regarding social transformation in post-socialist
societies: 1) continuing bureaucratic politics (power continuity;
2) market transformation (structural transformation); 3) the mix
solution of technocratic continuity (Tang & Parish, 2000: p.
83).
Nee’s market transition theory argues that “higher returns of
education, which is among the best indicators of human pro-
ductivity” (Nee, 1989: p. 666). The thesis of “power persis-
tence” (Bian & Logan, 1996) contends that political power of
party cadres can be transformed into economic advantages on
the course of the transition to a market economy. The politi-
cally-based privilege is still “deeply embedded in the economic
situation” (p. 741). The argument of technocratic continuity
suggests that the old technocratic managers with specialized
skills would regain their advantages in the socialist economy
and emerging as the new entrepreneurs in the market economy.
The technocratic cadres “can maintain their positions through
the acquired expertise” (Rona-Tas, 1994: p. 45).
Based on the literature, the whole theoretical debate comes
down to considering competing hypotheses whether human
capital or political capital is more important in determining
personal income in urban China. Human capital include educa-
tion, work experience, skills, parental education, etc. Political
capital refer to party membership, working in the state sector,
government and other power agencies, parental party member-
ship, social contact that can get access to political capital. My

Q. WU ET AL.
research hypotheses are as follows:
Hypothesis 1: Human capital is the best indicator of income
China today. In other words, higher educational credentials and
more work experience will lead to higher earnings.
Hypothesis 2: Political capital (party membership) remains
the best predictor of income in China today.
“Communist Party membership continues to yield an income
advantage to workers and workers whose jobs hold redistribu-
tive power earn more” according to Bian and Logan’s (1996)
analysis on survey conducted in Tianjin, China in 1988 and
1993. Bian, Shu, and Logan (2001) also found that during the
post-1978 reform era, “party membership had a significant
effect on mobility into elite positions of political and manage-
rial authority, and college education increased party members’
chances of moving into positions of political authority but not
into managerial positions within the state sector” (p. 832).
Hypothesis 3: The role of work unit sector and state owner-
ship remains significant in determining income.
With an analysis of data survey collected in Shanghai, Xi’an
and Wuhan in 1999, Xie and Wu (2008) indicates that “the
danwei (work unit) continues to play a very important role in
determining the economic well-being” (p. 13), and it still serves
as “a major agent of social stratification in urban China” (p. 6).
In this paper, the first author addresses the issue of the theo-
retical debate in the literature on the research on social inequal-
ity in China by using a newer and different national dataset
from CGSS, 2006 as a way to the competing hypotheses related
to human versus political capital. The fundamental questions in
this study are focused on: 1) Do income returns more on politi-
cal capital (party membership) or human capital (education and
work experience)? 2) How do these changes related to trends in
aggregate inequality?
Data and Variables
In this paper, the first author employs individual-level data
from the urban samples of the 2006 China General Social Sur-
vey (CGSS, 2006) under the joint sponsorship of Survey Re-
search Center, Hong Kong University of Science and Technol-
ogy, and Department of Sociology, Renmin University of
China.
The CGSS is an annual or biannual questionnaire survey of
China’s urban and rural households. It aims to “monitor sys-
tematically the changing relationship between social structure
and quality of life in urban and rural China”
(http://www.ust.hk/~websosc/survey/GSS_e.html). The survey
program started from 2003, and the first dataset only covered
the urban areas. In 2005, rural areas were added. The data of
2006 encompasses three sections: urban, rural and family ques-
tionnaires. For this paper, the first author only used the urban
data of 2006, for analysis.
The surveys were conducted during September 2006 to Oc-
tober 2006 with 1610 variables and 10,151 cases (6013 cases in
urban areas). A multistage cluster sampling procedure selected
28 provinces and municipalities. The respondents are from the
age of 18 to 69, in randomly selected 10,000 households in 28
provinces and cities nation-wide. The urban questionnaires
contained personal general information, work experience, cur-
rent work situation, family situation, and attitudes towards the
society.
In order to estimate the relationships between income distri-
bution and several socio-demographic characteristics of indi-
viduals, my analyses rely on OLS regression to predict total
individual income in urban China. Table 1 lists all the variables
used in the study.
“Hukou” is a particular household registration system in
China. Dating back about 2000 years ago, when Qin Dynasty
united the whole China, and set up this household registration
system to collect taxes according to the number of people. After
the Communist Party established the People’s Republic of
China, the Communist regime revived it in 1955 to keep poor
rural farmers from flooding into the cities in case that the “ex-
tensive rural-to-urban migration would undercut the attempt to
develop an urban welfare state”. The “Hukou” registration sys-
tem “classified each member of the population as having agri-
cultural (rural) or nonagricultural (urban) status (Hukou), with a
sharp differentiation of rights and privileges and extremely
stringent conditions for converting from rural to urban status”
(Wu & Treiman, 2004: p. 363).
