+Model RSSM-92;

No. of Pages 18

ARTICLE IN PRESS
Available online at www.sciencedirect.com

Research in Social Stratification and Mobility xxx (2010) xxx–xxx

Economic transition, school expansion and educational inequality in China, 1990–2000
Xiaogang Wu ∗
Social Science Division, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong SAR, China Received 16 January 2008; accepted 23 December 2009

Abstract This paper examines the trends in educational stratification during China’s economic reforms in the 1990s. Based on the sample data of population censuses in 1990 and 2000, school-age children were matched to their parents’ background information, and the effects of family background on their school enrollment and continuation were investigated. Results show that despite the substantial expansion of educational opportunities in the decade, family background continues to play an important role in determining school enrollment status and school transitions. During the decade, children of rural-hukou status became more disadvantaged compared to their urban counterparts, and the effect of their father’s socioeconomic status on school enrollment was enhanced. Despite the fact that children of rural-hukou status gained relatively more opportunities at junior high school level, as a result of nationwide saturation at the 9-year compulsory education, the rural–urban gap in the likelihood of transition to senior high school level enlarged, and the effect of their father’s socioeconomic status increased—even after controlling for regional variations in economic development. © 2010 International Sociological Association Research Committee 28 on Social Stratification and Mobility. Published by Elsevier Ltd. All rights reserved.
Keywords: China; Educational inequality; Market transition; Social stratification

This paper was presented at the ISA Research Committee on Social Stratification and Mobility (RC28) Spring Meeting in Brno, the Czech Republic, May 24–27th, 2007. The author would like to thank Sam Lucas of University of California, Berkeley, and other conference participants for comments and suggestions, and Yuxiao Wu and Zhigang Nie for their assistance in data analysis. Special thanks are due to the National Bureau of Statistics of the People Republic of China and John Z. Ma for his assistance in data access. This project is funded by a grant from Research Grants Council of Hong Kong (HKUST6424/05H) and a post-doctoral fellowship from National Academy of Education/Spencer Foundation. Please direct all correspondence to Dr. Xiaogang Wu, Social Science Division, the Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong (email: sowu@ust.hk). ∗ Fax: +852 23350014. E-mail address: sowu@ust.hk.

Education plays an important role in modern societies, both as an avenue of social mobility and as a tool for social reproduction. On the one hand, formal schooling can help children from disadvantaged backgrounds to change their fate; on the other hand, the schooling that individuals have received also depends on the advantages/disadvantages that their parents confer on them throughout childhood (Ishida, Muller, & Ridge, 1995). In other words, access to educational opportunities is unequally distributed among different social strata. The increasing importance of education, together with long-term growth of enrollment in a school system of a country in the process of economic development, has led some scholars to claim that educational achievement has become more and more independent of family background (Boudon, 1974; Treiman, 1970). However, linear regression analyses of educational attainment reveal that

0276-5624/$ – see front matter © 2010 International Sociological Association Research Committee 28 on Social Stratification and Mobility. Published by Elsevier Ltd. All rights reserved.

doi:10.1016/j.rssm.2009.12.003

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;
2

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx

the effect of family background has stabilized over time in many industrialized countries (e.g. Featherman & Hauser, 1978). This is because the expansion of education and the distribution of educational opportunities are two separate processes (Mare, 1980): the former may not necessarily lead to more equal access to education among different social strata. The expansion of the school system in many countries in the 20th century, reinforced by educational reforms, seems to have had little impact on the role played by family background on an individual’s educational attainment (Shavit & Blossfeld, 1993). Just as income growth does not necessarily lead to a more equal distribution of income, educational expansion has no intrinsic implications on the change in educational inequality. Instead, the distribution of educational opportunities may resemble the distribution of other scarce resources that affect educational outcomes, which are both embedded in the fundamental social structure of a particular nation at a particular time. Since education plays an increasingly important role in attaining a better job and receiving more economic benefits in a modern society, the question of “who gets educated” assumes a central place in stratification research (Deng & Treiman, 1997; Shavit & Blossfeld, 1993). To understand the change of stratification outcomes in a society that is undergoing dramatic transformation in the mechanism of resource distribution, it is necessary to investigate how the transformation has altered the allocation of educational opportunities among different social strata, which may have a longterm impact on the change in social structure. The dramatic institutional changes in former state socialist countries stimulated a lively debate among sociologists in the 1990s on how the social stratification order is reshaped by the shift from state socialism to market capitalism as the main mechanism of resource distribution (Bian & Logan, 1996; Cao & Nee, 2000; Gerber & Hout, 1998; Nee, 1989; Parish & Michelson, 1996; Róna-Tas, 1994; Szelényi & Kostello, 1996; Xie & Hannum, 1996; Zhou, 2000). Much of the existing literature in this area, nonetheless, is largely focused on income outcomes (e.g. Bian & Logan, 1996; Gerber & Hout, 1998; Nee, 1989; Xie & Hannum, 1996; Zhou, 2000). Despite the growing importance of education (human capital) in determining income (e.g. Bian & Logan, 1996; Zhou, 2000) and controversial interpretations of the evidence (Xie & Hannum, 1996; Wu & Xie, 2003), few scholars have explicitly examined the impact of economic reforms on educational inequality per se—an important issue to examine if one would like to understand the changes in the patterns of job shifts,

career mobility, and intergenerational transfers in the era of market transition (Gerber & Hout, 2004; Walder, Li, & Treiman, 2000; Zhou, Tuma, & Moen, 1997). This study investigates the change in educational stratification in China’s late reform period, during which substantial socioeconomic transformations were undertaken. Based on the samples of population census data in 1990 and 2000, I match school-age children aged from 6 to 18 to their parents’ background information, and investigate the impact family backgrounds on their school enrollment and transitions over the decade. In particular, I focus on the effects of the household registration (hukou) status and father’s socioeconomic status on a child’s educational outcomes. In the remaining part of this paper, I will first provide the historical background on economic reforms and school expansion in China since the 1980s, and explain how the census data can be employed to address the temporal trend in educational inequality. I then demonstrate how family socioeconomic backgrounds have affected children’s educational outcomes in the context of economic marketization and school expansion. Finally, I discuss the implications of the change in inequality structure in reform-era China. 1. Economic reforms and school expansion Few nations have undergone changes as dramatic as China has since the 1970s. China’s GDP per capita has consistently grown from 379 RMB yuan in 1978 to 14,040 RMB yuan in 2005 (see Column A of Table 1). At a fixed price in 1978, the per capita GDP increased by 5.8 times in 2000 and 8.8 times in 2005, with an annual growth of about 9 percent (Nation Bureau of Statistics 2006). The economic growth has been especially phenomenal since 1992 when Deng Xiaoping called for further market reform in his famous tour to southern China. The market economy had been fully legitimized by the Chinese Communist Party’s ideology and started playing an increasingly important role in China’s economic growth. The government had retreated to a large extent from the provisions of housing, education, health care, and other social services in the 1990s. Accompanied with China’s economic miracle was a rapid growth of inequality. As Column C of Table 1 shows, the Gini coefficient, a common measure of income inequality, increased from 0.317 in 1978 to 0.449 in 2005 for the nation as a whole. Income inequality between urban and rural population, institutionalized by the household registration (hukou) system (Wu & Treiman, 2004, 2007), was particularly prominent: the

