Abstract

This paper seeks to identify the effect on post-release recidivism of low-risk night curfew inmates who were granted a collective pardon in Chile on July 1st, 2012. Exploiting the exogeneity of the pardon date as a natural experiment, it compares the recidivism rates of pardoned convicts with similar convicts who did not receive the pardon. The evidence shows that pardoned convicts are less likely to commit a new offense compared with the control group, suggesting a pardon prevention effect, probably induced by convicts’ “gratitude” for being released, or the latent threat of the pardon revocation because of a new offense. Furthermore, the evidence also suggests a non-significant specific deterrence effect of longer night incarceration. These findings support that a rational use of collective pardon might be an effective policy tool, which serves both to decongest prisons, and to prevent recidivism.