In 2009, Edith Windsor found herself saddled with an estate tax bill of almost $400,000. She sold some assets, paid the bill, and then sued the U.S. government for a refund. The Second Circuit found for Windsor, but the case is now before the Supreme Court, which may decide this June whether she is entitled to her refund.

Saying that United States v. Windsor is a case about a tax refund is like saying that Moby Dick is a book about a whale. Windsor’s late spouse was not a man but a woman—her partner of 40 years. The New York couple had married in Canada and then returned to their home state, which recognized their union under New York’s Marriage Equality Act. The federal government, however, did not recognize the union due to the terms of the Defense of Marriage Act, or DOMA, which had been passed by Congress and signed into law by President Clinton in 1996.

Amid the turbulent public debate over gay rights in America and the emotive public spectacle surrounding California’s Proposition 8 (in which 52 percent of voters statewide banned same-sex marriage in 2008 even as Barack Obama was winning California by 25 points), the facts in United States v. Windsor are relatively banal. But of the two gay-marriage cases now before the Supreme Court, it is the one most likely to make significant inroads toward the legal recognition of gay marriage across the nation.

Windsor may well win, but not on the grounds that many gay-rights advocates hope. If the oral arguments heard by the court on March 27 are any indication, Windsor is likely to win not because a majority of justices find the denial of gay marriage to be a violation of the Constitution’s guarantee of equal protection under the law, but rather on principles of federalism—sound federalism.

Until the Defense of Marriage Act became law in 1996, marriage was one of the last bastions of state sovereignty. As the Supreme Court noted in 1890, “the whole subject of the domestic relations of husband and wife, parent and child, belongs to the laws of the States and not to the laws of the United States.” The family has become a good deal more mobile and dynamic since 1890, so federal law has kept up with the need to enforce child-support payments across state lines, manage adoptions, and regulate child welfare. But much deference has been afforded to the state as the level of political authority that is closest to the family and most representative of the community values animating marriage as an institution.

This is not to say that federal laws relating to marriage did not exist before DOMA. The 1862 Morrill Act forbade any married person in a territory from marrying any other person. The statute was meant to suppress polygamy among members of the Church of the Latter-Day Saints. In fact, it proved so difficult to enforce the prohibition that Congress had to pass a subsequent statute in 1874 to dilute the power of Mormon-dominated juries in cases relating to marital relations by removing criminal and civil cases to federal district courts.

These two statutes applied only to territories of the United States—meaning, areas that were not states. Territories were governed under territorial laws enacted on the sufferance of the federal government. Whatever autonomy Congress gave to the territories Congress could take away by a simple majority vote—even if it meant taking away men’s wives in the process. That could not happen to the residents of the states. The Constitution’s Tenth Amendment—which says that “the powers not delegated to the United States by the Constitution, nor prohibited by it to the States, are reserved to the States respectively, or to the people”—gives the states an independent authority that the federal government cannot breach.

But of course the federal government tries to do so all the time, and it has seldom done so as brazenly as it did in the Defense of Marriage Act. In an attempt to nationalize a uniform definition of marriage, the Republican 104th Congress injected the federal government into an area of law historically left to the states. The DOMA statute stipulates that for the purpose of federal law, “marriage” means a legal union between one man and one woman as husband and wife. It also stipulates that no state is obligated to recognize same-sex marriages performed in another state.

President Clinton signed DOMA into law in the early hours of September 21, 1996, before the sun had even risen on Washington D.C. There was no public ceremony, no fanfare, not even a press conference. The White House waited until the last of the Friday-night news programs had aired before issuing a written statement from the president. “I am signing into law H.R. 3396, a bill relating to same-gender marriage,” Clinton wrote. “I have long opposed governmental recognition of same-gender marriages and this legislation is consistent with that position.”

Fast-forward to March 7, 2013. In a Washington Post op-ed titled “It’s time to overturn DOMA,” Clinton glossed over his stated opposition to same-sex marriage and instead suggested that he had signed the bill into law 16 years earlier to forestall a constitutional amendment banning it. Such an amendment, he suggested, would have set back gay rights for generations.

