The Incidence, Determinants and Consequences of Female Headship in Rural Bangladesh∗

Using an unusually rich dataset from rural Bangladesh, this paper explores the incidence of female headship, the characteristics of female-headed households and the effects of female headship on children’s education. The empirical analysis takes into consideration the heterogeneity in the group of female-headed households, and the possible endogeneity of the female headship variable. The endogeneity of female headship is addressed in two ways. First, female headship is instrumented using widowhood status of mothers in the sample. Next, both female headship and widowhood are treated as endogenous variables and instrumented with information on the marriage market at the time of a woman’s first marriage. The results indicate that female headship has a negative effect on children’s education. Treating female headship as exogenous however, leads to the spurious result that female headship has a positive effect on children’s education. Since women almost everywhere are disadvantaged relative to men in their access to assets, credit, employment, and education, it is often suspected that female-headed households are more vulnerable to risk, economically less viable, and less able to invest in the health and education of their children (a review of these arguments is in Folbre, 1991; UNDP, 1995; United Nations, 1996; Buvinic and Gupta, 1997; World Bank, 2001). Surprisingly, robust empirical evidence supporting these claims is scarce. Though a variety of case studies from around the world have documented the disadvantages faced by female-headed households (Mencher and Okongwu, 1993; Kumari, 1989), results from empirical studies have been far from conclusive. For example, Buvinic and Gupta (1997) review 61 studies on headship and poverty and find female-headed households to be disproportionately represented among the poor in 38 cases. Quisumbing, Haddad and Pena (2001) however, find that in their sample of 10 developing countries, the relationship between female headship and poverty is strong in only two countries, namely Ghana and Bangladesh. The literature on the effects of female-headedness on children’s welfare is similarly inconclusive. Several researchers have observed that in the United States children who grow up in female-headed households experience lower educational and occupational attainment, and, for female children, higher risks of teenage pregnancy ∗I would like to thank Paul Schultz, Chris Udry, Joseph Altonji, Andrew Foster, the participants of Yale’s Development Lunch and the participants of Yale’s Trade and Development Workshop for valuable comments and feedback. I am particularly grateful to Paul Schultz for guidance in the empirical methods used in this paper, and to Andrew Foster for giving me access to the data.