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.
He cites a study of the Boston rental market as the latest example. Rental agents were less likely to show a particular apartment to a black applicant than to white applicant with identical qualifications.
I would point out that the Boston rental market is peculiar, in that it seems that the prospective renter must use an agent. In other cities, the owner is allowed to advertise the apartment and renters are allowed to respond directly to ads. My guess is that Boston is different because the rental agents were able to lobby for some legal requirements that are not present elsewhere.
When there is no agent in the picture, the incentive of a landlord or a home seller is to rent or sell to the highest bidder. If you exclude customers, based on race or any other factor, you risk leaving money on the table.
But agents have weird incentives. It works out that they want to complete a transaction with as little effort as possible. Maximizing traffic into the property is not the way to go, especially if we are talking about rent-controlled apartments that are scarce.
But why would an agent discriminate on the basis of skin color? The agent may have an instinct that the black person will not “feel comfortable” in a neighborhood, so that it is not worth spending time showing that person the apartment, given the alternative of showing a white person the apartment.
The study shows that renters with housing vouchers were actually more likely to see the apartment of choice if they were black. That might be because the real estate agent is more confident that a transaction will take place in the case of a housing voucher if the prospective renter is black. The agent figures that the black renter will not have as many options.
My hypothesis is that there would be less racial discrimination in housing and rental markets if agents were out of the picture. I suspect that some agents have some preconceptions that are effectively racist, and I doubt that this can be overcome as long as the incentives of agents are what they are.
I am pretty sure that if you don’t have rent control, rental agents won’t gain a foothold, and the rental market will operate without them. Landlords have an incentive not to discriminate, so my hypothesis would be that in markets without rent control an audit study will not show as much discrimination.
What about buying a home/ If you could design the real estate transaction process from scratch, you could make it as easy to buy and sell a home as it is to sell a used car. In doing so, you would reduce the need for real estate agents, and you might reduce racial discrimination.
Many a young techie has salivated over the prospect of solving this problem. But it is not a technology problem. It is a public choice problem. Uber was able to come out pretty well against the lobbyists for the taxicab industry. Airbnb was able to come out ok against the lobbyists for the hotel industry. The real estate lobby is a tougher nut to crack.
“But agents have weird incentives. It works out that they want to complete a transaction with as little effort as possible.”
This is probably wrong – they want to maximize their take home pay, which implies maximizing the number of closed transactions (or whatever their commissions are based on).
Landlords on the other hand are trying to maximize their risk adjusted rental income.
In base cases, the incentives are race neutral (holding everything else constant).
*both
AirBnB is constantly being accused of racial bias because the people renting through the platform (not agents, owners) are less likely to want black renters.
Likely, the owners believe black renters will be detrimental in some way relative to whites. We can claim that is racism, or perhaps race just captures a piece of real information that owners correctly perceive and act on.
Not wrong. Showing a place to 3 people and a quick close at, say, 5 or even 10% less, when it all drops down through the funnel, doesn’t mean very much loss of income. Many transactions at a slight discount beats fewer transactions at full price.
In unrelated news, how is your sister doing in her new job?
Please send her my best! Hoping that it’s been awesome so far!
Extra pay for working the plague tent comes in handy. It’s not a real tent anymore, though. They moved back inside.
I’ll tell her a stranger on the internet was asking about her. I am sure she will be thrilled.
I live in and serve on the Board of Directors of a 556-unit twin tower hi-rise condominium complex in Northern Virginia. Agents perform essential services for our community, both in sales and in leasing, in the latter role frequently serving as an interface between association management and governance bodies and tenants from foreign lands unfamiliar with American culture and language. I don’t know how our community could get along without agents. Their activity has helped shape the evolution of our community to the extent that any gathering of our residents resembles a meeting
the United Nations General Assembly.
Buyers must not only avoid black communities, they must avoid areas that will experience white flight. The government can enforce racist integration policies, but they have yet to figure out how to prevent white flight.
Massachusetts ended rent control in the mid-90s, FWIW. My theory is that agents exist because it allows landlords to transfer compliance risk (hey, I’m not discriminating, I used a licensed agent and took the first tenant they recommended) while still being picky in a market where problem tenants are hard to evict. There’s also the fact that outside of deep recessions, the landlord can usually get away with making the renter pay the broker fee, so why not get someone else to do all the footwork?
Maximizing the number of closed transactions implies completing transactions with as little effort as possible.
