Rationality and Chevron (3/22/22)

I am reading Steven Pinker’s book Rationality: What It Is, Why It Seems Scarce, Why It Matters. (Penguin 2021) here, to lead the discussion in a book club.  At the same time, I am preparing an outline—yet another-- on Chevron deference.  The two exercises (reading Pinker’s Rationality and trying to make sense of Chevron deference) have converged. 

Chevron deference, as normally articulated, requires a court sometimes (at least as a conceptual possibility) to apply a “not-best” but “reasonable” agency interpretation of a statute in lieu of what the judge believes is the best interpretation.  In order to apply the best interpretation in lieu of the agency not-best interpretation, a judge can use several escape routes where the judge can fuss around about Chevron and still apply the judge’s best interpretation.  The two key ways built into the Chevron Two-Step Framework to avoid deferring to a “not-best” agency interpretation is (i) at Step One, to find the statutory text not ambiguous and (ii) at Step Two, to find the agency interpretation unreasonable.  There are other ways to avoid as well, such as the “major questions” doctrine and interpretations not within the agency’s expertise where the judges read the tea leaves to conclude that surely Congress did not intend the agencies should have primacy of interpretation authority.  E.g., King v. Burwell, 576 U.S. 473 (2015).

I will focus here on the escape hatches formalized in the Chevron Two-Step.

First, a judge can avoid the agency not-best interpretation by finding it unreasonable at Chevron Step Two.  Judges are trained to derive the best interpretation.  I think that is a judge's preference by education and practice.  I have wondered whether judges having a bias in favor of their own best interpretations might be motivated to find the agency not-best interpretations unreasonable.

Consider this recent observation by Second Circuit Judge Newman, one of the most esteemed Circuit Judges in the recent past (Jon O. Newman, On Reasonableness: The Many Meanings of Law’s Most Ubiquitous Concept, 21 J. App. Prac. & Process 1, 83 (2021).

It is difficult to know how the Supreme Court or other federal courts determine whether an agency’s interpretation of an ambiguous statute is “reasonable.” No weighing process appears to be involved. It would probably be too cynical to suggest that the courts are just accepting agency interpretations with which they agree and rejecting those they disfavor, but in some cases that almost seems to be what is happening. Clearly there is no one meaning of “reasonable” in the context of Chevron deference. Perhaps this is simply a context where there is 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.”

Does this sound like the definition of pornography which Potter Stewart famously proclaimed he could not define pornography but knew it when he saw it?  Could a similar process pre-empt any supposed Chevron command to defer to a “not-best” but reasonable (whatever that is) agency interpretation.

 Approaching the issue from Step One, all the judge has to do is rigorously interpret the statutory text and find the statutory text is not ambiguous.  Ambiguous suffers from the same uncertainty of meaning and application that Judge Newman finds in Chevron’s reasonableness test.  E.g., Aditya Bamzai, Delegation and Interpretive Discretion: Gundy, Kisor and the Formation of Future Administrative Law, 133 Harvard L. Rev. 164, 187 n. 140 (2019) (“The question of how to identify ‘ambiguity’ is a long-running one in both administrative law and elsewhere,” citing scholarly discussions); Brett M. Kavanaugh, Book Review: Fixing Statutory Interpretation, 119 Harv. L. Rev. 2118, 2118-2119 (2016) (“judges often cannot make that initial clarity versus ambiguity decision in a settled, principled, or evenhanded way”); and Ryan D. Doerfler, The “Ambiguity” Fallacy, 88 Geo. Wash. L. Rev. 1110 (2020) (“‘ambiguity,’ is critically ambiguous.”).

Now, going back to Pinker’s Rationality, Pinker presents (p. 90) an offering from the Dilbert series  that might even illustrate some arbitrariness in distinguishing the best argument from the not-best argument.

Of course that is hyperbole if applied to deference.  But perhaps Dilbert’s proffered decisionmaking process can be imported into Chevron deference (perhaps masked in some way). Some have noted that, in recent cases, Chevron is rarely invoked by the Supreme Court—indeed, some have even claimed even the death of Chevron in the Supreme Court.  E.g., Nathan Richardson, Deference Is Dead (Long Live Chevron), 73 Rutgers U. L. Rev. 441, 445 & 485-493 (2021).   My recent research of Courts of Appeals commoting around about Chevron for the one-year period ending 11/30/21 suggests that the commotion masks that the judges are applying their own best interpretation (as Judge Newman suggests).  As suggested, they can do that by finding no ambiguity in the statutory text and unreasonableness in the agency interpretation, thus permitting judges to apply the best interpretation whether the agency is the best interpretation or not.  There are other ways such as the major questions doctrine that for some undefined category of ambiguous statutory text, Congress cannot be assumed to have wanted the agency to have interpretive authority.  E.g., King v. Burwell, 576 U.S. 473 (2015).  And, courts can just ignore Chevron in a case in which it seems applicable.  

Postscript:  I highly recommend Pinker's Rationality.  You should either have some background in statistics or be prepared to struggle with statistics concepts, math and logic and symbols used in those disciplines.  Having read Rationality once, I don't have the feeling that I understood all or even most of what he says in the book.  I do plan to read most of it again, focusing particularly on the part I like best, Bayesian reasoning which requires also understanding many of the other concepts.  For a particularly good video discussion the book by Pinker himself, see here.