This study examines whether implementing Urban Residents Medical Insurance Scheme decreased an individual's risky lifestyle behavior before illness, termed ex-ante moral hazard. Ex-ante moral hazard is predicted by the classical economic theory suggesting that health insurance coverage reduces an individual's incentive to take preventive efforts to remain healthy. Studies have provided mixed evidence for this prediction. China's 2006 nationwide social experiment of implementing the Urban Residents Basic Medical Insurance Scheme offers an excellent opportunity for examining the effect of the transition from uninsured to insured on an individual's health behaviors. We exploit the longitudinal dimension of a representative survey data for 2007-2010 and employ the instrumental variable technique, thereby addressing the issue of self-selection into voluntary health insurance schemes. The results do not provide evidence for and contrast the prediction of the ex-ante moral hazard. Significant differences exist between insured and uninsured groups with respect to smoking, drinking habits, and being overweight. People with insurance care more about their health than people without insurance do. The main results still hold if we use alternative estimation methods and other robustness tests.