Statistical Analysis of Area-wide Alcohol-related Driving Crashes: A Spatial Econometric Approach

Abstract

The article analyzes area-wide alcohol-related driving crash rates, with an emphasis on neighborhood effects, edge effects, and spatial effects arising from shared roadways that traverse area boundaries. Using township data for the state of Indiana, spatial Durbin models of alcohol-related driving crash rates are presented. The results suggest that a township's population composition and its abundance of alcohol-related businesses impact the alcohol-related driving crash rates. Moreover, positive spatial dependence is found to be highly significant, cautioning against the reliance on possibly biased OLS estimators. Due to restrictions in access to the crash and alcohol-related businesses data of the neighboring states, an alternative approach is adopted to address the spatial edge effects that may arise from alcohol consumption of residents from the edge areas of the neighboring states (possibly leading to crashes inside the border townships). Given the variation in alcohol laws and regulations across states, an empirical assessment of the edge effects is particularly relevant where some border crossing may be deliberate so as to avoid more stringent alcohol restrictions.