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx 3

Table 1 Selected indicators of economic growth and income inequality in China, 1980–2005. Year 1978 1980 1985 1990 1995 2000 2005 A. GDP per capita (RMB yuan) 379 460 853 1643 4854 6392 14,040 B: GDP per capita compared to 1978 price as 100 100.0 113.0 175.5 237.3 398.6 575.5 878.9 C: Gini index 0.317 0.295 0.331 0.357 0.290 0.390 0.449 D: urban–rural ratio of income ratio per capita 2.35 2.75 2.14 2.51 2.79 3.10 3.22

Data sources: A, B, D: Comprehensive statistical data and materials on 50 years of new China, China Statistics Publishing House, also available at http://www.stats.gov.cn/tjsj/ndsj/. C: World Income Inequality Database http://www.wider.unu.edu/wiid/wiid.htm.

urban–rural ratio of income per capita declined slightly in the early 1980s, but has increased dramatically since then, from 2.5 in 1990 to 3.1 in 2000 and 3.2 in 2005 (Table 1: Column D). Urban–rural income inequality has contributed 43 percent to overall income inequality in China (Cai & Wan, 2006, p. 3). Sociologists have always been interested in investigating who wins and who loses in the institutional transition (Nee, 1989; Szelényi & Kostello, 1996). While a large body of literature has been devoted to the discussion of changing returns to human capital (education) as a result of the market transition (Bian & Logan, 1996; Gerber & Hout, 1998; Wu & Xie, 2003; Xie & Hannum, 1996; Zhou, 2000), few scholars have explicitly examined the impact of economic reforms on unequal access to educational opportunities. Despite the fact that the pattern of educational stratification was relatively stable compared to the change in the distribution of economic resources, it was by no means immune to the economic reform in China, especially since the 1990s. Economic reform affected educational stratification in three respects. First, sustainable economic growth demanded skilled labor. The commencement of the reform era was marked by the complete dismantling of the educational policies adopted during the Cultural Revolution, which severely condemned the system of evaluating student performance by examinations (Tsui, 1997; Wang, 2002). Despite the fact that the pattern of educational attainment in China was found to vary across different historical periods associated with major shifts in government policies (Hannum & Xie, 1994; Zhou, Moen, & Tuma, 1998), educational inequality observed in the 1980s after the Cultural Revolution was largely seen as reflecting a return to the generic practice under socialism (Gerber & Hout, 1995; Simkus & Andorka, 1982; Wong, 1998), rather than the effect of market transition (Deng & Treiman, 1997; Tsui, 1997; Zhou et al., 1997).

Second, economic growth afforded more resources for educational development and school expansion. The government budgetary expenditure on education has been increasing dramatically since 1978 (see Table 2). In 1980, the Chinese government set the target of universalizing primary education by the end of the 1980s; and the implementation of 9-year compulsory education in the 1990s (Tsui, 1997). These goals were largely attained by 1998. As indicated in Fig. 1, the enrollment rate had reached over 98% in the 1990s. The progression rate to junior high school, given the completion of primary school education, was almost 100% by the mid-1990s; the progression ratio to senior high school given the completion of junior high school increased from 30% in the 1980s to 60% in 2005. Higher education has also been opening up since 1998 (Min, 2007). Within the next few years, the progression ratio to tertiary education, given the completion of senior high school, increased dramatically from 40% to 80%. Although there is no doubt that the central government intended to promote educational opportunities for all its citizens, economic reforms in rural areas slowed

Fig. 1. Educational expansion in China, 1978–2005.

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;
4

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx

Table 2 Government educational spending and educational expansion in China, 1978–2005. Year 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Government budgetary education expenditure (100 million yuan) 76.23 93.16 113.19 122.22 137.20 154.72 180.14 224.89 267.30 276.57 330.91 397.72 563.99 617.83 728.76 867.76 1174.74 1411.52 1671.70 1862.55 2032.45 2287.18 2562.61 3057.01 3491.40 3850.62 4465.86 –a Enrollment rate of school-age children % 95.5 93.0 93.9 93.0 93.2 94.0 95.3 96.0 96.4 97.2 97.2 97.4 97.8 97.8 97.2 97.7 98.4 98.5 98.8 98.9 98.9 99.1 99.1 98.3 98.6 98.7 98.9 99.2 Transition rate to junior high school % 87.7 82.8 75.9 68.3 66.2 67.3 66.2 68.4 69.5 69.1 70.4 71.5 74.6 75.7 79.7 81.8 86.6 90.8 92.6 93.7 94.3 94.4 94.9 95.5 97.0 97.9 98.1 98.4 Transition rate to senior high school % 40.9 40.0 45.9 31.5 32.3 35.5 38.4 41.7 40.6 39.1 38.0 38.3 40.6 42.6 43.4 44.1 46.4 48.3 48.8 44.3 50.7 50.0 51.1 52.9 58.3 60.2 62.9 69.7 Transition rate to tertiary school % 27.3 28.7 34.9 43.3 46.7 49.9 51.0 48.6 46.1 63.8 73.2 78.8 83.5 83.4 82.5 76.3

Sources: Comprehensive statistical data and materials on 50 years of new China, Beijing: China Statistics Publishing House. The data after 1998 from http://www.stats.gov.cn/tjsj/ndsj/. a Data unavailable for this year.

down progress to a certain extent and yielded a negative impact on school enrollments. On the one hand, the household responsibility system implemented in rural China since 1978 drove rural children out of school for agriculture labor and employment in the rural industry (as shown in the decline in the school enrollment rate in the mid-1980s in Fig. 1, even though the aggregate government statistics would not allow a breakdown by rural and urban areas). Moreover, the fiscal reform in education in the early 1990s exacerbated the situation. In the context of the decentralization of public finances in China since the early 1980s, the responsibility of funding primary and secondary education was shifted to local governments who had a strong incentive to invest in projects that could quickly reap profits and generate tax revenues, resulting in a low priority for investment in education. The uneven regional economic development further differentiated local governments’ capacity in funding education. In many poor and rural areas, local governments could hardly raise suf-

ficient revenue to cover teachers’ salaries, not to mention other non-instructive costs. In contrast, local governments in developed areas could mobilize significantly more resources, both government and non-government, for education (Tsang & Ding, 2005). This has resulted in the substantial disparities in per-student educational expenditure across areas and regions.1 Hence, to accommodate the increasing number of enrollments and increasing educational costs, schools have been allowed to charge tuition and other fees, even for 9-year compulsory education. For example, in 1999, the surcharges and miscellaneous fees together accounted for 62% of all out-of-budgeted revenue for

Among the 2070 Chinese counties and county-level cities (containing rural population) in 2000, the educational expenditure per capita in 2000 ranges from 3.4 RMB Yuan to 1,474 RMB Yuan, with an average of 164 RMB Yuan and standard deviation of 94 RMB Yuan (1 RMB yuan ≈ 0.128 USD) (Ministry of Education and National Bureau of Statistics, 2002).

