De-policing and Crime: Evidence from Colombia

• In Event: Law and Politics in Contemporary Latin America

Abstract

Recent declines in discretionary police activity in U.S. cities have sparked a debate about the effect of “de-policing” on crime, especially on homicide rates. Existing results are mixed. We provide new evidence from the staged rollout of criminal justice reform in Colombia. The reform caused an immediate 30% decline in arrests, driven by even larger declines in arrests for minor offenses. This sharp drop in discretionary arrests did not affect Colombia’s homicide rate: using vital statistics, we estimate a null effect of the reform on lethal violence. Nor did it cause economically significant increases in vehicle theft (according to data that we obtained from insurers) or self-reported victimization (in survey data). These findings are hard to reconcile with models in which discretionary arrests deter or incapacitate felony offenders, suggesting instead that discretionary arrests can be socially inefficient.

Authors

• Dorothy Kronick, University of California, Berkeley

• Nicolas Idrobo, University of Pennsylvania