Due to the restriction of “Hukou”, those who move to large
cities to work or study but do not have the local “Hukou” can-
not enjoy all kinds of benefits as the citizens, and have to go
back to their hometown to get a marriage license, apply for a
passport or take the national university entrance exam. Rather,
the “Hukou” system creates unfair advantages for those who
live in large cities especially Beijing and Shanghai. Because in
China, most highly regarded universities and hospitals locate in
large cities, and those institutions provide more preferential
policies to the local Hukou-holders. Moreover, most local en-
terprises tend to favor in those who are local residents. Thus,
those who have the urban “Hukou” of large cities tend to have
advantages over those who are originally from smaller places.
In pre-reform China, Chinese urban society was organized by
each work unit dominated by the state. “In Chinese official
statistics, the danwei1 or work unit is defined as an independent
accounting unit with three characteristics: 1) administratively,
it is an independent organization; 2) fiscally, it has an inde-
pendent budget and produces its own accounting tables of
earnings and deficits; 3) financially, it has independent ac-
counts in banks and has legal rights to sign contracts with gov-
ernment or business entities” (Bian, 1994: p. 23). The role of
danwei or work unit was extremely significant that it defined
one’s social, economic, and political life. Individuals depended
on danwei for almost everything. Without a work unit, it was
difficult to survive in a city because housing, food, and other
social services were hardly available through the market.
After the reform, with the emerging of private sector include-
ing private enterprises, foreign companies, joint-ventures, and
the self-employed, the role of danwei has lost some of its im-
portance compared to the era of pre-reform, because through
danwei is no longer the only way to get all social services, the
market has made it more diverse. However, danwei does not
disappear with the challenge of the market, and remains the
main agent of social stratification in contemporary urban China.
Except danwei or work unit, “ownership type has always
been an important factor in determining income”, (Wang, 2008:
p. 113). According to the questionnaire in CGSS, 2006, types
of work unit and ownership are two separate but close-related
questions. The types of work unit include government and party
agencies, enterprises, institutions, social organizations, and
individual operation or self-employed. Among these work or-
1“The term danwei or work unit refers to all work organizations in general,
but was often used to refer to state economic enterprises in particular” (Wu,
2002: p. 1073).
374

Q. WU ET AL.
Table 1.
Description of predictors for the analysis of individual income inequality in urban China.
Variables Description
Total income (income 2005) Personal yearly total income in 2005 (Yuan)
Gender (gender) 1 = female; 2 = male
Work experience (workexp) Work experience is measured by subtracting the end year of a job from the start year (in years)
Education level (education)
Education is measured by eight levels
1 = never schooled
2 = classes for eliminating illiteracy
3 = elementary School
4 = middle School
5 = high School
6 = junior college
7 = college/university
8 = graduate
Foreign language skill (lanskill)
Four categories:
1 = not at all
2 = know a little
3 = somewhat fluent
4 = very fluent
Type of “Hukou” (“Hukou”)
Four categories:
1 = urban “Hukou” in small cities/towns
2 = urban “Hukou” in middle cities
3 = urban “Hukou” in large cities (Municipalities and Provincial capital)
4 = rural “Hukou”
Party membership (party)
Two categories:
1 = member of communist party of China or communist youth league of China;
2 = non-communist party member (other parties or no party)
Type of workplace (including danwei and other
workplaces in the market sector) (workplace)
1 = government agencies and state-owned enterprises (SOEs)
2 = collective enterprises
3 = private enterprises
4 = foreign-invested enterprises (including Hong Kong, Macao, and Taiwan)
5 = institutions
6 = social organizations or public organizations
7 = other
Geographic or residential location (location)
1 = eastern coastal regions
2 = central regions
3 = western regions
Source: data from CGSS, 2006.
ganizations, only those who answered enterprises and institu-
tions have to answer the second question about the type of sec-
tor or ownership. The options are state-owned, collective, pri-
vate enterprises, enterprises from Hong Kong, Macao and Tai-
wan, and foreign-invested or owned enterprises. Since all insti-
tutions are government-sponsored, the first author combine the
type of work unit and ownership into one variable Workplace to
distinguish the different types of enterprises. I distinguish the
following type of workplace in urban China:
1) Government agencies and SOEs, which include all levels of
government and Communist party agencies and state-
owned enterprises is the reference group.
