Social Science Research 53 (2015) 177–190

Contents lists available at ScienceDirect

Social Science Research journal homepage: www.elsevier.com/locate/ssresearch

Population migration and children’s school enrollments in China, 1990–2005 q Xiaogang Wu a,⇑, Zhuoni Zhang b a b

Division of Social Science, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong Department of Applied Social Sciences, City University of Hong Kong, Kowloon Tong, Hong Kong

a r t i c l e

i n f o

Article history: Received 5 August 2013 Revised 28 February 2015 Accepted 20 May 2015 Available online 30 May 2015 Keywords: China Children Migration School enrolment

a b s t r a c t This paper examines the impact of migration on children’s school enrollment by analyzing the micro-data from Chinese population censuses in 1990 and 2000 and mini-census in 2005. We match school-age children (7–14 years old) with their parents, and examine how migration status and parents’ absence affect children’s school enrollment in urban China. We also compare rural–urban migrant children with their peers in both origin counties and destination districts. Results show that migrant children are less likely to be enrolled in school than urban local children and that children of rural registration status are particularly disadvantaged in school enrollment over the whole examined period in urban China. Rural–urban migrant children fare significantly worse than non-migrant children in both origins and destinations and noticeably they are even less likely than left-behind children to be enrolled in school. The likelihood of being enrolled in school increases for rural–urban migrant children as they spend more time in destinations. Ó 2015 Elsevier Inc. All rights reserved.

1. Introduction Market reforms and ensuing economic growth have brought about a surge in internal migration in China. In the pre-reform period, by virtue of the household registration system (hukou), the Chinese government set up an ‘‘invisible wall’’ to demarcate different places of residence, and especially to set the urban and rural sectors apart, with the aim of controlling population migration (Chan, 1994; Wu and Treiman, 2004). Economic reform in the past three decades has relaxed such administrative control. Consequently, geographic mobility has risen and changing jobs has become much easier than before (Hao, 2012; Ou and Kondo, 2013). The size of the ‘‘floating population’’, which consists of migrants who have resided at the place of destination for at least six months without local household registration status, reached 144 million in 2000 (Liang and Ma, 2004) and 147 million in 2006 (National Bureau of Statistics in China, 2006). These migrants, predominantly farmers from inland provinces, tend to not only move further and stay longer for better economic opportunities, but also to bring their spouses and children once they have settled down in cities. The younger ones

q This paper was presented at the annual meeting of the Population Association of America in Dallas on April 15–17, 2010, and at the summer meeting of the International Sociological Research Association—Research Committee on Social Stratification and Mobility (RC28), University of Queensland, Australia, on July 16–19, 2013. We thank the Research Grants Council of Hong Kong for financial support (General Research Fund Projects 644510 and 646411) and for a Prestigious Fellowship awarded to the first author in 2013(HKUST602-HSS-12). We also thank the China National Bureau of Statistics for providing data access, Dr. Yuxiao Wu for assistance in initial data preparation, and Zai Liang and other participants for comments and suggestions. ⇑ Corresponding author. E-mail addresses: [email protected] (X. Wu), [email protected] (Z. Zhang).

http://dx.doi.org/10.1016/j.ssresearch.2015.05.007 0049-089X/Ó 2015 Elsevier Inc. All rights reserved.

178

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

may even choose to start a family in the destination cities. In the 1990s, the city-born children of early migrants were beginning to reach school age. The new tidal wave of migration is also bringing to cities more school-age children migrating with their parents. Without local permanent registration status, migrant parents have to fork out extra to get their children into local schools, posing a huge burden on many migrant parents (Lu, 2007). Consequently, most migrants had to leave the care of their children to grandparents in their origin homes because they could not afford the extra costs to bring their children with them. These children, known as liushou ertong (left-behind children) in Chinese, have been found to exhibit various psychological and developmental problems (Liang et al., 2008; Lu and Zhou, 2013; Xiang, 2007). Hence, population migration has yielded significant impact on the well-being of children, be they migrant children or left-behind children. The provision of educational opportunities for children affected by migration has become an important issue concerning education policy makers and the public at large in China, which has fundamental implications for the country’s sustainable development in the future (Duan and Zhou, 2005; Liang et al., 2008; Yang and Fan, 2012). Most literature on internal migration in China tends to focus on the demographic patterns of migration and the socioeconomic consequences for migrants and community development (e.g., Liang, 2001; Liang and White, 1996; Ma, 2001; Yang and Guo, 1996; Zhang and Wu, 2013; Zhao, 2000). With few recent studies on the wellbeing of an increasing number of children affected by migration (e.g., Lee and Park, 2010; Liang and Chen, 2007; Lu and Zhou, 2013; Yang and Fan, 2012), systematic and rigorous empirical analyses of the issues remain relatively limited. First, most studies focus either on migrant children or on left-behind children, rather than placing them in a broad context for comparisons (with Liang and Chen, 2007 being an exception). Second, previous studies mainly employ either the census data from one province or the survey data on migration in selected regions; a systematic analysis of the situation at the national level has yet to emerge. Finally, the existing literature is limited to a snapshot analysis of how migration affects children’s school enrollment in the 1990s only, without updated information on the changing situations in the context of massive population migration since then. In this article, by analyzing the micro-data from population censuses in 1990 and 2000 and mini-census in 2005, we attempt to fill the gap and address the educational consequence of population migration to school-age children in China. Specifically, we examine how migration status and parents’ absence (family structure) affect the likelihood of children’s school enrollment, paying special attention to comparing rural–urban migrant children with their peers in both origin counties and destination districts. Our findings could shed lights on the roles of migration status, family structure, and rural hukou, and underlying mechanisms in determining children’s school enrollment in contemporary China. 2. Social exclusion, family structure, and educational consequences of migration to school-age children in China In a broader and comparative perspective, the issue of migration/immigration and children’s educational outcomes has been one of the major topics in sociological inquiries. In the United States, scholars have shown that residential and school moves would lead to an increase in high school drop-outs and poor academic performance (Long, 1975; Pribesh and Downey, 1999). Why migration has negative impact on children’s education? The prevailing explanation is that migration could results in the loss of social capital in school, the neighborhood, and community of origins, as social capital plays an important role in preventing school dropouts and enhancing learning outcomes (Coleman, 1988; Teachman et al., 1999). Compared to internal migrants in developed countries, the impact of transnational immigration on children’s educational outcome is less clear, largely contingent upon the human capital, social capital and other resources of their parents and communities (e.g., Xie and Greenman, 2011; Portes and Zhou, 1993; Zhou and Bankston, 1998). Empirical analyses show that, after taking into account of various background characteristics, the bulk of immigrant children often reach parity with or fare even better than their native-born peers (see reviews in Zhang, 2014). Social policies and civil rights legislations have largely afforded immigrant children opportunities for socioeconomic advancement in the host countries. For instance, immigrant children, even undocumented, are by law entitled to free public education from kindergarten to grade 12 in the United States (Lu and Zhou, 2013). Even if some immigrant children have poorer educational outcomes than their native-born peers, the disparity is mainly due to immigrant parents’ lower socioeconomic status, rather than their immigration status per se. Studies of the educational consequences of migration to children in the United States typically adopt non-migrant children as the comparison group, without further differentiating between those in origins and in destination, probably due to the fact that regional inequality is not a severe issue in the country. Immigrant children are seldom compared to their peers in origin countries, simply because the latter’s information is not available in most empirical analyses. While the extensive literature on internal migration or transnational immigration in western countries, especially in the United States, can inform our studies of similar issues in China, there are two distinct features that characterize the experience of Chinese children affected by migration. The first is the persisting role of the household registration system (hukou) in social exclusion of migrants and their children in destinations. The second is the changing family structure resulted from the massive population migration. Social exclusion refers to the process in which individuals or social groups are systematically blocked from or denied full access to various rights, opportunities and resources that are normally available to other members (Silver, 1994). It is a useful concept to help us understand the disadvantages encountered by the Chinese internal migrants in cities. The household registration system (hukou) has been served as a unique institution since the 1950s to divide Chinese rural and urban populations for the distribution of resources and life chances and the control of migration under state socialism (Wu and Treiman, 2004). After decades of economic reform, while its control over migration has been relaxed, the hukou continues to be employed by many local governments as the basis of providing employment opportunities and allocating subsidies and

