[Part II is available here.]

Universal relief is dead; long live universal relief! With the recent decision in Trump v. CASA, nationwide injunctions have been taken off the table as a way to control agency power. But as innocuous footnote 10 of the majority opinion notes, there remains another way for courts to grant universal, non-party remedies: section 706 of the Administrative Procedure Act, a.k.a. vacatur.

Since the latter half of the twentieth century, section 706 — which allows courts to “hold unlawful and set aside agency action” — has been read to authorize universal vacatur. And indeed, some have already argued that, in lieu of nationwide injunctions, vacatur will be the tool of choice for attacking executive orders. But not lacking a bit of controversy in its own right, vacatur will have to jump several hurdles before it is cemented as the undisputed successor to bread-and-butter equitable relief.

In our own article, we dive into the two questions that loom large over section 706. First, does the history and text of the APA support universal relief? And second, if it does, is it even constitutional? This post will tackle the first question, and Part II will deal with the latter problem.

As attentive readers of Notice & Comment will know, there is a lively debate about whether vacatur is permitted by section 706. There have been thought-provoking contributions both for vacatur — by Mila Sohoni, Ron Levin, Emily Bremer, Elliot Gaiser, and Fred Halbhuber — as well as against it — by John Harrison, Aditya Bamzai, Samuel Bray, and now ourselves.

The proper interpretation of section 706 depends on what “set aside” means, which hinges largely on the obscure administrative law of the early twentieth century. The anti-vacatur camp interprets the language to be a rule of decision: when a court sets aside agency action, it is disregarding it, then applying a remedy, like an injunction. Adhering to the traditional limits of equity, these remedies would be party-protective, not universal. On the other hand, the pro-vacatur camp reads “set aside” as an action upon the agency order itself. Much like when a court vacates a lower court’s judgment, the object is annulled in rem, thus binding the whole world.

The strongest evidence for the pro-vacatur thesis goes something like this: the APA’s “set aside” language is not unique, but rather draws from earlier statutes, like the 1908 Hepburn Act and the 1913 Urgent Deficiencies Act. When an order was set aside under these “special statutory proceedings,” the court would call the order “void,” “annulled,” or “without effect.” Because a voided order is not just enjoined, but rather destroyed in toto, it universally lacks legal cognizance.

Our article raises two objections to this account. First, it is not clear that calling an order “void” is intended to surpass the normal restrictions on equity. With a quick scan through James Lambert High’s 1905 treatise on injunctions, one finds dozens of examples where the object of suit is called “void” or “annulled,” yet it’s hard to believe that High was calling to obliterate the standard limits on in personam relief.^[1] Likewise, early set aside cases often construe the court’s power in terms of enforcement against parties, rather than some type of erasure power. For example, the court in Proctor & Gamble v. ICC construed its power to “enjoin, set aside, annul, or suspend” orders made by the ICC as no more than the ability to “stay the enforcement of an illegal order,” much like with traditional equitable relief.

Going a step further, some have tried to ground universal remedies not in equity at all, but rather in prerogative writs. James Pfander, writing with Jacob Wentzel, has argued that writs like mandamus determine the parties’ duties, not their rights, and therefore have been used to effect non-party relief. Fred Halbhuber has taken a similar tack, arguing that “set aside” provisions derive from the writ of certiorari, which could quash a lower-court judgment altogether.

We do not think these arguments get vacatur past the finish line. Professor Pfander notes that prerogative writs rarely acted beyond the pale of indivisible relief, which is not a “universal remedy” as understood today. To his credit, Halbhuber has convincingly shown that certiorari sometimes operated upon claims where relief was divisible — but vacatur is not certiorari. Whereas he notes that certiorari is “in the nature of a writ of error or appeal,” vacatur is not an appeal; it is not a review proceeding at all. Rather, vacatur is a “plenary suit in equity,” which is an original proceeding. This was understood even in the 1940s, when William Dempsey testified at the Walter–Logan hearings that “an appeal which is taken from an agency, whether commission, department, or any other executive officer… may be denominated an ‘appeal’ in the statute, [but] in all legal aspects… is an original proceeding.”

