Highway fatality rates vary significantly among jurisdictions. Before attributing these variations to policy differences, it is important to account for exogenous factors (e.g. weather) that are beyond the control of policy makers. A new method is developed to identify states that are doing better or worse than expected. The method isapplied to the 48 contiguous American states using highway fatalitydata for the years 1975 to 1986. A fixed effects linear model is used to estimate the fatality rate for each state, taking into accountexogenous factors. These factors were: a) LETHPC14 (the natural logarithm of ethanol consumption per capita for those aged 14 or older); b) DRVLE24P (percentage of drivers under 24 years of age); c) UEGT16 (unemployment rate for those aged 16 or older); d) LPCI (natural logarithm per capita income); e) TEMP (annual average temperature); and f) PRECIP (annual total precipitation). Results indicate that many states have higher or lower fatality rates than is expected basedon a ranking of states according to crude fatality rates. Policy implications are discussed
Abstract