Vacatur and United States v. Texas

Yesterday the Supreme Court heard argument in United States v. Texas. This case brings together standing, merits, and remedial issues of unusual complexity. Here are a few points of commentary on the vacatur question, about which there was a range of views on the Court, to put it mildly.

One of the occupational hazards of studying remedies is that the Court will often grant certiorari on remedial questions only to have them disappear because of how the question of standing or merits was resolved (e.g., Summer v. Earth Island Institute, Trump v. Hawaii). Here that is very possible given the challenges to Texas's argument on standing. But even if the Court does make it through standing and the merits, the Texas Solicitor General conceded to Justices Alito and Sotomayor that the Court didn't need to reach the question of whether the Administrative Procedure Act (APA) authorizes vacatur because the remedial question can be resolved entirely under Section 1252 (Transcript p. 120).

The argument against vacatur as a remedy is straightforward:

1. There is no historic remedy of vacatur.
2. The APA didn't add a remedy of vacatur.
3. That result has not been changed by Supreme Court precedent.

I think the arguments for points 1 and 2 are very strong. On 1, just ask a simple question: this remedy of "set aside" or "vacatur," is it legal or equitable? If there is no answer to that question (and there isn't), we're clearly not dealing with a traditional remedy. On 2, that the APA didn't add a remedy of vacatur is shown by text (the objects of "set aside" and its applicability across all kinds of actions including habeas), context (it's not in the remedies section), canons (hiding elephants in mouseholes), legislative history (no one noticed the elephant), treatises like Davis and Jaffe (still no one notices the elephant), and cases like Abbot Labs (still no one sees the elephant—why again do we think there's an elephant?).

I'll leave the more detailed analysis of 1 and 2 to John Harrison, whose work on these questions is superb (see here and here and here). But what's striking in United States v. Texas is that the state AG concedes point three: Supreme Court precedent does not settle whethere there is a vacatur remedy under the APA (Transcript pp. 109-110).

Assume points 1 and 2 for the sake of argument. If Supreme Court precedent doesn't require a different result, what are the other pathways to get to a vacatur remedy? And this is where most of the action was in the discussion of the remedy in oral argument. Three major theories seemed to emerge.

First, vacatur is a standard remedy in the D.C. Circuit. Yet the learned judges of that circuit, like all judges, apply their precedent. It's reasonable to ask where that precedent came from and when. The use of vacatur as a remedy in the D.C. Circuit doesn't develop until decades after the APA. Moreover, it's easy to see why the D.C. Circuit would be the one where the federal government would have the least incentive to argue about the scope of relief. If the D.C. Circuit rules against the federal government, in practical terms for the government it doesn't matter whether the scope of the remedy is universal or not, because anyone who wants can sue the government in D.C. and take advantage of the favorable precedent. So the D.C. Circuit is where the government is least likely to argue the scope of the remedy, and it is also where vacatur flourishes.

Second, the law of remedies changed. It is true that remedies can and do change. But there are interrelated problems with remedial change as a basis for vacatur. For a start, the federal courts need to connect their remedies to traditional practice (see, e.g., Grupo Mexicano; Atlas Life Insurance Co.). That means judicial change of remedies is supposed to be accretive, not avulsive. That's why the federal courts couldn't adopt the declaratory judgment remedy until it was authorized by the Federal Declaratory Judgment Act. And everyone admits that the existence or not of vacatur is a big deal (at least outside the D.C. Circuit—see previous point). Which means the legitimacy of this remedial innovation circles back to the question of whether the APA, as enacted, authorized a vacatur remedy. Moreover, statutes are read as incorporating traditional remedial principles. If the APA does make this avulsive change in remedies, it must be stated clearly. This principle is taught by numerous modern cases, including Weinberger v. Romero-Barcelo and Nken v. Holder, as well as pre-APA cases like Hecht Co. v. Bowles. I think "no vacatur" is the best reading of the APA, but even for someone who disagrees, it's really hard to believe that the "yes vacatur" reading could survive a clear-statement requirement. Finally, there's the difficulty that vacatur has not evolved to be a new remedy in the toolbox of the judge. Think of tort. Someone can't sue a company for making a defective product and ask for the remedies of (1) damages and (2) vacatur of the company's safety guidelines. Even to its supporters, vacatur is not a remedy that exists today outside of where it is authorized by statute. So it stands or falls on the basis of whether it was actually authorized.

Third, the legal theory can require vacatur: that is, if the plaintiffs are challenging the rule or other agency action on grounds that would suggest it is invalid as to everyone, then the remedy should have the same scope as the legal theory. But this confuses the merits with the remedy, or to put it another way, the holding with the judgment. A court can decide the case for A against B, and give a remedy to A against B, on grounds that would also apply to anyone else who brought a similar suit. Those grounds are what gives precedent its bite: A's case is a precedent for C's case. But A's case is not a judgment in C's case; A's remedy is not a remedy for C (see, e.g., Gill v. Whitford, Lewis v. Casey). As I put it in Multiple Chancellors, "Precedent should be the ordinary way one case ripples out to others." And the argument proves too much, because if "grounds that go beyond the plaintiff" meant that the remedy actually invalidated the rule for everyone, then why wouldn't the same be true of a challenge to the enforcement of a statute? Yet Massachusetts v. Mellon and California v. Texas both teach that courts don't invalidate statutes; they prohibit the enforcement of statutes against parties.

