Schooling and Parental Death

Loss of a parent is one of the most traumatic events a child can face. If loss of a parent reduces investments in children, it can also have long-lasting implications. This study uses parametric and semi-nonparametric matching techniques to estimate how one human capital investment, school enrollment, is affected by a parent's recent death. We analyze data from 600,000 households from Indonesia’s National Socioeconomic Survey (SUSENAS) during 1994-96. We find a parent's recent death has a large effect on a child's enrollment. We also use this shock to test several theories of intra-household allocation and find little differential treatment based on the gender of the child or the deceased parent. ______________________________ We have benefits from comments from Mark Gilkey and Daniel Levine, and able research assistance from Sebastian Martinez. Gertler and Ames gratefully acknowledge financial support from NICHD. This paper is forthcoming in the Review of Economics and Statistics. CIDER papers are produced by the Institute of International Studies and the Institute of Business and Economic Research. This paper can be found online at the new UC eScholarship Digital Repository site: http://repositories.cdlib.org/iber/cider with links to the CIDER publications page: http://www.haas.berkeley.edu/groups/iber/wps/ciderwp.htm 2 Loss of a parent while one is still young is likely to be the most traumatic event in a child’s life. If loss of a parent reduces investments in children’s human capital, it can also have long-lasting implications on their quality of life and livelihood. A significant proportion of school-aged children in less industrialized nations have always lost parents to accidents, childbirth, and illness. Unfortunately, the scourge of HIV/AIDS has greatly increased death rates of young adults in much of the world and, thus, increased the importance of understanding how parental loss affects investments in children. For example, one in ten African children under the age of 15 has lost one or both parents (Hunter and Williamson, 2000). In this paper we study how loss of a parent affects children’s school enrollment in Indonesia. The international community has become increasingly concerned about the effect of adult mortality on children’s schooling (Copson, 2002; World Bank, 1999). Many programs, especially in Africa, have been launched or proposed to support the school fees, uniforms and other schooling-related costs of orphaned children (Hunter and Williamson, 2000; Reid, 1993). However, the empirical evidence to support these policies is weak. Lloyd and Blanc (1996) use population surveys with limited socioeconomic controls from 7 African countries and find mixed results. Ainsworth et al. (2000) analyze a well-designed panel survey of 757 households from Northwestern Tanzania and find that adult mortality delays school entry, but otherwise does not affect enrollment. Indeed, there are conditions under which theoretical models predict no effect of parental death on enrollment (Becker and Tomes, 1979) One of the difficulties in studying the effects of parental mortality on children’s schooling is finding data sets large enough to capture sufficient cases of prime-age adult mortality linked to the socioeconomic status of the household. We analyze data from Indonesia’s National Socio-Economic Survey, known by its Indonesian acronym Susenas. The Susenas is an annually repeated cross section of approximately 200,000 households. We examine three years of Susenas surveys, from 1994 to 1996, and therefore have a sample of over six hundred thousand households. Susenas collects information on the general welfare of each household member, including school enrollment and mortality in the 12 months prior to the survey. Our bereaved sample (3,119 observations) includes individuals age 6 through 20, who live in households where either their mother or father died within the year prior to the survey.