What is the Best Interpretation for Purposes of Determining a Not Best Interpretation for Chevron Deference? (10/21/22; 11/8/22)

Last week, I participated in a panel at the ABA Tax Section Fall Meeting in Dallas. The panel was offered by the ABA Tax Section Teaching Taxation Section. The program was titled “Classification of Tax Regulations and the APA.” The panel participants were Les Book (moderator), Kristin Hickman, Gil Rothenberg, and me. The program dealt in part with some issues that Professor Hickman and I have engaged in the past. On those issues, Hickman and I continue to engage without agreement. 

I say "without agreement," but in fact Hickman and I did agree (i) on the starting point that, at the enactment of the APA in 1946, the APA permitted both legislative rules (must be notice and comment regulations) and interpretive rules (including, although not required, notice and comment regulations such as Treasury used for interpretive rules adopted as regulations) and (ii) that for well after the APA, the category of interpretive notice and comment regulations was alive and well, without controversy. Our point of continuing disagreement is whether something has changed that original meaning of the APA, so that notice and comment regulations that do no more than interpret ambiguous statutory text no longer exist but are now legislative regulations. That disagreement is not the focus of this posting but may underly some of the analysis in this posting.

In the panel discussion, I repeated my recent claim that Chevron deference is inapplicable if the agency interpretation is the best interpretation. The corollary to this is that Chevron’s domain in an outcome determinative sense involves only deference to “not best” agency interpretations. We only touched on that issue lightly in the panel, but one panelist questioned whether "best interpretation" was a meaningful concept. (Added 10/25/22 11:00am:  On Chevron's domain as limited to not best interpretations, see e.g., Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 980 (2005) (“Chevron requires a federal court to accept the agency’s construction of the statute, even if the agency’s reading  differs from what the court believes is the best statutory interpretation.”).)

I thought it might be helpful to say further what I mean by “best interpretation.”  The best (or most persuasive) interpretation is the interpretation which the court would apply based solely on interpretive tools in the absence of an agency interpretation or, if there is an agency interpretation, in the absence of Chevron deference. The court’s job in statutory interpretation is to determine and apply the best interpretation of the statute to the facts before the court. (BTW, it is important to note that courts cannot legislate, so courts’ exercise of interpreting a statute to determine the best interpretation is not legislative.  See Is Statutory Interpretation a Legislative Act When Agencies Do It But Not When Courts Do It? (Federal Tax Procedure Blog 4/8/22; 10/23/22), here.). 

If the agency interpretation is that best interpretation, the court applies that interpretation because it is the best interpretation and not because it is deferring to the agency interpretation. Of course, in determining the best, most persuasive interpretation, the court should consider respectfully the agency interpretation. Skidmore v. Swift & Co., 323 U.S. 134, 140 (1944) (Skidmore respect is not deference (although frequently mislabeled as deference) and applies at Chevron Step One as well as in generally determining the most persuasive interpretation.) See Really, Skidmore "Deference?" (Federal Tax Procedure Blog 5/31/20; 2/14/21), here. Deference—real deference--is only outcome determinative if the court invokes Chevron to apply a “not best” interpretation over the court’s own best interpretation.

That best interpretation concept is simple and, I think, irrefutably true. There will always be a “best interpretation” so defined because, in the absence of an agency interpretation, it is the interpretation the court would apply to decide the case. In deciding whether or not to apply Chevron deference to an agency interpretation, a court will have to know what the best interpretation is (i.e., what the court would do without deference to the agency interpretation). How else could a court determine whether an agency interpretation is “reasonable” or “permissible” even though not best (the requirement for Chevron deference)?

I suppose it is conceptually possible that a court may be in a state of interpretive equipoise, where the court cannot determine the best interpretation of the statutory text. First, I infer that judges (at least good judges capable of making decisions) are rarely (if ever) in interpretive equipoise. For example, Justice Scalia said (Antonin Scalia, Judicial Deference to Administrative Interpretations of Law, 1989 Duke L.J. 511, 520):

“If nature knows of such equipoise in legal arguments, the courts at least do not * * * [Positing the following question] ‘When are the arguments for and against a particular statutory interpretation in equipoise?,’ I am certain that the response would be ‘almost never.’”