1

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx 5

primary schools and 57% of that for lower secondary schools (Tsang & Ding, 2005: Table 5). Recent surveys conducted by some sociologists in selected rural counties revealed that Chinese farmers with an annual per capita net income of 3200 yuan in 2005 had to pay about 800 yuan a year for a child’s education in primary and lower secondary education. Excessive charges by schools have become a major reason behind the increasing rural school dropouts in recent years. In 2004, the rural average dropout ratios for primary and junior high schools were 2.45% and 3.91%, respectively. Schools charged even higher for schooling beyond the compulsory levels, thereby economic considerations significantly affect the decision to continue schooling (Min, 2007). Such policy reforms have had important implications on how family socioeconomic resources affect children’s educational opportunities in China’s expanding school system. Educational affordability has become one of the greatest public concerns (Kahn & Yardley, 2004). There is also a reported decline in the number of student enrollments from disadvantaged family backgrounds at several elite universities (Liu, 2004; Min, 2007; Yang, 2006). In the era of rapid educational expansion and economic marketization in the 1990s, how are increasing educational opportunities distributed among different social groups? Based on the analyses of school enrollment and transitions in the population census data of China in 1990 and 2000, this paper will examine the recent trend in the impact of family background on educational opportunities in reform-era China. 2. Social differentiations in access to educational opportunities: research hypotheses Regarding the consequences of educational expansion on educational inequalities, early scholars argued that if school attendance rates increased over time, the inequalities in educational opportunity would decline steadily, because children from disadvantaged backgrounds could increase the attendance rates by a larger percentage than those from the upper classes whose rates were already high (Boudon, 1974). This prediction, however, has received little empirical support. Instead, linear regression analyses of educational attainment reveal that the effect of family backgrounds has been stable over time in many industrialized countries (Featherman & Hauser, 1978). Mare (1980) distinguished the processes of selection and allocation of students from the expansion of the educational system per se, and proposed a logit model of change in inequality of educational opportunity whose parameters are not affected by the degree of educational

expansion. Comparative studies of educational attainment in 13 industrialized societies have confirmed that the logit effects of social origins on educational transitions remain largely stable across cohorts, even in the context of long-term educational expansion (except for Sweden and the Netherlands where the effects of father’s occupation and education on the low and intermediate transition decline).2 Most relevant to Chinese educational inequality are the cases in former state socialist countries. Simkus and Andorka (1982) analyzed educational stratification in Hungary for the period from 1923 to 1973 and reported an actual decrease of the effect of social origins on earlier transitions, accompanied by stable effects in the later transition. Mateju found similar results in Czechoslovakia (op. cit. Shavit & Blossfeld, 1993). These results suggest that the institutional shift to state socialism immediately after the revolution, along with the educational expansion, does bring more equality in school transitions at lower levels (also see Russia in Gerber & Hout, 1995) for a certain period of time, but educational stratification would subsequently resume to the normal order, in which family background would exert a stable influence, as found in many other modern societies. The above analysis of educational stratification under socialism did not cover the market transition era, when the institutional mechanism of distributing educational resources was undergoing a dramatic shift. Using the data collected in 1998, Gerber (2000) extended an earlier study of educational stratification in Russia (Gerber & Hout, 1995) and reported that the political chaos and economic crisis in transitional Russia increased the magnitude of origin-based inequalities in access to academic secondary schools for the cohorts who completed their education in the tumultuous late-Soviet and post-Soviet years when school enrollment contracted. Evidence from all countries other than post-Soviet Russia demonstrated either a stable effect or declining effect (for some welfare states and state socialist countries) of family origins on educational attainment. Together with the post-Soviet Russia case, it suggests that the distribution of educational opportunity is more related to the rules that govern the educational selection rather than the expansion of the education system per se. The former, to a large extent, is reflected in the broader inequality structure of a society. Hence, even without post-Soviet Russia’s experience of enrollment

2 These countries include USA, West Germany, England, Wales, Italy, Switzerland, the Netherlands, Sweden, Japan, Taiwan, Poland, Hungary and Czechoslovakia.

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;
6

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx

contraction, the educational expansion in China in the 1990s may not necessarily lead to more educational equality. Instead, the rapid marketization of education and withdrawal of the state in provision of education as a public good may lead to more unequal access to the enlarged educational pie. Hence, it is expected that the effect of family background on educational opportunity will increase over time in China. Educational expansion, if it has an effect, only increases equality at low level of transitions. Given the change of the inequality structure in reformera China described above, in the following analysis, I will focus on the trend in the effects of household registration status and father’s socioeconomic background on school enrollment/continuation and transitions from 1990 to 2000 for young cohorts between 6 and 18 years old in respective years. 3. Data, variables and methods

From the variable “relationship to the household head,” a child’s father and mother can be identified; the child’s individual records can be matched to his/her parent’s occupational and educational backgrounds and used as the main measures of family background. Other individual characteristics (gender and nationality) and household characteristics (hukou type) are available for the multivariate analyses in both census data sets. The data analyzed here is a sub-sample (0.1%) of the micro-data of population censuses in China in 1990 and 2000. I first extract those individuals aged between 6 and 18, and then match them with their parents or caretakers, based on the variable indicating the relationship of the respondent to the household.4 As a result, children–parent (or caretaker) records, as well as household records encompassing geographical location, household registration (hukou) status, father’s education and occupation, gender, and ethnicity were all obtained. 3.2. Variables

3.1. Data As far as we know, no national survey data are available on young cohorts who completed their primary and secondary education in the period when China proceeded deeply into marketization, including the marketization of the educational system. This paper analyzes a sample of micro-data from the China population censuses in 1990 and 2000. The decennial census is a unique tool studying social changes, because it provides a rich set of data for the detail analysis of social and demographic groups. For the most part, the census employs a constant set of measures for each decade, thereby avoiding the problem of confusing changes in the population in the way that the population is measured (Mare, 1995). The 1990 Chinese census data includes two variables on education: educational level and enrollment status, which can be combined, together with age/cohort information, to define whether a person of a certain age group (6–18) is enrolled in school or not. While the questions on education in the 2000 census are slightly modified, the variables are basically comparable to those in 1990.3 The dependent variable is the enrollment status and transition of the young cohorts at certain ages, which is coded as a dummy variable. Given the fact that primary school education is almost saturated in both rural and urban China, I focus on the determinations of enrollment status at secondary school level (junior high school and senior high school). While tertiary enrollment is of great interest, family background information for most tertiary students are not available in the census, because most college students moved out of their parents’ homes to live in student dormitories where their universities are located. In addition to school enrollment, I also examine the transition rate at two specific levels, from primary to junior high school, and from junior high school to senior high school. From 1990 to 2000, the Chinese school system remained largely the same. As Fig. 2 shows, a student typically starts school at age 7, proceeds to junior high school at 13 after 6 years of primary school, and then proceeds to senior high school/vocational school at age 16. Because there is no information about the particular grade/level that a student is attending, I approximate the transition rate at specific levels by referring to respon-

3 Educational questions differed slightly in 1990 and 2000 censuses. For example, illiteracy/semi-illiteracy was a category of the educational attainment variable, while illiteracy was asked as a separate question in 2000. This discrepancy suggests that results for educational attainment for the same cohort in 1990 and 2000 may not be directly comparable. This paper deals with school enrollment rather than educational attainment for the relatively young cohorts, who are almost impossible to fall in the group of illiteracy/semi-illiteracy (see Hannum, 2005, p. 290).

Tabulation of the 2000 population census data shows that 90% of Chinese children aged between 6 and 18 years old are living with their parents. About another 7.8% of children live with grandparents as household head, 2.1% with others as household head. In cases where the parents’ information is not identifiable, the household head and spouse are used to replace father’s and mother’s characteristics. Children who are household heads themselves (1.2%) were excluded in the analysis.

