The Volokh Conspiracy
Mostly law professors | Sometimes contrarian | Often libertarian | Always independent
In previous posts, I've criticized the Sixth Circuit's analysis of animus (see here) and its use of originalism (see here) to reject equal-protection claims for same-sex marriage. (You can read the Sixth Circuit's full opinion in DeBoer v. Snyder here.) In a future post, I will respond to the court's rejection of heightened scrutiny of laws that exclude gay couples from marriage.
In this post, I want to lay out concerns with the Sixth Circuit's analysis of what should be the easiest marriage-limitation hurdle for the states to jump, the requirement that a law need only be rationally related to a legitimate governmental objective ("rational-basis review"). Among many others, supporters of same-sex marriage like Eugene have noted that opposite-sex-only marriage laws should be able to satisfy rational-basis review even if they are unconstitutional on heightened-scrutiny grounds. I've long thought that was probably right, though with every passing decision and brief defending the limitation on that basis I've become less convinced that the issue is so cut-and-dried.
My conclusion is that the DeBoer court framed the question of rational justification in a way that avoids the precise issue being presented. It asks whether there's a rational basis for including opposite-sex couples in marriage rather than asking whether there's a rational basis for excluding same-sex couples from it. This way of analyzing the rational basis for a law proceeds from what I call an inclusion framing rather than an exclusion framing. The approach inverts the analysis and produces the correct answer to the wrong question. DeBoer doesn't actually respond to the constitutional challenge before the court.
Of the two legitimate interests the Sixth Circuit offers for opposite-sex-only marriage laws, responsible procreation and Burkean caution, I will deal here only with the former. I may address the Burkean concerns in a future post.
I. Inclusion v. exclusion framing in rational-basis review
In every case where the state has made benefits or rights available to some and denied them to others, it is possible to frame the inquiry in at least two ways. The DeBoer court asked, does it rationally advance a legitimate state interest to provide marriage to opposite-sex couples? Under this approach, we measure the rationality of the law by reference only to those who are included, ignoring who's left out and why. This framing respects the state's ability to address problems piecemeal so long as the law at least advances the interest to some extent. Let's call this the inclusion framing of the rational-basis test.
But there is another way to frame the analysis that also respects the need to legislate incrementally. Suppose we ask the rational-basis question in DeBoer this way: does it rationally advance a legitimate state interest to exclude same-sex couples from marriage while conferring marriage on opposite-sex couples who are identically situated to same-sex couples with respect to that very interest? Under this approach, we care about who's included in the law and why, but we can't escape some consideration of who's excluded and why. That's the group whose very claim is under consideration. Let's call that an exclusion framing of the rational-basis test.
It's fair to say that there is at least some support in the case law for both ways of framing rational-basis review. But DeBoer is an especially poor candidate for inclusion framing for two reasons.
First, in a series of decisions touching on personal and familial decisions, the Court has most often used exclusion as the frame for its rational-basis analysis. Eisenstadt v. Baird; Moore v. City of East Cleveland; Cleburne v. Cleburne Living Center; Romer v. Evans; Lawrence v. Texas; United States v. Windsor. On an inclusion framing, each of the laws at stake in these cases could have been regarded as rationally serving a legitimate interest by including one group within legal protection that served the state's interest. The "included" groups would be married couples using contraceptives in Eisenstadt, traditional family units in Moore, multiple-person dwellings other than those for the cognitively disabled in Cleburne, racial and other groups needing anti-discrimination protection in Romer, heterosexual sexual partners in Lawrence (a due-process rational-basis case), and married opposite-sex couples in Windsor. (You could also expand that list to include transitional rational-basis cases like Reed v. Reed and Frontiero v. Richardson.)
To take one of these examples, the food-stamp program in Moreno certainly served the twin congressional goals of feeding poor people and boosting the agricultural economy, regardless of whether "hippie communes" were also eligible beneficiaries along with more traditional families. The law satisfied an inclusion framing.
But in each of these cases, the Court instead adopted an exclusion framing, asking whether there was a rational basis for excluding a group from legal protection and benefit. The excluded groups whose exclusion was irrational were unmarried people using contraceptives in Eisenstadt, households of unrelated persons in Moore, a dwelling for the cognitively disabled in Cleburne, those claiming legal protection from discrimination based on homosexual orientation in Romer, same-sex sexual partners in Lawrence, and married same-sex couples in Windsor.
