Children who grow up in a home with one parent rather than two fare materially worse on almost every metric, which gives rise to the question of what the government could or should do about it.

Occasionally conservatives will assert — as Brad Wilcox and Chris Bullivant did in a recent USA Today piece — that this topic “cannot be uttered in our national conversation” and is “verboten” in elite circles. I don’t agree with that; I think progressives are simply skeptical that conservatives have any real ideas for promoting stable families and feel that the right typically brings this up to prevent the adoption of proven anti-poverty policies, like a child allowance.

We saw another troubling aspect of this show during the George W. Bush administration. He placed marriage promotion firmly at the center of his domestic policy agenda, and it turned out to be a mostly high-minded cover for homophobic demagoguery, with conservatives arguing that preventing gay and lesbian couples from getting married somehow defended the sanctity of marriage. Their efforts also included some very desultory pro-marriage campaigns aimed at low-income women memorably profiled in Katherine Boo’s article “The Marriage Cure,” one of the best pieces of policy journalism I’ve ever read.

There was just fundamentally nothing there, so as anti-gay rhetoric flipped from being a huge vote-getter in 2004 to an embarrassment by 2016, Republicans slinked away from the whole topic.

Meanwhile, the college-educated professional types who increasingly dominate American progressive politics do tend to lead neo-traditional family lives with no kids until after marriage and divorce being relatively rare. On the right, as the conservative base has shifted socio-economically, Republican Party politicians have become less inclined to be so vocal in their criticisms of “irresponsible” personal behavior — and Donald Trump is hardly a credible leader for a movement in favor of family stability.

And so the debate is kind of stuck.

Family structure seems to matter

Discussions about the impact of family structure on life outcomes are tricky in the United States because most white kids grow up living with two biological parents and most Black kids do not.

This leaves almost everyone disinclined to discuss the issue. One camp doesn’t like to talk about family structure as a source of disadvantage because they believe it detracts from the idea that any racial gap in life outcomes shows the need to fight racism. And another camp consists of racists who don’t like to talk about family structure as a source of disadvantage because they believe it detracts from the idea that Black people are inferior.

The Institute for Family Studies (which emphasizes the significance of family structure) often publishes charts like the one below that show a large race-independent statistical influence of family structure on child poverty.

With a little math, they show that for both Black and white kids, whether or not you grow up with two parents is a strong predictor of outcomes like future incarceration and college graduation as well as poverty.

For much of the 20th century, this conversation was dominated by conservatives arguing that people should stop complaining about racism because the real problem driving racial gaps in the United States was unmarried parents. But as the share of solo-parent families kept rising, driven by a decline in marriage rates among non-college whites, the subject became less racialized and therefore less politically convenient for Republicans.

And, of course, correlation is not causation. In this particular instance, we have so many correlations between income, education, race, and marital status that it’s extremely challenging to draw clear conclusions about the direction of causation.

Is this a causal relationship?

There are two basic problems with attempting causal inference about family structure.

One is that since married parents in America are richer, better educated, and more likely to be white, it’s not particularly surprising that their kids are better off. You could pick any attribute that is statistically associated with rich, well-educated white people and find that people like that have kids who end up with better life outcomes.

The second and larger problem is unobserved variables. You can apply statistical controls to obvious demographic characteristics, but even if you limit yourself to white, college-educated, 41-year-old dads, it’s probably the case that there are some differences on average between the WCE41YODs who are married to their children’s mother and those who are not. And it could be that those underlying personal differences are driving the differences in outcomes.

Imagine a world where there are only two kinds of dads: married dads and alcoholic abusive dads whose wives divorce them. It’s probably true that the divorced dads’ kids are going to be worse off than the married dads, and that would hold up even if you controlled for income, educational attainment, race, and other relevant variables. But does it follow that the kids in question would be better off still living with their abusive, alcoholic fathers? I’m skeptical! In the other ideological direction, a lot of progressives got mad last spring about a paper published in American Economic Review that looked at criminal defendants with children who were quasi-randomly assigned to harsher versus more lenient judges. Defendants who drew a lenient judge were less likely to end up in jail and could go back home to their kids. Bad luck and a harsh judge meant they likely went to prison, leaving their kids short a parent.

The paper found that “parental incarceration has beneficial effects on some important outcomes for children, reducing their likelihood of incarceration by 4.9 percentage points and improving their adult neighborhood quality.” And it’s not the only study to find this kind of result. Only a crazy person would conclude from this that you could improve the life of the average child by incarcerating her father. These studies look at a range-restricted sample of people convicted of crimes and show that in this population, the average impact of incarceration is positive. This is just to say that the human experience contains multitudes.

