Vertical Precedent in the Challenge to Male-Only Draft Registration (and Beyond)

by Michael C. Dorf

Last week, a panel of the US Court of Appeals for the Fifth Circuit reversed a district court judgment that had found male-only draft registration to be an unconstitutional denial of equal protection. The terse opinion in National Coalition for Men (NCM) v. Selective Service System was based on the following reasoning: (1) the Supreme Court rejected the contention that male-only draft registration was unconstitutional sex discrimination in the 1981 case of Rostker v. Goldberg; and (2) lower courts are bound to follow the holding of a precedent by a higher court "even where subsequent decisions or factual developments may appear to have ‘significantly undermined’ the rationale for [the] earlier holding."

Proposition (2) includes a quotation from a dissent by Justice O'Connor in Roper v. Simmons, but as the Fifth Circuit opinion correctly notes, the same proposition can be found in majority opinions in State Oil Co. v. Khan (1997) and Rodriguez de Quijas v. Shearson/Am. Express, Inc. (1989).

The Fifth Circuit's reasoning looks sound, which leaves us with a puzzle. Federal district courts are bound by both appeals court precedents in their circuits and SCOTUS precedents. Why, then, didn't Judge Miller likewise make short work of the plaintiffs' arguments? The answer is that the State Oil/Rodriguez rule is highly problematic.

In one of my earliest academic papers--a 1995 article in the UCLA Law Review--I noted that the State Oil/Rodriguez rule is essentially unenfoceable. Suppose a federal appeals court says that various developments since the decision in OldSCOTUSCase undermine the authority of OldSCOTUSCase, and the appeals court therefore declines to follow OldSCOTUSCase. What will the Supreme Court now do? If the Court grants cert to review the decision, it too will have to decide whether OldSCOTUSCase should be overruled or followed. If the Court now reaffirms OldSCOTUSCase (perhaps on stare decisis grounds), it will reverse the federal appeals court; if the Court instead concludes that the subsequent developments warrant overruling OldSCOTUSCase, it will affirm the appeals court. Yet these are exactly the same alternatives that would exist in the absence of the State Oil/Rodriguez rule. To be sure, regardless of the outcome, the SCOTUS opinion might include an admonition to lower courts not to anticipatorily overrule seemingly obsolete SCOTUS precedents, but the admonition will be toothless. It will not include a fine or other punishment, and in future cases, anticipatory overruling will be treated exactly the same way.

To be sure, most lower court judges will act in good faith and apply seemingly obsolete precedents despite the lack of any sanction for failing to do so. Yet we might also question the wisdom of State Oil/Rodriguez rule on the ground that it wastes judicial resources. If it is clear that a SCOTUS precedent is obsolete, a practice that permits lower courts to disregard them would free the Supreme Court of the need to grant cert and hear argument in cases in which the result is a foregone conclusion. To be sure, the Court might eventually want to take a case to declare OldSCOTUSCase officially overruled and thus avoid any lingering confusion, but the State Oil/Rodriguez rule--if followed stringently by the lower courts--requires it to choose between doing so earlier than it would like and allowing lower courts to apply rules of law that are, by hypothesis, incorrect.

Fair enough, you might say, but the State Oil/Rodriguez rule is currently obligatory, so why didn't the district court follow it in NCM? That brings me to the most fundamental objection to State Oil/Rodriguez: It relies on a problematic conception of precedent.

Although you wouldn't guess it from the Fifth Circuit opinion, Judge Miller's district court opinion did not say that Rostker was no longer good law. On the contrary, Judge Miller said he was applying Rostker, which held that laws that discriminate based on sex, even with respect to the military, are subject to intermediate scrutiny. Judge Miller said that now, as in Rostker, the government has an important interest in ensuring sufficient personnel for national defense but that the male-only registration requirement is not now substantially related to that interest. He explained that in Rostker the Court found that the second prong of intermediate scrutiny--substantial advancement of the important government interest--was satisfied given that women were excluded from combat. However, because the military has abandoned the combat exclusion, a blanket excuse from draft registration for women no longer substantially advances the government interest in securing access to combat-useful military personnel.

But wait. Didn't Justice O'Connor say that the State Oil/Rodriguez obligation persists even in the face of new legal or factual developments? Whether we regard the abandonment of the combat exclusion as a change in the law or facts, it shouldn't make a difference, right?

Wrong. For one thing, I don't see how we can possibly take Justice O'Connor's statement literally, either with respect to law or fact. Suppose a federal statute says that intentionally dumping benzene in navigable waters is a felony. Then suppose that in Case X the Supreme Court rules that for purposes of this particular statute navigable waters include wetlands. Now let's imagine that Congress amends the statute to provide that for purposes of the statute navigable waters do not include wetlands. The government prosecutes Doe for dumping benzene in wetlands. Doe says the current version of the statute doesn't apply. Is the district court required to reject Doe's argument because the Supreme Court has not yet ruled that the authority of Case X has been fatally undermined by the statute's amendment?

Of course not. And, to be fair, Justice O'Connor didn't say otherwise. She said that "subsequent decisions" that have seemingly fatally undermined a SCOTUS precedent should be disregarded until the SCOTUS itself delivers the coup de grace. She didn't say disregard all legal changes.

But even so, Justice O'Connor overstated the point. Brown v. Board of Education did not formally overrule Plessy v. Ferguson, leaving open the possibility that "separate but equal" might still be the law with respect to transportation. Yet in a series of per curiam opinions, the Supreme Court affirmed lower court decisions that--notwithstanding what would become the State Oil/Rodriguez rule--had treated Brown as overruling Plessy. No one thinks the lower courts were wrong to do so. 

