7
Which Girls Stay in School? The Influence of Family Economy, Social Demands, and Ethnicity in South Africa

Bruce Fuller and Xiaoyan Liang

Introduction

Considerable evidence demonstrates the influence of formal schooling on women's fertility and child health practices, as reviewed in earlier chapters. Notwithstanding important exceptions observed in Africa, longer schooling for girls and young women appears to alter an important range of fertility and child health behaviors (Cochrane, 1979; Hill and King, 1993). Accumulating evidence also shows that spending more of one's childhood in school usually results in higher levels of learning—despite uneven levels of school quality and a still hazy understanding of how classroom processes alter girls' knowledge or beliefs in ways that affect fertility and maternal health practices (Lockheed and Verspoor, 1991).

We are operating largely in the empirical dark, however, on a parallel issue, which is the focus of this chapter: What local forces—operating from both within the family and the surrounding economic and institutional environment—lead some girls to stay in school longer while many others drop out?

International agencies are moving ahead rapidly with ambitious programs aimed at increasing female school enrollment. But when we focus on impoverished communities and low-income families, we find very little empirical evidence on the factors that explain why some daughters stay in school while others exit. Western models of school attainment, relying on mobility traditions within industrialized societies, are partially useful in illuminating the economic and cultural forces that may explain girls' educational attainment in developing-country settings (Fuller et al., 1995a; Haveman and Wolfe, 1995). The purpose of this



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 181
--> 7 Which Girls Stay in School? The Influence of Family Economy, Social Demands, and Ethnicity in South Africa Bruce Fuller and Xiaoyan Liang Introduction Considerable evidence demonstrates the influence of formal schooling on women's fertility and child health practices, as reviewed in earlier chapters. Notwithstanding important exceptions observed in Africa, longer schooling for girls and young women appears to alter an important range of fertility and child health behaviors (Cochrane, 1979; Hill and King, 1993). Accumulating evidence also shows that spending more of one's childhood in school usually results in higher levels of learning—despite uneven levels of school quality and a still hazy understanding of how classroom processes alter girls' knowledge or beliefs in ways that affect fertility and maternal health practices (Lockheed and Verspoor, 1991). We are operating largely in the empirical dark, however, on a parallel issue, which is the focus of this chapter: What local forces—operating from both within the family and the surrounding economic and institutional environment—lead some girls to stay in school longer while many others drop out? International agencies are moving ahead rapidly with ambitious programs aimed at increasing female school enrollment. But when we focus on impoverished communities and low-income families, we find very little empirical evidence on the factors that explain why some daughters stay in school while others exit. Western models of school attainment, relying on mobility traditions within industrialized societies, are partially useful in illuminating the economic and cultural forces that may explain girls' educational attainment in developing-country settings (Fuller et al., 1995a; Haveman and Wolfe, 1995). The purpose of this

OCR for page 181
--> chapter is to advance theoretical understanding of the role of the family's character and practices, especially variability in the work and social demands parents place on daughters, in determining when girls leave school. First, we review how sociologists and economists have represented the school attainment process within developing-country settings. Second, we report on our household-level study of female school attainment among young black South Africans. After reviewing the economic and social factors that explain local variation in girls' school attainment, we examine how those factors help explain whether and when girls leave school. The final section presents a summary and conclusions. We find that the propensity of daughters to remain in school is associated with the family's economic consumption levels, the family's social structure, and labor and social demands placed on young females. We also find significant variability among black ethnic groups in when daughters typically leave school, variation that corresponds in part to regional histories and school supply. Our analysis suggests that unless policy and program designers take local institutions and particular conditions into account, their well-meaning interventions may be costly and ineffective. Household And Institutional Determinants Of Girls' School Attainment We often confound how parents make decisions about fertility behavior with whether, or how, parents press their children to stay in school. Working from a rational-choice framework, household economists and demographers argue that (1) parents come to see that more individualized ("higher-quality") childrearing holds greater economic utility, and (2) this realization leads to more determined family planning, lower birth rates, and stronger motivation to keep daughters in school for longer periods of time (Easterlin and Crimmins, 1985). Beyond the household are macroeconomic and institutional forces: evolving labor demand for female workers, shifts in family income, modern forms of social status, and women's variable commitment to reproductive independence (Caldwell and Caldwell, 1990). A major ingredient of this shifting institutional context is the expansion of mass schooling. Through mass education, the state attempts to define the legitimate social roles that girls should normatively fill. Yet only recently has a family-level literature emerged from sociology that attempts to link the household's immediate economic situation to the broader institutional and cultural context by examining how forces at these two levels of social organization shape girls' school attainment (e.g., Walters and James, 1992). The first conceptual step is to separate household-level processes from both the economic context and social institutions, such as the state and school, that may variably condition family-level action.

