The short and long run effects of daylight saving time on fatal automobile crashes

Prior literature suggests that Daylight Saving Time (DST) can both increase the risk of automobile crashes in the short run and decrease the risk of automobile crashes in the long run. We use 28 years (1976-2003) of automobile crash data from the United States, and exploit a natural experiment arising from a 1986 federal law that changed the time when states switched to DST to identify the short run and long run effects of DST on automobile crashes. Our findings suggest that (1) DST has no significant detrimental effect on automobile crashes in the short run; (2) DST significantly reduces automobile crashes in the long run with a 8-11% fall in crashes involving pedestrians, and a 6-10% fall in crashes for vehicular occupants in the weeks after the spring shift to DST.