Ban the Box and Statistical Discrimination

By Peter Van Doren of Cato.

"With 25 percent of the world’s prison population, the U.S. has the highest incarceration rate of any country in the world. Over 600,000 people are released from American prisons each year and, sadly, about two-thirds of them will be rearrested within three years. Creating opportunities for people released from prison to reintegrate into society has rightly become a key focus of criminal justice reformers.

In recent years, “Ban The Box” policies and legislation, which require companies to delay asking whether job applicants have a criminal record until later in the hiring process, have become a popular policy response to the reintegration problem. The reasoning in support of BTB is well-intentioned: make sure that employers evaluate candidates and their skills carefully rather than arbitrarily eliminating a whole class of applicants, ex-cons, at the early stages of the hiring process. But evaluations of BTB policies have concluded that the effect is to reduce minority employment rather than induce firms to engage in a costly search and interview effort.

Despite the negative academic evidence, BTB laws continue to have support. Last week, Maryland Governor Larry Hogan vetoed a bill that would have prohibited Maryland employers from asking applicants about their criminal history until the first in-person interview. Writing in the Washington Post, Representative David Trone contended that Gov. Hogan was wrong to veto the bill, and endorsed a bill sponsored by fellow Maryland Representative Elijah Cummings, which would ban the box for federal jobs (an Obama-era regulation has already banned the box for federal government jobs, but this bill would apply the rule to federal contractors as well). Part of Hogan’s justification for his veto is that the law would have imposed a new costly and time-consuming regulation on Maryland businesses, which is true. But more importantly, the unintended consequences of BTB rules mean both the Maryland bill and the Cummings legislation would hurt the intended beneficiaries.

In a working paper I reviewed in the fall 2016 issue of Regulation, Amanda Agan and Sonja Starr showed that BTB increases statistical discrimination. Unable to ask whether an applicant has a criminal record, employers unfortunately use race as a proxy leading to discrimination, particularly against young black men. The authors sent 15,000 fictitious online job applications to employers in New Jersey and New York City and found that before BTB white applicants received 7 percent more callbacks than black applicants. After BTB white applicants received 45 percent more callbacks.
Economists Jennifer Doleac and Benjamin Hansen found similar results when looking at individual-level data from the Current Population Survey and exploiting variation in when BTB laws were adopted by state and local jurisdictions. The authors found that young, low-skilled black men are 3.4 percentage points less likely to be employed after BTB laws take effect.

Along with race, BTB leads employers to rely on other factors to eliminate those who have been incarcerated. For example, research has shown that occupational licenses, which often require criminal background checks, offer a way for African-American men without a criminal record to signal their non-felony status. But the negative impacts of licensing as a barrier to entry outweigh these benefits. The best policy to assist people with criminal records to find employment is to eliminate licensure as well as BTB laws.

Many large corporations, including Walmart, Target, and Koch Industries, have already voluntarily banned the box. Such voluntary action has different effects than mandates. Voluntary adoption is consistent with firms engaging in extra effort to reintegrate ex-prisoners into the labor force. 

The intentions of BTB supporters are laudable. When adopted voluntarily by firms, BTB signals a commitment by the firm to take the employment of ex-convicts seriously. But when mandated by government BTB laws induce firms to engage in statistical discrimination that negatively affects the employment prospects of minorities."