Community Poverty, Industrialization, and Educational Gender Gaps in Rural China

This paper investigates community impoverishment and industrialization as explanations for educational gender gaps in rural China with analysis of a multi-province household survey and a longitudinal study of youth in one impoverished province. We consider attributes of poor communities that might shape gaps and the related roles of household and community poverty. Three major results emerge from this paper: community impoverishment, not industrialization, correlates with gaps; poverty and isolation shape gaps differently at different educational levels; and girls in relatively wealthy households fare better than boys at the transition to high school. Results suggest the importance of theorizing differences by educational stage and the need for research that conceptualizes the non-local dimensions of industrialization as potential considerations in educational decisions. Comments Cherng, H. and Hannum, E. (2013). Community Poverty, Industrialization, and Educational Gender Gaps in Rural China (Working Paper). Gansu Survey of Children and Families. This working paper is available at ScholarlyCommons: http://repository.upenn.edu/gansu_papers/41 Community Poverty, Industrialization, and Educational Gender Gaps in Rural China May 1, 2013 This paper investigates community impoverishment and industrialization as explanations for educational gender gaps in rural China with analysis of a multi-province household survey and a longitudinal study of youth in one impoverished province. We consider attributes of poor communities that might shape gaps and the related roles of household and community poverty. Three major results emerge from this paper: community impoverishment, not industrialization, correlates with gaps; poverty and isolation shape gaps differently at different educational levels; and girls in relatively wealthy households fare better than boys at the transition to high school. Results suggest the importance of theorizing differences by educational stage and the need for research that conceptualizes the nonlocal dimensions of industrialization as potential considerations in educational decisions. INTRODUCTION Across the world, one of the most consistent findings in educational research has been the declining disadvantage for girls, and, indeed, that a girls’ advantage has emerged in many settings (e.g., Blossfeld and Shavit 1993; Dorius and Firebaugh 2010; Fiske 2012; Grant and Behrman 2010). Globally, gender gaps are concentrated in certain regions, and among the poorest households in those regions (Filmer 2005). A secular trend of declining gender inequality in education is also apparent in China (Filmer 2005; Hannum and Xie 1994; Hannum 2005). Some recent evidence raises questions about the degree to which gaps favoring boys persist at all in some areas of China, and at some educational transitions, including the transition to higher education (Hannum et al. 2010; Tsui and Rich 2002; Wu and Zhang 2010). Also paralleling research on the global concentration of gender disparities in certain regions, research in China indicates considerable geographic variability in the degree of gender disparity, with girls’ disadvantage concentrated in rural areas, which are much poorer than urban areas, and in poor households (Connelly and Zheng 2007a; Connelly and Zheng 2007b; Hannum 2005; Hannum et al. 2009; Hannum et al. 2010; Wu and Zhang 2010). We know little, however, about why gender gaps persist in some poor communities: is the source rooted more in labor market explanations and incentives these contexts provide to parents, or in dimensions of community poverty? Do the effects of community poverty operate through household economic conditions, and do attributes of community disadvantage associated with community poverty matter? 1 The regions cited are South Asia and North, Western, and Central Africa (Filmer 2005). This study was based on analysis of the Demographic and Health Surveys, which do not cover China. 2 For the sake of brevity, we will refer to the “traditional” gender gap – girls’ disadvantage – as “gender gap” in the rest of this paper. This paper investigates the state of gender disparities in education in rural areas. First, to test the community correlates of gender disparities, we analyze a multi-province rural household income survey from 2002, the Chinese Household Income Project (CHIP) (Shi, Chinese Household Income Project, 2002). Second, to investigate how particular dimensions of poor communities relate to gender gaps, we analyze a singleprovince case study of one of China’s poorest provinces and one that has persistent gender disparities. For this purpose, we use the Gansu Survey of Children and Families (GSCF), a longitudinal survey of rural children that contains extensive measures of the household and community conditions of the childhood home of respondents (University of Pennsylvania 2010). Community Poverty and Gender Gaps in Education Cross-National Perspectives A substantial body of research has examined community effects on education in contexts across the world (Ainsworth 2002; Crane 1991; Garner and Raudenbush 1991; Harding et al. 2010; Kling, Liebman, and Katz 2005). One line of this research investigates how dimensions of community context are associated with gender gaps in educational outcomes. For example, one mixed-method study argues that the origins of gender gaps in math among high-scoring middle schoolers in the United States may lie in access to neighborhood academic programs (Entwisle, Alexander, and Olson 1994). Another line of research investigates how community poverty influences educational outcomes, more broadly (Brown and Park 2002; Connell 1994; Brooks-Gunn, Duncan, and Aber 1997; Hannum 2003). 