OCR for page 181
--> Neoclassical Household Decision Making Recent work by Haveman and Wolfe (1995:1832) reiterates the neoclassical claims by which variation in children's educational attainment is often explained: Economists have viewed the process of children's school attainment to be an aspect of the theory of family behavior. The family is viewed as a production unit which employs real inputs in order to generate utility for its members. Adults ...make decisions regarding the generation of family economic resources [and] determine the uses of these resources. Parents make a variety of other choices such as fertility, location, and family stability that influence the returns to productive efforts. The amount of family resources allocated to children, the nature of these resources, and the timing of their distribution influence the school attainment of children. This model is cross-generational with regard to school attainment: it argues that parents' decisions about their fertility behavior and family structure occur concurrently with choices about their children's school attainment. These decisions, although bundled by the rational parent, aim to advance the individual child's human capital, thereby advancing the family's long-term welfare (Becker, 1976). Caldwell's earlier work (1976) borrows heavily from this framework, arguing that the "emotional and economic nucleation" of the family occurs as parents come to see, within modernizing contexts, that bearing more children will not necessarily lead to greater long-term wealth or security in old age. As firms and institutions, especially the state and modern schools, exogenously come to associate social status with greater human capital investment, parents factor this into their calculus for allocating household-level resources and labor demands in ways that boost their children's school attainment. Becker and Tomes (1976:S148) note that as parents become upwardly mobile, they invest more heavily in the quality rather than the quantity of children, and they often move to communities where "public contributions to their children's schooling would be greater." Here the argument shifts to the contextual or institutional level: the commercializing and modernizing context, highly variable across local regions, is changing the structure of wage labor and economic opportunity. In addition, the formal school gains the authority to allocate youths to status and social-class positions. Even household economists are now struggling to incorporate context, rather than simply postulating that preferences are exogenous to the intrafamily utility-maximization process. Haveman and Wolfe (1995:1837), for example, move beyond the family unit by arguing that governments make choices to improve poor families' economic and social context. Parents are viewed as "choosing" various family practices from their environment: "the sort of monitoring, disciplinary, nurturing, and expectational environment in which their children are raised." Drawing on the notion of ''social capital" advanced by Loury (1977) and Coleman (1988), Haveman and Wolfe argue that the state can act in ways that

OCR for page 181
--> strengthen family and kin support and guide both the character and intensity of parents' investments in their children. The family thus mediates information and resources emanating from institutions to improve intrahousehold socialization practices and advance school attainment.1 The households and surrounding institutional arrangements are therefore acting in tandem. Institutional Resources and Constraints That Bound Family Choices We must delineate more crisply those institution-level forces which condition intrafamily decision making. Indeed, many of the quandaries over the education-fertility link expressed by demographers and others interested in population issues may stem from an inadequate understanding of how household-level processes and institutional arrangements operate simultaneously. In demographic circles, for instance, the grand theory of demographic transition has proven to be of limited utility in some African contexts: in countries such as Botswana or Kenya, economic commercialization unfolds, modern rules emerge, female enrollment rises, but fertility behavior changes only slightly (Isiugo-Abanihe, 1994). Similarly, school attainment levels have been highly uneven among societies and within particular regions, despite steady modernization and huge investments in mass education since the 1960s. Females comprise about 25 percent of all secondary school pupils in Chad, versus 56 percent in Namibia. Fewer than half of all children in sub-Saharan Africa reach grade 5, and girls complete 1.5 fewer school years on average relative to boys (World Bank, 1995). Such variability in the way local parents make decisions about their children's schooling, and over time their own fertility behavior, might be explained by three institution-level forces. First, the organization of job opportunities varies in ways that establish long-term or immediate returns to staying in school. Labor demands operate differently on girls versus boys. In many parts of Botswana and South Africa, for example, we observe that daughters actually stay in school longer than sons, a phenomenon that can be explained in part by centuries of male labor migration to either commercial centers or cattle fields. In addition, subsistence crop yields are often quite low in arid regions of Southern Africa, so that there is ample discretionary time for daughters to pursue schooling without cutting into time required for domestic chores (Kossoudji and Mueller, 1983). 1   These arguments also depart from the earliest mobility or status-school attainment studies of sociologists from the 1960s forward. This work has focused on how parents, with variable education levels and occupational status, attempt to reproduce the family's position through human capital investment. The question of mobility has been central to this line of work: Can offspring move up the status hierarchy independently of their parents' class position? This work has not, however, been able to incorporate changes in context, other than to see context as a residual set of exogenous factors that cannot be captured by within-family variables (e.g., Campbell, 1983).

OCR for page 181
--> Second, the state's educational policies, formulated over time, can induce family demand for female schooling. African countries have expanded their secondary school systems at widely varying rates since the 1960s. The Demographic and Health Surveys of the late 1980s, for instance, showed that just over half of all women aged 20-24 in Zimbabwe had completed some secondary schooling, versus just 2 percent in Mali and 10 percent in Ghana (Bledsoe and Cohen, 1993). The state also can influence norms about the legitimate location and role of young females. When the Botswana government reduced the terminal year of junior-secondary school from grade 10 to grade 9, female enrollment rates quickly began to decline after grade 9 (Fuller et al., 1995a). In addition, the state can influence parents' expressed demand for schooling by lowering the direct costs of enrollment, relieving labor demand for young girls (by investing in water wells or child-care centers), offering free lunches, and encouraging female enrollment through all-girls schools or targeted scholarship programs (Hill and King, 1993; Glewwe, 1994; Bradshaw and Fuller, 1996). School quality and student assessment processes can influence enrollment demand, as well as the extent to which girls are encouraged to achieve at high levels. In one study from Kenya, more than a third of all girls achieving in the bottom quartile dropped out of secondary school because of pregnancy, compared with just one-seventh of the top students (Division of Family Health, 1988, cited in Bledsoe and Cohen, 1993). Distance to school and the presence of female teachers also may encourage higher school attainment among young females (DiStefano, 1993; Fuller and Clarke, 1994). Third, parents' cultural commitments and social practices, extracted from a modernizing institutional environment, may further explain variability in female school attainment. A variety of cultural practices have received empirical attention in Africa: the argument that sexual and marriage practices are culturally determined, occasionally altered by modernizing values (Whiting and Whiting, 1991); the role of polygyny and resulting demands placed on young women (Lesthaeghe, 1989); bride price and the extent to which schooling is factored in; initiation rites and role conflicts for young females who are now culturally ordained as "adults"; and the role of kin, not only the husband and wife, in making reproductive decisions (see Bledsoe and Cohen, 1993; Hyde, 1993). Locally Situating the Household Model How can we situate the family within variable local contexts to better explain girls' school attainment? A family-economy model has emerged recently in sociology. This model situates the household's actions within (1) local economic demands and (2) more localized social or cultural commitments that stem from a shared set of evolving institutions, at times situated within a common ethnic history. The household is still seen partly in neoclassical terms as a social collective that actively seeks to optimize its members' welfare, not dissimilar to

