We estimate the causal effect of prison conditions on recidivism rates by exploiting a discontinuity in the assignment of federal prisoners to security levels. Inmates housed in higher security levels are no less likely to recidivate than those housed in minimum security; if anything, our estimates suggest that harsher prison conditions lead to more post-release crime. Though small sample sizes limit the precision of our estimates, we argue that our findings may have important implications for prison policy, and that our methodology is likely to be applicable beyond the particular context we study
A similar study from Italy that looks explicitly at overcrowding is here. There is an earlier paper by Lawrence Katz, Steven Levitt, and Ellen Shusterovich that has the opposite finding but it relies on death rates in prison as a proxy for conditions and does not have a credible identification strategy. Much of this work is plagued by difficulties identifying causal effects. The Chen and Shapiro study stands out in that regard.
Recidivism is but one of many factors to consider but it featured in the Court’s opinion. The majority quoted governor Schwarzenegger:
“‘overcrowding causes harm to people and property, leads to inmate unrest and mis-conduct, . . . and increases recidivism as shown within this state and in others.’
In his dissent, Scalia criticized the District Court’s for relying “largely on their own beliefs about penology and recidivism.” There must be more relevant social science out there. Feel free to link to it in the comments.