More Thoughts on APA and Legislative and Interpretive Regulations Inspired by Recent Cases (4/8/21; 4/11/21)

As readers of this blog know, I have recently been considering and even obsessing over the distinction between legislative and interpretive regulations for purposes of the APA and Chevron deference.  Just in the past few days, I have focused on four opinions that I think highlight the confusions in this area.  The first three are decisions by three different Circuits on the issue of the validity of a recently adopted regulation including bump stocks within the statutory definition of “machineguns,” thus prohibiting their possession with potential criminal penalty consequences.  Those decisions are:

• Guedes v. BATFE, 920 F.3d 1 (D.C. Cir. 2019), cert. denied, 140 S. Ct. 789 (2020), D.C. Cir. here and GS here; 
• Aposhian v. Barr, 958 F.3d 969 (10th Cir. 2020), 10th Cir. here and GS here; and
• Gun Owners of Am., Inc. v. Garland, 992 F.3d 446  (6th Cir. 3/25/21), 6th Cir. here; GS here.  [Note that the lengthy decision was published on the same day as oral argument.]

The regulation was sustained in Guedes and Aposhian as a Chevron-entitled interpretation of the ambiguous statutory text; the regulation was invalidated in Gun Owners as (i) violating separation of powers and the rule of lenity and (ii) in any event not passing Chevron analysis as a reasonable interpretation of ambiguous statutory text.

The Fourth Opinion, a truly monster opinion, was an en banc opinion in Brackeen v. Haaland, 994 F.3d 24 (5th Cir. 4/6/21), 5th Cir. here; GS here.  The opinions are, in the aggregate, 325 pages long and the portions of the opinions that represent the en banc opinion are scattered in two separate opinions – that of Judge Dennis and that of Judge Duncan.  Here is the Court’s description:

Dennis, J., delivered the opinion of the en banc court with respect to Parts II(B), II(C), and II(D)(2) of his opinion (except as otherwise noted in the Per Curiam opinion, supra). 

Duncan, J., delivered the opinion of the en banc court with respect to Parts III(B)(1)(a)(i)–(ii), III(B)(1)(a)(iv), III(B)(2)(a)–(c), III(D)(1), and III(D)(3) of his opinion (except as otherwise noted in the Per Curiam opinion, supra).

The exercise of working through that stuff to figure out what the en banc holdings were would be daunting indeed.  So, the Fifth Circuit helpfully offers the first opinion, a per curiam opinion, to provide a guide to the en banc opinions embodied in the opinions with the named authors.  Fortunately, for purposes of this blog entry, I and readers interested in the legislative – interpretive distinction do not have to dig through all of that morass.  In relevant part, Judge Dennis’ opinion has the en banc opinion at outline paragraph II.D.2., titled The Scope of the BIA’s Authority, on pages 138-147, beginning here.

I offer this high level summary with limited citations (I am revising my article on the subject for posting to SSRN later).  I begin with what I think should be the starting point for discussing the issue—the original meaning of the APA distinction between legislative and interpretive regulations and how that original meaning played out in deference.  I will then address the cases which evidence the distortions and distractions that have obscured the original meaning.

Original Meaning of the APA

Prior to the APA, there was a well-understood distinction between legislative regulations and interpretive regulations.  Legislative regulations were those regulations in which, pursuant to express delegation from Congress to make the law, the agency made the law.  The classic tax example is express delegation in § 1502 to the Treasury to make the law of consolidated return reporting.  Treasury exercised that authority in the consolidated return regulations.  Section 1502 is, as statutes go, crisp (2 sentences, 111 words).  The key statutory text is: “The Secretary shall prescribe such regulations as he may deem necessary in order that the tax liability of any affiliated group of corporations making a consolidated return and of each corporation in the group * * * *.”   The resulting regulations go from § 1.1502-0 Effective dates to § 1.1502-100 Corporations exempt from tax (with some gaps in the numbering); the regulations offer mind-numbing complexity.  The consolidated return regulations are the law; they are not interpretations of the law.  A court could not, by interpreting § 1502, derive the consolidated return rules in the absence of the regulations which, by delegation, are statute-equivalent laws.  Since legislative regulations are the law and not interpretations of the law, courts must respect those regulations “more than mere deference or weight.”  United States v. O'Hagan, 521 U.S. 642, 673 (1997).  That is because deference or weight is an oxymoron for legislative regulations.  This will become clearer when I discuss deference below.

