Why is the Defense of Marriage Act unconstitutional? Here are two propositions that United States v. Windsor might be thought to stand for:
(1) The federal government’s decision not to recognize state-sanctioned marriages of same-sex couples was an unconstitutional intrusion on federalism (a structural claim); or
(2) The federal government’s decision not to recognize state-sanctioned marriages of same-sex couples was an unconstitutional infringement on a substantive right, e.g., the right to marry (a liberty claim).
These are among the many plausible interpretations of Windsor, but some explanations are more plausible than others. After consuming several pages discussing the interests of the states in controlling family law, Justice Kennedy expressly states that the Court is not relying strictly on federalism. Surely a statement in a decision suggesting what it means should have some bearing on what it means. The Chief Justice, in dissent, thinks federalism is nevertheless critical to the result and would help to distinguish the case from one that involved a claimed constitutional right to state recognition of same-sex marriages. He might be right about that. But the Chief Justice’s explanation may also be more a hope about the limited consequences of an alternative Windsor than a reading of the actual Windsor (see Justice Scalia’s dissent).
As for the second proposition, the Court certainly mentions liberty several times. And the context is one in which the plaintiffs claim that “liberty” protects a right to have their marriages fully recognized by government. The Court sets for itself the task of deciding “whether the resulting injury and indignity is a deprivation of the liberty protected by the Fifth Amendment.” Slip op. at 19. It concludes that Congress “cannot deny the liberty protected by the Due Process Clause of the Fifth Amendment.” Op. at 25. But my sense is that reliance on the word “liberty” here is more a doctrinal necessity — this is a decision rooted in the Fifth Amendment’s Due Process Clause, which protects “liberty” against certain deprivations — than a substantive holding. As my friend and fellow SSM-supporter Neomi Rao notes in her inaugural guest post, if the states themselves did not recognize same-sex marriages the federal government would not be required to issue federal marriage licenses to same-sex couples. Marital recognition starts in the states, and it’s hard to argue that Windsor holds otherwise, though you could say that some principle in Windsor (perhaps “dignity”) should be used to force state recognition of such marriages.
That leaves us with a third possibility, according to Neomi:
(3) The federal government’s decision not to recognize state-sanctioned marriages of same-sex couples was an unconstitutional denial of the dignity that comes with a general right to recognition (a liberty claim).
Neomi argues that this holding would be problematic and unprecedented in its recognition of an unmoored right to liberty, connected neither to an independent affirmative nor negative right. The first time I read this interpretation of Windsor I was taken aback. Of all the ways we could characterize the rationale, I hadn’t even thought of this one. So I read it again, looking for this free-floating freedom that is independent of a substantive liberty like a right to marry. Certainly the Court emphasizes the demeaning nature of DOMA for same-sex couples. There is an affront to their dignity as validly married when their federal government tells them it will not recognize their marriages. At one point, in the most striking single word in the entire opinion, the Court says that DOMA “humiliates” their children. But this dignity is not something they enjoy because they have a general right to recognition; they have it because the state has chosen to confer it. They “are deemed by a State entitled to recognition and protection to enhance their own liberty.” Op. at 25 (emphasis added). Similarly, Lawrence v. Texas pointed out the dignitary harm done to gay people by a “homosexual conduct” law, but the dignitary harm was connected to the substantive liberty of adult sexual autonomy. I do not see any single sentence, or paragraph, in Windsor declaring an independent “right to dignity” or “right to recognition” that DOMA has trampled. So I’m still puzzled why we should think that’s what Windsor stands for, even before I can get agitated about why such a declaration would be troubling.
Here are two more ways to characterize the proposition embraced by Windsor:
(4) The federal government’s decision not to recognize state-sanctioned marriages of same-sex couples was an unconstitutional intrusion on the state’s ability to protect the liberty of its citizens (a hybrid claim based on liberty and structural federalism principles); and/or
(5) The federal government’s decision not to recognize state-sanctioned marriages of same-sex couples was an unconstitutional denial of equal protection (nominally, a “liberty” claim under the substantive demands of the Fifth Amendment’s Due Process Clause).
My own sense is that Windsor is primarily an equal protection decision heavily influenced by concerns about structural federalism as an important guarantor of liberty. After all, what was the “liberty” at stake under the Fifth Amendment? Having treated as “unlike” what New York treated as “like” in a federal law that was “designed to injure” the class. Op. at 19. At least since Bolling v. Sharpe, the liberty in the Fifth Amendment’s Due Process Clause “contains within it the prohibition against denying to any person the equal protection of the laws.” Op. at 25. (There is a scholarly, but not judicial, dispute about this.) In Windsor, dignity is denied insofar as it denies equal recognition of same-sex and opposite-sex marriages. That sounds a lot like an equal protection concern. The Court says that “by seek[ing] to injure” married same-sex couples DOMA “violates basic due process and equal protection principles applicable to the federal government.” Op. at 20. Whatever the Court means by listing “due process” and “equal protection” separately, I can’t see how this is protection of a free-form “right to recognition.” The purpose of DOMA was to ensure that same-sex marriages would be treated as “second-class,” says the Court, which is what raises “a most serious question under the Constitution’s Fifth Amendment.” Op. at 22. The effect was to “identify a subset of state-sanctioned marriages and make them unequal.” Id. Again, that’s an equal protection concern.
Equal protection claims are commonplace, not unprecedented or problematic as a group (though this particular equal protection claim might be weak). Brown v. Board did not declare a fundamental right to education because it is not a liberty decision; it’s an equal protection decision. But the demeaning nature of segregation informed the Court’s understanding of why equal protection was important, of what was at stake in the denial.
One may disagree that denying a class of validly married people the full recognition of the law is in fact an affront to their “dignity,” but that’s a different claim than that dignity has nothing to do with the demands of equal protection. It is central to equal protection jurisprudence that the government cannot create castes of citizens — not simply because to do so would deny them specific tax advantages or other benefits, but because creating a second-class status is itself a harm to their dignitary interests. It would have been no answer in Loving v. Virginia to say that the state was required to recognize a civil union status for inter-racial couples with all the rights, but not the status, of marriage. That’s because, at least in the racial context, the separate recognition itself was an unconstitutional insult to them. At the very least, the affront to their dignity more completely informs what harm they’ve suffered in being denied equality.
Now there are many questions left about how, precisely, same-sex couples are denied “equal protection” when the federal government refuses to honor their valid state marriages. Heightened scrutiny of classifications based on sexual orientation, or sex, would have been one way to make the claim. The Court did not explicitly follow either route. Or perhaps the government did not even satisfy rational basis review because the effects of the law were too sweeping in relation to any legitimate purpose behind it. Romer v. Evans.
Here’s my own working thesis about what Windsor instructs. Perhaps, as the Court argued in the important concluding section of the opinion, there was unconstitutional “animus” behind the law, which itself is an impermissible legislative purpose and a breach of the government’s duty to treat every individual as though he possessed some worth represented by the word “dignity.” The government failed even to consider the interests of future married gay couples when it passed DOMA, a serious failure of the political process that should have been expected given the subject matter. See Carolene Products. And perhaps federalism principles, by assisting the states in protecting the substantive liberties of their own citizens, by limiting the reach and substance of federal legislation frustrating the implementation of state policy on family relations, and by confirming a historical practice from which DOMA dramatically and suspiciously departed, lead to the conclusion that equal protection was denied. As Justice Jackson once wrote, equality structurally protects liberty because it means that the majority can’t impose on some what it would not impose upon itself. Perhaps Windsor is what happens when Carolene Products, Romer, Lawrence, and the federalism revolution converge.