3 Using a case study of a US metropolitan area, the authors found that girls spent less time in their community environment than boys, and as a consequence, were less able to access community resources that could foster academic success. As a result, the authors argue, high-scoring boys out-scored their female counterparts. However, few sociological studies have investigated how community poverty, particularly in middle and low income countries, is linked to educational gender gaps. One dimension of communities that has been linked theoretically to gender gaps in developing country contexts is the degree of industrialization of labor markets. Empirical work on trends in educational gender gaps has explicitly or implicitly explained trends with a modernization, industrialization, or development framework, which suggests that the shift of the work force out of agriculture and into industry, or the development of a “modern” economy, decreases girls’ disadvantage in education. This framework suggests that as patriarchal societies move out of agriculture and into industry, old cultural or family-based incentives for prioritizing boys over girls will decline (Forsythe, Korzeniewicz, Durrant 2000; Treiman 1970, but see Hannum 2005 for inconsistencies with the Chinese case). Of course, many scholars have contested modernization theory over the years, starting most notably with the work of Ester Boserup, who argued that men were best positioned initially to take advantage of new opportunities for work posed by industrialization and development (Boserup 1970; 1989). Emerging labor markets under industrialization and development might reinforce incentives to families for investments in boys, at least in the short term (Boserup 1970; 1989; Boserup et al. 2007; Forsythe et al. 2000a; Forsythe et al. 2000b). Although communities that experienced movement out of agriculture and into industry are likely to be less poor than those that remain primarily agricultural, a modernization or industrialization framework does not directly address the issue of whether dimensions of community poverty, beyond effects via industrialization, matter for gender inequality. This omission is potentially important, as economists studying developing countries have found, in many national settings, that community poverty is associated with educational gender gaps (Colclough, Rose, and Tembon 1999; Filmer 2000; World Bank 1999). Community poverty is commonly assumed to shape gender disparities through household decision-making processes. The household economy framework, or household decision-making framework, common in economics, suggests that parents make decisions about human capital investments in their children based on perceived monetary costs and benefits associated with educating children (for recent articulations of this framework, see Filmer 2005; Huisman and Smits 2009). These rational calculations are assumed to differ for sons and daughters, for example, if parents expect daughters to be less likely to earn returns on their education, or if those returns are deemed less likely to flow to the natal family due to traditions of co-residence with sons (Filmer 2005; Huisman and Smits 2009). Assuming that parents with sufficient resources attempt to provide education for all of their children, one might expect this calculation to disadvantage girls most notably among poor families, for whom resources to invest in children are constrained. To the extent that poor communities have higher numbers of the poorest households than other communities, the effect of community poverty may simply reflect a preponderance of the poorest households, where girls are educationally disadvantaged. However, it is also possible that girls in general, not just girls in the poorest households, are disadvantaged relative to boys in poor communities. This pattern could emerge if households are responding not just to their own resource constraints, but also to elements of the broader community context that shape incentives and risks for investing in boys and girls. For example, as implied by the industrialization framework, the presence or absence of local employment opportunities outside of agriculture may create or impede incentives to educate girls, relative to boys. Research has highlighted a number of other characteristics of poor communities that matter for educational access, and may matter for gender disparities. For example, scholars have argued that community isolation influences overall enrollment for boys and girls, but may disadvantage girls more than boys if parents are unwilling to send daughters to far away schools due to safety concerns (Colclough, Rose, Tembon 1999; Jacobs 1996, UNESCO 1998). An investigation of determinants of primary school enrollment in 30 developing countries found that when children have to travel longer distances to school, they are less likely to go to school (Huisman and Smits 2008). In Turkey, Tansel (2002) found that the distance between a community and a regional metro center was associated with lower probabilities of school enrollment for boys and girls, but that the disadvantage for girls was greater. Community infrastructure has also been shown to be associated with enrollment, and could be linked to gender gaps in some contexts. In one study of rural Bangladesh, rural electrification was associated with increased school participation for all rural children (Khandker 1996). One study using household surveys found that Pakistani women who resided in villages with convenient access to water spent less time gathering water each day and more time employed in these villages than women in villages with weaker infrastructure (Ilahi and Grimard 2000). To the extent that girls are also responsible for domestic tasks, it may be the case that girls who reside in villages with stronger infrastructure spend less time on household chores and more time in school. Community support for education could