2) Collective enterprises are not directly supported by the state
but are mostly sponsored by local governments.
3) Private enterprises include private firms and individual
operation or self-employed.
4) Foreign enterprises include foreign-owned, foreign-invested
companies and the enterprises from Hong Kong, Macao,
and Taiwan.
5) Institutions or public institutions include schools, research
institutions, libraries, museums, hospitals and publishing
houses, are the backbone of public service providers in
China.
6) Social organizations or public organizations are sets of as-
sociations emerged in the late 1980s with official encour-
agement, consisting of genuine NGOs and government-or-
ganized NGOs.
7) Others.
Residential location is a control variable that the first author
will use in the analysis. In the survey data, it covers all the
provinces and municipalities in China except Qinghai, Tibet
and Ningxia, which are all located in the west. The first author
recoded the cities by geographical location into three categories:
eastern coastal (=1), central (=2), and western regions (=3).
In the study, the dependent variable is the natural logged
personal total income in 2005. The independent variables in-
clude gender, education level, foreign language skill, years of
work experience, party membership, type of workplace, type of
“Hukou” and residential location. The analyses rely on OLS
regression to predict the total individual income in urban China.
In the analysis, I attempt to find out “trends in the importance
of individual-level earnings determinants and their cones-
quences for trends in overall inequality” (Hauser & Xie, 2003:
p. 52).
Methods
In order to estimate the relationships between the logged an-

Q. WU ET AL.
nual income and several predictors including gender, work
experience, education, foreign language skill, party member-
ship, type of “Hukou”, geographical location, and workplace,
my analyses rely on Ordinary Least Squares (OLS) regression
to predict total individual income in urban China.
Before developing a multiple regression, the first author did
several preliminary analyses, including univariate descriptive
analysis, bivariate scatterplots of the income with age and years
of education. Table 2 summarizes the descriptive statistics of
all the variables in the analysis (see Table 2).
The mean of personal total yearly income in 2005 is
18383.343 RMB (yuan), the standard deviation is 23214.25.
The mean of education level is 4.8378, which roughly reaches
high school level, and the standard deviation is 1.185. The
mean of level of foreign language skill is 1.5873 (approxima-
tely the level of knowing a little of foreign language), and the
standard deviation is .58826. Among all the respondents, there
are 17.7% are members of the Communist Party of China or the
Table 2.
Descriptive statistics.
N Mean
Standard
Deviation
Standard
Error
Total Income 2005 3109 18383.343 23214.25416.336
Education Level 3109 4.8378 1.185 .0213
Foreign Language Skill 3109 1.5873 .58826.01055
N Percent
Party Membership
1 = Communist Party &
Communist
Youth League
550 17.7
Gender
1 = male 1697 54.6
2 = female 1412 45.4
“Hukou”
1 = Small cities 844 27.1
2 = Middle cities 635 20.4
3 = Large cities 950 30.6
4 = Rural 680 21.9
Workplace
1 = Government Agencies
and SOEs 1020 32.8
2 = Collective Enterprises 334 10.7
3 = Private Enterprises 993 31.92
4 = Foreign Enterprises 48 1.52
5 = Institutions 519 16.68
6 = Social Organizations 74 2.38
7 = Others 122 3.94
Residential Location
1 = Eastern Coastal Regions 1731 55.7
2 = Central Regions 880 28.3
3 = Western Regions 498 16
Note: used the results from averaging the five imputations. Source: data from
CGSS, 2006.
Communist Youth League, 82.3% are from other political par-
ties, and those who do not belong to any parties. There are
54.6% of males, and 45.4% of females. For the type of “Hu-
kou”, 27.1% are from small cities, 20.4% are from middle-size
cities, 30.6% are from large cities, and 21.9% hold the rural
“Hukou”. In terms of the type of work place, 32.8% of the re-
spondents work at government agencies or state-owned enter-
prises, 10.7% work at collective enterprises, 31.92% are em-
ployed at private enterprises, 1.52% work for foreign enter-
prises, 16.68% work at institutions, 2.38% work at social or-
ganizations, and 3.94% work for other workplace.
Then the first author ran the regression model and tested the
residuals for normality, and found that the residuals of the de-
pendent variable income are not normal distributed based on a
significant Kolmogorov-Smirnov test. Accordingly, I logged
income, and used lnincome as the dependent variable in subse-
quent analyses. Though according to the residual of the regres-
sion model using the natural logged income variable were still
not perfectly normal distributed, the distribution looked much
closer to normal. With only a slight departure from normality
and a very large sample size, the first author is confident that
the results of the regression analysis are robust.
