Crime Time: How Ambient Light Affect Criminal Activity

We study the effect of daylight on crime. We take advantages of Daylight Saving Time policy (DST) which imposes exogenous variations in daylight exposure at different hours of the day. We use a rich administrative database managed by the Chilean national police which is a very centralized agency and has collected detailed information regarding each crime incident in the last ten years. We find a significant 18% decrease in overall crime associated to the DST transition that increases in one hour the amount of sunlight for the 7-9pm period. Consistently we find a 17% increase in crime when DST transition sharply decreases the daylight exposure for the same period of the day. We find no significant effects at other moments of the day in any of these two different DST transitions. Our results are consistent under two strategies that rely on different identification assumptions: sharp regression discontinuity design and a diff-in-diff regression analysis. We also analyze heterogenous responses for different crime categories and our results suggest that most of the variation is driven by robbery which decreases 30% during the evening hours of the day. Finally we find no significant effect on crime due to the DST transition on daily crime.