US presidents tend to replace departing federal judges with successors of the same race and gender, and companies do the same with departing board directors. This tendency applies when women and minorities depart as much as when white men leave, find Harvard’s Edward Chang and Chicago Booth’s Erika Kirgios.

While eliminating bias is a significant challenge in hiring, “one-time actions that change the demographic composition of a group or organization might produce progress toward demographic diversity,” they write.

To test the degree to which people try to preserve the demographic constitution of a group—or perhaps attempt to increase diversity—the researchers analyzed two real-world scenarios and conducted four lab experiments. In looking at federal judge appointments between 1945 and 2020, they analyzed whether a departing judge’s race or gender helped predict that of the person nominated as a replacement. They looked at the same dynamic among board appointments between 2014 and 2019 at companies in the S&P 1500—about 5,600 appointments.

Chang and Kirgios controlled for variables including the party affiliation of the US president, the party of the US Senate at the time (since nominations must be approved by the Senate), and the demographics of the district in which the judge would be working. For the board appointments, they controlled for size, since larger boards have more opportunities to nominate new members, and for the demographic diversity of the current directors.

An example of ‘demographic stickiness’

The most significant predictor of a US federal judge’s race or gender was whether the judge’s predecessor had the same demographic identity, the research finds.  

Meanwhile, in their first two lab experiments, they asked subjects (600 in one case, 800 in another) to choose a replacement for a departing member of an already diverse six-person team at a management consultancy. Participants in the first study were randomly assigned to learn that the departing individual was white or Black; in the second study, they were either told a woman was departing or were not given the gender of the departing individual. Each participant of each study was then given the chance to choose one of three candidates as a replacement: a white man, a white woman, or a Black man—all with equal qualifications.

In the field, Chang and Kirgios find that a “demographic stickiness” effect was statistically significant for women, minorities, and white men. When a white, male judge left, a white man was appointed to replace him about 77 percent of the time. When the departing judge wasn’t a white man, the likelihood of a white, male judge getting the job was 45 percent. When a Black judge was being replaced, the new judge was Black 25 percent of the time, versus 7 percent when the judge wasn’t Black. Similarly, white men were about 10 percentage points more likely to be appointed to a corporate board after the departure of another white man, and Black directors were 9 percentage points more likely after a Black board member left.

This pattern extended to the lab, where “participants were significantly more likely to select a Black person when the departing group member was Black,” the study explains. When white men departed senior roles, participants were more likely to select a white man.

However, the final two lab experiments picked up on a disconnect between the lab and reality. One experiment tested the choices made around who should replace a departing white man versus a departing person whose gender and race were not known. The other looked at how participants chose candidates when replacing a white man or white woman versus expanding a team with new members. In these cases, the demographic stickiness effect disappeared for white men.

This, suggests Kirgios, might have been because white men are the presumed default in most white-collar, high-status settings, and therefore participants may have assumed that the unknown person leaving was a white man. Alternatively, the result might have indicated a form of virtue signaling—that is, the participants may merely have been trying to show the researchers that they embrace diversity, Chang and Kirgios propose. A third possibility: participants might have identified that the study was focused on discrimination and therefore tried to produce an answer either in line with or counter to a suspected hypothesis. Regardless of the reason, it signals a reality-lab divergence that she is probing in further research.

More from Chicago Booth Review

More from Chicago Booth

Your Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.