On May 13, 2010, Paula England was inducted as the Frances Perkins Fellow of the Academy at the Newseum in Washington, DC. In inducting her, Heidi Hartmann, AAPSS chair and president of the Institute for Woman’s Policy Research, said, “For much of her 30 year career in sociology, Professor England has focused on one of the great puzzles of modern life—why women earn less than men. While this puzzle is still not completely solved, we know much more than we did 30 years ago, in no small part due to Paula England’s outstanding body of work.” The following is a transcript of Paula England’s remarks.
“It is appropriate that I am inducted by Heidi Hartmann, because she was ahead of me in recognizing a type of sex discrimination that most people never notice, but that I spent a couple of decades documenting. We all know about lack of equal pay for equal work in the same job. And we all know about hiring discrimination.
Data analysis done early in my career convinced me that there was also a third type of discrimination. I discovered that while women and men were in sex segregated jobs, on average, female-dominated jobs required as much education and general cognitive skill as male- dominated jobs. There were both male and female jobs at most skill levels. So why did female-dominated jobs systematically pay less?
I came to believe that employers implicitly take the sex composition of jobs into account when they set wages. If a job is filled mostly by women, they set a lower wage than they otherwise would. It is as if there were a cognitive bias toward thinking that if jobs are done by women, they cannot be worth much. This bias, I believe, reflects a general cultural devaluation of women and, by extension, roles associated with women. Institutional inertia cements this bias into wage structures.
Assembling convincing evidence that the sex composition of jobs actually affects pay absorbed me for 20 years. No matter what I controlled, I kept finding a net negative effect of the percent female of an occupation on wages. Critics suspected omitted variable bias: crudely put, maybe the low pay was because the women and men who selected into female jobs were just losers on some dimension our datasets do not measure. Against this claim, I showed with fixed-effects models that the same person gains money when moving from a female to male job and loses money moving the other way. More recently, in aggregate pooled occupation/year data, I used fixed-effects models to show that the same occupation pays relatively less as it feminizes, and that wages follow sex composition rather than the other way around.
In the 1980s and 1990s there was active policy discussion of this issue, dubbed “comparable worth,” or “pay equity.” Some, myself included, advocated requiring employers to show that they had used a consistent set of criteria to set the wages in male and female jobs. I crisscrossed the country participating in the debates with academics, students, lawyers, and personnel professionals. I testified before the U.S. Civil Rights Commission. I especially treasure the interdisciplinary dialogue I got to participate in. Many economists were convinced that the low pay of female occupations was due to crowding—although they could not measure crowding—or to compensating differentials. I argued that the latter begs the question of why the nonpecuniary job qualities that the typical man likes have wage premia, while those more women like are penalized. Once, an economist I was debating said something to the effect that most arguments for comparable worth make no sense but that I made the most reasonable case that she had heard (not that she was convinced). I decided to take it as a compliment.
I wish I could say I had a big effect on policy. Some states did change the relative pay of their female- and male-dominated civil service jobs. But the federal appellate courts decided that, in most cases, Title VII cannot be used to challenge the low pay of female- relative to male-dominated jobs. New legislation would be needed, and it has languished in Congressional committees for decades. Alas, female-dominated jobs still pay less than jobs of similar skill levels filled by men. Care work is especially badly paid, although I have argued that is distinctive in the positive externalities it produces.
While the devaluation of women’s work remains, other things changed a lot. Segregation of occupations declined, especially for the college educated, because of decreased hiring discrimination and women’s changed aspirations. This combined with women’s increased continuity of employment to narrow the sex gap in pay dramatically. These are huge changes. But a recent review I wrote of trends in gender inequality reminded me that we cannot assume that equalizing change, once started, is inexorable. Indeed, gender change on most fronts stalled since about 1990. Declines of the segregation of college majors and of occupations virtually ceased. While the “opt-out revolution” is a myth, increases in women’s employment have, indeed, largely stalled.
But I don’t want to be overly grim. There was, indeed, a gender revolution or many of us would not be here tonight.