OCR for page 181
--> a neoclassical model similar to Becker's (1976). But the family's "rationality" and opportunities are set by both local labor demands and institutional norms, and the family is seen as mediating these macro forces in allocating children's time between work and schooling. How has this model—attempting to balance economic and social forces—been applied to the question of school attainment at the child level? As an example, the new family-economy model has helped explain an early contradiction in the school attainment literature: as societies industrialized from an agricultural or commercial base, enrollment rates often grew most rapidly in rural areas, not within industrializing centers where neoclassical theory accurately predicted higher economic returns to more schooling (Fishlow, 1966). In the United States, the historical evidence on household decision making shows that parents took into account immediate labor demands and income opportunities, not longer-term optimization or human capital strategies. Urban parents could move their children into factory or early service jobs, pulling them out of school to advance the household's immediate welfare. In rural areas, children experienced more discretionary time between planting and harvest seasons. This economic dynamic interacted with rural parents' Calvinist commitment to formal education, boosting secondary school attainment levels for girls and boys above those observed in urban centers (Walters and O'Connell, 1988). These family-economy theorists emphasize the state's contextual role in expanding the supply of schooling, reducing private costs, and regulating child labor—factors that boosted school attainment (Walters and James, 1992). We recently extended the family-economy framework to explain rising female school attainment observed in Botswana villages since independence (Fuller et al., 1995a). Two major findings emerged from this study. First, economic dynamics mediated by the household must be culturally situated and seen in gender-specific terms. Daughters were more likely to persist further through secondary school, for instance, when the number of household chores they were assigned was closer to the number assigned to sons. The presence of younger children and additional child-care obligations contributed to adolescent daughters' greater propensity to leave school. We also found a tendency to stay in school longer among daughters in father-absent households in the two southernmost villages, those located closest to the South African job market. While not always statistically significant, this finding is consistent with Schultz's (1990) report that when mothers serve as the household head, resources for schooling are allocated more equally to daughters. Studying several African societies, Lloyd and Blanc (1996) found that school attainment for both sons and daughters was higher in female-headed households, after taking into account economic factors (including generally lower income levels).2 2   The presence and authority of the father also have implications for the mother's relative power over fertility decision making, as discussed by LeVine et al. (1994a) and Mburugu (1994).

OCR for page 181
--> Our second major finding was that mothers' attributes and social demands helped explain how far their daughters went in secondary school, again after taking into account the household's economic characteristics (within a survival analysis framework). In these Botswana villages, variability in the mother's schooling, assessed literacy level, perceived utility of formal education, family planning practices, and childrearing objectives all contributed to her daughter's school attainment. As posited by the new family-economy model, the discretionary time afforded by the economic context and parents' allocation of children's time between work and school contributed to daughters' school attainment. Next, we clarify how the family-economy model can be applied to variable regions and ethnic groups within South Africa. We then assess empirically how household-level factors help explain why girls' school attainment varies among poor black families. This family-level analysis reveals cross-ethnic and regional differences in female school attainment, leading us back to the institution-level force of school supply and action by local states. We turn first to a consideration of how economic and social forces, observed within the household, can be more clearly conceptualized. South Africa: Explaining Local Variation In Girls' School Attainment Poor Families Mediating Economic and Social Linkages The family-economy model posits that as parents consider the utility of keeping their children in school, they operate from preferences and anticipated streams of monetary benefits. Yet these "preferences" are viewed as stemming from variable economic demands and institutional norms, observable within the household, that pertain to childrearing aims and commitments to schooling. Parents attend to these (variable) local economic and social demands on the basis of the material and human resources they can draw from the environment. The family's linkages to the economic and social context vary in character across local regions. For example, in historically diverse and balkanized societies, such as South Africa, ethnic membership frequently corresponds to the family's regional location or designated "homeland," a point to which we will return. Economic Linkages Borrowing from the neoclassical household model, sociologists have emphasized the importance of economic resources and labor demands as parents allocate their children's time between work and school. Families are situated within specific labor structures that provide variable levels of income and job opportunities. These labor demands may compete for youths' time that otherwise would be available for schooling. Under the family-economy model, the resources and

