MPIDR Working Paper 2004-016The impact of SES on mortality is an established fact. I examine if this impact decreases with increasing age. Most research finds that it does so but it is unknown whether this decrease is due to mortality selection. The data I use come from the US-Health and Retirement Study, which surveyed 9376 persons aged 59 and over from 1992 to 2000. The variables allow for a time varying measurement of SES, health and behavior. Eventhistory- analysis is applied to analyze differences in mortality rates. My results show that socioeconomic mortality differences are stable across ages whereas they clearly decline with decreasing health. My first finding, that health rather than age is the equalizer combined with the second finding, that good health itself is unequally distributed, leads to the conclusion that in old age, the impact of SES is transferred to the health status and hence it is stable across ages.