Research

Ten million people pass through U.S. jails each year, yet we know little about how the experience inside shapes what happens after. I study whether the offense composition of an individual's cellmates affects their post-release criminal behavior. Using daily cell rosters from multiple county jails linked to statewide criminal histories, I observe exactly who shares a cell with whom on every day of every incarceration spell. I estimate how an individual's recidivism depends on their exposure to peers with different criminal histories. I find significant cross type and same type effects on recidivism. These patterns are larger for certain types of offenders and are robust. A counterfactual simulation suggests that simple diversification of cell assignments could substantially reduce peer-driven recidivism.

We study the impact of reducing the cost of jail phone calls on jail time, case outcomes, and recidivism. Phone access plays a critical role in early case resolution, yet its high cost and surrounding administrative frictions often limit communication with legal counsel and loved ones. We exploit policy changes in several states that substantially reduced the price of jail calls, combined with detailed pricing data and individual-level jail booking records linked to state court disposition files. Our preliminary findings suggest that cheaper calls shorten jail spells, reduce guilty plea rates, and shift case outcomes toward dismissals — consistent with detainees gaining bargaining power through better access to outside support. We also find evidence that recidivism increases, likely reflecting reduced incapacitation. Welfare analysis suggests that the policy is net positive.