How much of racial discrimination stems from the customers?

Here is some reporting on a new paper by John Nunley, Adam Pugh, Nicholas Romero, and Richard Seals, here is one bit:

Black applicants faced major discrimination when applying for jobs with a customer focus. Researchers looked for jobs with words like “customer,” “sales,” “advisor,” “representative,” “agent,” and “loan officer” in the description. For jobs such as these, the discrimination gap soared. Instead of facing a 2.8 percentage-point gap between callback rates for whites and blacks, they faced a 4.4-point gap.

For jobs with descriptions that lacked those terms and were instead focused on interaction with coworkers, the level of discrimination collapsed. Descriptions with terms such as “manager,” “administrator,” “coordinator,” “operations,” and so forth, the difference in callback rates was 0.1 to 0.3 percentage points.

In other words, the problem isn’t that Joe Smith doesn’t want to hire young African-Americans, but that he is worried that if he hires a black sales associate, old Mrs. Jones may take her business elsewhere.

You will find the paper, and related work, here.  By the way, the discrimination effect was weakest in the cities of Baltimore and Portland.

For the pointer I thank the esteemed Samir Varma.


The job performance of blacks is consistently poorer than that of whites, so larger racial differences in callback rates indicate greater racial fairness, not greater racism:

You use of the word "consistently" is misleading, the link you provide only speaks about a consistent trend, not a consistently poorer performance. What this means is that individual differences still dominate race differences. Which means customers and employers can select individuals, which is not only more economical but also fairer.

In addition, it's unlikely such an effect explains the full difference. Racism and xenophobic prejudices are real, and they are unjust to individuals of any race who do a good job individually.

Yes, of course there are individual differences within races and the distributions overlap between races. The point is that blacks are worse employees, on the average, than whites, whereas the discrimination hypothesis predicts the opposite. I'm sure that there are employers who discriminate against blacks, but their negative effect on blacks is overwhelmed by the fact that a lot more employers are biased in favor of blacks.

"Worse" employees at what sort of work and in which sort of situations?

I don't have the two minutes it takes to make a cursory examination of every individual I come across. Heuristics FTW!

All sorts, Art Deco. Read the meta-analysis:,%20Huffcutt,%20&%20Bobko%20(2003)%20JAP.pdf

C'mon, P. That would only make sense if employers systemically abjured common performance criteria in hiring. They do that in higher education, not your local grocery store. Your whole like of argument is malicious.

Apparently worse in customer service roles, worse meaning results. While the hypothesis that the problem could be racist customers, Im disappointed that research still lacks the integrity to hypothesize that, for whatever reason, blacks do poorly in providing customer service. I think it is obvious that any deficit is genetic, so a worthwhile examination should look at interpersonal skills, trust forming, communication, education, and negative spillover from endemic criminal behavior.

I think it is obvious that any deficit is genetic,

This is a put-on, right?


blacks do poorly in providing customer service.

Some people have deficient people skills and follow-up, some people do not. If you've got adequate screens, you can sort your applicants. You'll make some errors one way and another. Unless you're positing that there's some sort of occult property that black applicants have which make the employer's error rate in hiring more severe, your posited scenario is nonsense.

Malicious? Nope, just the facts. Employers do not use common performance criteria when hiring applicants of different races. See affirmative action, disparate impact. Moreover, even if the same criteria are used, they are often biased in favor of black applicants versus whites (for example, all cognitively loaded measures are biased in favor of lower scoring groups unless the scores are corrected for unreliability).

@Art Deco

Oh, Good Lord! Not a put on, a brain cramp while typing.

It is obviously NOT genetic.

@Art Deco

Ah, I see your point! Yes, you are correct.

This experiment consisted of sending fake resumes with randomly chosen names that sounded either "typically white" or "typically black". There was no difference in qualifications on average between black and white candidates, by design.

If you have a lower-class name, employers can infer things about you. The Fryer study cited below showed that blacks with made-up black names do not have worse career outcomes than blacks with "white" names once you control for class background. These studies should use lower-class-sounding names in white resumes, too.

