There is a wide class gap in non-marital childbearing, which has the potential to damage social mobility. A substantial research literature has shown that growing up in a two-parent, married home confers observable benefits on children. (The issue has been addressed many times on this blog, including here, here, here, and here). In a longer-form Brookings report, Richard Reeves and Kimberly Howard observe that “the benefits of marriage in terms of children’s outcomes and life chances seem clear…both the higher incomes and the more engaged parenting of married parents count for a good deal.” We build on this perspective by reporting on the variation in the observed benefits of marriage to children, using the term the “marriage premium for children” to refer to the difference in outcomes between children who are born to married versus unmarried parents.

The premium is not the same across all marriages

Not all marriages will benefit children equally. The average difference in outcomes between children from married and unmarried parents disguises considerable variation across circumstances. For a 19-year-old woman without a high school degree, having a child outside of marriage likely has different implications for her child than it does for a 25-year-old with some college, or a 35-year-old college graduate. Older and more educated unmarried mothers may be able to provide sufficiently for their children alone. The men who father children with younger, less-educated women may make not be able to make a sufficient contribution to improve a child’s long-term economic prospects significantly.

In a new paper released today by the NBER, “The Economics of Non-Marital Childbearing and ‘The Marriage Premium for Children,’” we document the relationship between family resources and the marriage premium for children. Importantly, “resources” should here be understood in a very broad sense, including income, but also time, maturity, stability, family and other networks, and all the varied inputs that parents use to raise children well. Age and education can act as proxies for resources, since older and more educated mothers tend to have higher levels of income, better networks, and other advantages to share with their children.

Specifically, we show that the gains from marriage (for a child) depend on (a) the mother’s own level of resources, (b) the additional resources that her marriage would bring to the household, and (c) how those additional resources translate into benefits for children. These factors, and the interactions among them, may generate heterogeneous marriage premiums.[1] For instance, even if a teen mother is married to the father of her child, their combined resources might not be sufficient to increase their child’s educational attainment. Similarly, a highly educated older mother might have sufficient resources as a single mother such that her child may not see much benefit.

The influence of marriage may also vary for different outcomes at different levels of resources. Perhaps the child of a well-resourced single mother would see little benefit from marriage in terms of clearing a low educational hurdle, such as finishing high school. But the story might be different when it comes to college graduation.

When would marriage help most? It depends.

Using data from the Panel Study of Income Dynamics (PSID), we estimate the difference in incomes and educational outcomes between children of married and unmarried parents as a function of the age at which a woman gives birth and her education level, controlling for gender, race, ethnicity, and year of birth. We find that the children of married mothers are more likely to complete college, with the premium increasing with each year of maternal age, up to the late 20s. For children whose mothers are in their early 30s, there is nearly a 20 percentage point gap in the likelihood of college completion between children born to a married versus unmarried mother (37 versus 17 percentage among this PSID sample).


The same pattern holds for maternal education; even among well-educated mothers, there are benefits associated with marriage in terms of their children finishing college. (We obtain similar results when we examine the likelihood of the child having family income at least 400 percent of the poverty line as an adult.) In terms of achieving reasonably high levels of education and income, then, it looks as though marriage brings benefits to the children of even quite well-resourced mothers. This may reflect the fact that in this case, the father/husband is bringing in higher levels of resources to the household. But it is important to note that household income in childhood is only partially responsible for these observed gaps—other resources matter, too.

What about less demanding outcomes? We also examine the relationship between marriage and high school completion, and find a different picture. The influence of parental marriage on a child’s likelihood of high school graduation is greatest for mothers giving birth in their early 20s. There is no difference in the high school graduation rate of children born to married and unmarried mothers either over the age of 27 or under the age of 18:


This suggests that at the low end of the resource distribution, the gains from marriage are not sufficient to increase substantially a child’s probability of graduating high school, but are unnecessary at the high end. For mothers in their early to mid-20s and those with a high school degree, marriage is associated with the largest differences in these two outcomes. We see a similar pattern for the outcome of avoiding poverty as an adult.

Marriage may help children, but can policy help marriage?

The so-called “marriage premium” for children depends critically, then, on both the level of resources that mothers and fathers bring to marriage, and on the particular outcome of interest. When it comes to graduating high school and avoiding poverty, the largest benefits appear to be for mothers in their 20s and those with high school degrees. So it is worrisome that these are precisely the groups among which non-marital childbearing has been rising most rapidly.

All of this, of course, raises the important issue of what role policy might play. At the very least, having a clearer sense of how marriage relates to children’s outcomes, and how that depends on circumstances, should help focus subsequent policy conversations, even if it leaves open the critical question of what might be done about it.


The authors did not receive financial support from any firm or person with a financial or political interest in this article. They are currently not officers, directors, or board members of any organization with an interest in this article.


[1] Our focus on resources sets aside the issue of partner compatibility and how well parents would get along if they were to marry. Certainly some marriages would be detrimental and would have a negative effect on children. The point of this analysis is to consider how the observed benefit of marriage on average masks differences across the resource distribution.