A substantial contributor to prison admissions is the return of individuals recently released from prison, which has come to be known as prison's “revolving door.” However, it is unclear whether being sentenced to prison itself has a causal effect on the probability of a subsequent return to prison or on criminal behavior. To examine the causal effect of being sentenced to prison on subsequent offending and reimprisonment, we leverage a natural experiment using the random assignment of judges with different propensities for sentencing offenders to prison. Drawing on data on all individuals sentenced for a felony in Michigan between 2003 and 2006, we compare individuals sentenced to prison to those sentenced to probation, taking into account sentence lengths and stratifying our analysis by race. Results show that being sentenced to prison rather than probation increases the probability of imprisonment in the first 3 years after release from prison by 18 percentage points among nonwhites and 19 percentage points among whites. Further results show that such effects are driven primarily by imprisonment for technical violations of community supervision rather than new felony convictions. This suggests that more stringent postprison parole supervision (relative to probation supervision) increases imprisonment through the detection and punishment of low-level offending or violation behavior. Such behavior would not otherwise result in imprisonment for someone who had not already been to prison or who was not on parole. These results demonstrate that the revolving door of prison is in part an effect of the nature of postprison supervision.