Measuring the Propensity to Drink and Drive

Laws such as .08 blood alcohol content, open container, and license revocation provide a policy framework for reducing drinking and driving, yet policy effectiveness relies on their enforcement. The relationship between enforcement and drink-driving behavior is difficult to assess. Unlike property and violent crimes, where incidence statistics can approximate behavior, most drink-driving trips go undetected. We develop a novel measure of drink-driving propensity by estimating an item response model using national survey data. Aggregating to the state level, we find that much of the variation in enforcement is not in proportion to residents predilections