Four years ago, Michael Herz published a fascinating article in which he argued that separation-of-powers devotees should stop worrying about Chevron deference. However it might once have been understood, Herz argued that Chevron deference descriptively was and normatively ought to be understood in 2015 as a two-step process that allowed both the judicial and the executive departments to discharge their respective constitutional duties and capitalize on their respective institutional advantages.

Here’s how: At Chevron Step One, courts engage in the interpretation of statutory text, “figuring out what constraints are actually imposed by the statute.” That’s their constitutional duty; that’s what they’re good at. Only if the text is genuinely ambiguous and the “law runs out” do courts engage in deferential review of agency policymaking at Step Two. Completing ambiguous legislation is the stuff of executive duty, and agencies are good at it. Herz concluded:

Chevron is not a revolutionary shift of authority from the judiciary to the executive. That Chevron is dead. Rather, Chevron is an appropriate allocation of decisionmaking responsibility among the three branches, relying on the judiciary to enforce congressional decisions, but protecting agency authority and discretion where Congress has left the decision to the executive. Long may it reign.

I’ve borrowed and amended the title of Hertz’s article because—as described/clarified/reengineered by Justice Kagan in Kisor v. Wilkie—much the same thing might be said about Auer deference. I’ll pass on the question of how many steps Auer should now be understood to have—for my purposes, it’s critical to understand that Auer has at least two steps, and that the first step should be taken very carefully.

First, we know that Auer isn’t triggered unless a regulation is “genuinely ambiguous.” Auer “Step One” entails ascertaining ambiguity (or lack thereof). We also know that genuine ambiguity means that judges are to throw “all the standard tools of interpretation” at a regulation before they even think about deferring. They are to do something like what Justice Gorsuch has done in recent statutory cases at Chevron Step One—hard-nosed, context-sensitive, canon-firing textual inquiry that is difficult to distinguish from de novo review. If the meaning is clear, that clear meaning controls. Only if no clear meaning emerges does a judge proceed to Step Two—the determination of whether the agency’s policy choice within what Justice Kagan calls the “zone of ambiguity” is reasonable.

Doctrines of statutory or regulatory deference that are triggered only after that kind of analysis yields no clear answers seem unobjectionable on separation-of-powers grounds. Post-Kisor, it will be open to any judge who, after making a “conscientious effort to determine, based on indicia like text, structure, history, and purpose whether [a] regulation really has more than one reasonable meaning”, concludes that one possible interpretation is more likely to be correct than any alternative to hold that the law has not run out and the regulation is not ambiguous.

The formalization of Auer Step One and Justice Kagan’s emphasis on genuine ambiguity is a big deal. It’s true that even before Kisor, judges could and did use “the traditional tools of statutory construction” to determine whether regulations were ambiguous before extending Auer deference. Cynthia Barmore’s empirical work suggests that judges in recent years have rarely abandoned what they regarded as the “best” interpretation of regulatory text because of Auer. The message of Kisor, however, is unmistakable: There was a “bad” Auer—one that led the Federal Circuit to reflexively defer to the Board of Veterans’ Appeals upon being confronted with reasonable-sounding competing claims about the meaning of “relevant” without any substantial inquiry of its own— and that bad Auer is now bad law. Judges are to begin by independently interpreting regulatory text, just as they would without an agency interpretation in the backdrop. They are only to defer if and when “the law”—and with it, their constitutional and statutory duties to independently interpret the law—runs out.

Is there anything else that Auer’s critics (and I’ve been one) have cause to complain about? The judicial-duty objection no longer appears to have bite. It’s difficult to dispute Justice Kagan’s claim that agency officials are institutionally better-positioned than judges to choose between linguistically permissible means of implementing ambiguous regulations that fall within the scope of agency officials’ expertise. The only systematic empirical study of the question whether Auer—as critics have claimed—has incentivized agencies to draft ambiguous rules has concluded that this “self-delegation” concern is a “false alarm.” What’s left?

Here’s a possibility: There’s no place (yet) for hard-look arbitrary-and-capricious review within Auer’s (five?) steps. Even if agency officials don’t promulgate mush because of Auer, there remains space within the interstices of regulations for them engage in opportunism—whether at the behest of interest groups, the White House, or Congress, or on their own initiative—when the law runs out. Hard-look review enables judges to smoke out such opportunism by determining whether agencies have taken relevant factors into consideration and have sought to accomplish contextually legitimate goals. As Chris Walker observed, Justice Kagan has contributed to Court’s identification of Chevron Step Two with hard-look review. But she did not suggest in Kisor that hard-look review ought to be part of Auer Step Two. Instead, she suggested that any interpretation that “come[s] within the zone of ambiguity the court has identified after employing all its interpretive tools” is reasonable.

One might object at this point that hard-look review is already available and there is thus no need to incorporate it into Auer Step Two. As Catherine Sharkey has observed in the parallel context of Chevron Step Two, however, this response is too quick. Some courts have applied hard look review only to agency rules that suffer from procedural defects or used it only to evaluate changes in agencies’ interpretive positions. Absent the incorporation of hard-look review into Auer Step Two, then, hard-look review may not be available at all in a nontrivial number of cases.

Incorporating hard-look review into Auer Step Two wouldn’t be very difficult. Some lower courts have drawn upon pre-Auer decisions for the proposition that interpretations must “sensibly conform[] to the purpose and wording of . . . regulations” in order to be reasonable. Agencies are required by law to issue preambles along with final rules, and those preambles must set forth purposes in their text. These preambular statements of purpose do not carry the force of law, but —as Jennifer Nou has emphasized—all of the institutional decisionmakers who must sign off on proposed rules must also sign off on the preambles that accompany those rules. No complex inferences from the text or structure of the law may be necessary to identify purposes which can guide decisionmaking within the zone of ambiguity. By requiring agency officials to show that their ultimate choices were based on, and have a reasonable connection to, purposes delineated after a bargaining process in which outsiders participated, judges could counteract opportunistic construction of regulatory language at Auer Step Two, just as they counteract opportunistic construction of statutory language through ordinary hard-look review.

With that caveat, however, there’s something for just about everyone to like about Kisor. Auer’s supporters can take comfort in the fact that the Court reaffirmed Auer’s sound institutional judgement that agencies are better positioned to make policy than are judges. At the same time, the Court in Kisor expressed a mood that should be understood as requiring judges to do precisely what Auer’s critics have long charged that Auer effectively prohibits judges from doing: saying what the law is. Justice Kagan recognized that the Court bore some responsibility for creating the impression that Auer entailed the abdication of judges’ interpretive duty. If once it was, however, Auer is no longer “a revolutionary shift of authority from the judiciary to the executive.” That Auer is dead. Long may Auer reign.

Evan Bernick is a visiting lecturer at Georgetown University Law Center and a fellow of the Georgetown Center for the Constitution.