This holds even within some companies. In the restaurant business in my town, a majority of cooks are black, but there are very few black waitstaff. I'm not sure if self-selection has anything to do with this. From a cultural standpoint, I could easily understand why blacks might be loathe to wait on people.

"I could easily understand why blacks might be loathe to wait on people."

Most black waiters are clearly effeminate, while that's not true of most Mexican waiters.

Somebody should do an economic study of gay black men. I suspect they do pretty well in terms of income; they are more likely to take on jobs that straight black men just won't do.

Most black waiters are clearly effeminate, while that’s not true of most Mexican waiters.


Well said, Simon Cowell: "It's rubbish!"

I very rarely see black waiters of any stripe, but Steve's observation is actually true of my (limited) experieces dining in California.

And you lose one point off your IQ every year you stay in California.

"Most black waiters are clearly effeminate, while that’s not true of most Mexican waiters."

Even if that were true, all it would mean is that black waiters behave that way in order to get higher tips.

That is a good observation. There was a time when ALL waitstaff in some areas were black. And in the kitchen they have been replaced by Mexicans.

I wonder if the fact that the customer service jobs are typically lower level than the other jobs has anything to do with it. Just looking at the list, I'd think they would be different in terms of salary, ed requirements, etc.

Well it certainly didn't take long for the rationalizing racist commenters of MR to come forth and grace us with their thoughts.

The sanctimony was right on schedule too.

...aaaaaand, the snarky response. We should do this every day!

Time for a cigarette

It's funny that people who are not racist, i.e. don't discriminate people solely on what race they were born into, are now "sanctimonious".

As though, somehow, racists weren't evil and antisocial.

It's this great new trend that every sociopathic idiot declares himself a moral right to amoralism.

Not gonna happen. But we both know you'll keep whining about it.

What is really amusing is that people who are obviously holier-than-thou, sanctimonious, religious zealots and who have a blind faith in the behavioural, and intellectual equality of all races, not only can‘t recognise themselves as such but often think of themselves as irreligious freethinkers. It is going to be amusing to see their reactions to the coming deluge of genetic data.

Increasingly, the academic paper-citing, fact-spouting, calm and rational 'racists' are overtaking you sanctimonious puritans, and you won't like it when you are completely irrelevant except as street preachers condemning 'darned racists' between drooling bites of donated food.

Thanks for sparing me from the urgent need to respond.

Oof, time for another ciggy

if only liberal sanctimony made black people smarter, we'd have menthol e-cigs by now

Is it racial discrimination if an employer tries, consciously or not, to match its employees and the customers with whom the employees interact, match not only the profile of the customers but the profile the customers expect? Why so many Asians working in Apple stores? Is it because Asians like working in Apple stores, because Apple has so many Asian customers, or because the customers believe, consciously or not, that Asians know more about the products Apple sells.

If the profile is explicitly racial? Could well be illegal. Give Title VII of the Civil Rights Act of 1964 a read.

Although I am a lawyer, I try to restrict my comments to economics, including behavioral economics, since this blog is not a legal blog. Of course, I'd advise my clients not to profile, regardless of the client's good faith intentions. What I described in my comment is discrimination that favors Asians but inherently disfavors everybody else, including African Americans. Many fine dining restaurants in the South favor African American waiters, which inherently disfavors everybody else. As I like to say to clients: it's complicated. Or in the case of firing an employee: you can fire an employee for no reason (we are an at will state) but not for many reasons.

I'm not a lawyer, but isn't the line whether or not they are discriminating against certain specifically because they are not Asian, rather than some other proxy (e.g. they prefer tech savvy staff and most of the tech savvy applicants happen to be Asian)? If they are choosing certain applicants because of their race, that would seem to clearly run afoul of the law. Proving it would of course be very difficult, unless the business owner is a total idiot.

Chevron burden shifting.

A complainant must establish a prima facie case of discrimination by disparate treatment or disparate impact. They can do this with statistical information.

Once they eatablish the prima facie allegation, the burden then shifts to the employer to demonstrate that the disparity was caused by some factor other than race.

Once they do that, the burden falls on the charging party to show that the legitimate business reason is, in fact, pretextual or false.

