Yesterday I reported on Bryan Caplan’s thoughtful criticisms of chapter six of The Logic of Life. One of the things that worried me most in that chapter is the possibility that employers don’t value qualifications or experience of minority groups, which in turn means that the victims rationally invest less in education and work experience, which in turn feeds statistical discrimination.
Bryan is quite right to point out that there is nothing inevitable about this. It might well be that a victim of discrimination had a higher return to education, not a lower one. I argue that this is true for women and that is a likely explanation of the fact that women tend to be better educated than men. Bryan thinks the same story is true for African-Americans:
I tested these claims using one of the world’s best labor data sets, the NLSY. The results directly contradict Tim’s self-fulfilling prophesy story. Blacks actually get a substantially larger return to education than non-blacks! The same goes for experience, though the result is not statistically significant. The real lesson of the data is that if you are young, gifted, and black, you should get a ton of education, because it has an exceptionally large pay-off.
Bryan is very smart so I take that claim seriously. But it’s informal work, not published in a journal. Bryan says he will tell us more, and I’ll look forward to that. I set it against the results of a large (n=5000) randomised audit trial from Marianne Bertrand and Sendhil Mullainathan, who find that qualifications and experience on an apparently-black CV do not result in a higher chance of being invited to interview. Qualifications and experience on an apparently-white CV, of course, substantially improve those chances.
Let’s recognise the limitations of that trial. It is just one experiment, and focuses on employers only in Boston and Chicago. It’s also true that a distinctively black name may indicate something about social class as well as race. And I admit that an interview call-back is not the same thing as a job. But I still find a randomised trial awfully persuasive even when set against peer-reviewed econometrics, let alone Bryan’s informal analysis.
One commenter (at Bryan’s blog, EconLog), writes:
Bryan, this is a good post, but you need to cite your sources…If this isn’t published, Hartford [Harford, please - TH] can’t be faulted.
I wouldn’t put it quite that way. Bryan can be unpublished but right. I can still be wrong even if I “can’t be faulted”. For now, I think Bryan’s critique is powerful but unproven.
For CV-audit junkies, this paper by Judith Rich and Peter Riach is well worth a read: shows discrimination in favour of men in a traditionally-male occupation (engineering) and in favour of women in both traditionally-female and mixed occupations. All the firms surveyed are based in England.
Update: Bryan links to further research supporting his calculations.