4

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx 7

Fig. 2. Age-specific full-time school enrollment rates in China, 1990 and 2000.

dents’ age. For the transition to junior high school, it is defined as those aged between 13 and 15 still enrolled in school divided by those of the same age group who have completed primary school education (i.e. those in junior high school and those who completed primary school but were not enrolled in school). For the transition to senior high school (i.e. continuing school after compulsory education), it is defined as those aged between 16 and 18 still enrolled in senior high school divided by those of the same age group who have completed junior high school education, namely, those who are currently in senior high school plus those who have completed junior high school but are currently not in school. The main independent variable in the following analysis is family background, measured by father’s occupation, education, and mother’s education. The father’s occupation is converted into a socioeconomic status scale, which is a continuous variable. To make the measurement consistent, I first convert the Chinese standard classification of occupation to international standard classification of occupations (1968 version), and then map them to international socioeconomic index (Ganzeboom, de Graaf, & Treiman, 1992). Father’s education and mother’s education are measured in three levels (1 = primary school; 2 = junior high school; 3 = senior high school or above). They are treated as a set of dummy variables in the multivariate analysis. The effects of household registration (hukou) status were also taken into account. Hukou type captures not only the effect of family background, but also the regional inequality that reflects the fundamental divide in the country (Wu & Treiman, 2007). Hukou type indicates whether one holds agricultural (rural) or non-agricultural (urban) hukou. It is also coded as a dummy variable (rural = 1 and urban = 0). The available resources need to be distributed among all children in a family. While scholars have demonstrated that the number of siblings has a negative impact

on educational attainment in western society (e.g. Mare & Chen, 1986), the census data allows us to identify a child’s relationship only with the household head. And because of the Chinese one-child policy which has been strictly implemented since the early 1980s, the effect of sibling size was not considered in the following analysis. To capture regional variations in socioeconomic development (Wu & Ma, 2004), all 31 province-level jurisdictions in China have been grouped into three regions based on their levels of economic development: 1 = East; 2 = Middle; and 3 = West. The eastern region includes Liaoning, Beijing, Tianjin, Hebei, Shandong, Jiangsu, Shanghai, Zhejiang, Fujian, Guangdong, and Hainan. The middle region covers Heilongjiang, Jilin, Inner Mongolia, Shanxi, Henan, Anhui, Hubei, Hunan, Jiangxi, and Guangxi; and the rest of the provinces belong to the western region. There exist great disparities in the level of economic and social development as well as education among the three geographical regions. County-level statistics on GDP per capita in both 1990 and 2000, and educational expenditure in 2000 were also compiled. The educational expenditure statistical data for 1990 was not available. In addition to geographical region, residential type has been coded as a dummy variable (rural = 1 and urban = 0). Residential locale need not be identical to household registration status (see footnote 3 in Wu & Treiman, 2007). People with rural-hukou status can live in cities, as exemplified by the increasingly large numbers of migrant workers since the early 1980s. Similarly, people with urban-hukou status can live in rural areas, as exemplified by agricultural technicians and school teachers.5 Because previous studies have shown that both gender and ethnicity are important predictors of school enrollment (Bauer, Wang, Riley, & Zhao, 1992; Hannum, 2002, 2005; Hannum & Xie, 1994), they are included in the models as the control variables. Gender is coded as a dummy variable (boy = 1), so is ethnicity (Han Chinese = 1 and non-Han minority = 0). 3.3. Methods To model the probability of enrollment, omitting subscripts denoting the ith person of jth birth cohort in t

5 According to the author’s own tabulation on the 2000 census data, about 20% of rural hukou holders resided in cities and townships, constituting 33% of the population in cities and 54% of the population in townships. On the other hand, 12% of urban hukou holders actually resided in villages, constituting only 4.7% of the rural population.

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;
8

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx

period (census year), a general model is specified as ln p 1−p = α + β X,

where p is the probability of being enrolled in school of certain level/age range, X is the vector of independent variables measuring family backgrounds (more control variables are to be added when necessary), and β is the vector of estimated coefficients. Note that in this specification β is estimated separately for each cohort in each of the two periods. To examine the temporal trend, the model can be expressed equivalently as ln p 1−p =α+βX+δS

where S = tX, t is s scalar dummy variable (2000 = 1), and δ is a vector of parameters representing the interaction effects between family background variables and time (t) (Wooldridge, 2003, Chapter 13). Because the sample was clustered within city districts/counties, an adjustment of standard errors is needed in regression analyses. All the models reported were estimated using Stata 9.2, with robust standard errors corrected for clustering on sampling units (districts/counties).

4. Descriptive statistics Fig. 2 plots age-specific enrollment rates in China from 6 to 18 years old in 1990 and 2000, respectively. Except for age 6–7, the enrollment rate at age 12 or below was quite high in 1990 and was almost at saturation level in 2000, which is consistent with the government statistics from the Ministry of Education presented in Table 1 (although the latter may be over-reported). This evidence suggests that enrollment in primary school has been near saturation in China since 1990. From age 13 to 15 (typically junior high school years), the rate dropped from 81.7% to 54.4% in 1990 and from 94.4% to 75.4%, suggesting the successful expansion of compulsory education at lower secondary level. From age 16 to 18 (typically upper secondary school years), the rate dropped further from 38.9% to 16.9% in 1990, and 58.9% to 24.1% in 2000. By comparing the statistics across the 2 years, we can observe a significant increase in enrollment rates within the decade, thanks to the successful implementation of the 9-year compulsory education law in the 1990s.

Table 3 presents descriptive statistics for those aged between 6 and 18.6 The rate of full-time school attendance increased from 64.9% in 1990 to 82.5% in 2000. Gender and age structure, ethnic composition, and residence remain largely the same between the two samples. However, while the father’s occupational status index has changed little, both father’s and mother’s education have improved significantly. For example, fathers who have junior school education or above increased from 40% in 1990 to 65% in 2000; mothers who have junior high school education increased from 20% to 44% within the decade. At the bottom of Table 3, the rate of transition to junior high school given the completion of elementary school for those aged between 13 and 15 and the rate of transition to senior high school given the completion of junior high school for those aged between 16 and 18 in both 1990 and 2000 have been calculated. The rate of transition to junior high school in both years match government statistics quite closely, as shown in Table 2 (75.9% vs. 74.6% in 1990 and 93.2% vs. 94.9% in 2000); whereas the rate of school advancement beyond the compulsory level is much lower than that reported in government statistics (31% vs. 41% in 1990 and 41% vs. 51% in 2000). The discrepancy confirms previous claims that the official net enrollment rate may have overestimated the number of students actually attending classes, because it only recorded enrollment status at the beginning of the school year (Tsui, 1997). As of 2000 in China, there still exists a huge variation in school attendance rate. Among 2,870 counties and urban districts, the enrollment rate ranges from 69.5% (Nimu county, Tibet) to 100% among those children aged between 6 and 15 years old (the national average is 94.6%), as plotted in Fig. 3a. Beyond compulsory education, the spatial differentials in school attendance among children aged between 16 and 18 years old are even more prominent (Fig. 3b). On average, only 45% of Chinese aged between 16 and 18 years were still staying in school in 2000.7 In the following analysis I first examine the effect of family background on the enrollment status for children aged 6–18, who mostly live with their parents. I then analyze school transition for those aged between 13 and 15 (from primary to junior high school) and those between 16 and 18, separated by urban and rural resi-

Children of age 6–7 who have not started school were excluded in calculating the rate of full-time school attendance. 7 City/county-level enrollment rates are computed by the author based on 0.9% of the 2000 census micro-data.

6

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx 9

Fig. 3. (a) County-level school enrollment rate in China (age 6–15) 2000. Notes: prefecture boundary shown in the map. Each dot represents a county or a city. (b) County-level school enrollment rate in China (age 16–18), 2000. Notes: prefecture boundary shown in the map. Each dot represents a county or a city.

