Beyond the obvious catastrophic effects of the HIV/AIDS pandemic on mortality, demographic changes, and the suffering of individuals and their families, we are still only learning about the complex longer-term effects of the pandemic on poverty and vulnerability. For example, the HIV/AIDS pandemic has substantially increased the number of widow-headed households in Africa. Many narratives and qualitative studies highlight gender inequalities in property rights and the difficulties that widows face in retaining access to land after the death of their husbands. HIV/AIDS has undoubtedly exacerbated such problems. However, to date, there is virtually no quantitative evidence on the proportion of widows who lose their land after the death of their husbands; whether they lose all or part of that land; and whether certain characteristics of the widow, her deceased husband, and/or her household influence the likelihood of her losing land rights. Because the number of widows is growing rapidly in areas hard hit by HIV/AIDS, there is an urgent need to understandthe magnitude of the problem and the degree to which it is exacerbating rural poverty.1 Such information may have important implications for poverty alleviation programs. If widow-headed households constitute a relatively large group whose ability to retain land is imperiled, then this would suggest the need for much greater attention to gender issues underlying local institutions and property rights as part of comprehensive poverty reduction programs. This study uses nationally representative panel data of 5,342 rural households in Zambia, surveyed in 2001 and 2004, to measure changes in landholding size among households experiencing the disease-related death of the male head of household between the two surveys.We estimate matched doubledifference average treatment effects models using households not experiencing any prime-age (ages 15-59) mortality during the survey interval as the control group. The longitudinal data used in the study are particularly well suited for understanding the factors associated with changes in widow-headed households' conditions, first, because of their nationally representative nature and, second, because the surveys include a rich set of information on individual kinship ties, the length of settlement of the household in the village, and other retrospective information not commonly collected in economic surveys. This social information provides an ability to examine whether changes in landholding size over time differ by initial household characteristics, attributes of the widow, social capital, and community characteristics such as matrilineal versus patrilineal inheritance institutions. The remainder of this article is organized as follows. Section II briefly describes land inheritance patterns in Zambia and perceptions of pressures for change. This is followed by a brief discussion of the conceptual framework and hypotheses to be tested in Section III. The data, sample attrition issues, and estimation methods are discussed in Section IV. Section V presents esti-mation results and their interpretation. Finally, Section VI discusses the conclusions and implications for donor and government policy. © 2011 by The University of Chicago.