In my view none of these theories can get us from an APA that was originally enacted without vacatur to an APA that should be interpreted today by the Supreme Court as authorizing courts to vacate agency rules etc. That will probably not matter in United States v. Texas. After the oral argument and the concessions of the Texas Solicitor General, I don't expect this case will be a major decision about whether the APA confers on federal courts the power to grant a remedy of vacatur.

(In the previous paragraph I said "authorizing." There would be a further irony if the Court did hold that Texas had standing, that "shall" means "shall," and that Texas could receive a remedy of vacatur. Section 706 says a "reviewing court shall . . . hold unlawful and set aside agency action, findings, and conclusions" that are arbitrary and capricious. Is vacatur required? Does this "shall" also mean "shall"? As others have recognized, that would be an invasion of traditional remedial flexibility and discretion. But it's precisely the kind of question that emerges from trying to force Section 706 to be about remedies. It's not an issue once we see Section 706 as being about the scope of review: of course a court can be told, with no impinging on its remedial discretion, to disregard unlawful agency action, findings, and conclusions.)

But if there is no major vacatur decision in United States v. Texas, the question will not be going away. It will continue to draw the attention of scholars and courts. Ultimately the question will be about how to square two propositions that seem to be tension:

1. There is a statute, the APA, for which text, context, prior practice, and subsequent practice all converge to show no vacatur remedy (on Harrison's reading, which I find persuasive).
2. There is a practice of a vacatur remedy under the APA that develops several decades later in the lower courts.

There are two main ways we could try to square these these two propositions.

One is that the first proposition is wrong. There was an APA remedy of vacatur, but it's just that no one noticed it when the statute was enacted or at any time over the next thirty years.

The other is that the practice reflected in the second proposition is understandable but incorrect. There was no APA remedy of vacatur, and the lower courts backed into it because they were used to special review proceedings, the scope of the remedy didn't matter in the DC Circuit, and there was a shift (since reversed?) in thinking about courts as acting directly on statutes and other legal norms (see pages 451-452 of Multiple Chancellors).

Which is more plausible?

Comments

[Dr. Ed 2]

The problem is that the right is now starting to use the courts the way that the left has for some time, and the questions being asked now ought to have been asked when the left started this.

[ReaderY]

The Solicitor General’s position on state standing does seem rather inconsistent with Massachussetts v. EPA.

[ReaderY]

It’s not the aspect of the case Professor Bray is focusing on, but I think there’s a good argument that “shall means shall” has constitutional problems.

The President has an enumerated pardon power. It seems to me that this power necessarily implies a power not to prosecute. It would be a wasteful and silly gesture to require the Executive to prosecute fully and then pardon at the last minute. I think a reasonable reading of the Constitution, one that recognizes implied powers when they really are directly related to the enumerated powers, would say that the enumerated power to pardon implies a power not to prosecute in the first place.

This means that prosecutorial discretion really is embedded in the constitution. And it means the courts cannot force the President to obey a statute mandating prosecutorial enforcement.

The issue is hardly new. Many criminal statutes say, and have long said, that any person who commits certain conduct “shall” be adjudged guilty of a felony, imprisoned, etc. But such language has never been interpreted to mean there is no prosecutorial discretion.

This does not mean the President is entitled to grant comcrete benefits to aliens Congress has ordained shall be removed. But it does mean he is entitled to exercise discretion regarding whether to arrest and remove them or not. The consequences of his exercising this kind of discretion are a matter for the political process, not the courts.

One of the ways to address strong disagreement within a polity is to put laws on the books but not enforce them. It has a long historical pedigree. It may seem hypocritical and intellectually unsatisfying. But sometimes it keeps the peace and to compromise by saying that society’s official position is you should do X or not do Y, without really requiring people to do it.

[Rossami]

re: "It seems to me that this [pardon] power necessarily implies a power not to prosecute."

That is not obvious to me at all. There are five classic objectives of a justice system - Retribution, Deterrence, Incapacitation and Restoration and Rehabilitation. Three of those can be supported or enhanced (in the right circumstances) by the judicious use of pardon power but only after formal recognition that a serious crime was committed. Prosecutorial discretion deprives victims, society and even the criminal of the formal recognition that what was done was wrong.

Prosecutorial discretion may be necessary in a world of scarce resources but it does not serve the same functions and should not be confused with the pardon power.

[David Nieporent]

The President has an enumerated pardon power. It seems to me that this power necessarily implies a power not to prosecute.

It does not.

It would be a wasteful and silly gesture to require the Executive to prosecute fully and then pardon at the last minute.

Why would the executive have to prosecute fully and pardon at the last minute, rather than just pardoning at the first minute?

To be clear, I do think the executive has the power not to prosecute, but it has nothing whatsoever to do with the pardon power. Prosecutorial discretion is simply an inherent part of the executive power. (It has to be, because how could a requirement that one must prosecute be enforced? A court could theoretically issue an order a district attorney to bring a case, but it couldn't order the DA to do a good job.) Prosecutorial discretion exists whether or not the executive has the power to pardon.