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;
10

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx

Table 3 Descriptive statistics of school-age children (6–18) in China, 1990 and 2000. Variables Full-time enrolled in school (yes = 1) Sex (female = 1) Age Region East Middle West Ethnicity (Han = 1) Hukou (rural = 1) Residential (county = 1) Father’s ISEI Father’s schooling Less than Elementary school Elementary school Junior high school Senior high school or above Mother’s schooling Less than elementary school Elementary school Junior high school Senior high school or above Number of cases Advance to junior high school given completion of elementary school (aged 13–15) Advance to senior high school given completion of junior high school (aged 16–18) Sources: 0.1% micro-data of 1990 and 2000 censuses. 1990 0.649 0.485 12.20 (S.D. = 3.80) 0.335 0.438 0.227 0.905 0.842 0.689 24.18 (S.D. = 15.32) 0.147 0.451 0.288 0.115 0.380 0.414 0.150 0.057 290,860 0.759 (N = 37,406) 2000 0.825 0.472 11.92 (S.D. = 3.45) 0.364 0.403 0.232 0.897 0.818 0.709 24.12 (S.D. = 14.01) 0.041 0.302 0.472 0.184 0.128 0.420 0.343 0.109 289,769 0.932 (N = 58,611)

0.305 (N = 27,686)

0.410 (N = 33,977)

dence. Special attention is given to the changing role of hukou status and father’s socioeconomic status in affecting the status of enrollment and the likelihood of school transition within the decade. Finally, I specifically investigate school transition in rural areas in the local context of economic development and educational financing in 2000. 5. Empirical findings from a multivariate analysis Table 4 presents the results from binary logistic regression predicting the likelihood of enrollment in school for all children aged 6–18 in 1990 and 2000. Model 1 includes a dummy variable for year 2000 to capture the increase in enrollment rate, gender, region, hukou status and residential place. The variables of father’s occupation and education, and mother’s education are added to Model 2. Finally, Model 3 includes an interaction between the hukou status and father’s occupational

status with the dummy variable for year 2000 to test whether these effects have changed over time. Not surprisingly, year, sex, ethnicity, hukou status, residential place and region, are all significant predictors of enrollment status, and so are family background variables. Children whose father holds a high-status occupation and whose parents have higher education are more likely to be enrolled in school. The interaction terms in Model 3 indicate that, despite the significant improvement in school enrollment in the past decade, the effect of father’s socioeconomic status on the likelihood of school enrollment was stronger in 2000 than in 1990, and the change is statistically significant (p < .05). Moreover, children of rural-hukou status were even more disadvantaged in 2000 than in 1990, as indicated by the negative coefficient of the interaction term. Other things being equal, the odds of being enrolled in school for rural-hukou holders are 86% (=e−0.155 ) of those for urban-hukou holders in 1990; such figures decreased to 73% (=e−0.155–0.160 ) in 2000.

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx 11

Table 4 Logistic model predicting full-time school enrollment for those aged 6–18, 1990 and 2000. Variables Year of 2000 Female Hukou (rural = 1) Regiona Middle West Ethnicity (Han = 1) Residence (rural area = 1) Father’s schoolingb Elementary school Junior high school Senior high school or above Mother’s schoolingc Elementary school Junior high school Senior high school or above Father’s socioeconomic index (ISEI) Father’s ISEI* year of 2000 Rural hukou* year 2000 Constant Pseudo-R2 Observations 1.216** (0.017) 0.057 579546 −0.053 (0.028) 0.078 477605 Model 1 0.945** (0.007) −0.256** (0.007) −0.678** (0.014) −0.159** (0.008) −0.344** (0.009) 0.395** (0.011) −0.138** (0.010) Model 2 0.626** (0.008) −0.276** (0.008) −0.212** (0.018) −0.122** (0.009) −0.154** (0.010) 0.272** (0.012) −0.065** (0.011) 0.413** (0.014) 0.724** (0.015) 0.767** (0.020) 0.370** (0.010) 0.662** (0.014) 0.663** (0.023) 0.007** (0.000) Model 3 0.735** (0.042) −0.276** (0.008) −0.155** (0.024) −0.120** (0.009) −0.151** (0.010) 0.270** (0.012) −0.063** (0.011) 0.415** (0.014) 0.729** (0.016) 0.765** (0.020) 0.374** (0.010) 0.667** (0.014) 0.651** (0.023) 0.007** (0.000) 0.002* (0.001) −0.160** (0.033) −0.101** (0.033) 0.078 477605

Notes: a East region as the reference. b Less than elementary school as the reference Robust standard errors in parentheses. c Less than elementary school as the reference Robust standard errors in parentheses. * Significant at 5%. ** Significant at 1%.

Because descriptive statistics in Table 2 and Fig. 1 suggest that enrollment in primary school almost reached saturation in the 1990s as a result of successful implementation of the compulsory education in China, to specifically investigate the social differentials in school attendance, school transition models for those aged between 13 and 15 have been formulated (see Table 5)

with the same independent variables and modeling strategies as in Table 4, but separately according to urban and rural residence. The results show that the patterns are quite similar to those previously observed in Table 4, except for the changing effects of father’s socioeconomic status and hukou status. In both rural and urban

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;
12

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx

Table 5 Logit models predicting transition to junior high school given the completion of elementary school (for those aged 13–15), 1990 and 2000. Urban Model 1 Year of 2000 Female Hukou (rural = 1) Regiona Middle West Ethnicity (Han = 1) Father’s schoolingb Elementary school Junior high school Senior high school or above Mother’s schoolingc Elementary school Junior high school Senior high school or above Father’s ISEI Father’s ISEI* year 2000 Rural hukou* year 2000 Constant Pseudo-R2 Observations 3.404** (0.130) 0.185 32856 1.531** (0.205) 0.229 25404 1.711** (0.061) −0.435** (0.056) −1.987** (0.076) −0.361** (0.066) −0.656** (0.073) 0.426** (0.102) Model 2 1.270** (0.075) −0.491** (0.065) −1.178** (0.107) −0.279** (0.075) −0.508** (0.085) 0.297** (0.112) 0.227 (0.118) 0.739** (0.129) 0.967** (0.177) 0.343** (0.086) 0.964** (0.119) 0.859** (0.192) 0.020** (0.003) Model 3 0.692* (0.297) −0.492** (0.065) −1.310** (0.128) −0.279** (0.075) −0.512** (0.085) 0.298** (0.113) 0.228 (0.119) 0.741** (0.130) 0.987** (0.180) 0.337** (0.086) 0.953** (0.120) 0.896** (0.195) 0.019** (0.003) 0.005 (0.006) 0.543** (0.209) 1.679** (0.223) 0.230 25404 Rural Model 4 1.654** (0.025) −0.667** (0.024) −2.141** (0.135) −0.451** (0.030) −0.826** (0.032) 0.650** (0.039) Model 5 1.331** (0.030) −0.762** (0.028) −1.242** (0.178) −0.396** (0.034) −0.645** (0.037) 0.577** (0.044) 0.304** (0.048) 0.805** (0.053) 1.137** (0.080) 0.297** (0.033) 1.052** (0.053) 1.398** (0.138) 0.032** (0.002) Model 6 0.578 (0.366) −0.762** (0.028) −1.537** (0.246) −0.396** (0.034) −0.644** (0.037) 0.578** (0.044) 0.304** (0.048) 0.806** (0.053) 1.137** (0.080) 0.297** (0.033) 1.051** (0.053) 1.396** (0.138) 0.031** (0.002) 0.004 (0.004) 0.690* (0.348) 1.148** (0.264) 0.192 52637

3.013** (0.140) 0.131 63123

0.838** (0.197) 0.192 52637

Notes: a East region as the reference. b Less than elementary school as the reference; Robust standard errors in parentheses. c Less than elementary school as the reference; Robust standard errors in parentheses. * Significant at 5%. ** Significant at 1%.

areas, father’s socioeconomic status has a significant impact on the likelihood of transition to junior high school given one has completed primary school education, but there seems to be no significant change between 1990 and 2000. Children of rural-hukou status, however, have indeed gained relatively more advantages, given their low starting point. This pat-

tern implies that the expansion of education, and in particular of compulsory education, has benefited rural children and overcome their disadvantages compared to urban children. Educational expansion has reduced urban–rural inequality at lower levels, but not the inequality associated with family socioeconomic backgrounds.

