Interpreting Tax Treaties

I.     Introduction

The international law of treaty interpretation, as codified in the Vienna Convention on the Law of Treaties (“Vienna Convention”), prescribes general rules of interpretation, based on a plain meaning approach, that apply uniformly to all treaties.1 Yet courts, states, and scholars seem to agree that some treaties warrant special interpretive rules.2 Theorizing this differentiated approach to treaty interpretation, however, remains elusive. Explanations based on treaty subject matter or purpose fail to satisfy.3 Instead, examination of the objective features shared within a treaty category provides a more promising avenue for justifying specialized interpretive methods. One such characteristic is the treaty’s degree of completeness, or its degree of specificity and operationality. Specifically, all else being equal, standalone instruments call for less reliance upon extrinsic materials; interstitial instruments demand more.

Applying this insight to the income tax treaty context,4 such instruments should not be viewed as complete; accordingly, reference to plain meaning or even the treaty makers’ mutual intent is often incoherent. Specifically, because tax treaties function to limit the taxing reach of treaty countries rather than prescribe substantive rules, they are closely intertwined with domestic law. Indeed, tax treaties explicitly state that domestic law provides the meaning of any undefined term. Gaps are intentionally left open by treaty drafters due to the complexity of the tax system and the close connection between fiscal policy and sovereignty.5

Tax treaties are also not heavily negotiated. They strongly resemble one another and refer to concepts that are embedded in the model treaties, domestic laws, as well as other international instruments. Moreover, the necessity of combatting tax abuse retrospectively reduces the utility of the treaty’s text and also mandates continuous involvement by the executive branch in interpreting the treaty. Accordingly, discerning the intent of the parties or plain meaning should not be the only goal of tax treaty interpretation. In cases of ambiguity, domestic law and other extrinsic sources thus fill the gaps in the language of the tax treaty. It would therefore be a mistake to dismiss these materials when, in fact, the nature of tax treaties makes reference to them inevitable.

It could be argued that application of sources outside of the treaty process may, at times, produce double taxation, against which tax treaties are purported to guard. This danger, however, is low. Typically, only one country is interpreting the treaty on any given issue. Treaties themselves also provide further protection against double taxation through mutual agreement procedures and binding arbitration clauses, and the domestic laws of countries have mechanisms, like foreign tax credits, that protect against double taxation. It is also unclear whether a plain meaning approach would produce less variation in interpretations. Additionally, a loose ordering of interpretive sources, as well as certain interpretive presumptions that I propose, could ameliorate this danger.

Looking to extrinsic sources best reflects the realities of the international taxing system. A non-formalist methodology also guards against tax abuse and assists in modernizing international agreements to reflect changing technologies and economies. Indeed, recent projects undertaken by both the Organisation for Economic Co-operation and Development (“OECD”) and the United States to update the existing treaty network reinforce the need for flexibility in the interpretation of tax treaties.6

There remains “limited scholarly attention” upon the interpretation of treaties in the United States,7 but the American literature specific to tax treaties is particularly sparse.8 A central goal of this Article, then, is to theorize an approach to the interpretation of tax treaties.9 United States courts, for the most part, embrace an eclectic mix of national, international, bilateral, and unilateral sources, and the unique characteristics of tax treaties and the processes by which they are enacted support the courts’ pragmatic approach to interpretation.10 In many respects, this Article is a defense against accusations that such reliance upon extrinsic materials is misplaced in the tax treaty context.11 This defense is particularly timely in light of: (1) recent international consensus that treaties should not be used to achieve tax abuse,12 making a plain meaning approach problematic; (2) Treasury’s renewed interest in issuing broadly applicable rules for tax treaties13 and its imminent release of a new model tax treaty for the first time in a decade;14 and (3) the Supreme Court’s recent willingness to interpret other treaties in a more textualist manner,15 a development that should not be extended to the income tax treaty context.16

Part II provides an overview of tax treaties and the tax treaty process. Part III discusses the domestic and international backgrounds of interpretive principles that guide treaty interpretation, both generally and with respect to tax treaties. Part IV presents my view of a differentiated interpretation of treaties using the instrument’s degree of completeness as a criterion. It then argues that tax treaties, as incomplete instruments, need to be interpreted using a variety of extrinsic sources in a pragmatic fashion. Part V discusses implementation issues, such as double taxation, and sets forth harmonization devices, such as a loose ordering of sources and interpretive presumptions that can ameliorate these concerns.

II.     Introduction to Tax Treaties

A.     The Role of Tax Treaties

1.     Avoidance of Double-Taxation

The stated purpose of tax treaties is to ameliorate or eliminate double taxation.17 The phenomenon of double taxation primarily occurs when both a person’s country of residence (the “residence country”) as well as the country from which the income is generated (the “source country”) claim jurisdiction to tax an item of income. Tax treaties resolve such conflicts in two ways. First, they lower or eliminate the source country’s withholding tax on certain categories of income. They also require that the residence country either exempts foreign source income or offers a foreign tax credit for taxes paid in the source country. In the latter case, residence countries are able to tax the foreign source income on a residual basis after allowing for a credit for the tax paid in the source state.18 Second, tax treaties establish competent authority procedures such that the tax authorities endeavor to resolve taxpayer complaints of double taxation.19

Although tax treaties require residence countries to alleviate double taxation, countries typically offer such relief unilaterally under their domestic law. Domestic law, however, may not completely eliminate double taxation. For instance, domestic rules allocating income among related entities might result in the denial of a foreign tax credit. The tax treaty regime purportedly attempts to synthesize the taxing jurisdictions of the treaty partners to fill in these types of gaps created by their domestic laws; however, its success in doing so is questionable given the generality of treaty language.20 The treaties themselves, as well as treaty commentaries, refer to the elimination of double taxation as their primary goal, yet some commentators have emphasized that the reduction of withholding taxes, as well as anti-avoidance concerns, have become the primary focus of modern tax treaties.21

2.     Prevention of Fiscal Evasion

Another stated goal of tax treaties is to limit fiscal evasion. One way treaties achieve this is through information sharing provisions, which require one country to disclose information regarding persons who have tax obligations in the other country.22 These provisions override domestic confidentiality laws that typically prevent governments from releasing tax information. They thus provide the country that is trying to tax foreign source income an easier means to identify such income. Treaties might also facilitate the collection of unpaid taxes of nonresidents by enlisting the collection authorities of the other country.

B.     The Tax Treaty Process

The model treaties are the starting point for the tax treaty process. In 1928, the League of Nations released the world’s first draft model double-income tax treaty. Today, more than 3000 such treaties exist in the world, nearly 70 of which include the United States as a partner.23 The general form of these treaties is heavily influenced by the League of Nations model tax treaty, which is the predecessor of the OECD Model Tax Convention on Income and on Capital.24 The United States Model Income Tax Convention, in turn, is derived from the OECD Model.25

These model treaties provide the basic structure of the United States’ tax treaties, but there is some variation among the particular provisions in individual treaties. Representatives from the United States, generally officials from the Treasury Department’s Office of International Tax Counsel and the Office of Tax Analysis (Business and International Taxation), along with staff from the Internal Revenue Service and the State Department, negotiate with officials from the other treaty country on the terms of the treaty. The negotiating history represents the official oral and written communications between the treaty partners.26

Once finalized, the State Department approves the treaty text, and the treaty is signed by officials of the United States and the other country before it is given to the President along with a letter describing the general terms of the treaty. The President then forwards the treaty to the Senate for its “Advice and Consent” as mandated in Article II, Section 2 of the Constitution.27 In the Senate, the Foreign Relations Committee reviews the treaty, as well as technical explanations provided by the Joint Committee on Taxation (“JCT”) and the Treasury Department. Staff from both the JCT and Treasury may testify during hearings. After this review, the Senate Foreign Relations Committee votes to recommend the treaty to the Senate. A Senate Report from the Committee often accompanies this recommendation.28

After the Committee sends the treaty to the Senate floor, the full Senate debates the recommendation of the Committee. The Senate’s consent to ratification is obtained upon a vote of two-thirds of the Senators present. The President then signs the treaty, which enters into force once each country notifies the other that ratification has taken place.29

The Senate’s constitutional role in the treaty-making process can prove frustrating to our treaty partners. Not only does its involvement slow down the treaty process, but sometimes the Foreign Relations Committee or an individual Senator inserts a “reservation,” thus predicating consent on the removal or modification of a treaty provision. In such cases, the treaty partners must renegotiate the treaty. In the tax treaty context, Senate reservations are not a matter of course, but they do arise somewhat regularly.30 In recent years, the Senate has altogether failed to provide its consent to numerous tax treaties, adding further uncertainty to the process.31

C.     The Interaction of Treaties and Domestic Law

The Supreme Court has interpreted the Supremacy Clause to mean that treaties and statutes have equal legal status. If a conflict results between the two, then the instrument enacted last trumps under the last in time rule.32 In the tax context, this canon has essentially been codified in section 894(a) of the Internal Revenue Code (“Code”).33 A companion to the last in time rule is the interpretive rule, called the “Charming Betsy canon,” which states that statutes should be construed as to avoid conflicts with international law.34

III.   Background of Interpretive Principles

A.    Treaty Interpretation in the United States

Scholars have long debated whether to interpret a treaty like a contract—that is, interpreting treaties to give effect to the drafters’ intent—or like a statute, where interpretive methodologies such as textualism or purposivism may come into play. The Court has purported to do the former,35 although some have argued that “while the prevailing rhetoric of [treaty] interpretation is contractual, the underlying idiom and approach is statutory.”36 Others contend that treaties are sui generis, and that neither contractual nor statutory interpretive principles guide the Court; instead, it decides treaty cases largely out of deference to the executive branch, which thus favors a flexible interpretive method.37 Nonetheless, the contractual approach still remains influential among U.S. courts as a whole. As I will discuss, however, neither analogy is particularly useful or descriptive in guiding the interpretation of tax treaties.

Contract interpretation leads courts beyond the four corners of the contract to effectuate the parties’ subjective intent. The Restatement (Third) of Foreign Relations Law embraces the relevancy of the contractual methodology to treaty interpretation in concluding that in the “United States tradition the primary object of interpretation is to ‘ascertain the meaning intended by the parties’” rather than focus simply on the text.38 The Supreme Court has also embraced intentionalism, at times emphasizing that treaty interpretation lends itself to even more liberal use of external sources than contract interpretation in order to effectuate such intent.39 Later courts have regularly turned to the negotiating history of the treaty (travaux préparatoires) and the parties’ subsequent practice as evidence of intent.40 American courts have scrutinized correspondence between nations, domestic law, parallel treaties, and subsequent treaty modifications to discern subsequent practice of the treaty parties.41

Advocates for interpreting treaties as contracts sometimes emphasize the consistency such an approach provides. Sital Kalantry argues that the contractual approach would more likely lend itself to uniform interpretations by U.S., foreign, and international courts.42 According to Kalantry, this is because these courts will review “the same body of extrinsic information” and will also look to each other’s interpretations as persuasive authority in adjudicating their own interpretation.43

Kalantry also contends that treaties are structurally more similar to contracts than to statutes. She argues that although “treaties have the [legal] effect of statutes,” their characteristics “should guide what interpretive principles courts apply.”44 Thus, because both contracts and treaties have parties that sign, negotiate, draft, approve, amend and are bound by the instruments, treaties are more akin to contracts.45 Statutes, on the other hand, are approved by legislators and drafted by them and their staff and are intended to govern people within the relevant jurisdiction rather than the parties involved in the legislative process.46 Another argument advanced in favor of a contractual approach is that, although the Senate is involved in treaty approval, the executive represents the nation’s interests in negotiating the treaty.47 Finally, unlike the statutory context where Article 1, Section 7 requires only majoritarian agreement to the text, which in turn represents compromises among competing interest groups and legislators, treaties represent a true “meeting of the minds” with regard to the treaty’s meaning.48

Although a contractual approach to treaty interpretation has dominated the discourse of federal courts, some argue that the Senate’s special role in the treaty-making process transforms treaties beyond a simple contract between the executives of nations.49 Express Senate reservations almost certainly bind courts interpreting the treaty,50 and formal Senate interpretations also seem to be heavily influential.51 What role the Senate’s informal views on the interpretation of the treaty should have in the judicial interpretative process raises a more difficult question. Such views may not be formally declared, but they may be unearthed in the treaty’s “legislative history,” which records the back-and-forth between the Senate and the Executive during the treaty-making process.52 Emphasis on the Senate’s role in treaty-making would allow a greater role for informal materials, such as reports by the Senate Foreign Relations Committee, statements of Senate leaders, as well as executive branch sources that influenced the Senate. While many courts have relied upon such materials, few have explicitly discussed their proper influence.53

Courts also give deference to bilateral and unilateral executive branch materials.54 This position is rare in the international community but is rationalized by the special constitutional role of the Executive in treaty-making and traditional administrative law principles of congressional delegation and institutional expertise.55 Some U.S. courts have even deferred to the executive branch over the objections of the treaty partner.56 Courts have, however, refused to confer such deference in the face of prior conflicting executive interpretations, constitutional concerns, or poor reasoning on the part of the executive branch.57

Despite the Court’s emphasis upon the parties’ intent in interpreting treaties, Justice Scalia made efforts to abandon such an approach in favor of a textualist reading of treaties, thus interpreting treaties more like statutes. For instance, in United States v. Stuart, the majority drew support from the Senate pre-ratification debate, negotiating materials, and past practices of the treaty parties in interpreting the treaty.58 In his concurring opinion, Justice Scalia vigorously argued that separation of powers and rule of law concerns dictate that the Court restrict its inquiry in treaty interpretation cases to the four corners of the agreement.59 He further argued that the Constitution provides the Senate with only formal means to act upon treaties and to give credence to informal materials lies in tension with constitutional structure.60

Later that year, in Chan v. Korean Air Lines, the majority opinion, written by Justice Scalia, took a textualist approach to treaty interpretation.61 In his concurring opinion, Justice Brennan asserted that the majority was wrong in ignoring extrinsic evidence in ascertaining the government’s intent.62 A few years later, however, the Court reaffirmed a non-textualist approach to treaty interpretation in Sale v. Haitian Centers Council.63 There, the majority implicitly invoked the canon of absurd results to break away from the treaty’s text,64 instead relying upon domestic legislation that used parallel language to that in the treaty,65 as well as negotiating and legislative history.66 Justice Scalia joined in the majority opinion, while Justice Blackmun, in dissent, decried the majority’s rejection of plain language and reliance on extrinsic evidence.67 The Court again applied a more flexible interpretive approach in Olympic Airways, where the majority looked to not only dictionary definitions but also the views of other countries in interpreting the Warsaw Convention.68 Even in his dissent, Justice Scalia cited to decisions of foreign courts.69

More recently, however, the Court returned to a more formalist approach to treaty interpretation in Medellín v. Texas, which famously held that decisions of the International Court of Justice are not binding under domestic law since the operative treaty was not self-executing.70 In the majority opinion, Chief Justice Roberts wrote that the text of the treaty anticipated future legislative action because it required only that the parties “undertake[] to comply with the decision of the [International Court of Justice].”71 In so doing, the Court largely focused on treaty text as opposed to other possible sources of meaning in ascertaining whether the treaty was self-executing.72 As a result, many point to Roberts’ majority opinion as further evidence of the Court’s formalist leanings in interpreting treaties.73 Even in an opinion criticized for its wooden textualism, however, the majority cited to Senate ratification debates74 while also condoning the use of negotiating history and the treaty partners’ post-ratification understandings (without explicitly relying upon them).75Medellín could thus be understood as placing textualism first but also as sanctioning “[a] new eclecticism in the selection of extrinsic sources . . . .”76

B.     The Vienna Convention

The Vienna Convention sets forth rules of interpretation governing international agreements and guides the judicial approach to treaty interpretation in many jurisdictions.77 The United States is a signatory to the Convention, but the Senate has never consented to its ratification.78 Nonetheless, many lower federal and state courts routinely turn to the interpretive principles set forth in the Vienna Convention79 and the State Department accepts it as an “authoritative guide” to customary principles of international law.80 Still, the Convention’s authority in the United States is questionable. The Supreme Court has only cited to it twice and in an incidental fashion.81 Indeed, the plain meaning approach of the Vienna Convention lies in tension with the Court’s predominantly intentionalist approach.

Article 31 of the Vienna Convention provides that a treaty be interpreted “in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.”82 “Purpose” is not the subjective intent of the parties but rather constitutes the goal of the treaty.83 “Context” can be unearthed using any corollary agreement made by the treaty partners in connection with the treaty. Article 31 further provides that the interpreter take into account subsequent agreements by the treaty parties, subsequent practices that reveal the parties’ agreement as to interpretive issues, and any relevant rules of international law. The Convention prescribes deviating from the plain meaning only if a special meaning reflects mutual intent.84 Article 32 provides additional rules of interpretation, which allow for the use of supplementary materials such as the negotiating history of the treaty (1) in order to confirm the meaning resulting from the application of article 31; or (2) in order to construe the text if the meaning is “ambiguous or obscure” or leads to “absurd” or “unreasonable” results.85

The drafters of the Convention intended to adopt a plain meaning approach because they thought this methodology reflected customary practice86 and would give nations less flexibility to pursue unilateralist interpretations.87 Thus, the Vienna Convention steers courts away from a contractual approach, relying upon intent only in narrow circumstances.88 The drafters of the Convention also explicitly rejected an amendment by Professor Myres McDougal, a developer of the “New Haven School” of interpretation. The New Haven School emphasized the importance of fluidity in the interpretive approach in order to discern “the value and policy considerations” that give rise to the obligations of the parties.89 This interpretive method was also guided by the importance of upholding “basic norms of world legal order.”90 McDougal’s proposal would have “eliminate[d] the rigidities, restrictions and hierarchical distinctions” by prescribing “[n]o fixed hierarchy . . . among the elements of interpretation.”91 It also would have adopted a wide range of interpretive sources “including ordinary meaning, subsequent practice and preparatory work, but not excluding others that might be also relevant.”92

Instead, the Vienna Convention largely discredits extra-textual materials beyond those used in the preparation of the treaty.93 It does not explicitly accept unilateral interpretations of the treaty, for instance Treasury Technical Explanations or Senate Foreign Relations Committee reports, and it prescribes customary international law as a gap-filling mechanism.94 The Convention thus adopts a stance that formalists might embrace, although its acceptance of a limited array of extrinsic materials might still prove controversial to pure textualists.

The Vienna Convention’s predominant focus on plain meaning has been generally accepted abroad and has led to a potential disparity between the interpretative approach of the United States and that of other nations.95 Although lower U.S. courts sometimes look to the interpretive principles of the Vienna Convention in construing treaties, the effect of the Convention upon U.S. treaty interpretation has been only “to check slightly the use of extra-textual means of interpretation.”96

C.     U.S. Pragmatism in Tax Treaty Interpretation

Generally, U.S. courts enthusiastically embrace the use of extrinsic sources in the tax treaty context.97 The starting point of interpreting a U.S. income tax treaty is the treaty language, which controls unless the result is inconsistent with the intent of the treaty-makers or the purpose of the treaty.98 But in piecing together intent or purpose, U.S. courts often look outside the treaty text. They employ canons of interpretation.99 They resort to negotiating history materials to arrive at the treaty’s meaning.100 They also look to materials from the ratification process, such as Senate committee reports,101 interpretations from the executive branch,102 the subsequent practice of treaty partners,103 and commentaries on model treaties.104 When treaties are silent on an issue, they may look to the domestic law to fill gaps in meaning, even where not specifically directed to do so by the treaty,105 and may also employ common law anti-abuse doctrines to combat tax shelters.106 Resort to these varied, often unilateral, sources means that U.S. courts are unmoored from both the plain meaning and intentionalist approaches. Instead, they pursue a pragmatic approach to tax treaty interpretation, which is best exemplified by exploring a few cases.

