The current research seeks to answer the question whether tough sentences decrease the probability of recidivism. The specific deterrence literature posits that increased sentence severity decreases the probability of recidivism. The results of previous studies on the incarceration decision and recidivism were mixed with some studies claiming that sentence length has no impact on recidivism and some claiming that sentence length increases the probability of recidivism. Relatively few past studies have focused exclusively on the impact of time served in prison on the recidivism of serious violent offenders. I use judge assignment as an instrumental variable to correct for omitted variable bias. It is found to be an exogenous variable that is not related to recidivism, but is related to predicting time served in prison. Using two-stage least squares regression, I find that longer sentence length increases the probability of recidivism. Implications for specific deterrence and labeling theory are discussed.