2020-03-032020-02-26http://hdl.handle.net/10784/15901From the US to Colombia, from India to Uganda, many inmates suffer from underprovision of services such as surveillance or rehabilitation in overcrowded prisons. Yet, we know little about how prison quality affects long-term inmate outcomes. I study a prison construction program in Colombia and find t hat q uasi-random assignment of inmates to less crowded, and higher service facilities reduced recidivism. Criminal capital is an important mechanism. Less crowded and better service facilities are associated with a lower level of unsupervised criminal contact within prisons. The program led to substantial welfare gains, even when assuming a low social cost per crime.spaworkingPaperinfo:eu-repo/semantics/openAccesscrimeviolencerecidivismprison conditionsColombiaAcceso abierto2020-03-03D04H41J24K14K42Tobón Zapata, Santiago