Second, in each of the cases just mentioned, the constitutional litigation was directed at the exclusion of some-not at the inclusion of others. It made sense, therefore, to demand the rational basis for the exclusion, not the rational basis for the inclusion. It's the way the line was drawn, the classification itself, that counts: "Even in the ordinary equal protection case calling for the most deferential of standards, we insist on knowing the relation between the classification adopted and the object to be attained." Romer (emphasis added). More on this in the state marriage-law context in a moment.
The choice between inclusion framing and exclusion framing is not a choice between deference to the legislature and no deference to the legislature, or between allowing piecemeal reform and not allowing it, or between "ordinary" rational-basis review and "rational basis with bite." These are false choices. Both inclusion and exclusion framing are highly deferential to state legislative choices, allowing substantial under-breadth and over-breadth. But as the Supreme Court has repeatedly demonstrated, highly deferential does not mean absolutely deferential.
II. The Sixth Circuit's inclusion framing in DeBoer
Marriage, the Sixth Circuit argues, is anchored to the biological reality that heterosexual sex may result in pregnancy. Judge Sutton actually refers specifically to the concepts of "nature's laws" and sexual "complementarity," language that is common in so-called natural-law writing. According to DeBoer, this rationale for opposite-sex marriage
starts from the premise that governments got into the business of defining marriage, and remain in the business of defining marriage, not to regulate love but to regulate sex, most especially the intended and unintended effects of male-female intercourse. . . .
Once one accepts a need to establish such ground rules, and most especially a need to create stable family units for the planned and unplanned creation of children, one can well appreciate why the citizenry would think that a reasonable first concern of any society is the need to regulate male-female relationships and the unique procreative possibilities of them. One way to pursue this objective is to encourage couples to enter lasting relationships through subsidies and other benefits and to discourage them from ending such relationships through these and other means.
This is a form of inclusion framing, which goes like this: Marriage for opposite-sex couples makes sense. Making this status and accompanying benefits available to them rationally advances the government's interest in dealing with the high social consequences of heterosexual irresponsibility. Judge Posner's opinion for the Seventh Circuit characterizes this "responsible procreation" argument in a much less flattering way, but it's the same thing.
As the Sixth Circuit notes, it doesn't matter for the rational validity of extending marriage to opposite-sex couples that the modern understanding of marriage as recognizing and encouraging commitment supports extending it to same-sex couples. It also doesn't matter for this analysis that the Sixth Circuit's explanation of marriage as evolving from "nature's law" is very far from a complete historical understanding of why marriage developed and why the state has been involved. The need for responsible procreation is enough to explain why we extend marriage to opposite-sex couples.
The inclusion framing would be the crucial question if someone were so bold as to challenge the constitutional validity of all opposite-sex marriages. The inclusion framing would also be relevant if some litigant challenged the extension of marriage to non-fertile heterosexual couples on the grounds that allowing them to join an institution designed to deal with heterosexual procreative potential was over-inclusive. If there were a challenge to the constitutional validity of including non-procreative heterosexual marriages, the answer under an inclusion framing would surely be that their inclusion is a permissible dash of over-inclusiveness.
The problem for the Sixth Circuit's inclusion framing is that the validity of marriage for heterosexual couples was not at all the issue in the cases before it.
III. An exclusion framing of DeBoer
There is an alternative. The Sixth Circuit could have asked whether it was rational to exclude same-sex couples given the state's legitimate procreation interest and its decision to include opposite-sex couples who cannot procreate.
Consider that we have three kinds of couples who might qualify for marriage: (1) fertile opposite-sex couples, (2) infertile opposite-sex couples, and (3) infertile same-sex couples. If the state's interest in marriage is channeling procreation in a responsible way, it would make sense to draw the line between the first and second groups. But once the second group is admitted, what sense does it make to draw the line between the second and third groups, who are identically positioned with respect to the state's interest in procreation?
Something must give. Either the state has additional interests in promoting marriage that justify extending it to the second group despite the fact that they can't procreate (like encouraging stability in people's lives and assisting them in raising their non-biological children), in which case there isn't an obvious reason to fence out the third group (because, as the Sixth Circuit agrees, they can serve all of the non-procreation interests in stability and child-rearing). Or there must be some further reason beyond procreative potential for drawing a line between the second and third groups. It's not enough, even on rational-basis review, for the state to say, "we draw the line here because we want it that way." (And despite the Sixth Circuit's suggestion, it's also not enough, as I've explained before, that the line was drawn where it was in order to punish or to defang activist judges.)
But what might that reason be, and is it rationally related to any legitimate state interest?