OCR for page 181
--> demands inherent in the organization of work are augmented by state action, especially government's ability to legitimate school attendance, rather than fulltime work, as the norm for youths (Horan and Hargis, 1991; Walters and James, 1992). In our empirical model of daughters' school attainment, we include the household's variable economic resources and indicators of labor demand. Social Linkages The desire to go beyond macrostructural explanations of family decision making has renewed interest in situating children within a local social or cultural network. Anthropologists, of course, have long been focused on the situational norms and cultural models that guide childrearing and early socialization (Ogbu, 1978; Whiting and Whiting, 1991; LeVine et al., 1994). Rather than focusing on the content of socialization, however, Coleman (1988) examines the basic architecture of social relations, especially levels of obligation and trust within the family, and how these normative obligations may influence children's propensity to stay in school and achieve at higher levels. Under his construction of "social capital," he delineates structural features of the family that he and Schneider relate empirically to children's school attainment (Schneider and Coleman, 1993). The presence of parents or kin and their obligatory expectations are posited to constrain (or advance) children's school attainment levels through a number of mechanisms: the absence of one parent, a mother who works outside the home, a greater number of siblings, little talk at home about the child's life or schoolwork, and parents' low expectations and lack of pressure on the child to go on to college (Coleman, 1990:595). The presence of parents and the pressure they place on children to do well in school have in turn been linked to the household economy. Astone and McLanahan (1991), for instance, detail how working single parents spend less time supervising their children's homework or discussing their school lives. This framework assumes that family-level linkages offer social resources to children: moral encouragement and specific strategies for doing well in school. Household economists are trying to incorporate the social capital construct, defining parental time as a resource or an investment made in children's development (Haveman and Wolfe, 1995:1836). As originally coined by Loury (1977), the term "social capital" referred to social linkages and obligations that enhance (or constrain) initial entry and advancement in the job market. Yet certain social linkages and obligations place social demands on children that pull them out of school. This is especially so in impoverished contexts where secondary schooling is not a fully legitimated institution, including developing-country settings where girls' school attendance is not taken for granted. The dampening effect of fathers' presence on their daughters' school attendance in Botswana, for example, may be due to domestic and marital obligations imposed by the father (see also Lloyd and Blanc, 1996).

OCR for page 181
--> A second element of the social capital construct—the content of social obligations and expectations between parent and daughter—has yet to be specified or related to female school attainment. While endorsement of the modern school is acquired from a modernizing institutional environment, it is manifest and observable within the household (Caldwell, 1986; United Nations, 1995). We take modest steps here to assess whether observable social resources and social demands placed on daughters, stemming from their parents' education and literacy levels, contribute to daughters' school attainment in South Africa. In the discussion, we return to what little is known about the content of these evolving social expectations. Variability Among Poor Provinces and Households The above intrafamily dynamics may be conditioned by the regional economic and institutional context in which the household is situated. Given its history of apartheid and racist migration controls, South Africa shows significant variation across regions and ethnic groups in the strength of regional economies, rates of father absence, and maternal schooling levels (see also Thomas, this volume). We focus on variability among regions populated by black ("African") families. Population growth rates, for instance, currently range from 2.6 to 3.9 percent in regions populated by Xhosa versus North Sotho and Venda communities (Erasmus, 1994). The separation of black ethnic groups into "homelands" also led to variability in the availability and quality of secondary schooling. Outside white communities, South African families suffer from extreme levels of poverty. Fifty-three percent of all South Africans currently live below the poverty line, recently established as US$97 in monthly household expenditures per capita (1993 prices),3 and blacks comprise 94 percent of this total. Indeed, two-thirds of all blacks, totaling 21 million adults and children, live in poverty. About two-thirds of all poor households are located in the three provinces on which our analysis focuses: the Eastern Cape (populated primarily by Xhosa speakers), KwaZulu/Natal (Zulu speakers), and the Northern Transvaal Province (North Sotho, Venda, and smaller groups). Figure 7-1 shows locations. Almost 70 percent of all female-headed households subsist below the poverty line, as compared with 44 percent of families with a resident father. Note that the old provincial boundaries were still in effect in 1993. Secondary school enrollment rates are much lower among youths living in poverty. For households in the first and second poorest quintiles, just 46 and 57 percent, respectively, of children aged 13-17 are enrolled in school. In contrast, 3   The poverty line is based on the 40th percentile of all households, using household income per adult equivalent, adjusting for family composition. Data and exchange rates are for 1993 (Reconstruction and Development Programme, 1995).

OCR for page 181
--> FIGURE 7-1 Map of South Africa. this rate is 83 percent for families in the most affluent quintile.4 Among the poor, just one-fifth have electricity, and almost half use wood as their main source of energy. South Africa's infant mortality rate remains at 53 per 1,000 live births; the total fertility rate stands at 4.1 (Erasmus, 1994; Reconstruction and Development Programme, 1995).5 Variability Across Regional Contexts Our analysis focuses on girls' school attainment among the largest black ethnic groups. We begin by focusing on the context and attributes of the three 4   The percentage of females enrolled in secondary schools exceeds that of males in most provinces. In the old black homeland areas in 1990, 58 percent of all grade 10 students who sat for the matriculation exam were females; females comprised just over half of all who passed the national exam. But relatively few black women enter a college or university. At the University of Cape Town and the Witwatersrand University, African females comprise just 9 and 7 percent of total enrollment, respectively (Truscott, 1994). 5   White and colored populations are relatively affluent, compared with African families. But income inequality is stark: the richest 10 percent of all households, comprising just 6 percent of the population, accounts for 40 percent of all consumption. The lowest 40 percent, representing 53 percent of the population, accounts for less than 10 percent of total consumption (Reconstruction and Development Programme, 1995).

