In a recent post on this blog, Jonathan Adler has written an imaginative contribution to the currently lively controversy over the propriety of nationwide injunctions. Other prominent scholars have expressed interest in his argument for curtailing these injunctions.

In this post, I will briefly explain why I agree with Adler’s policy position up to a point but disagree with the thrust of his legal analysis. I will expound my perspective more fully in a forthcoming article in the Notre Dame Law Review. A draft of that article is already available on SSRN.

At the oral arguments in two currently pending Supreme Court cases—United States v. Texas and Department of Education v Brown—Solicitor General Elizabeth Prelogar contended that “[t]he [Administrative Procedure Act (APA)] did not create a novel remedy of universal vacatur.” Adler’s post offers an unconventional argument in support of the SG’s thesis. As he recognizes, Chief Justice Roberts declared during the Texas argument that when he was on the D.C. Circuit, the judges of that court made frequent use of vacatur—“five times before breakfast”—to set aside agency rules. But Adler asserts that the Chief Justice was wrong.

More specifically, he contends that “when the D.C. Circuit vacates an agency action, it does not enter nationwide relief.” The perception that it does do so is an “artifact of the D.C. Circuit’s jurisdiction, combined with the binding effect of circuit precedent.” He elaborates:

As it happens, if one challenger has the ability to challenge an agency action in the D.C. Circuit, it is almost always the case that every challenger does. What this means is that once the D.C. Circuit has held that an agency action is unlawful, every other would-be challenger may rely upon the precedent in a challenge of their own, and those challenges will also occur in the D.C. Circuit. So once the D.C. Circuit upholds a challenge to an agency action and vacates or ‘sets aside’ the agency action, as a practical matter it has been set aside or vacated for the nation as a whole (unless, of course, the D.C. Circuit decides to remand without vacatur . . .).

But, Adler continues, the same reasoning doesn’t apply when an appeal is taken to a regional circuit. In those courts, “unless there are particular circumstances that weigh heavily in favor of nationwide relief (as might be the case with some immigration questions where less-than-nationwide relief may be tantamount to no relief), it seems to me we should let court jurisdiction and circuit court precedent do the work of determining whether successful challenges to agency actions have nationwide effect and be comfortable with the idea that individual courts offering universal or nationwide relief should be the exception and not the rule.”

To a large extent, I agree with Adler’s policy position. As he recognizes, the increasing prevalence of nationally binding injunctions issued by a single district judge is troubling. Aside from the inherent incongruity of vesting that much authority in a single individual, this practice impedes dialogue among the circuits and amplifies litigants’ incentives to maneuver in order to be able to bring their cases before the most favorably inclined adjudicator. Courts should, therefore, somehow be induced to issue nationwide relief less frequently. This is not to say that I endorse the one-plaintiff-at-a-time approach that Adler and the SG apparently prefer. In many situations, however, courts can and should try to head off the disadvantages that I just mentioned by, for example, limiting the scope of a vacatur or injunction, so that it applies only within the issuing court’s circuit.

On the other hand, I part company from Adler insofar as he discerns no appropriate role for vacatur of a rule except in “exceptional” situations. I especially question his suggestion that the judges who think they are vacating rules simply “do not usually give the question much thought” (maybe because they have not yet had their morning coffee?). To the contrary, I believe that Roberts’s description of current judicial practice accurately reflects the general understanding.

Back in 1985, a law review article by a lawyer named Merrick Garland stated that “[t]raditionally, [when] faced with an arbitrary and capricious decision . . . the court normally vacates the decision and remands the matter to the agency for further proceedings ‘consistent with’ the court’s opinion.” I do not mean to suggest that Attorney General Garland (to whom, of course, SG Prelogar reports) is estopped by a remark he made decades ago in a private capacity. Descriptively speaking, however, his statement was quite accurate. Similarly, the third edition of the leading administrative law treatise stated in 1994 that “[t]raditionally, a circuit court vacated a rule whenever it concluded that the agency action was arbitrary and capricious” (Davis & Pierce, § 7.4). You can find the same language, almost verbatim, in successive editions of the same treatise, all the way down to the sixth edition, published in 2020 (Hickman & Pierce, § 11.9).