Interpretive regulations, by contrast, are not the law but are interpretations of ambiguous statutory text.  The law is the statutory text.  The interpretation is just that – an interpretation, an interpretation as to which Courts may or may not agree and when they don't agree may sometimes defer to under the doctrine of deference which now appears in the Chevron deference framework.  Deference to an agency interpretation does not make the interpretation legislative in character, just as a court’s interpretation of ambiguous statutory text does not mean that a court is legislating (a power it does not have).

Further, with respect to interpretive regulations, at the  time of enactment of the APA, regulations adopted under the Treasury’s general rulemaking authority in the predecessor to § 7805(a) were interpretive regulations that might attract deference if otherwise appropriate.  When I say otherwise appropriate, there was no clear framework to test of when to defer, deference was often given to interpretations that were reasonable interpretations of ambiguous statutory text.  This is key:  whatever motivated a court to defer, a court would only defer to reasonable interpretations of the ambiguous statutory text  Chevron, decided in 1984, offered that framework long after the APA but, as before, Chevron permitted deference only to reasonable agency interpretations.

In my mind, a good example of interpretive regulations under § 7805(a) is United States v. Correll, 389 U.S. 299 (1967).  In a regulation, Treasury interpreted that statutory text to require stopping for sleep or rest, an overnight rule; in Correll, the Supreme Court adopted–via deference–that interpretation.  The Correll Court said: “The role of the judiciary in cases of this sort begins and ends with assuring that the Commissioner's regulations fall within his authority to implement the congressional mandate in some reasonable manner.”  There were several possible interpretations of the away from home statutory requirement; the agency chose one and adopted it in § 7805(a) regulation; the interpretation was a reasonable interpretation; the Court deferred to it.

At the time of the APA (and I think thereafter), deference only applied to interpretations and thus only applied to interpretive regulations.  Deference did not apply to legislative regulations.  Indeed, conceptually, deference was an oxymoron for legislative regulations which were the law and not interpretations of the law.

There is a lot of detail behind this high-level summary of what the law was at the enactment of the APA, but I think it is a correct summary that virtually everyone observer agreed upon contemporaneously with the APA and for most of the life of the APA.

Blurring the Distinction Between Legislative and Interpretive Regulations

As exemplified in the cases I cite and link at the start, this original meaning of the APA and deference no longer applies at least if one believes dicta.  In each of the cases, the Court said that a regulation that did no more than interpret ambiguous statutory text was a legislative regulation as a predicate to applying the Chevron framework to determine whether deference was appropriate.  As noted, however, deference to a legislative regulation which is the law and is not an interpretation  of the law is an oxymoron.  Deference applies to interpretations not law.

So, if that is true, why did these courts feel compelled to get to Chevron deference through first characterizing the regulations as legislative?  That is because courts (these and others) have confused Chevron and by repeating the confusion over and over have come to believe that Chevron applies to legislative regulations, with some even extravagantly claiming that it applies only to legislative regulations.  That sort of nonsense would be avoided if they just focused on the actual holding in Chevron, for as Justice Scalia noted “Chevron in fact involved an interpretive regulation.”  Christensen v. Harris County, 529 U.S. 576, 589-90 (2000) (concurring in part and concurring in the judgment).  Thus, Chevron fit squarely into the original meaning.  No one, including the author of Chevron at the time thought that Chevron departed from the original meaning.

In fact, in many (perhaps most) in which this claim of Chevron deference for legislative regulations is made, it is probably dicta because the courts apply Chevron deference to an interpretation in the same way that they would if they properly characterized the regulation as interpretive rather than legislative.  That is true in each of the cases  I cite above.

It is interesting that this same claim was made by the Tax Court in Altera Corp. v. Commissioner, 145 T.C. 91, 110-112 (2015), rev’d 926 F.3d 1061 (9th Cir. 2019), reh. en banc den. 941 F.3d 1200 (9th Cir. 2019), cert. denied, 591 U.S. ___, ___ S.Ct. ___, 207 L. Ed. 2d 1078 (2020).  The claim was nonsense for the reasons I have noted, and fortunately the Ninth Circuit just avoided dealing with the issue, instead focusing on the Chevron analysis without characterizing the nature of the regulation.

But other courts have not avoided the claim that regulations that do no more than interpret ambiguous statutory text are legislative on their journey to applying the Chevron framework.  Again, that is nonsense.

Perhaps I should stop there but I do want to step through some of the statements in the monster opinion in Brackeen v. Haaland because there are some nuggets in there as the court stumbled through its legislative regulation dicta in getting to Chevron.  I am focusing on part II.D.2. of Judge Dennis’ opinion which states the en banc holding on these issues.