Some relevant research

There are in fact a lot of different questions you can ask:

• Would the average American child be better off if her parents got divorced? (Nobody seems to think this.)

• Would the marginal child of divorced parents — the ones who were closest to not breaking up — be better off if her parents had stayed together?

• Would the average child of divorced parents be better off if her parents had stayed together?

• You can then ask the marginal and average question about kids with never-married parents — would they have been better off with a shotgun marriage?

There are also a bunch of philosophically thorny questions you can ask about non-existence. Especially for children born to never-married parents, oftentimes the most plausible counterfactual is one where the child simply isn’t born — either thanks to abstinence or contraceptive use or abortion. I am a little bit annoyed that the researchers who’ve tried to work on this problem are mostly a bit fuzzy as to what counterfactual they are actually considering.

But there is some interesting research that, despite areas of disagreement, tends to push in the same direction:

• A Bush-era literature review by Mary Parke concludes that “up to half” of the disadvantage faced by the children of solo parents in educational attainment is attributable to lack of income, but that instability in living situation, attenuated relationship with the non-custodial parent, and weaker ties outside the immediate family (i.e., to aunts and cousins and dad’s broader social network) are also big factors.

• Kimberly Howard and Richard Reeves for Brookings think lower incomes for solo parents are responsible for about a third of the disadvantage measured in terms of how much money kids earn when they grow up. They say that a somewhat larger difference is accounted for by continuously married parents doing more intensive actual parenting and suggest we adopt policies to promote the “more engaged parenting of married parents.”

• Bjorklund, Ginther, and Sundstrom use Swedish administrative data and find that cohabitating parents vs. married parents make no difference. My understanding is that this finding does not generalize to the United States because cohabitating parents here are much more likely to split up. The difference (I think) is that in Sweden, there’s a strong tendency for secular people to not marry even if they’re very committed, whereas in the U.S. the marriage/cohabitation boundary reflects a commitment difference, and plenty of non-religious couples get married.

I think this research essentially confirms what common sense would tell you, which is that a second parent usually brings a lot to the table. Some of that is money. But some of it is that parenting, while delightful, is also difficult work, and kids benefit from parents’ ability to tag-team. The second adult generally also brings connections to a whole larger universe of adults whose acquaintance with the kid can be useful. The level of involvement of the second parent exists on a spectrum, but in practice, given the actual social conditions of the United States, dads are more involved with kids when they are married to their children’s mother and that brings benefits.

Again, I hasten to add that in extreme cases, this may not apply. Among child protective services investigators, some are more aggressive in removing kids from potentially abusive homes than others, and researchers find that more aggressive removals are beneficial — especially to girls. As with the parental incarceration finding, we see that some parents are really bad and “absent father” is not the worst possible outcome. But at the margin, it does seem like it would be useful to encourage marriage.

What can we actually do here?

The most obvious pro-marriage move would be to remove marriage penalties in the welfare state.

A lot of programs are means-tested in an effort to maximize the alleviation of material suffering while minimizing expense. But because marriage itself alleviates material suffering, programs are often structured such that a mom with two kids who earns $20,000 per year gets less help than a mom and a dad with two kids and combined earnings of $40,000 per year. There is a certain logic to that, but it discourages people from getting married. The most straightforward solution is to consolidate and simplify these social assistance programs while also making them more generous overall. Universal benefits do not create marriage penalties.

But conservatives dislike spending money in general and specifically dislike spending money on programs that benefit people who don’t or can’t work. So Angela Rachidi of AEI recently devised a program to consolidate and reform the EITC and Child Tax Credit into a single Working Family Credit. Her plan would, by her own admission, do almost nothing to reduce child poverty. The idea instead is to reduce the implicit marginal tax rates on the poorest families while essentially redistributing the tax burden from single parents in the top 40 percent of the income distribution to below-average-but-not-the-poorest married parents.

I think this seems too penny-ante, but I wouldn’t object if Republicans wanted to do it.

For the Social Capital Campaign, Wilcox outlines a five-point plan to promote marriage. One point largely echoes Rachidi on marriage penalties, but he also calls for:

1. More career and technical education (CTE) programs in public schools

2. School choice and education savings accounts

3. A national campaign to promote the “Success Sequence”

4. A wage subsidy program

We can debate the merits of (1) and (2), but they are education policy controversies. People disagree about the best approach to CTE and school choice, but that’s because they disagree about what is going to have the best impact on kids’ education. The wage subsidy idea strikes me as a good one, though again the connection to marriage per se is tenuous. His point is that a wage subsidy “would not phase out as workers increase their hours, since it is calculated from hourly wages and not total income. So, the program could incentivize more work hours for men on the lower end of the wage scale.” That is definitely true, but it is not obvious to me that it would increase marriage rates.