OCR for page 181
--> most populous communities: Zulu, Xhosa, and North Sotho speakers. Table 7-1 presents descriptive statistics for the three groups and for the respective province within which the majority of each resides. "Sample count" refers to the number of households for which data were reported for a mother and daughter pair, that is the eldest daughter still resident in the household, age 11-25. Our subsample, stemming from a 1993 national probability sample, equaled 1,503 households, including households belonging to one of the five largest black groups and reporting household consumption levels at or below two times the poverty line (which included over 90 percent of all sampled households from these ethnic groups).6 Mean per capita income for 1988 across the three provinces was uniformly low but variable. The Northern Transvaal province showed the lowest income level, equaling just US$319 (1988 exchange rate and prices), compared with $841 in KwaZulu/Natal. Per capita income averaged $1,116 nationally. An estimated 61 percent of all Zulus live below the poverty line, versus 92 percent of Xhosas and 83 percent of North Sothos. Despite a more rural and weaker economy than the other two provinces, North Sotho adults display similar school attainment levels, which may historically have been specific to males. In 1993, the proportion of adults who had completed secondary school equaled 13.4 percent among North Sothos, versus 10.8 percent among Zulu households. Young adolescents in North Sotho communities currently display much higher rates of persistence through secondary school: in 1993, 54 percent of all children who began grade 1` had persisted to grade 11, whereas this figure equaled just 26 and 11 percent, respectively,7 among Zulu and Xhosa youths. Our earlier analysis showed no literacy advantage among young North Sotho males or females, and the lower level of selectivity associated with lower dropout rates probably contributes to a lower matric pass rate among North Sotho secondary graduates (28 versus 43 percent for Zulu graduates; see Fuller et al., 1995b). Household and Daughter Sample The attributes of young females, shown in Table 7-1 for eldest daughters residing in sampled households differ across the three ethnic groups. The percentage of females enrolled in school ranged from about 60 percent and 57 6   These data are from a national probability sample from 360 randomly selected clusters that yielded data on 8,848 households nationwide. The sampling plan is detailed in Project for Statistics on Living Standards and Development (1994). Gender-related analyses of school attainment and assessed literacy levels appear in Fuller et al. (1995b). 7   For North Sotho communities, grade 12 enrollment counts are actually higher than those for grade 11, as youths reportedly stay in school to sit repeatedly for the matriculation exam. This rise in grade 12 enrollments is not observed in Zulu or Xhosa communities overall.

OCR for page 181
--> risk of leaving school; therefore, negative coefficients indicate that the predictor lower the risk of dropping out. Household expenditures and access to credit (family has a loan from a neighborhood shop) significantly lower the daughter's risk of leaving school. Recall that all these families have quite low incomes. Yet consumption levels show variations that have implications for which daughters stay in school longer. Model 3 adds the social factors that may influence female school attainment. Overall, these social factors exert a greater influence on reducing the model's chi-square statistic than do the family-economy factors (bottom of Table 7-2). Particularly influential in reducing the risk of leaving school are mother's and father's school attainment levels, father's absence from the home, target daughter's age, and age of the youngest daughter. As suggested earlier, the father-absent effect tends to confirm Schultz's (1990) finding that female-headed households, despite lower income levels, generally allocate resources more equitably to daughters relative to sons than do households headed by the father. In Model 3 we also now observe significant effects from the mother's involvement in the wage sector: such involvement places daughters at greater risk of leaving school, perhaps because these daughters face greater labor demands within the household. Girls are also more likely to leave school if more adult females are resident in the household. This finding suggests either that adult females do not share the labor demands of the household, or that these families perpetuate more traditional social roles for young females, including less integration in the female labor market. Model 4 adds another social factor: the daughter's ethnic membership. Girls in North Sotho communities are much less likely to leave school. Xhosa females display a lower risk of dropping out as well, although the difference is less than for North Sothos (Zulus form the base group). These patterns are consistent with descriptive results reported earlier (Figure 7-3). Thus the economic and social factors studied at the family level are not substituting for these striking ethnic differences. In Model 5 we add the control predictor indicating whether the young female has ever experienced a pregnancy. This factor understandably raises the risk of leaving school. However, it substitutes for just two earlier observed effects: the number of resident adult females in the household and (partially) Xhosa ethnic membership. The former substitution effect supports the argument that households with more resident females either face more limited job opportunities or hold more traditional beliefs about gender roles. Comparative Risks of Leaving School Figure 7-6 presents comparative hazards of leaving school for eldest daughters situated in different kinds of households. Panel A shows a significantly higher likelihood of completing the matric (10 percent lower) or completing

OCR for page 181
-->

OCR for page 181
--> FIGURE 7-6 Comparative hazards of leaving school by social characteristics. some post-secondary education (11 percent lower) for daughters whose mothers had completed primary school relative to those whose mothers had no schooling. Note that these hazards are cumulative across time periods. Thus, in survival terms, 51 percent of daughters whose mothers completed primary school had completed some post-secondary education, versus 35 percent of those whose mothers had no schooling. Panel B shows the modest advantage of daughters in father-absent households—an 8-point gap for daughters who enter post-secondary education. In survival terms, 75 percent of daughters in father-absent households completed secondary school and the matric, versus just 67 percent of those in father-present households. Ethnic differences in school attainment are illustrated in Panel C. The probability that North Sotho females will leave school after matric is 15 points lower than that for Zulu daughters. This gap rises to 21 points for Zulu and North Sotho daughters who enter post-secondary education. Overall, 67 percent of North Sotho daughters complete some post-secondary schooling, versus just 35 percent of Zulu females. Does Ethnic Membership Mask Regional and Institution-Level Forces? The ethnic differences discussed above return us to the theoretical question of how household and institution-level forces may together shape female school attainment. The household factors on which we have focused certainly help

