We estimate the causal effect of parole on recidivism by exploiting the random assignment of parole board members to hearings in Quebec (Canada). Parolees in Quebec undergo a thorough selection process and are often required to stay in halfway houses for the remainder of their sentences, where they are provided with substantial rehabilitation assistance. We find that parole, in this context, increases reincarceration but decreases new sentences. To explain these results, we decompose the effect of parole on incarceration time into i) the direct reduction of the incarceration spell due to parole, ii) technical violations of parole conditions leading to reincarceration, and iii) future sentences. The positive reincarceration effect arises from technical violations, resulting in parolees being reincarcerated for a period of time on average shorter than the additional time they would have served without parole. Consequently, parole increases rearrests but still decreases time spent in prison. We find that, despite parolees being incapacitated for a shorter period, parole decreases the probability of committing new crimes. Our findings suggest that well-designed and targeted parole policies can simultaneously reduce incarceration and crime.