The diffusion of cost-effective life saving technologies has reduced infant mortality in much of the developing world. Income gains may also play a direct, protective role in ensuring child survival, although the empirical findings to date on this issue have been mixed. This paper assembles data from Demographic and Health Surveys (DHS) in 59 countries to analyze the relationship between changes in per capita GDP and infant mortality. The authors show that there is a strong, negative association between changes in per capita GDP and infant mortality- in a first-differenced specification the implied elasticity of infant mortality with respect to per capita GDP is approximately -0.56. In addition to this central result, two findings are noteworthy. First, although there is some evidence of changes in the composition of women giving birth during economic upturns and downturns, the observed changes in infant mortality are not a result of mothers with protective characteristics timing fertility to correspond with the business cycle. Second, the association between infant mortality and per capita GDP is particularly pronounced for periods of large contractions in GDP, suggesting the inability of developing country households or health systems (or both) to smooth resources. Simple back-of-the-envelope calculations using the estimates suggest that there may have been more than 1 million "excess" deaths in the developing world since 1980 as a result of large, negative contractions in per capita GDP.