OCR for page 181
--> TABLE 7-3 School Supply and Distinguishing Features of North Sotho Communitiesa   North Sotho Clusters [n = 16] Other African Clusters [n = 113] F value Average number of secondary schools 1.1 0.7 4.14* Percentage of clusters with tarmac roads 0% 24% 2.53ζ Index of health care workersb 3.7 3.9 0.64 Percentage of clusters with resident doctor 13% 20% 1.11 a North Sotho communities are defined as those with more than 80 percent of households with North Sotho speakers. Overall multivariate analysis of variance is significant: F = 2.65, * p < .05, ζp <.10. b Count of whether the cluster has a midwife, family planning, and/or health worker in a local clinic. explain which daughters are more likely to be enrolled and how far they persist through school. Yet even after taking into account a variety of family-level economic and social factors, ethnic effects remain strong and significant. What is it about ethnic membership—or the region in which a group predominates—that affects levels of female school attainment? One argument is that the availability of secondary schooling may vary across provinces in ways that help explain these ethnic effects. We assessed this possibility by drawing on community-level data collected in each of the local communities from which sampled households were randomly drawn. Our subsample of the five major black groups includes 136 local clusters with complete data. We split these clusters between those where the share of North Sotho households equaled or exceeded 80 percent versus 20 percent or less. The former group is located mostly in the Northern Transvaal province, primarily within the old Lebowa homeland. Table 7-3 compares North Sotho and other clusters. We found that the average number of secondary schools located in North Sotho communities equaled 1.1, compared with 0.7 per local cluster in the other black communities, a statistically significant difference. Interestingly, North Sotho communities have less infrastructure in terms of tarmac roads and healthcare workers. But the greater supply of secondary schooling is clear as compared

OCR for page 181
--> with the other black locales. Future research should examine such contextual differences more deeply, especially given that government policy and investments can address these institutional factors. Summary And Conclusions We have seen how low-income South African families are far from homogeneous. Elements of the household's economy, social resources and demands, and ethnic experience vary significantly and in ways that explain daughters' school attainment levels. After summarizing our basic findings, we speak to how this kind of evidence can inform stronger designs for the program interventions of governments and international agencies. Our findings are generally consistent with claims made by researchers who employ the new family-economy model. Even when restricting our analysis to low-income African households, we find that the family's financial strength—indicated by household expenditures and access to local credit—is positively related to the likelihood that eldest daughters remain in school and advance their school attainment. One important finding is that mothers' linkage to the formal wage sector may suppress their daughters' school attainment, perhaps because of the associated increase in domestic labor demand placed on the eldest daughter. Taken together, these findings confirm a basic family-economy process whereby resources and labor demands are associated with daughters' school attainment. Attainment appears to be linked to immediate and short-term labor demands, not necessarily to long-term optimization or human-capital investment, as neoclassical economists would have it. Thus far it seems that economic or structural interventions, such as income supplements, scholarships, or strategies for relieving female labor demand, might effectively boost girls' school attainment. We also find that social resources and demands placed on daughters contribute to female school attainment. The positive influence of father absence is consistent with earlier findings in several African societies (Fuller et al., 1995a; Lloyd and Blanc, 1996). The way this influence operates inside households deserves more attention, through both survey and ethnographic research. The influence of mothers' school attainment on daughter's attainment is not surprising, but the finding is notable given that we are focusing on impoverished black families. Ethnic membership is consistently telling for North Sothos, independently of the influence of family-economy factors. We also document the expected negative effect of becoming pregnant on daughters' school attainment levels. And it appears that many young females return to school after giving birth, providing an opportunity for more focused policy initiatives and interventions inside schools. Given the rising rate of adolescent pregnancy in some African societies (Bledsoe and Cohen, 1993), efforts to delay the age of first and

OCR for page 181
--> subsequent pregnancies for young females could effectively slow population growth in South Africa. The strong influence of social resources and demands suggests that program interventions in some local regions and with some ethnic groups must focus on more than economic constraints and incentives. Where parents' own schooling levels are low, encouraging daughters to stay in school will be especially challenging. Influencing the attitudes of mothers alone may not be sufficient. When fathers are present in the household or rural labor demands are inextricably linked to daughters' social obligations, interventions might better be aimed at males at both the household and village levels. Our analysis accents the higher school attainment levels found among North Sotho daughters, despite high rates of poverty and lower school attainment among older mothers. The more recent expansion of secondary education in North Sotho communities appears to be part of the key to understanding this advantage among young North Sotho females. This finding highlights the sharp variations among regional and ethnic situations, even across low-income black communities. Families of the same ethnicity—be it Zulu, Tswana, or North Sotho—do share important individual and household-level attributes: language, family structure, and patterns of child socialization. But ''ethnicity" in South Africa also manifests shared local institutional histories, thanks to the apartheid regime's migration controls and forced segregation of black groups under homeland administrations. The arid and rural Northern Transvaal province, where the former Lebowa homeland for North Sothos is situated, has a depressed regional economy, even relative to other poor regions. We saw earlier how per capita income is 2.5 times higher in KwaZulu/ Natal than in the North. Among all North Sotho speakers, 83 percent live beneath the poverty line, versus 61 percent of Zulus. Father absenteeism is twice as high in the Northern Transvaal province, in part because of males' movement to stronger labor markets. Thus the economic context lowers the opportunity cost for females who remain in school. Job opportunities are few; many males are leaving the household. Females are left with the task of raising a growing number of young children, as population growth remains a point higher in the Northern Transvaal relative to KwaZulu/Natal—3.9 percent versus 2.8 percent per annum, respectively. In addition to a shared economic context, particular ethnic groups have faced a distinct set of institutions over time. Demographers and anthropologists have tried to specify how changes in the institutional context are related to gains in maternal education, lower birth rates, and changed childrearing practices (Caldwell, 1986; LeVine et al., 1994a). But precisely how school, state, and church institutions act to legitimate female school attendance and novel fertility patterns remains a mystery. We do know that family demand for schooling is high in the Northern Transvaal province relative to other black communities. We reported earlier how a much higher percentage of North Sotho youths persist to