In Great-West Life Assurance Co. v. United States, for instance, the Court considered whether a provision in the U.S.–Canada treaty, which prevented the United States from taxing interest paid between Canadian corporations, applied when a U.S. permanent establishment received the interest.107 The parties stipulated that the literal language of the treaty applied, but the Court allowed the United States to tax the interest after relying upon the treaty’s negotiating history, subsequent statutory changes, interpretations by the Treasury and State Departments, and treaty purpose.108

United States v. Stuart is a case that, as discussed above, is well known for its discussion of Senate materials. It involved an exchange of information provision under the 1942 U.S.–Canada treaty, which obligated a treaty partner to obtain and send tax information to the other treaty partner that it was “in a position to obtain under its revenue laws.”109 After the treaty was signed, the U.S. domestic law had been changed to prohibit the IRS from issuing a summons order if the underlying investigation was criminal and had been referred to the DOJ.110 Pursuant to the treaty, Canadian officials asked the IRS to provide bank records that were pertinent to a Canadian taxpayer’s domestic liability.111 The taxpayer argued that the U.S. domestic law allows the IRS to exercise its summons power only if it determines that the Canadian tax investigation had not yet reached the stage that would be comparable to a referral to the DOJ for criminal prosecution.112

The Court rejected the taxpayer’s argument, reasoning that the ratification history, including Senate legislative history and the President’s transmittal message and proclamation, was silent as to any such restrictions.113 The Court also looked to the subsequent practice of the treaty partners to conclude that, since the treaties’ enactment, the IRS had regularly complied with the information exchange provision without inquiring whether the Canadian investigation was criminal.114 As mentioned above,115 Stuart is notable for its concurring opinion by Justice Scalia, who strongly objected to the majority’s use of unilateral materials in the interpretation process since such statements did not reflect the shared expectations of the parties.116

Other courts have ostensibly relied upon the plain meaning of the statute but generally look to extrinsic materials to support their analysis. Xerox Corp. v. United States involved the timing of when the U.K. Advanced Corporate Tax (“ACT”) could be considered a deemed paid tax under section 902 of the Code, which allows the U.S. corporation to credit foreign taxes paid by a foreign subsidiary upon receipt of a dividend from such subsidiary.117 On appeal, the Federal Circuit reversed the Claims Court’s decision, holding in favor of the taxpayer’s position that the credit could be claimed currently.118 In so doing, the Federal Circuit relied upon extrinsic evidence since the treaty was silent on the issue of timing.119 The court observed that U.K. officials contested receipt of the Technical Explanation120 and that the Treasury’s approach was rejected in the Senate Executive Report.121 The court also concluded that the revenue procedure was at odds with the treaty’s plain language and the treaty’s purpose against double taxation.122 The court reasoned that revenue procedures did not carry the same weight as a regulation and did not bind the court’s interpretation.123

Snap-On Tools v. United States also concerned timing issues involving the payment of the ACT under the 1945 United States–United Kingdom tax treaty.124 The court rejected the IRS’s denial of a carryback of the associated tax credit.125 Domestic law allowed such a carryback, but according to the IRS, the treaty overrode the domestic rule.126 In reaching these conclusions, the Claims Court relied upon several principles of treaty interpretation. After noting that treaties are to be construed liberally and in a manner akin to contracts, the court also observed that reliance upon extra-textual materials is more common in the case of treaties.127 Nonetheless, the court cited Scalia’s concurrence in Stuart to reject consultation of the Technical Explanation, which set forth the IRS’s argument, since it did not evince the mutual intent of the treaty parties.128 In justifying its interpretation, the court also noted that, although the later enacted instrument generally takes precedent as between statutes and treaties, the court has a duty to give effect to both if it is so possible (essentially applying the Charming Betsy canon).129 Still, the Court did rely upon a statement in the Senate Foreign Relations Committee Report that the treaty was not intended to repeal the foreign tax credit rules and concluded that the legislative history contained nothing to support the IRS’s interpretation.130 Furthermore, the court examined the subsequent conduct of the United States in noting that the domestic 60-day rule was repealed years later, indicating that Congress thought legislative action was needed to reconcile the treaty with domestic law.131

These two cases, Xerox and Snap-On Tools, reveal the difficulty in making blanket statements regarding U.S. treaty interpretation. Although both lower courts expressed hesitation to rely upon certain unilateral documents like the Technical Explanation, seemingly embracing the contractual approach, they examined treaty purpose and unilateral Senate legislative history.132 Other lower courts have relied on the treaty’s Technical Explanation,133 and the rejection of the Technical Explanations in Xerox and Snap-On Tools may be explained by their conflict with other unilateral extrinsic evidence.

The official Commentaries that accompany the OECD Model Tax Treaty provide another extrinsic source of tax treaty interpretations. These are most helpful when the treaty language at issue is taken from the OECD Model Treaty. U.S. courts have relied heavily upon the commentaries.134 Complex issues arise involving the commentaries since they are ambulatory in nature. Because the commentaries are often restated or revised, courts must decide whether to apply the interpretation from the commentaries in effect at the time of the treaty’s negotiation, at the time of the treaty’s application, or at the time of the court’s decision. U.S. courts have generally followed the ambulatory approach, applying the commentaries currently in effect.135

This discussion included only a sampling of the U.S. case law on tax treaty interpretation, but the heavy reliance on varied extrinsic sources illustrates that the American judiciary does not strictly adhere to the Vienna Convention. Additionally, reliance on unilateral or ambulatory sources suggests that the courts are not exclusively focused on mutual intent either. Instead, tax treaty interpretation in the United States approximates a pragmatic approach. The remainder of this Article defends this methodology on both positive and normative grounds.

IV.     The Necessity of Extrinsic Evidence in Tax Treaty Interpretation

A.     Differentiated Treaty Interpretation

The Vienna Convention’s interpretive regime governs all treaties and is considered to be customary international law. Yet, there is also consensus that certain treaties warrant differentiated interpretive rules.136 Such claims often resort to a bald assertion of “special subject matter.”137 Scholars and advocates make the case for individualized treaty interpretation in the context of human rights, the environment, and the protection of civilians in armed conflict, among others, resting upon “a mere assertion about the importance of the treaty’s subject matter.”138 Problematically, it is difficult to unearth normative values from the category of the treaty alone. Thus, subject matter becomes an unprincipled rationale for differentiated interpretation.139 This is not to say that human rights treaties do not deserve liberal interpretation. But deeper analysis of the norms driving such an interpretation is more helpful in guiding treaty-makers since such rights are likely not limitless. Additionally, such an analysis is helpful when treaties involve several subject matters, as is often the case.140

Each treaty’s specific object and purpose provides another justification for differentiated interpretation, and hence the intent of the treaty-makers.141 This approach would apply on a treaty-by-treaty basis and hence would not differentiate on a categorical basis. The problem with this reasoning is that the objectives and purposes of a treaty are difficult to discern and are often conflicting. For instance, the stated and unstated goals of tax treaties—the avoidance of double taxation, the general reduction of taxes, the protection against tax avoidance, and the prevention of fiscal evasion—may be in conflict with one another.142 Moreover, even if there is a readily ascertainable single purpose, one must also inquire into the parties’ intent as to how to achieve that purpose or what should be sacrificed in order to do so.

More promising avenues look at objective characteristics in validating interpretive differentiation. Thus, one may consider the reliance interests of third parties in considering whether to give weight to subsequent practices and agreements.143 One may also consider whether a treaty imparts reciprocal obligations between states, in which case an evolutionary approach may be justified, or instead attempts to set forth a more lasting commitment, which is insulated from changes in context or intent.144

Adding to these latter types of inquiries, I contend that a treaty’s degree of plenitude has implications for the appropriate interpretive methodology. Certain types of treaties are more comprehensive than others, in which case reference to extrinsic source is less justifiable. Other treaty categories are more interstitial and thus require reliance on extrinsic sources. Comprehensiveness can be defined along two dimensions—degree of specificity and operationality. As to specificity, some treaties contain open-ended or vague language, as a matter of necessity and/‌or to further the drafters’ intent.145 As to operationality, treaties may contain only provisions that require other legal sources to have effect.146 Below, I contend that tax treaties fall far towards the incomplete end of the spectrum on both dimensions. They lack specificity and are jurisdictional in nature, thus validating liberal use of extrinsic sources. Because of these characteristics, tax treaties themselves often direct that interpretive gaps be filled in by the domestic law.147

B.     Jurisdictional Tax Treaties and the Complexity of Tax Law

Tax treaties mostly lack operative provisions of law.148 Instead, they function to restrict a state’s claim to taxation. To be more precise, tax treaties limit the reach of the domestic tax laws with respect to international transactions either by allocating the right to tax to one treaty country or by requiring foreign tax credits.149 The country that receives the tax claim applies its domestic taxing rules in conjunction with the principles set forth in the treaty. Accordingly, a tax treaty does not create a tax obligation that does not otherwise exist in the domestic law. Thus, the tax rules found in treaties are more akin to “rules of limitation of law” as seen in “international administration law” as opposed to rules of conflict seen in private international law.150 Although they do have a separate legal underpinning, these “rules of limitation” are “embodied in, or closely related to” the underlying domestic law of the treaty states.151

Because of their separate functions, “Code provisions are tailored to specific situations, whereas treaties state general taxing principles.”152 Consequently, treaties often refer to or piggyback upon domestic concepts.153 Treaties also have a gap-filling provision, based on the OECD Model Treaty’s article 3(2), stating that terms that are not defined by the treaty have the meaning given to them under the law of the State that is applying the treaty.154 As will be discussed in further detail below, this is typically the source state but could also be the residence state.155 The use of domestic law is often necessitated given that most tax treaties define only a handful of terms.156

At times, specific treaty provisions dictate that domestic law applies when defining a term. For instance, “company” under the U.S. Model Treaty is defined as “any body corporate or any entity that is treated as a body corporate for tax purposes according to the laws of the state in which it is organized.”157 The determination of residence is a weighty threshold question that tax treaties have also, by and large, delegated to the domestic law. In determining whether a person is “residing” in the United States, the treaty first asks whether that is the case under domestic law. The treaty only steps in to provide tie-breaking rules when both countries treat the taxpayer as a resident.158

Neither do tax treaties define the term “business profits,” which presumably encompasses income derived from any “trade or business”—a phrase that is in turn defined by the domestic law.159 Even the definition of “business expenses,” for which the treaty mandates the source country to allow deductions,160 is undefined. Yet, it would be difficult to argue that the absence of a treaty definition means that the source country cannot impose limitations that otherwise apply in the domestic context.161 Thus, a business expense must likely be “ordinary and necessary” in the U.S. treaty scenario.162

Tax treaties also invoke domestic law concepts by leaving terms vague or by providing definitions that themselves employ undefined terms. The characterization of income is one such area, the outcome of which has a profound effect on tax liability. Before applying the treaty rules for capital gains, interest, or dividends, the treaty, having left these terms vague, requires consultation of domestic law.163 For instance, in the case of interest, the OECD Model Treaty defines the term as “income from debt-claims of every kind, whether or not secured by mortgage and whether or not carrying a right to participate in the debtor’s profits.”164 Of course, this begs the question of what constitutes a debt-claim, and this may require consultation of domestic law, even though technically the treaty itself has defined the term.

If both countries characterize the item differently, then the treaty merely provides for competent authority procedures to attempt to avoid double taxation. Boulez v. Commissioner, a well known case illustrating this conundrum, addressed whether the receipt of payments for U.S. performances by a music conductor residing in Germany constituted royalties, in which case they were exempt under the U.S.–Germany Treaty and taxable in Germany, or whether they were compensation for personal services, in which case the payments were subject to U.S. taxation.165 Both countries characterized the payments differently, and because each country asserted taxing jurisdiction based on source and residence, article 3(2) could not resolve the conflict.166 The Tax Court held that the income was personal service income under domestic rules and therefore taxable in the United States.167

To provide another example of the treaty relying upon domestic law, under U.S. domestic rules, a foreign person will be taxed on active income in the United States if they have U.S. source income that is effectively connected with a U.S. trade or business.168 Tax treaties narrow this jurisdiction to business profits that are “attributable to [a] permanent establishment.”169 Some argue that the “attributable to” treaty standard is essentially equivalent to the domestic “effectively connected” test.170 Even if the two concepts are independent, they do overlap. Article 5(1) of the 2006 Model Treaty defines a “permanent establishment” as “a fixed place of business through which the business of an enterprise is wholly or partly carried on.”171 U.S. treaties, however, do not define the term “fixed place of business” and hence the domestic law largely provides its meaning.172 Under section 864, certain foreign source income of a foreign corporation is taxed if earned through a domestic office or “fixed place of business.”173 Accordingly, the regulations under section 864 serve as a gap-filling mechanism for the treaty. The Service has also stepped in to interpret the term for purposes of the Code and treaties in revenue rulings,174 recognizing that, in accordance with OECD commentary, the concept has both a geographic and a temporal element.175

But why is this gap filling exercise even necessary? Can we not just assume the treaty-makers intended “fixed” to mean immobile and permanent? If that were the case, then a taxpayer’s business that is inherently mobile, such as a logging business or a floating sea facility, would escape taxation.176 A seasonal restaurant operating during the two-year World’s Fair would also be exempt despite the fact such presence is not simply temporary.177 These results would seem contrary to the intent of the treaty-makers and to the treaty’s purpose, and yet drafting a treaty to this level of detail would be impractical, particularly when domestic interpretive sources can be leveraged. The treaty-makers’ choice not to define “fixed place of business” is essentially a choice to incorporate the domestic meaning of the term—a decision that is quite common in the tax treaty context. Thus, we are left with instruments that have “generalized ‘treaty speak’ that has relatively little connotative value standing alone.”178

Hence, in addition to the non-substantive nature of tax treaties, we arrive at another reason for drafting (and interpreting) tax treaties in an incomplete manner—the complexity of tax law and the astonishing range of fact patterns to which tax law must apply. Unlike other areas of law, tax law applies to nearly all economic activities, all individuals, and all business enterprises, each with different attributes. Tax law does this while also providing critical revenue-raising and redistribution functions. Thus, tax law is often complexly drafted. Even so, the treaty, the statute, and even the regulations, cannot anticipate all the variations of these situations; thus a rich body of interpretive sources develops to accommodate these limitations.

Some of the literature on the interpretation of domestic tax statutes is helpful in illustrating the difficulties interpreters face in construing tax treaties. Much of it focuses on whether the interpretation of tax statutes is qualitatively different from other areas of law. As Michael Livingston has discussed, committee reports take on a significant role during the enactment process of tax legislation due to the conceptual, technical, and complex nature of such legislation.179 Accordingly, Livingston argues that the traditional intent-based and process-based objections to legislative history are not appropriate in the tax context.180

Relevant to the argument here, Livingston argues that the detailed nature of the Code and its frequent revisions may justify a “contextual approach to interpretation,” rather than a plain meaning approach, when it is necessary to create consistency within the complex statutory regime.181 The plain meaning approach, he argues, is also inappropriate since the self-contained nature of tax law creates specialized meanings for tax terms that do not correspond with their meanings in everyday parlance.182 Moreover, the complexity of tax law means that the text likely does not address the specific fact pattern at issue and thus reliance upon a variety of different sources is usually justified.183 Indeed, tax legislative history takes on a unique role in that committee reports typically go beyond explaining a statute and instead prescribe “an entire range of results that would otherwise be left to administrative or judicial discretion.”184 This legislative dynamic flows directly from the complexity of tax legislation.185

These observations have interesting implications in the tax treaty context. If the nature of the tax law requires reliance upon extrinsic interpretive aids in the domestic context, then the case for nonliteral interpretation should be even stronger when we interpose a treaty that borrows many of its concepts from the underlying domestic systems. The detailed nature of the domestic law means that treaty-makers may be unable to incorporate concepts directly, thus relying upon the rich array of extrinsic sources to fill in the details. Specialized meanings of tax terms make trying to ascertain a plain meaning shared by both treaty partners a difficult task. And, as in the statutory setting, the treaty will likely not address the factual scenario at issue. Finally, like legislative history in the domestic context, the technical explanations that accompany tax treaties, as well as the commentaries on the model treaties, are often very detailed and address specific factual patterns.

The highly complex demands upon tax law are one reason for the contemplated gaps in treaty drafting; another is the connection between taxation and state sovereignty.186 If nations are to retain some control over this important governmental role, they must build into tax treaties some degree of ambiguity and reference to the domestic law.187 Although sovereignty enters into most debates over international law, its connection to tax is particularly strong.188 Taxation implicates the revenue function of a nation, which is, in turn, the means by which public goods, such as national defense, are funded.189 Taxation also is an integral part of the state’s fiscal policy and is employed to affect consumer demand, economic growth, price stability, and unemployment.190 Tax treaties restrict a nation’s taxing jurisdiction over foreign persons. Although this foregone revenue may be offset by the residual taxation that the United States asserts over its own citizens and residents who pay lower foreign withholding taxes under the treaty, tax treaties undoubtedly alter the nation’s tax policy and mix of revenues. Moreover, international and domestic tax policies are interrelated; tax treaties may, for instance, implicate the debate over the proper corporate tax rate since multinationals receive benefits under these agreements.191

It is also important to note that taxation differs from the trade context, where economists claim that the elimination of trade barriers contributes to not only worldwide efficiency but also to national efficiency. In the international tax context, no such synergies exist.192 We therefore should expect a greater degree of retention over national policy in the tax treaty context. And we do, both implicitly, through ambiguity, and explicitly, through reference to the domestic tax laws.193

C.     The Similarity of Tax Treaties

The similarity between individual tax treaties casts further doubt upon the appropriateness of a plain meaning or intentionalist approach in tax treaty interpretation. Typically, we think of treaties as pacts between nations that tailor laws to the specific needs of the signatories. Under this view, the United States would take into account the particular investment dynamics in negotiating concessions from different partners. Thus, if the United States is a capital importer in relation to the treaty partner, it would argue for higher withholding rates since the revenue on taxing nonresidents on their source-based income would outweigh the loss of U.S. residual taxation on its residents transacting abroad.

Yet despite differences in relative capital flows and the taxing environments of our treaty partners, tax treaties are quite similar to one another.194 Although many countries have their own model treaty, they largely follow the OECD Model. This is because a nation has to justify any such departure during treaty negotiations and is required to pay a “price” for its inclusion.195 Another scholar has concluded that the OECD Model Tax Treaty has such an influence on tax treaties that “[o]ne can pick up any modern tax treaty and immediately find one’s way around, often even down to the article number.”196 Although the treaties are individually negotiated, the negotiations produce fewer deviations than one might expect.197 Effectively, the OECD Model has spawned a single multilateral regime, albeit comprised of numerous treaties.198

What does the similarity among tax treaties mean for their interpretation? First, because the treaties look less like individually negotiated contracts, limiting interpretation to the treaty text and negotiating materials seems less relevant. Second, a rich array of extrinsic sources is available to interpret the nearly uniform set of treaty principles.

D.     Tax Abuse

The reality of tax abuse requires non-treaty based solutions and hence also contributes to the interstitial nature of tax treaties. Combatting tax abuse is a major goal of domestic tax systems worldwide. To some extent, treaties further these goals through the use of information sharing provisions. In other respects, treaties enable, or at least fail to prevent, tax avoidance. The discussion below provides examples of tax abuse involving treaties and argues that invocation of domestic anti-abuse doctrines and administrative interpretations is necessary for each.

1.     Illustrations of Treaty Abuse

Residents of a non-treaty country may inappropriately invoke a treaty, a phenomenon known as treaty shopping. For instance, such residents may organize an entity in a treaty country without having economic ties to the treaty country in order to take advantage of the lower treaty withholding rate. Because the intention of the bilateral treaty structure is to provide reciprocal benefits to the residents of the two treaty countries and not to residents of a third country, treaties contain rules, known as limitation on benefits provisions, mandating some economic nexus to the treaty countries before treaty benefits can be utilized.199

Various other taxpayer abuses involve the avoidance of treaty limitations. For instance, treaties often contain a 12-month permanent establishment threshold for construction projects.200 Taxpayers may try to get around this rule by dividing their contracts into short-term periods, attributing each to a related company.201 The hiring out of foreign labor by an intermediary to obtain the benefit of exemption from source country taxation on compensation is another abuse.202 Taxpayers may also attempt to alter the characterization of income in order to obtain a favorable treaty result.203

Taxpayers also use tax treaties to avoid domestic taxation. Such abuses may include financing transactions that reduce the costs of borrowing through tax deductions; strategies involving arbitrage between the residency rules for the domestic system and the treaty; transfer pricing abuse; transactions that arbitrage the differences found in the domestic law of one or both treaty countries related to the characterization of income, the treatment of taxpayers, or timing differences; and transactions that abuse relief of double taxation mechanisms to achieve non-taxation in both states (double non-taxation).204

2.     Measures to Combat Treaty Abuse

Strategies to reduce treaty abuse take several different forms, and I argue here that resort to domestic anti-abuse laws and doctrines is necessary because a comprehensive treaty-based solution is impractical. The many opportunities for abuse of the treaty system and those created by it again reinforce the characterization of tax treaties as incomplete devices, thus requiring resort to extrinsic sources. Among these sources are legislative or administrative rules curtailing specific abusive situations and also the U.S. common law doctrines such as economic substance, substance over form, the sham transaction doctrine, and the step transaction doctrine.205 The economic substance doctrine has perhaps been the most influential, requiring that tax benefits not be extended to transactions that lack non-tax purposes and economic consequences.

