Kisor v. Wilkie, in which the Court narrowly declined to overrule Auer deference while distinctly constraining its reach, shares with United States v. Mead Corp. and Skidmore v. Swift & Co.—so prominently mentioned in its opinions—the characteristic of being about an administrative action that, at root, is soft law, guidance. Quite appropriately, it did not address Chevron deference, although readers of the opinions, with Justices’ prior opinions in mind, will have no problem foreseeing a similar confrontation in the coming Term of Court, or perhaps the one to follow.

Gundy v. United States, in which the Court barely escaped finding a non-delegation problem in a statute permitting the Attorney General to render applicable to persons convicted of sex offenses before its enactment, a federal statute imposing serious criminal penalties if they failed to register their status (SORNA), shares with Chevron v. NRDC and NLRB v. Hearst Publications the characteristic of addressing the hard law consequences of administrative actions taken under statutory authority that had the consequence of constraining, but not eliminating, the judiciary’s role.

Justice Kagan wrote the prevailing opinions in both Kisor and Gundy. Justice Gorsuch wrote both the four-Justice concurrence in Kisor and the three-Justice dissent in Gundy (Justice Kavanaugh not participating), strongly arguing that separation of power considerations should have led to the overruling of Auer (in favor of “Skidmore deference”) and required a finding of non-delegation in Gundy. Chief Justice Roberts provided the decisive fifth vote in Kisor, emphasizing the majority’s constraints on Auer and invoking stare decisis; but he joined the dissent in Gundy. Today’s conservative assault on Chevron is, essentially, grounded in separation of power concerns that the decision denies to judges their “exclusive” final authority to determine issues of statutory meaning—an authority recognized again and again in 5 U.S.C. 706, the relevant provision of the Administrative Procedure Act. So, then, is Chevron about to go?

Start with the dissent in Gundy. A crucial passage for understanding it, in the author’s judgment, one that seems to him fully to explain—indeed, justify—the Chief Justice’s vote, appears in its concluding paragraphs:

Nothing found here can come as a surprise. In Reynolds, the government told this Court that SORNA supplies no standards regulating the Attorney General’s treatment of pre-Act offenders. This Court agreed, and everyone proceeded with eyes open about the potential constitutional consequences; in fact, the dissent expressly warned that adopting such a broad construction of the statute would yield the separation-of-powers challenge we face today. Now, when the statute faces the chopping block, the government asks us to ignore its earlier arguments and reimagine (really, rewrite) the statute in a new and narrower way to avoid its long-predicted fate. No wonder some of us are not inclined to play along.

The only real surprise is that the Court fails to make good on the consequences the government invited, resolving nothing and deferring everything. In a future case with a full panel, I remain hopeful that the Court may yet recognize that, while Congress can enlist considerable assistance from the executive branch in filling up details and finding facts, it may never hand off to the nation’s chief prosecutor the power to write his own criminal code. That “is delegation running riot.”

The government had made a similar no-standards argument in a case involving the Department of the Interior’s consideration of Native American applications to permit casinos, and in the Eighth Circuit that (unsurprisingly) produced a non-delegation holding. South Dakota v. Department of the Interior, 69 F.3d 878 (8th Cir. 1995). As Judge Leventhal pithily remarked in Ethyl Corp. v. EPA, 541 F.2d 1 (D.C. Cir. 1976), “Congress has been willing to delegate its legislative powers broadly and courts have upheld such delegations because there is judicial review to assure that the agency exercises the delegated power within statutory limits, and that it fleshes out objectives within those limits by an administration that is not irrational or discriminatory” (emphasis supplied). Thus, to deny the possibility of standards for review in a case in which judicial review is a requisite of constitutionality, is to invite a non-delegation holding. Over Justice Scalia’s dissent demanding argument, the Supreme Court summarily vacated the judgment, 519 U.S. 919 (1996), on receiving the Solicitor General’s representation that the Secretary of the Interior had belatedly concluded that there were in fact statutory standards by which the legality of his judgment could be judicially assessed. Gundy appears to have permitted the argument Justice Scalia—long a committed Chevron enthusiast—had hoped for; “delegation” issues in general are not threatened by the proposition that the government takes risks with inconsistent positions about the existence of reviewability standards, finding them only belatedly, when it serves to provide litigating cover.

