Audit studies are one of the most persuasive ways to show real-world discrimination. The idea is to come up with pairs of potential applicants or buyers, where each pair is given background information that makes them essentially identical--except for a difference in race. Then these pairs of people enter into the economy by taking actions like renting an apartment, applying for a job or a loan, or buying a car. In some audit studies, like applying for certain job, an audit study can be done without actual people, just by constructing resumes and social media links, and seeing who gets a response at all and who gets invited for an actual interview. In these studies, the difference in race just involves using names or differences in interests that can act as a signal of race to the person (or the software?) looking at the job applications. In other audit studies, actual pairs of people are matched up.
For a just-published a group of researchers from Suffolk University Law School and the consulting firm Analysis Group have just published an audit study of rental housing in the Boston area ("Qualified Renters Need Not Apply
Race and Voucher Discrimination in the
Metro Boston Rental Housing Market," July 2020, funded by the Boston Foundation).
One additional factor in the rental markets is whether prospective renters say that they are planning to use a voucher for subsidized housing, or whether they were just planning to pay the full market rate. Thus, this study involved 200 "testers," who were divided into groups of four who were given similar characteristics and background: "Specifically, the test coordinator created matched pairs who were demographically similar (i.e., cisgender, same sex, no visible disabilities, age) and assigned the testers similar characteristics like income, family size, and credit score." The 50 potential rentals were randomly selected from public lists. The testers were also told to communicate with housing providers within the same fairly short time window, and to make the original communication in the same way (for example, by phone call or text). Within each group of four testers, two were black and two were white, and two said they were planning to pay full market rate while the other two were planning to use a housing voucher.
With this background:
The study measured a number of data points including:
- whether testers were able to make appointments to see the properties;
- how many units housing providers told testers about or showed them;
- whether housing providers offered financial incentives;
- whether housing providers made positive or negative comments about the housing units; and
- whether housing providers offered testers an application.
Here's a summary of some results:
Results indicate that White market-rate testers— meaning White testers not using vouchers—were able to arrange to view apartments 80% of the time. Similarly situated Black market-rate testers seeking to view the same apartments were only able to visit the property 48% of the time. Testers who had vouchers, regardless of their race, were prevented from viewing apartments at very high rates. White voucher holders were able to view rental apartments only 12% of the time. Black voucher holders were able to view apartments they were interested in renting only 18% of the time. ... In addition, housing providers showed White market-rate testers twice as many apartment units as Black market-rate testers, and provided them with better service as measured by a number of different variables. The results also showed that testers who were offered a site visit by the housing provider received differential treatment at the visit based on race and voucher status.
Of course, no social science methodology is truly bulletproof. One might wonder, for example, if the sample size is large enough, or the matched pairs of rental applicants were really the same, or if the problem of discrimination is especially severe in Boston rental markets. Because any single study can be questioned into exhaustion, one looks for other studies--and it turns out that these findings are broadly similar to the results of previous audit studies of housing markets done at other places and times.
For example, Newsday did its own audit study last fall looking at real estate agents. ""Newsday conducted 86 matching tests in areas stretching from the New York City line to the Hamptons and from Long Island Sound to the South Shore. Thirty-nine of the tests paired black and white testers, 31 matched Hispanic and white testers and 16 linked Asian and white testers." The testers interacted with 93 real estate agents in the Long Island area, and found pervasive evidence of racial bias. For a number of other studies, I've in the past discussed "Audit Studies and Housing Discrimination" (September 21, 2016).
There's an old saying in social science that data is not the plural of anecdote: that is, you can pile up a lot of anecdotes, but it's inevitably hard to know if such stories tell what happened, or what people think might have happened, or whether they reveal a common or a rare event. The growing body of audit studies in US housing markets is not a bunch of anecdotes: it's data showing that racial discrimination which is illegal under existing law is in fact disturbingly pervasive in US housing markets. I would love to see a wave of these audit studies of housing market discrimination carried out around the country, with loud publicity for the results and also with some legal consequences attached. It would be socially useful if rental agents and real estate agents needed to take seriously the possibility that the ways in which they are treating their minority customers could come under public scrutiny.