In some instances of treaty abuse, the treaty could arguably be changed to close its loopholes. And indeed many treaties have evolved to provide for rules aimed at specific anti-abuse situations. For instance, the limitation on benefits rules require some nexus to the treaty countries before a taxpayer can invoke the treaty.206 Yet in a pattern familiar to tax professionals, the creativity of taxpayers ensures that the current rules will be evaded through new transaction forms. Indeed, because of this issue, the OECD and G20 have recently recommended through the BEPS Project that, in addition to revamped limitation on benefits rules and other specific anti-abuse limitations, treaties include a general anti-abuse rule that would disregard treaty benefits where one of the principal purposes of the transaction was to obtain such benefits.207

Going forward, the United States and other countries would be wise to close loopholes and include anti-abuse rules in its treaties. At least in certain abusive contexts, there is some international support for doing so. Yet altering the vast network of existing treaties in a timely manner will not be easy as a practical matter. Currently, there is a BEPS proposal for a multilateral treaty that may have the capacity to update existing treaties automatically, in which case this task may become easier.208 The details and scope of the project, however, have not yet been determined and participation by countries is optional. Also, participating members may be able to opt in and out of certain updates.209 Even if nations incorporated the general anti-abuse rule into existing treaties, either individually or through the multilateral treaty, judicial resort to the common law doctrines would be essential to understanding the scope of any such rule given its necessarily fact-intensive application.

Several courts have indeed invoked anti-abuse doctrines to deny taxpayers their anticipated tax results. For instance, in a recent case, Bank of New York Mellon Corp. v. Commissioner, a taxpayer engaged in a complex tax shelter and reported income from trust assets as foreign source income based on a resourcing provision in the U.S.–U.K. Income Tax Treaty.210 The court disregarded the transaction, which involved circular cash flows, and therefore resourced the income as domestic, effectively ignoring the treaty language.211 In so doing, the court relied upon the judicial economic substance doctrine as well as its codification by Congress in section 7701.212

In a leading case on limitation on benefits, Aiken Industries v. Commissioner, the court held that a transaction using back-to-back loans among related taxpayers lacked a business purpose and therefore that a treaty exemption did not apply to interest payments.213 In another case involving limitation on benefits, the court invoked the step transaction doctrine to disregard the receipt of interest payments that were exempt under a treaty.214 Congress’s recent codification of the economic substance doctrine arguably gives courts more authority to employ the doctrine in domestic and treaty contexts.215

The justification for relying upon broad anti-abuse rules to curb treaty abuse is similar to the justification for relying upon such rules in the domestic context. Anti-abuse rules displace certainty, yet without them the tax base would be eroded. Long ago, Stanley Surrey wrote: “It is clear that [anti-abuse provisions] save the tax system from the far greater proliferation of detail that would be necessary if the tax avoider could succeed merely by bringing his scheme within the literal language of substantive provisions written to govern the everyday world.”216 Closing individual loopholes through treaty text is a similarly futile and resource-demanding task, one in which the taxpayer is always one step ahead.217

Although the incorporation of anti-abuse rules is in tension with the plain language approach blessed by the Vienna Convention, similar accusations have been rejected in the purely domestic context. Specifically, the departure from a strictly textualist approach to combat abusive tax shelters has been largely accepted by American academic literature and the judiciary as necessary to defend the Code against tax avoidance. Noel Cunningham and James Repetti emphasized the connection between the rise of textualism and the proliferation of such shelters.218 Indeed, the Court itself seems to have shied away from a textualist approach in tax cases after a brief experimentation.219 Furthermore, the failure of British and Canadian courts to look beyond a statute’s text except in limited circumstances has resulted in an inability to develop robust judicial anti-avoidance doctrines and has arguably contributed to the growth of such shelters within those systems.220

My argument here is that, in interpreting treaties, taxpayer abuse necessitates that courts have the authority to rely upon both general and specific domestic rules aimed at abuse, including those administered at the agency level, as well as domestically refined judicial doctrines. Tax treaties function as limits upon the substantive provisions of the domestic law. The determination of true economic substance and hence the assessment of a tax liability is a necessary predicate to the application of a treaty, and therefore there is arguably no conflict between anti-abuse rules and the treaty.221 Additionally, it could be contended that tax anti-abuse rules have even been elevated to the status of international law.222 In addition to the OECD and G20’s recent acceptance of such rules in the tax treaty context,223 the International Court of Justice has applied an anti-abuse concept in interpreting other types of treaties, the Vienna Convention incorporates a good faith requirement, and numerous countries have developed anti-abuse doctrines of their own.224

E.     Types of Extrinsic Sources

This Article has argued that tax treaties are unsuited to interpretation as standalone vehicles; instead, their incomplete and uniform nature, the technicality of the tax system, and the reality of tax abuse require reliance on extrinsic sources, including ones outside of the negotiating process. This Subpart further discusses the types of extrinsic sources that are appropriate in the interpretive endeavor.

1.     Domestic Law

After accounting for the interstitial nature of treaties, one must determine which country’s laws should govern when domestic law is implicated. Article 3(2) of the OECD Model Treaty provides that the undefined term “shall, unless the context otherwise requires, have the meaning that it has at that time under the law of that State for the purposes of the taxes to which the Convention applies.”225 The U.S. Model Treaty employs a similar rule.226 Some have interpreted this provision to require incorporation of the source state’s laws since it is only that state’s laws that technically “apply.”227

Although it is correct that it is typically the source state whose laws are at issue, in meeting the duty to alleviate double taxation, the residence country must apply its own domestic foreign tax credit rules. The residence jurisdiction should thus be able to apply its own foreign tax credit limitations, and even its own characterization rules, to determine whether the source jurisdiction should have taxed the item.228 I believe this is the better reading of the “applies” language in article 3(2). Thus, article 3(2) of the Model Tax Treaties, by incorporating domestic law concepts, necessarily “accepts double taxation as a result.”229 This result is not inappropriate given the “pragmatic consideration that the authorities and the courts quite naturally understand their own law best.”230 Additionally, relying primarily on the interpreting countries’ resources ensures that taxpayers need only analyze a more limited world of interpretive principles and domestic laws.231

Also at issue is to what extent the domestic law should be used in interpreting tax treaties. The treaties themselves, as generally set forth in article 3(2) of the U.S. and OECD Model Treaties, command that the domestic law be employed whenever the term is left undefined “unless the context otherwise requires.”232 Some have taken the position that domestic law should only be used as a last resort.233 Klaus Vogel, on the other hand, argues that the “requires” language in article 3(2) mandates use of the domestic law unless there exists a strong argument not to do so.234 Still others contend that “given the general and noncomprehensive nature of tax treaties,” the “unless the context otherwise requires” language simply directs “the interpreter to render the most appropriate meaning.”235

In Samann v. Commissioner, the Court of Appeals interpreted a provision in the U.S.–Switzerland treaty that exempted royalties of taxpayers not having a permanent establishment in the United States.236 The court held that the taxpayer could not have a permanent establishment “at any time during the taxable year” because of a similar concept in the Code, which Treasury had incorporated into regulations.237 The court reasoned that the gap-filling provision for undefined terms in the treaty is indicative of the “signatories’ desire to retain their own scheme of taxation.”238 Article 3(2) does seem to suggest an effort to coordinate the tax treaty with the internal tax systems of the signatories. It therefore seems straightforward to agree with Vogel in concluding that this provision commands the interpreter to draw upon internal law to define an undefined term unless an interpretation can be found by other particularly convincing means.

But this does not end the inquiry; the generalist nature of treaties requires further use of domestic law. It may also be necessary to consult domestic law when the definition provided by the treaty is not sufficiently detailed or itself has undefined terms. Reference to the domestic law may also be appropriate when an undefined concept is analogous to, but worded differently than, a domestic concept. Finally, treaties often specifically defer to the state’s definition of a term, such as is the case with residency and real property.239

A further question arises as to whether the interpreter should apply the domestic law at the time the treaty was executed (a static approach) or instead apply current law (an ambulatory approach).240 U.S. courts generally follow the ambulatory approach,241 and the OECD commentaries have also endorsed looking to current law.242 An ambulatory approach is preferable for several reasons. Most importantly, updated domestic rules are necessary to fill the treaty’s gaps because of the need not only to combat tax abuse but to grow with a changing global economy. For instance, treaties were written before the digital economy. Does doing business on network servers located in the United States constitute a permanent establishment? It is impractical to draft treaties anticipating future events but also to renegotiate treaties for every necessary update.

An ambulatory approach also somewhat mitigates the worrisome phenomenon that occurs when the vast network of bilateral tax treaties locks in domestic policy.243 Moreover, taxpayers and courts may find it burdensome to apply outdated concepts and terms. Also, because it is the approach favored by the OECD, and one that is naturally contemplated given the interstitial nature of tax treaties, one can presume that it represents the intent of many contracting nations. A static approach might also contribute to double taxation because “the treaty would have no relevance to the new rule.”244 That is, if tax treaties lack a mechanism to remain current, harmonization of the treaty partners’ laws will be short-lived. Finally, the obligations set forth in tax treaties are reciprocal in nature, “reflecting mere exchanges of rights and duties between parties,” rather than absolute duties that are “resilient against shifts in party intention.”245 Accordingly, they are more amenable to an ambulatory interpretive approach.246

It could be argued that a new multilateral instrument updating the existing network of bilateral treaties, as mentioned above, diminishes the need for dynamic interpretation (and perhaps reliance on extrinsic materials in general). Aside from the implementation questions of such an instrument, which I have discussed,247 a multilateral instrument will not be able to comprehensively address the constant need to renovate treaties in light of changing circumstances. Indeed, the very proposal of such a radical instrument highlights the dramatic need for a fluid approach to interpretation, and the introduction of new changes to the treaty system through the multilateral instrument will make dynamic interpretation even more important.

Somewhat relatedly, new draft U.S. Model Treaty provisions may signify a trend towards dynamic interpretation built into the treaties themselves. Two new provisions that target double non-taxation hinge directly on evolving domestic law. The first denies treaty withholding rates on investment income to taxpayers eligible for “special tax regimes” that are introduced by the treaty partner subsequent to the treaty signing.248 The second turns off the treaty withholding rates for all taxpayers if a treaty country reduces its tax rate on investment income below 15% or shifts to a territorial regime.249 The U.S. Model Treaty is heavily influential given the United States’ importance in the markets and thus these types of provisions may very well impact the treaty policies of other nations.250 Treasury officials have even stated that the new model treaty provisions were intended to influence the OECD and its BEPS project.251 As with the introduction of the multilateral instrument, the evolutive character built into these provisions will almost certainly require reference to dynamic, domestic law sources in the interpretive process.

2.     Senate Materials

As discussed above,252 there are constitutional arguments for considering the Senate’s views, and even the late Justice Scalia signed on to a post-Stuart opinion that relies upon Senate hearings in its treaty interpretation.253 Apart from constitutional considerations, however, should courts rely upon Senate materials? Although such materials are unilateral, they generally represent an informed view of one of the treaty partners and thus are legitimate interpretive aids. Additionally, in the tax treaty context, the Senate reports are reflective of the technical expertise brought to bear by Treasury. Of course, such materials present dangers similar to those in the statutory context. For this reason, more weight should be given to those materials that best represent the views of the chamber, such as committee reports, and less weight to Senate materials that may represent the opinions of a lone Senator, such as floor statements.

3.     Executive Materials

In addition to the constitutional arguments for reliance upon Executive materials, the expertise upon which they draw makes the case for using them in the treaty context stronger. Administrative materials, however, range in usefulness. Courts generally give significant weight to executive branch interpretations of U.S. treaties, although, as illustrated in the discussion of the case law above, this does not mean that an agency’s interpretation always carries the day.254 Indeed, administrative interpretations may be given less weight in the treaty context because the interests of the U.S. government are adverse to that of the other treaty state. This differs from the domestic context where the interests of the executive branch and Congress are aligned.255

The Treasury’s Technical Explanation of the model treaty and each individual treaty is an integral part of the materials considered by the Senate in providing its advice and consent. Although Technical Explanations are almost never agreed to by the treaty party, they are sometimes transmitted to them and known to the other party in advance.256 Their importance in the enactment process and in the interpretive endeavor is well known in other jurisdictions. Although they likely do not have formal legal status that would entitle them to deference under administrative law principles,257 they are properly consulted as evidence of one treaty partner’s interpretation of the statute (or even both if the partner did not negotiate a deviation from a known interpretation).

The IRS rarely issues Treasury regulations interpreting tax treaties, but in the domestic context, properly promulgated Treasury regulations are entitled to Chevron deference.258 Informal guidance such as revenue rulings and revenue procedures are much more common in the treaty context. They also may be entitled to deference, although what type of deference is in flux, particularly in light of recent Supreme Court case law.259

Lowest on the hierarchy are revenue rulings or other administrative pronouncements made in the course of litigation. In the statutory context, the Court has precluded Chevron deference for an agency’s litigation position. However, the question of whether such a position, manifested in amicus briefs or revenue rulings, receives Skidmore deference is the subject of a circuit split.260 It may be problematic to rely on such guidance given that the incentive effects that occur at the litigation stage have the potential to be more self-serving than those at the interpretive stage.261 The interpretation also might represent the views of agency counsel rather than the views of the agency itself, it might have been developed without expertise and under time pressure, and there may have been no opportunity to consider differing viewpoints and evidence.262 Such retroactive guidance may also take the taxpayer by surprise. Of course, a court could still choose to apply Skidmore deference after applying these considerations to the specific factual scenario.263

Another question arises as to whether a later treaty reinterpretation by the executive branch should take precedence over an interpretation upon which the Senate relied in providing its advice and consent. This issue resulted in intense debate between the two branches when President Reagan reinterpreted the Anti-Ballistic Missile Treaty. In reaction to the President’s interpretation of the treaty, the Senate adopted the “Biden Condition” for all subsequent treaties, which stated that the treaty would be interpreted based on the shared views of the President and the Senate during the advice and consent stage.264 In response, the State Department pronounced that the President was not bound to prior interpretations, citing the important role given to the President by the Treaty Clause.265 In later litigation involving the treaty, the Senate Foreign Relations Committee stood its ground, and the executive branch appeared to retreat from its earlier position.266

The point here is not to decide these complex questions involving deep ambiguities within administrative and constitutional law doctrine but is instead to present a general overview of materials upon which courts may rely in the treaty interpretation process. Generally speaking, my approach does not differ from that of the Court’s in that I conclude that administrative materials are entitled to significant weight when created outside of the course of litigation.267 This conclusion is partially based on the traditional deference afforded the executive branch over foreign affairs and in administrative law generally, but, as discussed above, flows also from the incomplete nature of tax treaties, the related need to keep the international tax system abreast with a changing global economy, and the associated necessity of staving off tax abuse.

4.     International Materials

I have thus far emphasized domestic materials as aids of interpretation, but I do not mean to foreclose the use of extrinsic international materials. In particular, the OECD commentaries should be tapped as extremely important gap filling mechanisms.268 They are written by representatives of OECD nations, including the United States, and are binding upon OECD states to the extent reservations are not entered. Thus, they represent the specific intent of OECD members. Klaus Vogel has even argued that, because the commentaries are well known and easily obtainable, they have a status above that of preparatory work and thus courts need not invoke the exceptions to the plain meaning rule of the Vienna Convention to consult them.269

Although Vogel’s reasoning, in the technological age, could extend to insulate many types of extrinsic materials from the plain meaning rule, Vogel correctly notes that awareness of the commentaries, both specific and general, gives them legitimacy as tools of interpretation.270 The treaties themselves are based heavily on the OECD Model, which is itself a skeletal document, making reference to the official interpretations of the model vital. That being said, if a member country enters reservations with regard to a particular commentary (such reservations are recorded in the commentaries themselves), then the reservation is relevant for interpretative purposes.

The commentaries are designed to be ambulatory in nature. They themselves recommend applying the most current commentaries,271 and there is some support in the Vienna Convention for so doing since article 31 allows reference to subsequent agreements and practice.272 Additionally, an ambulatory approach assists to modernize an aging legal infrastructure, to adjust the treaties to new policies, and to guard against tax abuse. For these reasons, U.S. courts, as well as courts in other countries, have embraced the ambulatory approach.273 Indeed, it may be that nations follow the OECD Model so closely in drafting their own treaties precisely because they want to take advantage of the opportunity for ambulatory construction provided by the commentaries. Even certain countries that have explicitly rejected the ambulatory approach have later embraced it “simply to cope with the difficulty of some treaty interpretation questions that arise.”274

Outside of the commentaries, there are many other international sources available. First, if a domestic court construes a treaty provision, the interpretation may be persuasive even in cases involving other treaties with the shared language.275 A domestic court may also find relevant other countries’ interpretations of the treaty at issue or even parallel treaties.276 Subsequent practices by the treaty partners, such as diplomatic communications regarding a treaty interpretation or even silent acquiescence to a unilateral interpretation,277 are also widely accepted extrinsic sources and specifically condoned by the Vienna Convention. Finally, the Vienna Convention also allows use of the negotiating history, although such material is usually limited in the case of bilateral treaties.278

V.     Implementation Issues

A.     Objections

1.     Interaction with International Law

For obvious reasons, it is important that an interpretive methodology not lie in derogation of international law. The use of extrinsic sources, particularly unilateral ones, may be controversial from the perspective of international law but is accepted in some circumstances by article 32 of the Vienna Convention, which provides that courts look to supplemental materials to either: (1) confirm the meaning resulting from article 31; or (2) arrive at a meaning in the face of “ambiguous or obscure” text or “absurd or unreasonable” results from a literal application.279 Of course, article 32 begs the question of the threshold level of ambiguity or absurdity. It is quite difficult and even counterproductive to determine a threshold that is applicable to all cases. Nonetheless, because of the highly technical nature of the international tax system that causes frequent ambiguity and often results in special meanings, as well as the regularity of abusive transactions that will likely lead to absurd results, courts should invoke the exceptions to the Vienna Convention relatively frequently.

Another hurdle to the pragmatic approach in treaty interpretation is that the drafters of the Vienna Convention rejected a similar proposal by Professor McDougal, who argued for a less textual interpretive approach than what was eventually adopted. Although this approach would have allowed more flexibility in application and sources consulted, much like my own, the New Haven School of treaty interpretation, of which McDougal was a founder, also heavily focused on world policy norms, prescribing furtherance of “value distribution and the fundamental institutions that are compatible with the preferred system of public order.”280 In this respect, the New Haven School’s approach is quite different from the one I suggest.

Additionally, although commentators sometimes contend that the United States’ reliance on extrinsic sources deviates from the predominant interpretive approach with regard to tax treaties abroad,281 other countries consistently diverge from the textualist approach adopted by the Vienna Convention in the tax treaty context. For instance, courts in New Zealand and Canada have cited to unilateral materials, such as U.S. Technical Explanations and Senate reports.282 Numerous countries, such as Australia, Belgium, Denmark, Germany, Japan, the Netherlands, New Zealand, Sweden, Switzerland, and the United Kingdom rely upon the OECD commentaries.283 Indeed, the very existence of the OECD commentaries reflects international rejection of a strictly textualist based approach to tax treaty interpretation.

Although some countries are more formalist than others,284 the particular features of tax treaties and the realities of the international tax system have created substantial deviation away from the Vienna Convention. Additionally, the OECD’s embrace of domestic anti-abuse rules and ambulatory interpretation indicates a less formalist trajectory for tax treaty interpretation in the broader international community. Strict adherence to a plain meaning approach would thus lie in derogation to trends in the international community rather than bolstering international law.

2.     Double Taxation

It could also be argued that relying upon extrinsic sources in tax treaty interpretation would produce double taxation due to conflicting interpretations resulting from methodological flexibility and the many sources that may be consulted. The plain meaning approach, however, does not guarantee predictable, harmonious results. After all, the ordinary meaning of a term is often far from clear, especially when different legal cultures and languages are involved.285 Moreover, the extensive detail of the OECD commentaries supplies clear answers to many interpretive questions, thus providing more certainty and less variation than the plain meaning approach in many cases.

Additionally, tax treaties are usually applied only by the source country, leaving few instances where two countries are actually interpreting the same term.286 Even when the residence country is also the interpreter, tax treaties themselves have additional mechanisms, such as the mutual agreement procedure and arbitration clauses, that attempt to resolve disputes resulting in double taxation. These devices minimize the risk of double taxation that any discordant interpretations may bring.

Moreover, the domestic laws of developed countries already provide mechanisms, like the foreign tax credit and exemption of foreign source income, that prevent double taxation. As far back as 1963, Elisabeth Owens argued that “U.S. income tax treaties play a very marginal role in relieving double taxation . . . [since] the U.S. has unilaterally provided for the avoidance of double taxation.”287 More than a decade ago, Tsilly Dagan argued that, in fact, the domestic tools of nearly all countries eliminate double taxation of business taxpayers and thus make obsolete the stated function of tax treaties.288

3.     Planning Difficulties

Another objection to the use of extrinsic materials in tax treaty interpretation is that taxpayer planning would be disrupted by unsettled interpretations. For instance, the countries might arrive at different interpretations depending on the sources they consult or the interpretive methodology they adopt. This puts taxpayers in the position of navigating between multiple interpretations.

