This paper studies the effect of sunlight on crime, taking advantage of a daylight saving time (DST) policy. We find a 30% decrease in robberies when the DST transition increases the amount of sunlight by 1 h during the 7–9 p.m. period. We document a symmetric reaction when DST decreases sunlight exposure. We complement these findings by showing that the response induced by DST is not associated with a plausible demand-side response such as the population’s commuting patterns and we find no substantial short-term displacement of crime. Our results show that ambient light is a key driver of criminal activity in urban spaces.