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

179

welfare conferred to the permanent residents (Li et al., 2015; Roberts, 1997; Wang et al., 2002; Yang and Guo, 1996; Zhang and Wu, 2013). Due to the lack of local hukou, migrants must pay extra to go to the hospital, to rent an apartment, or to have their children attend local schools (Cai, 2002, p. 215). Such discrimination policies have created special hurdles to socioeconomic attainment not only for the adult rural migrants themselves (Wu, 2009; Zhang and Wu, 2013), but also for their offspring, particularly in regard to the latter’s access to educational opportunities in urban destinations (Liang and Chen, 2007). Therefore, whereas Chinese internal migrants are similar to transnational immigrants in western countries to a certain extent because of the institutional divide associated with the persisting hukou system, children of Chinese migrants may fare even worse compared to immigrant children, as the latter are protected from discrimination and afforded equal opportunities via various of legal, social and economic policies (Farley and Alba, 2002). Moreover, Chinese migrant children to cities may not necessarily fare better in educational opportunities, due to social exclusions in destination, than their non-migrant peers who stay in villages of origins. Sociologists have shown that family structure plays a significant role in determining children’s educational outcomes and subsequent status attainment. The prevalence of divorce, non-marital child bearing, and cohabitation in western societies made the family a less stable unit of two parents in which children were reared (Tach, 2015). In the United States, children who grow up in single-parent families are less likely to complete high school or to attend college than children who grow up with both parents, even after taking into account of the precarious economic position of the single-parent families (McLanahan, 1985). Some scholars suggested that child-rearing practice and parent–child relationship affected by the family disruption and single-parenthood could have yielded negative consequence in the long-term wellbeing of children (Astone and McLanahan, 1991). This account is consistent with the narrative on the role of social capital in the creation of human capital (Coleman, 1988), as family structure implies social capital a family could lend to support children’s education. In the literature on the negative consequence of migration to children’s education, scholars further pointed out that the association between family structure and children’s educational outcomes can be partly explained by migration, as children of non-intact families are more likely than children of two-parent families to change residence (Astone and McLanahan, 1994). In the case of China, social exclusions associated with the hukou and changing family structure resulted from migration have created a dilemma for migrant parents. On one hand, if they wish to bring their children with them, they face substantial obstacles in destination cities because of the lack of supporting services. It is difficult and also expensive to get their children into school and a great challenge for their children to be assimilated to urban life, as many rural migrant children live in villages-in-cities (cheng zhong cun) or on the outskirts, and are spatially segregated in the place where most residents are also migrants. The children’s guardianship and safety are an issue that concerns their parents, who are too busy with earning a living. On the other hand, if migrant parents leave their children in their rural origin homes, their children grow up in absence of one or two parents and their time inputs. The rural education reform in the 1990s has imposed an extra economic burden on families and has driven some parents to migrate in search for more economic resources to support their children’s education at home (Du et al., 2005; Wu, 2010; Zheng and Gu, 2012). While they may be more likely to get their children into school by sending home remittance, their absence has led to the loss of social capital and parental attention, negatively affecting the development of left-behind children at large (ACWF Research Team, 2013; Chang et al., 2011; Ren and Treiman, 2013). This dilemma has confronted a substantial amount of migrant families in China. A research report recently released by the All-China Women’s Federation, based on the 2010 population census data, estimates that the number of migrant children under age 18 has reached 35,810,000, and over 80% of them hold rural hukou status. Because of the social exclusion they face in cities, the number of left-behind children under age 18 who live with at least one parent absence has reached 61,025,500 (ACWF Research Team, 2013). Migration status may be intertwined with family structure (i.e., parents’ absence) in affecting children’s education in China. In this paper, we focus on the impact of migration on the school enrollment of children between 7 and 14, an age group required to receive 9-year compulsory education in China.1 According to the literature we have reviewed above, we expect that migration has negative impact on children’s school enrollment. In other words, migrant children are less likely to be enrolled in school than non-migrant children. Because of the hukou-based exclusion, rural–urban migrant children are disadvantaged in school enrollment compared to their non-migration peers not only in destination cities but also in rural origins, including left-behind children with at least one parent absent (presumably due to migration). Furthermore, migrant children living with at least one parent absent are even more disadvantaged because they are subject to both social exclusion and loss of social capital (parental guardianship). Finally, we expect that the migration timing affect the likelihood of school enrollment for rural–urban migrant children. While migration may have disruptive impact on children’s school enrollment, the impact is expected to fade away as migrants spend more time and adjust themselves in the destinations. 3. Data, analytical subsamples, and variables 3.1. Data The data sets analyzed here are samples of the micro-data from population censuses in both 1990 and 2000 and of the mini-census in 2005 in China. We first extracted those children aged between 7 and 14, and then matched them with 1 We exclude children aged 6 and 15 because some children may delay in school beginning and some may complete compulsory school early. Analyses based on the data including those aged 5 and 15 years old yield similar results (available upon request).

180

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

Table 1 Hukou, current residence, and migrant status of children aged 7–14 in China, 1990–2005 (cell percentages). Current residence