These problems, along with the problem of double taxation, will occur unless each country applies the identical interpretive methodology, a far-fetched goal. Still, one could argue that if all nations adopted the plain meaning approach set forth in the Vienna Convention, their interpretations would be closer to harmonization. As previously mentioned, however, such an approach does not promise a determinate meaning.289 This is especially true in the tax treaty context. As one author has stated: “There can be few legal documents, short of a national constitution, that cover more ground with fewer sentences than a tax treaty. Where is a court to turn in the numerous situations where the treaty will have no ‘plain meaning’?”290 A plain meaning approach would also lie in derogation of the wishes of treaty partners who intentionally design tax treaties interstitially. Nor does the plain meaning rule likely reflect the expectations of taxpayers, whose activities are carried out in reliance upon robust extrinsic sources that support the treaty.

B.     Harmonization Devices

For the reasons mentioned above,291 I am not particularly troubled by the risks of double taxation and planning difficulties presented by a pragmatic approach to interpretation. I do recognize, however, that if there are ways to harmonize interpretations to reduce double taxation and unpredictable results, then efforts should be made in that direction so long as they allow for the flexible use of extrinsic sources in the interpretive process. To assist in this endeavor, I have developed a set of harmonization devices that may serve to guide interpreters to a more uniform approach in the interpretation of tax treaties.

1.     A Loose Ordering of Sources

Ordering the available interpretive sources would ameliorate the critique that incorporation of extrinsic materials leads to unpredictable results and double taxation. Providing a positive account of the Court’s use of interpretive sources in domestic statutory interpretation cases, William Eskridge and Philip Frickey developed the famous funnel of abstraction.292 Statutory text is, of course, the most concrete of available interpretive sources, and current policy is the most abstract, with legislative history, purpose, and statutory evolution in between (in order of abstraction). The more concrete sources provide a more limited set of arguments than the more abstract. Eskridge and Frickey argue that this funnel represents a rough hierarchy, with a compelling textual argument winning over an equally compelling purposivist argument.293 That being said, the interpreter is not strictly bound by the hierarchy and will instead “move up and down” the funnel of abstraction, “evaluating and comparing the different considerations represented by each source of argumentation.”294 In addition to presenting the model as a descriptive account of the Court’s interpretive approach, Eskridge and Frickey defend the funnel from a theoretical perspective. Specifically, they argue that the insights of legal hermeneutics tell us that an interpreter draws upon an array of arguments and cannot avoid present-day policies and principles.295

From a practical perspective, the funnel of abstraction has become a useful blueprint for formulating and ordering interpretive arguments, as well as for predicting the range of possible rationales and outcomes in a statutory case. Extending this model to the tax treaty interpretation context may similarly assist the judiciary and taxpayers in organizing their reasoning, thus harmonizing treaty interpretations. The model I have created orders commonly used interpretive sources from most concrete to most abstract. The order of the sources roughly corresponds with their degree of bilateralness. Sources that are clearly available and shared between treaty parties are most concrete. Unilateral sources that represent policy norms or that are not contemporaneous are the most abstract. Within each category, there are subcategories, which I explain in the text. Additionally, the other treaty partners’ affirmative embrace or rejection of the interpretation set forth in each source will either bolster or discredit its usefulness.

Like the original model, it is intended to be both descriptive and normative. Although it provides a loose hierarchy of sources, it acknowledges that the interpreter will test her interpretation of the statute against the considerations presented at each level. Thus, although there is an order of priority in the model, it is one that represents only the typical case. The precise synthesis of sources would depend on the particular issue and facts at hand. I have chosen bilateralness as a metric in order to give effect to the parties’ shared intent, a goal that is central to treaty interpretation, but the flexibility inherent in the model recognizes that such an endeavor is unlikely to produce certain results and may, under certain circumstances, be overridden by other concerns.


Figure 1. A Practical Reasoning Model of Tax Treaty Interpretation296



The treaty text is, of course, the most concrete inquiry. Under the Vienna Convention, and in most U.S. tax treaty cases, the starting point is the treaty language. The text is primary since it is international law and parties have relied upon it. That being said, courts must be attuned to the special meaning of terms, as opposed to the plain meaning, since the former is common in the tax context. Other textual considerations, such as treaty structure, may enter into this analysis. A treaty’s direct incorporation of domestic law will be highly instructive and should be thought of as part of the treaty’s text. For instance, when a treaty refers to the laws of the treaty state in defining residency, then the local residency rules must be consulted. Additionally, when a treaty defers to local law in the case of an undefined term, then local law must supply the definition.

The model next considers materials that are closest to the treaty process—bilateral agreements and negotiating materials, in order of concreteness. Bilateral agreements may consist of: (1) contemporaneous agreements as to the meaning of a treaty term or concept; or (2) subsequent agreements by the competent authorities as to such meaning. The former may include affirmative assent to Senate conditions. The latter might also include the recent phenomena of joint technical explanations.297 Negotiating materials, or travaux préparatoires, are also included within this category.

Contemporaneous legislative and executive materials are in the next category due to the essential involvement of the two branches in the treaty-making process. Senate conditions are the most notable and, under some views, must be adhered to as a constitutional matter. Typically, the President must put the other treaty country on notice of any such condition.298 Treasury Technical Explanations, Joint Committee reports, and Senate Finance Committee Reports are also sources that are visible to our treaty partners and can assist in interpretation. Less useful are those types of materials that traditionally receive less deference in the domestic context, such as hearings, statements, or comments from executive officials. Similarly, the U.S. Model Treaty explanation is rarely updated, and thus may not reflect current U.S. negotiating positions, but it may nevertheless be useful in supplying meaning.

The OECD commentaries are also widely available to the international community and should be highly influential in treaty interpretation. Although contracting states do not directly adopt the commentaries, they bind OECD countries that do not make reservations to them. The commentaries also benefit from input by OECD countries. These commentaries are ambulatory in nature; as discussed above, it is appropriate to rely on recent interpretations so that treaties can reflect changing economic and technological environments, as well as ward off taxpayer abuse.299

The next level of the funnel prescribes attention to treaty purpose. The academic literature has cast doubt upon statutory purpose as an interpretive aid, questioning purposivism’s underlying assumption that legislatures produce reasonable, purposive results and that a single, public purpose can be articulated.300 These concerns extend to the treaty context.301 Also, as discussed above, the stated purposes of tax treaties—the relief of double taxation and avoidance of fiscal evasion—are not the only factors guiding treaty-makers.302 I therefore do not accept a wholesale “purposive” approach to tax treaty interpretation as some commentators recommend.303 Nonetheless, concerns about double taxation and other treaty goals should be a factor in the interpretive process, depending on the reliability of other available sources.

The remaining two levels might be thought of as “dynamic,” as it is referred to in the literature.304 The evolution of the treaty will primarily be the subsequent practices of the parties and its implementation by Congress, the courts, and agencies. Domestic law is thus a large component of this category. As mentioned, direct incorporation of domestic law is akin to a treaty supplement, in which case its status is that of the text. Domestic law, however, may also be consulted when the treaty employs ambiguous terms, such as business profits discussed above.305 The interpreter may also consult case law and administrative materials that apply and interpret such concepts, keeping in mind the aforementioned discussion of deference levels, with minimal deference to materials published in connection with pending cases.306 Court decisions of the treaty partner, or even of other countries, may be consulted, although practical barriers to doing so may exist, such as the difficulty in obtaining and translating foreign materials. Finally, general policy concerns, such as fairness, or tax policy concerns specific to the international tax context307 may also influence the judiciary’s interpretation.

2.     Interpretive Presumptions

Interpretive presumptions in the form of substantive canons could also be employed as harmonization devices. In this Subpart, I propose one such presumption and discuss presumptions that have been recommended in the literature. Note that, as interpretive presumptions, they should set up a probable interpretive outcome that can be overcome through support for a conflicting interpretation in other sources such as text, negotiating history, and model treaty commentaries. They thus offer more interpretive guidance than mere tiebreaker rules and less constraint upon the treaty drafters than clear statement rules.308

i.     Presumption Against Double Taxation

One way to ameliorate the problem of double taxation that is arguably exacerbated by the incorporation of extrinsic materials is for courts to adopt a presumption against interpretations that would create double taxation. Such a presumption would also assist in the harmonization of treaty interpretations across nations, therefore promoting predictability. There is precedent for such an interpretive canon. Private international law recognizes a principle of “common interpretation” in the interpretation of conflict of law rules, as well as in the areas of commercial paper and international sales, among others.309 Additionally, implicit within articles 31 and 32 of the Vienna Convention is the principle that parties endeavor to apply the treaty consistently.310

Indeed, some courts apply a rule similar to common interpretation in the context of tax treaties. Canadian courts have cited U.S. interpretations as influential, emphasizing that inconsistencies between U.S. and Canadian decisions might produce double taxation.311 Likewise, the U.S. courts have noted the interpretation of another country and the desirability to avoid double taxation in forming their own interpretation.312 Other foreign courts have gone so far as to treat foreign decisions as binding precedent,313 although in my view such an approach would violate national sovereignty. An interpretive presumption against double taxation would merely call upon courts to consult foreign authority on the subject if reasonably available and, if persuasive, interpret the treaty in harmony with such authority. Thus, this presumption has particular application when a foreign court has directly spoken on the issue or when a foreign nation has expressed its views diplomatically or through an amicus brief.

ii.     Interpreting Tax Treaties Liberally

Another interpretive presumption prescribes that courts interpret tax treaties liberally.314 This canon applies generally to treaty interpretation and was first embraced by the Court in Geofroy v. Riggs, which stated that

[i]t is a general principle of construction with respect to treaties that they shall be liberally construed, so as to carry out the apparent intention of the parties to secure equality and reciprocity between them. . . . [W]ords are to be taken in their ordinary meaning . . . and not in any artificial or special sense impressed upon them by local law . . . . [W]here a treaty admits of two constructions, one restrictive of rights that may be claimed under it and the other favorable to them, the latter is to be preferred.315

Thus, the canon has been employed largely to promote private rights bestowed by treaties,316 although it has fallen out of favor with the Court in the last several decades.317

Liberal interpretation is a poor fit in the tax treaty context. For one, specialized meanings in the tax context abound, as do express and implicit references to domestic law, thus constraining the interpreters. Liberal interpretation may also be in tension with another international law principle—that an agreement be undertaken in good faith.318 The requirement of good faith may be especially relevant in the context of tax abuse. Applying the two concepts may devolve into dueling purposes, producing outcomes that would catch treaty partners off-guard.319 Liberal interpretation might also lend itself to double non-taxation by the treaty parties, a phenomena of increasing concern.320

A liberal presumption is also at odds with the notion of sovereignty. To be sure, this is the case with all treaties and I do not mean to return to a world where sovereignty concerns required clear expression of all treaty obligations.321 Yet the tension is especially apparent in the tax context. Given the tie between taxation and the fisc,322 the relinquishing of taxing jurisdiction is not something that a sovereign would likely do implicitly or lightly.323 More importantly, a liberal presumption ignores the view that tax treaties primarily exist not to benefit taxpayers but to “sort out the interfacing aspects of [the treaty partners’] tax systems and to cede jurisdiction from the source nation to the taxpayer’s residence nation when the taxpayer’s contacts with the source nation are minimal.”324 Although my recommendation for the presumption against double taxation is similar to the liberal presumption rule because it results in benefitting taxpayers, it would do so in narrower circumstances—when there is direct foreign law on point. A presumption against double taxation is a softer version of the liberal presumption rule and more successfully navigates between the interests of sovereignty and harmonization.

iii.     Presumption in Favor of Source/‌Residence Country

Commentators have urged the adoption of other possible presumptions to encourage harmonization. John F. Avery Jones, for instance, has argued that when a treaty uses legal terms that comprise the substantive law of both treaty partners, the source country’s rules should apply.325 Similarly, Russell Osgood has contended that ambiguities be resolved in favor of the source country since “the basic message of tax treaties is to permit source country taxation.”326 Others, however, view the OECD and U.S. Model Treaties as confirming “taxation in the state of residence is the rule, while taxation in the country of source is the exception.”327 There is truth in both of these descriptions. For instance, one could point to the fact that the source country retains its right to tax business profits under tax treaties as support for the view that the source country has primary jurisdiction to tax active income. Indeed, this has been the consensus view for nearly 100 years.328 Nonetheless, treaties also lower or eliminate the withholding taxes imposed by the source country on investment income, thus supporting the understanding that tax treaties benefit residence countries, which are able to tax the investment income on a residual basis.329

From a normative perspective, there are also arguments on both sides as to which country has the primary right to tax income. T.S. Adams, one of the founders of our international system of taxation, argued that the source country had primary jurisdiction to tax.330 Under his view, source countries had the right to tax income since this was analogous to a sovereign’s claim over natural resources and because foreigners doing business within a nation used its resources.331 Edwin R.A. Seligman, a contemporary and intellectual opponent of Adams, argued that the residence country should receive exclusive power to tax since fairness concerns demand that ability-to-pay principles be based on worldwide income.332

A more nuanced canon might resolve ambiguities with regard to active business income in favor of the source country and those with regard to passive income unrelated to a taxpayer’s business in favor of the residence country. This would be in keeping with Adams’ view that a type of benefits theory justifies the source country’s jurisdiction to tax, a view that is not as convincing when applied to investment income because such income is often located in a source country purely for tax-motivated reasons. While still respecting the source country’s primary jurisdiction, this rule would limit it primarily to situations in which there is meaningful contact with the source state. It would also correspond with the dual nature of tax treaties, which benefits both source and residence countries depending on the type of income.

iv.     The Last in Time Rule and the Charming Betsy Canon

Another way to encourage harmonization across jurisdictions is to reduce the number of treaty overrides. To achieve this, one might be tempted to join those American scholars who have condemned the aforementioned last in time rule.333 Doing so, however, would present particular problems in the tax treaty context. I have previously written about the constitutional tension between the Treaty Clause, which contemplates a role only for the Senate and the Executive in treaty-making, and the Origination Clause, which mandates that revenue-raising laws originate in the House of Representatives.334 The House’s ability to override legislation likely contributes to the current stability over this precarious arrangement. Taking that power away from the House would not only compound the constitutional problems, it would likely result in the House reasserting its prerogative in this area through more direct means.335 Moreover, undoing the last in time rule would, in the tax context, require both judicial and legislative participation since the rule has been codified in the Code.336

Still, the last in time rule is not without limits. For instance, there is some authority for requiring legislative intent to override before abrogating a treaty.337 More importantly, a longstanding rule of interpretation, the Charming Betsy canon,338 in which courts construe statutes in a way that does not conflict with international law, softens the harshness of the last in time rule. The canon is an important piece of the puzzling U.S. relationship with international law,339 and its regular deployment in the tax treaty area could potentially reduce the number of treaty overrides that occur.340 There is authority to apply the canon forcefully in this context because Congress has affirmatively approved of its usage with regard to tax treaties.341

VI.     Conclusion

This Article argues that a differentiated approach to treaty interpretation is merited based on objective characteristics shared within a treaty category. Specifically, the completeness of a treaty informs whether extrinsic materials should be heavily relied upon. Tax treaties are skeletal in nature and the reality of the international taxing system requires them to be so. Accordingly, they mandate a fluid interpretive methodology that encompasses many actors and sources, in a sense, their flesh and blood. The risk of double taxation that such an approach entails is overstated and can be minimized through harmonization devices. In contrast, a plain meaning approach to tax treaty interpretation carries consequences: unintentional encroachment upon the sovereign’s domestic tax system, policy ossification in a rapidly changing global economy, and abusive transactions that reduce the effective tax rate to zero. These are all predictable and serious dangers. The use of a pragmatic approach to tax treaty interpretation diminishes these risks.









  1. [1]. Vienna Convention on the Law of Treaties art. 31, May 23, 1969, 1155 U.N.T.S. 331 (entered into force Jan. 27, 1980) [hereinafter Vienna Convention].

  2. [2]. See Julian Arato, Accounting for Difference in Treaty Interpretation over Time, in Interpretation in International Law 205, 205–12 (Andrea Bianchi et al. eds., 2015).

  3. [3]. Id.

  4. [4]. Note that my analysis is limited to double income tax treaties rather than other types of tax treaties and agreements, such as estate and gift tax treaties or tax information exchange agreements. My analysis also does not apply to treaties that have ancillary tax effects or contain isolated tax provisions. I limit my thesis to the jurisdictional elements of double tax treaties relating to income, which comprise the backbone of the treaty, rather than those aspects that have substantive, operative effects—specifically, nondiscrimination, exchange of information, and mutual agreement provisions. Although I focus on American sources and case law, my general conclusion—that liberal use of extrinsic sources is appropriate in the interpretation of tax treaties—is applicable to other legal cultures.

  5. [5]. Treaty-based gap-filling mechanisms exist; however, these are incomplete. For instance, article 3(2) of the Model Treaty prescribes that domestic law steps in to provides definitions for undefined terms “unless the context otherwise requires.” U.S. Model Income Tax Convention of November 15, 2006 art. 3(2) (U.S. Dep’t of Treasury 2006), Gap-filling may nonetheless be necessary when the domestic definitions themselves contain vague terms, as is often the case, or to determine if “the context otherwise requires.” Id. Article 21 attempts to close jurisdictional gaps by providing that types of income not dealt with by specific treaty provisions be taxed only by the country of the taxpayer’s residence. Id. art. 21(1). Nonetheless, extrinsic resources may still need to be consulted to determine how to categorize the income at issue.

  6. [6]. See infra notes 247–50 and accompanying text (discussing the OECD/G20 Base Erosion and Profit Shifting Project (“BEPS”) and the 2015 U.S. draft model treaty provisions).

  7. [7]. Michael P. Van Alstine, The Judicial Power and Treaty Delegation, 90 Calif. L. Rev. 1263, 1266 (2002) (recounting the paucity of treaty interpretation scholarship); see also Curtis J. Mahoney,Note, Treaties as Contracts: Textualism, Contract Theory, and the Interpretation of Treaties, 116Yale L.J. 824, 828 (2007) (same). For notable American contributions in the field of treaty interpretation, see generally David J. Bederman,Revivalist Canons and Treaty Interpretation, 41 UCLA L.Rev. 953 (1994) (recounting that treaties’ status as sui generis instruments has created a chaotic approach to their interpretation by the Court resulting largely in deference to the executive branch); Curtis A. Bradley,ChevronDeference and Foreign Affairs, 86Va. L. Rev. 649 (2000) (arguing for Chevron deference with respect to the executive’s foreign affairs interpretations, including that of treaties); Robert M. Chesney,Disaggregating Deference: The Judicial Power and Executive Treaty Interpretations, 92Iowa L. Rev. 1723, 1727 (2007) (arguing that the deference doctrine is “unsettled” and suggesting “link[ing] the practice of deference to a defensible theoretical foundation”); Evan Criddle, The Vienna Convention on the Law of Treaties in U.S. Treaty Interpretation, 44 Va. J. Int’l L. 431 (2004) (arguing against the nationalist approach to treaty interpretation on the basis of interpreting treaties like contracts); John Norton Moore,Treaty Interpretation, the Constitution and the Rule of Law, 42Va. J. Int’l L. 163, 174 (2001) (arguing against the “the ‘dual’ approach” in treaty interpretation); Martin A. Rogoff, Interpretation of International Agreements by Domestic Courts and the Politics of International Treaty Relations: Reflections on Some Recent Decisions of the United States Supreme Court, 11 Am. U. J. Int’l L. & Pol’y 559 (1996) (critiquing the Rehnquist Court’s restrictive approach to treaty interpretation); Scott M. Sullivan,Rethinking Treaty Interpretation, 86Tex. L. Rev. 777 (2008) (contending that Skidmore deference, rather than Chevron deference, is appropriate in the treaty context); Andrew Tutt, Treaty Textualism, 39 Yale J. Int’l L. 283 (2014) (setting forth an originalist argument for a textualist interpretation of treaties); Michael P. Van Alstine, TheDeath of Good Faith in Treaty Jurisprudence and a Call for Resurrection, 93Geo. L.J. 1885 (2005) [hereinafter Van Alstine, The Death of Good Faith] (arguing for the revival of a liberal interpretation of treaty obligations); Michael P. Van Alstine,Dynamic Treaty Interpretation, 146 U. Pa. L. Rev. 687 (1998) [hereinafter Van Alstine, Dynamic Treaty Interpretation] (arguing that certain private law treaties delegate to courts the authority to supply substantive law as gap-filling mechanisms in line with general treaty principles); and Mahoney,supra (proposing a relational contract approach to treaty interpretation). The foreign treaty interpretation literature is vast. For notable works, see generally 1 Treaty Interpretation and the Vienna Convention on the Law of Treaties: 30 Years on (Malgosia Fitzmaurice et al. eds., 2010) (collecting treaty articles); Rudolf Bernhardt, Evolutive Treaty Interpretation, Especially of the European Convention on Human Rights, in 42 German Yearbook of International Law 11 (Jost Delbrücke & Rainier Hoffman eds., 1999); Rudolf Bernhardt, Interpretation in International Law, in 2 Encyclopedia of Public International Law 1416 (Rudolf Bernhardt ed., 1995); Malgosia Fitzmaurice, The Practical Working of the Law of Treaties, in International Law 172 (Malcolm D. Evans ed., 3d ed. 2010) (discussing “structural questions and fundamental problems” of treaty law); and I.M. Sinclair, Vienna Conference on the Law of Treaties, 19 Int’l & Comp. L.Q. 47 (1970) (providing background information on the Vienna Conference and discussing issues that arose during the conference debate).