Then the first author generated new scatterplots with the
logged income, and found a nonlinear relationship between
logged income and years of work experience. Thus, the first
author used curve estimation to check for the nonlinearity. By
doing the curve fit analysis and incremental F-test between
linear and quadratic models; the first author found that the
quadratic model is the best in this case. After detecting and
correcting for nonlinearity, I ran a regression and performed the
White’s test for homoskedasticity and found that the first author
needed to correct for heteroskedasticity using weighted least
squares regression which yielded homoskedastic residuals.
According to the results of collinearity diagnostics, all the
indexes, including VIF, square root of VIF, Tolerance, Eigen-
value, and condition index, show that there is no problem of
multicollinearity when excluded the variable workexp.
Results
Having fulfilled all the assumptions of OLS regression and
corrected for the violation, my regression now is the best linear
unbiased estimator.
Table 3 presents the main results from the final regression
model with location as the control variable. From the table, we
can see that the adjusted R2 is .2652, which indicates that
26.52% of the variation in logged income in 2005 is explained
by the sets of independent variables. Also, R is .5192, which
shows that there is a statistically significant and moderate rela-
tionship between logged income in 2005 and the sets of inde-
pendent variables (See Table 3).
Table 3 also shows the coefficients of each independent
variable. The unstandardized slope B for Education is .2136.
Taking the antilog and multiplying by 100, shows that for each
additional level of education, there is a 23.8 percent increase in
earning. The unstandardized slope B for Lanskill is .0742. Tak-
ing the antilog and multiplying by 100, shows that for each
additional level of foreign language skill, there is a 7.7 percent
increase in earnings. The unstandardized slope B for Female is
–.2582. Taking the antilog and multiplying by 100, shows that
females earn 22.8 percent less than males. The unstandardized
slope B for Small is –.2654. Taking the antilog and multiply-
376

Q. WU ET AL.
Table 3.
Regression results for LN (Income05) with location as control variable.
Variable B SE B Beta T Sig
(Exp(B) – 1)*100
Education .2136 .0156 .309 13.813 0** 23.8
Lanskill .0742 .0272 .0544 2.7176 .0158* 7.7
Small –.2654 .0348 –.143 –7.6298 0** –23.3
Mid –.1458 .0358 .0754 4.0694 .0002** –13.6
Rural –.1668 .0476 .0686 3.4948 .0016** –15.4
female –.2582 .0268 –.1562 –9.6698 0** –22.8
Private .0898 .0382 .0464 2.3554 .0456* 9.4
Foreign .554 .1126 .0808 4.966 0** 74.0
Central –.3014 .0314 –.1676 –9.5776 0** –26.022
Western –.3802 .0376 –.1726 –10.119 0** –31.6
(constant) 8.8012 .0916 95.873 0 **
Collective –.0806 .0458 –.0308 –1.7658 .0974
Institution .0252 .0356 .013 .7056 .497
Socialorg –.2042 .1074 –.0302 –1.8678 .128
Nonccp –.072 .0348 –.0352 –2.0518 .0738
Other –.01 .093 .0076 –.4274 .2584
Workdev .0008 .002 .0092 .427 .6752
Workdev2 0 0 –.0256 –1.26 1.121
R .5192
Adjusted R2 .2652
Std. Error of the Estimate 1.00427
Note: used the results from averaging the five imputations. *p < .05, **p < .01. Source: data from CGSS, 2006.
ing by 100, shows that that those who have the urban “Hukou”
of small cities tend to have 23.3 percent lower income than
those who hold the urban “Hukou” of large cities. The unstan-
dardized slope B for Mid is –.1458. Taking the antilog and
multiplying by 100, shows that that those who have the urban
“Hukou” of middle cities tend to have 13.6 percent lower in-
come than those who hold the urban “Hukou” of large cities.
The unstandardized slope B for Rural is –.1668. Taking the
antilog and multiplying by 100, shows that that those who have
the rural “Hukou” tend to have 15.4 percent lower income than
those who hold the urban “Hukou” of large cities. The unstan-
dardized slope B for Private is .0898. Taking the antilog and
multiplying by 100, shows that those who work at private en-
terprises or engage in the private business earn 9.4 percent
more than those who work for government and SOEs. The un-
standardized slope B for Foreign is .1126. Taking the antilog
and multiplying by 100, shows that those who work at foreign
enterprises, including the enterprises from Hong Kong, Macao,
and Taiwan, earn 74 percent more than those who work for
government and SOEs. The unstandardized slope B for Central
is –.3014. Taking the antilog and multiplying by 100, shows
that those who live in the central regions earn 26.02 percent less
than those who live in the eastern coastal areas. The unstan-
dardized slope B for Western is –.3802. Taking the antilog and
multiplying by 100, shows that those who live in the western
regions earn 31.6 percent less than those who live in the eastern
coastal areas. The rests of predictors, Collective, Institution,
socialorg, nonccp, Other, Workdev, Workdev2, are not statisti-
cally significant (p > .05).