Hearst, then, was a case decided contemporaneously with Skidmore. Here, the relevant statute concerned “employees,” and the Court began by independently concluding, on its own, that using state law definitions would be inappropriate in a national law, and that Congress in statutes using the term had not used it as if it had a single meaning, uniform across all federal statutes. In the National Labor Relations Act, then, its meaning should be influenced by labor policy concerns, and that made the task primarily one for the NLRB—that is to say, a delegated task, but one occurring under judicial supervision. Here, the task of judicial review was not fully independent judicial judgment influenced by persuasive agency reasoning, Skidmore, but a policing of the boundaries of possible meaning (the NLRB could not say an ordinary independent contractor hired for an occasional job was an employee, or that someone paid hourly wages was not) and, for judgments within those boundaries, “an administration that is not irrational or discriminatory.” One could describe the Hearst-Skidmore difference in a manner similar to the way Chief Justice Roberts described the difference between Auer and Skidmore in his concurring opinion in Kisor:

That is not to say that Auer deference is just the same as the power of persuasion discussed in Skidmore v. Swift & Co., 323 U. S. 134 (1944); there is a difference between holding that a court ought to be persuaded by an agency’s interpretation and holding that it should defer to that interpretation under certain conditions. But it is to say that the cases in which Auer deference is warranted largely overlap with the cases in which it would be unreasonable for a court not to be persuaded by an agency’s interpretation of its own regulation.

And Chevron is little more than Hearst generalized. Its first step corresponds to the initial independent inquiry to be found in Hearst, independently bounding what the statute in question could mean; and then finding within the space thus delimited a permissible delegation to the agency authorized to administer it by “an administration that is not irrational or discriminatory,” and under judicial supervision, the authority to determine its reasonable application.

Justice Kagan’s Kisor opinion supplies analyses strangely missing from the Chevron opinion, and in doing so suggests what would be a welcome outcome of the current kerfuffle. Chevron, oddly, failed entirely to mention the APA’s provisions structuring judicial review of administrative action. Her opinion’s reasoning, applied in the Chevron context, would remedy that defect, and in doing so provide a persuasive Chevron framework. Its Step One is the independent judicial determination of questions of law the APA (and Marbury) call for, informed by the traditional tools of statutory interpretation, including such tools as “narrowing construction” and questions too important to leave for the agency that the caselaw has developed. A catalog of constraints has emerged from the caselaw and the literature not identical to, but parallel with, those her opinion finds qualifying Auer—fully joined in these respects by the Chief Justice. Within the space independent judicial judgment finds after using those tools and observing those constraints, an agency empowered to act with the force of law may do so by “an administration that is not irrational or discriminatory,” Step Two.

The exercise of independent judicial judgment at step one, and oversight at step two, are the apotheosis of Justice Gorsuch’s “delegation running riot.” In his Kisor concurrence, posted after Gundy, he remarks without apparent concern:

To be sure, some cases involve regulations that employ broad and open-ended terms like “reasonable,” “appropriate,” “feasible,” or “practicable.” Those kinds of terms afford agencies broad policy discretion, and courts allow an agency to reasonably exercise its discretion to choose among the options allowed by the text of the rule. But that is more State Farm than Auer. See Motor Vehicle Mfrs. Assn. of United States, Inc. v. State Farm Mut. Automobile Ins. Co., 463 U. S. 29 (1983).

Indeed, this is Step Two, 5 U.S.C. 706(2)(A).

Back for just a moment to Justice Scalia—so enthusiastic about Chevron, though in his last years he perhaps began to waiver. Recall his angry dissent to Mead: Skidmore was dead; there could only be Chevron. But Mead concerned soft law, Chevron hard law. If Skidmore required reviving, this Term’s work plainly has accomplished that. And Hearst/Chevron, even so, may (and in the author’s judgment should) remain good law.

Peter L. Strauss is the Betts Professor Emeritus of Law at Columbia Law School.