Whether migrants

1990

2000

Hukou status

Total

Nonagricultural

Agricultural

No Yes

7.50 0.09

2.52 0.42

No Yes

3.78 0.06

No Yes Total

3.15 0.14 14.73

N

153,864

2005

Hukou status

Total

Nonagricultural

Agricultural

10.02 0.51

9.80 0.30

4.98 0.90

2.27 0.11

6.05 0.17

4.89 0.10

79.55 0.40 85.27

82.70 0.54 100

1.71 0.10 16.90

Hukou status

Total

Nonagricultural

Agricultural

14.79 1.20

11.27 0.77

6.95 1.61

18.22 2.38

6.89 0.29

11.77 0.39

6.15 0.16

10.55 0.46

16.70 0.63

69.53 0.50 83.10

71.25 0.60 100

1.06 0.06 19.47

60.37 0.58 80.53

61.44 0.63 100

City

Town

Village

195,741

41,613

household heads (typically fathers, hereafter referred to as parents), based on the variable indicating the relationship of the respondent with the household head. As a result, we were able to obtain child-parent records, as well as household records including geographic location, household registration status, and children’s migration status and living arrangement (as an indicator of family structure). It shall be noted that the definition of migration has been revised several times for the enumeration of the increasingly mobile population since the 1990s. In the 1990 census, the migrant population was defined as those who have stayed in the residence for over 12 months but have their hukou registered in other counties. In other words, only cross-county movements for over a year were considered as migration. In both the 2000 census and the 2005 mini-census, the migrant population is defined as those who have stayed in the residence for over 6 months but have their hukou registered in other townships/street committee. Therefore, the increase in the reported migration counts over years is partly due to the change in criterion (Wu and He, 2015). In this article, we omit the within-county migration in 2000 and 2005, which accounts for about 15% of all migration population, to make the definitions and temporal comparison more consistent. This restriction strategy also suits the issues we aim to tackle in this article, as many cases of residence-hukou separation (renhu fenli) within the county/city boundaries are indeed local residents, but choose to live in other residence close to their children’s school locations. The definition, together with information on place of origin, facilitates the construction of variables indicating the different migration statuses of children in subsequent multivariate analyses. The process of internal migration in China has become increasingly heterogeneous since 1990.2 While migration is literally defined based on the discrepancy between place of hukou registration and place of residence, researchers typically refer to migration as a one-way process from villages to cities among agricultural hukou holders. In fact, migration exists also within urban areas and within rural areas. Based on the micro sample data, Table 1 presents the cell percentages in tabulations between hukou status (non-agricultural vs. agricultural), residence type (city, town vs. village) and migrant status (no vs. yes) from 1990 to 2005 for children aged between 7 and 14 in China. Results in Table 1, supplemented to the published figures we cited earlier, indicate the slow pace of hukou reform and the gradual relaxation of the hukou system in controlling population migration in China: children with non-agricultural hukou accounted for 14.73% of all children of the same age in 1990, 16.90% in 2000, and 19.47% in 2005. However, the percentage of children living in urban areas increased much faster: children living in urban areas accounted for 16.75% (10.02 + 0.51 + 6.05 + 0.17) in 1990, 28.15% in 2000, and 37.93 in 2005. Further examination of the migration status of the school-age children suggests that, as expected, migration is mainly an adults’ game and most children aged 7–14 are non-migrants. In 1990 only 1.22% (0.51 + 0.17 + 0.54) of children was migrants; the figure increased to 2.19% in 2000 and 3.64% in 2005. Not surprisingly, rural children accounted for the majority of migrant children: 76.2% ([0.42 + 0.11 + 0.40]/1.22) in 1990, 77.2% in 2000, and 72.8% in 2005. If we break down migrant children by urban destinations (i.e., cities and towns) vis-à-vis rural destinations (i.e., villages), the percentage of children migrating to urban areas multiplied from 0.68% (0.51 + 0.17) in 1990 to 3.01% (2.38 + 0.63) in 2005, while that of children migrating to rural areas remains stable at around 0.6% over the period. Rural to urban migrant children (hereafter referred to as those with agricultural hukou), who typically received much public concern, increased from 0.53% (0.42 + 0.11) in 1990 to 2.07% (1.61 + 0.46) in 2005.

2 While the different duration in defining migration (one year in 1990 and half a year in 2000 and 2005) suggest that the size of migrant children reported in 1990 would have to be adjusted upward according to the new criterion in 2000 and 2005. Nevertheless, the temporal trend in migration is evident (Wu and He, 2015).

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

181

3.2. Analytical subsamples Our key interest is in the educational consequence for migrant children in urban destinations, with a focus on rural–urban migrant children. In the following analyses, there are three different analytical subsamples. The first subsample includes all children aged 7–14 in urban China from 1990 to 2005, which we employ to examine school enrollment of migrant children relative to that of urban non-migrant children (i.e., the urban children subsample). In a further step, we adopt an origin–destination approach (e.g., Liang and Chen, 2007) and focus only on those who migrate from rural to urban areas. Apart from these rural-to-urban migrant children, we also include non-migrant children both (including the left-behind) in the rural counties where this group of children came from and in the urban districts where they were living at the time of survey (i.e., the origin–destination subsample).3 Since information on migration origin is available only in the 2000 census and the 2005 mini-census, the origin–destination comparison is conducted only in these two years. The third analytical subsample is the rural-to-urban migrant children per se in the 2000 census and 2005 mini-census data, which we analyze to examine the impact of migration timing on the school enrollment (i.e., the rural–urban migrant subsample). 3.3. Variables The dependent variable in this study is the school enrollment status of children aged between 7 and 14. It is coded as a dummy variable, 1 if enrolled in school and 0 otherwise. There are three sets of main independent variables. In urban children subsample, the main independent variable is children’s migration status, coded 1 for those whose hukou is registered in other county (i.e., cross-county) of the same province of residence or in a different province (i.e., cross-province),4 and 0 for those whose hukou is registered in within the same district of residence.5 In the origin–destination analysis, the main independent variable includes three groups: rural–urban migrant children, urban non-migrant children in the districts where rural–urban migrant children are currently staying, and rural non-migrant children in the counties where migrant children came from. Here, rural–urban migrant children refer to those whose hukou is registered in rural areas outside the district of residence (i.e., cross-county/provincial migrant children in urban destinations with hukou registered in rural areas).6 Note that in this approach we uses place of hukou registration (rural areas) to define rural–urban migrant children, to obtain mutually exclusive samples of non-migrants in urban vs. rural areas. Place of hukou registration is different from hukou status, but the vast majority of people registered in rural areas hold an agricultural hukou. We also control hukou status in multivariate analyses. The three groups of children in the origin–destination sample are coded as two dummy variables with rural–urban migrant children as the reference group in multivariate analyses. To examine the interaction effect between migration status and family structure and to compare rural–urban migrant children to left-behind children (liushou ertong) specifically, we further divide the origin–destination sample into six groups by combining information on children’s migration status and living arrangement. The modified independent variable includes rural–urban migrant children with at least one parent absent, rural–urban migrant children living with both parents, urban non-migrant children with at least one parent absent, urban non-migrant children living with both parents, rural non-migrant children with at least one parent absent (liushou ertong), and rural non-migrant children living with both parents. In the last analytical sample of rural–urban migrant children, the main independent variable is time since migration. In the 2000 census and the 2005 mini-census, the timing of migration was sought by asking respondents a question on ‘‘arrival time at the current location’’ (Lavely, 2001; Liang and Ma, 2004). It is a categorical variable: less than 1 year, 1–2 years, 2– 3 years, 3–4 years, 4–5 years, or more than 5 years. They are coded as five dummy variables with less than 1 year as reference in multivariate analysis. We do not use continuous values of time since migration because information on the exact year of migrants who had left their place of hukou registration for more than 5 years by 2000, and more than 6 years by 2005 is not available. Other independent variables are the same for the three analytical subsamples. Hukou status is a dummy variable, with agricultural coded as 1 and non-agricultural as 0. Living arrangement includes four categories (i.e., with both parents; with one parent; with grandparents; and with others) coded as three dummy variables with the first category as the reference group. The household head’s occupation and education are included to proxy the impact of family background on children school enrollment. Household head’s occupation is coded into six categories, including managers and professionals, clerks, 3 A reviewer suggests us to include urban-to-urban migrant children as a group for comparison, which accounts for a tiny proportion of school-age children (see Table 1). Because our focus is on rural-to-urban migrant children and three subsamples are defined according to the analytic purpose to compare migrant children to non-migrants in both origins (including left-behinds) and destinations. Adding another group of migrants would distract the focus of our analysis. 4 Here county refers to counties in rural areas and districts in urban areas. To be certain, as an important part of China’s urbanization, some rural areas might have been converted to urban areas over years, in addition to the population migration from rural to urban areas. Because we consider the movement across county and provincial boundaries as migration in this article, the incorporation of villages neighboring the cities would have minimum consequences on the count of rural migrant children in cities in different census years. Indeed, cross-county and cross-provincial migration dominated the population migration in China since 1990 (Wu and He, 2015). 5 Note that in the 1990 census there is no information on the place of hukou registration and we used place of residence on July 1st, 1985 as a proxy. We deem the approximation appropriate since population migration and separate of hukou registration and residence were not prevalent in the mid-1980s. 6 Note again that we used place of residence on July 1, 1985 as a proxy of place of hukou registration in the 1990 census.

