Physical activity likely improves health. Yet, even as health outcomes are positively associated with physical activity, confounding bias exists. We examine a large-scale government campaign to encourage physical activity among women in the United Kingdom as a source of an exogenous increase. Our difference-in-differences estimates confirm a large increase in activity associated with the campaign. We use this as the basis of IV estimation indicating that physical activity leads to a reduction in BMI and a lower chance of obesity, but we find no change in other health conditions over our four-year window. Implications for policy and future research are discussed.