Table 4 displays the OLS regression coefficients for the
model without geographic variables. In Table 5, I report the
OLS regression estimates for two models of income determina-
tion. Model 1 is a model with all the predictors. In Model1,
only the variables education level, foreign language skill, Hu-
kou dummies, Gender dummy, Workplace dummies (Private
and Foreign) have significant effects on earnings. In Model 2, I
exclude place of residence as a set of dummy variables and find
that the estimates of all the predictors increase slightly, but
variables party membership dummy, work experience, and
workplace dummies (Collective, Institution, socialorg, Other)
are not statistically significant (See Table 4).
Based on my results, in both models (See Table 5), educa-
tion is the best indicator to predict personal income, and in my
findings, education has a rate of 24.8%, which is much higher
than previous estimates (Xie & Hannum, 1996; Wu & Xie,
2002; Zhou, 2000). In addition, as part of education, foreign
language skill enjoys a 7.7-percent advantage, which also con-
firm the significance of human capital in determining earnings.
Work experience, another conventional measurement of hu-
man capital, has no linear relationship with the dependent vari-
able in the regression model. After conducted curve estimation,
I set up a quadratic model for work experience by computing
workdev and workdev 2. However, the result shows that work-
dev and workdev 2 are not significant. Thus, overall, work ex-
perience is not significant in either model. This result is differ-
ent from Xie and Hannum’s findings that work experience has a

Q. WU ET AL.
Table 4.
Regression results for LN (Income 05) without location as control variable.
Variable B SE B Beta T Sig
(Exp(B) – 1)*100
Education .2216 .016 .3196 13.992 0** 24.8
Lanskill .1076 .0276 .0786 3.8708 .0004** 11.4
Small –.34 .0348 –.183 –9.761 0** –28.8
Mid –.2414 .0358 –.124 –6.7642 0** –21.4
Rural –.2032 .0488 –.083 –4.1594 0** –18.4
female –.254 .0272 –.1538 –9.307 0** –22.4
Private .1238 .0392 .0634 3.1598 .0064** 13.2
Foreign .618 .114 .0918 5.5116 0** 85.5
(Constant) 8.5718 .0918 93.3516 0**
Collective –.035801 .0456 –.014 –.791 .4778
Institution .0292 .0364 .0156 .805 .437
Socialorg –.1396 .1092 –.0206 –1.2568 .3308
Nonccp –.061 .036 –.0296 –1.69 .1432
Other .0158 .0948 –.0064 –.334 .0828
Workdev –1.47E – 05 .002 .0012 .0464 .861
Workdev2 –5.16E – 05 0 –.0076 –.3762 .6888
R .4828
Adjusted R2 .233
Std. Error of the Estimate 1.00349
Note: used the results from averaging the five imputations. *p < .05, **p < .01; Source: data from CGSS, 2006.
Table 5.
OLS coefficients from multiple linear regression of logged income in 2005 on selected independent variables and control variables.
Variable Model 1 (geographic variables controlled) Model 2
Education level .2136** .2216**
Foreign language skill .0742* .1076**
Hukou dummy (small = 1) –.2654** –.34**
Hukou dummy (mid = 1) –.1458** –.2414**
Hukou dummy (rural = 1) –.1668** –.2032**
Gender (female = 1) –.2582** –.254**
Workplace dummy (private = 1) .0898* .1238**
Workplace dummy (foreign = 1) .554** .618**
Residential location dummy (central) –.3014** -
Residential location dummy (western) –.3802** -
Workplace dummy (collective = 1) –.0806 –.0358
Workplace dummy (institution = 1) .0252 .0292
Workplace dummy (socialorg = 1) –.2042 –.1396
Party dummy (nonccp = 1) –.072 –.061
Workplace dummy (other = 1) –.01 .0158
Work experience (workdev) .0008 –1.47E–05
Work experience (workdev 2) 0 –5.16E–05
(Constant) 8.8012 8.5718
R .5192 .4828
Adjusted R2 .2652 .233
Note: *p < .05, **p < .01. Source: data from CGSS, 2006.
378

Q. WU ET AL.
positive but concave effect on logged income. Thus, I partially
approve my hypothesis that education has the greatest impact in
determining income distribution, while work experience does
not show much significance.
