Effects of daylight saving time on accidents in Brazil: A Discontinuous Regression Approach

Abstract

Daylight saving time is a policy to reduce energy consumption instituted in Brazil each year from the second week of October. This study aims to evaluate the possible effects of DST policy on accidents in Brazilian federal highways. The data of accidents of the federal highway police used cover the period 2007-2013, aggregated by day and time for each Brazilian state. The short-term effects were first appreciated on the number of accidents in the summer time input transition in states that underwent the intervention, through the regression discontinuity model. The transition to daylight saving time, about the first twenty-four days, reduce the risk of accidents on federal highways by reallocating an hour of morning light for the afternoon. Soon after it was measured the duration of the effect of the same long-term policy in all states through the differences in differences model. Both models have shown that the application of DST policy presented evidence in the reduction of the number of accidents in Brazilian federal highways. To complement the analysis, robustness tests were conducted which determined the effects of previous trends, states with higher risk of accidents, decomposition hours light and final period of daylight saving time. The results showed that the States where the DST policy was used the accidents were reduced by 10 %.