Friday, September 16, 2016

How Ban the Box Reduces Job Opportunities for African-Americans

Some job applications have a question which ask if you have a criminal history; if so, you are supposed to put a checkmark in a certain box. Given that African-Americans are statistically more likely to have a criminal history, it might seem obvious that this question tends to reduce job opportunities for African-Americans. But some evidence suggests that this intuition may be wrong; indeed, banning the box might actually reduce job opportunities for African-Americans. The study is called "Ban the Box, Criminal Records, and Statistical Discrimination: A Field Experiment," by Amanda Agan and Sonja Starr (Princeton University International Relations Section, Working Paper #5998, July 2016).  [I learned about a couple of additional studies from responses to this original post, which are now discussed briefly at the end.] are As Agan and Starr write at the start of the paper (citations omitted):
In an effort to reduce barriers to employment for people with criminal records, more than
100 jurisdictions and 23 states have passed “Ban-the-Box” (BTB) policies. Although the details vary, these policies all prohibit employers from asking about criminal history on the initial job application and in job interviews; employers may still conduct criminal
background checks, but only at or near the end of the employment process. Most BTB policies apply to public employers only, but seven states (including New Jersey) and a number of cities (including New York City) have now also extended these restrictions to private employers. These laws seek to increase employment opportunities for people with criminal records. They are often also presented as a strategy for reducing unemployment among black men, who in recent years have faced unemployment rates approximately double the national average ... Thus, a policy that increases the employment of people with records should disproportionately help minority men.
Agan and Starr carried out an experiment. They sent out about 15,000 fictitious online job applications to entry-level positions in New Jersey and New York city, both before and after the "ban-the-box" policy went into effect. The resumes were set up in pairs, so that they were largely the same resume except for a difference in race; in particular, out of each pair, one job applicant could be identified as  white and one as black. In addition, some of the pairs of hypothetical applicants checked "the box" early on, while others did not; some had a high school diploma, or a GED high-equivalency, or neither; some had a gap in their job  history, while others did not.

 The study found that whites with the same credentials are more likely to get a call-back than blacks: as they write, "white applicants overall received about 23% more callbacks compared to similar black applicants." Before "ban-the-box" went into effect, admitting to a criminal record definitely made it harder to be hired: that is, "among employers that asked about criminal convictions in the pre-period, the effect of having a felony conviction is also significant and large: applicants without a felony
conviction are 62% ... more likely to be called back than those with a conviction, averaged across races ..."

However, when ban-the-box (BTB) was enacted, the black-white gap in the chances of being called back got larger, not smaller. "Our estimates of BTB’s effects on callback rates imply that BTB substantially increases racial disparities in employer callbacks. We find that BTB expands the black-white gap by about 4 percentage points, multiplying the gap at affected businesses by a factor of about six. In our main specification, before BTB, white applicants to BTB-affected employers received 7% more callbacks than similar black applicants, but after BTB this gap grew to 45% ..."

The authors suggest that what economists call "statistical discrimination" is a possible explanation for these findings. The idea of statistical discrimination is that people might make decisions that have a discriminatory effect not out of animus against a particular group, but because they are using group membership as a marker for a higher probability of certain outcomes. Thus, it is a statistical fact that more blacks have a criminal history than whites. Consider an employer who is both mildly biased against blacks, but also would strongly prefer not to hire someone who has a criminal record. If that employer has information on whether someone has a criminal record, they will continue to be biased against blacks. But if this employer is banned from collecting information on criminal record, they will tend to act on the statistical knowledge that blacks are more likely to have a criminal record than whites. As a result, blacks without a criminal record will have a lower chance of a job callback, and whites with a criminal record will have a higher chance of a job callback.

Of course, one study with fictional resumes isn't the final word on any subject. One can concoct scenarios where even if ban-the-box means that blacks got fewer call-backs, perhaps this doesn't translate into fewer actual jobs.  But the evidence does suggest that advocates of ban-the-box should open their minds to the possibility that their good intentions about improving employment prospects for low-skilled black workers might in this case be leading to counterproductive results.

Addendum #1: Thanks to Catherine Rampell for pointing out to me that there's another recent empirical study of ban-the-box, different methods, but similar results. The study is "Does "Ban the Box" Help or Hurt Low-Skilled Workers? Statistical Discrimination and Employment Outcomes When Criminal Histories are Hidden," by Jennifer L. Doleac and Benjamin Hansen, published as NBER Working Paper No. 22469 (July 2016). (These working papers are not freely available online, but many readers will have access through a library subscription.) Instead of using fictional resumes, this study looks at variations in the details and timing of ban-the-box policies. They conclude:
We find that BTB policies decrease the probability of being employed by 3.4 percentage points (5.1%) for young, low-skilled black men, and by 2.3 percentage points (2.9%) for young, low-skilled Hispanic men. These findings support the hypothesis that when an applicant's criminal history is unavailable, employers statistically discriminate against demographic groups that are likely to have a criminal record.
Addendum #2: Thanks to Stan Veuger for pointing out yet another recent working paper on this subject, which uses a different approach and emphasizes a different set of tradeoffs. In "No Woman No Crime: Ban the Box, Employment, and Upskilling," Daniel Shoag and Stan Veuger look at employment with a focus on the outcome of ban-the-box an employment rates of those living in high-crime neighborhoods. They study the effects by looking at variations in employment rates that arise from the differences in timing of when cities, counties, and states adopt ban-the-box policies. They find:
"Using LEHD Origin-Destination Employment, a novel dataset on millions of job postings, and American Community Survey data, we show that these bans increased employment of residents in high-crime neighborhoods by as much as 4%. This effect can be seen both across and within Census tracts, in employment levels as well as in commuting patterns. The increases are particularly large in the public sector and in lower-wage jobs. At the same time, we establish that employers respond to Ban the Box measures by raising experience requirements. While black men benefit on net from these changes, a perhaps unintended consequence of them is that women, who are less likely to be convicted of crimes, see their employment opportunities reduced."