Beyond my expectation, party membership is not significant
in either model. This suggests that party membership has little
impact on earnings, and weak support for hypothesis 2. Com-
pared to government agencies and state-owned enterprises,
where accumulate the redistributive power and political capital,
collective enterprises, public institutions and social organiza-
tions, which have more or less connections or relationships
with the state reveal no remarkable advantages in earnings.
However, private sector (private and foreign enterprises) dem-
onstrates considerable disparity on income. Beyond the regional
income differences in urban China, “the gap incomes between
the different state and non-state sectors has become more im-
portant in explaining social inequality as whole, with the rapid
growth of the foreign-invested and domestic private econo-
mies” (Guan, 2001: p. 246).
My findings also suggest that gender difference in earnings is
also estimated to be large, with females earning 22.8 percent
less than males. “Hukou” is still playing a crucial role in that
large cities’ residents earn 23.3 percent more than “Hukou”-
holders in small cities, 13.6 percent more than citizens in mid-
dle cities, and 15.4 percent more than those who originally from
rural areas. Regional income disparities are also evident. Resi-
dents in eastern coastal areas tend to earn 26.02% more than
those who live the central China 31.6% more than the people in
the west.
I do find high returns to education, but fail to find high re-
turns to work experience and party membership. And I did not
find the significant effect on work unit sector and state owner-
ship either. These findings are consistent with Nee’s prediction
that the significance of political power declines with the proc-
ess of the marketization, and “the income determination will
depend more on market credentials (such as education), and
less on political factors as economic reform advances” (Xie,
2008: p. 195).
Discussion
In this paper, I have examined the determinants of income in
urban China based on the data of 2006. My hypotheses regard-
ing the role of educational credentials was generally supported
in both analyses and held up when various controls were intro-
duced. According to the results from the regression models,
working at market sector firms, especially foreign enterprises
are the most predominant in determining the income distribu-
tion in urban China.
Does Political Capital or Power Really Decline
Significantly?
Returns to political capital or power “is operationalized in
three ways: a) party membership, b) cadre position, and c ) jobs
with redistributive power” (Bian, 2002: p. 100). In China, not
everyone can become a member of Communist Party. There are
mainly two ways to apply for a membership of Chinese Com-
munist Party. One way is that first one should join the Commu-
nist Youth League in middle school or high school, and until
when he becomes an adult (≥18 years old) and enters a college
or university, he can write an application letter to show his
desire and loyalty to the party. A party membership can be an
advantage to find a job in government or party agencies after
graduation. Another way to be a party member is to apply at
work units, such as public institutions, SOEs. For both ways,
“to achieve Chinese Communist Party membership, individuals
must pass through five ‘loyalty filters’ (Walder, 1995): 1)
self-selection; 2) political participation; 3) daily monitoring; 4)
closed-door evaluation; and 5) probationary examination”
(Bian, Shu, & Logan, 2001: p. 813). Nowadays, the Chinese
Communist Party tends to recruit educated youths and profes-
sional, which indicates that the role of educational credentials
has become more and more important.
While variables related political capital did not turn out to be
significant, things does not mean that party membership ceases
to be an important factor in determining income. For example,
“grey income is not included in the survey data and the limita-
tion of my current research that does not partition cadre posi-
tion into the party officials, government bureaucrats, and man-
agers in SOEs”.
Income distribution in the foreign enterprises and private
companies are directly reflected in salaries, while in the gov-
ernment agencies and SOEs, the base wages may be lower than
the workers in foreign and private enterprises, but the hidden
bonuses and other forms of welfare benefit including allowance
for transportation as well food, a housing packages, medical
insurance, unemployment insurance and annuity. Moreover,
many SOEs assumed monopoly positions in the new market
economy after the structural reforms. Those monopolized en-
terprises, such as China Mobile, State Grid, China Telecom and
China National Petroleum Corporation occupy the most impor-
tant and profitable industries, such as mining industry, banking,
communication and telecom. With the powerful supporting
polices and ample and stable financial support from the state,
the profits of these SOEs rose tremendously given the size and
importance of these enterprises in the state sector it would be
hard to conclude that political capital has no influence on in-
come.
Moreover, the “grey income” of the state bureaucrats has
great widen the income gap that 54% of the respondents of
CGSS, 2006 recognize the huge gap between the cadre and the
mass (poor vs rich has 57.7%). In light of this, most people do
realize the existence of the “grey income”. According to Xiaolu
Wang’s research, “the government’s statistics omit roughly
RMB 9.26 trillion (about US$1.36 trillion) in ‘invisible’ in-
come—that is, money earned illegally and under the table or
not declared to tax authorities”
(http://www.knowledgeatwharton.com.cn/index.cfm?fa=viewA
rticle&Articleid=2284&languageid=1).