Although appellate review and plenary suits served a very similar function, there yet remained a formal legal distinction between the two. The latter is defined as a suit where the court may “vary, qualify, restrain, and model the remedy so as to suit it to mutual and adverse claims, controlling equities, and the real and substantial rights of all the parties” — the relief therefore acted upon parties, not proceedings. By the by, this excludes analogizing special statutory proceedings too closely to prerogative writs (which are legal, not equitable) as well as to the “appellate-review model” (Tom Merrill’s pioneering article itself calls the bill in equity an “original action”).

Even if all these arguments were correct — vacatur is some kind of proceeding in rem, which annuls the order “against the whole world” — we arrive at the next bramble. What was the effect of such a judgment; was it claim-preclusive on other jurisdictions? If not, then vacatur might be conceptualized as universal, but would not in fact be universal, because the government could simply engage in non-acquiescence, continuing to litigate its rule in other courts.

The hearings on the Walter–Logan Bill, the predecessor to the Administrative Procedure Act, are highly revealing on this point. The Bill provided for a judicial review provision, where the D.C. Circuit could “render a declaratory judgment holding [a] rule legal and valid or holding it contrary to law and invalid.” If the rule was invalid, “the rule thereafter shall not have any force or effect.” In response to this provision, Roosevelt’s acting Secretary of the Interior, Harry Slattery, complained that it would “confer upon the [D.C. Circuit] a veto power over administrative regulations,” and that “a defeat for the government would invalidate the rule once and for all.” Notably, there is no hint that this proposed scheme of universal relief was common or even precedented, despite the pro-vacatur thesis maintaining that it was a staple of all special statutory proceedings.

But what is most telling is the response given by O.R. McGuire, the chief draftsman of the American Bar Association’s Administrative Law Bill, which was the prototype of the Walter–Logan Bill and thence the APA. Slattery had things all wrong: “a judgment of the courts is never controlling beyond the territorial jurisdiction of such courts.” Strike one. “A judgment of the courts is never binding beyond the territorial jurisdiction of such courts and particularly upon individuals which were not parties to the proceedings resulting in such judgment.” Strike two. And if we couldn’t be surer, McGuire would say in a law review article that same year: “any such determination by [the D.C. Circuit] may not foreclose the raising of the issue in the merits of any controversy in some other court.” That’s strike three. Judgments could not bind other parties; they certainly could not bind other courts. And to the extent that Walter–Logan tried to flaunt this principle with universal review, it appears to have leaned into the D.C. Circuit’s “quasi-legislative” jurisdiction, a result of Article I rather than Article III. Therefore, this scheme should not be read into the APA.

Notably, no party in this hearing took the position that universal relief was a routine practice. This helps explain why the first unmistakable traces of vacatur do not sprout up until decades after the APA is passed. The scent did not go cold; it simply didn’t exist to begin with.

Despite the formidable historical objections to reading vacatur into section 706, others have noted that the Court might do so anyways. But even so, the race isn’t over: the courts will still have to tackle a far greater, even more daunting challenge. Even if vacatur is authorized, is it constitutional?

Jameson Payne is a graduate student at Hillsdale College’s Van Andel School of Statesmanship. GianCarlo Canaparo is a senior legal fellow in the Edwin Meese III Center for Legal and Judicial Studies at the Heritage Foundation where he studies equal protection, administrative law, and the separation of powers.

---

^[1] There is one exception, when he discusses suits to enjoin the levying of taxes. See pp. 545–48. He says that in taxpayer suits, one plaintiff may effect an injunction against the entire tax. Yet he is careful to note that his position is contradicted by other authorities, who say all parties seeking relief must be joined in the suit. The fact that he singles this case out as an example of “universal relief” implies that in other cases, the normal rule is maintained. And anyways, the “void” assessment in question is simply enjoined against all parties; the relief itself is not in rem. Cf. Bray, Multiple Chancellors, at 246.