The Truth of The Truth of Erasure, by Samuel L. Bray
Universal remedies are hot, and a number of scholars have weighed in on the specific question of whether the Administrative Procedure Act authorizes a separate remedy of vacatur (sometimes called “universal vacatur”). The leading contributions so far are by, in alphabetical order, Adler, Bamzai, Bremer, Harrison (x2, x3), Levin, Mizelle, and Sohoni (x2). At the core of this emerging literature are difficult questions about pre-APA practices, how a statute can be developed by courts over time, the scope of the federal courts’ remedial powers, and the translation of a model designed for review of agency adjudication to review of agency rulemaking.
A new entrant is The Truth of Erasure: Universal Remedies for Universal Agency Actions, an article by T. Elliot Gaiser, Mathura Sridharan, and Nicholas Cordova. It rejects universal injunctions but argues in favor of what it calls “the universal final remedy of vacatur” (*8). It argues that this remedy is legal, not equitable, and that it is not only constitutionally permissible but constitutionally required.
I am going to leave largely to one side the arguments from The Truth of Erasure that sound in administrative law. One is that the APA draws in substantial part on an appellate-review model. (That theme has been explored by Bamzai, Bremer, and Sohoni, but also note the cautions in Harrison’s piece on Corner Post about the cases where that model runs out. And there is foundational work by Merrill.) Another is that the APA model was focused on agency adjudication, since that was expected to be the norm—and was the norm until perhaps the late 60s and early 70s. (That theme has been explored by Schiller.) And there are other arguments in the article about pre-APA history, APA text, and APA structure, siding more with Sohoni than with Bamzai and Harrison.
Regardless of what one thinks about them, on these points the article is not breaking new ground. Where the article does innovate is on remedies, equity, and constitutional interpretation. Unfortunately, the original arguments are wide of the mark. Here are four concerns.
First, the authors’ argument for why the Constitution requires vacatur is thin gruel (*3; cf. *10, 11). The argument is not that the Constitution requires some judicial review of agency action, but that the Constitution requires vacatur as a remedy. For starters, it is not remotely an originalist argument. Its actual content seems to be that rulemaking is too easy, that nonparties need protection, and that equal protection (the authors refer to “full protection” at *10) requires courts to step in and do something. That argument rests on contestable empirical premises: it is not true that legislation is always “far more daunting” (*13) than rulemaking. And it proves too much—if there is an equality basis for constitutionally required universal remedies with respect to rules, then the same argument can be generated for statutes. (I am aware the authors try to distinguish rules and statutes, but in my view it doesn’t work. At least for rules there are two federal branches that can ensure a measure of accountability, Congress and the courts; but for statutes there is only one, the courts.)
And the article carves out presidential action as not being subject to vacatur (*2, 7-8, 17). The authors’ arguments about constitutionally required vacatur for agency rulemaking—because of too-easy-action and unequal treatment—apply with as great or greater force to the actions of the president alone. (Cf. Kovacs.) Moreover, the scope of the APA is irrelevant for determining what remedies are required by the Constitution. It is hard to see why the Constitution would require vacatur all through the height and breadth of the administrative state excepting only one desk in the Oval Office.
Second, it is rather startling that the article would argue that vacatur and stay of agency action are legal remedies (*11). As an initial matter, they aren’t remedies. You won’t find them in remedies treatises or casebooks. They have never been considered that way. “Vacatur” is a remedy like remand is a remedy or certiorari is a remedy. Similarly, there is no traditional remedy of “set aside.” The traditional remedies are well known—damages, injunctions, constructive trust, accounting for profits, specific performance, and so on—and when a new remedy is created by statute, such as the declaratory judgment, there is no doubt about what is happening.
But let’s assume for a moment that vacatur and stay of agency action should be conceptualized as remedies. What would be the implications of them being legal remedies? There would likely be a jury trial right in all APA suits for vacatur—an astonishing change from current practice. And none of the equitable defenses would apply, so the equitable conduct and misconduct of the parties would be irrelevant (including laches—consider how that irrelevance would fit with Corner Post). The court could not make vacatur and stay of agency action conditional—a hallmark of equitable remedies. Moreover, the availability of vacatur would wind up sharply limiting the availability of injunctions, since vacatur would provide an adequate remedy at law. There would be no ability to modify or dissolve the order using equitable powers. And there likely would be no contempt enforcement (the authors say vacatur does not act in personam). The authors recognize none of these likely implications of their argument that vacatur and stay of agency action are legal remedies.
Third, one should be skeptical of the authors’ suggestion that equity poses a special separation of powers problem (*11-12). The authors say “the Chancellor could ‘make law’ independent of the courts and Parliament” (*11), and from that the authors draw a direct line to the federal courts, concluding that for judges to develop equitable remedies would be an exercise of lawmaking power. That argument relies on medieval practices from a time when there were no printed reports of equity decisions. Eventually it became clear that chancery was operating as a court—so it’s odd for the article to describe the chancellor as if he were operating outside of the courts. And what the chancellor did (particularly in equity’s concurrent jurisdiction) was not understood as lawmaking, but as controlling the exercise of legal powers. In fact, the equitable focus on exceptions would seem to make equity, if anything, less legislative. Bottom line: equity was no more legislative than the common law, which is to say it was not.
