Which Is More Effective at Reducing Discrimination, Competition or Law?

This was the question that Gary Becker (left) and Judge Richard Posner (right) exchanged viewpoints on at the Becker-Posner blog earlier this week.  Becker's post and comments to it can be found here, and Posner's here. As a preliminary matter, neither disagreed that employment discrimination laws were justified on grounds of morality and justice. But both explored the role of law and of economics on reducing discrimination.

Becker began by suggesting that competition has been more effective than law in India to reduce discrimination against untouchables. That assertion was contested in the comments to Posner’s post, but the analysis he goes through under that assumption is interesting, nonetheless.

Becker defined discrimination somewhat oddly as "when [an employer] refuses to hire [minority candidates] even though they are cheaper relative to their productivity than the persons he does hire." I don't think that is what he meant to define discrimination as--solely in market terms, because he continued with the explanation of why discrimination is economically inefficient, suggesting that he meant to link discrimination (refusing to hire minority candidates) with the inefficiency--that these candidates are cheaper relative to their productivity. He continues to explain the economic inefficiency of discrimination: 

Discrimination in this way raises his costs and lowers his profits. This puts him at a competitive disadvantage relative to employers who maximize their profits, and hire only on the basis of productivity per dollar of cost. Strongly discriminating employers, therefore, tend to lose out to other employers in competitive industries that have easy entry of new firms.

Becker further describes how discrimination will promote segregation in the workforce by catering to the discriminatory preferences of employees.

Posner seems to take a dimmer view of the power of competition to promote integration or diminish discrimination, stating that "[t]here is no reason for competition to affect that aversion [discriminators have], other than by bringing the costs of it home to employers and through them to their [discriminating] workers and customers."

At the same time, he discusses why antidiscrimination laws will not be efficient in economic terms.

Suppose white employees have a strong aversion to working with blacks. Then forbidding discrimination will impose a heavy cost on the white employees. If there are more of them than there are blacks, the cost to the white employees may exceed the benefits to the black employees. Of course, an antidiscrimination law may rest on a political or moral judgment that costs imposed by thwarting a taste for discrimination should not count in the social calculus, but that is a judgment outside of economics.

In other words, discriminating employees may have negative feelings that they get a lot of utility out of--that aversion may be a powerful positive feeling for the employees who hold it. Posner also suggests that antidiscrimination laws aren’t likely to be effective without very serious penalties, which we don’t impose:

[An] employer who wants to continue discriminating against blacks can (within limits) reconfigure his work force to reduce his demand for skills likely to be possessed by black applicants for employment, can substitute capital for labor, and can relocate to areas in which the applicant pool contains few blacks. Second, felt legal pressure to hire blacks results in "affirmative action," which both creates resentment among whites and casts some doubt on the average quality of black employees and so in effect stigmatizes the entire class. And third, because a discrimination law makes it more difficult to fire a member of the class protected by the law, it increases the cost of hiring members of the class and so increases the incentive to discriminate in hiring.

A fourth reason, I would submit, is that given that discrimination can operate below the fully self-aware level and can be very difficult to detect, it is too easy for discrimination to be hidden by reasonable sounding motivations.

The comments are interesting too–some touch on how cognitive bias makes it particularly difficult for “enlightened self-interest” to work so that the market alone can eliminate discrimination, although as Posner points out, neither he nor Becker has argued that it can be.

The exchange is very thought-provoking as most of what Becker and Posner write is. I'm not sure that I agree with all of the assumptions or conclusions, and I also think that the economic analysis of law is of limited usefulness, but I highly recommend reading it.

MM