Friday, June 7, 2019

Ninth Circuit Reverses Unanimous Tax Court in Altera (6/7/19; 6/20/19; 7/2/19)

I have blogged on the Ninth Circuit's prior opinion reversing the unanimous Tax Court in Altera Corp. v. Commissioner, 145 T.C. 91 (2015) (reviewed opinion), here. Developments - Federal Tax Procedure Book 2018 Editions and Altera (7/25/18; 7/27/18), here. That opinion was reversed because it was rendered after one of the panelist died.  Ninth Circuit Withdraws Altera Opinions (8/7/18; 8/13/18), here.  Another judge was substituted for the deceased judge and oral argument was heard by the reconstituted panel.

The Ninth Circuit reconstituted panel, with all members apparently still alive, issued its opinion reversing the unanimous Tax Court.  Altera Corp. v. Commissioner, ___ F.3d ___, 2019 U.S. App. LEXIS 17143 (9th Cir. 2019), here.

Altera has been quite a saga, including the strange concept of a dead judge joining a majority opinion.  At the outset, it might be worth doing a tally of the judges on the merits.  In just the win-lose category.  There are two judges giving the win to the IRS, but they are the most important judges -- two of the three judges on the reconstituted panel.  All the other judges (other than the deceased Ninth Circuit judge who apparently voted before his death) who voted on merits held against the IRS.  Those judges are the dissenting judge on the reconstituted panel and all of the judges (15 in number) who voted on the reviewed opinion in the Tax Court.  So, just counting heads, two judges thought the IRS should win; 16 thought the IRS should lose.  (And this is not counting the dead judge's vote for the original panel opinion, which, if counted, would have been 3 for the IRS and 16 for the taxpayer.)  For those with the time to review an anecdote from my earlier appellate career at DOJ Tax for a Government appeal, like Altera, from a reviewed Tax Court opinion with most of the judges voting for the taxpayer, see Developments - Federal Tax Procedure Book 2018 Editions and Altera (7/25/18; 7/27/18), here.

Now to the current opinions from the reconstituted panel with living panel members.  The split is as it was in the withdrawn opinion.  Judge Thomas was for the IRS; Judge O'Malley from the Federal Circuit (by designation for the original and reconstituted panel) was for the taxpayer.  The swing judge was Judge Graber from the Ninth Circuit, designated to the panel to replace the deceased Judge Reinhardt.  Like Judge Reinhardt, the swing judge voted with Thomas whose opinion thereby became the majority just as with the withdrawn opinion.

I am focusing here only on the new panel majority and dissenting opinions.  I make no attempt to compare the differences between the withdrawn opinions and reconstituted panel current opinions; I just assume that, in broad strokes, the positions are the same (with some interim tweaking) since the same judges wrote the panel majority and dissenting opinions. (Readers interested in the withdrawn panel majority and dissenting opinions can look at my prior blog or Google any other comment on them.)  Readers interested in a discussion of the differences between the withdrawn and the current opinions might watch the Miller & Chevalier Tax Appellate Blog, here, because, in a quick posting on the blog on Friday, there the author said:  "Although it borrows heavily from the withdrawn opinion (indeed, much of the language remains similar if not the same), there are some notable differences between today’s opinion and the withdrawn opinion. We will post some observations after a more careful comparison."  Steve Dixon, Ninth Circuit Again Upholds Cost-Sharing Regulation in Altera (Tax Appellate Blog 6/7/19), here.

In broad outline, the panel majority opinion holds:

1.  Chevron Analysis.

   a.  Chevron Step One. Section 482 is ambiguous on the issue presented (whether the qualified cost sharing arrangement ("QCSA") must include employee stock option costs in allocating income from the intangible ).  Accordingly, Chevron Step One is passed.

   b.  Chevron Step Two.  The regulations' requirement that employee stock option costs be included in the QCSA costs is reasonable and therefore the interpretation that the court applies, by Chevron deference, in Chevron Step Two.

2.  State Farm Analysis.  The promulgation of the regulation requirement met the reasoned decisionmaking requirement and was not procedurally defective.

I am contemporaneously finishing up an article that I initially made available at the Virginia Tax Study Group at UVA law School in April that will include more discussion of the new opinions.  JAT addition on 6/20/19: The article has been posted on SSRN:  Townsend, John A., The Report of the Death of the Interpretive Regulation Is an Exaggeration (June 6, 2019). Available at SSRN:   The discussion of Altera begins on p. 100.

