In Department of Homeland Security v. Regents of the University of California, the Supreme Court held that DHS had not properly rescinded Deferred Action for Childhood Arrivals (DACA), a program offering some unauthorized migrants some immigration relief. The majority opinion offers a treasure trove of implications that we will be unpacking for some time. Here, I discuss what Regents may augur for Chevron waiver (a topic I addressed on this blog a while ago).

Chevron waiver involves an agency that declines to seek deference for a regulation or order that would normally warrant it. Say an agency issues a regulation (or a formal adjudication order) that interprets a statutory provision—something that would normally get Chevron deference if challenged in court. Now say someone challenges that interpretation in litigation. But instead of arguing against the challenger, the agency agrees that its own interpretation is indeed invalid. And—without going through a rulemaking or adjudication to change its interpretation—the agency asks the court not to defer to the (still operative) rule.

This is not the paradigmatic way agencies behave in litigation. What explains it is an election. New agency leadership, appointed or installed by a new Presidential Administration, does not wish to keep an interpretation promulgated under previous leadership in a previous Administration. But that new agency leadership has not gotten around to, or has no intention of ever, going through the procedures to officially change that interpretation.

A court in this situation basically has three options. It can defer to the agency’s current position that the interpretation is invalid. It can grant the agency’s request and interpret the statute de novo, with no deference to anything. Or it can defer to the promulgated interpretation anyway, even if the agency doesn’t want it to. The idea that an agency might be able to reject a court’s deference to the agency’s own interpretation has come to be called “Chevron waiver.”

In essence, Chevron waiver allows an agency in one Presidential Administration to weaken, undermine, or invalidate agency action undertaken in a previous Presidential Administration without going through the procedures normally required to make such a change. Instead, the agency at Time 2 can harness private party litigation to encourage courts to undo the work of the agency at Time 1. Some scholars have started discussing Chevron waiver, but the Supreme Court has not weighed in on this relatively recent phenomenon.

As is now widely known, the Regents Court held that, although DHS may revoke DACA, the way it did so in this instance was improper. Since DACA took the form of guidance, not a regulation or formal adjudication, neither the program nor the decision to end it would be eligible for Chevron deference. Regents thus says nothing directly about Chevron waiver. But the reasoning in Regents suggests that a majority of the Court would be wary.

Like Chevron waiver cases, Regents involves an agency in one Presidential Administration trying to undo its work under a predecessor. DHS instituted DACA during the Obama Administration and announced its rescission early on in the Trump Administration. That process, and the litigation that followed, involved three key documents, which I’ll just introduce here.

First, then-Attorney General Sessions sent a one-page letter to Elaine Duke, then-Acting Secretary of DHS, describing DACA as “an open-ended circumvention of immigration laws [that] was an unconstitutional exercise of authority by the Executive Branch.” He discussed the fate of a related program, Deferred Action for Parents of Americans (DAPA), which had been subject to a nationwide injunction, issued by the Fifth Circuit and affirmed by an equally divided Supreme Court, before being rescinded (under the Trump Administration) in 2017. “Because the DACA policy has the same legal and constitutional defects that the courts recognized as to DAPA,” the letter concluded, “it is likely that potentially imminent litigation would yield similar results with respect to DACA.” Therefore, “DHS should consider an orderly and efficient wind-down process” for the program, given “the costs and burdens that will be imposed on DHS associated with rescinding this policy”—presumably meaning the point in the future when DACA would be invalidated by courts in the course of litigation that was “potentially imminent.”

Second, having received this letter, then-Acting Secretary Duke issued a memorandum rescinding DACA. The memo reviewed the fate of DAPA and laid out a plan for gradually eliminating DACA. Most significantly for the purposes of Regents, it quoted key parts of the Sessions letter and articulated this rationale for ending DACA: “Taking into consideration the Supreme Court’s and the Fifth Circuit’s rulings in the ongoing litigation, and the . . . letter from the Attorney General, it is clear that the . . . DACA program should be terminated.” Litigation ensued.

In one of several cases born of this decision, the D.C. District Court held that the rescission memo’s “conclusory statements were insufficient to explain the change in [the agency’s] view of DACA’s lawfulness.” The court gave DHS the opportunity to give “a fuller explanation for the determination that the program” was not lawful.

Third, in response to this order, then-DHS Secretary Kirstjen Nielsen issued a memorandum “declin[ing] to disturb the . . . rescission,” and finding Duke’s “decision to rescind . . . sound” for three reasons. First, she noted the Attorney General’s conclusion that DACA was illegal, which was binding on DHS, and quoted the Fifth Circuit’s conclusion (in a footnote in the grant of preliminary injunction) that DAPA violated immigration law. Secretary Nielsen stated that “any arguable distinctions” between DAPA and DACA were “not sufficiently material to convince [her] that the DACA policy is lawful.” In the alternative, she said, there were at least serious questions as to DACA’s legality, and DHS should not pursue programs of questionable legality. Finally, there were good enforcement policy reasons to end the program. (The D.C. District Court held that this memo did not “elaborate meaningfully” on DHS’s reasons for finding the program unlawful.)