You dont have to profile to end up with racial disparity, but the presence of a disparity establishes a prima facie case - the first prong of the burden shifting analysis.

A sophisticated employer, such as Apple, will conduct periodic barrier analyses. One purpose of this analysis is to identify legitimate business practices that, nonetheless, have a discriminatory impact. Second, it helps establish its affirmative defenses.

Yes, explicit racial profiles are almost always illegal.

I say almost always because, for example, a motion picture company might hire persons of only one race because the film requires that race of characters. Nothing prevents a business from having a foreign language requirement if it is a business necessity. But if challenged, they will have to prove their business necessity and defend it as non-pretextual.

Is it racial discrimination if an employer tries, consciously or not, to match its employees and the customers with whom the employees interact

No, anybody would do that except Government; see the police force of Ferguson MO. Only government could ignore efficacy that much and remain in business.

News flash. The range of manners, modes of speech, and non-verbal social signals differs from one side of the color bar to another. That affects performance in jobs which require client contact. It's disagreeable for aspirant sales reps in the black population, who may have satisfactory people skills for most purposes but do not quite fit with the market they're seeking to work with.

When we travel or vacation, I make a point of looking online at the demographics, unemployment, etc. at the place we visit then, while I'm there at the resort, hotel, visiting restaurants, or observing publics service workers, If there is a great disparity at the hotel, I write the management and have made the observation in TripAdvisor that there is this imbalance at the hotel or restaurant.

Thanks for this service. This reporting lets me arbitrage towards places that operate efficiently and away from places who piously signal egalitarianism to the detriment of their customers.

He should start a service called Bill's List, and you can start a service called AC's List that simply inverts Bill's ratings.

So, black names are rejected 16% more for "sales" but called back at the same rate for "manager".

It's just a conjecture, I may be wrong but it's really uncommon to see black sales people in any shop related to luxury. This may account for the 16% difference in callback rate. When people become comfortable of receiving sales advice for jewelry, fashion or a boat from black people, this will end. Cultural changes take time, the endorsement of luxury brands by famous blacks today will allow other blacks in the future to get these jobs.

In the end, no difference for managerial positions shows a lot of evolution compared to a few decades ago.

'To signal ethnicity, the researchers gave half the candidates the “typically white” names of Cody Baker, Jake Kelly, Claire Kruger, and Amy Rasmussen. They gave the other half “typically black” names of DeShawn Jefferson, DeAndre Washington, Ebony Booker, and Aaliyah Jackson.'

I would not be surprised if black men named DeShawn differ from black men named Robert in ways that matter to employers.
Black parents who name their boys Robert instead of DeShawn are deciding that their fitting in and getting ahead is more important than racial pride. Employers will prefer to hire the children of parents with the former mentality. Here is a relevant study:

The Causes and Consequences of Distinctively Black Names
Roland G. Fryer Jr. and Steven D. Levitt
+ Author Affiliations
Harvard University Society of Fellows and NBER
Department of Economics, University of Chicago
In the 1960s Blacks and Whites chose relatively similar first names for their children. Over a short period of time in the early 1970s, that pattern changed dramatically with most Blacks (particularly those living in racially isolated neighborhoods) adopting increasingly distinctive names, but a subset of Blacks actually moving toward more assimilating names. The patterns in the data appear most consistent with a model in which the rise of the Black Power movement influenced how Blacks perceived their identities. Among Blacks born in the last two decades, names provide a strong signal of socioeconomic status, which was not previously the case. We find, however, no negative relationship between having a distinctively Black name and later life outcomes after controlling for a child's circumstances at birth.

This is a cultural signalling issue. MLK won the civil rights battle, but Malcom X won the battle for black culture. Black America became convinced that equality meant exacerbating differences, ebonics, rejecting assimilation, demonizing assimilated blacks as "Uncle Toms" and handicapping their children with stereotypically underclass names. Much is made of "black" names, but I'd like to see a comparative correlation with stereotypically low-class white names. How does Thurston Powell III compete against RayShawn Burris and Jebediah Hinkel Willis? The Freakonomics lads did some early work on "upper class" versus "lower class" naming. I'd be interested in further research.

