Symposium: Shadow boxing with the administrative state

Thomas W. Merrill is the Charles Evans Hughes Professor at Columbia Law School. He filed an amicus brief in support of the challenger in Kisor v. Wilkie.

An important subcurrent in today’s Supreme Court concerns the legitimacy of the administrative state. The court’s conservative wing wants to take down the administrative state a notch or two. The liberal wing is anxious to shore it up. This was abundantly clear in Gundy v. United States, where the conservative four announced a willingness to reconsider, at least in some future case, the “extraordinarily capacious” standard for considering whether Congress has impermissibly delegated legislative power to the executive branch. The liberal four insisted that the existing, highly deferential standard was settled law and should not be disturbed.

The same sort of division was very much in evidence in Kisor v. Wilkie. The Supreme Court granted cert to consider whether to overrule the highly deferential standard of review of agency interpretations of their own regulations, first set forth in Bowles v. Seminole Rock & Sand Co. in 1945 and more recently associated with Auer v. Robbins. When the dust settled, the case revealed very broad agreement among the justices. No one defended the unadorned standard in Auer, which says the agency interpretation is entitled to controlling weight unless it is “plainly erroneous or inconsistent with the regulation.” At the same time, no one defended de novo review, which would give no weight to the agency view. Instead, all nine justices appeared to believe that a more contextual approach is required, applying traditional tools of interpretation to determine whether the regulation is genuinely ambiguous, and even then, deferring to the agency view only after considering such factors as whether the agency interpretation is authoritative, the agency has expertise in the matter, the agency interpretation is well considered, and the interpretation has been consistently maintained. The judgment was unanimous to remand to the U.S. Court of Appeals for the Federal Circuit to reconsider its decision under the new contextual standard.

Shadow boxing over the administrative state emerged in a sharp disagreement about how to characterize the new standard of review. Writing in part for the court and in part for a plurality consisting of the four liberals, Justice Elena Kagan argued that Auer should not be overruled but merely clarified as having adopted the contextual approach. The fifth vote to reject overruling was provided by Chief Justice John Roberts, who nevertheless declined to join parts of the Kagan opinion. The Kagan approach (which effectively garnered five votes on this point) would thus preserve the label “Auer deference” for the new standard of review. Writing for the remaining four conservatives, Justice Neil Gorsuch argued that Auer should be overruled, and should be replaced by the contextual standard of review associated with Skidmore v. Swift & Co. This would have eliminated the name “Auer deference” and substituted in its place “Skidmore deference.”

The shadow boxing over the legitimacy of the administrative state also emerged in the argumentation advanced in support of the competing choice of characterizations of the contextual approach. In arguing for preserving Auer in name if not in substance, Kagan had many supportive things to say about the importance of administrative discretion and expertise. In support of overruling, Gorsuch emphasized the anomalous nature of agency government from a traditional separation-of-powers perspective. The fact that everyone ended up in largely the same place warrants calling the disagreement shadow boxing.

Lawyers will want to know if there is any meaningful difference between Kagan’s contextualized Auer and Gorsuch’s contextualized Skidmore. I would characterize Kagan’s approach to contextualization as a kind of step zero for Auer (or more accurately, a combination of step zero and step one), borrowed from the jurisprudence associated with Chevron U.S.A. Inc. v. Natural Resources Defense Council. Before deferring to the agency interpretation, the court must make a number of threshold determinations such as whether the regulation is truly ambiguous, whether the agency interpretation falls within the scope of the ambiguity identified, whether the issue is a technical one as to which the agency has superior familiarity, whether the agency has flip-flopped in its interpretation, and so forth. If the tumblers all line up in support of deference, then the court should adopt the agency view. This imports the sequencing familiar in the Chevron context from United States v. Mead Corp., although the characterization of step zero here is clearly of the multi-factor, all-things-considered variety, as opposed to the single-factor approach (did the agency speak with the force of law?) which is at least one reading of Mead. The key point would be that, if the court grinds its way through all the factors relevant to step zero and step one, then the agency view must be adopted.

