Comments

Mark Duggan: During the past two decades, the U.S. prison population increased by a factor of four, with 1.36 million individuals in state and federal prisons by the end of 2002. An additional 0.80 million were held in local jails, juvenile facilities, and other correctional institutions during that same year.⁴⁶ One consequence of this growth has been a steady rise in the number of individuals released from prison in each year. For example, approximately 550,000 individuals were released from state prisons in 1999, nearly half the number residing in state prisons at the end of the year. If, as previous studies suggest, these individuals are significantly more likely than the average person to commit crimes, it is plausible that the release of these prisoners has a substantial impact on area-level crime rates.

Despite the obvious importance of this issue, no previous study has tackled this question head on. Most research has instead explored whether recently released prisoners violate the terms of their parole and are returned to custody within a certain time after release.⁴⁷ While these studies have contributed a number of important findings, they cannot be used to estimate the effect of recently released prisoners on area-level crime rates for three reasons. First, individuals released from prison may commit a crime and yet not be caught by the authorities, which tends to understate their effect on crime. Or those released from prison may be targeted by authorities and arrested for crimes that they did not commit, thus to some extent overstating their contribution to crime rates. And finally, a person released from prison may commit a significant number of crimes before being caught. Thus the outcome variables used in previous studies are clearly an imperfect measure of total criminal activity among recently released prisoners.

In the current study, Steven Raphael and Michael Stoll address the research question outlined by estimating the contribution of recently released prisoners to state-level crime rates. To do this, the authors utilize a data set from the Bureau of Justice Statistics that includes the number of individuals released from prison and the number newly committed to prison in each state and in each year from 1977 to 1999.⁴⁸ Combining these data with crime data from the FBI, demographic information from the U.S. Census Bureau, and unemployment data from the Bureau of Labor Statistics, the authors investigate the hypothesis that crime increases in response to an increase in the number released from prison and declines when the number newly committed to prison increases. In essence, the authors are testing a reverse incapacitation effect in the first case and a pure incapacitation effect in the second case.

There are several important findings in this study. First, the authors convincingly demonstrate that years in which states release a relatively high number of prisoners are followed by significant increases in crime while the opposite is true for the number imprisoned for the first time. Additionally, the authors find that the magnitude of these two effects is not equal—a person who is imprisoned appears to reduce crime by more than a person released from prison increases it. This is consistent with a number of possible hypotheses that predict criminal tendencies decline following imprisonment, including a positive effect of prison programs on earnings potential or a negative effect of parole supervision on the propensity to commit crime. Raphael and Stoll also find that the impact of prison releases was substantially greater early in the period (1978-84), suggesting that the person released from prison today is much less criminally active than his or her counterpart from two decades ago when prison populations were much lower. Finally, the authors demonstrate that prison releases are not significantly related to changes in crime in those states with powerful parole boards, perhaps because these states can selectively release those individuals who are least likely to commit crimes in the future.

The...