Malcolm X 'won the battle for black culture'? There are 40 million blacks in the United States and W.D. Mohammed's denomination has a membership in the low six digits; Louis Farrakhan's is in five digits.

As for 'ebonics', its a vernacular dialect. Blacks switch back and forth between standard English and dialect throughout the day, sometimes addressing one co-worker in one and then turning around and addressing another co-worker in another. It's no bar to acquiring trade skills.

It was not done fifty years ago to give your child mock-up names. Blacks favored ordinary English names, Biblical names ("Obadiah"), Southern familiars ("Hattie"), or off-beat but recognizably Indo-European names ("Monique"). The first person I ever met with a synthetic name carried around the moniker "DeMorge". "DeMorge" was born in 1966 or thereabouts. (Guy works in Hollywood, pretty consistently from what I can tell in modest acting and writing gigs). These sort of ugly pseudo-African names are bad taste and likely do cost some people interviews. Not sure they're an indicator of separatism or defiance in any serious way. I've had co-workers who were perfectly functional and reasonable who stuck their daughter with appellations like "Lakeisha".

Art Deco I don't know how much time you've spent in the south, but in the South especially Atlanta Ebonics or its equivalent is very much used even when addressing customers. There certainly are different levels of slang used but the general Ebonics grammar and phonics is simply how some blacks talk. Maybe this is a product of the Black Redneck/ White Redneck system that Sowell claims to have identified but there are definitely times especially late at night at a fast food restaurants were the clerk and I have been reduced basically to having rspeat each thing we say to each other for comprehension.

There certainly are different levels of slang used but the general Ebonics grammar and phonics is simply how some blacks talk.

It's how most blacks talk to each other. No one's speech is perfectly grammatical and ungrammatical speech by blacks is likely to have elements of dialect in it. If you think your food service employees should sound like C.S. Lewis, you're bound to be disappointed.

I always wonder why these studies use white names like Cody and Claire and Brandon. Why dont they use Cletus and Mary Sue and Candi to be more similar to the socioeconomic status of the typical black names they use.

I think this is an excellent point. The study may well have detected a difference in socio-economic class versus race. But if they used the same set of names, why wasn't there a large difference among managers?

Excellent point.

By the way, my vague impression is that some middle class black families try to pick Christian names for their sons beginning with "D" but without any apostrophes or pseudo-Frenchisms, to show they are both black and not ghetto: e.g., Darren.

While you nailed it, I'd just point out that as someone born in 1971 everything went wrong with America. Really, that damn moon landing was the peak.

They could have simply included pictures of the fictitious applicants with the application. They probably used the name method to maximize the discrimination they would observe.

The article implies the study found identical bias in Atlanta, with a 63% African American population as was found in Portland, which is less than 7%. Either Fortune can't get the facts straight or the study is flawed. I would not be surprised if both were true.

"Joe Smith doesn’t want to hire young African-Americans....": is that true even if Joe is black? In fact, isn't it racist for you to have assumed that he isn't?

In a study of car dealerships and racial discrimination, it was found that car salesmen charged Blacks higher prices than Whites, and Black salesmen had the largest gap.

No they do not 'charge' prices unless they're the odd-dealer who does not haggle. Salesman are more resistant in negotiations with black customers under the assumption that blacks have a different set of utilities and will ceteris paribus, pay more for display goods. That's third-degree price discrimination; it's not sweet but business often is not. Blacks who wish to be free of that and live in Central New York can buy their care here

Where the prices are take it or leave it.

I read once, a long time ago, that about 60% of car buyers pay MSRP which is an eggregious markup.

I got taken to the cleaners on my first two cars. I think the problem is informational and experiencial. If blacks are, on average, less informed (less access to Edmunds, KBB) and less experienced (fewer car purchases) that could provide a non-racial explanation.

Somewhere upwards of 70% of people who walk out the dealership door never come back. Commissions provide incentives. I find it hard to believe that salesmen havent found the equillibrium set of tactics that extract and exploit information from the customer. I see no incentive for discrimination.