also benefit students and potentially narrow gender gaps. Development research has highlighted the importance of community involvement for improving school enrollment rates. One review of World Bank projects designed to improve access to quality primary education in sub-Saharan Africa argued that community support is vital for successful interventions, and effective programs often incorporate the viewpoints of community members (Heneveld and Craig 1996). Collaborations between project and community leaders also strengthen the community's commitment to education, which in turn increases enrollment rates. Community leaders and members who support the education of youth in their communities may also share progressive views of girls’ education. Evaluations of USAID projects in Mali found that efforts to increase community participation in education, both by financial contributions and involvement in school planning, led to marked increases in girls’ enrollment and entry into primary school (Rugh 2000; Tietjen 1999). Although this study investigated girls’ enrollment and not gender gaps, per se, it is likely that villages that share norms that encourage girls’ education may also have smaller gender gaps. Research in China Many studies attest to secular declines in girls’ educational disadvantage in China (e.g., Hannum and Xie 1994; Hannum 2005; Hannum et al. 2010; Wu and Zhang 2010; Zhou et al. 1998). However, temporal patterns in the decline are not very consistent with a story of development or industrialization: much of the narrowing of the gap emerged during the Cultural Revolution (Hannum and Xie 1994; Hannum 2005). In rural areas, where the majority of the population resides, incentives persist for son preference in educational investments. In most rural areas of China, sons are the primary source of old age support, and for this reason, parents face strong incentives to invest in sons as long-term insurance. Research in a rural county in Yunnan province indicates that expectations of support from sons are more pronounced among mothers in poorer, more remote rural areas (Li and Lavely 2003). In rural Gansu Province, which is one of China’s poorest provinces, only about 15 percent of mothers did not expect to get support from children, and the vast majority of those who did expected to get support from sons (Hannum et al. 2009). Further, expectations of future support were tied to aspirations for children’s education, which, in turn, were predictive of children’s subsequent school persistence (Zhang et al. 2007). However, recent research among families of primary and junior high school age children in Gansu Province has suggested few overt signs that families were substantially biased toward boys in their educational attitudes or practices, though modest gender gaps did exist in parental aspirations for children and years of schooling attained by children (Hannum et al. 2009). The same study found no evidence of performance differences between boys and girls. There is some evidence linking poverty to gender inequality. Song et al. (2006) report that female schooling is a “luxury good:” it is more income elastic than male schooling. Consistent with this viewpoint, other work with older data has suggested that higher income households make more egalitarian decisions about enrolling sons and daughters in school (Hannum 2003; 2005). A small body of literature has also explored the links between community poverty on the education of boys and girls in rural China. Prior scholarship investigates how village resources condition educational opportunities for both boys and girls in rural China (Adams 2006; Adams and Hannum 2005; Knight, Li, Deng 2009). Knight et al. (2009) and Adams and Hannum (2005) find that county and village economic resources are strongly associated with school enrollment, and Adams (2006) finds that average village income is also associated with higher test scores. However, little research focuses on community poverty and gender gaps. One study using older data found that girls' enrollment in schools benefited more than boys’ in terms of enrollment from living in villages with higher average income (Hannum 2003). Scant research explores the implications of particular characteristics of poor communities, such as labor market opportunities, isolation, infrastructure, or support of education, for gender gaps in China. Further, little research has considered whether gender disparities in access associated with community and household wealth operate differently at different educational stages. In China, compulsory education encompasses nine years of education, which includes primary and middle schools (Ministry of Education 1986). High schools in rural China serve more villages and charge much higher fees than middle schools, and scholars have argued that higher fees influence rural girls’ disadvantage at this transition (Wu and Zhang 2010). Therefore, household economic constraints may matter more at the transition to high school than at compulsory stages. Summary and Research Questions Overall, research on gender gaps points to a long-term decline in female disadvantage in many global contexts. Prior work has shown that there are still persistent gender gaps – for example, in parts of rural China – and there is some evidence that pockets of female disadvantage are associated with poverty. However, little research has systematically explored the link between female disadvantage with degree of industrialization—emerging non-agricultural labor markets—or with other aspects of community poverty 4 . It also remains unclear whether the link between community poverty and gender gaps simply reflects the aggregated effect of girls’ disadvantage in poorer households, and whether the story may differ by educational stage. In this paper, using the case of educational gender gaps in rural China, we address