The Gorsuch approach to contextualization would replace Auer with Skidmore. This draws upon roughly the same contextual factors invoked by Kagan. But the difference would seem to be that under Skidmore, deference exists on a sliding scale, rather than an all-or-nothing conclusion that emerges after a sequential inquiry. The court remains responsible for the interpretation, and whether the court adopts the agency view depends on how the various contextual factors stack up, either for or against the agency. The more the factors favor the agency, the more “persuasive” the agency view becomes, but at no point is the court compelled to adopt the agency view.

If this characterization is correct, there are arguably two differences between Kagan’s version of contextualization and Gorsuch’s. One difference is that the Gorsuch approach adopts an established standard of review – Skidmore. Like other multi-factor standards, this is highly indeterminate, and subject to different outcomes in the hands of different judges. But at least Skidmore is a standard that has been around for a long time – since 1944 to be exact – and has accumulated a body of precedent and gained a degree of familiarity with judges. Kagan’s new contextualized Auer, although it draws upon roughly the same factors as Skidmore, is an unknown animal at this point. Consequently, it is likely to produce significant uncertainty among lower court judges, agencies and persons contemplating a challenge to agency interpretations. This difference, in my view, counseled in favor of adopting Skidmore rather than rewriting Auer.

The second difference involves whether the agency is free to change its interpretation. Under Kagan’s sequencing approach, the agency should be able to change its interpretation, provided the sequencing continues to favor deference to the agency. Under Skidmore, the interpretation is ultimately the court’s, which means the agency may not be able to change its interpretation. (Justice Antonin Scalia made this point in his dissent in Mead.) Of course, insofar as agency consistency is one of the contextual factors under either approach, the agency’s ability to change its interpretation may be constrained under either approach. So I would not give this difference great weight one way or the other.

If the legitimacy of the administrative state was the subcurrent of Kisor, the Supreme Court’s willingness to overrule its prior decisions was a strong cross-current. The court has undoubtedly become sensitized to overruling. The ultimate cause is the agitation on the left about the prospect of the overruling of Roe v. Wade. The liberal bloc of justices stokes this concern, by appending ominous-sounding “what’s next?” codas to dissenting opinions, as in Kagan’s dissent in Knick v. Township of Scott and Justice Stephen Breyer’s in Franchise Tax Board of California v. Hyatt.

My view is that the agitation about overruling Roe is largely artificially contrived. Roe was effectively overruled in Planned Parenthood v. Casey, which adopted a contentless “undue burden” standard for reviewing abortion regulations. This allows the Supreme Court to give a thumbs up or thumbs down to virtually any regulation, obviating the need to revisit ultimate questions. Unless the court wants to make itself the object of extreme controversy, it will continue to follow Casey, and leave Roe to gather dust.

Whether I am right about this or not, it is conceivable that aversion to overruling helps explain why Kagan and the liberals opted for rewriting Auer rather than overturning it. It may also explain why Roberts decided (perhaps at the last minute?) to cast his lot with Kagan rather than Gorsuch. The chief justice clearly viewed the two approaches as largely equivalent. Rewriting Auer rather than overruling it may have seemed a small price to pay to avoid another editorial claiming that the court has become a partisan institution bent on overruling Roe.

A final point concerns the implications of Kisor for the future of Chevron. Roberts went out of his way to say that Kisor has no impact on Chevron, and Justice Brett Kavanaugh, joined by Justice Samuel Alito in a short concurrence, agreed. But I think Kisor foreshadows the ultimate fate of Chevron. Chevron, like Auer, will be displaced by Skidmore. The only question is whether Chevron will be progressively “Skidmorized,” while remaining authoritative in name, or will be overruled and replaced by Skidmore. Judging by what happened in Kisor, the court for institutional reasons is likely to find the former strategy more congenial.