In fact, lets suppose that, on average, blacks have poorer credit. The dealer is going to be making more money from the financing than profit on the car. That makes them more likely to do what it takes to close the deal.

People used to say that every NBA had to have a least one white starter to keep fans. So their is definitively some discrimination, Never the less I think it is a myth that blacks do bad in the USA. Most blacks in the USA are not poor(only 1/3rd are considered poor). In fact blacks do great here, for example they dominate the most desired jobs in the USA (those in basketball and football) and sports are very important to people. If a guy plays college sports he will talk about it for his whole life and it will give him higher status.

And who can blame blacks for not getting into the schooling rat race and then the work rat race. You spend your youth studying so you can do a boring (though high paying) job for the rest of your life. Maybe it is better to take it a little slower and easier and have less stuff. Just because blacks do not live like some want them to live does not mean that they are not doing well.

Now government needs to stop persecuting blacks through the war on drugs and needs to hire more black police.

Now IMHO white people are excessively fearful of blacks most blacks are not that hostile to whites. In fact the median black makes a great citizen. Also blacks in the USA invented as, blues rock & roll.The world would be poorer without those. They have added a grace to sports. I heard on one of the popular news magazine shows that a black person is 8 times more likely to help in an emergency than a white person.

Nobody cares what somebody else heard on some popular news show.

What's been called "rational discrimination" will continue to be a problem.

The taxi driver knows young black men commit more robberies than young white women, and passes up the fare: it's against regulations, but it's not entirely irrational if, statistically, there is a significant correlation between race or ethnicity and the risk of getting robbed..

Rational discrimination may or may not be unlawful (A pedestrian breaks no laws by crossing the street to avoid a crowd of young, black men) but, lawful or not, it seems likely to persist simply because it is, in fact, rational. Even (or despite) the also-true argument that it's unfair to the young black man who wishes to be judged "not be judged by the color of his skin but by the content of his character."

As Jesse Jackson once remarked, "There is nothing more painful to me at this stage in my life than to walk down the street and hear footsteps and start thinking about robbery. Then look around and see somebody white and feel relieved.... After all we have been through. Just to think we can't walk down our own streets, how humiliating."

I don't know if rational discrimination is part of the story here, and I'd not be surprised if (due to human's notoriously lazy thinking habits that substitute heuristics for logic) rational discrimination tends to also produce irrational discrimination.

But I do think the problem is not quite so simple as it's sometimes made out to be.

Yeah, but all of your examples refer to situations where folks may find themselves in imminent peril and act with extreme prejudice.

This is a tiny subset of the interactions we have with one another. I refer you to the post, which discusses evaluating candidates for jobs.

So no, I don't see your examples of rational discrimination as being germane to this conversation.

Whenever you can indulge the concept of treating individuals as individuals, you should do this. It's basic common sense. Within group variation swamps between group variation.

You are assuming the objective is profit maximization instead of minimizing maximum regret. A robbery can erase many days of pay, and a murder does worse.

It is certainly true that most blacks are honest and law abiding, but taxi drivers are not picking up fares randomly drawn from the population. Drivers have lots more information than you give them credit for.

No, I understand the rational discrimination of the cab driver, I just don't see the relevance here, where presumably you have the opportunity to learn something about the individual in question without imperiling yourself.

Good point.

Common sense? You first. You can individually canvass everybody in a majority black school district and tell your wife how much money you'll save on the mortgage. Or biz and gov to stop preferring Ivy League.

I have no idea what you are talking about. I'm talking about evaluating individual job applicants. I understand some employers prefer to recruit Ivy Leaguers, but I don't understand what in the world that has to do with the conversation at hand.

It's common sense to generalize. That's why your realtor doesn't bother showing your wife certain neighborhoods. Pattern recognition evolved over a long time. It's common sense, not Title VII.

AG, It's common sense to generalize in the absence of specific information. Why do you keep talking about real estate?

Interesting that the discrimination effect was smallest in two cities, Baltimore and Portland, that stand on opposite ends of the demographic spectrum.