182

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

commercial and service workers, farmers, production workers, in addition to a missing category, with farmers as the reference group. Household head’s education is coded into five categories, including primary school or below, junior high school, senior high school, college or above, as well as a missing category, with the first category as reference. Gender and ethnic inequalities in education have long been documented by scholars in China (e.g. Hannum and Xie, 1994; Hannum, 2002; He and Wu, 2014; Wu and He, forthcoming; Wu and Song, 2014). We take into account these factors as controls in our multivariate analyses. Gender is coded as a dummy (female = 1), and so is ethnicity (Han = 1). Age is also a controlled variable, coded as seven dummies with age 7 as the reference group. Studies show that there exists a huge regional variation in both the migration rate and the school enrollment rate in China (Connelly and Zheng, 2003; Lu, 2007; Wu, 2010; Wu and Zhang, 2010). To consider the geographical heterogeneity, we include province as a set of dummy variables in the multivariate analyses. We also adjust for the clustering effect of counties of residence and report the robust standard errors, except for the origin–destination analysis, which does not allow us to do. 4. Empirical analyses 4.1. Descriptive statistics Table 2 presents descriptive statistics for the variables used in the urban children analysis. Overall, the school enrollment ratios have increased from 93% in 1990 to around 98% in 2000 and 2005. This is largely attributed to the educational expansion, and specifically to the successful implementation of nine-year compulsory education. In 1980, the Chinese government set the goals of universalizing primary education by the end of the 1980s and implementing nine-year compulsory education in the 1990s (Tsui, 1997). In 1985, the Decision on the Reform of the Education Structure was launched, followed by the 1986 Compulsory Education Law. With the increase in educational resources, these goals were largely achieved by the mid-1990s (Connelly and Zheng, 2003; Wu, 2010; Wu and Zhang, 2010). Table 2 Descriptive statistics for the sample of children aged 7–14 in urban China, 1990–2005 (%). 1990

2000

2005

93.0

97.9

97.5

97.8 2.2 30.9

94.4 5.6 46.4

92.1 7.9 51.6

Living arrangement With both parents With one parent With grandparents With others

80.4 8.6 9.1 2.0

79.0 7.2 11.4 2.4

68.8 8.1 16.3 6.9

Household head’s occupation Managers, professionals Clerks Sales and service workers Famers Production workers Missing

21.9 6.5 12.2 15.5 33.0 10.8

13.2 6.8 16.8 23.4 25.6 14.3

11.5 5.7 17.2 21.1 23.8 20.8

Household head’s education Primary school or below Junior high school Senior high school College or above Missing Han ethnicity Female

34.9 37.2 20.0 7.6 0.3 93.7 48.0

22.8 43.0 23.7 10.2 0.4 93.2 47.4

22.7 43.2 16.7 11.1 6.2 93.1 46.4

Age 7 8 9 10 11 12 13 14

14.0 14.5 11.0 12.0 12.6 12.1 11.6 12.2

9.9 10.5 11.0 13.5 13.7 13.8 14.5 13.2

11.4 11.1 11.4 12.6 12.5 13.5 13.4 14.3

N

25,296

55,109

16,463

Enrolled in school Migrant status Non-migrants Migrants Agricultural hukou

Notes: data are weighted for the 2005 mini-census.

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

183

The surge in internal migration in the 1990s was dominated by cross-county and cross-provincial mobility (Fan, 2008). Children’s migration is also featured with a similar pattern. Migrant children, as defined by either-county or cross-provincial ones, accounted for 2.2% of urban children in the same age bracket in 1990, 5.6% in 2000, and 7.9% in 2005. The proportion of children with agricultural hukou in urban areas increased from 30.9% in 1990 to 46.4% in 2000 and to 51.6% in 2005. Such an increase is a product of both the increase in internal population migration and the urban expansion in which certain rural areas are converted to urban areas over census years. Pattern of living arrangements changed for urban children over the years. The percentage of children living with both parents in urban areas dropped from 80.4 in 1990 to 68.8 in 2005. On the other hand, the percentage of children living with grandparents increased from 9.1% in 1990 to 11.4% in 2000 and to 16.3% in 2005; the percentage of children living with others increased from 2% in 1990 to 2.4% in 2000 and further to 6.9% in 2005. This change is unlikely due to the notable increase in divorce rates, as the percentage of children living with a single parent remained quite stable over the 15-year period.7 In the following, we first conduct a series of multivariate analyses to investigate how children’s migration status, hukou status, and living arrangement, as well as other demographic variables, affect their likelihood of school enrollment in urban China. We then restrict our targeted subjects to rural-to-urban migration children, and compare their school enrollment to their counterparts’ in both origins and destinations, with some specific information available in the 2000 census and 2005 mini-census data. Finally, we examine how migration timing affect school enrollment for the rural–urban migrant children. 4.2. Migrant children in urban China Table 3 presents the results from binary logistic regression predicting the likelihood of being enrolled in school for urban children aged 7–14 in 1990, 2000 and 2005. We include children’s migration status, hukou status, gender, ethnicity, living arrangement, household head’s occupation and education, age dummies and province as independent variables in the models.8 Models 1, 3 and 5 are additive models for 1990, 2000, and 2005, respectively. Migrant children are disadvantaged in school enrollment compared to non-migrant urban children. In 1990, migrant children’s odds of being enrolled in school are 45.7% lower (1 e 0.610) than the odds for non-migrant children, holding constant all other factors. In 2000 and 2005, migrant children’s odds of being enrolled in school are 55.3% (1 e 0.805) and 49.7% (1 e 0.687), respectively, lower than the odds for non-migrant children. Hukou status has a significant impact on school enrollment. Children with agricultural (rural) hukou are particularly disadvantaged in school enrollment over the whole examined period in urban China. They are significantly less likely to be enrolled in school than those with non-agricultural (urban) hukou. For instance, the odds of being enrolled in school for children with rural hukou are 23.6% (1 e 0.269) lower than the odds for children with urban hukou in 1990, 30.2% (1 e 0.359) lower in 2000, and 22.7% lower (1 e 0.257, p < .10) in 2005, holding constant all other factors. These findings suggest that the rural–urban inequality in educational opportunities was not alleviated in the 1990s with the hukou system at work, despite the significant progress that China had made in economic development and relaxation of control over population migration during that period (Connelly and Zheng, 2003; Wu, 2010, 2011; Wu and Zhang, 2010). In all three years, children living with grandparents and those living with their own parents do not differ significantly in the likelihood of attending school, but children living with others are significantly disadvantaged in this respect in both 1990 and 2000, although such disadvantages diminished and became insignificant in 2005. After introducing control variables, gender inequality still exists in 1990 and 2000 but disappears in 2005 (also see Wu and Zhang, 2010). Ethnic inequality exists between Han Chinese and ethnic minorities in 1990 and 2000 but disappears in 2005 as well. As previously shown in Table 1, there are both rural and urban hukou holders among migrant children in urban destinations. In Models 2, 4 and 6 of Table 3, we include interaction terms between hukou status and migration status. While the other effects largely remain the same, the interaction coefficient is negative and significant in 1990 and it becomes statistically insignificant in both 2000 and 2005. The negative effects of migration apply to all children in urban destinations irrespective of their hukou status. 4.3. Rural-to-urban migrant children: An origin–destination approach Migrant children are significantly less likely to be enrolled in compulsory school than non-migrant children in urban China. Chances in this respect are worsen for those with agricultural hukou because they are furthered punished by holding a rural hukou status. An ensuing question is: why do migrants bring their children given that migrant children are significantly disadvantaged in school enrollment compared with non-migrant children? To address the question, we focus now on rural–urban migrant children and compare them with those staying in the origin counties and those in the destination cities/towns in terms of the likelihood of school enrollment (see Liang and Chen’s 2007 paper on origin–destination analysis on the impact of migration on children’s education in Guangdong province using 7 Despite the notable increase in the divorce rate in China in recent years, especially in coastal cities, the national divorce rate remains low by international standards. The United Nations Statistics Division reports that in 1985 merely 0.4 out of 1000 marriages in China ended in divorce, and the rate increased to 1.6 out of 1000 in 2007. In comparison, the rate is 2.0 per 1000 in Japan in 2007, 4.8 per 1000 in Russia in 2007, and 5.2 per 1000 in the US in 2008 (Weber, 2012). 8 The coefficients of age dummies and province dummies are not reported in multivariate analyses to conserve space.