1. The regulation in question was adopted pursuant to a general authority statute to promulgate “rules and regulations as may be necessary to carry out the provisions” of ICWA. 25 U.S.C. § 1952.  That is basically the same as § 7805(a) which is not like a legislative authority delegation.  See e.g., § 1502 for consolidated return regulations.  At all relevant times (enactment and each amendment), § 7805(a) has been treated by Congress consistent with the historic interpretation that it permitted interpretive regulations only.  (I get into that in great detail in revised version of my article, where I trace the original enactment of the predecessor to § 7805 through its most recent relevant amendment in 1996.) As noted a similar general authority statute was involved in Brackeen.  The problem is that, because the courts somehow bought into the notion that a legislative regulation was required for Chevron deference, they began claiming that such general authority delegations were legislative rather than interpretive.  That was just wrong for the reasons noted above.

2. The Brackeen Court does not use the terms legislative regulation or interpretive regulation but does talk about “binding regulation” which might be viewed as a proxy for legislative regulation.  But I think the court’s analysis is that the regulation is or can be a "binding" regulation only if it passes the Chevron test of reasonable interpretation.  But, if the court were using “binding regulation” as a proxy for legislative regulation, that was dicta because, whether the regulation is legislative or interpretive, the same Chevron test applies in the same way (as the court articulated it to test the reasonableness of an interpretation).  If the court were not using “binding regulation” as a proxy for legislative regulation but was simply saying that the regulation was binding if it passed muster under Chevron, then well of course, because Chevron applies only to interpretive regulations.

Finally, although veering from current cases, consider this (which I cut and paste from my working draft of my revised article (footnotes omitted:

In United States v. Home Concrete & Supply, LLC, 566 U.S. 478 (2012) the Court considered the Treasury attempt to apply retroactively interpretive regulations (both temporary and final) issued under § 7805(a).  Since the statute being interpreted was enacted prior to the effective date of the 1996 § 7805(b) amendments, the prior version of § 7805(b) applied to permit retroactive application without those limits consistent with the general treatment of interpretive regulations. The Supreme Court treated the regulations as interpretive regulations susceptible to the Chevron Framework, but stopped the analysis at Chevron Step One, holding that its prior interpretation in Colony, Inc. v. Commissioner, 357 U.S. 28 (1958) made the statute unambiguous, thus precluding any Chevron interpretive space.  But, if the regulation in question were legislative, the Court would (at least, should have) have tested under the arbitrary or  capricious standard rather than under Chevron and, most importantly, would have not permitted retroactivity for a § 7805(a) “legislative” regulation which requires Prospectivity.  So, in effect, if the regulations were legislative, the IRS would have had no possibility of prevailing in the flurry of cases that worked through the system culminating in Home Concrete, and the courts, including the Supreme Court, just did not know what they were doing and were wasting their time dealing with the issue that would not exist if the regulation were legislative rather than interpretive.  But, as an interpretive regulation, the regulation could be tested under Chevron and could be retroactive if they passed Chevron reasonable interpretation test   Hence, there was a real issue at stake in Home Concrete and the other cases, but only because the regulations were § 7805(a) interpretive regulations.  

This was precisely Justice Kennedy’s point in dissent (cleaned up):

The Department's clarification of an ambiguous statute, applicable to these taxpayers, did not upset legitimate settled expectations. Given the statutory changes described above, taxpayers had reason to question whether Colony's holding extended to the revised § 6501(e)(1). Having worked no change in the law, and instead having interpreted a statutory provision without an established meaning, the Department's regulation does not have an impermissible retroactive effect. It controls in this case.

Justice Kennedy thus clearly understood that the regulation was interpretive, that interpretive regulations exist and that as interpretive regulations they were not subject to the Prospectivity requirement for legislative regulations. The point of disagreement between the majority and Justice Kennedy was not whether the regulations were interpretive but whether, as interpretations, they failed at Chevron Step One.  And, to close that loop, Chevron Step One could not have even been reached if the regulations were legislative rather than interpretive.

Added 4/9/21 1:30pm  I add the following to extend the discussion for some points that are not explicit in the earlier discussion.

The spurious notion that § 7805(a) regulations that reasonably interpret ambiguous statutory text are legislative regulations is a central feature of the following claim, quoted prominently in the short single paragraph Conclusion of CIC Servs LLC v. United States, 925 F.3d 247, 258 (9th Cir. 2019), reh. en banc den. 936 F.3d 501 (6th Cir. 2019), cert. granted 206 L.Ed.2d 916 (2020) (which in turn quoted Kristin E. Hickman & Gerald Kerska, Restoring the Lost Anti-Injunction Act, 103 Va. L. Rev. 1683, 1686 (2017)):

Defendants [Treasury and IRS] "do not have a great history of complying with APA procedures, having claimed for several decades that their rules and regulations are exempt from those requirements."

