While I was heading to the APEE conference in Cancun, I learned that economist Barbara Bergmann had died. I reviewed her book The Economic Emergence of Women in Fortune on March 2, 1987. Here are some excerpts from my review:
The liberal [DRH note: that was Dan Seligman’s adjective. I would not have used it myself but I had many battles to fight on this one] feminist agenda has beep pushed hard in the U.S., but the pushing has mainly been done by politicians and “movement” stalwarts. Not many professional economists have been promoting the agenda, which emphasizes the Equal Rights Amendment, “comparable worth,” and affirmative action. Now, at last, a trained economist has weighed in heavily on the feminists’ side. The Economic Emergence of Women, BasIc Books, $19.95), by Barbara R. Bergmann, presents the most sustained economic caseI know of for much of the agenda.
Bergmann, professor of economics at the University of Maryland, argues persuasively that women are occupationally segregated and major victims of job discrimination. She is less persuasive about other matters, however. For example, what causes the famous 40% male-female wage gap? Bergmann says she is confident that discrimination explains much of it. She rests her case mainly on statistical evidence suggesting that differences in experience, trairnng, etc., explain only half the wage gap–about 20%. The other 20%, she argues, must be mainly discrimination.
But that doesn’t follow. The wages of white males also vary widely, and the same kinds of statistical analyses typically explain only half of the variation there too. Economists do not therefore conclude–and Bergmann herself certainly would not–that the unexplained variation mainly reflects discrimination. Instead, they come away from the data saying that our statistical tools are imperfect and that much wage variation reflects factors that we simply do not know how to measure (of which discrimination may well be one).
I’ll publish a further excerpt in the next day or two.
READER COMMENTS
John Goodman
Apr 23 2015 at 3:09pm
Hasn’t June O’Neill effectively demolished the idea that wage rates reflect gender discrimination?
And if so, why are we still talking about it?
JLV
Apr 23 2015 at 3:31pm
There’s a subtle difference here though that the last paragraph elides.
Basically, regressions != Oaxaca-Blinder decompositions.
The unexplained part of a Oaxaca-Blinder decomposition does identify discrimination under certain assumptions, while (to my knowledge) no such interpretation of the size of the R2 in a regression is tenable.
Comments are closed.