We study the results of a massive nationwide correspondence experiment sending more than 83,000 fictitious applications with randomized characteristics to geographically dispersed jobs posted by 108 of the largest U.S. employers. Distinctively Black names reduce the probability of employer contact by 2.1 percentage points relative to distinctively white names. The magnitude of this racial gap in contact rates differs substantially across firms, exhibiting a between-company standard deviation of 1.9 percentage points.

This is from the abstract of Patrick M. Kline, Evan K. Rose, and Christopher R. Walters, “Systemic Discrimination Among Large U.S. Employers,” Becker-Friedman Institute Working Paper No. 2021-94.

The authors focus solely on discrimination but don’t speculate about why that discrimination takes place. I wonder if they would consider the following hypothesis.

An employer puts out a job description and asks for potential employees to submit applications. Due to the high cost of knowing, without hiring and experiencing, whether a potential employee will be productive, employers look for low-cost proxy variables to help them make relatively quick decisions. One such variable is the first name of the applicant. The authors of the study point out, correctly, that people with a given first name such as Antwan, Kareem, and Tyrone (they list 19 such names in their Appendix B) have a high probability of being black. The authors find discrimination against applicants with such names. But why might an employer who cares only about productivity and how the employee gets along with other applicants discriminate against job applicants with what the authors call black names? Could it be because people with such names are more likely than the average black person to come from a home with only one parent, typically the mother, present? We know from other data that people brought up in single-parent households do less well in school and are more likely to engage in crime. Could this be what potential employers are trying to avoid?

The authors also list 19 last names that are more likely to be black names. Here would have been an interesting way to test my hypothesis. Given their large sample size, they could have done it. Pair “black” first names, not with “black” last names, but half and half. That is, pair half of the black first names with black last names and half of the black first names with “white” last names. Then see what difference the black first name makes. If my speculation is correct, then they would find more discrimination based on black first names than based on black last names.

For the article on discrimination in The Concise Encyclopedia of Economics, see Linda Gorman, “Discrimination.”