And so we reach the “Success Sequence.”

Who’s afraid of the Success Sequence?

The Success Sequence has launched furious online wonk battles, but basically it says that if you want to do well in life, you should:

1. Graduate from high school

2. Maintain a full-time job or have a partner who does

3. Have children if and only if you are married and over the age of 21

For all the fighting about this topic, it seems like clearly good life advice.

What’s less clear is what actually follows from the fact that this is good life advice. Sometimes conservatives say that because it’s good advice, government and society ought to state the advice more clearly — that’s Wilcox. But then you have Rich Lowry saying that Ta-Nehisi Coates should stop complaining about racism because if all Black people followed the Success Sequence, their average economic outcomes would be better. That Lowry column makes it clear why progressives are reluctant to run around singing the praises of the Success Sequence — we are afraid it’s just a thin conservative pretext to ignore racism and gut the welfare state!

But I’m also not really sure what conservatives want the Success Sequence propaganda campaign to look like. Oftentimes they observe the fact that college-educated liberals have very low rates of unmarried parenting and urge American elites to “preach what they practice.” But is that actually what conservatives want?

I think we are hung up on the Sexual Revolution Trilemma where you can have two out of the three of “no shotgun weddings,” “no unmarried parenting,” and “no abortions,” but not all three.

The ethos that liberal elites practice but don’t really preach that aggressively is “use contraceptives, but if that fails for whatever reason then get an abortion rather than derailing your life plan.” I’m very certain that if conservatives wanted to start preaching this message, liberal cultural elites would be happy to join them. But they don’t want to preach that message — they want to preach a different message where we eliminate both abortions and unmarried parenting by returning to the stigmatization of premarital sex and re-establishing a norm that if you did have sex before marriage and a pregnancy arose, the man would be expected to “do the right thing” and marry his paramour.

Years ago, Janet Yellen and her husband George Akerlof wrote an article positing that effective birth control plus legal abortion had undermined these norms, creating a situation where even though it was easier than ever to avoid pregnancy, it was now also socially acceptable for an unmarried woman to become pregnant, for the father not to marry her, and for the woman to carry the baby to term.

Conservatives like this paper and consider it embarrassing to liberals — here’s Ashley McGuire’s gloss on it:

Undoubtedly there are plenty of Democrats who wish they could “disappear” an article that Yellen co-authored with her husband George Akerlof for The Quarterly Journal of Economics in 1996. An adaptation of the article, entitled, “New mothers, not married: Technology shock, the demise of shotgun marriage, and the increase in out-of-wedlock births,” was featured by the Brookings Institution and is a relic of a day when, despite increasingly liberal social policy, honest discussion of its impacts was not censored, labelled with trigger warnings, or a cause for “cancellation.” It’s almost unthinkable today that a progressive would criticize the impact of abortion, contraception, and the concurrent rise in single motherhood on family stability, and consequently, the economy. And yet that is exactly what Yellen and her husband did, and their warnings have proven devastatingly accurate. 

But Yellen is not criticizing legal abortion! She says rolling back the technological changes is “probably not possible” and “would almost surely be counterproductive,” and “efforts should be made to ensure that women can use the new technologies if they choose to do so.” And just earlier this month she testified about abortion in a Senate Banking Committee hearing. Yellen is doing the thing that conservatives accuse educated liberals of not doing: preaching what we practice, that both contraceptives and legal abortion are good and women should have them available, while also acknowledging that individual autonomy is an important value.

A difficult problem

Ultimately, I absolutely agree that we should take a hard look at our social assistance programs and rework them to eliminate marriage penalties, but I’m not convinced that this is a high-leverage tool for promoting marriage. And I’m very skeptical that the other ideas on the list are.

Comparing 1990 to 2020 in the United States, wages went up, educational attainment went up, life expectancy went up, and all-around things got better. But the rate of unmarried parenting rose because, I think, in a more prosperous society people become more individualistic and less dependent on each other. Conservatives seem to be promoting magical thinking about the ability of liberal elites to radically alter society if we would just agree that kids are better off, on average, when raised by two parents. So I will say it: kids are better off, on average, when raised by two parents.

I’m just not 100 percent sure where that leaves us. If unmarried women become more aware of this fact (but is this something people don’t know?), that doesn’t magically conjure up husbands. And causing possible children not to exist doesn’t benefit the hypothetical children who are now not born. So what are we actually achieving down this route? Let’s fix the marriage penalties and hope for the best!