OCR for page 181
--> grade 11 as a proportion of entering grade 1 pupils. Part of this advantage is attributable to the variable actions of local governments. For example, within schools that fell under the pre-1994 Lebowa administration, 54 percent of all youths persisted into grade 9, versus just 21 percent of black students attending schools run by the national education department (Verwey and Munzhedzi, 1994).15 And prior research has revealed how school improvement efforts in the old Bophuthatswana bantustan have resulted in measurable gains in matric exam scores and literacy levels among young females (Education Foundation, 1993; Fuller et al., 1995b). It is the actions of local governments and institutions that make a difference. Beyond secular action by the state, the Tswana and Sotho peoples of the Western and Northern Transvaal have been exposed to a common linguistic and religious history, one that may have affected female commitments to formal schooling over time. Of particular importance, missionaries from competing churches moved aggressively into the Transvaal during the mid-nineteenth century. Literacy came to be associated with one's faith and loyalty to the village church. To compete for members, churches continue to translate the Bible and other texts into indigenous languages (Hofmeyr, 1993). Mission and, eventually, government schools came to signal a new form of status for young women (see, e.g., Bozzoli, 1991). We certainly have much to learn about how parents' beliefs and practices vary among households and in ways that contribute to daughters' school attainment levels. Clearly these micro household processes are rooted in parents' own schooling experience and the benefits of education they perceive. Yet our fundamental point is that the local economic situation and social institutions—contexts that differ across regions and ethnic groups in South Africa—further advance female school attainment. International agencies continue to promulgate means by which female enrollment can be raised or fertility behaviors altered (e.g., Odaga and Heneveld, 1995). But in many African societies we have scarce empirical knowledge of variations in the actions of families that are situated within specific regional histories, institutional contexts, and ethnic experiences. Where institutional conditions vary beneficially—such as the greater supply of secondary schools in North Sotho communities—opportunities exist for policy makers to act with greater precision. But as long as development agencies and governments assume that universal factors constrain female school attainment, costly policies and school interventions may continue to yield disappointing results. 15   In the larger Northern Transvaal province there are fewer than 3.0 residents for every one primary or secondary school student; this ratio is 3.7 in KwaZulu/Natal (calculated from data appearing in Verwey and Munzhedzi, 1994). Note that the Transvaal is now called Northern Province.

OCR for page 181
--> Acknowledgments The data analyzed in this chapter are from a household survey conducted in 1993-1994 by Francis Wilson and Pundy Pillay, University of Cape Town. Our work was supported by the World Bank with aid from the governments of Denmark, The Netherlands, and Norway. Special thanks go to Neeta Sirur at the World Bank for organizing the overall project. We are grateful to Yisgedullish Amde and Carlo del Ninno for their advice and essential help with data management. Caroline Bledsoe, Linda Chisholm, Barney Cohen, Duncan Thomas, and two anonymous reviewers provided very helpful comments on earlier drafts. References Astone, N., and S. McLanahan 1991 Family structure, parental practices, and high school completion. American Sociological Review 49:359-382. Becker, G. 1976 A Treatise on the Family. Cambridge: Harvard University Press. Becker, G., and N. Tomes 1976 Child endowments and the quantity and quality of children. Journal of Political Economy 84:S143-S247. Bledsoe, C., and B. Cohen, eds. 1993 Social Dynamics of Adolescent Fertility in Sub-Saharan Africa. Washington, D.C.: National Academy Press. Bozzoli, B., with M. Nkotsoe 1991 Women of Phokeng: Consciousness, Life Strategy, and Migrancy in South Africa, 19001983. Portsmouth, N.H.: Heineman. Bradshaw, Y., and B. Fuller 1996 Policy action and school demand in Kenya: When a strong state grows fragile. International Journal of Comparative Sociology 37:72-96. Caldwell, J. 1976 Toward a restatement of demographic transition theory. Population and Development Review 2(3&4):321-366. 1986 Routes to low mortality in poor countries. Population and Development Review 12(2):171-214. Caldwell, J., and P. Caldwell 1990 High fertility in sub-Saharan Africa. Scientific American 262(5):1 18-125. Campbell, R. 1983 Status attainment research: End of the beginning or the beginning of the end? Sociology of Education 56:47-62. Cochrane, S. 1979 Fertility and Education: What Do We Really Know? Baltimore, Md.: The Johns Hopkins University Press. Coleman, J. 1988 Social capital in the creation of human capital. American Journal of Sociology 94:S95-S120. 1990 Foundations of Social Theory. Cambridge, Mass.: Belknap Press.

