Bench Memos
Advocates of Same-Sex Marriage Are Wrong to Claim It Poses No Risk to the Marriage Ecosystem
The pro-same-sex marriage blogosphere has reacted with furor to a friend-of-the-court or “amicus” brief that I recently filed in the Supreme Court marriage case on behalf of 100 scholars of marriage. That brief concludes that a decision forcing same-sex marriage on unwilling States would likely harm the children of heterosexuals by, among other things, reducing marriage rates among opposite-sex couples.
In the end, though, critics such as those writing in the Washington Post’s Wonkblog and the Huffington Post do not really dispute the keystone of the scholars’ brief: that same-sex marriage will fundamentally alter the institution of marriage by undermining or destroying several social norms attached to man-woman marriage. And their empirical criticisms of the scholars’ predictions are misguided.
Marriage as an Ecosystem
As these critics surely understand, the institution of marriage is similar to a natural ecosystem, which can be substantially changed by seemingly small things. A degree or two higher annual temperature, or a moderate decrease in water flows, can have huge long-term effects. Social institutions are no different, and marriage is society’s most important social intuition focused on the welfare of children.
As with environmental changes, moreover, we often do not intend or even foresee the consequences of changes to a social institution. For example, no-fault divorce, which was implemented with the best of intentions, substantially altered the institution of marriage, turning it more adult-centric and undercutting the traditional norm of permanency in marriage. No longer was it “till death do us part,” but instead “till we grow tired of each other.” And children and women have suffered the most from the “new,” non-permanent marriage institution that resulted.
Normative Effects of Removing Marriage’s Man-Woman Definition
Similarly, while no one can foresee the future, institutionalizing same-sex marriage has the potential to do to the institution of marriage what no-fault divorce did—or worse. Whereas no-fault divorce undermined just one norm of marriage—permanency—the genderless redefinition of marriage that is legally necessary to accommodate same-sex couples undermines at least five norms of marriage: biological bonding, gender diversity, the postponement of procreation, the value of begetting and rearing children, and partner exclusivity. It also undermines the more fundamental norm that marriage is principally about the welfare of children, and only secondarily about the wants and needs of adults.
With these norms gutted or gone, marriage will not be the same. Even same-sex marriage advocates realize this. As Lambda Legal Defense Fund attorney Tom Stoddard declared a quarter of a century ago: “enlarging the concept to embrace same-sex couples would necessarily transform [the institution of marriage] into something new.” And by undermining traditional marital norms, that “transformation” of marriage from an inherently gendered to a genderless institution would significantly increase the risk that those on the margins of marriage—who are most in need of the social reinforcement these norms bring—will increasingly choose to forego marriage altogether.
Concrete Effects of A Redefinition
If that happens—and these 100 experienced marriage scholars believe it very likely will—the consequences can be estimated in what demographers call decomposition analysis. The logic is simple: As fewer women marry, more will either cohabitate or live singly, but still have intimate relations. Given the differing birthrates and abortion rates between married and unmarried women, three results would follow:
- Fewer children would be born
- More children would be born out of wedlock, and functionally fatherless
- More children would be aborted
The last observation is based on census data showing that unmarried women have an abortion rate nearly five times higher than that of married women. And data from the pro-abortion Guttmacher Institute shows that the abortion index—a group’s relative abortion rate calculated by determining that group’s proportion of total abortions compared to that group’s proportion of the total population—is ten timers higher for cohabitating women than married women, and over three times higher for women living singly than married women.
The Critics’ Response
The amicus brief’s critics don’t dispute these three consequences of decreased opposite-sex marriage rates. What they dispute—consistent with same-sex marriage advocates’ narrative that “my marriage won’t affect your marriage”—is that transforming marriage from a gendered to a genderless institution actually poses a significant risk of depressing opposite-sex marriage rates. They then make two rhetorical moves.
First, they mischaracterize these 100 scholars as claiming all of this will come to pass. In fact, all the brief states is that it is likely—even very likely—that all three of these things will happen. But no one has a crystal ball, least of all same-sex marriage advocates. And to say these things will not occur is to slip into the very error that the critics mistakenly lay at the scholars’ doorstep.
Second, the critics attack the data by either invoking other studies, bringing in new data, or downplaying the data the scholars cite. All three of these responses are flawed.
Other Studies
One study mentioned by the critics is from two researchers at Portland State University. But that study included no control variables; instead it looked only at marriage rates and whether the state had same-sex civil unions or marriage. And that doesn’t tell us anything; indeed, it’s the statistical equivalent of putting on blinders. Further, the study looked only at data through 2009—a time when only four states had same-sex marriage, with three of those adopting it only that year or late in the year before. As a result, there were not enough observations to generate sufficient statistical power to detect differences. Does anyone really think one or two years of data is enough?
Another study cited by the critics is a study by Marcus Dillender—which is also addressed in the amicus brief. Dillender was looking at the data at a time when there had been only a few years of same-sex marriage and thus not enough statistical power to detect change. Worse, Dillender didn’t look at states individually, but instead attempted to look at effects in the entire United States at a time when only a handful of jurisdictions had adopted same-sex marriage, and most still overwhelmingly maintained the man-woman definition. No wonder his study found no effects.