  8. [8]. Although the American work on tax treaty interpretation is quite useful in application, by and large, it addresses specific interpretive issues rather than recommending a general interpretive approach. See, e.g., David H. Brockway,Interpretation of Tax Treaties and Their Relationship to Statutory Law—A U.S. Perspective, in35 Report of Proceedings of the Thirty-Fifth Tax Conference 619 (Canadian Tax Found. 1983); Michael S. Kirsch, The Limits of Administrative Guidance in the Interpretation of Tax Treaties, 87 Tex. L. Rev. 1063 (2009) (suggesting that, to achieve deference in the judiciary, the U.S. Treasury should issue narrowly tailored regulations instead of technical explanations to support its treaty interpretations); Russell K. Osgood,Interpreting Tax Treaties inCanada, the United States, and the United Kingdom, 17Cornell Int’l L.J. 255 (1984) (rejecting contractualism and reciprocity as bases for interpretation and arguing instead for a presumption in favor of the source country); Sidney I. Roberts, Great-West Life Assurance Company v. United States:Exploration of the U.S. Interpretation of Treaties, 30Canadian Tax J. 759 (1982); Robert Thornton Smith, Tax Treaty Interpretation by the Judiciary, 49 Tax Law. 845, 887–88 (1996) (contending that a purposive approach should apply to tax treaty interpretation); John A. Townsend, Tax Treaty Interpretation, 55 Tax Law. 219 (2001) (advocating for more liberal use of Senate and executive materials). But see Federal Income Tax Project: International Aspects of United States Income Taxation II (Am. Law Inst. 1992) (condoning a plain meaning approach).

  9. [9]. The foreign academic literature on tax treaties is more comprehensive. See, e.g.,7 Franciscus Antonius Engelen, Interpretation of Tax Treaties Under International Law (2004); 54 Fundamental Issues and Practical Problems in Tax Treaty Interpretation (Michael Schilcher & Patrick Weninger eds., 2008); 3 Tax Treaty Interpretation (Michael Lang ed., 2001); Klaus Vogel et al., Klaus Vogel on Double Taxation Conventions: A Commentary to the OECD-, UN- and US-Model Conventions for the Avoidance of Double Taxation of Income and Capital (1991); John F. Avery Jones et al.,The Interpretation of Tax Treaties with Particular Reference to Article 3(2) of the OECD ModelI, 1984Brit. Tax Rev. 14; Int’l Fiscal Ass’n,Interpretation of Tax Treaties,40 Bull. for Int’l Fiscal Documentation 75, 75–86 (1986); Sidney I. Roberts & Peter A. Glicklich,U.S. Interprets Netherlands–U.S. Treaty by Reference to Later Treaties with Other Nations, 34Canadian Tax J. 228 (1986); David Rosenbloom, Tax Treaty Interpretation, 34 Bull. for Int’l Fiscal Documentation 543 (1980); Ian Sinclair, Interpretation of Tax Treaties, 40 Bull. for Int’l Fiscal Documentation 75 (1986); Klaus Vogel,Double Tax Treaties and Their Interpretation, 4 Int’l Tax & Bus. Law. 1 (1986); David A. Ward,Abuse of Tax Treaties, in15 Essays on International Taxation397 (Herbert H. Alpert & Kees van Raad eds., 1993).

  10. [10]. I label this approach “pragmatic” since it rejects adherence to the foundationalist theories of interpretation—intentionalism, purposivism, and textualism. Instead, it follows the view of the pragmatists that no single interpretive source or methodology is adequate, that no single plain meaning exists, and that no single legislative intent or purpose can be discovered. Early pragmatists include, William Eskridge, Philip Frickey, and Daniel Farber. See William N. Eskridge, Jr. & Philip P. Frickey, Statutory Interpretation as Practical Reasoning, 42 Stan. L. Rev. 321, 321–22 (1990); Daniel A. Farber, Book Review, The Hermeneutic Tourist: Statutory Interpretation in Comparative Perspective, 81 Cornell L. Rev. 513, 522 (1996).

  11. [11]. Federal Income Tax Project: International Aspects of United States Income Taxation II 27, 30–31.

  12. [12]. See OECD, OECD/G20 Base Erosion and Profit Shifting Project: Explanatory Statement 4–7 (2015), For the final BEPS reports, see BEPS 2015 Final Reports, OECD, (last visited Mar. 13, 2016). For a comprehensive review of the BEPS proposals, see Yariv Brauner, What the BEPS?, 16 Fla. Tax Rev. 55 (2014).

  13. [13]. Treasury officials have repeatedly suggested that the agency plans to work on general guidance on treaties at the request of many practitioners. See, e.g., Alison Bennett, General Guidance on Treaties Expected for Coming Business Plan, Treasury Official Says, Daily Tax Rep., Nov. 7, 2012, at G-1 to G-2.

  14. [14]. Treasury Releases Select Draft Provisions for Next U.S. Model Income Tax Treaty, U.S. Dep’t Treasury (May 20, 2015),

  15. [15]. See infra notes 58–76 and accompanying text.

  16. [16]. In the statutory interpretation context, James Brudney and Corey Ditslear have similarly cautioned against adoption of a monolithic approach toward interpreting statutes without considering subject matter. James J. Brudney & Corey Ditslear, The Warp and Woof of Statutory Interpretation: Comparing Supreme Court Approaches in Tax Law and Workplace Law, 58 Duke L.J. 1231, 1251–52 (2009). As evidence of this trend toward uniformity, the authors point to the decreasing use of legislative history in the Court’s tax law cases, in spite of a historical tradition of relying on such materials in order to leverage the rich expertise of the tax-writing committees.

  17. [17]. Almost all tax treaties emphasize their purpose of avoiding double taxation by stating in the recital of the treaty the following: “Convention Between the United States of America and ___ for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income.” Philip F. Postlewaite & David S. Makarski, The A.L.I. Tax Treaty Study—A Critique and a Modest Proposal, 52 Bull. Sec. Tax’n 731, 734 n.2 (1999).

  18. [18]. The lower the withholding tax is in the source state, the higher the residual tax will be in the residence state. For instance, assume Country A, the residence state, taxes income at a 35% rate with a credit for foreign taxes paid. If a Country A resident earns $100 income in Country B that is taxed at a 30% withholding rate, then Country B will receive $30 of tax revenue and Country A will receive only $5 of revenue. Now assume a treaty lowers the Country B withholding tax to 15%. In that case, Country B will receive $15 of tax revenue, and Country A will receive $20 of revenue as opposed to $5 without the treaty. Why would Country B enter into this deal? For one, it hopes that the lowering of withholding tax rates will increase foreign investment within its borders. Additionally, because the benefits are reciprocal, it may receive more revenues with the treaty than without it, depending upon the balance of investment flows between the two countries.

  19. [19]. U.S. Model Income Tax Convention of November 15, 2006 art. 25 (U.S. Dep’t of Treasury 2006). More recent treaties also provide for binding arbitration processes to resolve double taxation issues. See, e.g., Model Convention with Respect to Taxes on Income and on Capital art. 25(5) (Org. for Econ. Co-operation & Dev. 2014),
    treaties/2014-model-tax-convention-articles.pdf; Protocol Amending the Convention Between the Government of the United States of America and the Government of the French Republic for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and Capital, U.S.–Fr. art. 26(5)–(6), Jan. 13, 2009, S. Treaty Doc. No. 111-4 (2009),; Convention Between the Government of the United States of America and the Government of the Kingdom of Belgium for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, U.S.–Belg. art. 24(7)–(8), Nov. 27, 2006, S. Treaty Doc. No. 110-3 (2007),; Protocol Amending the Convention Between the United States of America and the Federal Republic of Germany for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income and Capital and to Certain Other Taxes, U.S.–Ger. art. 25(5), June 1, 2006, S. Treaty Doc. No. 109-20 (2006),

  20. [20]. See Postlewaite & Makarski, supra note 17, at 790.

  21. [21]. Patrick Driessen, Is There a Tax Treaty Insularity Complex?, Tax Notes, May 7, 2012, at 745, 748.

  22. [22]. Postlewaite & Makarski, supra note 17, at 802.

  23. [23]. OECD, Addressing Base Erosion and Profit Shifting 8 (2013); Stafford Smiley, Mutual Administrative Assistance in Tax Matters, Corp. Tax’n, May–June 2013, at 21.

  24. [24]. Michael J. Graetz & Michael M. O’Hear, The “Original Intent” of U.S. International Taxation, 46 Duke L.J. 1021, 1066 (1997).

  25. [25]. Id. at 1066 n.181.

  26. [26]. Townsend, supra note 8, at 226.

  27. [27]. U.S. Const. art. II, §2, cl. 2 (“[The President] shall have Power, by and with the Advice and Consent of the Senate, to make Treaties, provided two thirds of the Senators present concur....”).

  28. [28]. See Rebecca M. Kysar, On the Constitutionality of Tax Treaties, 38 Yale J. Int’l L. 1, 22 (2013) (providing a brief overview of the tax treaty process).

  29. [29]. U.S. Model Income Tax Convention of November 15, 2006 art. 28 (U.S. Dep’t of Treasury 2006).

  30. [30]. See Kevin C. Kennedy, Conditional Approval of Treaties by the U.S. Senate, 19 Loy. L.A. Int’l & Comp. L.J. 89, 112–14, 114 n.118, 157–59 app.9 (1996) (noting that reservations were attached to 18 out of 97 tax treaties).

  31. [31]. Diane Ring, When International Tax Agreements Fail at Home: A U.S. Example, 41 Brooklyn J. Int’l L. (forthcoming 2016) (on file with author).

  32. [32]. Whitney v. Robertson, 124 U.S. 190, 194–95 (1888).

  33. [33]. Section 894(a)(1) requires that “[t]he provisions of this title shall be applied to any taxpayer with due regard to any treaty obligation of the United States which applies to such taxpayer.” I.R.C. §894(a)(1) (2012). Section 7852(d) provides that “neither the treaty nor the law shall have preferential status by reason of its being a treaty or law.” I.R.C. §7852(d)(1) (2012). According to the conference report, the “due regard” language in section 894(a) “simply provides for giving the treaty that regard which it is due under the ordinary rules of interpreting the interactions of statutes and treaties,” thus “where a treaty obligation has been superseded for internal U.S. law purposes, no effect need be given to the treaty under the agreement’s provision.” Technical and Miscellaneous Revenue Act of 1988, H.R. Rep. No. 100-1104, vol. 2, at 12–13 (1988) (Conf. Rep.). This interpretation comports with section 894’s cross-reference to section 7852, which essentially codifies the last in time rule. Id. The enacted language of sections 894(a) and 7852(d) was a compromise between the House’s position that later enacted statutes would always take precedence over treaties, regardless of intent and Treasury’s position that Congress must explicitly override a treaty in order for the statute to trump. See Irwin Halpern, United States Treaty Obligations, Revenue Laws, and New Section 7852(d) of the Internal Revenue Code, 5 Fla. Int’l L.J. 1, 6 (1989); see also Kathleen Matthews, Treasury Encouraged by Finance Treaty Override Substitute, Tax Notes, Aug. 15, 1988, at 662, 662–64.

  34. [34]. Murray v. Schooner Charming Betsy, 6 U.S. (2 Cranch) 64, 118 (1804).

  35. [35]. See, e.g., Olympic Airways v. Husain, 540 U.S. 644, 650 (2004) (noting that it is “the Court’s responsibility to read the treaty in a manner ‘consistent with the shared expectations of the contracting parties’”); Air Fr. v. Saks, 470 U.S. 392, 399 (1985) (same); Washington v. Wash. Commercial Passenger Fishing Vessel Ass’n, 443 U.S. 658, 675 (1979) (noting that a treaty is a contract between two sovereign nations); Sullivan v. Kidd, 254 U.S. 433, 439 (1921) (“Writers of authority agree that treaties are to be interpreted upon the principles which govern the interpretation of contracts in writing between individuals, and are to be executed in the utmost good faith, with a view to making effective the purposes of the high contracting parties.”); Tucker v. Alexandroff, 183 U.S. 424, 437 (1902) (citing 1 James Kent, Commentaries on American Law 174 (1826), and analogizing treaties to private contracts); Fourteen Diamond Rings v. United States, 183 U.S. 176, 182 (1901) (Brown, J., concurring) (stating that “[i]n its essence [a treaty] is a contract. It differs from an ordinary contract only in being an agreement between independent states instead of private parties”); Worcester v. Georgia, 31 U.S. (6 Pet.) 515, 581 (1832) (“What is a treaty? The answer is, it is a compact formed between two nations or communities, having the right of self government.”); Foster v. Neilson, 27 U.S. (2 Pet.) 253, 314 (1829) (recognizing that Article II treaties are not acts of legislation but contracts between nations); Ware v. Hylton, 3 U.S. (3 Dall.) 199, 219 (1796) (same); Coplin v. United States, 6 Cl. Ct. 115, 126 n.11 (1984) (same), rev’d, 761 F.2d 688 (Fed. Cir. 1985), aff’d sub nom. O’Connor v. United States, 479 U.S. 27 (1986).

  36. [36]. Bederman, supra note 7, at 1022; see also Van Alstine, Dynamic Treaty Interpretation, supra note 7, at 691 (“[T]he Court’s treaty jurisprudence has fallen under the strong influence of a resurgent strain of formalism in domestic statutory interpretation.”).

  37. [37]. Bederman, supra note 7, at 1026.

  38. [38]. Restatement (Third) of Foreign Relations Law of the United States §325 reporters’ note 4 (Am. Law Inst. 1986).

  39. [39]. Choctaw Nation of Indians v. United States, 318 U.S. 423, 431–32 (1943) (“[T]reaties are construed more liberally than private agreements, and to ascertain their meaning we may look beyond the written words to the history of the treaty, the negotiations, and the practical construction adopted by the parties.”).

  40. [40]. Bederman, supra note 7, at 970–71.

  41. [41]. Id. at 972.

  42. [42]. See generally Sital Kalantry, The Intent-to-Benefit: Individually Enforceable Rights Under International Treaties, 44 Stan. J. Int’l L. 63 (2008).

  43. [43]. Id. at 100.

  44. [44]. Id. at 89–90.

  45. [45]. Id. at 85–87.

  46. [46]. Id.

  47. [47]. See Mahoney, supra note 7, at 835–36.

  48. [48]. See id. at 840–42.

  49. [49]. See Townsend, supra note 8, at 244.

  50. [50]. Restatement (Third) of Foreign Relations Law of the United States §314(1) & cmt. b (Am. Law Inst. 1986); Bederman, supra note 7, at 958. Since the treaty partner allowed the treaty to enter into force with the express conditions, it could even be argued that such Senate conditions are binding on the treaty partner. Stefan A. Riesenfeld & Frederick M. Abbott, The Scope of U.S. Senate Control over the Conclusion and Operation of Treaties, 67 Chi.-Kent L. Rev. 571, 602–03 (1991).

  51. [51]. Nw. Bands of Shoshone Indians v. United States, 324 U.S. 335, 351–53 (1945); Haver v. Yaker, 76 U.S. (9 Wall.) 32, 35 (1869); Restatement (Third) of Foreign Relations Law of the United States §314 cmt. d; Bederman, supra note 7, at 958. But see N.Y. Indians v. United States, 170 U.S. 1, 23 (1898) (failing to rely on an express Senate interpretation when the President had rejected the interpretation).

  52. [52]. Bederman, supra note 7, at 958.

  53. [53]. Air Fr. v. Saks, 470 U.S. 392, 397, 403 (1985); Immigration & Naturalization Serv. v. Stevic, 467 U.S. 407, 417–18 (1984); Trans World Airlines, Inc. v. Franklin Mint Corp., 466 U.S. 243, 250, 257 (1984); Warren v. United States, 340 U.S. 523, 527 n.5 (1951); Cook v. United States, 288 U.S. 102, 119 n.19 (1933); In re Korean Air Lines Disaster of Sept. 1, 1983, 932 F.2d 1475, 1489–90 (D.C. Cir. 1991); Frolova v. Union of Soviet Socialist Republics, 761 F.2d 370, 376 (7th Cir. 1985). But see Coplin v. United States, 6 Cl. Ct. 115, 131 n.16 (1984) (declining to follow an interpretation that tracked an exchange between a Senate and an executive branch representative because the interpretation was not transmitted to the other treaty partner), rev’d, 761 F.2d 688 (Fed. Cir. 1985), aff’d sub nom. O’Connor v. United States, 479 U.S. 27 (1986).

  54. [54]. The origins of executive branch deference in the treaty context can be traced to Foster v. Neilson, 27 U.S. (2 Pet.) 253, 307 (1829). Courts have continued to embrace it throughout the decades. See, e.g., Sumitomo Shoji Am., Inc. v. Avagliano, 457 U.S. 176, 184–85 (1982); Kolovrat v. Oregon, 366 U.S. 187, 194 (1961) (“[T]he meaning given [to treaties] by the departments of government... charged with their negotiation and enforcement is given great weight.”); United States v. Pink, 315 U.S. 203, 220–21 (1942); Nielsen v. Johnson, 279 U.S. 47, 52–53 (1929); Terrace v. Thompson, 263 U.S. 197, 223 (1923); Charlton v. Kelly, 229 U.S. 447, 468 (1913); Ortman v. Stanray Corp., 371 F.2d 154, 157 (7th Cir. 1967). But see Trans World Airlines, Inc., 466 U.S. at 276 n.5 (noting that treaty interpretation of agency “not entitled to any special deference”).

  55. [55]. See, e.g., Bradley, supra note 7, at 702–03 (arguing that treaty interpretation by the executive branch is within the ambit of Chevron analysis).

  56. [56]. Factor v. Laubenheimer, 290 U.S. 276, 298 (1933) (“Until a treaty has been denounced, it is the duty of both the government and the courts to sanction the performance of the obligations reciprocal to the rights which the treaty declares and the government asserts, even though the other party to it holds to a different view of its meaning.” (citing Charlton v. Kelly, 229 U.S. 447, 472–73 (1913))).

  57. [57]. See Bederman, supra note 7, at 962.

  58. [58]. United States v. Stuart, 489 U.S. 353 (1989).

  59. [59]. See id. at 371 (Scalia, J., concurring).

  60. [60]. Id. at 375. Scalia also argued that the Supreme Court, prior to Stuart, has never consulted Senate materials to interpret a treaty and that two lower court decisions in 1988 are the first federal decisions to do so. He then criticized the Restatement (Third) on the Foreign Relations Law for approving of such materials. Id. at 375–76. A Harvard Law professor subsequently wrote a scathing critique of Scalia’s description of the law, unearthing several such cases in both categories. See generally Detlev F. Vagts, Editorial Comment, Senate Materials and Treaty Interpretation: Some Research Hints for the Supreme Court, 83 Am. J. Int’l L. 546 (1989).

  61. [61]. Chan v. Korean Air Lines, Ltd., 490 U.S. 122 (1989).

  62. [62]. Id. at 136 (Brennan, J., concurring).

  63. [63]. Sale v. Haitian Ctrs. Council, Inc., 509 U.S. 155 (1993).

  64. [64]. Id. at 179–80.

  65. [65]. Id. at 180.

  66. [66]. Id. at 176, 184–87.

  67. [67]. Id. at 194–95 (Blackmun, J., dissenting).

  68. [68]. Olympic Airways v. Husain, 540 U.S. 644, 656–57 (2004).

  69. [69]. Id. at 659–60 (Scalia, J., dissenting).

  70. [70]. Medellín v. Texas, 552 U.S. 491 (2008).

  71. [71]. Id. at 508 (emphasis omitted).

