Can police deter drunk driving?

Economic studies using aggregate data generally find that higher taxes are the most effective policy to reduce drunk driving while criminologists report strong evidence supporting law enforcement measures in policy evaluations. This paper evaluates these differing perspectives using the aggregate data that is typically used in the economic literature. OLS and fixed effects models show that police can affect the probability of arrest for drunk driving and, in combination with evidence from DUI deterrence experiments, this suggests that the failure of economic models to detect deterrence reflects the lack of systematic and sustained police efforts against DUI.