OCR for page 181
--> Comaroff, J. 1985 Body of Power, Spirit of Resistance: The Culture and History of a South African People. Chicago, Ill.: University of Chicago Press. DiStefano, J. 1993 The Demand for Primary Schooling in Rural Ethiopia. Addis Ababa: United States Agency for International Development. Easterlin, R., and E.M. Crimmins 1985 The Fertility Revolution: A Supply-Demand Analysis. Chicago, Ill.: University of Chicago Press. Education Foundation 1993 Teacher qualifications and classroom effectiveness. Edusource Data News (December):5. Erasmus, J. 1994 A Human Development Profile. Halfway House, South Africa: Development Bank of Southern Africa. Fishlow, A. 1966 Levels of 19th century investment in education: Human capital formation or structural reinforcement? Journal of Economic History 26:418-436. Fuller, B., and P. Clarke 1994 Raising school effects while ignoring culture? Review of Educational Research 64:119157. Fuller, B., J. Singer, and M. Keiley 1995a Why do daughters leave school in southern Africa? Family economy and mothers' commitments. Social Forces 74:657-682. Fuller, B., P. Pillay, and N. Sirur 1995b Literacy Trends in South Africa: Expanding Education While Reinforcing Unequal Achievement? Cambridge, Mass., and Cape TowHarvard University and University of Cape Town. Glewwe, P. 1994 Student achievement and schooling choice in low-income countries: Evidence from Ghana. Journal of Human Resources 29:843-864. Haveman, R., and B. Wolfe 1995 The determinants of children's school attainment: A review of methods and findings. Journal of Economic Literature 33:1829-1878. Hill, M.A., and E. King 1993 Women's Education in Developing Countries: Barriers, Benefits, and Policies. Baltimore, Md.: The Johns Hopkins University Press. Hofmeyr, I. 1993 We Spend Our Years as a Tale That Is Told: Oral Historical Narrative in a South African Chiefdom. Johannesburg: University of Witwatersrand Press. Horan, P., and P. Hargis 1991 Children's work and schooling in the nineteenth-century family economy. American Sociological Review 56:583-596. Hyde, K. 1993 Sub-Saharan Africa. Pp. 100-135 in A. Hill and E. King, eds., Women's Education in Developing Countries: Barriers, Benefits, and Policies. Baltimore, Md.: The Johns Hopkins University Press. Isiugo-Abanihe, U. 1994 Demographic transition in the context of Africa's development. Pp. 61-73 in U. Himmelstrand, K. Kinyanjui, and E. Mburugu, eds., African Perspectives on Development. London: James Curry.

OCR for page 181
--> Kossoudji. S., and E. Mueller 1983 The economic and social status of female-headed households in rural Botswana. Economic Development and Cultural Change 31:825-838. Lesthaeghe. R. 1989 Reproduction and Social Organization in sub-Saharan Africa. Berkeley, Calif.: University of California Press. LeVine, R., S. LeVine, A. Richman, F.M. Tapia Uribe, and C. Sunderland Correa 1994a School and survival: The impact of maternal education on health and reproduction in the Third World. Pp. 303-338 in L. Chen, A. Kleinman, and N. Ware, eds., Health and Social Change in International Perspective. Boston, Mass.: Harvard University Press. LeVine, R., S., Dixon, A., Richman, P., Leiderman, C., Keefer, and B., Brazelton 1994b Child Care and Culture: Lessons from Africa. New York: Cambridge University Press. Lloyd, C.B., and A.K. Blanc 1996 Children's schooling in sub-Saharan Africa: The role of fathers, mothers, and others. Population and Development Review 22(2):265-298. Lockheed, M., and A. Verspoor 1991 Improving Primary Education in Developing Countries. New York: Oxford University Press. Loury, G. 1977 A dynamic theory of racial income differences. In P. Wallace and A. LeMund, eds., Women, Minorities, and Employment Discrimination. Lexington, Mass.: Lexington Books. Mburugu, E. 1994 The persistence of high fertility in Africa and prospects for fertility decline. Pp. 74-83 in U. Himmelstrand, K. Kinyanjui, and E. Mburugu, eds., African Perspectives on Development. London: James Curry. Odaga, A., and W. Heneveld 1995 Girls and Schools in Sub-Saharan Africa: From Analysis to Action. Africa Technical Paper Series 298. Washington D.C.: The World Bank. Ogbu, J. 1978 Minority Education and Caste: The American System in Cross-Cultural Perspective. New York: Academic Press. Project for Statistics on Living Standards and Development 1994 Preliminary Abstract. Cape Town: University of Cape Town, South African Development Research Unit. Reconstruction and Development Programme 1995 Key Indicators of Poverty in South Africa. Pretoria: Office of the President. Schneider, B., and J. Coleman 1993 Parents, Their Children, and Schools. Boulder, Colo.: Westview Press. Schultz, T.P. 1990 Testing the neoclassical model of family labor supply and fertility. Journal of Human Resources 25:599-635. Singer, J., and J. Willett 1993 It's about time: Using discrete time survival analysis to study duration and the timing of events. Journal of Educational Statistics 18:155-195. Truscott, K. 1994 Gender in Education. Johannesburg: University of Witwatersrand, Education Policy Unit. Tuma, N., and M. Hannan 1984 Social Dynamics: Models and Methods. New York: Academic Press.

OCR for page 181
--> United Nations 1995 Women's Education and Fertility Behavior: Recent Evidence from the Demographic and Health Surveys. Department of Economic and Social Information, Population Division. New York: United Nations. Verwey, C., and E. Munzhedzi 1994 Education in the Eastern Cape, Northern Transvaal, and KwaZulu/Natal (3 volumes). Halfway House, South Africa: Development Bank of Southern Africa. Walters, P., and D. James 1992 Schooling for some: Child labor and school enrollment of black and white children in the early twentieth century South. American Sociological Review 57:635-650. Walters, P., and P. O'Connell 1988 The family economy, work, and educational participation in the United States, 1890-1940. American Journal of Sociology 93:1116-1152. Whiting, B., and J. Whiting 1991 Adolescence in the preindustrial world. Pp. 814-892 in R. Lerner, A. Petersen, and J. Brooks-Gunn, eds., The Encyclopedia of Adolescence. New York: Garland. World Bank 1995 Priorities and Strategies for Education: A World Bank Review. Washington D.C.: The World Bank.