Cohabitation dustup followup thoughts

Just a few roundup thoughts and graphs on the cohabitation-is-bad-for-kids thing.

First, W. Bradford Wilcox sent me an email reply, which he asked me to post, and I have added it to the comments section of yesterday’s post, here. (It doesn’t include complete references to the passages he quotes or research he cites, which is a violation of my blog style requirements, so I’m not putting it in its own post.) Some of these comments are inspired by that reply, and some are just further thoughts.

Who are these cohabitors?

Host Pat Morrison, on the radio show with Wilcox and Stephanie Coontz, asked, “Are you looking at long-term relationships, Mr. Wilcox, or more transient relations, when you look at these unmarried couple figures?”

Wilcox, to his credit, does sometimes highlight this distinction, but it is too often lost. For example, on the show he mentioned the names of R. Kelly Raley and Shannon Cavanagh. It looks to me like he meant this Raley paper and this Cavanagh paper, both of which studied cohabiting step-family arrangements, not those in those in which biological or adoptive parents cohabit instead of marrying.

To clarify, there are two different family situations going under the name of “cohabitation” here:

1. Some parents have children (through procreation or adoption) without being married. These are sometimes called “unmarried parents.” In 2009, 69% of children lived with “two parents,” including 4% who lived with “unmarried parents.”
2. Sometimes people (usually women) who already have children move in with a man to whom they are not married. These could be thought of as step-cohabitating arrangements — but that cohabiting partner may or may not “parent” those children, and from a distance you can’t assume they are “cohabiting parents” or “unmarried parents.” In Census lingo, they are a single parent and that parent’s “unmarried partner.” In 2009, 27% of children lived with one parent, and 10.3% of them — or less than 3% of all children — live with a single parent and that parent’s partner. Of course, because these are relatively unstable relationships, more than 3% of children will have this experience at some point in their lives.

Anyway, these two “cohabiting” situations are very different. The New York Times Room for Debate on this was titled, “Should Parents Marry for the Kids?”, which seems like a reference to the first scenario. In fact, the URL for the item at NYT included, “shotgun-weddings-vs-cohabitating-parents,” as if the issue is about whether couples should marry when they are having kids. The great majority of the research in this area is not about that issue.

For example, the report that Wilcox cited in the NYT debate does show higher rates of abuse among children whose biological parents are living with a partner, controlling for basic demographics. It is not surprising that children are at greater risk of abuse when they live with unrelated adults (usually men), who are not committed to them as parents. Does that mean these cohabitation situations “cause” that abuse, or might there be, as I and others suggested, a selection mechanism? Contrary to what Wilcox suggested in the NYT piece, we can’t say from this research, but I am concerned that these families are experiencing tumult, uncertainty and insecurity on a scale unfamiliar to the majority of two-bio-parent-married families with whom they are being compared in these analyses. That is, the women who actually face the issue of whether their boyfriend should move in with them and their children do not have lives that are otherwise similar to the relatively low-risk reference groups in the research. In fact, even simple statistical controls for education, race and income, for example, are unlikely to capture the life experience and history of these mothers.

Single parents and crime

At the end of yesterday’s post I said that, “Today’s cohabitation-causes-bad-parenting is the 2010s version of the single-mothers-and-welfare-cause-crime hype from the 1980s.” Wilcox’s response to that is:

Evidently, Cohen believes that single parenthood per se has nothing to do with crime. Say what? What about Harper and McLanahan (2004), LaFree et al. (2010), and Sampson et al. (2005), among others? What does Cohen make of Sampson’s (1995) observation that “family structure is one of the strongest, if not the strongest, predictors of… urban violence across cities in the United States?”

This actually convinces me even more of what I said (not what he says I said). I don’t know what causes crime, but I do know something about hype. That Sampson quote, for example, seems to be from a chapter in a 1995 book, in which he analyzed the pattern of crime rates across large cities in 1980 as a function of family structure and other variables. That analysis, besides being 30 years old (see below), is a perfect example of something that can’t show a causal effect of family structure on crime — besides a few demographic controls, things that cause (or are correlated with) single-parent families could be causing the crime. The same holds for the LaFree et al. 2010 paper, which I believe is this: “Still Separate and Unequal? : A City-Level Analysis of the Black-White Gap in Homicide Arrests since 1960.”

But on the hype issue: I hope many of you readers are too young to remember the 1980s, when (Black) single parenthood was the bogeyman behind the crack-homicide craze. So, let me ask conservative marriage people this: Why aren’t we still fighting that single-parenthood battle? Are single mothers gone? Of course not.

Two reasons come to mind: first, crime rates collapsed even though single parenthood didn’t; and second, the 1990s welfare reform used punitive economic sanctions to try to discourage single parenthood, and failed completely. Here are the trends.

Up until 1991 or so, it sure looked like single-parent families were the harbinger of a collapsing civilization:

Sources: Violent crime from Uniform Crime Reports, single mother families from U.S. Census Bureau.

In fact, from 1960 to 1991, the trends for violent crime rates and single-mother-headed families were correlated at .95! And then, in what can only be described as one of the greatest trend-correlation reversals of modern times, the bottom fell out from under violent crime, plunging American back into the light ages of the early 1970s — while more and more families continued to come under the reign of unmarried women. The correlation since 1991 is reversed: -.51. That’s something.

For what it’s worth, the same pattern holds if you use births to unmarried women instead of living arrangements of children. In this figure the blue line for violent crime is the same. Same story: family continues heading toward hell-in-a-hand-basket, yet peace now guides the planet.

Sources: Violent crime from Uniform Crime Reports, single mother births from these: a, b, c, d.

Maybe single-parenting does have a causal effect on crime — I’m not saying it doesn’t, and there reasons it might (such as lower levels of supervision). But what these trends show is that it was possible to reduce crime drastically without reversing the family structure tends.

How’s that change-family-structure-through-policy thing workin’ out for ya?

Finally, I argued previously that even if marriage is giving some people an advantage, I’d rather work to undermine that advantage (or, the disadvantage it implies for everyone else) than change marriage behavior.

In fact, the social engineers of traditional family salvation have already taken a very big whack at trying to redirect family structure trends, and it hasn’t worked. Remember welfare reform? Remember this stuff from the 1990s?

Welfare also plays a powerful role in promoting illegitimacy …  Being born outside of marriage and raised in single parent homes … doubles the probability a boy will become a threat to society, engage in criminal activity, and wind up in jail. … Steps must be taken to reduce future illegitimacy, beginning with restricting cash welfare to unmarried teen mothers.

Well, they got what they wanted. They kicked millions of families off welfare while the number in poverty remained virtually constant, ripping away that toxic incentive to have children in an unmarried state. And?

Source: TANF data from here.

I hope it was worth it.

Addendum: Paul Krugman today is thinking along similar lines…

If you’re an intellectual of a certain age, you remember that in the 80s and maybe a bit of a way into the 90s it was common on the right to see American society as being in a process of catastrophic moral decline, descending into social anarchy. Crime would continue to rise, chaos would continue to spread, until and unless we returned to the Victorian virtues — and more specifically, to Dickensian social policies, in which only the deserving poor — as so designated by faith-based charities — received help. … But then, in the 90s, a funny thing happened: in many ways, American society began healing. True, out-of-wedlock births continued to rise, although at a much slower pace. But crime plunged, and in general our society began to look a lot more functional… [OK, as a card-carrying sociologist I wouldn’t say “functional”… -pnc]

h/t Neal Caren for the link.