How do economists think about discrimination? Economists have many theories of discrimination, why it happens, and what its effects are, but one way to think about discrimination is that it stops people from specializing. For example, a society that discriminates against women may not allow a woman to be a doctor even if she is well suited to doctoring.
Economics Professors Hseih, Hurst, Jones, and Klenow used this way of thinking to estimate how the US economy has benefited from reducing discriminating against women and minorities since the 1960s.* Using a model that thinks about discrimination as stopping people from taking the job they are best suited for,** they find that 40% of US per person economic growth since 1960 can be explained by decreasing discrimination.
Graph of US economic output (GDP) per person. The green line is what the authors find US GDP per person would be if the US had remained as discriminatory as it was in 1960.
The authors also do a few sub-tests that are interesting in their own right. For example, they find that it is more correct to say that the US discriminated against women and African Americans than it is to say the the US discriminated in favor of white men. How can they tell the difference? Well, discrimination in favor of a group can be thought of as a subsidy. How big would a subsidy would white men have to receive to explain why they held almost all high income jobs? The authors find that the subsidy would have to be several times the US' entire economic output in 1960. That's an impossibly large subsidy, so the authors conclude that it is more accurate to say that the US discriminated against women and minorities rather than in favor of white men.
I'm going to close out this post with two other very interesting graphs from the paper. The first shows the female-male and black-white wage gaps falling over time, and the second shows how occupational barriers have changed over time for white women in several professions.
US wage gap over time. The y axis is basically "what percent less on average can this group expect to earn relative to white men?" Note that the wage gap gets smaller over time for both white women and black men, and that the wage gap for women gets worse as they get to middle age, which is consistent with papers finding that having a child has a big negative impact on women's wages.
Graph of the occupational barriers faced by white women over time. A "1" on the y axis means women face no occupational barriers relative to white men. Note that while the barriers in some occupations, like law and medicine, have fallen significantly over time, other occupations have seen only a small reduction in barriers (such as construction, where, on average, men have a physical advantage in lifting heavy materials).
Pay attention to the scale on the Y axis! By 2010, women face about 16 times higher occupational barriers in entering construction than the do in entering teaching.
*Why do the authors think that discrimination has gone down since 1960? The data support that pretty strongly. For example, 94 percent of doctors and lawyers were white men in 1960. By 2010, the fraction was just 62 percent. It may not be gone, but it's much less than it used to be.
** Job I'm best suited for = job where my comparative advantage is highest