Most economic models consider sleeping as a pre-determined and homogeneous constraint on individuals' time allocation neglecting its potential effects on health and human capital. Several medical studies provide evidence of important associations between sleep deprivation and health outcomes suggesting a large impact on health care systems and individual productivity. Yet, there is little causal analysis of the effects of sleep duration. This paper uses a spatial regression discontinuity design to identify the effects of sleep on health status, weight, and cognitive abilities. Our results suggest that delaying morning work schedules and school start times may have non-negligible effects on health.