I have watched the real estate industry fail to disintermediate with fascination for decades. I have come to two conclusions.
The political power of Realtors dwarfs that of taxi companies or hotel chains. I’m not talking about the agent you deal with in buying a house, but the brokers they work for who are frequently also developers. They are deeply intertwined with government planners and zoning boards. We’re talking power on the level of the teachers’ union or the NRA.
Second, a real estate transaction is far more complex than buying a used car. One 2016 Camry is not that different from another. And the differences are finite. Houses vary dramatically. Even a Levittown after 20 years is no longer a row of cookie cutter houses. Each owner has individualized the original house to their needs.
And a real estate transaction is harder to complete than buying a used car. You can drive the car off the lot the same day. Typically it takes 60-90 days to close on a house. Lots of things can derail the deal. Realtors earn their money by keeping the train on the track and getting the deal closed.
Purchasing a house is the largest financial transaction most people make, at least until college tuition went out of control. It is a very emotional process. Realtors manage that. I’m convinced you’ve got a better shot at replacing investment banks’ M&A departments with software than you do realtors.
Nice comment. Thanks!
Always strikes me that if these glaring inefficiencies exist, far more study authors should be leaving academia to start businesses.
If women make 10% less…start a business that hires primarily women…you’ll have a dominant competitive advantage. If real estate agents are ignoring black customers (all else being equal) then start an agency that focuses on black clients.
I think it’s very telling that these “inefficiencies” are seemingly rarely exploited.
If it were possible to reduce the roundtrip cost on a home from say 10% to 5%, it would be a dramatically better world.
But if it were that easy, people would already have done it. Occasionally you get some platform that will cut you a break of 1-1.5% out of the selling agents fee, but rarely more.
Could it be that agents are better equipped to do more due diligence?
One nearby home was rented out by the landlord. After a couple of renters in a row disappeared in the middle of the night without paying rent, with one leaving several thousand dollars in damage, the landlord started using an agent. The quality of renter increased dramatically.
With another nearby home, I used to see homemade signs pop up between renters. It was always a crapshoot to see what would show up. After a few skipped town and a couple left severe damage, I noticed that the homemade signs do not appear in between. The renters have improved, but I’m not sure what mechanism the landlord is using to filter them.
My hunch is that it’s not racism.
“Rental agents were less likely to show a particular apartment to a black applicant than to white applicant with identical qualifications.”
This is not quite right and is the root of the analytical and methodological problem at the heart of all these studies. Indeed, since it is the same error made in every similar study, when Taylor says they are no longer anecdotes but now consistent data, he is just compounding the error. (#NotMyFavoriteBlogger).
There is a different between “identical qualifications” and “identical *applications*”
Let’s say I have two objects which can described with three variables.
If the third variable is important to the choice, but is correlated with the second variable, then I can perform a factor analysis and reduce my descriptions to two dimensions.
If I am comparing two of these objects which have the same value for the first variable, then when I prefer one particular value in the second variable, you could say I am discriminating on the basis of that value for that dimension.
But that’s not quite right. I could be completely indifferent to the value of the second dimension, but what I care about was the value in the third *unreported* dimension, which only correlates with the value in the second dimension. The objects weren’t “identical except for the value in variable two”, they were different in variable two *and* another important variable that was *omitted* from the characterization.
So let’s say I am buying golf balls. The first variable is color (there are those wacky orange and yellow balls out there), the second variable is “quality” (practice, regular, premium, etc.), and the third variable is “typical flight distance”. You could have a fourth variable of “price” too.
No surprise, the flight distance and price dimensions *correlate strongly* with the quality dimension. When you do a factor analysis, you can reduce the description down to just two dimensions.
And someone looks at your choices and sees you tend to pick (white, premium) over (white, practice) most of the time. “Ah, you are discriminating on the the labeling! You are judging balls by their box covers!”
No, you don’t care at all about the labeling or the box covers. You care about the flight distance. But the way the study compressed and oversimplified complex, messy reality into fewer discrete dimensions and choices forced your legitimate preference into being embedded in an illegitimate category.
If the study is going to do this, it is only ethical and rigorous to announce at the very start that “we are assuming that practice balls and premium balls are identical in every possible respect *except* for the other reported dimension of color.”
And then that study is not going to get published or even get started because that would be an incredibly stupid and embarrassing thing to say which could and would get shot down immediately by anyone who knew anything about golf.