This is factitious, to put it mildly. Clinton signed DOMA into law because, had he failed to do so, according to his then pollster Dick Morris, he would have handed his Republican rival Bob Dole a potentially fatal weapon in the final weeks of his reelection campaign. But it is also true that the 104th Congress, with policy ideas pent up over the 40 years in which Republicans had never held the majority in the House, had entertained no fewer than 120 amendments in the first six months of 1995 alone.¹ The failure of a gay-marriage amendment to gain political traction had less to do with DOMA than with a healthy national aversion to writing temporal policy preferences into the Constitution.

But the political posturing that made DOMA the law of the land is child’s play compared with the contortions that now accompany calls for its demise. For example, President Obama has instructed the attorney general not to defend DOMA in the Windsor case because it is now the executive branch’s belief that the law is unconstitutional. At the same time, though, the executive branch continues to insist that the Justice Department (led by the attorney general) must enforce DOMA—because it is the constitutional function of the executive branch to take care that the laws are faithfully executed.

During oral arguments on March 27, Chief Justice John Roberts wondered: “I don’t see why [the president] doesn’t have the courage of his convictions and execute not only the statute, but do it consistent with his views of the Constitution, rather than saying, oh, we’ll wait till the Supreme Court tells us we have no choice.” Justice Antonin Scalia was characteristically blunt: “I’m wondering if we’re living in this new world where the attorney general can simply decide, yeah, it’s unconstitutional, but it’s not so unconstitutional that I’m not willing to enforce it.”

The administration’s position is intellectually and legally indefensible. This constitutional double-standard for DOMA renders the executive branch impotent when it comes to arguing its view of the statute as unconstitutional. It is a position, moreover, that may well backfire if the court rules against Edith Windsor.

All lawsuits are based on the twin principles of adversity and standing. The parties to a case must be adverse: They must have a genuine dispute that they can entrust to the court for resolution. But because the administration actually agrees with Windsor, the Supreme Court could conceivably find there is no dispute for it to resolve and simply dismiss the case.

When the administration declined to defend the statute, the Bipartisan Legal Advisory Group, or BLAG, stepped into the fray. BLAG is a standing panel of the House of Representatives. It comprises the speaker of the house, the majority leader, the majority whip, the minority leader, and the minority whip. The panel’s 3–2 vote to defend DOMA was along straight party lines. In voting as they did, Republican leaders who ordinarily deplore federal overreach chose to defend the extension of federal power into an area historically reserved to the states. And the Democrats, whose administration has all but invited the judiciary to decide the matter, voted not to deliver DOMA to the steps of the Supreme Court.

The Supreme Court will have to determine whether BLAG has standing to defend the statute in the first place. The court is not obligated to decide the case simply because BLAG disagrees with the appellate court’s ruling. Rather, BLAG must demonstrate that it stands to suffer direct harm if the court does not decide the case. Here, BLAG will have to show that the House of Representatives will sustain some harm if the decision of the Second Circuit stands as the final word on DOMA. The court has recognized Congress’s interest in the validity of its own statutes in the past and so is likely to accept that BLAG has standing to defend DOMA now.

That BLAG, and not the Justice Department, is defending the statute underscores the insupportable arrogance of the administration’s strategy. If the Supreme Court determines that BLAG lacks standing to defend DOMA, the constitutionality of the statute will remain unsettled. The Second Circuit’s decision is binding only within the Second Circuit, so those who want to advance similar challenges will have to do so in the other 12 circuit courts. This presents the possibility that some circuits will rule that DOMA is constitutional while others rule that it is unconstitutional. Then what? The nation would encounter a bizarre scenario: a federal statute that is being enforced in some circuits by the executive branch, but that cannot be meaningfully challenged nationally because the executive branch will not defend it.

After the oral arguments, it seems rational to expect that if the court does decide Windsor on the merits, it will do so in a narrow fashion—for example, by striking down DOMA on the ground that Congress encroached upon the police powers of the states by formulating a federal definition of marriage. In other words, it will not take up the question of whether banning same-sex marriage violates the equal-protection component of the Fifth Amendment, the desideratum of the gay-rights movement at the present time.

