Many have pointed to criminal records as a substantial barrier to employment that could exacerbate racial inequality in the United States. Recent research from UChicago economists shows that retroactively reducing felony convictions to misdemeanors does not, on average, change employment. In this paper, the authors test the possibility that policies that clear entire records—rather than simply reducing their severity— might improve defendants’ labor market outcomes.
To answer this question, the authors first examine defendants’ employment trajectories before and after they encounter the criminal legal system. The authors link criminal records from four jurisdictions (Maryland, New Jersey, Pennsylvania, and Bexar County, Texas) to tax data from the Internal Revenue Service, and show the following:
- There are large and persistent drops in employment at the time of both misdemeanor and felony charges. These patterns are observed for convictions and, perhaps more surprisingly, for non-convictions.
- A survey of firm hiring decision-makers supports the interpretation that these patterns at least partially reflect the negative impact of having a record. Hiring professionals report markedly reduced likelihood of hiring someone with a drug or theft charge, even if it resulted in a non-conviction.
The authors next study whether removing criminal records can reverse these patterns. They measure the impacts of three policies that limit the information about criminal history reported in employment background checks: the Fair Credit Reporting Act (FCRA), which prohibits reporting of criminal charges that did not lead to a conviction (mainly dismissals) after seven years for jobs that pay less than $75,000 a year; the Maryland Credit Reporting Law, which prohibits the reporting of convictions after seven years for jobs that pay less than $20,000 a year; and Pennsylvania’s Clean Slate Law of 2018, which legislated automated sealing of all non-convictions.
The authors detail their methodology for measuring the impact of each policy in their working paper; in sum, they use administrative tax records to compare individuals who had their records cleared to otherwise-comparable defendants. The authors find similar results across the three policies and jurisdictions that they study:
- There is little evidence that clearing criminal records from background checks (or removing non-convictions, in the case of the FCRA) improves labor market outcomes, on average.
- A notable exception is evidence that record remediation policies increase the rate of electronically mediated gig platform work, albeit from a very low base.
How is it possible that criminal records are associated with large drops in employment and that remediation policies aren’t effective at mitigating this harm? The authors hypothesize that criminal charges scar defendants’ labor market trajectories in a way that can be difficult to undo later, for instance by creating resume gaps, loss of experience, discouragement, and reduced search. Since there are fewer barriers to participating in platform gig work, the authors posit such work is less susceptible to scarring.
Importantly, criminal record remediation policies may have benefits that are not reflected in the earnings observed in tax records. Remediation of records could directly impact access to housing, civic engagement, quality of life, and other policy-relevant outcomes. At the same time, this research suggests that if reintegration and labor market participation are primary objectives, existing policies are not achieving these goals. These results may help explain why recent evaluations of Ban-the-Box policies failed to find improvements in labor market outcomes for individuals with records.