Physical activity is crucial for maintaining and improving health, especially at advanced ages. While retirement increases the amount of time available for physical activity, there is only limited evidence regarding the causal effect of retirement on recommended levels of physical activity. Addressing this gap in the literature, we use data from the U.S. Health and Retirement Study to estimate the causal impact of retirement on meeting the federal government's 2008 Physical Activity Guidelines for Americans. Using official early and normal retirement ages as instruments for retirement, our causal IV analyses suggest significant positive effects of retirement on meeting the Guidelines. These effects are robust with regard to the treatment of unobserved individual-specific heterogeneity, the measurement of guideline compliance, the definition of retirement and respondents' health insurance status. We also show that the effects of retirement on physical activity are larger for persons with higher levels of education and wealth.