Married women earn less than single women.  In the NLSY, married women make 10% less, even after controlling for education, experience, IQ, race, and number of children.  How is this possible?

As I explained in my post on the male marriage premium, there are three competing economic explanations, each of which may be partly true: Ability bias, human capital, and signaling.  What is the breakdown for the female marriage penalty?  My tentative opinion:

1. Ability bias goes in the “wrong” direction.  Married women are, on average, more conscientious, ambitious, and cooperative than single women – and were so long before their weddings.  And at least in the modern world, high income makes it easier for women to find a spouse.  Income matters less for women than it does for men in the modern mating market, but female income is usually still a plus.  And even if income has little direct effect on women’s perceived desirability, higher income indirectly puts women into close contact with marriageable men.  Adjusting for ability bias, I think the true marriage penalty for women is roughly 20%. 

2. Human capital explains about 10 percentage points of the marriage penalty.  Even ignoring children, marriage causally reduces women’s focus on their careers.  Once they’re married, women want more work-life balance.  Some even reallocate their energy from promoting their own careers to promoting their husbands’ careers.

3. Signaling explains the remaining 10 percentage points of the marriage penalty.  Several commentators pointed out the fact that it’s illegal to ask job applicants about their marital status.  But people aren’t legal robots.  Interviewers frequently ask illegal questions about applicants’ personal lives, and applicants often volunteer their personal information just to make polite conversation.  When employers learn that a woman is married, they assume – correctly on average – that she will be slightly less focused on her job than an otherwise identical single woman.  No matter what the law says, employers have a strong temptation to factor this information into their hiring and promotion decisions.

In traditional societies, ability bias was probably close to zero: When almost everyone marries, there’s little pre-existing difference between the conscientiousness, ambition, and cooperativeness of married and single women.  The human capital and signaling effects, in contrast, probably used to be much stronger.  Many women – including my maternal grandmother – simply quit their jobs right after they married.  Even married women who kept their jobs typically made homemaking their top priority.  As you’d expect, the employers of yesteryear responded with a strong presumption against married women.  Remember: Signaling is just a special case of statistical discrimination.

Final thought: If the marriage-class correlation continues to increase, future employers might actually start to see female marriage as a positive signal.  Ceteris paribus, marriage will still predict a stronger desire for work-life balance; but marriage will also predict all the professional class traits that Murray discusses in Coming Apart.

Is my breakdown for the female marriage penalty correct?  If not, what’s the correct breakdown between
ability bias, human capital, and signaling?  Please show your work.