Is Chevron on Life Support; Does It Matter? (4/2/22; 4/3/22)

In teaching tax procedure for many years, I introduced my students to Chevron deference while discussing the false claim of “tax exceptionalism.” I define Chevron deference as a court deferring to an agency statutory interpretation that is less persuasive than the interpretation determined by the court. After all, if a court determines that the agency interpretation is the best interpretation, it does not defer to the agency interpretation. In a recent article titled  See John A. Townsend, The Report of the Death of the Interpretive Regulation Is an Exaggeration (SSRN last revised 12/15/21), here, I addressed the key feature for Chevron deference, the continued existence of the APA interpretive regulation category.  As a Postscript to that article I dug in deeper on Chevron deference.  Readers can view that Postscript separately here. In the Postscript, I review my evolved thoughts on Chevron, developing analytical Categories for potential deference to agency interpretations.  (Those Categories are below.)  I inferred that Chevron deference to lesser agency interpretations is rarely outcome determinative. I base my inference on how I understand judges work (do they really apply less persuasive interpretations if they can avoid it?) and my reading of many cases citing Chevron.

An author posted an empirical analysis of cases where he identified 56 Federal Courts of Appeals cases as “granted Chevron deference” for the year ended November 30, 2021. Damonta D. Morgan, Chevron’s New Step Zero?: Measuring the Impact of Justice Gorsuch’s “Pecuniary Interest” Query (Notice and Comment 1/4/22), here. Mr. Morgan graciously shared his data set and permitted me to use it to test my inference.

I used the Categories to identify which cases in Mr. Morgan’s dataset involve deference to lesser agency interpretations. My categorization of the cases is in the table linked here; and explanations for the Categories assigned in the table are here [Note: the linked file when this blog was originally published was the wrong one; on 4/3/22 at 1:30 pm, I posted the correct linked file.] The aggregate results by Category are:

Category 0 - Cannot characterize but no deference; some could be like Chevron Step 0; this Category was not in the Postscript.	8
Category 1 - Statute unambiguous with no interpretive space for agency interpretation; no deference.	2
Category 2 - Statute ambiguous; no agency interpretation; court interprets; no deference	0
Category 3 - Court determines agency interpretation is best interpretation and applies it; no deference	8
Category 4 - Court determines agency interpretation unreasonable & applies the court's better interpretation; no deference	2
Category 5 - Court determines agency interpretation is reasonable but not best interpretation and defers to the agency interpretation anyway; deference (only category where court expressly deferred to a less reasonable agency interpretation)	0
Category 6 - Court determines agency interpretation is reasonable but makes no determination of best interpretation; applies agency interpretation; inconclusive deference	36
Category 7 - Court determines agency interpretation is reasonable but there is at least one other reasonable interpretation and court is in equipoise as to the best interpretation; court applies agency interpretation; no deference (because the court does not defer to a less persuasive agency interpretation)	0

The only relevant Categories to test my inference are Categories 5, 6 and,7 possibly, 7.  I wanted to present the other Categories (0-4) to illustrate a large swath of cases involving interpretations where there is no deference.  Actually 56 identified cases in a year that could possibly involve Chevron deference is a very small percentage court of appeals cases in the courts of appeals.  Mr. Morgan’s data set already excluded the bulk of the Category 0-4 cases because they involve no possible deference.  I have cases in those Categories where I disagree with Mr. Morgan that they were “granted Chevron deference.” Based on my assignment of cases to Categories, I conclude: 1. No court said or implied that it deferred to a less persuasive agency interpretation  (Category 5). 2. Courts deferred in 36 cases to “reasonable” agency interpretations without indicating that the agency interpretations were not the best; deference is inconclusive (Category 6). In this Category, my reading of the cases is that many and even most could be Category 3 no deference cases if the courts had determined the best interpretation. 3. In 20 cases (Categories 0-4), the courts gave no deference, although they noised about Chevron.   My broader observations are: 1. This data set is consistent with my inference that courts rarely defer to lesser agency interpretations. I recognize that there may be more unspoken deference in Category 6 than I infer.  2. Scholars have noted the decline, even the death, of Chevron deference in the Supreme Court, but some claim that Chevron deference is more common in the Courts of Appeals.  E.g., Nathan Richardson, Deference Is Dead (Long Live Chevron), 73 Rutgers U. L. Rev. 441, 445 & 485-493 (2021).  My analysis suggests that the claim may not account for Category 6 which is inconclusive as to deference.   3. Many of the Chevron discussions in cases in the data set and in cases I have read tend to be formulaic pablum without focus on what drives outcomes. This may be what Justice Kennedy lamented as “reflexive” Chevron in Pereira v. Sessions, 585 U.S. ___, ___, 138 S. Ct. 2105, 2120-2021 (2018) (Kennedy, J., concurring); see also Kisor v. Wilkie, 588 U.S. ___, 139 S.Ct. 2400, 2415 (2019) (Kagan majority opinion invoking Kennedy’s caution for reflexive Auer deference, like Chevron deference for subregulatory interpretation of agency regulations). Justice Kennedy’s description and other oft-voiced concerns about Chevron deference should result in less deference. Perhaps this phenomenon was in play in the post-Pereira data set from which I worked.  See Justices Discuss Limited Ambiguity Role for Lenity and by Analogy Chevron Deference (Federal Tax Procedure Blog 3/11/22), here. 4. In Chevron lingo, Step Two’s “reasonable” concept is ambiguous and flexible enough to mask a court’s approving the agency interpretation because it is the best interpretation.  Judge Newman of the Second Circuit recently observed that: (i) reasonable for Chevron has no set meaning; (ii) in some cases, courts “almost” seem to be “just accepting agency interpretations with which they agree and rejecting those they disfavor,” which is not deference; and (iii) Chevron “perhaps” offers “a narrow range of acceptable agency interpretations, on either side of the disputed issue, that courts are willing to uphold, but they are ready to assert the power to reject others that, for stated, or more often unstated, reasons, they deem beyond an amorphous notion of ‘reasonable.’” Jon O. Newman, On Reasonableness: The Many Meanings of Law’s Most Ubiquitous Concept, 21 J. App. Prac. & Process 1, 83 (2021),, downloadable pdf here. 5. If the Supreme Court or Congress were to reverse Chevron and any vestige of deference to lesser agency interpretations, case outcomes may not differ much in the aggregate. This would be particularly true if Skidmore respect is given to agency interpretations.  Although courts sometimes noise about Skidmore “deference,” Skidmore is not deference but an approach for a judge to determine and apply the best interpretation. See John A. Townsend, Really, Skidmore "Deference?" (Federal Tax Procedure Blog 5/31/20; 6/3/20), here.  If the judge applies the best interpretation there is no deference even if it is the same as the agency interpretation. Added 4/3/22 2:00pm: As originally published on 4/2/22, I included a Postscript to explain my description in the first sentence above of tax exceptionalism as being false.  Since that tax exceptionalism was not directly related to the principal topic of this blog, I have deleted it and made it a separate blog entry.