On several occasions, Justice Anthony Kennedy (the key swing vote) steered the oral argument away from equal protection to federalism. Perhaps most telling, when Solicitor General Donald Verrilli Jr. began arguing that homosexuals should be protected from discriminatory legislation under a standard of heightened scrutiny, Justice Kennedy interrupted him by saying that Verrilli was “insisting that we get to a very fundamental question about equal protection, but we don’t do that unless we assume the law is valid otherwise to begin with.”

Similarly, when Roberta Kaplan, the attorney for Edith Windsor, launched into an argument based on equal protection, Chief Justice Roberts stopped her: “No, no….You’re following the lead of the solicitor general and returning to the Equal Protection Clause every time I ask a federalism question.” Justice Sonia Sotomayor formulated the fundamental federalism question perfectly when she asked: “[W]hat gives the Federal government the right to be concerned at all at what the definition of marriage is?” Neither Verrilli nor Paul Clement, BLAG’s attorney, could give a satisfactory answer.

Clement was forced into logical contortions no less sinuous than those of the administration. When invited to offer precedents of federal involvement in marriage, he offered the federal prohibition of polygamy in the territories and then, alas, a statute passed during Reconstruction to recognize marriages between newly liberated slave couples—couples that had frequently been paired off involuntarily by their former masters.

Nor did the justices seem satisfied with Clement’s argument that DOMA was a mere “definitional provision,” and that the federal government defines marriage for the purpose of its own regulations all the time. The statutes to which Clement pointed do nothing of the sort; they simply incorporate state-law definitions of marriage for federal purposes. They then establish how those individuals will be treated under federal law. DOMA goes one step further by making a state definition of marriage irrelevant to the treatment of individuals under federal law.

Just last year, the Supreme Court confirmed the importance of state-level definitions of family relations for the purpose of enforcing federal law. In its unanimous opinion in Astrue v. Capato, a case that turned on the definition of a “child” for the purpose of Social Security benefits, the court examined the often tautological definitions of familial categories to which federal statutes resort. Under the Social Security Act, a “child” is “the child of an individual”; a “wife” is “the wife of an individual”; and a “husband” is “the husband of an individual.” In all instances, state law supplies the substantive definitions for these tautologies. But where the Social Security Act offers definitions additional to state law for the purposes of the receipt of benefits, those definitions do not contradict state definitions.

DOMA may well be the first occasion for a familial definition offered by Congress to come into direct conflict with the definitions states are adopting. If anything, this conflict serves to underscore one of the central reasons that the federal government deferred the regulation of marriage to the states for so long. Like all contractual relations, marriage is predicated on the free choice of the parties and bounded by the legal norms of the community. But the marriage contract is a contract unlike any other. It involves the affective bonds, religious beliefs, reproductive capacity, and sexual desires of the individual. It is an institution that exists in sacred and secular realms. Because state legislators are most proximate both physically and ideologically to the values of the community, our Constitution entrusts them with the exclusive right to regulate marriage in a manner consistent with the values of their respective communities.

DOMA sought to do something else. According to the House Report (a description of the legislative process leading to the act’s passage) that accompanied DOMA in 1996, which Justice Elena Kagan quoted during oral arguments, the act represented Congress’s desire “to reflect an honor of collective moral judgment and to express moral disapproval of homosexuality” for the nation as a whole. Collective disapproval of homosexuality was certainly nothing new in 1996 (to understate matters enormously), but using federal law to express it was new. Since the founding of the United States, it has been the role and function of individual states to identify the relationships they wish to honor or disapprove, not the federal government.

To allow the federal government to formulate a national standard for what has long been a state decision would drive one of the last nails into the coffin of federalism. In Windsor, the Supreme Court has an opportunity to reanimate the dying spirit of the nation’s many-layered and many-leveled political system. Striking down DOMA and leaving the recognition of gay marriage to the political process in the states is not just the federalist thing to do, it is the conservative thing to do.

1 This was not surprising. For those who believe, as conservatives do, in the importance of the federal system that enshrines the power of the states, amending the Constitution (which requires 38 states to agree to a new amendment) is the most legitimate way to enact a national policy that will trump state law.