Last month’s Starr v US involves a $38 million 2007 refund suit in the DC district court brought by Starr International (Starr), the then-largest shareholder in American Insurance Group (AIG). In most refund suits there is little question that a district court has the power to conduct a de novo review of the IRS’s decision to deny a refund claim and determine whether a taxpayer is entitled to a refund. Why did the IRS view the case differently?
The issue involves the IRS’s failure under the US-Swiss income tax treaty to grant Starr a discretionary reduction in withholding on US-sourced AIG dividends. IRS argued that its decision to not grant Starr the discretionary treaty-based withholding relief was for it alone to make, and that its decision was not subject to judicial review. The district court rejected the IRS position, and in so doing discussed an aspect of the Administrative Procedure Act that the IRS claimed applied (at least in principle) to the proceedings.
In this post I will summarize the dispute and offer my view as to why as a policy matter IRS should be reluctant to argue for absolute discretion over most substantive decisions it makes, especially one that goes to the merits of a tax liability such as in this case.
What is the APA provision at issue? The APA states that agency action is generally reviewable “except to the extent that . . . [it] is committed to agency discretion by law.” 5 U.S.C. § 701(a) (note that another statute such as the Anti-Injunction Act may also remove an IRS decision from court review; that was not at issue in this case). Starr did not bring an action under the APA, however, as it brought a refund suit under Section 7422 and 28 U.S.C. § 1346(a)(1). As such, Starr argued that the “committed to agency discretion” rule did not apply because it did not bring an APA claim.
The district court rejected Starr’s position and essentially provided that even if the claim itself were not brought under the APA (and by implication even if the APA did not apply to agency determinations of this type), the principle of nonreviewability in 5 USC § 701(a) applies to all agency actions because the APA merely codified traditional principles of nonreviewability. In other words, if the IRS could establish that the decision to grant treaty benefits at issue in the case were committed solely to the IRS’s discretion, then its decision would be unreviewable despite the fact that there was no APA cause of action.
How does an agency establish that something is solely in its discretion and this not subject to court review? The legal inquiry asks generally if there are judicially manageable standards for a court to review. While that seems a somewhat vague principle, it is safe to say that courts are reluctant to let agency actions escape judicial scrutiny.
That limitation on nonreviewability is for good reason, as many administrative law scholars have claimed and common sense dictates. In an article I wrote awhile back on CDP (around page 1167) I have discussed both the legal and policy reasons why unreviewability of agency action should be the rare exception. In that piece I quote extensively from a 1990 article by Ronald Levin in the Minnesota Law Review called Understanding Unreviewability in Administrative Law. Professor Levin summed up the policy reasons nicely:
Scrutiny of administrative action by an independent judiciary is an integral part of the American checks and balances system—a powerful deterrent to abuses of power and an effective remedy when abuses occur. By helping maintain public confidence that government officials remain subject to the rule of law, judicial review also bolsters the legitimacy of agency action. . . . Finally, judicial review can enhance the quality of administrative action by exposing partiality, carelessness, and perverseness in agencies’ reasoning.
Despite the policy reasons against doing so, over the years, in addition to being able to put an invisibility cloak around its actions due to the Anti-Injunction Act, IRS has claimed unreviewability based on the “committed to agency discretion” defense over a wide range of agency decisions. For example, prior to legislation, it argued with success that decisions to not abate interest and to not grant equitable relief from joint and several liability were exempt from court review, and it continues to assert unreviewability when taxpayers seek judicial review of IRS denials of collection alternatives outside of CDP.
The Discretionary Relief in the Swiss Treaty
As the opinion describes, Starr relocated its corporate headquarters to Switzerland from Ireland. It did so because of a highly-publicized split between former AIG CEO Ace Greenberg and AIG, and disputes over Starr funding AIG’s compensation plans. It alleged that its move to Switzerland was done to “protect its assets from an AIG lawsuit claiming that Starr was contractually obligated to fund the plan.”