And here’s a quote from a 1997 report by the ABA Section of Administrative Law and Regulatory Practice. Laying a foundation for an endorsement of the then-relatively new practice of remand without vacatur, the Section wrote: “Until recently, reviewing courts routinely vacated agency actions that they found to have been rendered unlawfully. That result has been generally accepted and relatively uncontroversial.” Flash forward to 2013, and you can find similar language in the Administrative Conference’s recommendation on remand without vacatur: “Traditionally, courts have reversed and set aside agency actions they have found to be arbitrary and capricious, unlawful, unsupported by substantial evidence, or otherwise in violation of an applicable standard of review.”

I played an active part in the deliberations leading up to these latter two institutional statements. Many members of the two organizations participated in those deliberations, and the debates were wide-ranging, vigorous, and sometimes contentious. But the issue the members debated was whether vacatur of a remanded regulation should be mandatory or discretionary. As best I can recall, not a single participant argued that vacatur was not an option in the first place.

Moreover, the courts often have compelling reasons to vacate rules as a whole, rather than only with respect to a victorious plaintiff. In a wide range of situations, the plaintiff-only model (which, to be fair, the SG largely inherited from the preceding administration) simply would not work.

Such situations are especially common when the litigation arises in an extensively regulated industry governed by a host of complex rules. If the agency is to be able to administer its program in a coherent manner, let alone a well-considered manner, it needs to be able to develop and implement these rules on a uniform basis, or at least one that considers relevant differences holistically. If, for example, a court were to find that a Federal Aviation Administration regulation prescribing safety standards for aircraft was arbitrary and capricious, we would expect the court to vacate the regulation with respect to all aircraft manufacturers, not merely the one company that may happen to have initiated the challenge in court. Moreover, the court could not simply revise the rule with respect to that manufacturer, because such an edict would require an exercise of discretion that has been vested in the agency. Rather, the court would have to remand the rule, so that the agency could consider whether and how to revise the regulation with respect to all manufacturers.

The key point here is that, for the system to work properly, the court’s directive to the agency has to be legally effective with respect to the regulation as a whole, not just the individual victorious plaintiff. Reliance on stare decisis and acquiescence, as Adler’s and the SG’s model contemplates, would not suffice, because the rule would still be on the books with respect to addressees other than the victorious plaintiff. To eliminate the portions not addressed by the court’s order, the agency would have to conduct additional rulemaking proceedings, which might take months or years.

This is certainly not how judicial review works today, and I fail to see how our system would be improved if the Supreme Court were to embrace the SG’s model going forward.

Much of the debate over universal injunctions has revolved around the question of whether they violate the APA. My article for the Notre Dame Law Review will address this question, and interested readers may consult the posted draft for my views on that subject; but Adler’s post does not probe the issue, so I will not discuss it here. The point I want to emphasize is that, in my view, the power to vacate a rule need not be exercised in all cases. It is discretionary. In that regard it is akin to the power to remand a rule without vacating it.

This brings me back to the terrain on which Adler and I have at least some agreement. The ABA’s 1997 resolution on remand without vacatur, which grew out of the project that I mentioned above, declared that “a reviewing court should normally strike the balance in favor of vacating the agency’s action, unless special circumstances exist.” Recent experience with the problems that nationwide injunctions can cause suggests that this statement may have been too strong. On the other hand, Adler’s claim that nationwide relief should be “the exception, not the rule” may go too far in the opposite direction. In any event, the spectrum between those two polar positions is a space within which reasonable dialogue can be carried on.

My forthcoming article makes some suggestions as to presumptions that the Court could adopt to clarify the circumstances in which universal relief is acceptable, and also to favor staying the effective date of nationwide injunctions pending appeal, to give other courts an opportunity to agree or disagree. I do not profess to have all the answers to these policy questions. I do believe, however, that moving beyond the unwarranted claim that such injunctions are categorically impermissible would be a good first step.

Ronald M. Levin is the William R. Orthwein Distinguished Professor of Law at Washington University in St. Louis.