A new tool for investigating employer discrimination

Blacks earn less than whites and have higher unemployment rates.  This suggests that a critical mass of employers could be discriminating against Blacks, but how do you prove that?  

One way is to conduct research studies where employers are sent resumes of persons who are identical except for their names.  Studies have shown that there are some names where 90 percent or more of those with that name identify as Black (Darnell or Precious) or white (Brad or Claire).  If an employer tends to call back applicants at a vastly different rate based on the race-identification of their name, that would signal discrimination by race.  The same approach can be used for gender discrimination as well.  

A team of economists at Cal-Berkeley and Chicago used a brigade of undergraduate volunteers to apply to 108 of the largest employers in the US.  The odds of a callback were two percentage points lower for resumes with Black-sounding names as opposed to those with white-sounding names.  Overall the callback rate was around 25 percent, so this means that if 1000 Blacks applied they would get 240 callbacks and if 1000 whites applied they would get 240 callbacks.  This difference by itself is not likely to explain the gaps in wages and employment that we observe in the labor market, but it certainly can be a contributing factor.  

This finding was not uniform across all of the companies in the study; in fact it was concentrated among 23 of them.  The authors question whether these companies have intended to discriminate against Blacks; they may very well lack internal controls to ensure equal treatment.