JAT Comments:

1.  I think the majority opinion gets it right.

2.  The majority opinion does not address the Tax Court's notion that the regulation in issue was a legislative regulation under the traditional APA distinction between legislative and interpretive regulations.  In my article, I argue that the regulation was an interpretive regulation under the APA, although one of the IRS's stipulations in the Tax Court may have, in effect, treated the regulation as legislative.  Instead of addressing the APA distinction, the reconstituted majority panel opinion just treats Chevron as a proper framework for analysis.  As I argue in my article, Chevron is a proper framework only for interpretive regulations and not for legislative regulations, except in the interpretive question as the scope of the legislative authority Congress delegated the agency.  Chevron is a test of reasonableness of an agency interpretation.  Since legislative regulations make law, testing law (rather than interpretation) for reasonableness of an interpretation of law (except within the scope of the delegation) is an oxymoron.  I will let that issue go for now, but I hope those interested will read my article when published on SSRN and offer me their insights on this fascinating issue.

3.  The Chevron Step Two analysis of reasonableness of the interpretation, I think, is fairly routine holding (Slip Op. 31]
In sum, Treasury reasonably understood § 482 as an authorization to require internal allocation methods in the QCSA context, provided that the costs and income allocated are proportionate to the economic activity of the related parties. These internal allocation methods are reasonable methods for reaching the arm’s length results required by statute. While interpreting the statute to do away with reliance on comparables may not have been “the only possible interpretation” of Congress’s intent, it proves a reasonable one. Entergy Corp. v. Riverkeeper, Inc., 556 U.S. 208, 218 (2009). Thus, Treasury’s interpretation is not “arbitrary, capricious, or manifestly contrary to the statute,” and it is therefore permissible under Chevron. 467 U.S. at 844.
That holding gets one nuance wrong.  The "arbitrary or capricious" standard should not apply in Chevron Step Two which is solely a test of the reasonableness of the agency interpretation, whereas, as the court noted earlier, the arbitrary or capricious/State Farm test is for procedural regularity not reasonableness of an interpretation.  As the panel majority said (Slip Op. p.33) quoting Catskill Mountains Chapter of Trout Unlimited, Inc. v. EPA, 846 F.3d 492, 521 (2d Cir. 2017) (cleaned up): “If an interpretive rule was promulgated in a procedurally defective manner, it will be set aside regardless of whether its interpretation of the statute is reasonable.”  The important point is that the procedurally defective regulation is set aside under State Farm, not under Chevron. I suppose one could say that an unreasonable interpretation is arbitrary or capricious, but that has not been the way that courts have applied Chevron except in some opinions, including Supreme Court opinions, that really botch Chevron with dicta and dicta of the worst sort, erroneous dicta (I go into this in detail in the article).  But, just to refocus, Chevron as articulated is solely a test of the reasonableness of the agency interpretation.

4.  After concluding that, under Chevron," Treasury’s interpretation of its statutory grant of authority was reasonable" [note that  the grant of authority was § 7805(a), which in APA speak is a grant of interpretive authority], the Court then concludes that promulgation of the regulations interpretation was not procedurally defective (Slip Opinion pp. 33 ff.).  In its arbitrary or capricious exegesis, the Court has comment that, I think, can also be considered in the Chevron analysis.  That comment is (Slip Op. 43):
Treasury articulated why treating stock-based compensation as a cost led to arm’s length results. It first noted that stock-based compensation is a “critical element” of R&D costs for parties to a QCSA and noted that such compensation is “clearly related to the intangible development area.” Compensatory Stock Options Under Section 482 (Preamble to Final Rule), 68 Fed. Reg. at 51,173. Logic supports these conclusions. Parties dealing at arm’s length, as Treasury explained, would not “ignore” stock-based compensation if such compensation were a “significant element” of the compensation costs one party incurs and another party agrees to reimburse when developing high-profit intangibles. Id. Rather, “through bargaining,” each party would ensure that the cost-sharing agreement is in its best interest, meaning that the parties will consider the internal costs of stock compensation without requiring the other party to recognize those costs. Id.
5.  That's pretty much as I see the guts of the majority holding on a winding down Friday afternoon even without wine.  I am sure others will be weighing forth on the majority and dissenting opinions for some time now.  And, I suspect, the taxpayer will move for rehearing (probably en banc because this panel will be done with further argument except to correct some obvious error which I did not see in the  opinions) and then, if no relief is obtained in the Ninth Circuit, petition for certiorari.  I have mixed feelings as to whether the Supreme Court should weigh into Altera.  On the Court, Justices Breyer and Kagan are the only Justice that I think has deep expertise on the APA issues and Chevron issues.  And those Justices who weigh into those issues most frequently and noisily (like Justice Gorsuch) either do so with a political agenda or, if they don't have one, might be confused by the rhetoric of those who do have a political agenda.  Still, at least in theory, if the Supreme Court wanted to wade into these issues to bring light rather than confusion, Altera might be a good case.