In Regents, the government asked the Supreme Court to consider both the Duke and the Nielsen DHS memoranda. Including the Nielsen memo was appropriate, the government argued, because “it was invited by the District Court and reflects the views of the Secretary of Homeland Security—the official responsible for immigration policy.” The Supreme Court declined, treating the memo instead as a post-hoc rationalization. Considering the memo, the majority wrote, would violate the principle that courts may only uphold agency action based on the agency’s own justification for that action, as required by Chenery I. (On this blog, Chris Walker described the Court’s rejection of the Nielsen memorandum as a newly vigorous version of Chenery, with implications for the practice of remand without vacatur.)

A dissenting Justice Kavanaugh scoffed at the idea that Chenery I extended that far: “the post hoc justification doctrine merely requires that courts assess agency action based on the official explanations of the agency decisionmakers, and not based on after-the-fact explanations advanced by agency lawyers during litigation (or by judges).” Because then-Secretary Nielsen was the agency’s primary decisionmaker, her memorandum could not be simply a post-hoc rationalization in litigation; it was the statement of the agency about the rescission’s reasoning. Indeed, Justice Kavanaugh went beyond debating the reach of Chenery I. Quoting the Administrative Procedure Act’s definition of the term “rule,” he asserted that “the Nielsen Memorandum is itself a ‘rule’ setting forth ‘an agency statement of general . . . applicability and future effect designed to implement . . . policy.’” This description would seem to suggest that any pronouncement by an agency head could be a “rule” under the APA, regardless of the procedures it took to produce it.

If we read the APA the way Justice Kavanaugh does, then the Nielsen memo was not just a memo but a rule—in fact, it may be the rule that rescinded DACA. In that case, the Court should indeed consider its reasons. But this is a pretty idiosyncratic reading of the APA, at least given the way its terms have developed in precedent and legal convention, where the unmarked term “rule” usually indicates something that went through some sort of notice and comment process. We usually refer to agency programs like DACA, as well as its rescission, as policy statements, or sometimes as nonlegislative rules, or more generally as guidance. Although rules are promulgated under an agency head’s name, I’m not sure I’ve ever seen anyone else suggest that an agency head’s statement could be a valid rule—in the way we normally use the term without modification—simply because an agency head made it. Perhaps that is one reason why Justice Kavanaugh wrote this dissent for himself alone. On the other hand, since we’re dealing with guidance here, there is at least a colorable argument that the Nielsen memo itself rescinded DACA, and that therefore the Court should primarily review that memo’s reasoning. The Court did not take this option, and the memo’s own phrasing works against this argument somewhat: it does not purport to rescind DACA but merely declines to disturb the Duke memo’s decisions.            

In any event, the majority opinion rejected the possibility that statements by an agency head were necessarily acceptable under Chenery, as Justice Kavanaugh seemed to argue. “The functional reasons for requiring contemporaneous explanations apply with equal force regardless whether post hoc justifications are raised in court by those appearing on behalf of the agency or by agency officials themselves,” Chief Justice Roberts wrote. Because the Court is tasked with reviewing whether the agency’s action was reasonable, it evaluates the reasons the agency had for taking the action—not other possible reasons that may justify it. “[T]he problem,” the majority said, “is the timing, not the speaker.”

So post-hoc reasons given by the head of an agency cannot serve as a justification for an agency action—not even nonlegislative guidance like DACA. The agency, moreover, does not get to tell the court what its “real” position is. The court evaluating an agency action looks to the agency’s position when it took that action.

But if a later reason cannot support or legitimize an earlier action, it’s hard to see why it should be able to undermine or delegitimize one, either. And that, of course, is what Chevron waiver does. In a Chevron waiver situation, an agency has already gone through extensive procedures that mobilize agency expertise and public participation to produce and justify a statutory interpretation. Under Chevron itself, courts are supposed to defer to that result insofar as it is reasonable—a decision courts make by looking both at the statute and at the agency’s reasoning about it.

Chevron waiver seeks to supplant that reasoning with a new view that has not gone through the arduous vetting process of rulemaking or adjudication. That new view may be the opinion of new agency leadership. But “the problem is the timing, not the speaker.” An agency head’s later decision that a statutory interpretation is not viable cannot be the reasoning underlying that statutory interpretation—and what the court reviews, and defers to under Chevron, is that reasoning. It should hardly matter whether the later agency opinion supports or undermines the earlier one. If you can’t consider post-hoc reasons, you shouldn’t consider them in any case. And if that’s the case for courts evaluating agency decisions about guidance like DACA, as Regents holds, it should apply even more strongly to decisions that have gone through the procedures that put them in Chevron’s domain in the first place.

Anya Bernstein is a Professor of Law at SUNY at Buffalo School of Law.