The US criminal justice system is commonly referred to as a revolving door due to its high rates of recidivism—the tendency of people convicted of crimes to face more criminal charges in the future. A key question for policymakers is whether the criminal justice system itself contributes to this pattern or whether it is driven by external factors such as addiction, mental health, neighborhood disadvantage, or limited labor market opportunities. Previous research on the role of contact with the criminal justice system has focused primarily on how incarceration affects recidivism. However, noncarceral conviction (i.e., a conviction that does not result in incarceration) is also a frequent outcome. In 2010, 2.7 people were on probation for every person who was incarcerated.

Noncarceral convictions could affect recidivism in several ways. For example, a conviction could make it harder to find employment, making a person more likely to resort to crime. A conviction could also make future contact with the criminal justice system more severe, even if it has no impact on criminal behavior. For example, prosecutors may be more likely to pursue charges against someone with a recent conviction, and judges may sentence them more harshly. Conversely, a conviction could deter crime if it increases the expected penalties for future crime.

Our research provides new evidence on how felony noncarceral conviction and incarceration affect recidivism. It uses a dataset of felony cases in Virginia between 2000 and 2020. Our research design leverages the fact that people are randomly assigned to a judge to hear their case and issue a sentence and that judges vary in their propensities to convict and incarcerate people.

Our research finds that noncarceral conviction led to large and long-lasting increases in recidivism (relative to dismissal). For example, noncarceral conviction increased a person’s likelihood of being charged again within one year of conviction by 66 percent and within seven years of conviction by 47 percent. In contrast, incarceration decreased recidivism in the first year (relative to noncarceral conviction), a time when most of the people we studied were incapacitated by incarceration. Our research finds no evidence that incarceration reduces recidivism after the incapacitation period is over.

Previous research has explored how to reduce felony convictions or their impact. Policymakers could reduce convictions by expanding felony diversion programs, decriminalizing certain offenses, or downgrading felony charges to misdemeanors. Alternatively, policymakers could reduce the impact of felony convictions by limiting the accessibility or permissible uses of criminal records. For instance, limiting how long criminal records are publicly available could mitigate the employment effects of having a criminal record, potentially reducing recidivism by increasing formal employment options. Likewise, reducing feedback loops within the penal system, such as automatic charge upgrades or sentence increases for those with a felony conviction, could mitigate the impact of a criminal record.

Our findings suggest that these policies could reduce the penal system’s revolving door problem. Of course, there may be valid reasons for using past felony convictions in hiring or sentencing decisions. However, given the prevalence of felony convictions in the United States—8 percent of all adults and 33 percent of adult black men are estimated to have a felony conviction—the impact of convictions on recidivism is an important part of the policy debate.

Note
This research brief is based on John Eric Humphries et al., “Conviction, Incarceration, and Recidivism: Understanding the Revolving Door,” National Bureau of Economic Research Working Paper no. 32894, August 2024.