Photo courtesy of Flickr user rosmary, under a Creative Commons license.

We know that children raised by two parents tend to be more successful — at school, in the future labor market, in their own marriages — than children raised by a single mom or dad. And from this fact, it might seem easy to conclude that marriage wields some outsized power over a child's life — that its absence creates unstable homes and chaotic families, while its presence nurtures them.

In reality, though, the question of why children of married parents are more likely to thrive is an extraordinarily complicated one. From a new analysis by Kimberly Howard and Richard V. Reeves at the Center on Children and Families at the Brookings Institution:

Is it simply because they have, on average, higher family incomes? (Two earners are better than one, and one household is cheaper to run than two.) Or are two committed spouses better able to provide consistent parenting? Is it marriage itself that matters, or is marriage the visible expression of other factors, that are the true cause of different outcomes? And if so, which ones?

Parents who marry differ from parents who don't in many ways beyond the marriage itself. Today, better-educated, higher-income adults are much more likely to marry. That means their children benefit from the marriage, and the income, and the education of their parents. Howard and Reeves also point out that the same skills that make marriages work (like commitment and patience) also come handy for good parenting. And so perhaps it's not that children are better off when their parents marry — it's that the qualities that enable successful marriages also make good parents.

Among all of these factors, it's not easy to tease out what matters most. But the answers (as best as we can identify them) are crucial for public policy. If we believe that marriage itself is what matters for children, then we'd want to encourage parents to marry. If we believe it's the financial stability that matters, then we'd want to find ways to bolster the income of single parents outside of marriage. If we believe it's the good parenting skills so often present in married households that make the difference, we could try to instill those skills in parents regardless of whether they have spouses.

In their analysis, Reeves and Howard offer a large part of the answer. By their calculation, children whose mothers are continuously married grow up to make higher incomes at age 40 than children raised at some point by single parents. The difference amounts to about 14 percentiles in adult income rank (children with married parents grow up to make, at age 40, in the 57th income percentile, compared to the 43rd). How much of that difference might be attributable to factors other than — and perhaps obscured by — the marital status of their mothers?

Their analysis uses a model based on data from the National Longitudinal Survey of Youth. When they controlled for the income differences of married and single-parent households, the age-40 income gap shrank by 5 percentiles:

Reeves and Howard

Two-parent households don't just tend to have more money (which they might spend on tutors, museums, books or simply better health care and groceries). They also have more time (which they might spend on homework help, library visits and bedtime reading). Add the time factor to the parenting qualities I mentioned earlier (patience, commitment), and it's possible that part of the marriage effect is really a "parenting effect": Children with married parents also have more engaged parents, and it's the engagement that really matters.

When Reeves and Howard controlled for a measure of parenting based on home observation and self-reported behavior, that 14-percentile difference shrank even more dramatically, to 7.5 percentiles. "Parenting" here covers activities like regularly reading or eating meals with children.

Reeves and Howard
Reeves and Howard

When Reeves and Howard controlled for parenting and income at the same time, along with a few other characteristics like race and the age of the mother, that 14 percentile difference shrinks down to a little more than four percentiles. Reeves is confident the gap would shrink even further if researchers had reliable data to account for other factors, like the socioeconomic status of neighborhoods where children grow up, or the quality of schools they attend. At the end of the day, marriage itself might still have some effect on the adult outcomes of children. But it would be a small one.

Parenting skills and income levels are no doubt closely related, so we can't simply add the two effects together in the above analysis. But it's clear here that parenting skills and income levels drive much of the difference we often more simply attribute to marriage itself.

"Those two factors taken together explain most of the better outcomes for the children of married couples," Reeves says. "Not all. But most. And I think the takeaway here is not to mistake a commitment device – which marriage is – for an explanatory device."

Making single parents get married, in other words, won't fundamentally change the other characteristics about them that really drive their children's success. The good news in this is that family income and parenting skills are more realistically addressed through public policy than marriage anyway.