A third study looked at opposite-sex marriage rates and divorce rates in five European countries that had adopted same-sex civil unions or marriages. The problem with this study is that the data stops after 2002. And at that point only one country—the Netherlands—had adopted same-sex marriage, and then only for one year! Of course no impact was found.
As described in the amicus brief, the one study to correct these methodological and statistical shortcomings was done by Mircea Trandafir on the country that has had same-sex marriage the longest: the Netherlands. That study, using sophisticated statistical controls, found that after the introduction of same-sex marriage, opposite-sex marriage rates among young women aged 18-22 (the only age group studied) dropped five percent across the country compared to the rate of decline beforehand. The study also found a comparative decline of 13.4 percent across the country for native young women, and a 31.8 percent comparative decline for young women in the four largest—and least religious—urban areas.
The critics characterize these reductions as “a small drop.” But as explained above and in the amicus brief, even a small drop in opposite-sex marriage rates will have huge ramifications in the form of lower fertility rates, higher numbers of functionally fatherless children, and more abortions. The 100 scholars’ decomposition analysis in Appendix B of their brief was based on only a five percent decline in man-woman marriage rates—which is at the low end of the post-same-sex-marriage reductions seen in every U.S. state and European nation for which we have sufficient data. And that analysis predicts that over a generation, a forced redefinition of marriage would product nearly 600,000 additional functionally fatherless children and 900,000 abortions.
The critics also point out that the Netherlands study didn’t find such declines when the country a few years earlier adopted same-sex civil unions—which the critics claim are the functional equivalent of same-sex marriage. But that is misguided: Unlike same-sex marriage, a civil union regime doesn’t require a change in the definition of marriage or, therefore, its fundamental transformation from a gendered to a genderless institution. And the Netherlands data demonstrate that the two are not functionally equivalent.
Finally, while admitting “the Netherlands study is a very good study on its own,” the critics discount it as “just one study.” True. But it’s the study with the best methodology, and the study that has looked at the nation that has had same-sex marriage the longest. It therefore should be given a lot of weight—especially since it accords with the more limited experience of U.S. states and that there simply are no credible studies pointing the other way.
New Data
The critics also counter the observation that man-woman marriage rates have declined in states that have redefined marriage in genderless terms by pointing out that marriage has also declined in states that have not adopted same-sex marriage, like Texas and Utah. That depends on the state, and how far back one goes. For example, since 2008, Texas has had no change in its man-woman marriage rates, whereas Vermont and Iowa have dropped by 5.1% and 9.2% respectively. That may not seem like much, but imagine the impact on the Earth if global temperatures increased by five or nine degrees? Or even three?
It is also true that national marriage rates have been dropping for decades because of a mixture of increased urbanization, education, secularization and other cultural changes. However, as the brief explains, during the past four years the national rate has stable even as marriage rates in the same-sex marriage states discussed in the brief have continued to decline.
The critics also dispute the amicus brief’s prediction of increased abortions, arguing that abortion rates have been declining in the United States while the number of same-sex couples has increased. But their argument rests on at least two fatal errors: First, they appear to be naively assuming an instant impact: A same-sex couple marrying today means some woman will abort tomorrow. But the 100 scholars’ warning was that abortions would increase over the next fertility cycle—a 30-year window. Second, the critics end their analysis in 2011, when only a handful of states had legalized same-sex marriage. Cutting off analysis prematurely is an easy way to find no effect.
Criticizing the 100 Scholars’ Data
The 100 scholars also pointed to Spain, noting that after adopting same-sex marriage, opposite-sex marriage rates plunged 36 percent. The critics counter that that if one goes back to 1960, Spain’s marriage rate has been dropping significantly for fifty years. True enough. But in the nine-year period before Spain redefined marriage, after this long decline in marriage rates, opposite-sex marriage rates in Spain were actually on the rise. Then they fell off a cliff after the adoption of same-sex marriage.
The critics also take issue with the Massachusetts data, noting that its marriage rates had been in decline before it adopted same-sex marriage. But what is significant is that Massachusetts’ marriage rates continued to decline, and did so substantially even at a time when overall U.S. marriage rates had stabilized. Moreover, when one breaks down the Massachusetts data after the adoption of same-sex marriage, just like in the Netherlands, the more religious Massachusetts counties experienced only a 3.3% decline in opposite-sex marriage rates, whereas the less religious counties suffered a 16.8% reduction. That experience reinforces the concern that a redefinition of marriage will substantially weaken marital norms in communities or sub-populations that lack strong alternative voices—like religion—that continue to promote those norms once the state has abandoned them by removing the man-woman definition.
Conclusion: Forcing a Redefinition of Marriage Would Pose Enormous Risk to Children
Admittedly, as the amicus brief acknowledges, correlation is not causation. But the strong correlation between the adoption of same-sex marriage and reduced opposite-sex marriage rates points to an enormous potential risk to children. And the correlation is even more alarming when it is consistent with theoretical predictions (the undermining of norms), past history (the impact of no-fault divorce), and causal analysis (the Netherlands study). The critics can point to nothing in theory, history, or sound empirical studies to refute the correlation shown in the United States data—much less the causality established by the Netherlands data.
For all these reasons, the Supreme Court would be wise to avoid forcing unwilling States to redefine a key component of the marriage ecosystem—especially when all signs point to a host of negative outcomes for children.