  72. [72]. See Curtis A. Bradley, Self-Execution and Treaty Duality, 2008 Sup. Ct. Rev. 131, 177–78 (discussing the Court’s textual emphasis as relevant in ascertaining the views of the Senate); Carlos Manuel Vázquez, Treaties as Law of the Land: The Supremacy Clause and the Judicial Enforcement of Treaties, 122 Harv. L. Rev. 599, 659 (2008) (discussing the Court’s textual focus in supporting the conclusion that the Court attempted to reconstruct the treaty parties’ intent rather than the unilateral view of U.S. treaty-makers).

  73. [73]. See David J. Bederman, Medellín’s New Paradigm for Treaty Interpretation, 102 Am. J. Int’l L. 529, 532 (2008); Jean Galbraith, The Oxford Guide to Treaties Symposium: Comparing International and US Approaches to Interpretation, Opinion Juris (Nov. 9, 2012, 11:30 AM),

  74. [74]. Medellín, 552 U.S. at 554–55 (citing the Senate’s 1945 and 1946 hearings and debates).

  75. [75]. Id. at 506 (“The interpretation of a treaty, like the interpretation of a statute, begins with its text.”).

  76. [76]. Bederman, supra note 73, at 540.

  77. [77]. See generally Vienna Convention, supra note 1.

  78. [78]. Robert E. Dalton, The Vienna Convention on the Law of Treaties: Consequences for the United States, 78 Am. Soc’y Int’l L. Proc. 276, 276 (1984).

  79. [79]. See, e.g., Gonzalez v. Gutierrez, 311 F.3d 942, 949 n.15 (9th Cir. 2002) (“While the United States is not a signatory to the Vienna Convention, it is the policy of the United States to apply articles 31 and 32 as customary international law.”); Aquamar S.A. v. Del Monte Fresh Produce N.A., Inc., 179 F.3d 1279, 1296 n.40 (11th Cir. 1999) (same); Kreimerman v. Casa Veerkamp, S.A. de C.V., 22 F.3d 634, 638 n.9 (5th Cir. 1994) (same); see also Fujitsu Ltd. v. Fed. Express Corp., 247 F.3d 423, 433 (2d Cir. 2001) (explaining that the court applies customary international law principles elucidated in the Vienna Convention on the Law of Treaties when interpreting treaties even though the United States is not a signatory); Haitian Ctrs. Council, Inc. v. McNary, 969 F.2d 1350, 1361–62 (2d Cir. 1991), rev’d sub nom., Sale v. Haitian Ctrs. Council, Inc., 509 U.S. 155 (1993) (“Rather than having evolved from a judicial common law,... principles of treaty construction are themselves codified[] in Article 31 of the Vienna Convention on the Law of Treaties.”); Risinger v. SOC LLC, No. 2:12-CV-00063-MMD-PAL, 2014 WL 804802, at *2 (D. Nev. Feb. 27, 2014) (“Although the United States has not ratified the Vienna Convention on the Law of Treaties, our Court relies on it as an authoritative guide to the customary international law of treaties, insofar as it reflects actual state practices.” (quoting Mora v. New York, 524 F.3d 183, 196 n.19 (2d Cir. 2008))); Busby v. State, 40 P.3d 807, 814–15 (Alaska Ct. App. 2002) (applying the interpretive principles of the Vienna Convention).

  80. [80]. Report from William Rogers, Sec’y of State, to President Richard Nixon (Oct. 18, 1971), 65 Dep’t St. Bull. 684, 685 (1971).

  81. [81]. See, e.g., Sale, 509 U.S. at 191 (Blackmun, J., dissenting) (citing the Vienna Convention as incidental support for the common law notion that a treaty must be interpreted according to its “ordinary meaning”); Weinberger v. Rossi, 456 U.S. 25 (1982) (distinguishing between the definition of treaties in the Vienna Convention and the Constitution); see also Criddle, supra note 7, at 434.

  82. [82]. Vienna Convention, supra note 1, art. 31(1).

  83. [83]. Vogel, supra note 9, at 35.

  84. [84]. Vienna Convention, supra note 1, art. 31(4).

  85. [85]. Id. art. 32.

  86. [86]. Ian Brownlie, Principles of Public International Law 607 (2d ed. 1973).

  87. [87]. U.N. Conference on the Law of Treaties, Official Records: Summary Records of the Plenary Meetings and of the Meetings of the Committee of the Whole, U.N. Doc. A/CONF.39/11/Add.1 (1968).

  88. [88]. A drafter of the Vienna Convention stated that “the starting point of interpretation is the elucidation of the meaning of the text, not an investigation ab initio into the intentions of the parties.” Report of the International Law Commission on the Work of Its Eighteenth Session, 21 U.N. GAOR Supp. No. 9, U.N. Doc. A/6309/Rev.1 (1966), reprinted in [1966] 2 Y.B. Int’l L. Commission 220, U.N. Doc. A/CN.4/191.

  89. [89]. Richard A. Falk, On Treaty Interpretation and the New Haven Approach: Achievements and Prospects, 8 Va. J. Int’l L. 323, 340 (1967).

  90. [90]. Id. at 341.

  91. [91]. U.N. Conference on the Law of Treaties, supra note 87, at 168. I discuss the Vienna Convention’s rejection of McDougal’s amendment and its implications for the arguably similar pragmatic approach prescribed herein. See infra notes 281–84 and accompanying text.

  92. [92]. Id. at 168.

  93. [93]. See Van Alstine, Dynamic Treaty Interpretation, supra note 7, at 744 (stating that the Vienna Convention accepts extrinsic sources only in narrow circumstances).

  94. [94]. Vienna Convention, supra note 1, art. 31(3)(c); Vogel, supra note 9, at 34–35, 37–39.

  95. [95]. See Restatement (Third) of Foreign Relations Law of the United States §325 cmt. g (Am. Law Inst. 1986); Bederman, supra note 7, at 972.

  96. [96]. Bederman, supra note 7, at 975.

  97. [97]. The American Legal Institute’s Federal Income Tax Project also generally supports the American embrace of extrinsic materials in the tax treaty context. Federal Income Tax Project: International Aspects of United States Income Taxation II 28–29 (Am. Law Inst. 1992).

  98. [98]. See Sumitomo Shoji Am., Inc. v. Avagliano, 457 U.S. 176, 189 (1982) (deferring to the plain language of the treaty when both treaty partners agreed with such interpretation); Maximov v. United States 373 U.S. 49, 51 (1963) (adopting an interpretation that was “consonant with [the treaty’s] language, purpose and intent”).

  99. [99]. Restatement (Third) of Foreign Relations Law of the United States §325 n.4.

  100. [100]. Air Fr. v. Saks, 470 U.S. 392, 400 (1985) (“In interpreting a treaty it is proper, of course, to refer to the records of its drafting and negotiation.”).

  101. [101]. See, e.g., United States v. Stuart, 489 U.S. 353 (1989); Pierpoint v. United States, No. 83-0354-2, 1983 WL 1665 (D.S.C. Oct. 3, 1983) (reading committee report to determine nonexemption of U.S. income taxes); Great-W. Life Assurance Co. v. United States, 678 F.2d 180 (Ct. Cl. 1982) (relying upon U.S. legislative history to hold that interest paid by a Canadian corporation to a Canadian life insurance corporation was not exempt under the U.S.–Canada treaty); Crow v. Comm’r, 85 T.C. 376, 392 (1985) (examining committee reports to determine relationship between 26 U.S.C. §877 and the treaty’s savings clause); Burghardt v. Comm’r, 80 T.C. 705, 708–09 (1983) (interpreting exemption of estates of nonresident aliens to be liberal in light of committee report).

  102. [102]. See, e.g., Xerox Corp. v. United States, 41 F.3d 647, 656 (Fed. Cir. 1994); Snap-On Tools, Inc. v. United States, 26 Cl. Ct. 1045, 1065 (1992); Taisei Fire & Marine Ins. Co. v. Comm’r, 104 T.C. 535, 558 (1995).

  103. [103]. Subsequent practices may reflect bilateral agreement on interpretations, in which case they are highly influential. See, e.g., El Al Isr. Airlines, Ltd. v. Tseng, 525 U.S. 155, 167 (1999); Zicherman v. Korean Air Lines Co., 516 U.S. 217, 226 (1996).

  104. [104]. See, e.g., N.W. Life Assurance Co. of Can. v. Comm’r, 107 T.C. 363, 378 (1996); Taisei Fire & Marine Ins. Co., 104 T.C. at 548.

  105. [105]. For instance, in Maximov v. United States, the Supreme Court held that the treaty exemption for U.K. citizens and residents on their capital gains did not apply to U.K. beneficiaries of a U.S. trust that realized capital gains. Maximov v. United States, 373 U.S. 49, 56 (1963). The Court reasoned that, in the absence of treaty text on point, domestic law, which defines a trust as a separate U.S. taxpayer, applied to deny the treaty benefits to the U.K. beneficiaries. Id. at 51–53.

  106. [106]. See Bank of N.Y. Mellon Corp. v. Comm’r, 140 T.C. 15, 48 (2013) (denying a foreign tax credit since “U.S. tax laws and treaties do not recognize sham transactions... that [lack] economic substance as valid for tax purposes”), aff’d, 801 F.3d 104 (2d Cir. 2015).

  107. [107]. Great-W. Life Assurance Co. v. United States, 678 F.2d 180, 185 (Ct. Cl. 1982).

  108. [108]. Id. at 188–89.

  109. [109]. United States v. Stuart, 489 U.S. 353, 365 (1989) (quoting article XIX of the 1942
    U.S.–Canada treaty).

  110. [110]. Id. at 361­­–62.

  111. [111]. Id. at 356.

  112. [112]. Id. at 357.

  113. [113]. Id. at 366–68.

  114. [114]. Id. at 369.

  115. [115]. See supra notes 58–60 and accompanying text.

  116. [116]. Stuart, 489 U.S. at 373–77.

  117. [117]. See generally Xerox Corp. v. United States, 14 Cl. Ct. 455 (1988), rev’d, 41 F.3d 647 (Fed. Cir. 1994).

  118. [118]. Xerox Corp. v. United States, 41 F.3d 647, 660 (Fed. Cir. 1994).

  119. [119]. Id. at 652.

  120. [120]. Id. at 656.

  121. [121]. Id. at 655.

  122. [122]. Id. at 652–54, 656.

  123. [123]. Id. at 656–57.

  124. [124]. Snap-On Tools, Inc. v. United States, 26 Cl. Ct. 1045 (1992).

  125. [125]. Id. at 1075.

  126. [126]. Id. at 1052.

  127. [127]. Id. at 1064–65.

  128. [128]. Id. at 1066.

  129. [129]. Id. at 1068.

  130. [130]. Id. at 1068 & n.22.

  131. [131]. Id. at 1074.

  132. [132]. See Great-W. Life Assurance Co. v. United States, 678 F.2d 180, 190 (Ct. Cl. 1982) (relying upon U.S. legislative history to hold that interest paid by a Canadian corporation to a Canadian life insurance corporation was not exempt under the U.S.–Canada treaty); Pierpoint v. United States, No. 83-0354-2, 1983 WL 1665 (D.S.C. Oct. 3, 1983) (reading committee report to determine nonexemption of U.S. income taxes); Crow v. Comm’r, 85 T.C. 376, 392 (1985) (examining committee reports to determine relationship between 26 U.S.C. §877 and the treaty’s savings clause); Burghardt v. Comm’r, 80 T.C. 705, 708–09 (1983) (interpreting an exemption of estates of nonresident aliens to be liberal in light of a committee report).

  133. [133]. See, e.g., United States v. Stuart, 489 U.S. 353, 368 n.8 (1989) (noting the Technical Explanation’s silence on the issue at hand); Clayton v. United States, 33 Fed. Cl. 628, 654–56 (1995) (relying on the Technical Explanation of U.S.–Canada Treaty to interpret the source of trust income); Taisei Fire & Marine Ins. Co. v. Comm’r, 104 T.C. 535, 558 (1995) (relying upon the Technical Explanation of a prior 1970 U.S.–Belgium treaty in interpreting whether reinsurance was to be included within the coverage of the independent agent provision in the 1971 U.S.–Japan treaty); Riley v. Comm’r, 74 T.C. 414, 427 (1980) (comparing the Technical Explanations of two treaties in determining the residency of a temporary visitor to Canada).

  134. [134]. See, e.g., Nat’l Westminster Bank, PLC v. United States, 44 Fed. Cl. 120 (1999) (reading the OECD commentaries to support the taxpayer’s argument that the Treasury regulation’s formulaic allocation of interest expense was not intended by the treaty parties); Pekar v. Comm’r, 113 T.C. 158 (1999) (referring to the Technical Explanation to the U.S. Model Income Tax Treaty in interpreting whether a beneficiary of the U.S–Germany treaty had to reduce his foreign tax credit in accordance with the domestic limitations under the Alternative Minimum Tax); N.W. Life Assurance Co. of Can. v. Comm’r, 107 T.C. 363 (1996) (reasoning that guidance under the OECD model treaty supported the taxpayer’s interpretation and that the later enacted treaty language trumped the conflicting domestic statute).

  135. [135]. See, e.g., Taisei Fire & Marine Ins. Co., 104 T.C. at 558 (effectively adopting the interpretation supplied by later commentaries, which conflicted with the interpretation of the commentaries at the time the treaty went into effect).

  136. [136]. See generally Arato, supra note 2.

  137. [137]. Although I argue that taxation has a special connection to state sovereignty, I do so in the context of explaining why tax treaty drafters leave open treaty gaps rather than basing the methodology on that connection alone. See infra notes 186–93 and accompanying text.

  138. [138]. Arato, supra note 2, at 210.

  139. [139]. Id. Arato’s statement was made with regard to whether an evolutionary approach to interpretation is subject-specific, but his conclusion is applicable to the interpretive endeavor generally.

  140. [140]. Investment treaties, for instance, may be construed as partially implicating human rights since they bestow property rights upon individuals. Id. at 211.

  141. [141]. See Rosalyn Higgins, Time and the Law: International Perspectives on an Old Problem, 46 Int’l & Comp. L.Q. 501, 519 (1997) (arguing that the parties’ intent, object, and purpose—rather than subject matter—explain a differential approach to treaty interpretation). My hesitation with this approach is that, in the tax treaty context, the object and purpose is indeterminate. See infra notes 300–03 and accompanying text.

  142. [142]. See Driessen, supra note 21, at 747–48.

  143. [143]. See generally Anthea Roberts, Power and Persuasion in Investment Treaty Interpretation: The Dual Role of States, 104 Am. J. Int’l L. 179 (2010).

  144. [144]. See generally Arato, supra note 2.

  145. [145]. For instance, like tax treaties, earlier iterations of bilateral investment treaties contained vague terms. See Wolfgang Alschner, Interpreting Investment Treaties as Incomplete Contracts: Lessons from Contract Theory 9–10 (July 18, 2013),
    abstract_id=2241652 (arguing that first generation bilateral investment treaties are “highly incomplete” whereas second generation bilateral investment treaties are “complex and comprehensive”). On the other hand, treaties that represent single transactions between governments, such as a transfer of land, are more likely to contain specific terminology, thus justifying a more formalist approach to interpretation. See Mahoney, supra note 7, at 846–48 (arguing for a flexible interpretative approach with regard to treaties that resemble relational contracts and a more formalist approach with regard to treaties that resemble spot contracts).

  146. [146]. For instance, procedural treaties, like the Hague Service Convention, typically allow each signatory to comply with the treaty objectives through flexible use of their own domestic law, agencies, and processes. They are thus general and non-operative in their approach, much like tax treaties.

  147. [147]. See, e.g., Model Convention with Respect to Taxes on Income and on Capital art. 3(2) (Org. for Econ. Co-operation & Dev. 2014); U.S. Model Income Tax Convention of November 15, 2006 art. 3(2) (U.S. Dep’t of Treasury 2006).

  148. [148]. Tax treaties do contain some operative provisions—specifically, nondiscrimination, exchange of information, and mutual agreement provisions—but my analysis is limited to the jurisdictional provisions relating to income. These latter provisions form the backbone of the tax treaty and are, by far, more likely to present interpretive difficulties.

  149. [149]. Vogel, supra note 9, at 14.

  150. [150]. Id.

  151. [151]. Id.

  152. [152]. Klaus Vogel et al., United States Income Tax Treaties 26 (1989).

  153. [153]. Indeed, sometimes treaties deviate from Code terminology simply to correspond with the foreign text in which the treaty is written. Id.

  154. [154]. Model Convention with Respect to Taxes on Income and on Capital art. 3(2) (Org. for Econ. Co-operation & Dev. 2014); see also U.S. Model Income Tax Convention of November 15, 2006 art. 3(2) (U.S. Dep’t of Treasury 2006).

  155. [155]. See infra notes 224–28 and accompanying text.

  156. [156]. For instance, the definitions section of both the U.S. and OECD model treaties contain only ten and eight defined terms, respectively. U.S. Model Income Tax Convention of November 15, 2006 art. 3(1); Model Convention with Respect to Taxes on Income and on Capital art. 3(1).

  157. [157]. U.S. Model Income Tax Convention of November 15, 2006 art. 3(1)(b); see also Model Convention with Respect to Taxes on Income and on Capital art. 3(1)(b).

  158. [158]. Model Convention with Respect to Taxes on Income and on Capital art. 4(2); see also U.S. Model Income Tax Convention of November 15, 2006 art. 4(3).

  159. [159]. The model treaty acknowledges that business profits is left undefined but encompasses “income derived from any trade or business.” Technical Explanation Accompanying the U.S. Model Income Tax Convention art. 7, ¶ 1 (U.S. Dep’t of Treasury 2006). “Trade or business” is a technical term of art in the tax world. The Supreme Court has held that, for purposes of section 162(a) of the Code, a trade or business encompasses those activities that are continuous, regular, and primarily entered into for profit. This definition has been extended to other contexts, such as whether income is effectively connected to a trade or business for international tax purposes. See, e.g., Comm’r. v. Groetzinger, 480 U.S. 23, 27–36 (1987); Free-Pacheco v. United States, 117 Fed. Cl. 228, 259–60 (2014); Park v. Comm’r., 136 T.C. 569, 580–81 (2011), rev’d and remanded, 722 F.3d 384 (D.C. Cir. 2013).

  160. [160]. See, e.g., U.S. Model Income Tax Convention of November 15, 2006 art. 7(3).

  161. [161]. Townsend, supra note 8, at 266 (“[A] facial reading of the treaty would seem not to permit the U.S. Code limitations on business expenses, thus giving treaty partners doing business in the United States a competitive advantage over similarly situated U.S. taxpayers. Nevertheless, it has never been seriously argued that U.S. Code limitations do not apply in determining the treaty partner’s U.S. [permanent establishment] tax base.”).

  162. [162]. 26 U.S.C. §162(a) (2012) (“There shall be allowed as a deduction all the ordinary and necessary expenses paid or incurred during the taxable year in carrying on any trade or business....”). Facing a similar query, the IRS released a Technical Advice Memorandum concluding that section 882(c)(2)’s denial of a deduction to a foreign corporate taxpayer that failed to file a U.S. tax return did not violate the nondiscrimination provision of the treaty; instead the provision was merely an administrative provision that was necessary in light of foreign taxpayers being less discoverable by the Service. I.R.S. Tech. Adv. Mem. 1999-41-007 (Oct. 15, 1999).

  163. [163]. I.R.S. Tech. Adv. Mem. 1999-41-007, supra note 162, at 9–10.

  164. [164]. Model Convention with Respect to Taxes on Income and on Capital art. 11(3) (Org. for Econ. Co-operation & Dev. 2014).

  165. [165]. Boulez v. Comm’r, 83 T.C. 584 (1984).

  166. [166]. Id. at 588.

  167. [167]. Id. at 596.

  168. [168]. 26 U.S.C. §871(b) (2012).

  169. [169]. Model Convention with Respect to Taxes on Income and on Capital art. 7(a) (emphasis added); see also U.S. Model Income Tax Convention of November 15, 2006 art. 7(1) (U.S. Dep’t of Treasury 2006).

  170. [170]. See, e.g., Lee A. Sheppard, Why Do We Need Treaties?, Tax Notes, Nov. 19, 2012, at 825, 826; Reuven S. Avi-Yonah & Oz Halabi, Double or Nothing: A Tax Treaty for the 21st Century 1–2 (Univ. of Mich. Law Sch. Scholarship Repository, Law & Econ. Working Paper No. 66, 2012), http://

  171. [171]. U.S. Model Income Tax Convention of November 15, 2006 art. 5(1).

