We use data from the Wisconsin Longitudinal Study (WLS) to consider the distal and proximal correlates of happiness and eudaimonia in later life. Even after controlling for proximal covariates, outcomes at age 18 (IQ score, parental income and parental education) remain good predictors of well-being over 50 years later. In terms of the proximal covariates, mental health and social participation are the strongest predictors of well-being. Although some factors are important in explaining both happiness and eudaimonia, there are notable differences between the two measures: well-being policy will thus depend to an extent on which measure is preferred.