When women earn more but wives earn less (Mundy/Rosin edition)

We have been told many times now that in U.S. metropolitan areas, among 22-30 year-olds who have no children, have never been married, and work full-time and year-round, women’s median income is higher than men’s.

One implication both Hanna Rosin and Liza Mundy draw from this is that, when these people get married, there will be more marriages in which women are the higher earners.

I previously showed that women’s advantage in this group is partly the result of its odd race/ethnic composition — with a lot of White women and Latino men especially — reiterating  that men outearn women at all ages and all education levels.

But what if that group really was the marriage market, with more and more women looking down the economic ladder at the men they’re dating? Would we have more egalitarian married couples? Maybe, but not necessarily.

Atlanta is a good place to start, since, according to Liza Mundy’s book The Richer Sex, it’s the place where women’s advantage is greatest:

Of all the major cities where young women outearn young men, Atlanta is number one. Well do these [high-income women she’s interviewing] know the accuracy of that statistic. “I never had a boyfriend who made more than me,” says one of them.

Let’s look at this case, starting with a hypothetical income distribution of men and women:

These women have it made, with a median income $2,000 greater than men’s. Now let’s imagine that 80% of them get married — which is 8 marriages involving 16 people. Look how egalitarian the couples will be:

Whoops. Turns out men and women don’t get married randomly. If all men decide to marry women with lower incomes, they can — they just have to squeeze out the poorest two men and ignore the two richest women. Or, maybe it’s that the richest two women opt out of marriage, and the rest of the women taking advantage of the chance to marry up, ignoring the two poorest men. Either way, there’s a $2,000 male advantage in every couple, and the median incomes are now reversed, with married men having a $2,000 advantage, $41,000 to $39,000.

Is this what’s happening in Atlanta? Pretty close, I think.

Using the 2008-2010 American Community Survey (ACS) for the Atlanta metro area (three years pooled for larger sample size), I got the never-married, childfree, full-time and full-year employed men and women ages 25-34 (22-30 is not a good marriage market, since the average age at marriage is near the top of that range).

Sure enough, the women in this group have an earnings advantage, with median earnings of $37,473, compared with $35,000 for men. The distribution looks like this:

I’ve added a few calculations to the figure. The “index of net difference” (ND) is a handy tool for showing how two groups rank hierarchically along a single dimension (in this case income). Using the formula given by Lieberson here, and these categories, I reckon that the chance that a random woman will be in a higher category than a random man from this distribution is 45%. The opposite, that a man will be in a higher category, is 39%, so the ND is -.06. So, that’s good for women.

Note also that, despite the alleged statistical know-how of the high-income women at Mundy’s table, there are actually 1.6-times as many men in the the top income range of this marriage market as there are women. They have plenty to choose from at the high end, even though women’s median is higher. That’s because there are more men than women in the pool. I suppose there are two reasons for this: First, fewer women work full-time year-round. And second, lots of people are single parents, and when children live with their mothers, it’s the mothers who are excluded from this pool — the fathers, not the mothers, come up as childless.

Anyway, that’s the pool. What about the marriages? The ACS identifies people who got married in the past 12 months. In that 3-year sample I have about 200 couples to work with. So, here is the income distribution for men and women, ages 25-34, with no children present, who just got married. I dropped the full-time and -year restriction, since people could have quit working, and I limited it to couples where both are in the age range:

Just as in our hypothetical example, richer men and poorer women got married. Now the distribution skews decidedly male. The ND has reversed, and husbands’ median income is $12,527 higher than wives’.

How is that possible, when, as we are reminded so often, women are so much more likely to have graduated college? Two reasons: first, people overwhelmingly marry partners on the same side of the BA/no-BA divide; and second, men with BAs make more than women.

Here’s the Atlanta situation. First, education: Women in this group — 25-34, FTYR, never-married, no kids — are much more likely to have BAs: 62% to 41%.

However, in 72% of couples both spouses are on the same side of the BA divide (57% + 15%):

Setting aside all that educational endogamy, with so many more women BAs, women really are much more likely to marry down the educational ladder: 25% of childless, 25-34-year-old Atlanta newly-wed couples have a BA wife and a non-BA husband.

But men earn more at every education level. As a result of that — and by whatever additional machinations of partner-selection — only 38% of these couples have a higher-earning wife. Only a third of the BA-BA couples have higher-earning wives, and even when the wife has a BA and the husband doesn’t, that number is only 54%:

That’s how, even when you define the marriage market in such a way as to paint women’s situation as positively as possible — finding that rare niche in which women earn more than men — you discover the marriage system reproducing gender inequality.

3 thoughts on “When women earn more but wives earn less (Mundy/Rosin edition)

  1. If families where wives outearn husbands are more frequent at the lower income/education levels, and if that’s also where ideas about gender roles are less evolved, what happens? (BTW, your earlier post on earnings -“Mystery Solved” — is a very useful real-world example of Simpson’s Paradox, if anyone needs one.)


Comments welcome (may be moderated)

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s