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx 13

Table 6 Logit models predicting transition to senior high school given the completion of junior high school (for those aged 16–18), 1990 and 2000. Variables Urban Model 1 Year of 2000 Female Hukou (rural = 1) Regiona Middle West Ethnicity (Han = 1) Father’s schoolingb Elementary school Junior high school Senior high or above Mother’s schoolingc Elementary school Junior high school Senior high or above Father’s ISEI Father’s ISEI* year 2000 Rural hukou* year 2000 Constant Pseudo-R2 Observations 0.777** (0.083) 0.184 26121 −1.004** (0.172) 0.248 19142 1.010** (0.034) −0.014 (0.033) −2.077** (0.034) −0.433** (0.037) −0.138** (0.046) −0.020 (0.076) Model 2 0.817** (0.045) −0.047 (0.041) −1.411** (0.048) −0.414** (0.046) −0.048 (0.059) −0.047 (0.098) 0.343* (0.138) 0.494** (0.140) 0.960** (0.145) 0.066 (0.077) 0.323** (0.081) 1.045** (0.093) 0.021** (0.001) Model 3 0.805** (0.136) −0.045 (0.041) −1.183** (0.072) −0.402** (0.046) −0.042 (0.059) −0.062 (0.099) 0.356** (0.136) 0.512** (0.138) 0.986** (0.144) 0.106 (0.076) 0.356** (0.081) 1.052** (0.093) 0.019** (0.002) 0.005 (0.003) −0.420** (0.094) −1.053** (0.180) 0.250 19142 Rural Model 4 0.478** (0.029) −0.322** (0.028) −1.614** (0.056) −0.151** (0.031) −0.011 (0.039) 0.212** (0.055) Model 5 0.379** (0.037) −0.378** (0.032) −1.075** (0.075) −0.155** (0.035) −0.084 (0.044) 0.184** (0.063) 0.250** (0.083) 0.405** (0.085) 0.802** (0.093) 0.056 (0.046) 0.245** (0.054) 0.883** (0.081) 0.020** (0.001) Model 6 0.863** (0.179) −0.375** (0.032) −0.672** (0.132) −0.147** (0.035) −0.091* (0.044) 0.171** (0.063) 0.266** (0.083) 0.435** (0.085) 0.829** (0.093) 0.072 (0.045) 0.257** (0.054) 0.880** (0.081) 0.017** (0.002) 0.008** (0.002) −0.718** (0.161) −1.821** (0.172) 0.069 29575

−0.084 (0.079) 0.036 35505

−1.511** (0.127) 0.068 29575

Notes: a East region as the reference. b Less than elementary school as the reference; Robust standard errors in parentheses. c Less than elementary school as the reference; Robust standard errors in parentheses. * Significant at 5%. ** Significant at 1%.

Table 6 presents the school transition models for those aged between 16 and 18, who have completed junior high school education. The patterns are quite different from the early transition. For children of urban-hukou status, father’s socioeconomic status still plays an important role, but the effect remained the same from 1990 to 2000. For children of rural-hukou status, however, mak-

ing the transition to senior high school becomes even more difficult. In other words, in Chinese cities, people with rural-hukou status (namely, rural migrants denied urban citizens’ rights) face significant disadvantages in entering senior high school after completing compulsory junior high school education, compared to those with urban permanent hukou status (Liang & Chen, 2007).

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;
14

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx

Table 7 Logit model predicting school transitions for those living in rural areas (counties), 1990 and 2000. Variables Transition to junior high Model 1 Year of 2000 Female Hukou (rural = 1) Ethnicity (Han Chinese = 1) County per capital GDP Father’s schoolinga Elementary school Junior high school Senior high or above Mother’s schoolingb Elementary school Junior high school Senior high or above Father’s ISEI Father’s ISEI* year 2000 Rural hukou* year 2000 Constant Pseudo-R2 Observations −2.736** (0.225) 0.137 51561 −3.886** (0.280) 0.192 42888 0.761** (0.037) −0.673** (0.026) −2.101** (0.143) 0.530** (0.042) 0.750** (0.025) Model 2 0.589** (0.042) −0.759** (0.030) −1.254** (0.193) 0.467** (0.046) 0.637** (0.028) 0.299** (0.050) 0.808** (0.056) 1.190** (0.086) 0.241** (0.035) 0.941** (0.058) 1.389** (0.158) 0.029** (0.002) Model 3 −0.087 (0.404) −0.758** (0.030) −1.483** (0.250) 0.468** (0.046) 0.636** (0.028) 0.299** (0.050) 0.809** (0.056) 1.190** (0.086) 0.241** (0.035) 0.940** (0.058) 1.388** (0.158) 0.028** (0.002) 0.003 (0.005) 0.623 (0.384) −3.643** (0.325) 0.192 42888 Transition to senior high Model 4 0.266** (0.043) −0.330** (0.031) −1.577** (0.063) 0.109 (0.057) 0.144** (0.028) Model 5 0.173** (0.051) −0.383** (0.036) −1.049** (0.085) 0.053 (0.066) 0.128** (0.032) 0.271** (0.090) 0.433** (0.093) 0.836** (0.103) 0.048 (0.051) 0.252** (0.060) 0.862** (0.093) 0.018** (0.001) Model 6 0.650** (0.198) −0.381** (0.036) −0.707** (0.137) 0.043 (0.066) 0.128** (0.032) 0.284** (0.090) 0.457** (0.093) 0.855** (0.103) 0.062 (0.051) 0.265** (0.060) 0.859** (0.094) 0.016** (0.002) 0.007** (0.003) −0.693** (0.177) −2.664** (0.286) 0.065 24505

−1.157** (0.214) 0.033 29575

−2.394** (0.263) 0.064 24505

Notes: a Less than elementary school as the reference; Robust standard errors in parentheses. b Less than elementary school as the reference; Robust standard errors in parentheses. * Significant at 5%. ** Significant at 1%.

