Last week was Equal Pay Day in America, this year it fell on the 97th day of the year, marking the amount of additional time a woman needs to work to earn a man's wages. As I have asked before on this blog, it is a mystery to me why anyone would expect men and women to have similar averages in anything they do, since there are good evolutionary rationales why we might differ. Not that differences are necessarily large; the pay gap is not really the 77 cent on the dollar, as the organizers of Equal Pay Day presumably claim, but rather something like five cent once factors such as different occupations and job experience have been taken into account. No-one really knows why there remains a (much smaller) gap and it is an intriguing question.
But in this post I want to focus on attempts to equalize pay by legislation under the hypothesis that what really causes the pay gap is employer discrimination against women. Suppose women and men in some occupation possess equal amounts of human capital and are equally productive workers. If employers discriminate against women, they are willing to forego some money income in order to satisfy their taste for discrimination. How much profit depends on the strength of their preferences for discrimination. Since non-discriminating employers are willing to pay a man's wage to a woman, women will prefer working for such employers.
This means that positions available at non-discriminating firms be first to be filled, followed by positions at firms which discriminate only a little, followed by firms discriminating a little more, and so on. The employers with the strongest anti-woman preferences will never hire any women, because the women all choose to work for non-discriminating employers who are willing to offer higher wages.
The last position which was filled might have paid something like 85 cent on the dollar, but the woman who filled that position - the last woman in the work force - has no better options, so she takes it. Now what happens if all firms are ordered by the government to pay their female employees one hundred cent on the dollar ("Equal Pay for Equal Work")? The 85-cent-on-the-dollar woman is fired at the first opportunity. There is no way her employer will pay a man's wage for her job. In fact, all women are fired whose discriminating employers do not value their work the same as work by men. Maybe the government will make it illegal to fire women as well, but then fewer women would get hired in the first place, so unemploymet would still be the ultimate consequence of "equal-pay" legislation.
I wish this analysis were original with me, but it is actually found in Gary Becker's The Economics of Discrimination. It is for the above reason, also, that relatively great unemployment among women or among some minority is probably not explained by discrimination provided firms are not required to pay the same to all groups. Because if firms are allowed to reduce wages for discriminated groups, the incentive to hire members of non-discriminated and discriminated groups will be the same once different wages incorporate tastes for discrimination.
I don't know, but I think that, if women are significantly discriminated against by employers, "Equal Pay" legislation is bound to either fail or be fairly toothless given the above considerations. I would not expect governments to introduce legislation which results in significant and relatively great unemployment among women, and if they did I would expect other policies to follow which provide loop holes for discriminating employers, effectively annulling the first piece of legislation. Therefore, if a forceful "Equal Pay" act ever comes into existence, I would take that as evidence that women are not suffering significantly from discrimination in the labour market.