Children at Risk: The Role of Family Structure in Latin America and West Africa

class of their parents is frequently used to index children's access to resources. Thus, it is usually argued that children from higher-income households will receive better health care and nutrition than those from poorer households. Nevertheless, the evidence on the link between household income and nu-trient intake is surprisingly ambiguous, with some studies finding a strong positive relationship (Alderman, 1990), others finding an extremely weak link (Wolfe and Behrman, 1983; Behrman and Deolalikar, 1987), and still others finding moderate correlation at very low levels of income but relatively small correlation at middle to high income levels (Ravalli on, 1990). Although this lack of a consistently strong relationship between income and nutrient intake may be attributed to a variety of issues related to study design, I argue that a more fundamental reason underlies some of these counter-intuitive results. I believe that these studies are misleading because they focus on total household income, while excluding considerations of intra-household resource allocation and sources of income. This issue is particularly important to development planning and dis-aster relief programs in developing countries. Increasingly, governments in these countries are moving away from generalized food subsidy-such as subsidizing the market price of specific food grains or instituting universal rationing programs-frequently toward a more targeted form of assistance such as nutritional supplementation, rationing for low-income families, or food stamp programs. In some cases, this trend toward targeted subsidy has been a part of the International Monetary Fund's recommended structural adjustment programs. Nevertheless, appropriately identifying groups most at risk and directing assistance in a form that would reach children is extremel