Short-Circuiting Crime: Evidence from a Post-incarceration Field Experiment

Many individuals who were previously incarcerated are re-arrested in the months and years following release from prison. We investigate whether encouragement to use post-incarceration support services reduces re-arrest. Voluntary participants in our field experiment are offered a monetary incentive to complete either three or five visits to a support service provider. Those in the incentives groups increased visits compared to the control group, but the majority do not complete the required visits to obtain the incentive. We find no significant intent-to-treat or local average treatment effects of the visits on average re-arrest rates. We show that this is due to heterogeneous treatment effects for different complier types rather than a uniform null effect. An increase in visits causally decreases re-arrest for one complier response, the type complying with the easier goal (3 visits), but it increases re-arrests for another, the types complying with the harder goal (5 visits). Encouragement of support service usage can be effective for some people, but not others. Our results highlight the importance of uncovering latent response types to treatments to properly evaluate the effectiveness of programs aimed to help the previously incarcerated.