That was significant because the US –Ireland treaty automatically provided that the rate on withholdings on dividends was 15% rather than the normal 30%. The US-Swiss treaty on the other hand provided for a reduced 15% rate if certain conditions were satisfied.
Because it failed to satisfy the conditions for the non-discretionary Swiss treaty benefit Starr did not qualify for the automatic reduction under the US-Swiss treaty. In the absence of qualifying for the automatic reduction, the US –Swiss treaty allowed for the discretionary relief that is at the heart of this dispute. The relief provision provides that a taxpayer “may, nevertheless, be granted the benefits of the Convention if the competent authority of the State in which the income arises so determines after consultation with the competent authority of the other Contracting State.” US Swiss Treaty, Art. XXII(6).
As the opinion explains, Treasury has published a Technical Explanation of the US-Swiss treaty; the Technical Explanation is as the opinion mentions roughly analogous to legislative history. That explanation provides that the reason for the limitation on benefits (that is, the reason why not every Swiss taxpayer was entitled to the reduced withholding) was to prevent “treaty shopping.” As the opinion sets out, treaty shopping is the practice of ”moving … specifically to benefit from the lower U.S. tax rate offered by the U.S.-Swiss tax treaty.”
In 2007, Starr sent a request for discretionary reduced withholding as per the treaty to the US Competent Authority. Having not received a response in early 2010 it filed a protective refund claim. Starr forwarded the claim to the IRS analyst working the competent authority request. Shortly thereafter the competent authority denied the 2007 request for discretionary withholding relief, though interestingly it granted relief for the 2008 year.
Starr then sued for a refund in the fall of 2014. The complaint Starr brought claimed that the IRS abused its discretion for three reasons:
- Starr was not treaty shopping when it relocated to Switzerland,
- the IRS failed to consult with the Swiss Competent Authority before denying Starr’s request, and
- the IRS had no legal basis for issuing Starr a 2008 refund while denying its 2007 request based on the same material facts.
IRS filed a motion to dismiss on the grounds that its decision was unreviewable because it was committed to agency discretion. (alternatively, it also argued that the challenge raised an unreviewable political question, an issue that I will not discuss in this post but which the court found also did not apply to restrict court review).
How Did the Court Determine Whether the Matter Was Solely in the IRS Discretion?
As I discuss above, the court first disposed of Starr’s argument that only APA claims themselves are subject to the committed to agency discretion exception to judicial review. That IRS victory however was only short-lived though because the rest of the opinion discusses why in this case there was enough law for the court to apply, leading the court to the conclusion that it had the power to review the IRS’s decision to not grant Starr discretionary treaty relief. How does it get there? I suggest interested readers look at the opinion as there is lots there but I will hit a few of the highlights.
The court’s opinion goes through the black letter I also briefly describe above that generally asks if the applicable provision of law “is drawn so that a court would have no meaningful standard against which to judge the agency’s exercise of discretion,” or “in those rare instances where ‘statutes are drawn in such broad terms that in a given case there is no law to apply.’” (citations omitted).
In a footnote (note 7), the court notes the “strong presumption” that “Congress intends judicial review of administrative action” but also acknowledges that this case differs from many others given it was based not in a statute but a treaty provision. Despite the difference, it extends that presumption of reviewability to treaty provisions:
Federal courts “should normally apply the [background] presumptions supplied by American law” when ascertaining treatymakers’ intent in assessing entitlement to relief under a federal statute. One such presumption—an especially strong one—is that Congress intends for agency action to be reviewable. (citations omitted).
To look at the issue in question the court considered the prototypical agency matters that do get exempted from review, agency decisions whether to enforce or prosecute, and contrasted them with the IRS decision in this case. Despite accepting that distinction, IRS argued nonetheless that the treaty benefits’ decision warranted unreviewability because it implicated “complicated foreign policy matters” and courts have on occasion extended unreviewability to agency decisions sweeping in foreign policy.
