The #metoo movement has brought renewed focus on gender equity questions. Economics examines the pay gap between men and women, and a recent analysis from the Federal Reserve Bank of St. Louis links this gap to marriage, creating a puzzle for economics.

The gender pay gap is large: among workers with at least a high school diploma between ages 45 to 54, men earn almost 50 percent more than women, roughly $75,000 versus $50,000 annually. Is this evidence of discrimination against women we could address through comparable worth pay legislation? Perhaps, but first let’s dig deeper into the issue.

Labor economics explains wages and salaries based on productivity, or the extra output that a worker helps a business produce. Firms can afford to pay workers the value of this product and still make an adequate profit. Competition among firms to attract and retain good workers should drive salaries up to this level. If Alabama underpaid Nick Saban, other universities would happily compensate him fairly.

Salary differences should then depend on differences in productivity. Economists would want to make more nuanced salary comparisons by gender in narrower job categories before concluding that women are paid less. Education and skills requirements differ way too much across jobs requiring a high school diploma to be conclusively informative.

The St. Louis Fed analysis provides a different perspective: the gender pay gap is really a gap between married men and everyone else. Single men, single women, and married women all make around $50,000 in the prime earning years of 45 to 54; married men make almost $90,000. Interestingly, no pay gap seems to exist between single men and single women.

Can we make sense of this? First off, marriage may not necessarily make men more productive. Men who are more productive – that is, have more education, training, and drive to succeed – may be more likely to be married. We need not believe that reciting the marriage vows increases men’s (but not women’s) productivity.

Marriage could also make men focus seriously on work and a career. We might recognize that at some point we became much more serious about work; for me, this occurred in grad school. Marriage may have this impact on many men. Seriousness and focus could explain higher earnings, and since economists can’t easily measure a person’s seriousness directly, in the data this will look like a marriage effect.

There’s another possible explanation. In many workplaces, bosses have discretion over giving out raises, and an employee might have to ask for a raise. Suppose married and single male employees both ask for raises. The boss might believe that the married man “needs” the raise more – to pay for his kids’ braces, or to help take care of his in-laws. While plausible, salaries based on need violate the labor economics theory.

And it undermines a potential argument against comparable worth pay legislation to narrow the gender pay gap. The argument maintains that businesses must be given the freedom to pay employees based on productivity. But if compensation based on perceived need does not ruin our economy, then raises for women surely won’t cause an economic train wreck.

All the above factors likely contribute to married men’s higher earnings. Businesses can deviate at least some from productivity in setting wages and salaries without going bankrupt. The greater consequence of comparable worth legislation is shifting salary determination ultimately from businesses to bureaucrats. In the long run as politics determines more salaries across the economy, economic performance may decline significantly.

Labor economics seeks to explain salaries across different jobs, but productivity theory is also gender (and color) blind. Although women may indeed not be paid according to their productivity by every employer, competition should prevent pay from getting too far out of line with productivity. Hopefully bosses will reward underpaid women employees, because #metoo has sadly shown that politicians are not always gender blind.

Daniel Sutter is the Charles G. Koch Professor of Economics with the Manuel H. Johnson Center for Political Economy at Troy University.