Marijuana laws and pedestrian fatalities in the United States

Pedestrian fatality rates in the US began to increase in 2009, after three decades of decline. This increase is occurring at the same time policy makers are encouraging walking as a healthy and sustainable transportation mode. The increase in pedestrian fatalities is also concurrent with the spread of liberalization of marijuana use laws across the US. Has the liberalization of marijuana laws contributed to the increase in pedestrian fatalities?

Using a Poisson regression generalized difference in difference design, we measure the relationship between pedestrian fatalities and liberalization of marijuana laws. We employ randomization inference since persistent idiosyncratic shocks in the state time series may create spurious correlations that inflate type 1 error rates for conventional hypothesis tests. Consistent with the alcohol substitution hypothesis, we find both medical and recreational marijuana laws are followed by a statistically significant reduction in daytime fatalities involving alcohol. Both are also followed by a reduction in nighttime fatalities involving alcohol, but the declines are not statistically significant.