184

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

Table 3 Logit models predicting school enrollment of children aged 7–14 in urban China, 1990–2005. 1990

2000

2005

(1)

(2)

(3)

(4)

(5)

(6)

0.610*** (0.170) 0.269** (0.093)

0.467 (0.468) 0.227* (0.095) 1.277* (0.501) 0.096  (0.053) 0.302** (0.113)

0.805*** (0.109) 0.359*** (0.085)

0.687*** (0.180) 0.257  (0.153)

0.153* (0.063) 0.466*** (0.111)

0.567* (0.251) 0.334*** (0.088) 0.295 (0.274) 0.152* (0.063) 0.467*** (0.111)

0.122 (0.112) 0.237 (0.197)

0.793* (0.322) 0.275  (0.160) 0.151 (0.387) 0.123 (0.112) 0.236 (0.197)

Living arrangement (reference: with both parents) With single parent 0.035 (0.098) With grandparents 0.038 (0.128) With others 0.669*** (0.168)

0.036 (0.099) 0.036 (0.128) 0.696*** (0.166)

0.392*** (0.111) 0.097 (0.134) 0.603*** (0.174)

0.393*** (0.111) 0.096 (0.134) 0.605*** (0.174)

0.008 (0.214) 0.366  (0.213) 0.369 (0.363)

0.004 (0.213) 0.369  (0.213) 0.371 (0.365)

Household head’s occupation (reference: farmers) Managers and professionals 0.339** (0.131) Clerks 0.407** (0.156) Sales and service workers 0.243  (0.124) Production workers 0.350** (0.106) Missing 0.280  (0.150)

0.366** (0.131) 0.435** (0.156) 0.269* (0.124) 0.375*** (0.107) 0.303* (0.150)

0.649*** (0.150) 0.458* (0.185) 0.291** (0.098) 0.338*** (0.095) 0.345** (0.131)

0.660*** (0.151) 0.472* (0.185) 0.306** (0.099) 0.349*** (0.096) 0.355** (0.131)

0.500  (0.279) 0.021 (0.299) 0.083 (0.200) 0.213 (0.177) 0.171 (0.237)

0.492  (0.281) 0.031 (0.301) 0.075 (0.204) 0.205 (0.178) 0.163 (0.236)

Household head’s education (reference: primary or below) Junior high school 0.154* (0.073) Senior high school 0.131 (0.092) College or above 0.103 (0.159) Missing 0.033 (0.448) Age dummies Yes Province dummies Yes Constant 2.261*** (0.313)

0.155* (0.073) 0.132 (0.092) 0.095 (0.158) 0.082 (0.445) Yes Yes 2.215*** (0.311)

0.469*** (0.074) 0.687*** (0.102) 0.834*** (0.178) 0.841  (0.490) Yes Yes 2.891*** (0.279)

0.468*** (0.074) 0.684*** (0.102) 0.833*** (0.178) 0.837  (0.490) Yes Yes 2.872*** (0.279)

0.405** (0.144) 0.363  (0.205) 0.404 (0.267) 0.467 (0.397) Yes Yes 4.221*** (0.595)

0.407** (0.144) 0.367  (0.206) 0.411 (0.269) 0.472 (0.399) Yes Yes 4.235*** (0.598)

N

25,296

55,109

55,109

16,463

16,463

Migrant children Agricultural hukou Migrant children ⁄ agricultural hukou Female Han ethnicity

0.097  (0.053) 0.298** (0.113)

25,296

Notes: Sampling weights are used in the 2005 sample. Robust standard errors in parentheses adjusted for clustering effects on counties. *** p < 0.001. ** p < 0.01. * p < 0.05.   p < 0.10.

1995 1% population sample survey data). One may argue that, although rural migrant children face hurdles in attending schools in urban destinations, they still enjoy more educational opportunities than their peers back in their home villages. We test this claim by analyzing the data from the 2000 census and the 2005 mini-census, which contain the information needed to link the origins and destinations of rural–urban migrant children. Note again, here rural–urban migrant children refer to cross-county/province migrant children in urban destinations with hukou registered in rural areas. We employ three steps to obtain the analytical sample for the origin–destination analysis. First, we select rural to urban migrant children from the sample data of 2000 census and 2005 mini-census by current residence (i.e., cities/towns), migrant status (i.e., cross-county/province migrants), and place of hukou registration (i.e., rural areas: villages or village committees).9 Second, we list the districts where the rural–urban migrant children were living at the time of census and mini-census enumeration, and select all non-migrant children in those districts from the sample data. Third, we list the counties of origin where the rural–urban migrant children were registered, and select all non-migrant children living in those counties 9 Note again, the 1990 census contains no information on the place of hukou registration and we used place of residence on July 1, 1985 as a proxy. This approximation is reasonable as most people in China lived in the place where their hukou was registered before the 1990s.

185

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

from the sample data. Therefore, the analytical sample in the origin–destination approach includes rural–urban migrant children, urban non-migrant children in destination districts, and rural non-migrant children in origin counties. Descriptive statistics of the sample show that the school enrollment rate of rural–urban migrant children is lower than that of non-migrant children in both urban destinations and rural origins. In 2000, 93.8% rural–urban migrant children aged 7–14 (n = 1209) are enrolled in school, and 98.4% urban non-migrant children (n = 17,862) and 96.0% rural non-migrant children (n = 51,519) in the same age bracket are enrolled. In 2005, the school enrollment rate is 95.7% for rural–urban non-migrant children (n = 1026), and the corresponding figures are 97.8% for urban non-migrant children (n = 5289) and 96.9% for rural non-migrant children (n = 9191). Table 4 presents the results from the multivariate analysis adopting the origin–destination approach, which supports our hypothesis that rural–urban migrant children are disadvantaged in school enrollment compared to their non-migration

Table 4 Logit models predicting school enrollment of children aged 7–14 in China, 2000 and 2005 (rural–urban migrant origin–destination sample). 2000

Migrant status (reference: rural–urban migrants) Urban non-migrants (destination) Rural non-migrants (origin) Agricultural hukou Female Han ethnicity Living arrangement (reference: with both parents) With single parent With grandparents With others

