This study examines the effects of retirement on lifestyle habits to determine the relationship between retirement and health. Looking at panel data from the Japanese Study of Aging and Retirement, fixed effects with time effects methods are used to account for the endogeneity of retirement. We then apply the fixed effects with time effects instrumental variable methods, after checking that the endogeneity of retirement is excluded by using the Durbin–Wu–Hausman test. Basic pension eligibility age is the instrumental variable. Our results reveal that elderly Japanese people reduce their drinking and increase their walking, frequency of heavy exercise, and sleep time on weekdays after retirement. However, sleep time on holidays and smoking and appear unchanged. Only elderly Japanese women increase their frequency of light exercise after retirement.