SEMINAR HAS BEEN CANCELLED

Peer Effects in Prison

Abstract

Peer interactions likely play an especially important role in the criminal sector due to its secretive nature and lack of formal institutions. Crucially, a large part of criminal peer exposure happens in prison, and is thus directly under the influence of policy makers and prison administrators. This paper

provides a more comprehensive understanding of these prison peer effects, shedding light on how to reduce recidivism and criminal network formation through changes in the composition of inmates. Our research design causally identifies peer effects in prison using rich Norwegian register data on over 150,000 prison spells and leveraging within-prison facility variation in peers over time. We produce several novel findings. First, exposure to more experienced co-inmates increases recidivism. Second, our results underscore the role of homophily and suggests network formation rather than skill acquisition as an important mechanism. Third, exposure to ``top criminals'' plays a distinctive role in shaping recidivism patterns. Fourth, we provide the first explicit documentation of criminal network formation among prison co-inmates.