Tyler Cowen had a popular piece in the New York Times last week in which he argued that assortative mating (rich marrying rich, poor marrying poor) reinforces income inequality. As in his book "Average Is Over," I think Cowen makes an important mistake in his column, though it is a mistake nearly everyone involved in discussions of distribution makes all of the time.
The distribution of income in society is a function of two things: distributive institutions and everything else. By distributive institutions, I am referring mainly to the laws that pertain to the distribution of resources in a society. Distributive institutions include such things as bankruptcy law, corporate law, securities law, property law, contract law, tax law, benefits law, and so on. Put another way, distributive institutions are the distribution-specific subset of the "rules of the game" that structure economic activity in society.
When Cowen writes that assortative mating pushes up income inequality, he implicitly treats distributive institutions as necessarily static. Cowen does the same thing in his book when he predicts that future technological changes will cause income inequality to go up. But distributive institutions do not need to be static in the face of changes in "everything else." Where changes in "everything else" threaten higher inequality within an existing set of distributive institutions, those institutions can be reformed to neutralize the threat.
Without getting deep in the weeds of the philosophy of causation, it suffices to say that: changes in "everything else" only result in higher inequality when distributive institutions are accommodating of that result.
If a rise in "power couples" threatens to increase inequality within our existing set of distributive institutions, we could cut off that threat by increasing taxes on high-earners and using the revenue from that change to cut taxes on low-earners or finance various transfer payments. This is an especially easy things to do for families with children (a particular focus of Cowen's) because you could jack up tax on high-earners and use the proceeds to fund expensive in-kind services (such as free child care and health care) and direct cash payments (such as a child allowance).
The fact that distributive institutions are not a suprahistorical constant should be an obvious point. But it's a point that is constantly missed, especially in the poverty realm where I spend most of my time. In the poverty realm, most of the respectable discussion is about how to change "everything else" so that it delivers lower poverty within our existing set of distributive institutions. Little attention is paid to the fact that the national income of the US is so high that even slight changes in our distributive institutions could bring sweeping anti-poverty gains. In fact, anti-poverty efforts that focus on the country's distributive institutions are almost absent from mainstream poverty discourse. Even poverty experts who know better often exclude distributive reform because it is "against American values" (which is to say too counterhegemonic to happen in the short term).
The inattention to distributive institutions in inequality analysis is a critical mistake. If you want to create and, more importantly for the present point, maintain an egalitarian society, you must be prepared to reform distributive institutions fairly regularly. This is because "everything else" in society is dynamic: technologies change, demographics change, ways of living and relating to one another change. Some changes will improve distributive fairness within the existing distributive institutions while others will threaten it. Where a change threatens distributive fairness, institutional reform must meet the threat.
For libertarian types that think altering distributive institutions is tantamount to violent theft, the prospect of repeatedly tinkering with them may be a bleak one. But for everyone else, this kind of tinkering should be welcomed as a great facilitator of personal freedom and dynamism. Instead of trying to stop inequality by freezing ways of life (such as the oppressive way of life that left women completely locked out of higher education), reforming distributive institutions gives us the ability to accommodate social changes without risking distributive unfairness.