A final culprit is demographic trends—particularly changes in family composition. Since income is a “household” measure, the rising rate of single parent households is often invoked to explain inequality. Looking to U.S. trends [see FIG below], it is clear that the most prosperous family units (married with a wife in the labor force) are also declining as a share of all families, just as the least prosperous (single female head of households) are rising as a share of all families. Indeed the US is second in the OECD (to Sweden) in the shared of single-headed families.¹ Some estimates attribute as much as a quarter of the increase in overall inequality to the growing percentage of single-parent households.²

At the same time, social demographers have identified the rise in “assortative mating.” Much has been made of the fact that men and women are, over the last half-century, decreasingly likely to cross an educational or income barrier when they get married [see FIG above]. They are more and more likely to find themselves in homogomous marriages: in 1960 doctors married nurses; today they marry other doctors. This has tended to widen the gap between affluent dual-income families and all others. Some estimates attribute as much 10 or 15 percent of the increase in inequality (mostly between the lower and median incomes) to these shifting marriage patterns.³

We need to be careful about what is being observed or explained here, and the mechanisms by which demographic change might contribute to economic inequality. 

First, there is a certain tunnel vision about the demographic explanations. If single parenthood were the driving source of income inequality, why are the rates highest in the OECD settings [see FIG above] where inequality is severe (the United States) and where inequality is mild (Sweden)? And, by the same token, there is nothing to exceptional or distinctive about assortative mating in the United States—which falls in the middle of the OECD in terms of both the rate of homogomous marriage and its change over time. [see FIG below]. 

The differences, it turns out, are political rather than demographic. Against their OECD peers [see FIG below], two-earner families in the United States do fairly well (only 5 percent fall below the poverty line). But American families with only one worker—over a quarter of which fall below the poverty line—are at a much sharper disadvantage than similar families elsewhere.⁴ At the same time, the presence of kids (in the absence of policies like family leave and publicly supported child care found in much of the rest of the world) is a distinct economic burden in the United States. While childless American households have one of the lowest rates of poverty incidence in the OECD (at under 6 percent), households with children in the United States (nearly 20 percent of which fall into poverty) have one of the highest. It is not the rise of single-parent households that feeds inequality, it is the stigma attached to single motherhood—and the policies that flow from that--in the United States. ⁵

Second, these explanations also tend to confuse consequence and cause. Wages, income, and wealth are shaped by family structure but they also shape family structure. Those with lower incomes and lower educational attainment, for example, tend to marry earlier and divorce more often. Single motherhood is a reflection of inequality not a cause. And while greater homogamy (fewer individuals crossing educational or income boundaries in order to marry) might deepen income (household) inequality, such choices may themselves be driven by trends in wage (individual) inequality.⁶ 

And third, these explanations cannot account for—or glide by—the fact that most of the measurable inequality (and insecurity) is occurring within demographic categories. The gap is growing not between single moms and everyone else, but between (for example) some parents with kids and other parents with kids. Inequality and insecurity cut across family types—indeed most parents raising children on below-poverty incomes, are married. ⁷