Going Off Parole: How the Elimination of Discretionary Prison Release Affects the Social Cost of Crime

In order to lengthen prison terms, many U.S. states have limited parole boards' traditional authority to grant early releases. I develop a framework in which the welfare effects of this reform depend on (1) the elasticity of future recidivism with respect to time in prison, (2) the accuracy of boards in conditioning release dates on recidivism risk, and (3) the extent to which such conditioning encourages inmates to reform. Using micro-data from Georgia and quasi-experimental variation arising from policy shocks and institutional features of its criminal justice system, I find that longer prison terms decrease recidivism, boards assign higher-risk inmates to longer terms, and inmates' investment in rehabilitative activities falls -- and their recidivism rises -- when boards' discretion is limited. Back-of-the-envelope calculations suggest that the benefits of parole (the ability to ration prison resources based on recidivism risk and the creation of incentives) outweigh the costs (lost incapacitation due to shorter prison terms).