More College Educated Working Women Marrying College Educated Working Men Raised US Income Inequality Since 1960

From The New York Times, Economix, "Income Inequality in the U.S. Means Princes Don’t Go After Cinderellas" by Shaila Dewan:

But a new working paper from the National Bureau of Economic Research says that Americans increasingly engage in what economists call “positive assortative mating” — meaning they marry someone who is about the same as they are. Princes do not tend to go after Cinderellas.

The paper’s authors, led by Jeremy Greenwood at the University of Pennsylvania, mined census data from 1960 to 2005 and found that people’s tendency to marry someone of the same education level as their own increased steeply. After taking into account the increases in the education levels for men and women that have occurred between 1960 and 2005, the odds of a college-educated male marrying a college-educated female rose by 12 percentage points.***They then looked at the effect of this self-selection on income inequality. In 1960, they found, a couple without high school degrees would have made 77 percent of the mean household income. But if the woman without a high school degree married a man with a college degree, their household would have made 124 percent of the average.

In 2005, a woman with post-college education married to a man without a high school degree would have had a household income 92 percent of average. But if she married a man with a similar level of education to her own, they would have been making more than twice the average income.

The authors compared their marriage model to a measure of inequality known as the Gini coefficient — the higher the Gini coefficient, the greater the inequality in a given society. From 1960 to 2005, the Gini coefficient in the United States increased to 0.43, from 0.34, representing the dropping share of income controlled by poor families.

If people married completely randomly in 1960, instead of following the pattern they did, the Gini coefficient would have remained about the same. But if people married randomly in 2005 (or if they followed the preferences of couples in 1960) the Gini coefficient — and thus income inequality for the society at large — would have been significantly lower.

Not only are people more apt to marry someone similar to themselves today, but their choices also matter more to society. That’s because a far greater share of women work today than before.

Before, it didn’t matter so much whom you married if your household was going to have only one income, Mr. Greenwood said. “Plus, the higher educated women earn a lot more now relative to what they did in 1960, so that makes sorting more important in 2005,” he said. [Emphasis added.]