models where only years of education, but no quality measures, were included, returns to education declined for both blacks and whites in older cohorts. The interesting result from Welch's point of view was that the percentage returns to black education fell off appreciably faster compared to whites as the cohort aged. The implication is that blacks have been gaining with respect to whites as far as returns to education are concerned. Welch observed that this gain to blacks is likely to be due to a recent rise in the quality of education for blacks relative to whites. From a policy point of view, this leads to the conclusion that, however limited its role may be, education still is a potential policy tool with which to reduce observed racial earnings differences. The basic problem with Welch's results is that the observed decline in the returns to education for older blacks could have oc

Anne Case et al. (2002), using cross-sectional data, found a positive relationship between children's health and income, with income's protective effect increasing with age. Janet Currie and Mark Stabile (2003), using a panel of Canadian children, found that low- and high-SES children respond similarly to health shocks, but the low-SES children are subject to more shocks as they age. Our study examines this relationship using panel data for US children. We find some support for the latter result of Currie and Stabile, but also evidence that low- and high-SES children respond differently to specific health shocks.