  172. [172]. To be sure, the “permanent establishment” concept, at least as has been interpreted by the domestic laws of developed countries and the OECD commentaries, disadvantages developing countries by limiting their jurisdiction over source income. That being said, incorporating domestic law interpretations of the concept may enable developing countries to prevent the definition of “permanent establishment” from becoming too narrow. See Lee Sheppard, Beyond Tax Avoidance: Managing Multinationals’ Tax and Contractual Relationships with Developing Countries 28–33 (unpublished manuscript) (on file with author) (presenting ways in which developed countries can reinterpret or alter permanent establishment clauses to their advantage).

  173. [173]. 26 U.S.C. §864(c)(4)(B), (c)(5)(A) (2012).

  174. [174]. For example, the Service ruled that an office used by a group of salespeople for advertising and soliciting orders constituted a fixed place of business under a treaty. Rev. Rul. 62-31, 1962-1 C.B. 367.

  175. [175]. OECD Model Tax Convention on Income and on Capital art. 5, ¶ 5 (Org. for Econ. Co-operation & Dev. 2005).

  176. [176]. See Rev. Rul. 56-165, 1956-1 C.B. 849 (ruling that the demonstration and selling of logging equipment in several U.S. forests over a two-year period constituted a permanent establishment).

  177. [177]. See Rev. Rul. 67-322, 1967-2 C.B. 469 (ruling that such a restaurant constituted a permanent establishment).

  178. [178]. Richard L. Reinhold & Catherine A. Harrington, What NatWest Tells Us About Tax Treaty Interpretation, Tax Notes, Apr. 14, 2008, at 169, 203.

  179. [179]. Michael Livingston, Congress, the Courts, and the Code: Legislative History and the Interpretation of Tax Statutes, 69 Tex. L. Rev. 819, 836–37 (1991). In an empirical project, James Brudney and Corey Ditslear have demonstrated that the Court historically relied more heavily upon legislative history and structural canons of interpretation in the tax context than in other contexts, perhaps due to the recognized expertise of tax committees and the perception of tax law as a coherent system. See generally Brudney & Ditslear, supra note 16.

  180. [180]. Livingston, supra note 179, at 822–23.

  181. [181]. Id. at 827­–28.

  182. [182]. Id. at 828­–29.

  183. [183]. Id. at 829–30 (observing that, in most tax cases, resort to extrinsic sources is necessary since Congress will rarely have provided a specific interpretation); see also Lawrence Zelenak, Thinking About Nonliteral Interpretations of the Internal Revenue Code, 64 N.C. L. Rev. 623, 675 (1986) (arguing that a literal interpretation is justified, even in the face of complexity, if Congress has identified the specific set of facts in the case). I would contend that the likelihood that the drafters contemplated a specific set of facts is even less likely in the tax treaty context.

  184. [184]. Livingston, supra note 179, at 839.

  185. [185]. Id. at 842.

  186. [186]. Townsend, supra note 8, at 267–68 (“[G]iven the dynamics of treaty negotiation, particularly of tax treaty negotiation implicating one of the most vital governmental functions (the revenue function), the negotiators must build in some wiggle room that permits the respective countries to interpret and apply the same provision differently, limited by the overall parameters of the treaty text.”). Lee Sheppard puts this dynamic in a characteristically blunt fashion: “One of the many glitches... is that [tax] treaties are vague documents—deliberately so in order to encourage countries to sign them.” Lee A. Sheppard, Treaty Countries’ Right to Use Domestic Law, Tax Notes, Oct. 22, 2012, at 347.

  187. [187]. One might argue that the last in time rule offers sufficient protection over U.S. tax policy since a later-enacted statute can override any undesirable treaty results. Global politics and the realities of the legislative process make such an action unrealistic. First, such overrides violate international law. Vienna Convention, supra note 1, at 339; see also Reuven S. Avi-Yonah, Tax Treaty Overrides: A Qualified Defence of U.S. Practice, in Tax Treaties and Domestic Law 65 (Guglielmo Maisto ed., 2006) (“[Overrides] clearly violate[] international law as embodied by the Vienna Convention on the Law of Treaties.”); Org. for Econ. Cooperation & Dev., Comm. on Fiscal Affairs, OECD Committee on Fiscal Affairs Report on Tax Treaty Overrides, 2 Tax Notes Int’l 25, 26–28 (1990). Second, the legislative process has a status quo bias that makes congressional action difficult, even where legislative preferences support such action taking place. See generally Rebecca M. Kysar, Automatic Legislation (2015) (unpublished manuscript) (on file with author). Third, not every country has a last in time rule and thus cannot override treaty obligations through domestic legislation.

  188. [188]. See Diane M. Ring, What’s at Stake in the Sovereignty Debate?: International Tax and the
    , 49 Va. J. Int’l L. 155, 157 (2008) (noting the “particular strength to the claims for tax sovereignty”).

  189. [189]. Id. at 167.

  190. [190]. Id. at 168–69.

  191. [191]. See Kysar, supra note 28, at 32.

  192. [192]. See generally Michael J. Graetz, The David R. Tillinghast Lecture: Taxing International Income: Inadequate Principles, Outdated Concepts, and Unsatisfactory Policies, 54 Tax L. Rev. 261 (2001).

  193. [193]. Of course, this is not to say that cooperation among nations is undesirable in the taxing context; in fact, tax competition and tax abuse make such cooperation in the national interest. Some scholars would even go further to suggest a view of sovereignty that demands positive duties upon nations in exercise of their taxing function. See, e.g., Allison Christians, Sovereignty, Taxation and Social Contract, 18 Minn. J. Int’l L. 99, 101–02 (2009).

  194. [194]. Related to this observation, there is literature casting doubt upon whether developing nations should enter into tax treaties as they are currently drafted since their uniformity fails to take into account the divergent interests of developed and developing nations. Victor Thuronyi, Tax Treaties and Developing Countries, in Tax Treaties: Building Bridges Between Law and Economics 441, 450–55 (Michael Lang et al. eds., 2010) (suggesting a tax treaty for developing nations that would forgo lowered withholding rates and instead be limited to exchange of information, administrative cooperation, nondiscrimination, tie-breaker rules, and mutual agreements); see also Tsilly Dagan, The Tax Treaties Myth, 32 N.Y.U. J. Int’l L. & Pol. 939, 991 (2000).

  195. [195]. Michael Long, Tax Treaty Policy, in National Tax Policy in Europe: To Be or Not to Be? 191, 206 (Krister Andersson et al. eds., 2007).

  196. [196]. John F. Avery Jones, The David R. Tillinghast Lecture: Are Tax Treaties Necessary?, 53 Tax L. Rev. 1, 2 (1999).

  197. [197]. Differences, of course, exist between tax treaties. See generally 11 Departures from the OECD Model and Commentaries: Reservations, Observations and Positions in EU Law and Tax Treaties (EC & Int’l Tax Law Series, Guglielmo Maisto ed., 2014). Variances among the treaties, however, are vastly overshadowed by their similarities. See Federal Income Tax Project: International Aspects of United States Income Taxation 3–4 (Am. Law Inst. 1987); Yariv Brauner, An International Tax Regime in Crystallization, 56 Tax L. Rev. 259, 291 (2003) (discussing “the (almost) consensus about the impressive extent of similarity and even unity within the multitude of tax treaties”).

  198. [198]. See Reuven S. Avi-Yonah, Commentary, 53 Tax L. Rev. 167, 169 (2000) (discussing the existence of an international tax regime which “rests mainly on the bilateral tax treaty network” and “which has definable principles that underlie it and are common to the treaties”).

  199. [199]. See, e.g., U.S. Model Income Tax Convention of November 15, 2006 art. 22 (U.S. Dep’t of Treasury 2006).

  200. [200]. U.S. Model Income Tax Convention of November 15, 2006 art. 5(3); see also Model Convention with Respect to Taxes on Income and on Capital art. 5(3) (Org. for Econ.
    Co-operation & Dev. 2014).

  201. [201]. Org. for Econ. Co-operation & Dev., Preventing the Granting of Treaty Benefits in Inappropriate Circumstances, Action 6: 2015 Final Report 69 (2015), http://www.keep

  202. [202]. Id.

  203. [203]. For instance, a shareholder may enter into a transaction to avoid dividend characterization under domestic rules, thereby allowing treaty exemption as a capital gain. Comm. of Experts on Int’l Cooperation in Tax Matters, Treaty Abuse and Treaty Shopping, at 17–18, U.N. Doc. E/C.18/2006/2 (Oct. 30–Nov. 3, 2006). Taxpayers may also convert dividends to interest in order to take advantage of a lower withholding tax (or complete exemption) on interest. Id. at 18.

  204. [204]. Org. for Econ. Co-operation & Dev., supra note 201, at 78.

  205. [205]. The judicial anti-abuse doctrines often overlap and apply concurrently. The economic substance doctrine requires that a transaction have economic purpose, apart from tax savings, in order to be respected. See Coltec Indus., Inc. v. United States, 454 F.3d 1340, 1347 (Fed. Cir. 2006). The substance over form doctrine prescribes the tax consequences that flow from the economics of the transaction rather than its formalities. See BB&T Corp. v. United States, 523 F.3d 461, 471 (4th Cir. 2008). The sham transaction doctrine ignores transactions that have no economic consequences other than tax benefits. See Kirchman v. Comm’r, 862 F.2d 1486, 1490 (11th Cir. 1989). Finally, the step transaction doctrine integrates the steps of various transactions, basing the tax liability on a single event. See Gregory v. Helvering, 293 U.S. 465, 469–70 (1935).

  206. [206]. See, e.g., U.S. Model Income Tax Convention of November 15, 2006 art. 22 (U.S. Dep’t of Treasury 2006).

  207. [207]. Org. for Econ. Co-operation & Dev., supra note 201, at 55.

  208. [208]. See Org. for Econ. Co-operation & Dev., Developing a Multilateral Instrument to Modify Bilateral Tax Treaties, Action 15: 2015 Final Report 9 (2015).

  209. [209]. Id. at 42.

  210. [210]. Bank of N.Y. Mellon Corp. v. Comm’r, 140 T.C. 15, 16–17 (2013).

  211. [211]. Id. at 48.

  212. [212]. Id. at 30–35, 48.

  213. [213]. Aiken Indus., Inc. v. Comm’r, 56 T.C. 925, 933–35 (1971).

  214. [214]. Del Commercial Props., Inc. v. Comm’r, 251 F.3d 210, 215 (D.C. Cir. 2001); see also Johansson v. United States, 336 F.2d 809, 813 (5th Cir. 1964); Gaw v. Comm’r, 70 T.C.M. (CCH) 1196 (1995). In the 1990s, taxpayers engaged in “dividends stripping” transactions that were asserted to have no pretax profit but generated after-tax profit through foreign tax credits. In two well known opinions, the Fifth and Eighth Circuits reversed trial courts’ summary judgments for the United States, holding that the transactions did not lack business purpose or economic substance. Notably, however, the courts did not rely upon the argument that the transactions involved treaties that, by their plain language, required the United States to allow the foreign tax credit, nor did the taxpayers assert such arguments. Compaq Comput. Corp. v. Comm’r, 277 F.3d 778, 788 (5th Cir. 2001); IES Indus., Inc. v. United States, 253 F.3d 350, 356 (8th Cir. 2001).

  215. [215]. See 26 U.S.C. §7701(o) (2012); Erik M. Jensen, Legislative and Regulatory Responses to Tax Avoidance: Explicating and Evaluating the Alternatives, 57 St. Louis U. L.J. 1, 30–31 (2012) (“[F]or judges unsure about their authority to apply extra-statutory principles, codification of economic substance makes the extra-statutory statutory, and, for those judges unsympathetic to government arguments, codification obligates them to take economic substance seriously.”); Bret Wells, Economic Substance Doctrine: How Codification Changes Decided Cases, Fla. Tax Rev., 2010, at 411, 418 (“[I]t can be inferred that Congress believed that the codification of the economic substance doctrine would further enhance the successful application of this doctrine.”).

  216. [216]. Stanley S. Surrey, Complexity and the Internal Revenue Code: The Problem of the Management of Tax Detail, 34 Law & Contemp. Probs. 673, 707 n.31 (1969); see also David A. Weisbach, Formalism in the Tax Law, 66 U. Chi. L. Rev. 860, 862 (1999) (arguing that anti-abuse rules preserve simplicity in the Code).

  217. [217]. Specific rules are often formalistic. Their objective criteria can be easily evaded by taxpayers. See Anna A. Kornikova, Comment, Solving the Problem of Tax-Treaty Shopping Through the Use of Limitation on Benefits Provisions, 8 Rich. J. Global L. & Bus. 249, 264–65 (2008) (discussing bright-line anti-abuse rules as being easily circumvented by taxpayers in the context of treaty shopping). Moreover, addressing individual loopholes often results in taxpayers simply finding another. See, e.g., Marvin A. Chirelstein & Lawrence A. Zelenak, Tax Shelters and the Search for a Silver Bullet, 105 Colum. L. Rev. 1939, 1950 (2005) (“It is always possible, of course, to shut down particular shelter techniques with narrowly-targeted legislation.... The problem is that these targeted fixes are always made prospective only. As Congress closes one loophole, tax shelter designers find other glitches in the Code around which to build new shelters.”).

  218. [218]. See generally Noel B. Cunningham & James R. Repetti, Textualism and Tax Shelters, 24 Va. Tax Rev. 1 (2004). Larry Solan and Steven Dean argue that neither textualism nor intentionalism provides the means to combat abusive tax shelters since such shelters either fall within statutory language or take advantage of gaps in statutory language where legislative intent cannot be discerned. Steven A. Dean & Lawrence M. Solan, Tax Shelters and the Code: Navigating Between Text and Intent, 26 Va. Tax Rev. 879, 880, 889 (2007). Instead, they argue, anti-abuse doctrines, such as the business purpose doctrine, are effective because they allow judges to be attentive towards legislative purpose while still considering statutory language. Id. at 903–04.

  219. [219]. Lawrence Zelenak, The Court and the Code: A Response to The Warp and Woof of Statutory Interpretation, 58 Duke L.J. 1783, 1788–89 (2009).

  220. [220]. Cunningham & Repetti, supra note 218, at 27; see also Brian J. Arnold, Reflections on the Relationship Between Statutory Interpretation and Tax Avoidance, 49 Canadian Tax J. 1, 2, 13–19 (2001) (suggesting that a literalist approach by Canadian courts has led the Supreme Court of Canada to uphold “several blatant tax avoidance schemes”). Although the 1982 Ramsay case seemed to signal the development of a judicial anti-abuse doctrine in the U.K., it did not bear fruit. Judith Freedman, Interpreting Tax Statutes: Tax Avoidance and the Intention of Parliament, 123 Law Q. Rev. 53, 58 (2007); see also Roy Rohatgi, Basic International Taxation 31 (2002) (describing the domestic approach to statutory interpretation as literal in many countries). It could, however, be argued that the Article’s prescription to look to extrinsic sources in interpreting tax treaties may actually abet tax abuse since such sources may include textualist judicial opinions. Such courts adhering to textualist precedent, however, would most likely be hostile to my approach anyway. I therefore do not think that the prescribed methodology would exacerbate the problem unless it led to the spread of textualism to other interpretive bodies. Because I recommend the incorporation of a multitude of extrinsic source, the chances of such an interpreter focusing solely on a textualist source are low. If anything, my prescribed methodology should encourage courts and other bodies reluctant to consider extrinsic sources on the basis of international norms or law to reconsider their textualist leanings.

  221. [221]. See generally Comm. of Experts on Int’l Cooperation in Tax Matters, Treaty Abuse and Treaty Shopping, U.N. Doc. E/C.18/2006/2 (Oct. 30–Nov. 3, 2006).

  222. [222]. See Vogel et al., supra note 9, para. 121, at 56 (concluding that tax treaties “are subject to a general ‘substance v. form proviso’ based on international law”).

  223. [223]. Org. for Econ. Co-operation & Dev., supra note 201, at 54–55 (adopting a test that disregards treaty benefits where one of the principal purposes of the transaction was to obtain such benefits).

  224. [224]. Ward, supra note 9, at 403.

  225. [225]. Model Convention with Respect to Taxes on Income and on Capital art. 3(2) (Org. for Econ. Co-operation & Dev. 2014).

  226. [226]. U.S. Model Income Tax Convention of November 15, 2006 art. 3(2) (U.S. Dep’t of Treasury 2006).

  227. [227]. Avery Jones et al., supra note 9, at 25–48.

  228. [228]. See Vogel, supra note 9, at 7; Klaus Vogel & Rainer G. Prokisch, General Report, in LXXVIIIa Cahiers de Droit Fiscal International: Interpretation of Double Taxation Conventions 55, 77–79 (Int’l Fiscal Ass’n ed., 1993).

  229. [229]. Vogel, supra note 9, at 74.

  230. [230]. Id. at 63.

  231. [231]. See Smith, supra note 8, at 863.

  232. [232]. See, e.g., U.S. Model Income Tax Convention of November 15, 2006 art. 3(2) (U.S. Dep’t of Treasury 2006) (“[A]ny term not defined therein shall, unless the context otherwise requires, or the competent authorities agree to a common meaning pursuant to the provisions of Article 25 (Mutual Agreement Procedure), have the meaning which it has at that time under the law of that State for the purposes of the taxes to which the Convention applies.”).

  233. [233]. See Avery Jones et al., supra note 9, at 105 (noting that a number of courts and scholars have taken this position).

  234. [234]. Vogel et al., supra note 9, at 139–42; see also Avery Jones et al., supra note 9, at 108 (“The context must therefore be reasonably strong for the internal law meaning to be ousted.”).

  235. [235]. See Smith, supra note 8, at 882.

  236. [236]. Samann v. Comm’r, 313 F.2d. 461 (4th Cir. 1963).

  237. [237]. Id. at 464.

  238. [238]. Id. at 463.

  239. [239]. U.S. Model Income Tax Convention of November 15, 2006 art. 4(1), 6(2) (U.S. Dep’t of Treasury 2006).

  240. [240]. These concepts are illustrated in Revenue Ruling 56-446, which characterized the distribution from a pension fund for purposes of an older U.S.–Canada tax treaty that, as most treaties do, exempted gains from the sale or exchange of capital assets. Under domestic law at the time of the treaty’s effective date, such proceeds were ordinary income; however, Congress later changed the treatment to be capital gains. The ruling followed current law, characterizing the proceeds as capital gains and hence granted exemption under the treaty. Rev. Rul. 56-446, 1956-2 C.B. 1065, modified by Rev. Rul. 58-247, 1958-1 C.B. 623.

  241. [241]. See, e.g., United States v. A.L. Burbank & Co., 525 F.2d 9, 15–17 (2d Cir. 1975); Crow v. Comm’r, 85 T.C. 376, 395 (1985); Estate of Burghardt v. Comm’r, 80 T.C. 705, 717 (1983), aff’d sub nom., Estate of Burghardt v. Comm’r, 734 F.2d 3 (3d Cir. 1984); Rev. Rul. 80-243, 1980-2 C.B. 413; Rev. Rul. 56-446, 1956-2 C.B. 1065; I.R.S. Priv. Ltr. Rul. 85-26-005 (Mar. 8, 1985); I.R.S. Gen. Couns. Mem. 39,373 (June 24, 1985); I.R.S. Gen. Couns. Mem. 37,978 (June 15, 1979).

  242. [242]. Commentaries on the Articles of the Model Tax Convention art. 3(2) cmt. 13, 13.1 (Org. for Econ. Co-operation & Dev. 2010),
    43324465.pdf. The OECD’s embrace of ambulatory interpretation was heavily influenced by Avery Jones. See generally Avery Jones et al., supra note 9.

  243. [243]. Avery Jones, supra note 196, at 4 (“Treaties cannot be changed because there are so many of them and so the OECD tends to rewrite the Commentary instead of the Model Treaty, leading to problems that I consider later. And treaties inhibit changes in internal law, which would have no effect in cases where treaties exist, that is to say, in most of the cases that matter. This leads to either the changes not happening or to treaty override.”).

  244. [244]. Smith, supra note 8, at 883.

  245. [245]. Arato, supra note 2, at 218–19.

  246. [246]. See id.

  247. [247]. See supra notes 208–09 and accompanying text.

  248. [248]. Select Draft Provisions of the U.S. Model Income Tax Convention: New Article 3 Paragraph 1(1) Definition of “Special Tax Regime” (U.S. Dep’t of Treasury 2015), https://

  249. [249]. Select Draft Provisions of the U.S. Model Income Tax Convention: New Article 28 (Subsequent Changes in Law) (U.S. Dep’t of Treasury 2015),

  250. [250]. Allison Christians, Seeing STRs: A New Vision for Tax Treaties?, 41 Brooklyn J. Int’l L. (forthcoming 2016) (on file with author).

  251. [251]. Id.

  252. [252]. See supra notes 49–53 and accompanying text.