Fourth, the authors call class actions under Federal Rule of Civil Procedure 23 a legal remedy, and they seem to be saying that such a characterization is necessary for class actions to be constitutional (compare *12 and *13). This is incorrect. Rule 23 is a lineal descendent from the Federal Equity Rules. It’s not legal. It’s not a remedy. And there’s no constitutional problem at all.
Some concluding thoughts.
The authors reject universal injunctions and endorse “the universal final remedy of vacatur” (*8). That position has gained traction recently in some quarters, and it has been endorsed by one justice of the Supreme Court. But I am aware of no substantial scholarship that supports it, and for good reason, since it is the least likely of all the positions to be correct. The only exception I am aware of is one subsection in Mitchell’s article on erasure, and that is heavily relied on by The Truth of Erasure.
In broad outline, there are four positions in the recent scholarship on the APA and the scope of remedies. The four positions can be sketched this way, though note that I am abstracting from some of the complexities (e.g., which forms of relief?) in order to present the positions in parallel:
- The pre-APA background did not allow universal relief, and the APA followed suit and did not authorize universal relief (broadly, Bamzai and Harrison; Multiple Chancellors and Proper Parties, Proper Relief tend to support this position).
- The pre-APA background did allow universal relief, and the APA followed suit and did authorize universal relief (broadly, Sohoni).
- The pre-APA background did not allow universal relief, and the APA followed suit and did not authorize universal relief, but the APA is a framework statute that left the courts the task of filling in the details, and they have done so with universal relief (broadly, Levin).
- The question is not about universal relief at all, but is simply about judicial review of agency action: the APA authorizes review of rules as if the court were hearing an appeal from an adjudication, with the usual non-remedial outcomes, including vacating the “judgment” below (Bremer).
My view is that the threshold question for an interpreter is about the relationship of the judiciary to statutes and change. That determines whether option 3 is plausible. If courts can fill in the details of a framework statute, then Levin’s analysis works descriptively, and it explains why the D.C. Circuit later builds out vacatur (as discussed more in Adler). If not—and framework statutes are not au courant with the Court’s approach to statutory interpretation—then the other options come to the fore. I don’t think option 2 works, since the background principles of equity and of the declaratory judgment do not support universal relief.
So that leaves options 1 and 4 as the most plausible. I suspect that a lot of what makes one or the other persuasive depends on the perspective of the reader. If you come at it from a perspective of the fundamental principles of judicial power (the dispute resolution model, the judgment power, what remedies do, who are necessary parties, alignment of injunctive scope and capacity to enforce via contempt, preclusion doctrine, etc.), then the way Bamzai and Harrison put the pieces together will make the most sense. If you come at from a perspective of agency functioning and especially of agency adjudication as a neglected feature of the administrative state, then the way Bremer puts the pieces together will make the most sense.
But here’s the point that’s relevant for The Truth of Erasure. All four of the positions I enumerated are more plausible than a background rejection of universal relief with a dramatic legal remedy of vacatur created by the APA. One reason is that on many points the APA was a consolidation more than an innovation, and that seems to be the case on remedial scope. That’s why none of the big early treatise-writers understood the APA to be innovating on the scope of remedies. It would be surprising if they missed it. Another reason is that the APA explicitly refers to various kinds of traditional remedies, and statutes take the law of remedies as they find it unless they modify it with clarity (Weinberger v. Romero-Barcelo makes this point for equity; cf. Nken v. Holder for stays). Congress needs to speak clearly if it is going to add a super-remedy that does not have a traditional basis.
So if the remedy and judicial power framing is correct (in line with options 1-3 above, contra Bremer), and if the background law did not allow universal relief (in line with options 1 and 3 above, contra Sohoni), then the APA is just not going to make a huge change to remedies on the sly—light hidden under a bushel, or elephants in mouseholes, so to speak. So you get continuity pre-APA and post-APA. And it is striking that a similar continuity obtains with option 2 (universal relief both pre- and post-APA) and option 4 (appellate-style review of agency action both pre- and post-APA). The four scholarly positions enumerated above differ in how they see the pre-APA background and in whether they think a remedies and judicial power framing is correct, but all four positions emphasize continuity.
By contrast, it’s worth emphasizing how novel the position of the article is. In the 1990s, when Judge Williams was writing National Mining Association, he was giving a full-throated endorsement of courts setting aside agency rules—and he says that’s an injunction. What that means is that someone can distinguish the remedy of an injunction from a separate remedy of vacatur or “set aside.” Or someone can line up in continuity with all the DC Circuit practice going back to the 1980s. But one cannot do both.
The position of The Truth of Erasure—that universal injunctions are suspect, but a legal remedy of vacatur is home free and even constitutionally required—is fundamentally novel. The reason this novel position is emerging now is a direct response to a tension felt by federal judges: on the one hand, a desire for “strong medicine” remedies against federal agencies, but on the other hand, a recognition that universal injunctions are at odds with the traditional concepts of party-specific judgments and remedies. The Truth of Erasure is generated by that tension, but the tension remains.
Postscript: I have focused only on injunctions, as the authors do. Mandamus is also relevant and needs to a bigger part of the conversation. In particular, mandamus has a different logic that is less focused on the parties; its scope is determined by the duties of the defendant, not the protection of the plaintiff. Where an agency fails to perform a ministerial duty, mandamus allows a court to grant relief.
Samuel L. Bray is the John N. Matthews Professor of Law at the University of Notre Dame.