The relative income-health hypothesis postulates that income distribution is an important determinant of population health, but the age and sex patterns of this association are not well known. We tested the relative income-health hypothesis using panel data collected for 21 developed countries over 30 years. Net of trends in gross domestic product per head and unobserved period and country factors, income inequality measured by the Gini index is positively associated with the mortality of males and females at ages 1-14 and 15-49, and with the mortality of females at ages 65-89 albeit less strongly than for the younger age groups. These findings suggest that policies to decrease income inequality may improve health, especially that of children and young-to-middle-aged men and women. The mechanisms behind the income inequality-mortality association remain unknown and should be the focus of future research.