How Does Retirement Impact Health Behaviors? An International Comparison

Recent work has found that retirement may lead to improvements in health, although the literature has not yet reached a consensus. This could be due to actual differences in the relationship of interest between countries or due to methodological differences between studies. The first goal of this paper is to estimate the causal impact of retirement on self-reported health using consistent estimation techniques on three harmonized longitudinal data sets, representative of the United States, England, and continental Europe. Using panel data and instrumental variable methods exploiting variation in statutory retirement ages, this paper then estimates how retirement causally affects health and health-related behaviors. We find, in all settings, retirement leads to better self-reported health, but that magnitude of the effect varies considerably. We also find that retirement increases the amount of exercise for those retiring from nonphysical jobs in all settings. The effect of retirement on addictive behaviors (drinking and smoking) was more mixed across settings. These findings suggest that public health interventions targeted to get near retirees to exercise more could allow countries to reap the benefits of a longer-working life while minimizing the associated health decline