Justice Gorsuch apparently feels the same, quoting Justice Scalia in Kisor v. Wilkie, 588 U.S. ___, 139 S.Ct. 2400, 2429-2430 (2019) (Gorsuch concurring); see also Justice Kavanaugh's similar statement in Kisor v. Wilkie, at 2448 (concurring) (when "a reviewing court employs all of the traditional tools of construction, the court will almost always reach a conclusion about the best interpretation of the [law] at issue."); but see E.g., Cass R. Sunstein, Chevron as Law, 107 Geo. L.J. 1613, 1627 (2019) (discussing interpretive equipoise in the context of the Chevron Step One inquiry into whether a statute is ambiguous and noting the constriction of agency interpretive space if Chevron applied only in interpretive equipoise and expansion of interpretive space if Chevron applied (as Justice Scalia suggests) beyond the state of equipoise).  [Note the cite to Justice Kavanaugh's concurring opinion in Kisor w. Wilkie was added on 11/8/22 at 1:45pm.]

Even assuming interpretive equipoise can occur, what does a court do if the court is in interpretive equipoise but must decide the case at hand? Flip a coin? Check the court’s gut? Consult an augur? Sleep on it and hope the answer comes in a dream? Make the determination of the law by which side of the bed the court wakes up on? I could go on, but I am sure readers get the point. 

Interpretive equipoise (to the extent it exists) requires a tiebreaker to resolve the case before the court. Fact-finding equipoise is resolved by burden of proof assignments, where equipoise requires that the party having the burden of persuasion loses. Is there a burden of persuasion as to the best interpretation of the statute, so that the party bearing that burden loses if the court is not persuaded that the best interpretation supports that party? If there is such a statutory interpretation tiebreaker, then how is the burden assigned? Is it possible that interpretive equipoise is the domain of Chevron deference? And, if an agency interpretation wins via some tiebreaker rule, then the court still is not deferring to a “not best” agency interpretation.

If there is interpretive equipoise, what is the zone of equipoise? This issue came up for me as I considered the requirement for a “more likely than not” (“MLTN”) legal opinion issued for tax shelters. In relevant part, the MLTN opinion is supposed to interpret the law and apply it to the facts, much the way a court does in resolving a case. Many said that, well, for the tax shelter MLTN opinion, the interpretive likelihood need be only 51%. But how does one tell the difference between 51% likelihood and 49% likelihood? In truth, the range of equipoise, conceptually, would be from, say, 45% to 55%, with equipoise being in between because no finer calculation of the probabilities of the law interpretation can be made.   (I provide at the end of this blog some previous Federal Tax Procedure blogs dealing with aspects of the equipoise issue.)  So, in this sense, there might be a range of equipoise where "deferring" to the agency interpretation might be meaningful, but is not deferring to a not best agency interpretation.

The key point of this blog is that there can generally be a best interpretation of the statute. No deference is needed or given when a court applies the best interpretation. The further question then is what is the domain of Chevron deference to be outcome determinative. I dealt with the equipoise issue above, and I suppose that, if Chevron deference were limited to interpretive equipoise (a rare phenomenon), that might be very limited scope. But, given the escape hatches in Chevron, will a court really apply Chevron deference to bless a not best agency interpretation when it is persuaded there is a better interpretation? See The Impact of Chevron Deference is Exaggerated (Federal Tax Procedure Blog 8/19/21; 8/21/21), here.

More postings on Chevron and equipoise.

• More on Fact Finding Tools and Statutory Interpretation through Chevron Deference (Federal Tax Procedure Blog 1/21/19; 1/25/19), here;
• Even More on Skidmore (Including Equipoise as to Interpretation) (Federal Tax Procedure Blog 7/7/19), here; and
• Chevron and Equipoise In Statutory Interpretation (Federal Tax Procedure Blog 5/26/22), here.