Tyler Cowen points me to a paper today about the rise in assortative mating. Basically, this means that we increasingly marry people who are similar to ourselves. High school grads tend to marry other high school grads, and college grads tend to marry other college grads. The authors of the paper conclude that this has implications for rising income inequality:
If matching in 2005 between husbands and wives had been random, instead of the pattern observed in the data, then the Gini coefficient would have fallen from the observed 0.43 to 0.34, so that income inequality would be smaller. Thus, assortative mating is important for income inequality. The high level of married female labor-force participation in 2005 is important for this result.
The table on the right is a standardized contingency table that compares 1960 to 2005. The diagonal numbers show the percentage of each educational class who are married to others of the same educational class, and in every case the numbers are higher in 2005. This does indeed suggest that assortative mating has contributed to increasing income inequality. However, I’d offer a few caveats:
- Comparing observed GINI with a hypothetical world in which marriage patterns are completely random is a bit misleading. Marriage patterns weren’t random in 1960 either, and the past popularity of “Cinderella marriages” is more myth than reality. In fact, if you look at the red diagonals, you’ll notice that assortative mating has actually increased only modestly since 1960.
- So why bother with a comparison to a random counterfactual? That’s a little complicated, but the authors mainly use it to figure out why 1960 is so different from 2005. As it turns out, they conclude that rising income inequality isn’t really due to a rise in assortative mating per se. It’s mostly due to the simple fact that more women work outside the home these days. After all, who a man marries doesn’t affect his household income much if his wife doesn’t have an outside job. But when women with college degrees all started working, it caused a big increase in upper class household incomes regardless of whether assortative mating had increased.
- This can get to sound like a broken record, but whenever you think about rising income inequality, you always need to keep in mind that over the past three decades it’s mostly been a phenomenon of the top one percent. It’s unlikely that either assortative mating or the rise of working women has had a huge impact at those income levels, and therefore it probably hasn’t had a huge impact on increasing income inequality either. (However, that’s an empirical question. I might be wrong about it.)
This is interesting data, which is why I’m presenting it here. And it almost certainly has an impact on changes of income distribution between, say, the top fifth and the middle fifth. But the real drivers of rising income inequality, which have driven up the incomes of the top one percent so stratospherically, almost certainly lie elsewhere.