2005

(1)

(2)

(3)

(4)

1.147*** (0.149) 0.542*** (0.131) 0.741*** (0.105) 0.516*** (0.044) 0.753*** (0.075)

1.309*** (0.155) 0.928*** (0.145) 0.244* (0.115) 0.509*** (0.044) 0.716*** (0.075)

0.681** (0.242) 0.479* (0.199) 0.350  (0.210) 0.075 (0.108) 0.241 (0.170)

0.704** (0.254) 0.624** (0.231) 0.054 (0.229) 0.085 (0.108) 0.205 (0.172)

0.202** (0.070) 0.036 (0.083) 0.252* (0.108)

0.062 (0.071) 0.311*** (0.093) 0.513*** (0.131)

0.080 (0.185) 0.017 (0.174) 0.099 (0.178)

0.161 (0.191) 0.092 (0.206) 0.513 (0.313)

Household head’s occupation (reference: farmers) Managers and professionals Clerks Sales and service workers Production workers Missing Household head’s education (reference: primary or below) Junior high school

0.783*** (0.185) 0.581* (0.261) 0.501*** (0.126) 0.500*** (0.092) 0.337** (0.116)

0.421 (0.325) 0.589 (0.528) 0.356 (0.238) 0.096 (0.175) 0.340 (0.269)

Yes Yes 3.914*** (0.782)

0.427** (0.130) 0.997*** (0.258) 0.470 (0.445) 0.537 (0.402) Yes Yes 3.045*** (0.824)

15,506

15,506

Age dummies Province dummies Constant

Yes Yes 2.724*** (0.422)

0.621*** (0.049) 0.877*** (0.092) 0.819** (0.259) 0.847*** (0.252) Yes Yes 1.313** (0.435)

N

70,590

70,590

Senior high school College or above Missing

Notes: The sample for each year includes three groups of children: (1) rural to urban migrant children who were cross-county/province migrants living in cities/towns with their hukou registered in rural areas; (2) urban non-migrant children living in the destination districts where rural–urban migrant children were staying at the time of census and mini-census; and (3) rural non-migrant children who were living in the rural origin counties where rural– urban migrant children were registered. Data are weighted for the 2005 sample. *** p < 0.001. ** p < 0.01. * p < 0.05.   p < 0.10.

186

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

peers not only in destination cities but also in rural origins. In both 2000 and 2005, non-migrant children in either destination cities/towns or origin counties are significantly more likely to be enrolled in school than rural–urban migrant children, controlling for hukou status, living arrangement, gender, ethnicity, and age and province dummies. In 2000, as Model 1 shows, the net odds of being enrolled in school for urban non-migrant children in destinations and rural non-migrant children in origin counties are 3.1 (e1.147) and 1.7 (e0.542) times, respectively, the odds for rural–urban migrant children. In 2005, as shown in Model 3, the corresponding figures are 2.0 (e0.681) and 1.6 (e0.479) times the odds for rural–urban migrant children. Even after controlling for family socioeconomic status (household head’s occupation and education), the pattern still holds in Models 2 and 4. Therefore, the negative impact of migration on school enrollment is very likely due to the social exclusion in the destinations on the basis of place of hukou registration. Otherwise, these rural–urban migrant children do not differ from their peers back in the villages. As our results here reject the view that rural–urban migrant children enjoy more educational opportunities than their peers back in the villages, the question remains why their migrant parents bring them. Prospective data are needed for analyzing the family migration strategies and motivation to answer this question. To make a direct comparison between migrant children and left-behind children, we combine children’s migration status and family living arrangement to generate a new independent variable of main interest, and conduct further analysis on the

Table 5 Logit models predicting school enrollment of children aged 7–14 in urban China, 2000 and 2005 (rural–urban migrant origin–destination sample). 2000

Migrant status (reference: rural–urban migrants with at least one parent absent) Rural–urban migrants with both parents Urban non-migrants with at least one parent absent Urban non-migrants with both parents Rural non-migrants with at least one parent absent (left-behind children) Rural non-migrants with both parents Agricultural hukou Female Han ethnicity

2005

(1)

(2)

(3)

(4)

0.927*** (0.269) 1.521*** (0.254) 1.931*** (0.234) 1.193*** (0.225) 1.243*** (0.221) 0.764*** (0.106) 0.516*** (0.044) 0.752*** (0.075)

0.948*** (0.271) 1.938*** (0.259) 2.081*** (0.236) 1.705*** (0.231) 1.638*** (0.228) 0.279* (0.115) 0.513*** (0.044) 0.713*** (0.075)

0.639 (0.405) 1.307** (0.424) 1.130** (0.380) 0.976** (0.358) 0.977** (0.354) 0.356  (0.210) 0.074 (0.108) 0.232 (0.169)

0.574 (0.403) 1.458*** (0.426) 1.101** (0.379) 1.151** (0.369) 1.093** (0.363) 0.067 (0.230) 0.086 (0.108) 0.198 (0.172)

Household head’s occupation (reference: farmers) Managers and professionals Clerks Sales and service workers Production workers Missing Household head’s education (reference: primary or below) Junior high school

0.759*** (0.184) 0.548* (0.261) 0.468*** (0.126) 0.461*** (0.092) 0.421*** (0.113)

0.406 (0.327) 0.557 (0.526) 0.334 (0.239) 0.070 (0.175) 0.271 (0.235)

Yes Yes 3.398*** (0.834)

0.423** (0.133) 1.001*** (0.259) 0.483 (0.449) 0.091 (0.265) Yes Yes 2.584** (0.883)

15,506

15,506

Age dummies Province dummies Constant

Yes Yes 2.043*** (0.460)

0.599*** (0.048) 0.852*** (0.092) 0.785** (0.259) 0.200 (0.232) Yes Yes 0.667 (0.474)

N

70,590

70,590

Senior high school College or above Missing

Notes: The analytical sample is the same as that used in Table 4; data are weighted for the 2005 sample. *** p < 0.001. ** p < 0.01. * p < 0.05.   p < 0.10.

187

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

origin–destination sample. Results are shown in Table 5. Model specifications in the table are the same as those in Table 5, except for the main independent variable. Overall, results in Table 5 lend further support to our previous conjectures that rural–urban migrants are the most disadvantaged group, presumably because of the social exclusion they are subjected to in the destinations on the basis of their hukou registration and that migrant children living with at least one parent absent are even more disadvantaged because they are subject to both social exclusion and loss of social capital. Specifically, compared with rural–urban migrant children Table 6 Logit models predicting time effects on school enrollment of rural–urban migrant children aged 7–14 in China, 2000–2005 pooled data.

Time since migration (reference: less than 1 year) 1–2 years 2–3 years 3–4 years 4–5 years More than 5 years Agricultural hukou Female Han ethnicity Living arrangement (reference: with both parents) With single parent With grandparents With others

(1)

(2)

0.618* (0.278) 0.611  (0.314) 1.363** (0.492) 1.679** (0.589) 1.312*** (0.369) 0.074 (0.469) 0.184 (0.236) 0.650 (0.395)

0.523  (0.272) 0.622  (0.323) 1.301* (0.507) 1.692** (0.622) 1.327*** (0.380) 0.384 (0.471) 0.199 (0.246) 0.417 (0.375)

0.267 (0.363) 0.339 (0.601) 0.933** (0.302)

0.192 (0.376) 0.243 (0.682) 1.030*** (0.311)

Household head’s occupation (reference: farmers) Managers and professionals

0.279 (0.528) 0.544 (1.369) 0.680 (0.446) 0.331 (0.442) 0.011 (0.567)

Clerks Sales and service workers Production workers Missing Household head’s education (reference: primary or below) Junior high school

0.340 (0.232) Yes Yes 1.447  (0.771)

0.911*** (0.242) 1.043* (0.469) 0.520 (0.645) 0.209 (0.746) 0.370 (0.244) Yes Yes 1.400 (0.872)

2137

2137

Senior high school College or above Missing Year: 2005 Age dummies Province dummies Constant N

Notes: The sample includes only rural to urban migrant children who were cross-county/province migrants living in cities/towns with their hukou registered in rural areas; Robust standard errors in parentheses adjusted for clustering effects on counties. *** p < 0.001. ** p < 0.01. * p < 0.05.   p < 0.10.