6.  Added 6/8/19 2:30pm:  I just noticed that I did not say anything about the reconstituted panel dissenting opinion by Judge O'Malley (of the Federal Circuit sitting by designation).  The dissent's principal position is that Treasury failed the arbitrary or capricious State Farm test.  To the extent the dissent relied upon Chevron at all, it was based on the notion that Chevron Step Two incorporated the arbitrary or capricious standard which the dissent first held failed on procedural grounds.  The notion that Chevron is the same as State Farm arbitrary or capricious standard applicable to legislative regulations is just wrong, a misreading inspired by Judulang's misreading which in turn was inspired by Mayo's misreading which in turn was a misreading of Chevron.  My article goes into those misreadings in depth, so I won't repeat that discussion here.  Suffice it to say that the gravamen of the dissent's analysis related to the application of the State Farm arbitrary or capricious test rather than Chevron's test of the reasonableness of the agency interpretation.  Assuming the dissent's premises that procedural irregularity invalidated the regulation, Chevron could not apply and the Court then could interpret the statute de novo without any deference to the agency interpretation.  In its interpretation de novo, the dissent interpreted the ambiguous statute to exclude stock based compensation costs in QCSAs, focusing solely on the first sentence of § 482.  However, if the dissent's premise is incorrect (the premise being that procedural irregularity invalidates the regulations), then Chevron would have applied and the test under Step Two is not whether the judge thought another interpretation was better but whether the Treasury interpretation was reasonable even if less reasonable than the dissent's interpretation.  And, here is the kicker that the dissent failed to perceive, the dissent had already determined that the agency interpretation was reasonable within the scope of the ambiguity in the statute (otherwise the dissent could not have gotten past Step One which she clearly did).  In other words, an unreasonable interpretation while tested at Step Two under the standard Chevron Framework can be inherent in Step One because an unreasonable interpretation cannot make statutory text ambiguous.  (That is perhaps a reason some say that Chevron really can be conflated into a single step.)

Added 6/10/19 3:44pm:  To follow through on the end of the paragraph above, I will revise the article to develop that issue.  Here are some unformed thoughts on that revision.  Consider this example, the court finds that there are two possible interpretations of the statutory text, but one of the interpretations is unreasonable.  Does the unreasonable interpretation create sufficient ambiguity so that the court can go to Step Two and deny Chevron deference at Step Two because the interpretation is unreasonable?  Consider another example with, say 4 possible interpretations of the statute only one of which is unreasonable, and the agency then chooses the unreasonable interpretation but the other 3 interpretations are reasonable.  Of course, the agency did not choose a reasonable interpretation so Chevron deference is not in question (whether kicked out at Step One or Step Two) and the court makes and applies its own best or better interpretation.  But, the question is whether Chevron's Framework even applies at all.  Is there another Chevron Step Zero (or some such) where the court just asks at the threshold whether the agency interpretation is unreasonable and stops if it is.  But, the court would only know if it is unreasonable through some statutory interpretation of the type that Step One and even Step Two entail.  Seems like I am going in circles.  I am not fully committed to developing this via revision of the article.  Any thoughts?

6.  This is an interesting footnote (p. 24):
fn. 6 We afforded the parties the opportunity to file optional supplemental briefs on the question whether the six-year statute of limitations under 28 U.S.C. § 2401(a)—which generally applies to procedural challenges to regulations under the APA—applies to this case. The Commissioner responded that it had waived this non-jurisdictional defense by failing to assert it to the Tax Court. We agree with the parties that the Commissioner waived the defense. Day v. McDonough, 547 U.S. 198, 210 n.11 (2006) (“[S]hould a State intelligently choose to waive a statute of limitations defense, a district court would not be at liberty to disregard that choice.”); Whidbee v. Pierce County, 857 F.3d 1019, 1024 (9th Cir. 2017) (“[E]ven if a claim has expired under a state statute of limitations, a defendant can still waive this affirmative defense.”). Therefore, we need not address it.
I may write something on this later, although that is a pure APA issue.

