Explaining the Black-White Wage Gap


As of 2010, black men in America earned 74.5 percent of a typical white man’s wage; black women earned 69.6 percent. A new paper from Harvard’s Roland Fryer (certified genius), Princeton’s Devah Pager and Jorg L. Spenkuch of the University of Chicago examines some of the factors driving the black-white wage gap.

Using data from unemployed workers in New Jersey who sought employment for up to 12 weeks, the authors show that racial discrimination accounts for one-third of the wage difference. They also estimate that blacks have a 7 percent lower reservation wage than their white counterparts at a comparable job that demands a comparable skill level. Fryer and his colleagues control for skill level by measuring the job applicants’ wage at their previous job against the wage they were seeking.

Here’s the abstract:

The extent to which discrimination can explain racial wage gaps is one of the most divisive subjects in the social sciences. Using a newly available data set, this paper develops a simple empirical test which, under plausible conditions, provides a lower bound on the extent of discrimination in the labor market. Taken at face value, our estimates imply that differential treatment accounts for at least one third of the black-white wage gap. We argue that the patterns in our data are consistent with a search-matching model in which employers statistically discriminate on the basis of race when hiring unemployed workers, but learn about their marginal product over time. However, we cannot rule out other forms of discrimination.

The differences in wages between blacks and whites is between 25 and 30 percent, and the need to explain this remains a vexing debate in social science today. Throughout their research, the authors verified a story arc in which employers discriminate against blacks in terms of initial wage, but then slowly raise wages as they learn more about the individual employee.

To conclude our analysis, we explore the extent to which discrimination based on animus or differences in pre-market skills can account for our set of facts: (1) blacks incur larger losses than whites with job separations; (2) blacks have lower reservation wages; and (3) blacks garner higher returns to tenure in a firm.

The study shows that although the black-white wage gap widens by .9 percentage points per year of potential labor market experience, it decreases by 12 percentage points per year of tenure with a given employer. Meaning, blacks have much more of an incentive to stay at a firm rather than look for new employment since “the market provides less insurance than it does for equally skilled whites.”

The authors conclude:

This suggests that alleviating racial inequality may take a combination of policies to both eliminate barriers to investing in pre-market skills and anti-discrimination enforcement so that minorities are appropriately rewarded for those skills.

Which is an interesting thought, considering the recent cookie protests sweeping the media.


As an HR professional, I can see how this could take decades to remedy itself. Consider the following situation. A black candidate applies for a job where the salary range is $30-35K and the average salary for the existing team is $34K (all white, long tenure). If that candidate is currently making $25K, very few companies are going to offer $34K. That candidate is likely to see a $28-$30K offer. Why offer a person a 36% raise when you know you can hire them with a healthy 10-15% raise). On the surface, it looks discriminatory. However, it's more of a function of the previous salary. Historical salary gaps (racial, gender) take a while to level off in part because of this phenomenon.


As someone who has had various jobs in the same field at vastly different wages, I hate the "what was your previous salary" question. I feel that should be as verboten as "What is your religion?" or other irrelevant questions. If I take a low salary at a job because I get other intangible benefits there (like believing in their mission, etc), it shouldn't put me permanently on a lower pay scale for my whole career. Similarly, if I happen to score a very high paying job, it shouldn't price me out of future jobs.

I should be paid commensurate with the job I am taking. My previous salary is as irrelevant as is my race, gender, religion, etc.


I'm not a fan of this question either but someone's current salary is usually (but certainly not always) a good indication of what they're worth.

If you're underpaid (or overpaid for that matter), you can try to focus the discussion to what your are looking for or what you're worth. It takes some talent to do that.

I'm seeing a similar situation in college applications. They ask quite a few questions about your parents/siblings ...


The fascinating thing to me is we are highly willing to see things in others that we deny in ourselves. This is an example.

Consider Israel. If you read about Israeli Arabs - meaning Arab citizens of Israel - it's most often in the context of discrimination: they receive fewer services, have higher unemployment, etc. But on essentially all measures Israeli Arabs do better versus Israeli Jews than African-Americans do versus white Americans. The wage gap is smaller - and that's not excluding the Bedouin, who make very little and who aren't part of either society or accounting for larger Arab households. Every measure, from infant mortality on up. (I'm using UN figures, btw. You can look them up.)

We can also easily see racism in the worse off condition of Turks in Germany or Algerians in France.

The point is this: a big chunk of the African-American gap is racism. We deny it because we don't want to see it.


Bryan S

"The point is this: a big chunk of the African-American gap is racism. We deny it because we don’t want to see it."

Does 1/3 not count as a big chunk?

"Taken at face value, our estimates imply that differential treatment accounts for at least one third of the black-white wage gap."

I'm having a hard time seeing the denial...


Wouldn't some smart people figure out there's an inequality?

They would start businesses and staff them qualified but underpaid minorities (think MoneyBall). Give them equity stakes in the business.

This would be similar to what smart people do at big companies. They get fed up with not being appreciated and/or under-compensated. They either start their own companies or move to small companies where their efforts might be better rewarded.


I think the Freakonomics Blog also shared a story from The Economist showing that international companies were taking advantage of sexual discrimination in South Korea to get high quality, low cost female workers there.

So is there an opportunity for non-discriminating companies here too: to hire low cost but high quality black workers who are otherwise being discriminated against by employers?


And that resolves the issue how? If you pay them low wages, you are only contributing to the problem.


The South Korean women have choices. Local company that is going to pay them low wages because of sex discrimination, or multi-national company that is going to pay them 10-25% more than the local company (perhaps even more, I don't know.) The local company now, if it wants to have that high quality woman working for 'em, is going to have to meet the multi-national's rate, or provide some other incentive for the women. And thus competition enters...

Remember a core component of economics: it's about tradeoffs and alternatives. The South Korean woman has basically five alternatives available to her. Work for the local company. Work for the multinational. Work for herself. Not participate in the labor market. Go overseas. Each additional alternative works to weaken the impact of sex discrimination.


A good response to the direct salary question is to answer with the percent you expect to capture of the value you create. Admittedly this is tougher in jobs without direct link to sales. For example if you know expectation is $500k in net income then respond that you expect to capture 30%.

Steve Sailer

So, you could evidently make a killing by starting a company that hires only black people, since they are, according to a certified genius, undervalued by the market. Hey, it worked for Berry Gordy in 1959!

Bill McCullam

If you could break out the data, it would be interesting to compare differences between races based on white vs. black employers.