“Crime” on the Field

Does a greater police presence cause crime rates to decrease? To answer this question, we circumvent common simultaneity and reverse causality issues by exploiting an experiment from college football. In 2013, the Big XII added an eighth referee to conference games only. Unique features of this experiment allow us to identify the causal effects of increased “police force” (officiating crew) size on observed “crime” (penalty) rates, and to distinguish between alternative mechanisms behind those effects. Overall, we find that the increased police presence causes an increase in the observed crime rate. This increase is largest in teams' first treated games, suggesting a substantial detection effect. However, the penalty rate decreases sharply in teams' second treated games, suggesting a large general deterrent effect. Under mild assumptions, results suggest that the general deterrent effect is partly due to players learning to avoid detection, but that the learning process involves errors over time.