Occupant death: a study with directed graphs

We apply directed graphs to an empirical analysis of traffic occupant fatalities based on a model by Crandall. In this paper, we use Crandall's data on US traffic fatalities for the period 1947-81 and extend the sample to include 1982-93. Based on the 1947-81 annual data, directed graph algorithms reveal that occupant traffic deaths are directly caused by income, vehicle miles, and safety devices. Vehicle mileage is caused by income and rural driving. Estimation is conducted using three stage least squares regression. Results show a difference between the traditional regression methodology and causal graphical analysis. We also find that forecasts from the directed graph model outperform forecasts from the regression-based models, in terms of mean squared forecasts error.