#ThatsHowYouGotTheRoadToSociology
+1
Looking through the study cited by Taylor, I didn’t see anything to suggest that testers’ criminal records were matched or revealed to lessors.
In the employment discrimination literature we see studies showing that “ban the box” measures and other efforts to conceal criminal background adversely affect blacks without a criminal background as employers are unable to select based on clean criminal record.
A recent meta analysis of housing discrimination studies finds that the amount of discrimination has been declining over the last several decades, a period of time in which black incarceration rates has also declined. Related?
The meta analysis states “A significant part of the discriminatory behaviour can be attributed to missing information about the social status of applicants.” Many other factors could be at play and perhaps “matching “ does not get to such other relevant factors.
The net present value of anticipated discrimination related litigation might be such a factor. One wonders whether real estate agents steer black applicant renters to black lessors because they believe white lessors perceive a higher rate of such costs and are therefore less likely to rent to protected races.
The meta analysis:
Closed doors everywhere? A meta-analysis of field experiments on ethnic discrimination in rental housing markets
Katrin Auspurg, Andreas Schneck, Thomas Hinz
Journal of Ethnic and Migration Studies 45 (1), 95-114, 2019
Discrimination is long seen as a meaningful factor for ethnic inequalities on rental housing markets. Yet empirically, the extent of discrimination is still debatable. For the first time, this article provides a quantitative meta-analysis of field experiments (in person audits and correspondence tests) that were run over the last four decades in the United States, Canada and Europe (N = 71). Special focus is given to a possible inflation of effect sizes by publication bias; to time trends; and to evidence for statistical discrimination. Taken together, nearly all experiments document the occurrence of ethnic discrimination. Effect sizes are inflated by publication bias, but there is still substantial evidence left once the bias is removed. The analysis reveals a consistent decline in the extent of discrimination over time, from moderate levels of discrimination in the 1970s and 1980s, up to only small but still statistically significant levels in the 1990s and 2000s. A significant part of the discriminatory behaviour can be attributed to missing information about the social status of applicants, which supports theories on statistical discrimination. It is discussed how future research could move our knowledge on the underlying mechanisms forward.
Everyone gives lip service to studies being garbage, cherrypicked, backed by motivated reasoning, innumerate, sometimes actually lies or frauds, but everybody keeps obeying them.
It reminds me of the old joke about Catholics and Mormons.
Catholics think the Pope is infallible in theory, but ignore him in practice.
Mormons think the Prophet is fallible in theory, but slavishly obey him in practice.
We are all Mormons now.
I own multiple income properties in a city that does not require agents and this is not true in my experience because the process doesn’t involve bidding. A unit is listed at a given price and prospective tenants request to see it. The price is usually listed too high and needs to be reduced to the point where one will actually get calls. Would-be renters often ask for a discount but rarely do they volunteer a higher “bid”. Therefore the incentive is to screen for the highest quality tenant who will pay on time and require the minimum attending to.
To the extent that race may signal the quality of tenants, there is very much an incentive to discriminate and some unsavory rental property owners have told me that they do so.
It’s true that an agent’s incentive is to close quickly so it’s surprising that they discriminate. But then again, they are offering a product to their clients in the form of “quality tenants”, which again, has racial implications in the minds of many.
I’m usually suspicious of social science studies that find racism but my gut as a rental owner is that this one is on to something.
There are 2 agents, buyer and seller, who will only talk to each other, so the lobby is 2x as strong as others.
The problem is right here:
The qualifications are never identical.
“When there is no agent in the picture, the incentive of a landlord or a home seller is to rent or sell to the highest bidder. If you exclude customers, based on race or any other factor, you risk leaving money on the table” – To me, this captures the disconnect between economist and the real world. I can tell from several first hand experiences that landlords are willing to leave money on the table to satisfy their social preference.
Would a Trump-hating agent not show the good stuff to someone wearing a Make America Great Again hat?
Not necessarily. You want a low-risk client. I’d say a black person in a MAGA hat would win over one not wearing one. Much less likely to invite their gangster friends over and get the place shot up.
“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.”
To whom would it be useful?
As VanDerWeff would put it, this remark makes me feel less safe.
