Vioxx verdict seems outrageous

Some notes on comparable worth

My wife, among others, noted that the headline to a Washington Post story, "Roberts Resisted Women's Rights," was based in part on Roberts's strong opposition to "comparable worth". My wife briefly described some of the economic analysis of comparable worth but asked me to add some details.

I would have complied anyway, but inasmuch as she is currently an invited guestblogger on the top-ranked blog in the world, I am especially eager to help out.

In the late 1950s and early 1960s research had established that U.S. women working full-time outside the home earned, on average, only about 60% of what U.S. men working full-time outside the home earned. Research also suggested that in many workplaces women doing the same jobs as men were paid less. Observers therefore attributed some or all of the gender gap in wages to discrimination against women. (And these observers were not comforted at all when someone pointed out that, in a striking coincidence, God told Moses--Leviticus 27:3-4--that women were worth only 60% of what men were worth.)

In response, Congress passed the Equal Pay Act of 1963 and Title VII of the Civil Rights Act of 1964. Together they established in federal law the principle of "equal pay for equal work". A man and a woman doing the same job must be paid the same. I'm open to references claiming otherwise, but I've read that these laws are enforced strongly and that compliance has been, and is, high.

Imagine, then, the surprise and disappointment of the laws' supporters when for nearly twenty years following the laws' passage, the ratio of women's pay to men's pay remained almost exactly 60%. Here is a graph (from June E. O'Neill and Dave M. O'Neill, "What Do Wage Differentials Tell Us About Labor Market Discrimination?" NBER Working Paper No. 11420).

Activists then argued that the principle of equal pay for equal work should be extended to "comparable worth": different jobs, but jobs that were judged comparable, should also have to pay the same. I recall a television show long ago that used as an example tax assessors and librarians. Tax assessors were mostly males and they earned more than librarians, mostly female, did. But according to a multidimensional test devised by work experts, the education, intellect, and skills required to be a librarian were comparable to those required to be a tax assessor. A comparable worth law would thus make it illegal to pay librarians less than tax assessors.

I'll comment first on what economics has to say about what the likely effects of a comparable worth law would be and second on what we think we know about why women earn less on average than men.

A comparable worth law is analytically identical to a minimum wage law, a minimum wage imposed on jobs held primarily by women. Economists' theoretical and empirical research on minimum wage laws suggests the law would therefore have the following effects.

1) Some women would receive increased pay and, especially in the short run, would be unambiguously better off.

2) Other women, however, would be worse off. They would bear two kinds of costs: some would have their work hours cut below what they would like to have, or they would lose their jobs altogether; some would see their working conditions worsen (fewer fringe benefits and more effort demanded). These costs would tend to grow over time.

3) Which women would win and which would lose? Experience suggests that less educated, younger, and minority women would be more likely to lose.

4) The number of people needed to enforce the law would grow, and the legal apparatus needed to enforce the law would broaden and become more convoluted.

Why do women earn less than men to begin with? Discrimination is a possibility. Economics cannot rule it out. But theoretically and empirically, another explanation is more appealing: women earn less because of their choices. Specifically, many women choose to combine a career with family responsibilities. Those responsibilities induce women to accept lower wages in exchange for the  flexibility--"family friendliness"--offered by certain occupations.

This "choice" hypothesis seems more consistent with a number of observations than does raw discrimination. For one, never-married women who have never had children earn about the same amount as never-married men who have never had children. (See the O'Neill and O'Neill paper referenced above.) These women presumably are freer to devote the same amounts of time and energy to their careers as men, and the market seemingly rewards them accordingly.

(I know, discrimination might explain this, too. Maybe men--we filthy dogs--discriminate more against married women. Because they are less "available"? But this seems like a stretch. And I won't deal here with the objection instantly raised by one of my female MBA students: "That proves nothing other than that never-married men are losers!")

Two other stylized facts are that the male-female wage gap tends to increase up to about age 40 and then it starts to decline, and that holding constant the number of children a working woman has, her wage gap compared to men is larger the further apart in time her children are born. I'm sorry, but I don't see discrimination easily fitting with either of these facts, but these facts do follow rather readily from the choice theory. See Fischel and Lazear, "Comparable Worth and Discrimination in Labor Markets," University of Chicago Law Review, Summer 1986.

One of the country's leading experts on this topic, June O'Neill (former director of the CBO, former vice-president of the American Economic Association, and currently Wollman Distinguished Professor of Economics at Baruch College), has concluded that the vast majority of the wage gap can be attributed to differences between women and men in their work experience and occupational/family choices. A short, non-technical article on comparable worth by Professor O'Neill is here. In the concluding paragraph of the paper referenced above, the O'Neills state (p. 33):

. . . the gender gap is attributable to choices made by women concerning the amount of time and energy to devote to a career as reflected in years of work experience, utilization of part-time work, and workplace and job characteristics. There is no gender gap in wages among men and women with similar family roles. Comparing the wage gap between women and men ages 35-43 who have never married and never had a child, we find a small observed gap in favor of women, which becomes insignificant after accounting for differences in skills and job and workplace characteristics. What the average woman sacrifices in earnings from choosing jobs that allow for part-time work and flexible work conditions is presumably offset by a gain in the utility of time spent with children and family.

Should the reader not completely trust economists and their econometric tricks--perish the thought, but there's no accounting for taste--I suggest a recent column by Anne Applebaum in the Washington Post (2/23/05, p. A19). Ms. Applebaum writes:

It isn't ability or discrimination that hold women up most, in other words, but the impossibility of making a full-time commitment to work in a culture that demands 80-hour weeks, as well as to family in a society unusually obsessed with its children.

We all know this anecdotally, but research confirms it. A British sociologist, Catherine Hakim, recently concluded for example that out of 3,700 working-age women she surveyed, about a third were fully focused on their jobs, about a third were fully focused on their families, and about a third wanted a mix -- meaning, invariably, that they took the sort of job that doesn't lead to fast-track promotion. If these numbers hold there never will be a 50-50 split between men and women at the highest professional or managerial levels of anything: The ratio will always hover around 2 to 1.

Is this nature or nurture? I don't see that it matters. What matters is that those women who want to become high achievers can do so, but those who want to stay home some of the time aren't forced, by economics or social pressure, to take high-pressure jobs.

Finally, let's look briefly at long-standing reasoning in economics that suggests discrimination, alone, can't explain the wage gap. Suppose, for the sake of argument, we think the wage gap is due entirely to discrimination. Currently, women earn about 75% of what men earn. If discrimination alone explains this wage gap, an entrepreneur--Teresa Heinz Kerry? Barbra Streisand? Carly Fiorina?--could replace an all-male firm with an all-female firm and immediately reap a 25% reduction in labor cost with no loss in productivity. Even if labor cost equaled only 20% of the firm's revenues, this change would raise the firm's net income as a percentage of sales by five percentage points. Is that a lot? Yes. A rough estimate is that the average U.S. firm's earnings equals five to ten percent. (I don't have a cite for the overall average; a quick Googling turned up this data on chemical firms' earning about 9 to 10% on sales.)

That is, pure discrimination would create significant profit opportunities for non-discriminators, such that discriminating firms would tend to be replaced by non-discriminating firms. firms that discriminated against women would create significant profit opportunities for firms that did not discriminate against women, and as discminating firms failed, male employees would tend to be replaced by women. [Better language in response to James McMurrin's comment below. Thanks, James.] Profits are a powerful motivator. Economists, at least, don't bet against them.