The Effects of the Police on Crime: A Second Look

This paper is a replication of the analysis undertaken in 1978 by James Q. Wilson and Barbara Boland. We use time-series data for nine cities for portions of the period 1948-1978 to test the effect of aggressive policing on the robbery rate. Contrary to Wilson and Boland, we do not find moving violations to be a useful indicator of police aggressiveness. We also conclude that using the arrest/offense ratio introduces artifactual negative correlations. When one examines police expenditures, the size of the police force, the arrest rate, and the concentration of the police on robbery arrests, one generally finds positive relationships rather than negative ones. We interpret these results to indicate that a focus on the problem of robbery by the police leads the police both to make more arrests and to record more offenses, thus producing the positive correlations we found.