7.  One other nuance on the APA.  I refer to the State Farm standard as the "arbitrary or capricious" standard.  The State Farm standard is from the APA (5 U.S.C. § 706(2)(A)).  I explain in my article draft:
Legislative regulations are the law unless “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” fn1 (I refer to this standard as the “arbitrary or capricious” standard, but will use the longer form where quoting or context makes it important; I do caution readers that the same standard is stated by some, perhaps most, as the “arbitrary and capricious” standard. fn2)
   fn1 5 U.S.C. § 706(2)(A); see also Motor Vehicle Mfrs. Ass'n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29 (1983) (“State Farm”); Chevron, U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 844 (1984); and Batterton v. Francis, 432 U.S. 416, 425-426 (1977).
   fn 2 E.g., Perez v. Mortgage Bankers Assn., 575 U. S. ___, ___-___, 135 S. Ct. 1199, 1209 (2015) (“The APA contains a variety of constraints on agency decisionmaking—the arbitrary and capricious standard being among the most notable.”); but see Fed. Commc'ns Comm'n v. Fox Television Stations, Inc., 556 U.S. 502, 516 (2009) (“the Administrative Procedure Act's ‘arbitrary [or] capricious’ standard” (bracket “or” supplied by Court); and Bowen v. Georgetown Univ. Hosp., 488 U.S. 204, 220 (1988) (Scalia, J., concurring) ("'arbitrary' or 'capricious,' see 5 U.S.C. § 706").  Based on some quick searches on the Casetext platform, I extrapolate that courts use the conjunctive version more often than the disjunctive version.  I choose the disjunctive version because the statutory text is disjunctive. I am fairly confident that courts and commentators using either the conjunctive or disjunctive mean the same standard.
8.  Added 6/9/19 1:15 pm:  I am not sure what to make of the Tax Court's opinion in Altera.  As I noted, it was a reviewed opinion with 15 judges joining the opinion and no judge dissenting.  While I have not done any empirical study, I would think that is extremely rare that an appellate court would reverse a unanimous tax court reviewed opinion.  But then, Altera was not really a tax case.  It was an administrative law case where tax was the setting for the administrative law issues.  To say the least, the Tax Court judges are not known as administrative law experts.  (By contrast, although not a specialized court, the Supreme Court in recent years has had three recognized administrative law experts--Justices Scalia, Breyer and Kagan, each of whom taught administrative law; I don't really consider wannabee administrative law experts such as Justices Gorsuch and Kavanaugh.)  In this regard, I am reminded of Justice Scalia's famous sound bite:  “administrative law is not for sissies.”  Antonin Scalia, Judicial Deference to Administrative Interpretations of Law, 1989 Duke L.J. 511, 511. Others describe the difficulty of administrative law in less graphic but equally strong terms.  E.g., Sidney A. Shapiro, Eight Things Americans Can't Figure out about Controlling Administrative Power, 61 Admin. L. Rev. 5, 5-6 (2009) (“Administrative law is hard. We make this assertion with perfect confidence as it relates to American administrative law, which, in American law schools, is regarded as a notoriously difficult subject to teach and learn.”)  I am going to turn now to a comment on administrative law, and caveat that I too am not an administrative law expert, but I have spent an inordinate amount of time of a key subset of administrative law--the APA distinction between legislative and interpretive regulations and its trajectory in the courts and scholarly writing before and after the enactment of the APA in 1948.

Focusing on the Tax Court opinion, Professor Dan Shaviro states succinctly:  "the Tax Court unfortunately lost its way."  Dan Shaviro, Rationality wins, for a change (Start Making Sense Blog 6/8/19), here.  (Professor Shaviro subtitles his blog:  "Unfair but balanced commentary on tax and budget policy, contemporary U.S. politics and culture, and whatever else happens to come up.") Professor Shaviro was addressing the substantive tax issue, but I think the Tax Court also lost its way on the procedural issues.  For example, the Tax Court opens with what I think is clear error:  characterizing the regulation as a legislative regulation.  Most readers will know in a general senses what a legislative regulation is and looks like.  A legislative regulation is a congressional grant to an agency of authority to set the law by regulation.  The consolidated return regulations are the classic tax example of a legislative regulation.  Prior to 1928, the Revenue Act included consolidated returns and transfer pricing authority in the same provision without a grant of authority to set the law, but in the Revenue Act of 1928 split the consolidated return provisions (now § 1502) from the transfer pricing provision (now § 482), specifically giving Treasury statutory authority the law in the consolidated return section but excluding that authority from the transfer pricing section.  Congress knew how to grant legislative regulation authority and specifically did not grant it in the predecessor  to § 482 which has not changed in any way that is material to the legislative / interpretive issue.  There is a whole lot more on the legislative / interpretive regulation distinction under the APA.  But, like the Tax Court, many scholars think that essentially because of judicial developments without any congressional change to the APA (where the legislative / interpretive distinction is made), regulations that do no more than interpret the statute are now legislative regulations.  That is the principal issue that I address in the article that I hope to post this week on SSRN, titled "The Report of the Death of the Interpretive Regulation Is an Exaggeration."  The point is that, while the Tax Court might be in the mainstream of scholarly consensus in its statement that the § 482 regulation in issue was a legislative regulation, as I develop in the article the courts have not bought into that notion except that they have used certain loose language, dicta and worse (erroneous dicta) that might lead some to think they have.  That is why I believe the Altera reconstituted panel majority and even the dissent were right to avoid this issue that the Tax Court so readily leapt into.

As always, I appreciate comments.  Please leave a comment to this blog or email me at

For subsequent related items see:

  • Supreme Court Again Weighs In At the Edges on Legislative and Interpretive Rules (Federal Tax Procedure 6/23/19), here.

No comments:

Post a Comment