The notion is that § 7805(a) regulations, being legislative, are subject to all the APA requirements for legislative regulations – including in particular (i) either prospective only application (which the Supreme Court has said is the general requirement for statutes and legislative regulations which are like statutes) and (ii) in all events, no retroactivity to the date of the statute.  Specifically,  if that claim were true, a Treasury Temporary Regulation (and a subsequent Treasury Final Regulation) which reasonably interprets ambiguous statutory text (i) cannot be retroactive to the date of the statute enactment; and (ii) can only be immediately effective as of the date of the Temporary Regulation if accompanied by a good cause statement.  Treasury historic practice was that, as interpretive regulations, Treasury Temporary and Final Regulations were not subject to those APA requirements imposed on legislative regulations.  And courts have routinely either (i) approved interpretations in Temporary Regulations with retroactive effect or (ii) not approved them without suggesting that retroactivity violates the APA.  If indeed the § 7805(a) regulations that do no more than interpret ambiguous statutory text are interpretive regulations, then the claims  of Treasury and IRS noncompliance are false. 

Indeed, as I have  previously noted, although this claim was a central feature of the CIC Services opinion (which I quote above), neither the parties nor the amicus in their respective Supreme Court briefs chose to repeat this spurious claim.  Hopefully, it will suffer the fate of the similar spurious claim by the Tax Court in Altera, that the Court of Appeals chose to ignore in its Chevron analysis.

I should say something about the limitations on retroactive effect imposed in § 7805(b).  Those who claim legislative status for § 7805(a) regulations that only reasonably interpret statutory text have to claim then that the permitted retroactivity in subsection (b) are exceptions to the APA prospectivity requirement.  One facial  problem with that claim is that an exception to the APA’s requirements (including prospectivity) cannot apply “except to the extent that it does so expressly.”  5 USC § 559.  Section 7805(b) does not expressly state that they are exceptions to the APA requirement of prospectivity.  

Indeed, there is a reason that § 7805(b) does not provide the express exception – because all concerned in the process of enacting all of § 7805 and all relevant amendments understood that § 7805(a) permitted only interpretive regulations when ambiguous statutory text was being interpreted and chose the language of § 7805(b) as a limitation on retroactivity for interpretive regulations rather than an exception to prospectivity required  for legislative regulations.  

Even thus properly viewed as an exception to retroactivity for interpretive  regulations, despite § 7805(b), the interpretation can still have full retroactive effect if it is the best interpretation of ambiguous statutory text under a Skidmore analysis; the regulation is not retroactive but the interpretation can be regardless of § 7805(b). Section 7805(b) thus at best applies to mean that a regulation is invalid, not that an interpretation is invalid.  If the agency interpretation is the best interpretation, does the court ignore it simply because the regulation including the interpretation fails § 7805(b)?  Of course not.

In  that light, the only possible effect of § 7805(b) is when a reviewing court is in equipoise as to the best interpretation (although unable to find agency interpretation is affirmatively the best, there is at least one other reasonable interpretation that is neither better nor worse than the agency interpretation).   In that case, what is a court to do?  Flip  a coin  as to the interpretation?  Or check the court's gut?  Or consult an augur?  Or choose against the agency interpretation even when the nonagency interpretation is not better?  Would any of those be  justice?  Would a court really do that?  Would a court still not  be inclined to tilt to the agency interpretation in a state of equipoise for all the reasons the Chevron court used to justify deference (agency expertise and more holistic view of the statute than the court’s anecdotal view in a case)?

Consider the following from Justice Breyer for the majority opinion  County of Maui, Hawaii v. Hawaii Wildlife Fund, 590 U.S. ___, ___, 140 S. Ct. 1462, 1462 (2020):

Neither the Solicitor General nor any party has asked us to give what the Court has referred to as Chevron deference to EPA’s interpretation of the statute. See Chevron U. S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U. S. 837, 844 (1984). Even so, we often pay particular attention to an agency’s views in light of the agency’s expertise in a given area, its knowledge gained through practical experience, and its familiarity with the interpretive demands of administrative need. See United States v. Mead Corp., 533 U. S. 218, 234–235 (2001); Skidmore v. Swift & Co., 323 U. S. 134, 139–140 (1944).

And the following from Smiley v. Citibank (South Dakota), N. A., 517 U.S. 735, 744 n.3 (1996).

Where, however, a court is addressing transactions that occurred at a time when there was no clear agency guidance, it would be absurd to ignore the agency's current authoritative pronouncement of what the statute means.