Or consider facts. Suppose that in 2020 the Supreme Court rules in Shmacobson v. Shmassachusetts that mandatory vaccination against Disease X with Vaccine Y does not violate a substantive due process right to bodily integrity because there is a compelling interest in eradicating Disease X and Vaccine Y will provide herd immunity if administered in sufficiently high numbers, so mandating it is narrowly tailored. Ten years go by. The germ that causes Disease X has mutated so that Vaccine Y is no longer effective but a government agency nonetheless mandates Vaccine Y, even though there is now a cheaper, safer, more effective alternative, Vaccine Z. Must a lower court disregard the facts and simply uphold mandatory vaccination with Vaccine Y while it waits for the Supreme Court itself to overrule Shmacobson? That seems plainly nuts.

We might try to salvage the disregard-changed-facts part of Justice O'Connor's statement by assuming she had in mind changed legislative facts, not changed adjudicatory facts. That dichotomy distinguishes between facts that go into evaluating a proposition of law (legislative) and those that concern some particular dispute (adjudicatory). Consider a pair of examples.

In Shelby County v. Holder, the Supreme Court held that the coverage formula of the Voting Rights Act (VRA) was invalid partly on the factual ground that it imposed obligations in places where they were not needed. Suppose that next year the Biden administration attempts to enforce the VRA without first going back to Congress, arguing that new facts--evidence of widespread racially biased voter suppression--undercut the factual premise of Shelby County. Although it would be open to the Supreme Court to overrule Shelby County on that basis, a lower court could not do so anticipatorily. The existence of nonexistence of voter suppression is a legislative fact to which the State Oil/Rodriguez rule applies.

Here's a contrast. Suppose that Joan sues her employer, the Acme corporation, under Title VII, alleging that Joan's supervisor, Alex, sexually harassed her and that Acme did not take adequate preventive or remedial measures. The case goes up to the Supreme Court, which uses the case to establish a number of important propositions about employer liability. In the course of applying these principles of law, the Court finds that Acme's existing policy is inadequate. Joan wins and Acme loses. A few months later, Acme fires Alex and implements a new policy. A few months after that, Sharon sues Acme, alleging that her supervisor, Bob, sexually harassed her and that Acme did not take adequate preventive or remedial measures. Acme concedes that Bob committed the alleged acts, but as soon as the harassment was reported, it took action. Would a lower court be bound by the earlier SCOTUS ruling to find that Acme's policy was inadequate, notwithstanding the changes? Of course not. The adequacy of Acme's policy is a question of adjudicatory fact to which the State Oil/Rodriguez rule.

It looks like we have a sense of the domain of the State Oil/Rodriguez rule, at least with respect to facts: It applies to legislative facts but not adjudicatory facts. The problem is that many cases fall into a gray area in between. Was the existence of the combat exclusion in Rostker a legislative fact or an ajdudicatory act? It's tempting to say it was a legislative fact because it applied to all cases. But to say that requires us to delve into another gray area that also happens to be the subject of another one of my earliest articles: the scope of holdings.

In June Medical Services v. Russo, the SCOTUS reversed the Fifth Circuit's decision upholding Louisiana's law requiring doctors performing abortions to have admitting privileges in a nearby hospital. SCOTUS reversed on the ground that the Fifth Circuit failed to persuasively distinguish the essentially identical Texas admitting privileges law invalidated in the Whole Woman's Health case four years earlier. Now I happen to agree with the Court that the Louisiana and Texas laws were indistinguishable--but that's a contingent fact-based conclusion. We can imagine a world in which the facts on the ground in Louisiana and Texas are so different that the same legal text would satisfy the undue burden test in Louisiana but fail it in Texas. In such an imaginary world, would the Fifth Circuit have been required by the State Oil/Rodriguez rule to strike down the Louisiana law?

I would hope not. It's true that we could characterize Whole Woman's Health as holding that admitting-privileges laws are unconstitutional, full-stop, but we might better characterize it as holding that, on the record before the Court, the particular law was unconstitutional. In the course of reaching that conclusion, of course, the Court announced and applied propositions that have a bearing on the evaluation of similar laws from other states, but the holding itself is probably best characterized in less than sweeping terms.

Rostker looks similar. We could characterize the case as holding that excluding women from the obligation of draft registration is constitutional, full-stop, but we might alternatively characterize it as holding that excluding women from the obligation of draft registration is constitutional so long as women are also excluded from combat. If we understand the district court opinion in NCM as premised on the latter reading of Rostker, then it no longer appears that the district court was anticipatorily overruling Rostker; the court was simply distinguishing its facts and applying its holding.

So which is it? Should Rostker be read broadly? If so, then the Fifth Circuit was right that the State Oil/Rodriguez rule requires rejection of the challenge in NCM. Or should Rostker be read more narrowly? If so, then the State Oil/Rodriguez rule is irrelevant and the district court was entitled to consider the issue as not controlled by precedent.

I happen to think the district court was right. The Rostker opinion relies on the combat exclusion pervasively (using the term "combat" over a hundred times). But my main point is that there's no clear right answer to the question of how to characterize the holding of Rostker.

How to characterize the holdings of prior precedents is most of what appellate lawyers argue about in any given case. It's unrealistic to premise any rule on the supposition that the correct characterization of those holdings is self-evident. Yet the State Oil/Rodriguez rule rests on just that premise. Perhaps it's fortunate that the rule is unenforceable.