Control of drunken driving accidents in Norway: An econometric evaluation of behavior under uncertainty

This article makes use of a very simple choice-theoretic model and econometric analysis to evaluate the control of motoring accidents in Norway. The model attempts to take account of the impact of driving levels, traffic density, and road quality as well as alcohol consumption, the probability of convictions, and sanctions on the generation of accidents. Five alternative logit formulations, estimated within a simultaneous systems framework, provide insights on the impact of law enforcement on the levels of fatal and serious accidents. Results of this pooled cross-section-time-series analysis compare favorably with earlier time-series estimates of the strength of control effects.