Does this study really show racism? Or was it designed that it would never show anything else? According to the study, black applicants were assigned a “race-associated” name which was used right from the start. Such names would likely trigger concerns in anyone not knowing more. I suppose it depends on one’s definition of racism. Race-associated names carry with them….associations. Is it racist to acknowledge associations? If so, is any project to abolish associations at all realistic? Or is it totalitarian?
Were black applicants as likely to ultimately rent the place as white applicants, all else being equal? We can probably speculate reasonably that the likelihood of someone moving in to an apartment they’re shown may be affected by their race and/or racial makeup of the community. They should look at this effect as a function of the move-in rate by race across neighborhoods. It would tell us something about whether this is statistical discrimination or not.
There is no rent control in Boston. There is no rental agent requirement in Boston. The convention is that rental agents are paid by the tenant.
Almost all the posters seem to misunderstand the study. The study recruited the testers. They did their best to match the pair comparisons by visible demographics, and ASSIGNED the testers similar characteristics like income, family size and credit score. Sorry to all the hypothetical dismissals by B. Reynolds, Handle, Edgar, Yancey Ward, Thucydides and Mark Z.
The thing that jumped out at me was that they only looked at the very lowest end of the rental market. The researchers limited listings to one-bedroom apartments with monthly rents under $1,563 per month and studio apartments with monthly rents under $1,378 month and excluded short-term rentals. That is roughly $1000 or more below average for apartments that size here.
It is humorous that you think your comment really torpedoes all the previous ones, Tom.
Seriously- go back and actually read Handle’s comment, and Edgar’s reply to it.
Right.
Handle took several hundred words to explain why its all very complicated and unfair.
Edgar wanted us to understand that a lack of criminal information about the testers tends to disproportionately hurt the black applicants , which sounds kind of racist, but what do I know.
And you thought it was important to claim “qualifications are never identical”, even when someone makes them up so they are.
Yes, I suppose it is unforgivably racist to observe that there are disparities in identity groups and criminal records. Racist is such a meaningless epithet that I could care less.
Per Wikipedia:
“The data from 2008 reveals that black Americans are over-represented in terms of arrests made in virtually all types of crime, with the exceptions of “driving under the influence”, “liquor laws” and hate crime. Overall, black Americans are arrested at 2.6 times the per-capita rate of all other Americans, and this ratio is even higher for murder (6.3 times) and robbery (8.1 times).”
When a lessor who would prefer to rent to someone without a criminal record is deprived of criminal record information, they may then simply substitute race as a proxy. This adversely impacts individuals without a criminal record who are in the higher crime rate identity group.
Researchers have found such a pattern with in employment and and “ban the box” measures outlawing criminal record questions on employment applications:
“We confirm that criminal records are a major barrier to employment: employers that asked about criminal records were 63% more likely to call applicants with no record. However, our results support the concern that BTB policies encourage racial discrimination: the black-white gap in callbacks grew dramatically at companies that removed the box after the policy went into effect. Before BTB, white applicants to employers with the box received 7% more callbacks than similar black applicants, but BTB increased this gap to 43%. We believe that the best interpretation of these results is that employers are relying on exaggerated impressions of real-world racial differences in felony conviction rates.”
https://scholar.google.com/scholar?hl=en&as_sdt=0%2C21&q=ban+the+box&oq=ban+the+bo#d=gs_qabs&u=%23p%3DZbZcIzGIwvcJ
The question is whether matched testing in housing creates a similar result. I can’t tell from the study Taylor links to how that is taken into account. I vaguely remember having to answer criminal record questions on rental applications many decades ago. However, I do know that some jurisdictions have in fact criminalized such questions on rental applications so I would presume that the practice exists or might exist if legalized.
Edgar- I was not claiming that you making your observation was racism. I was claiming that the behavior you were observing was racism.
“When a lessor who would prefer to rent to someone without a criminal record is deprived of criminal record information, they may then simply substitute race as a proxy.”
Substituting race as a proxy is a perfect definition of racism.
That being said, the testers weren’t real and didn’t actually follow through to a stage where they committed to renting these apartments. No one runs background checks at the stage of the rental process being studied here.
And, it should be pointed out that Boston does allow criminal background checks. Massachusetts housing providers may refuse to rent to persons with criminal records, although there are some qualified protections in the law.
Given all this, it was hard to see exactly why Yancey Ward thought your comments were
very useful.