What’s more, “as private economic activities became legal
and market competition played a greater role in economic op-
erations, people with more human capital and political capital
began to be involved in business activities. Some cadres also
managed to convert their political privileges into new economic
advantages in this stage” (Wu, 2006: p. 391). In CGSS, 2006,
there is question asking “comparatively, speaking, in the recent
decade, which group of people in the following do you think
obtain the most benefit?” 38.5% of the respondents think state
cadres gain the most, 20.8% claim that it is private entrepre-
neurs, and 15% favor in foreign investors. Based on the an-
swers, we can clearly find that most people still deem that the
state cadres who hold the political capital and power benefit the
most. Even in the market system, the state cadres can transfer
their political power and skills to revive in the new economy.

Q. WU ET AL.
This is consistent with my third hypotheses of the technocratic
continuity. Thus, I advocate that not only capitalists are the
winners of the market transition in China, cadre still gain bene-
fits but not as remarkable as in the pre-reform era.
Impact of Marketization and Globaliza tion on
Income Inequality
Since the reform, especially after 2001 when China joined
the World Trade Organization (WTO), an increasing foreign
trade and investment has flown into Chinese market. Along
with this trend, the impacts of globalization and marketization
from the exterior forces have greatly influenced the patterns of
income equality.
First, from the Table 6 below, we can see that foreign in-
vestment is unevenly distributed which, to great extent, leads to
the regional income gap. There are 87.46% of foreign enter-
prises investing in the eastern coastal areas, while central and
western areas all together share 12.45%. To the extent that the
unbalanced development pace and unequal policy support in the
initial stage of the reform opened the gap between regions, then
the involvement of foreign investment has greatly increased the
disparity.
Second, with more and more foreign-owned enterprises en-
tering Chinese market, many SOEs face more challenges and
competitions. From my regression result, we can clearly find that
those who work at foreign companies earn much more than any
others on average. Moreover, the income advantage in SOEs that
gain all kinds of support from the state has declined greatly.
Third, “income inequality within foreign-invested enterprises
is generally much higher than in state and collective enter-
prises” (Guan, 2001: p. 249). According to a survey conducted
in Shanghai in 2005, the average annual wages of the highest
level managerial personnel, such as Chief Executive Officer
(CEO) and Chief Finance Officer (CFO), earn “over 400,000
yuan, which is 13.68 times higher than the ordinary workers
who only earn 28,000 yuan yearly”
(http://www.ccw.com.cn/work2/culture/clcw/htm2006/2006020
8_13SBO.htm). In the foreign enterprises, the unequal salary
structure is considered as a way to stimulate high efficiency
under the market mechanism. Thus, SOEs also adopted this
method during the structural reform in the mid-1990s and early
2000s, which further widen the income gap within the market
sector.
Table 6.
Regional distribution of the foreign-invested enterprises in China
(2005).
Number of foreign invested
enterprises (unit)
Total investment
(100 million USD)
Regions
No. % No. %
Eastern Coastal* 227401 87.46192 12729 86.9586
Central** 21464 8.255385 1393 9.516327
Western*** 11135 4.282692 516 3.525072
National Total 260000 100 14638 100
*Includes: Beijing, Tianjin, Hebei, Liaoning, Jilin, Heilongjiang, Shanghai, Ji-
angsu, Zhejiang, Fujian, Shandong, Guangdong, and Hainan; **Includes: Shanxi,
inner-Mongolia, Anhui, Jiangxi, Henan, Hubei, Hunan, Guangxi, and Chongqing;
***Includes: Sichuan, Guizhou, Yunan, Tibet, Qinghai, Shaanxi, Gansu, Ningxia,
and Xinjiang. Source: “Chinese Statistics Yearbook (2006), Calculated from the
data in the Tables 18-19 in 2006 Chinese Statistics Yearbook, Chinese Statistic
Publishing House (See references)”.
Conclusion
By using new data from 2006 China General Social Survey
(CSSS, 2006), I conduct an OLS regression analysis on the
logged annual income and gender, work experience, education,
foreign language skill, party membership, type of “Hukou”,
geographical location, and workplace. The results of the OLS
regression analysis suggest that there is estimated to be a large
gender-based difference, “Hukou” discrimination and regional
disparity in earnings.
The empirical results also reveal that education matters more
while the political advantage of party membership drops, so do
state ownership or non-market workplaces. This finding pro-
vides evidence to support Nee’s theory that market transition
lead to “a decline of the significance of redistributive power
and political capital, relative to market-based non-state eco-
nomic actors, higher return to human capital than under a cen-
trally planned economy, and new sources of economic advan-
tage associated with entrepreneurship and hybrid/private sec-
tor employment” (Nee & Cao, 1999: p. 807).
