Sectoral Differences in Statistical Discrimination: Evidence from Ban the Box’s Removal of Criminal History Information

Abstract

The growing number of Americans with criminal records has sparked the proliferation of Ban the Box (BTB) policies that prevent employers from asking about criminal history during initial job application stages. These policies may inadvertently generate statistical discrimination against certain populations, such as blacks and Hispanics, that are overrepresented in the criminal justice system, as the removal of criminal history information may lead employers to assume that members of these groups are more likely to have a conviction record. This paper uses data from the Panel Study of Income Dynamics and a triple-differences strategy to test whether the effects of public and private sector versions of BTB on minorities differ from the effects on whites, indicating the presence of statistical discrimination. Employment is the primary outcome variable I analyze, and I further study BTB’s effects on labor income and two other life outcomes, home ownership and having health insurance. I find that public and private BTB reduce the likelihood of employment for blacks relative to whites by 2.09% (1.55 percentage points) and 5.99% (4.71 percentage points), respectively, and these results statistically significantly differ at the 5% level. Additionally, I find that private BTB reduces both labor income and the likelihood of having health insurance for blacks relative to whites. I interpret these results as evidence of statistical discrimination against blacks as a result of both public and private BTB, with private BTB having greater effects than public BTB.