Misdirection on Assortative Mating and Income Inequality

This is a story about misdirection, how authors contort their analysis to answer a question no one is asking but make it seem as if they are answering an important current question, such as ‘why has income inequality increased?’. The consequence is to be grossly misleading or, worse, to present conclusions that are directly opposite of what one’s data show.

The paper in question is titled “US income inequality and assortative marriages” and written by Jeremy Greenwood, Nezih Guner, Georgi Kocharkov, and Cezar Santos for VoxEU.org. The research relates to an increase in positive assortative mating: “how likely a person is to marry someone of similar educational background. Since education is an important determinant of income, these patterns of matching might have an impact on the economy’s distribution of income.” Basically, if higher income men are now more likely to marry higher income women then household income inequality will grow.

The authors conclude that “rising assortative mating together with increasing labour-force participation by married women [emphasis added by me] are important in order to account for the determinants of growth in household income inequality in the US.” So, right out of the gate, a key influence not trumpeted in the headline (rising labour-force participation by married women) is introduced. But we’ll stick with the findings on assortative mating for now. The authors compare assortative mating in 2005 to that of 1960. The selection of the dates for comparison, 1960 and 2005, determines their story and they choose a misleading one. They show their key finding in the very first graph, presented below, which uses Kendall’s tau statistic to measure the relationship between husband’s and wife’s educational levels. The higher the tau statistic ‘the higher is the degree of positive assortative mating.”

The authors conclude from this graph that “while the series is not increasing throughout the entire period, the statistic is clearly higher in 2005 relative to 1960.” The increase in positive assortative mating is then shown to contribute to a rise in income inequality between 1960 and 2005.

I am confident that we will be reading conservative columnists and politicians using this research to claim that income inequality was caused by demographic trends and not by changes in the distribution of economic rewards. That would be an inappropriate conclusion. Why? Simply because the question that should be answered is why income inequality grew a great deal between 1979 and now and was fairly stable before then (the share of income, including capital gains, going to the top 10 percent was 32.2 percent in 1960, 33.5 percent in 1980 and rose to 46.6 percent by 2005). Greenwood et al.’s data show that positive assortative mating declined(!) from 1980 to 2005, which directly tells us that this phenomenon did not cause any of the income inequality generated after 1980: in fact, positive assortative mating was a force that equalized incomes after 1980. It was in the period from 1960 to 1980 that positive assortative mating lead to more unequal incomes. Consequently, their research in no way lifts up the role of ‘like marrying like’ in generating inequality since 1980: it actually means that economic inequalities overcame an equalizing demographic factor and that the inequality of economic outcomes had an even larger impact than we might have thought. The title of the VoxEU piece, “US income inequality and assortative marriages,” may seem like it’s answering a relevant question, but it isn’t: No one is trying to figure out why income inequality is greater now than in 1960, and interest in a factor that increased inequality between 1960 and 1980 but was equalizing post-1980 is definitely not informative about current inequality trends.

This is very similar to the sleight-of-hand used in articles bragging how incomes are greater now than in 1960, which is true but pretty misleading since the vast amount of that income growth for a typical family occurred between 1960 and 1973. That’s not quite relevant, right? The fact that assortative mating may have generated more inequality between 1960 and 1980 is similarly irrelevant to current day discussions of income inequality. Unless, that is, you are trying to mislead people.