  253. [253]. Medellín v. Texas, 552 U.S. 491, 499–500 (2008).

  254. [254]. See supra notes 117–29 and accompanying text.

  255. [255]. See William W. Chip, Interpreting Tax Treaties After NatWest, 37 Tax Mgmt. Int’l J. 321, 330 (2008).

  256. [256]. Smith, supra note 8, at 888–89 n.206.

  257. [257]. Kirsch, supra note 8, at 1095–97.

  258. [258]. Mayo Found. for Med. Educ. & Research v. United States, 562 U.S. 44, 53–58 (2011).

  259. [259]. Id. at 55–58.

  260. [260]. See Bradley George Hubbard, Comment, Deference to Agency Statutory Interpretations First Advanced in Litigation? The Chevron Two-Step and the Skidmore Shuffle, 80 U. Chi. L. Rev. 447, 448 (2013).

  261. [261]. Id. at 471–72.

  262. [262]. See id.

  263. [263]. See Leandra Lederman, The Fight over “Fighting Regs” and Judicial Deference in Tax Litigation, 92 B.U. L. Rev. 643, 644, 648 (2012) (proposing that the court take into account retroactivity and surprise issues in analyzing the effects of regulations and revenue rulings issued in the course of litigation).

  264. [264]. See 134 Cong. Rec. 12,844–49 (1988).

  265. [265]. The ABM Treaty and the Constitution: Joint Hearings Before the S. Comm. on Foreign Relations & the S. Comm. on the Judiciary, 100th Cong. 177 (1987) (statement of Sen. Joseph Biden, Jr., Chairman, S. Comm. on the Judiciary). Scholarly articles on this debate are numerous. See, e.g., Michael J. Glennon, Essay, Interpreting “Interpretation”: The President, the Senate, and When Treaty Interpretation Becomes Treaty Making, 20 U.C. Davis L. Rev. 913 (1987); Malvina Halberstam, A Treaty Is a Treaty Is a Treaty, 33 Va. J. Int’l L. 51 (1992); David A. Koplow, Constitutional Bait and Switch: Executive Reinterpretation of Arms Control Treaties, 137 U. Pa. L. Rev. 1353 (1989); Eugene V. Rostow, The Reinterpretation Debate and Constitutional Law, 137 U. Pa. L. Rev. 1451 (1989); Gary Michael Buechler, Note, Constitutional Limits on the President’s Power to Interpret Treaties: The Sofaer Doctrine, the Biden Condition, and the Doctrine of Binding Authoritative Representations, 78 Geo. L.J. 1983 (1990); and W. Michael Reisman, Comment, Necessary and Proper: Executive Competence to Interpret Treaties, 15 Yale J. Int’l L. 316 (1990).

  266. [266]. Rainbow Navigation, Inc. v. Dep’t of the Navy, 699 F. Supp. 339, 343 (D.D.C. 1988) (quoting the Government’s position in litigation briefs), rev’d, 911 F.2d 797 (D.C. Cir. 1990).

  267. [267]. See, e.g., United States v. Stuart, 489 U.S. 353, 369 (1989) (noting that executive branch materials are entitled to “great weight”); Sumitomo Shoji Am., Inc. v. Avagliano, 457 U.S. 176, 184–85 (1982) (same); see also Restatement (Third) of Foreign Relations Law of the U.S. §326(2) (Am. Law Inst. 1986) (“Courts... will give great weight to an interpretation made by the Executive Branch.”).

  268. [268]. A difficult question arises as to reliance upon the commentaries by developing nations. As Lee Sheppard recently has noted, such commentaries disadvantage source countries. See generally Sheppard, supra note 172. Care should be exercised in interpreting treaties with developing nations, paying particular attention to such nations’ reservations to the commentaries. Developing countries may also, in interpreting the treaties, wish to consult the relatively developed interpretations by large developing countries such as India. See Eduardo Baistrocchi, The Use and Interpretation of Tax Treaties in the Emerging World: Theory and Implications, 4 Brit. Tax Rev. 352, 390 (2008) (arguing that the standards-based approach of OECD Model Tax Treaties allows developing nation’s considerable flexibility in interpreting treaties that they will use to generally benefit the taxpayer due to the necessity of competing with other developing nations).

  269. [269]. Vogel, supra note 9, at 40.

  270. [270]. For this reason, considerably less deference to the commentaries is warranted in tax treaties with non-OECD countries. Id. at 42.

  271. [271]. OECD Model Tax Convention on Income and on Capital, ¶¶ 1, 3, 10–11, at 7, 9 (Org. for Econ. Co-operation & Dev. 2010).

  272. [272]. Richard Vann, Interpretation of Tax Treaties in New Holland 7 (Sydney Law Sch. Legal Studies Research, Paper No. 10/21, 2010).

  273. [273]. See, e.g., Taisei Fire & Marine Ins. Co. v. Comm’r, 104 T.C. 535, 550 (1995); see also Itai Grinberg, Breaking BEPS: The New International Tax Diplomacy, Geo. L.J. (forthcoming 2016) (manuscript at 38) (on file with author).

  274. [274]. Grinberg, supra note 273 (manuscript at 42 n.194) (citation omitted).

  275. [275]. See, e.g., Pekar v. Comm’r, 113 T.C. 158, 162–64 (1999).

  276. [276]. But see Townsend, supra note 8, at 297–98 (noting that the court failed to cite U.K. cases interpreting language from parallel treaties that were put forth by amici).

  277. [277]. See Julian Arato, Subsequent Practice and Evolutive Interpretation: Techniques of Treaty Interpretation over Time and Their Diverse Consequences, 9 Law & Prac. Int’l Cts. & Tribunals 443, 459–60 (2010) (discussing “subsequent practice” as encompassing acts by all of the treaty parties or acquiescence by the non-engaging party). For instance, in Stuart, the Court observed that the IRS had complied with information requests from Canadian authorities without regard to the stage of the criminal proceeding. United States v. Stuart, 489 U.S. 353, 369 (1989).

  278. [278]. Bederman, supra note 7, at 994.

  279. [279]. Vienna Convention, supra note 1, art. 32. Some have argued that supplementary materials exclude unilateral materials. See Bederman, supra note 7, at 973; Vogel & Prokisch, supra note 228, at 74. These arguments are based on the fact that article 32 mentions only negotiating materials and the circumstances of the treaty’s conclusion. The plain language of article 32, however, uses these examples as illustrative rather than exhaustive: “Recourse may be had to supplementary means of interpretation, including the preparatory work of the treaty and the circumstances of its conclusion....” Vienna Convention, supra note 1, art. 32.

  280. [280]. Myres S. McDougal et al., The Interpretation of International Agreements and World Public Order 40­–41 (1994).

  281. [281]. See Rohatgi, supra note 220, at 31.

  282. [282]. Philip Baker, Double Taxation Conventions and International Tax Law 42–43 (2d ed. 1994).

  283. [283]. Id. at 28–29.

  284. [284]. See, e.g., id. at 27–28 (recounting a trio of cases in the United States, Germany, and England expounding similar treaty provisions essentially upon intentionalist, purposivist, and literalist grounds respectively).

  285. [285]. In the domestic statutory interpretation context, there is also little consensus over the meaning of ambiguity. See United States v. Hansen, 772 F.2d 940, 948 (D.C. Cir. 1985) (containing any inquiry by then Circuit Court judge Scalia as to “how much ambiguousness constitutes an ambiguity”). Further, what constitutes an absurd result such that departure from the plain meaning is justified? See John F. Manning, The Absurdity Doctrine, 116 Harv. L. Rev. 2387, 2392 (2003) (discussing the vagaries in applying the absurdity doctrine).

  286. [286]. Boulez v. Commissioner, discussed earlier, is the classic counterexample where both the residence and source countries are interpreting the same term in the treaty. Boulez v. Comm’r, 83 T.C. 584 (1984). There is to my knowledge, however, only one such case in the American case law, indicating the rarity of this phenomenon.

  287. [287]. Elisabeth A. Owens, United States Income Tax Treaties: Their Role in Relieving Double Taxation, 17 Rutgers L. Rev. 428, 430 (1963).

  288. [288]. See generally Dagan, supra note 194; see also Joseph Isenbergh, International Taxation: U.S. Taxation of Foreign Persons and Foreign Income ¶ 55.2 (2000) (“[I]ncome tax treaties can easily be taken at first inspection as measures designed to confer tax relief on certain individuals or enterprises. In fact, that is rarely their function. Tax treaties are principally concerned with the apportionment of tax revenues between the treasuries of the treaty countries.”); Paul R. McDaniel et al., Introduction to United States International Taxation 178 (5th rev. ed. 2005) (remarking on the unilateral means nations employ to alleviate double taxation, using treaties to conform these measures “to the specifics of the tax relationships between the two countries involved”); Julie Roin, Rethinking Tax Treaties in a Strategic World with Disparate Tax Systems, 81 Va. L. Rev. 1753, 1763 (1995) (arguing that the domestic means for avoiding double taxation has eliminated the benefits treaties provide to taxpayers, and instead such treaties “intend[] to effect a roughly neutral exchange of tax revenues between the source and residence countries”).

  289. [289]. Because tax treaties contain ambiguities, textualism does not necessarily provide a plain meaning that would be known to all. See Smith, supra note 8, at 867.

  290. [290]. Chip, supra note 255, at 331.

  291. [291]. See supra notes 286–89 and accompanying text.

  292. [292]. Eskridge & Frickey, supra note 10, at 353–54; see also William N. Eskridge, Jr., Dynamic Statutory Interpretation, 135 U. Pa. L. Rev. 1479, 1496–97 (1987) (prescribing an interpretive “continuum” ranging from statutory text to “evolutive” context).

  293. [293]. Eskridge & Frickey, supra note 10, at 354.

  294. [294]. Id.

  295. [295]. Id. at 360.

  296. [296]. This schematic is adapted from Eskridge’s and Frickey’s funnel of abstraction. Id. at 353.

  297. [297]. For instance, Treasury conferred with Canada in drafting the technical explanation to the 2007 Protocol to the U.S.–Canada tax treaty. See Technical Explanation of the Protocol Done at Chelsea on September 21, 2007 Amending the Convention Between the United States of America and Canada with Respect to Taxes on Income and on Capital Done at Washington on September 26, 1980, as Amended by the Protocols Done on June 14, 1983, March 28, 1994, March 17, 1995, and July 29, 1997, at 1 (U.S. Dep’t of Treasury 2007),

  298. [298]. Restatement (Third) of Foreign Relations Law of the United States §314 cmt. b (Am. Law Inst. 1986).

  299. [299]. See supra notes 240–45 and accompanying text.

  300. [300]. See, e.g., Eskridge & Frickey, supra note 10, at 332–39 (criticizing purposivism); Cass R. Sunstein,Interpreting Statutes in the Regulatory State, 103 Harv. L. Rev. 405, 426–28 (1989) (same).

  301. [301]. Specifically, under public choice theory, legislators dole out benefits to rent-seeking interest groups, and the byproduct of these many compromises does not contain one public purpose, but is instead an aggregation of various purposes. See, e.g., Eskridge & Frickey, supra note 10, at 335. Similar observations can be made in the treaty process, which is not impervious to capture by interest groups. Indeed, a cynical view of tax treaties is that they “quietly bestow benefits on multinationals that would not be obtainable in the regular political process.” Sheppard, supra note 170, at 828 (noting comments of Stuart Chessman).

  302. [302]. See supra note 21 and accompanying text.

  303. [303]. See Smith, supra note 8, at 858–63. My conclusions here are thus similar to those of Michael Livingston’s in the domestic statutory context. Livingston rejects the claims of some scholars that the underlying logic and structure of the Code justifies a purposivist approach. Michael Livingston, Practical Reason, “Purposivism,” and the Interpretation of Tax Statutes, 51 Tax L. Rev. 677, 679 (1996). Accordingly, we both prescribe a practical reasoning approach to interpretation rather than purposivism. Id. at 720–24.

  304. [304]. See generally Eskridge & Frickey, supra note 10.

  305. [305]. See supra notes 159–62 and accompanying text.

  306. [306]. See supra notes 254–66 and accompanying text.

  307. [307]. For instance, the NatWest court might have considered whether, from a tax policy perspective, intracorporate lending transactions should be disregarded. See Reinhold & Harrington, supra note 178, at 195–96.

  308. [308]. See William N. Eskridge, Jr., et al., Cases and Materials on Legislation: Statutes and the Creation of Public Policy 884 (4thed. 2007).

  309. [309]. Vogel et al., supra note 152, at 29.

  310. [310]. Id.

  311. [311]. See, e.g., Canadian Pac. Ltd. v. The Queen, [1976] 30 D.T.C. 6120, 6135 (F.C.T.D.).

  312. [312]. See, e.g., Donroy, Ltd. v. United States, 301 F.2d 200, 207–08 (9th Cir. 1962).

  313. [313]. See, e.g., Corocraft Ltd. v. Pan Am. Airways Inc., [1969] 1 QB 616.

  314. [314]. See, e.g., United States v. Stuart, 489 U.S. 353, 368 (1989); Bacardi Corp. of Am. v. Domenech, 311 U.S. 150, 163 (1940); United States v. A.L. Burbank & Co., 525 F.2d 9, 14 (2d Cir. 1975); N.W. Life Assurance Co. of Can. v. Comm’r, 107 T.C. 363, 378 (1996).

  315. [315]. Geofroy v. Riggs, 133 U.S. 258, 271–72 (1890).

  316. [316]. This is the most common view of the canon, although newer iterations of the canon suggest that liberal construction simply instructs courts to interpret provisions using various interpretive aids. See also Van Alstine, The Death of Good Faith, supra note 7, at 1916–19. There is also case law that emphasizes protection of the rights of the sovereign through narrow treaty construction. See Kreimerman v. Casa Veerkamp, S.A. de C.V., 22 F.3d 634, 639 (5th Cir. 1994) (“[C]ourts should interpret treaty provisions narrowly—for fear of waiving sovereign rights that the government or people of the State never intended to cede.”); Stuart S. Malawer, Maryland and Supreme Court Treaty Interpretation: Paradox and Dilemma, 2 U. Balt. L. Rev. 35, 46 (1972).

  317. [317]. See Van Alstine, The Death of Good Faith, supra note 7, at 1915–16.

  318. [318]. See Vienna Convention, supra note 1, art. 26 (“Every treaty... must be performed by [the parties] in good faith.”); Bederman, supra note 7, at 968–69 (discussing the tension between good faith and liberal interpretation).

  319. [319]. Bederman, supra note 7, at 969–70.

  320. [320]. See Edward D. Kleinbard, Stateless Income, 11 Fla. Tax Rev. 699, 701–06 (2011).

  321. [321]. See Rogoff, supra note 7, at 616–46 (discussing the restrictive approach to treaty interpretation based on sovereignty concerns and its subsequent fall from relevance).

  322. [322]. See supra notes 186–93 and accompanying text.

  323. [323]. See Snap-On Tools, Inc. v. United States, 26 Cl. Ct. 1045, 1072 (1992) (“[A]n understanding of a position which forms the basis for a negotiated international agreement cannot be arrived at ‘tacitly,’ but must be achieved consciously and deliberately by both parties.”); see also Townsend, supra note 8, at 286 (concluding that implicit treaty benefits are inappropriate in the tax context essentially for sovereignty-related reasons).

  324. [324]. Osgood, supra note 8, at 294; see also Baker, supra note 283, at 21 (“[D]ouble taxation agreements have a purpose substantially differing from that of normal political or economic treaties because they are intended to reconcile two national fiscal legislations and to avoid the simultaneous taxation in both countries.” (citation omitted)); supra notes 288–89 and accompanying text.

  325. [325]. See Avery Jones et al., supra note 9, at 48; Osgood, supra note 8, at 297.

  326. [326]. Osgood, supra note 8, at 297. The ALI Federal Income Tax Project supports this view. Federal Income Tax Project: International Aspects of United States Income Taxation II, at 62 (Am. Law Inst. 1992) (prescribing tiebreaker rule in favor of the source country).

  327. [327]. Vogel, supra note 9, at 65. Specifically, Vogel points to article 21 of the OECD Model Treaty, in which items of income not dealt with specifically by the treaty are taxable only by the state of residence. Id. at n.343.

  328. [328]. Edward D. Kleinbard, Why Corporate Tax Reform Can Happen, Tax Notes, Apr. 6, 2015, at 91, 95.

  329. [329]. See, e.g., Dagan, supra note 194.

  330. [330]. See Graetz, supra note 192, at 298.

  331. [331]. Id.

  332. [332]. Id. at 297.

  333. [333]. See Louis Henkin, Treaties in a Constitutional Democracy, 10 Mich. J. Int’l L. 406, 425–26 (1989); Jules Lobel, The Limits of Constitutional Power: Conflicts Between Foreign Policy and International Law, 71 Va. L. Rev. 1071, 1096 (1985).

  334. [334]. See Kysar, supra note 28, at 7.

  335. [335]. For instance, prior to the advent of the modern tax treaty, the House used to regularly defend its right to participate in treaties involving tax revenues. Id. at 45. In fact, trade treaties take the form of congressional–executive agreements and, as such, are approved only with House participation. Scholars have speculated that the Origination Clause is one driver behind this development. Bruce Ackerman & David Golove, Is NAFTA Constitutional?, 108 Harv. L. Rev. 799, 923 (1995); see also Halpern, supra note 33, at 3–4 (suggesting that the last in time rule functions to alleviate the tension created by the incongruity between the Treaty Clause and the Origination Clause). An interesting question arises in the tax context because, in codifying the last in time rule with respect to tax treaties, the Senate indicated that the rule should apply without regard to whether Congress expressed an intent to override the treaty. S. Rep. No. 100-445, at 326 (1988). Since the rule is a constitutional interpretation of the Supremacy Clause, it is unclear whether the Court would give effect to the Senate’s interpretation in the tax treaty context. See Anthony Infanti, Domestic Law and Tax Treaties: The United States, Tax Network (Feb. 2, 2012),

  336. [336]. See supra notes 32–33 and accompanying text.

  337. [337]. If the later enacted, conflicting statutory language is clear, then it overrides the treaty. Roeder v. Islamic Republic of Iran, 195 F. Supp. 2d 140, 170 (D.D.C. 2002). If, however, such language is ambiguous, courts inquire into Congress’ intent with respect to the override before giving priority to the statute. Trans World Airlines, Inc. v. Franklin Mint Corp., 466 U.S. 243, 252 (1984) (holding that legislative silence was insufficient in establishing a treaty override).

  338. [338]. Murray v. Schooner Charming Betsy, 6 U.S. (2 Cranch) 64 (1804).

  339. [339]. Curtis A. Bradley, The Charming Betsy Canon and the Separation of Powers: Rethinking the Interpretive Role of International Law, 86 Geo. L.J. 479, 482 (1998).

  340. [340]. Traditionally, U.S. courts have extended the Charming Betsy canon to treaties. See, e.g., Santovincenzo v. Egan, 284 U.S. 30, 40 (1931) (stating that treaties are to be construed in accordance with international law); United States v. Rauscher, 119 U.S. 407, 419–20, 429 (1886) (interpreting a treaty as to not conflict with preexisting obligations under international law). Recent cases, however, suggest that this presumption is softening. Criddle, supra note 7, at 456 (citing United States v. Alvarez-Machain, 504 U.S. 655 (1992), and United States v. Lui Kin-Hong, 110 F.3d 103 (1st Cir. 1997), as examples of cases where the court rejects a treaty interpretation that would comport with international law).

  341. [341]. In amending sections 894(a) and 7852(d) to codify the last in time rule, legislative history indicates that Congress did not intend to override the presumption of harmony between treaties and statutes. S. Rep. No. 100-445, at 321 (1988).


Professor of Law, Brooklyn Law School. 

I am grateful to Hugh Ault, Anita Bernstein, Dana Brakman-Reiser, Perry Dane, Steve Dean, Robin Effron, Katie Eyer, George Geis, Dan Halperin, Edward Janger, Louis Kaplow, Anita Krishnakumar, Steve Landsmann, Michael Livingston, Ruth Mason, Thomas McDonnell, Greg Mitchell, Saikrishna Prakash, Diane Ring, Rand Rosenblatt, David Reiss, Fred Schauer, Rich Schragger, Kate Shaw, Steve Shay, Larry Solan, Bob Williams, and workshop participants at Harvard Law School, Pace Law School, Rutgers School of Law-Camden, and University of Virginia School of Law for valuable comments and discussion on earlier drafts. I am also thankful to Andrew Granek, Jana Hymowitz, Bob Nelson-Sullivan, and Grace Sur for excellent research assistance. Finally, I gratefully acknowledge the Dean’s Summer Research Stipend Program at Brooklyn Law School for funding portions of my research.