The opinion distinguished a tax treaty’s benefits’ provision from those cases where courts found that the complex considerations in foreign policy matters rendered unreviewable, such as the State Department’s decisions to deny a consular visa or delegate a decision to control arms exports, noting that “reviewing a denial of benefits under the discretionary provision does not involve “second-guessing executive branch decision[s] involving complicated foreign policy matters.”
After disposing of the IRS argument that treaty benefit decisions were inherently nonreviewable the opinion undertook the task of discerning whether the “language, structure and history” of the benefits provision in the treaty supported nonreviewability. I hit some of the highlights of the opinion below.
The opinion started with a focus on the treaty language itself I quoted above, which provides permissively that the taxpayer “may” be granted relief. At first blush the use of the permissive “may” cuts in favor of absolute agency discretion, but the opinion notes that many cases have held that permissive language such as may was not determinative, leading the court to look to sources other than the text itself.
Its next stop was the treaty’s Technical Explanation, which as I mention above is like legislative history. Prepared by Treasury to assist the Senate in the ratification process, it reflects Treasury’s view and explanation of the treaty. That explanation includes a statement on the discretionary provision, stating that the Treasury would “base [its] determination . . . on whether the establishment, acquisition, or maintenance of the person seeking benefits under the Convention, or the conduct of such person’s operations, has or had as one of its principal purposes the obtaining of benefits under the convention.”
The opinion looked to testimony of Treasury officials, which like the explanation puts some context on the decision to grant discretionary benefits. For example, the Treasury Assistant Secretary stated that in determining eligibility for discretionary relief IRS would look for a “a substantial non-treaty-shopping motive for establishing themselves in their country of residence.”
While not part of the treaty itself, the explanation and testimony suggested a standard by which the IRS would evaluate requests for treaty relief. In fact, in denying Starr’s claim, the IRS applied that standard, informing Starr that it could not “conclude that obtaining treaty benefits was not at least one of the principal purposes for moving Starr’s management, and therefore its residency, to Switzerland.”
What was the IRS’s rebuttal? It pointed to Senate report accompanying the treaty ratification which stated that “the Secretary of the Treasury may, in his sole discretion, treat a foreign corporation as a qualified resident of a foreign country[.]” Tax Convention with Switzerland, S. Exec. Rep. No. 105-10, at 54 (1997). The opinion notes that this “does not explicitly discuss judicial review, but it nonetheless provides some evidence of an intent to preclude such review.” Despite what the opinion felt were the mixed messages, it felt that the report was not enough for the IRS to prevail:
IRS has not presented clear and convincing evidence that the discretionary provision was intended to preclude judicial review. Indeed, the structured guidance set forth in the Technical Explanation—a long with the lack of any express preclusion of judicial review—renders the issue sufficiently ambiguous that the presumption of reviewability controls?
IRS also argued that the “principal purpose” standard was too vague to afford a court meaningful review:
The IRS claims that this standard is not specific enough to permit review. What, asks the IRS, does it mean to conclude that a company’s “principal purpose” is to obtain the benefits of a tax treaty?
The opinion likewise finds this argument unavailing and notes that unlike IRS it was “not so daunted by the prospect of reviewing the IRS’s determinations. Courts routinely face somewhat amorphous and open-ended standards…. At the margins, it may be difficult to determine whether a company moved to Switzerland principally to lower its tax rate. But this does not mean there are no manageable standards to apply.”
Having survived a motion to dismiss, the case now will proceed to the merits. While the presence of the treaty provisions complicates the analysis, I think the district court got this one right on the law and on the policy. Courts are pushing back against IRS claims that its process or decisions should be insulated from judicial review. For the reasons Professor Levin succinctly mentions in his 1990 law review article, courts are rightfully skeptical of agency blanket calls for unreviewability. IRS has faced plenty of allegations of abuse of power, and it is rocked still by questions pertaining to its politicized review of applications for exemption. For an agency like the IRS, which administers a law which depends in part on continued public respect for the integrity of the tax system, IRS needs to choose its battles on this issue carefully.