In truth, as Justice Scalia noted, courts are rarely in doubt as to legal interpretations so such a state of equipoise may be extremely unlikely.  (Courts have made similar observations about fact finders rarely being in a state of equipoise.)  The court will either find the agency interpretation the best and apply it for that reason wholly apart from Chevron or find some other interpretation the best interpretation which it can apply by concluding that, given the court’s best interpretation of the statute, the agency interpretation is not reasonable (which could defeat the agency interpretation at either Step One or Step Two, depending upon how you conceptualize the two steps).  All  of  that is to say that despite the claims about Chevron, it is a concept of limited effect in determining actual outcomes.  If that is right, the claims of horrors about Chevron are more about political angst over the perceived evils of the  administrative statute than about anything that is (or should be) practical.  (Other observers have noted that by constricting Chevron’s two steps (or one step if so conceived), the practical application of Chevron can be severely restricted to mitigate any claims of horrors; in a similar vein, others have said that much of the case discussion about Chevron is often not outcome determinative so that the claims of hits horrors are exaggerated.)

In my discussion above, when discussing Chevron outside the context of discussing its two steps, I am generally referring to the concept of  deference generally rather than its specific iteration in Chevron.  Chevron is just a specific framework for applying deference to reasonable agency interpretations that has been in the law for a very long time.  "Overruling" Chevron alone will not do anything (at least not much).  Eliminating deference, at least conceptually, would do something but would require overruling many more cases  than Chevron.  Even then, deference may rise up like a Phoenix.  See Lisa Schultz Bressman & Kevin M. Stack, Chevron Is a Phoenix, 74 Vand. L. Rev. 465 (2021), here.

JAT Notes 

1. (added 4/11/21 1pm):  On the 325 page length for the Brackeen opinion with all the ins and outs commotions, the ABA Journal (here) reports that one lawyer, Mary Kathryn Nagle, who specializes in federal Indian law, said "that the court could have answered the legal questions in a 10- or 20-page legal decision."  (That may not be an exact quote from the lawyer but is presented as the sense of what she said.)  I can't speak to that because the opinions wander around a lot of territory but if the Court had just cut to the chase on the Chevron analysis without detouring through the legislative rule nonsense, it could have probably cut 2 pages out and Chevron was a relatively small part of the overall opinion.  As I note above, the Ninth Circuit in Altera facing the same issue after the Tax Court found an interpretive regulation to be legislative, just ignored that issue and went straight to Chevron.  Chevron clearly applied, the only issue in both Altera and Brackeen was whether, since it clearly applied, any discussion of the legislative-interpretive distinction was necessary, meaning that the discussion of that issue by the Court of Appeals in Brackeen and the Tax Court in Altera was anything other than dicta because not necessary to the holding that Chevron applies.

2. I developed above an analysis that assumes that there was an original meaning of the APA as enacted by Congress and that the original meaning was recognized at that time the APA was enacted.  The overwhelming consensus of authority, certainly contemporaneous authority, was that there was such an original meaning.  Those who claim that the original meaning no longer applies, however, have considerable difficulty (in my opinion) justifying the claim.  Certainly, Congress could change the original meaning by amending the APA and making it clear that general authority delegations such as § 7805(a) are legislative delegations subject to legislative regulations’ burdens (notice and comment and prospectivity) and benefits (the law as opposed to interpretation).  Congress has not done that.  

Rather, as best I understand it, those claiming the original meaning no longer applies rely upon some creeping language in cases that, they claim, has authoritatively overruled the original meaning of the APA.  When they get specific about which cases did that (or even if cases could do that), however, there is no consensus.  Professor Hickman advised me that the watershed case was American Mining Cong. v. Mine Safety & Health Admin., 995 F.2d 1106 (D.C. Cir. 1993).  I am not sure that in terms of pecking order priority, the D.C. Circuit could announce a revision of the original meaning; I would think that if a court could do that at all, it would be the Supreme Court.  Further, as I develop in my article, however, American Mining Congress is totally consistent with the original meaning of the APA.  

There are other cases that seem to go back to some loose language in Chevron.  I have addressed that loose language in the article.  But one thing is clear is that no one at the time Chevron was decided—and certainly not its author or Justice Scalia (an administrative law expert) in his famous 1989 article—thought that Chevron was inconsistent with the original meaning or changed the original meaning.  And none of the cases since have done it either except by deploying loose language, often serving no other purpose to getting to Chevron by needlessly (and erroneously) calling the regulation legislative—a detour which is dicta and erroneous dicta.