My paper examines the public health consequences of the regulatory subsidy given to light trucks. The empirical challenge is to disentangle the causal effects of light trucks from the selection bias that may occur due to drivers sorting into different vehicle types depending on their unobservable characteristics. I address this by using state variation of snow depth as an instrumental variable for vehicle miles traveled of light trucks and cars. This instrument has strong first-stage explanatory power. Since snow depth is likely a direct determinant of crashes, I meet the exclusion criteria by restricting the dependent variable to those crashes that occurred in the summer. My findings suggest that, given a crash, light trucks are more dangerous to others but less dangerous for those driving them. However, I also find that light trucks are more likely to crash than cars, which neutralizes the safety advantage to those who drive them. My estimates for aggregate fatalities suggest that a world of light trucks leads to substantially more fatalities than a world of cars. (Author/publisher)
Abstract