Because they set out to find racism and it’s obvious that P(criminal or otherwise troublesome|black, can afford only the cheapest rents)>>P(criminal or otherwise troublesome|black, can afford expensive rents). In other words, in the higher-end rental market race provides much less of a clue to valuable but unobservable characteristics, and they wouldn’t get a strong result.
You do know we also have lots of criminal or otherwise troublesome white people here too, don’t you? We’re kind of famous for it. Hollywood loves to make movies about them.
Hollywood!? **HOLLYWOOD**?!?!
Hohohoho {wipes laughter-tears}. You have got to be kidding me with this stuff Tom. “Jump the Shark” does not even begin to describe it.
Are you seriously trying to argue that we can rely on Hollywood portrayals to accurately reflect the real state of affairs with regards to representing members of ethnic groups committing certain actions in the proper statistical proportion? You know, as opposed to going to absurd and extreme lengths for several generations in precisely the opposite? (See, e.g., Law and Order).
Good God man. Embarrassing.
“Are you seriously trying to argue that we can rely on Hollywood portrayals to accurately reflect the real state of affairs with regards to representing members of ethnic groups committing certain actions in the proper statistical proportion?”
Only you could write a question like that. And no. It’s a bit of a joke here that whenever they make a movie set in Boston, its usually about a criminal from Southie.
You do know that some numbers can be bigger than other numbers, don’t you?
As much as there is to criticize about the kind of stuff that gets published in journals these days (or retracted, or double-down-retracted), my experience is that they are *still* about an order of magnitude more serious and reliable and less likely to be a stinking pile of worthless cant with tables and charts, than these kind of polished and sharp-looking pdf “””studies””” put together by independent advocacy organizations and universities.
Trying to live an intellectually serious life these days in which one tests one’s cognitive mettle looking for subtle holes and flaws in arguments is like going to the gym and discovering that the weights increase by ounces and only go up to two pounds max. “What are these, dumbbells for ants? How am I supposed to get good exercise?”
I haven’t seen one of these in years which isn’t like shooting fish in a barrel, only publishable at all because they know everyone will pretend the floating, punctured remains are all still alive and swimming.
I offered a potential explanation, not a dismissal; you seemed to have totally missed my point. It had nothing to do with confounding variables or how well controlled the experiment was, but whether the variable of interest – race here – is predictive of a client’s likelihood of calling back. For example, if you’re selling an apartment in a predominately white neighborhood, you might find black applicants are less likely to buy because they put a premium on other black people living in the neighborhood, or vice versus with white applicants in a black neighborhood.
(1) Post June 2020, white flight, and consequently “racial discrimination” by landlords is about to increase greatly. Landlords who rent more than a small fraction of units in multi-unit developments to blacks will see many other tenants flee. Turnover is costly for landlords so they naturally don’t want to encourage it (falling rents due to reduced demand is even worse). What will Timothy Taylor say to all that?
(2) In recent but less agitated times, careful nationwide “paired testing” audit studies by the US Department of Housing and Urban Development revealed that “discrimination” against blacks was actually quite mild, even using study designs slanted to find “discrimination” by excluding highly relevant distinctions between testers in each pair from consideration. For one example, a huge 2012 study found that blacks were shown only 4.2% fewer rental units than whites and “told about” just 11.4% fewer other units. The study found worse discrimination against hispanics and asians than blacks. Of course testers were not really equally-qualified, for reasons explained much better by commenter Handle elsewhere on this page. As our host Dr. Kling explained, Boston may be an outlier for local structural reasons.
Speaking of Boston as an outlier for structural reasons, one wonders how much of the results were influenced by South Boston. As Christopher Caldwell writes in his book The Age of Entitlement about the school bussing protests:
“When the whites of South Boston and Charlestown protested, their neighborhoods were put under military occupation. “Southie” had a curfew and laws against public assembly, enforced by 1,600 police officers, 100 federal marshals, 50 FBI agents, and 600 National Guard troops.”
Seems like a powerful bit of unique history that would be culturally relevant.
This is your wish that landlords (and agents) were Homo-sapiens-onlyeconmicmaximicus; but that Titanic (including all rationalus, scientificus sub-species) ship comes across the actual Homo-sapiens-mygroupicus (myownrationalus/myscientificus) iceberg now and then but people still wish for an alternate reality.
Always rent or buy through an Asian agent. They definitely don’t sell or rent to blacks or near blacks. I rent in a large complex and 100 percent white and Asian. Not a single black. Malaysian Chinese-owned property. I imagine its company policy.