IV. The Sixth Circuit's flawed answer
The Sixth Circuit suggests a couple of reasons beyond mere procreation for why a state can constitutionally include one set of couples that don't serve the state's procreative interest in marriage while excluding another. The first is hopelessness. "History is replete with examples of love, sex, and marriage tainted by hypocrisy," asserts the Sixth Circuit in its description of the inconsistencies of marriage law, as if a history of hypocrisy is its own justification rather than a reason to reject it.
The DeBoer court then brings up a familiar slippery-slope argument, cautioning that "[i]f it is constitutionally irrational to stand by the man-woman definition of marriage, it must be constitutionally irrational to stand by the monogamous definition of marriage." But it is not the case that there can be no rational limits on marriage, even under an exclusion framing of the issue.
Take the specific issue of group marriage. Polygamy raises all kinds of questions not raised by (straight or gay) dyadic marriage that a state might rationally take into account, including legal responsibility for children, complexified dissolution and inheritance laws, tax issues, property division, legal authority to act for a loved one, and so on. Adjusting these aspects of marriage law to accommodate groups of married people will not involve mere form-alteration or technical or administrative burdens. It will require hard policy choices. Polygamy also has a very long and often troubling history that might give pause regarding broader issues of gender equality, children's welfare, violence and instability in the resulting larger pool of unmarried men, and more. These differences between polygamous and dyadic marriages relate to history, experience, logic, and policy that have been extensively discussed (my take is here, along with links to others' views), even if lawyers for same-sex couples do not always seems eager to highlight them.
Whether the concerns about polygamy are ultimately persuasive is one thing, but they would at least provide a minimally rational basis for excluding group marriage. We may, despite the concerns and the historical trend against polygamy, one day accept it. But the debate about accepting it will not turn on whether we first accept gay marriage.
The closest the Sixth Circuit comes to rationally justifying the line between infertile opposite-sex couples and infertile same-sex couples is its general observation that rational-basis review allows government to make policy based on "rough correlations."
Several cases illustrate just how seriously the federal courts must take the line-drawing deference owed the democratic process under rational basis review. Massachusetts Board of Retirement v. Murgia, 427 U.S. 307 (1976), holds that a State may require law enforcement officers to retire without exception at age fifty, in order to assure the physical fitness of its police force. If a rough correlation between age and strength suffices to uphold exception-free retirement ages (even though some fifty-year-olds swim/bike/run triathlons), why doesn't a correlation between male-female intercourse and procreation suffice to uphold traditional marriage laws (even though some straight couples don't have kids and many gay couples do)?
For the Sixth Circuit, it's the "rough correlation" between heterosexual intercourse and procreation that suffices to explain why the state can rationally include non-fertile opposite-sex couples while shutting out non-fertile same-sex couples. But that rough correlation doesn't quite work in DeBoer in a way that supports a rational over- or under-inclusiveness. Here's why.
The citation to Murgia exposes a basic flaw in the Sixth Circuit's rational-basis review. There is (I'm sad to report) indeed a rough correlation for fitness purposes between those who are under 50 and those who are over 50. This particular rough correlation constitutionally legitimizes a mandatory retirement age of 50 that is clearly both under- and over-inclusive.
Murgia was correctly decided under an exclusion framing because the state excluded a group (those over 50 years old) whose age was rationally related to the very problem the government was concerned about (physical fitness) without at the same time exempting from mandatory retirement a group whose age raised the very same problem in the same way. Think about it this way: if the law in Murgia had mandated retirement only for those police officers over 50 who were left-handed it would have been a very different case.
Under an inclusion framing, such a distinction between aged left-handers and aged right-handers would have been rational because it would at least result in the discharge of more aged officers and would thus advance the state's interest in fitness.
Under an exclusion framing, the handedness distinction would not have been rationally related to the state's legitimate interest in the fitness of law-enforcement personnel unless there was some further rational basis for distinguishing between left-handed people over 50 and right-handed people over 50. Absent such an explanation, the handedness distinction would have been entirely arbitrary and thus unconstitutional.
If I'm right that the better approach to rational-basis review in same-sex marriage cases is exclusion framing, there is a big problem for the analysis in the DeBoer decision. There is no "rough correlation" for procreation purposes between opposite-sex couples who can't procreate (included) and same-sex couples who can't procreate (excluded). It's not as if the former can procreate a little better than the latter. They are identically situated and yet only one is burdened by reference to an asserted responsible procreation interest. Unless the state has some further basis for making a distinction unrelated to its legitimate procreative interest, the law fails. This is the rational-basis issue the DeBoer court never really confronts.