Compared to 1990, the situation in 2000 deteriorated, given the surging wave of migration from rural to urban areas (Liang & Ma, 2004). In rural areas, the scenario was quite different from 1990 to 2000. Father’s socioeconomic status was still a significant predictor of the likelihood of transition to senior high school and the effect was even stronger in 2000 than in 1990. For rural-hukou holders in rural areas, their situation was worse in 2000 than in 1990. Therefore, thanks to the successful implementation of the 9-year compulsory education in China in the 1990s, fam-

ily background and registration status still play an even greater role in determining whether school-age children receive further education beyond the compulsory level, despite the fact that the educational expansion benefited children of rural hukou in lower secondary education. Does this reflect the uneven regional economic development in rural China? In Table 7, local economic development level, measured by (logged) GDP per capita of the county in 2 years, was controlled for. The results show that local economic development does play an important role in determining school atten-

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx 15

dance rate: children in more developed counties/areas have more educational opportunities (also see Fig. 3a and b). The effect of father’s socioeconomic status and family’s hukou status on school transition continue to affect school transition rates, and these effects, while unchanged on the transition to junior high school, were stronger in 2000 than in 1990 in favor of those who are at advantaged status, namely, those who have fathers of high occupational status and who hold urban-hukou status. Educational inequality at senior high school level increased within that decade. Local per capita GDP may not capture the exact amount of resources spent on education. Two countylevel indicators – educational expenditure per capita and

the percentage of educational surcharge on per capita income – are available only for the year 2000. They are included in the models of Table 8, which predict school transitions in rural China in 2000. As shown in the models, both are significant predictors of school transition rates, but again, the effects of family backgrounds and hukou status remain substantial. What might explain the above-mentioned difficulty faced by children of rural hukou are the exclusionary barriers they encounter for not holding local urban hukou (Liang & Chen, 2007). Societal norms and values might also have a role in making them less interested in continuing education beyond the compulsory levels. The latter factor could also explain why rural-hukou children have

Table 8 Logit model predicting school transitions for those living in rural areas, controlling for county educational expenditure, 2000. Variables Transition to junior high school Model 1 Female Hukou (rural = 1) Ethnicity (Han = 1) Father’s schoolinga Elementary school Junior high school Senior high or above Mother’s schoolingb Elementary school Junior high school Senior high or above Father’s ISEI Education spending per capita (logged) % Surcharge in per capita income (logged) Constant Pseudo-R2 Observations 0.770* (0.317) 0.108 24627 −0.597** (0.051) −0.762** (0.278) 0.837** (0.074) 0.177 (0.099) 0.669** (0.105) 1.088** (0.143) 0.503** (0.066) 1.356** (0.095) 1.894** (0.225) 0.040** (0.005) Model 2 −0.593** (0.053) −0.779** (0.302) 0.797** (0.078) 0.145 (0.109) 0.648** (0.116) 1.110** (0.157) 0.467** (0.070) 1.264** (0.100) 1.826** (0.235) 0.035** (0.005) 0.502** (0.126) 0.073* (0.034) −1.431* (0.704) 0.108 22837 Transition to senior high school Model 3 −0.326** (0.043) −1.428** (0.115) 0.211* (0.082) 0.122 (0.157) 0.272 (0.158) 0.677** (0.162) −0.004 (0.077) 0.258** (0.085) 0.767** (0.112) 0.026** (0.002) Model 4 −0.324** (0.044) −1.387** (0.118) 0.176* (0.087) 0.036 (0.168) 0.249 (0.168) 0.682** (0.174) −0.036 (0.080) 0.249** (0.088) 0.707** (0.117) 0.023** (0.002) 0.584** (0.096) 0.065* (0.032) −3.637** (0.544) 0.069 12569

−0.887** (0.201) 0.069 13457

Notes: a Less than elementary school as the reference; Robust standard errors in parentheses. b Less than elementary school as the reference; Robust standard errors in parentheses. * Significant at 5%. ** Significant at 1%

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;
16

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx

lower attendance rates at senior high school level compared to urban-hukou children in rural areas. As yet, no data are available to directly address this issue. 6. Summary and conclusions To summarize, this study examined the trend in educational stratification during China’s economic reforms in the 1990s. Based on the samples of the population census data for 1990 and 2000, school-age children were matched to their parents’ background information and the effects of family background on their school enrollment and transitions were investigated. Results show that, despite the substantial expansion of educational opportunities from 1990 to 2000, family backgrounds have continued to play an important role in determining school enrollment status and school transitions. During the decade, children of rural-hukou status became more disadvantaged compared to their urban counterparts and the effect of father’s socioeconomic status on school enrollment was enhanced. While children of rural-hukou status gained more opportunities at junior high school level as a result of saturation in the 9-year compulsory education at the national level, the rural–urban gap in the likelihood of transition to senior high school level was enlarged and the effect of father’s socioeconomic status on the transition rate increased, even after controlling for regional variations in economic development. Hence, educational expansion in China, accompanied by the rapid marketization in the 1990s, did not bring more equal access to educational opportunities among different social strata. Instead, uneven distribution of educational opportunities seems to have exacerbated in the context of market reforms in the educational sphere and rising inequality in the distribution of economic resources. The change in educational inequality to a large extent mimicked the change in the overall structure of inequality in reform-era China in the 1990s. China’s case is consistent with the thesis of “maximally maintained inequality” (Raftery & Hout, 1993), which argues that inequality in educational opportunity is maximally maintained, namely in modern societies, where the effect of social origin at all levels of education does not change. Only when the enrollment of advantaged groups is already high at a given level, could further expansion be feasible by increasing the opportunity of disadvantaged groups to make the transitions. The implications of the findings in this paper, however, may go further beyond the thesis of maximally maintained inequality. While the thesis predicts that educational expansion does not lead to better chances for disadvantaged groups to transition to a higher school

level, and that it will not change the association between family background and the given level of school transitions, the analysis of this paper has demonstrated that the effect of family background has indeed increased (rather than remained constant and decreased conditionally), and educational opportunities of the disadvantaged groups were less in 2000 than a decade prior. What are the implications of these findings for the change in social stratification order and the evolution of social structure in China in the future? While the available data do not allow us to examine the trend in tertiary school attendance rates of children from different social backgrounds, one can reasonably speculate that the expansion of higher education in the late 1990s will largely benefit urban children and children from betteroff families, further increasing educational inequality at higher levels (Min, 2007; Yang, 2006).8 The rising educational inequality among students of different socioeconomic backgrounds in the 1990s could lead to increasing earnings inequality after they complete education and enter the labor markets. In the long run, intergenerational transmission is enhanced in the course of market transition (as observed in post-Soviet Russia by Gerber & Hout, 2004); the role of education as an important channel for socioeconomic mobility is weakened. Future research should be devoted to assessing the far-reaching social consequences of the rising educational inequality in China in recent years. The reversed equalization of educational opportunities reported from China has echoed previous findings from post-Soviet Russia (Gerber, 2000) and Hong Kong (Wu, 2007). The political chaos and economic crisis had harmed the Russian educational system and increased origin-based inequality in access to secondary schools for certain cohorts, who completed education during the later Soviet and post-Soviet years when enrollment had contracted. In Hong Kong, despite a decline in the 1980s, family background affected more importantly the progression to higher levels of education (particularly to university) in 2001. The economic prosperity accompanying dramatic institutional changes seems to have brought limited equality in access to upper secondary education in China, which is still in short supply for the majority of the population. All three societies have experienced a rapid increase in income inequality over the past decades. In Russia, the Gini coefficients among all employed workers increased from 0.261 in

8 There were 2.04 million full-time students enrolled in colleges in 1991; the enrollment increased to 5.56 million in 2000 and 12 million in 2003. Also see the transition rate to tertiary school in Table 2.