LOL. Yes it's all those racist grannies blocking black advancement in the reverse mortgage and adult diaper sectors.

I think the next fruitful study would be why business elites seem so obsessed with employing immigrants, rather than seeing if some of the 1 in 4 to 1 in 5 unemployed black Americans are available to take up the slack. One might suspect they prefer employees who are vulnerable due to economic straits, language limitations and deportation threats to get themselves a more docile, less physically-threatening workforce. But nobody's that cynical, right? Right?

Globalism is pushing blacks down a very tall totem pole.

How much of it, though, is their unwillingness or inability to take low skilled jobs?

When I was younger, every hotel maid was black. Now, they are all speaking Spanish.

Where you live. Where I grew up, nursing home staff from top to bottom is plurality black. The orderlies are black, the dietary staff is black, the nurses aides are commonly black, a share of the RNs are black, the property clerk is black, and some portion of the treasury and control staff are black. The place my mother was billeted had a good institutional culture. The only problem employee was the diphead social worker (ethnicity unknown, married an Italian). "P" tells us all this can't happen.

The immigrants don't bring along failed background checks from their home countries. The unemployed black Americans' checkered past eliminates many of them from a job that's taken by an Honduran with a blank slate.

Race correlates with a number of work relevant variables of performance, and hiring is a process that relies on incomplete information. So it is rational to consider race as one of many variables in the hiring decision.

This is @ Brian Donohue

I don’t have the two minutes it takes to make a cursory examination of every individual I come across. Heuristics FTW!

efficiency is how we got here!

Nice to see some recognition that market forces can encourage racial discrimination.

No doubt some libertarians get the vapors at the idea, but it's true nonetheless.

So the biggest discrimination effect was less than 5 percent? That's tiny. In fact I would argue that such small callback variations represent spectacularly good news on employment racism.

My guess would be that it refers to a 4% increment over a foundation, so, perhaps 20% v. 16%.

Black job candidates really do come with a lot of baggage which can't be readily assessed from their paper qualifications or interview responses. For example, they are very much more likely to have close relatives and friends who are criminals, and who may "get them into trouble" down the road. Also, most college degrees claimed by black candidates must be heavily discounted to allow for affirmative action. Any attempt to zero in on these risks per-candidate will be parsed as intentional racial discrimination, forcing the would-be employer to engage in statistical discrimination which is harder to litigate. This is why the "identical resumes with the names changed" survey method is grossly flawed. Given affirmative action and so-forth, the "black name" resumes should be made considerably stronger to be "equal" in prospective value to an employer (except for government jobs, of course).

Hi, i feel that i noticed you vijsited my site thus i ggot here to return thhe want?.I'm trying to
find things to enhance my web site!I guess its ok to use some
of your ideas!!

This post is proposing the idea that blacks are less likely to be hired than whites for customer based work. This is hard to conclude because of the large amount of variability in factors that go into the hiring process, though this study is rather well developed by holding most of the influential factors constant and controlled.

According to this study, if an employer needs an employee to work customer-based work, they are less likely to hire a black than a white. In contrast, if the employer is looking to fill a more coworker-based position, the difference gap lessens significantly. The justification used in this post is that a black employee is potentially less appealing to an elderly white person, therefore the employer would want to hire a worker that will attract more business than they scare away.

A large factor that may not have been considered is the business’s targeted customer basis. If the employer’s targeted customers are primarily black, you would expect to see a much higher hiring rate for blacks than whites, and vise versa. Another argument; is it justifiable to have a hiring preference that fits your target customer basis, even if it is racially discriminated? It is understandable in my eyes to try to maximize business efficiency, but it is still a highly controversial topic that will cause a lot of conflict, regardless of the reasoning. You also have to consider those employer’s who actually are racially discriminative, and distinguish between the two (if the latter is seen as justifiable), which is a hard and large issue to tackle.

Good day very cool blog!! Man .. Excellent .. Amazing ..
I will bookmark your website and take the feeds also?

I'm glad to seek out so many useful info here in the
put up, we'd like work out extra strategies on this regard, thanks for sharing.
. . . . .

Comments for this post are closed