188

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

with at least one parent being absent, non-migrants are significantly more likely to be enrolled in school, regardless of whether they live with both parents or not. In particular, left-behind children in rural areas enjoy more educational opportunities than rural–urban migrant children with at least one parent absent. Side analyses on the coefficients in Models 2 and 4 show that, left-behind children are also more likely to be enrolled in school than rural–urban migrant children living with both parents, holding constant the other factors. Furthermore, left-behind children do not differ significantly in school enrollment from rural non-migrant children with both parents. Among rural–urban migrants, those living with both parents are significantly more likely to be enrolled in school than those with at least one parent absent in 2000, but the effect becomes insignificant in 2005, probably due to the fact that the remittance from the migrant parents could compensate to some extent for the absence of parental guardianship. 4.4. Rural-to-urban migrant children: Migration timing Previous analyses have shown the disadvantages rural–urban migrant children face in school enrollment. Now we turn to examine whether chances improve for this group of children as duration of stay increases. Table 6 presents the multivariate analysis on the effect of migration timing on the school enrollment of rural–urban migrant children. We pool the data from 2000 and 2005 to obtain a larger sample size. Confirming the previous observations, results show that, as rural–urban migrant children spend more time in the destinations, they are more likely to be enrolled in school. The results remain largely the same even after controlling for the characteristics of the family’s socioeconomic background. 5. Conclusions and discussions Population migration has been on the rise in China since the mid-1990s and continues to be extremely popular to this day. How migrant children fare in this large-scale social transformation has become an important issue receiving much attention from scholars and social commentators. The question bears a lot of policy implications for the government. We have examined the school enrollment of children aged between 7 and 14 in urban China, using the micro-data from population censuses in 1990 and 2000 and the mini-census in 2005. Our results confirm earlier studies that migrant children are less likely to be enrolled in school than non-migrant children in urban areas from 1990 to 2005. While children of rural hukou are particularly disadvantaged in school enrollment, the effect of migration status applies to all children in urban China regardless of their hukou status. The absence of parents or grandparents also negatively affects the likelihood of school enrollment in both 1990 and 2000 urban China. Children living with people other than parents and grandparents are much less likely to be enrolled in school. The effects become statistically insignificant in 2005, probably because of the near saturation of compulsory education in China in the 1990s. We further employ an origin–destination approach to compare rural-to-urban migrant children with their non-migrant counterparts in urban destination (districts) and rural origin counties, using 2000 and 2005 data. Results show that rural migrant children are significantly less likely to be enrolled in school than non-migrant children in both origin counties and destination cities/towns. In particular, they are significantly less likely to be enrolled in school than even the left-behind children in rural areas. Finally, we examined the impact of migration timing on school enrollment for rural–urban migrant children, and showed that, as they spend more time in destinations, the likelihood of being enrolled in school increased. Hence, the study reveals a broad picture of migration and children’s schooling and delineates the changing trend over the decades in contemporary China. The patterns we have identified, nevertheless, are descriptive at best and suffer from several limitations. First, the population census/mini-census data we have analyzed do not allow us to effectively address the endogenous nature of the migration decision, which could depend on unobserved characteristics experienced by households as well as the abilities and preferences of parents. These concerns likely affect whether migrant parents bring their children with them or not (Xu and Xie, 2013). While we are unable to address the endogeneity and selectivity issue of migration in the empirical analyses due to data limitation, our results on the disadvantages faced by migrant children would be more prominent, had we taken into account of the selection effect, presumably positive one. Second, while we have been able to examine one important aspect of child development only, i.e., school enrollment, migration and separation from parents may have profound implications for other aspects such as children’s academic performance, psychological well-being, behaviors and their development of non-cognitive skills (Lu and Zhou, 2013; Ren and Treiman, 2013), all of which are as important as schooling in determining their future labor productivity and status attainment in destination cities/towns. Finally, as we discussed, sociologists argue that the negative consequences of migration on children’s schooling are due to the loss of social capital. Our analysis in the context of China largely confirms the negative consequences but is unable to provide a sociological explanation. To what extent, is the cause of negative consequences for education attributable to the fact of not having a local hukou rather than the loss of social capital in the community of origin and migrant families’ lower socioeconomic status? We look forward to further data collection and analyses in the future to examine explicitly how social capital interacts with family economic resources, institutional constraints, and social exclusions in shaping children’s outcomes in urban China. As such, social policies to address the issues of migration children’s education should be directed not only to the removal of hukou system based on which the migrant children are excluded from educational opportunities but also to the social supports that assisting them in adapting themselves to the new environment.