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx 17 Hannum, E. (2005). Market transition, educational disparities, and family strategies in rural China: New evidence on gender stratification and development. Demography, 42, 275–299. Ishida, H., Muller, W., & Ridge, J. M. (1995). Class origin, class destination, and education: A cross-national study of ten industrial nations. American Journal of Sociology, 101, 145–193. Kahn, J., & Yardley, J. (2004). Amid China’s boom, no help hand for young Qingming. The New York Times, (August) Liang, Z., & Chen, Y. (2007). Educational consequences of migration for children in China. Social Science Research, 36(2007), 28–47. Liang, Z., & Ma, J. Z. (2004). China’s floating population: New evidence from the 2000 census. Population and Development Review, 30(3), 467–488. Liu, J. (2004). Education in transitional Chinese society. Liaoning Renmin Press. (in Chinese). Mare, R. D. (1980). Social backgrounds and educational continuation decisions. Journal of American Statistical Association, 75, 295–305. Mare, R. D. (1995). Changes in educational attainment and school enrollment. In R. Farley (Ed.), State of the union: American in the 1990s (pp. 155–213). Sage Russell, Economic Trends. Mare, R. D., & Chen, M. D. (1986). Further evidence on number of siblings and educational stratification. American Sociological Review, 51, 403–412. Min, W. (Ed.). (2007). China education and human resource development report 2005–2006. Beijing: Peking University Press (in Chinese). Ministry of Education and National Bureau of Statistics. (2002). China educational finance statistical yearbook 2001. Beijing: China Statistics Press. Nee, V. (1989). A theory of market transition: From redistribution to markets in state socialism. American Sociological Review, 54, 663–681. Parish, W. L., & Michelson, E. (1996). Politics and markets: Dual transformations. American Journal of Sociology, 101, 1042–1059. Raftery, A. E., & Hout, M. (1993). Maximally maintained inequality: Expansion, reform, and opportunity in Irish education: 1921–1975. Sociology of Education, 66, 41–62. Róna-Tas, A. (1994). The first shall be last? Entrepreneurship and communist cadre in the transition from socialism. American Journal of Sociology, 100, 40–69. Shavit, Y., & Blossfeld, H. (1993). Persistent inequality. Boulder, CO: Westview Press. Simkus, A., & Andorka, R. (1982). Inequalities in educational attainment in Hungary, 1923–1973. American Sociological Review, 47, 740–751. Szelényi, I., & Kostello, E. (1996). The market transition debate: Toward a synthesis. American Journal of Sociology, 101, 1082–1096. Tsang, M. C., & Ding, Y. (2005). Resource utilization and disparities in compulsory education in China. The China Review, 5, 1–31. Treiman, D. J. (1970). Industrialization and social stratification. In Edward Laumann (Ed.), Social stratification: Research and theory for the 1970s (pp. 207–234). Indianapolis: Bobbs-Merrill. Tsui, K. (1997). Economic reform and attainment in basic education in China. The China Quarterly, 149, 104–127. Walder, A. G., Li, B., & Treiman, D. J. (2000). Politics and life chances in state socialist regime: Dual career paths into the urban Chinese elite, 1949–1996. American Sociological Review, 65, 191–209. Wang, J. C. F. (2002). Contemporary Chinese politics: An introduction. Prentice-Hall.

1986 to 0.296 in 1991, 0.483 in 1996 and further to 0.521 in 2001. In Hong Kong, the Gini coefficient for household income rose from 0.453 in 1986 to 0.476 in 1991, 0.518 in 1996, and 0.525 in 2001 (World Income Inequality Database, 2004, for China statistics from the same source, see Table 2), and, finally, 0.533 in 2006 (Census and Statistics Department, 2007). These observations suggest that the distribution of educational opportunities may reflect the mechanism of resource distribution rather than a direct response to expanding opportunities. Thus, more comparative studies of educational stratification in rapidly changing societies are called for in order to establish such a direct linkage. References
Bauer, J., Wang, F., Riley, N., & Zhao, X. (1992). Gender educational inequality in urban China. Modern China, 18, 333–370. Bian, Y., & Logan, J. R. (1996). Market transition and persistence of power: The changing stratification in urban China. American Sociological Review, 61, 739–758. Boudon, R. (1974). Education, opportunities and social inequality. Wiley: New York. Cai, F., & Wan, G. (2006). Studies on income inequality and poverty in transition China: What do we know and what we should know? In F. Cai, & G. Wan (Eds.), Income Inequality and Poverty in Transition China (pp. 1–22). Beijing: Social Science Academic Press. Cao, Y., & Nee, V. (2000). Comment: Controversies and evidence in the market transition debate. American Journal of Sociology, 105, 1175–1189. Census Statistics Department, Hong Kong SAR Government. (2007). Thematic report: Household income distribution in Hong Kong. Hong Kong SAR Government Printing Department. Deng, Z., & Treiman, D. J. (1997). The impact of the cultural revolution on trends in educational attainment in the People’s Republic of China. American Journal of Sociology, 103, 391–428. Featherman, D. L., & Hauser, R. M. (1978). Opportunity and change. New York: Academic Press. Ganzeboom, H. B. G., de Graaf, P., & Treiman, D. J. (1992). An international scale of occupational status. Social Science Research, 21, 1–56. Gerber, T. P. (2000). Educational stratification in contemporary Russia: Stability and change in the face of economic and institutional crisis. Sociology of Education, 73(3), 219–246. Gerber, T. P., & Hout, M. (1995). Educational stratification in Russia during the Soviet Period. American Journal of Sociology, 101, 611–660. Gerber, T. P., & Hout, M. (1998). More shock than therapy: Employment and income in Russia, 1991–1995. American Journal of Sociology, 104, 1–50. Gerber, T. P., & Hout, M. (2004). Tightening up: Declining class mobility during Russia’s market transition. American Sociological Review, 69, 677–703. Hannum, E., & Xie, Y. (1994). Trends in educational gender inequality in China: 1949–1985. Research in Social Stratification and Mobility, 13, 73–98. Hannum, E. (2002). Educational stratification by ethnicity in China: enrollment and attainment in the early reform years. Demography, 39, 95–117.

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003

+Model RSSM-92;
18

No. of Pages 18

ARTICLE IN PRESS
X. Wu / Research in Social Stratification and Mobility xxx (2010) xxx–xxx occupational mobility. American Journal of Sociology, 103(2), 415–445. Wu, X., & Xie, Y. (2003). Does the market pay off? Earnings returns to education in urban China. American Sociological Review, 68, 425–442. Xie, Y., & Hannum, E. (1996). Regional variation in earnings inequality in reform-era urban China. American Journal of Sociology, 102, 950–992. Yang, D. (2006). Access to higher education: Widening social class disparities. Tsinghua Journal of Education, 27(1), 19–26. Zhou, X. (2000). Economic transformation and income inequality in urban China. American Journal of Sociology, 105, 1135–1174. Zhou, X., Moen, P., & Tuma, N. B. (1998). Educational stratification in urban china: 1949–1994. Sociology of Education, 71, 199–222. Zhou, X., Tuma, N. B., & Moen, Phyllis. (1997). Institutional change and job shift patterns in urban China, 1949 to 1994. American Sociological Review, 62, 339–365.

Wong, R. S. K. (1998). Multidimensional influences of family environment in education: The case of socialist Czechoslovakia. Sociology of Education, 71, 1–22. Wooldridge, J. (2003). Introductory econometrics. Thomson Southwestern. World Income Inequality Database (2004). http://www.wider.unu. edu/wiid/wiid.htm. Wu, X. (2007). Family resources and educational stratification: The case of Hong Kong, 1981–2001. Social Transformations in Chinese Societies, Volume 3 Wu, X. & Ma, J. Z. (2004). Regional inequalities in education of Chinese population: Evidence from China’s 2000 population census. Paper presented in the International Seminar on China’s 2000 Population and Housing Census, April 27–29, 2004, Beijing, China. Wu, X., & Treiman, D. J. (2004). The household registration system and social stratification in China 1955–1996. Demography, 41(2), 363–382. Wu, X., & Treiman, D. J. (2007). Inequality and equality under Chinese socialism: The Hukou system and intergenerational

Please cite this article in press as: Wu, X. Economic transition, school expansion and educational inequality in China, 1990–2000. Research in Social Stratification and Mobility (2010), doi:10.1016/j.rssm.2009.12.003