While my findings imply that political capital is less impor-
tant, I am not ready to reject the role of party membership in
determining earnings. First of all, there is a large deal of invisi-
ble income (grey income) and all kinds of welfare benefit
which are not covered in the survey data, I cannot simply rely
on the results from data analysis to make conclusions. Second,
my research is limited in that it 1) excludes the variables of
occupation and cadre status; 2) parental party membership,
parental education level, and the parental social capital link; 3)
“grey income” sources; and 4) welfare benefit.
For further research, I would like to take the variables of oc-
cupation and cadre status; take parental party membership,
parental education level, and the parental social capital link
(e.g., education) and how that turns into more market power
into account to improve the model, and investigate more in the
part of “grey income” and welfare benefit.
Acknowledgements
This work is based on Qiong Wu’s Master’s thesis. Wu’s
gratitude goes to her thesis committee, Dr. Barry Goetz, Dr.
David Hartmann, and Dr. Yuan-Kang Wang. They have been
her inspiration as she hurdles all the obstacles in the completion
this research work for the support and guidance. Also, Wu
thanks China General Social Survey Open Database to provide
the data of 2006 for free.
REFERENCES
Bian, Y. (1994). Work and inequality in urban China. Albany, NY:
State University of New York Press.
Bian, Y., & Logan, J. R. (1996). Market transition and the persistence
of power: The chaning stratification system in urban China. Ameri-
can Sociological Review, 61, 739-758. doi:10.2307/2096451
Bian, Y., Shu, X., & Logan, J. R. (2001). Communist party membership
and regime dynamics in China. Social Forces, 7 9, 805-841.
doi:10.1353/sof.2001.0006
Guan, X. (2001). Globalization, inequality and social policy: China on
the threshold of entry into the World Trade Organization. Social
Policy and Administration, 35, 242-257.
doi:10.1111/1467-9515.00231
Hauser, S., & Xie, Y. (2005). Temporal and regional variation in earn-
ings inequality: Urban China in transition between 1988 and 1995.
Social Science Research, 34, 44-79.
380

Q. WU ET AL.
doi:10.1016/j.ssresearch.2003.12.002
Lewis, W. A. (1976). Development and distribution. In A. Cairncross,
& M. Puri (Eds.), Employment, income distribution and development
strategy. London: Macmillan.
National Bureau Statistics (2006). China statistical yearbook. Beijing:
China Statistical Publishing House.
Nee, V. (1989). A theory of market transition: From redistribution to
markets in state socialism. American Sociological Review, 54, 663-
681. doi:10.2307/2117747
Rona-Tas, A. (1994). The first shall be last? Entrepreneurship and
communist cadres in the transition from socialism. American Journal
of Sociology, 100, 40-69. doi:10.1086/230499
Tang, W., & Parish, W. L. (2000). Chinese urban life under reform:
The changing social contract. Cambridge: Cambridge University
Press.
Victor, N., & Cao, Y. (1999). Path dependent societal transformation:
Stratification in hybrid mixed economies. Theory and Society, 28,
799-834. doi:10.1023/A:1007074013540
Wang, F. (2008). Boundaries and categories: Rising inequality in
post-socialist urban China. Stanford, CA: Stanford University Press.
Wu, X. (2006). Communist cadres and market opportunities: Entry into
self-employment in China, 1978-1996. Social Force, 85, 389-411.
doi:10.1353/sof.2006.0149
Wu, X., & Treiman, D. J. (2004). The household registration system
and social stratification in China: 1955-1996. Demography, 41, 363-
384. doi:10.1353/dem.2004.0010
Wu, X., & Xie, Y. (2003). Does the market pay off? Earnings Returns
to education in urban China. American Sociological Review, 68, 425-
442. doi:10.2307/1519731
Xie, Y., & Hannum, E. (1996). Regional variation in earnings inequal-
ity in reform-era urban China. American Journal of Sociology, 101,
950-992. doi:10.1086/230785
Xie, Y., & Wu, X. (2008). Danwei profitability and earnings inequality
in urban China. The China Quarterly, 1, 558-581.
Xinhua (2004). Survey of Chinese officials’ opinions on reform: Bei-
jing Daily. Xinhua Ne w s Bulletin.
Zhou, X. (2000). Economic transformation and income inequality in
urban China: Evidence from panel data. American Journal of Soci-
ology, 105, 1135-1174. doi:10.1086/210401