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

189

References All-China Women’s Federation (ACWF) Research Team, 2013. Woguo Nongcun Liushou Ertong, Chengxiang Liudong Ertong Zhuangkuang Yanjiu Baogao. (Research Report on the Situation of Rural Left-Behind China, Rural–Urban Migration Children in China). (accessed 10.06.13). Astone, Nan Marie, Mclanahan, Sara S., 1991. Family structure, parental practice, and high school completion. Am. Sociol. Rev. 56 (June), 309–320. Astone, Nan Marie, Mclanahan, Sara S., 1994. Family structure, residential mobility, and school dropout: a research notes. Demography 31, 575–584. Cai, Fang, 2002. Employment in Rural and Urban China: Issues and Options. Beijing, China: Social Science Academic Press (in Chinese). Chan, Kam Wing, 1994. Cities with Invisible Walls. Oxford University Press, Hong Kong. Chang, Hongqin, Dong, Xiao-yuan, MacPhail, Fiona, 2011. Labor migration and time use patterns of the left-behind children and elderly in rural China. World Develop. 39 (12), 2199–2210. Coleman, James, 1988. Social capital in the creation of human capital. Am. J. Sociol. 94, S94–S120. Connelly, Rachel, Zheng, Zhenzhen, 2003. Determinants of school enrollment and completion of 10 to 18 year olds in China. Econom. Educat. Rev. 22 (4), 379–388. Du, Yang, Park, Albert, Wang, Sangui, 2005. Migration and rural poverty in China. J. Comp. Econom. 33 (4), 688–709. Duan, Chengrong, Zhou, Fulin, 2005. Woguo Liushou Ertong Zhuangkuang Yanjiu’’ (A Study on Left-behind Children. Populat. Res. 29 (1), 29–36 (in Chinese). Fan, Cindy, 2008. China on the Move: Migration, the State, and the Household. Routledge. Farley, Reynolds, Alba, Richard, 2002. The new second generation in the United States. Int. Migration Rev. 36 (3), 669–701. Hannum, Emily, 2002. Educational stratification by ethnicity in China: enrollment and attainment in the early reform years. Demography 39, 95–117. Hannum, Emily, Xie, Yu, 1994. Trends in educational gender inequality in China: 1949–1985. Res. Social Stratif. Mob. 13, 73–98. Hao, Lingxin, 2012. Cumulative causation of rural migration and initial peri-urbanization in China. Chinese Sociol. Rev. 44 (3), 6–33. He, Guangye, Wu, Xiaogang, 2014. Gender Earnings Inequality in Reform-Era Urban China. University of Michigan PSC Research Report No. 14-830. October 2014 . Lavely, William, 2001. First impressions from the 2000 census of China. Popul. Develop. Rev. 27, 755–769. Lee, Leng, Park, Albert, 2010. Parental Migration and Child Development in China. Working Paper. . Li, Jun, Gu, Yanfeng, Zhang, Chuncen, 2015. Hukou-based stratification urban China’s segmented economy. Chinese Sociol. Rev. 47 (2), 154–176. Liang, Zai., 2001. The age of migration in China. Popul. Develop. Rev. 27 (499), 524. Liang, Zai., Chen, Yiu Por., 2007. Educational consequences of migration for children in China. Social Sci. Res. 36, 28–47. Liang, Zai, Ma, Zhongdong, 2004. The floating population of China: new evidence from the 2000 population census. Popul. Develop. Rev. 30 (3), 467–488. Liang, Zai, White, Michael, 1996. Internal migration in China 1950–1988. Demography 33, 375–384. Liang, Zai, Guo, Lin, Duan, Chengrong, 2008. Migration and well-being of children in China. Yale-China Health J. 5, 25–46. Long, Larry H., 1975. Does migration interfere with children’s progress in school. Sociol. Educat. 48, 369–381. Lu, Yao, 2007. Educational status of temporary migration children in China: determinants and regional variations. Asian Pacific Migration J. 16 (1), 29–55. Lu, Yao, Zhou, Hao, 2013. Academic achievement and loneliness of migrant children in China: school segregation and segmented assimilation. Comp. Educat. Rev. 57 (1), 85–116. Ma, Zhongdong, 2001. Urban labor force experience as a determinant of rural occupation change: evidence from recent urban–rural return migration in China. Environ. Plan. A 33, 237–255. McLanahan, Sara S., 1985. Family structure and the intergenerational transmission of poverty. Am. J. Sociol. 90, 873–901. National Bureau of Statistics in China, 2006. Report on 2005 National One Percent Population Sample Survey (accessed 24.06.2013). Ou, Dongshu, Kondo, Ayako, 2013. In search of a better life: the occupational attainment of rural and urban migrants in China. Chinese Sociol. rev. 46 (1), 25–59. Portes, Alejandro, Zhou, Min, 1993. The new second generation: segmented assimilation and its variants. Ann. Am. Political Social Sci. 530, 74–96. Pribesh, Shana, Downey, Douglas B., 1999. Why are residential and school moves associated with poor school performance? Demography 36, 521–534. Ren, Qiang, Treiman, Donald J., 2013. The Consequences of Parental Labor Migration in China for Children’s Emotional Well-being. In: Poster presented at the annual meeting of the Population Association of America, 11–13 April. Roberts, Kenneth D., 1997. China’s ‘Tidal Wave’ of migrant labor: what can we learn from Mexican undocumented migration to the United States? Int. Migration Rev. 31, 249–293. Silver, Hilary, 1994. Social exclusion and social solidarity. Int. Labour Rev. 133, 531–578. Tach, Laura, 2015. Social mobility in an era of family instability and complexity. Ann. Am. Acad. Political Social Sci. 2015, 83–96. Teachman, Jay D., Paasch, Kathleen, Carver, Karen, 1999. Social capital and dropping out of school early. J. Marriage Family 58, 773–782. Tsui, Kai Yuen, 1997. Economic reform and the attainment in basic education in China. The China Quart. 149, 104–127. Wang, Feng, Zuo, Xuejin, Ruan, Danching, 2002. Rural migrants in Shanghai: living under the shadow of socialism. Int. Migration Rev. 36, 520–545. Weber, Claire, 2012. Chinese Divorce Rate: China’s Divorce Rate is Rapidly Increasing. . Wu, Xiaogang, 2009. Household Registration, Social Exclusion, and Rural Migrants in Chinese Cities. In: Wu, Guoguang, Lansdowne, Helen (Eds.), Socialist China, Capitalist China: Social-Political Conflicts under Globalization, Routledge, pp. 29–54. Wu, Xiaogang, 2010. Economic transition, school expansion and educational inequality in China, 1990–2000. Res. Social Stratif. Mob. 28 (2010), 91–108. Wu, Xiaogang, 2011. The household registration system and rural–urban educational inequality in China. Chinese Sociol. Rev. 44 (2), 31–51. Wu, Xiaogang, He, Guangye, 2015. The evolution of population census undertakings in China, 1952–2010. China Rev. 15 (1), 203–238. Wu, Xiaogang, He, Guangye, forthcoming. Changing ethnic stratification in contemporary China. J. Contemp. China. Wu, Xiaogang, Song, Xi, 2014. Ethnic stratification amid China’s economic transition: evidence from the Xinjiang Uyghur autonomous region. Social Sci. Res. 44, 158–172. Wu, Xiaogang, Treiman, Donald J., 2004. The household registration system and social stratification in China, 1955–1996. Demography 41 (2), 363–384. Wu, Xiaogang, Zhang, Zhuoni, 2010. Changes in educational inequality in China, 1990–2005: evidence from the population census data. Res. Sociol. Educat. 17, 123–152. Xiang, Biao, 2007. How far are the left-behind left behind? a preliminary study in rural China. Popul. Space Place 13, 179–191. Xie, Yu, Greenman, Emily, 2011. The social context of assimilation: testing implications of segmented assimilation theory. Social Sci. Res. 40, 965–984. Xu, Hongwei, Xie, Yu, 2013. The causal effects of rural-to-urban migration on children’s wellbeing in China. In: Paper presented in the Population Association of America Annual meeting. New Orleans, LA, April 11–13. Yang, Juhua, Fan, Ni, 2012. Migration, family structure and high school enrollment: an analysis of China’s 2000 census. Chinese Sociol. Rev. 44 (4), 27–57. Yang, Quanhe, Guo, Fei, 1996. Occupational attainments of rural to urban temporary economic migrants in China, 1985–1990. Int. Migration Rev. 30, 771– 787. Zhang, Zhuoni, 2014. Contingent transition to triumph: socioeconomic attainments of Chinese immigrants’ children in Hong Kong. Chinese Sociol. Rev. 46, 68–88. Zhang, Zhuoni, Wu, Xiaogang, 2013. Occupational Segregation and Earnings Attainment of Rural Migrants in Urban China. University of Michigan PSC Research Report No. 13-809 .

190

X. Wu, Z. Zhang / Social Science Research 53 (2015) 177–190

Zhao, Yaohui, 2000. Rural to urban labor migration in China: the past and the present. In: West, Loraine A., Zhao, Yaohui (Eds.), Rural Labor Flow in China. University of California Press, Berkeley, CA, pp. 15–33. Zheng, Bingdao, Gu, Yanfeng, 2012. Tax-for-fee Reform as Exogenous Shocks: The Slowdown of Rural Migration in Chinese Villages, 1998–2002. Working Paper, Division of Social Science, Hong Kong University of Science and Technology. Zhou, Min, Bankston, Carl L., 1998. Growing up American: how Vietnamese children adapt to life in the United States. Russell Sage Foundation, New York.