Finding the Purpose of Tax Treaty Provisions Under GAAR: Lessons From Alta Energy

Jinyan Li

Jinyan Li is a professor of law and co-director of the LLM tax program at Osgoode Hall Law School, York University, in Canada. The author thanks Marshall Rothstein, former justice of the Supreme Court of Canada, for a conversation about the Alta Energy case, and Scott Wilkie for comments on an earlier draft of this article. She is extremely grateful to Allan Lanthier for his meticulous review of the drafts and sharing his insights. The views expressed are the author’s alone.

In this article, Li considers the implications of the Supreme Court of Canada’s Alta Energy judgment for determining the purpose of tax treaty provisions under general antiabuse rules and suggests that a treaty’s text and context should carry more weight than inferred negotiator intentions.

Establishing the object and purpose of tax treaty provisions lies at the heart of applying antiabuse rules, such as general antiavoidance rules under domestic law¹ and the principal purpose test (PPT) in tax treaties.² Tax planning arrangements like treaty shopping, designed to obtain treaty benefits, are not abusive unless they contravene the object and purpose of the provisions relied upon by the taxpayer.

Because the purpose of the applicable provision is rarely stated in the text, one must find it by looking beyond the plain words of the provision. But how much beyond? Given that a tax treaty is an agreement between countries, to what extent may courts in one country impute intention of treaty negotiators to both countries? Is the “intention” of treaty negotiators the same as the “purpose” of the treaty? What are the appropriate ways of finding purpose?

The 6-3 split decision of the Supreme Court of Canada in Alta Energy³ offers insights into these questions. The 189-paragraph judgment covers many aspects of the GAAR and treaty interpretation.⁴ This article focuses on aspects of the judgment related to the finding of “object, spirit and purpose” of the provision in dispute — the carveout in article 13(4) of the Canada-Luxembourg tax treaty.⁵

As the first decision by the highest court of any country on the intersection of a domestic GAAR and “treaty shopping” involving article 13 of the OECD model tax convention,⁶ Alta Energy may have significant implications for the application of the PPT in existing tax treaties and a future multilateral tax convention to implement pillar 1 on the taxation of large multinational corporations’ residual profit.⁷

Following an overview of the purpose test in Section I, this article discusses the Alta Energy case in terms of the facts, the main issue, and the split decision. It shows how the majority and dissent reach their decisions and highlights some takeaways. Section III argues that the Court adopted a problematic approach to finding the purpose of the carveout by conflating the treaty purpose and negotiator intention and not giving sufficient weight to the language and structure of the treaty. It explains that the treaty provisions related to the carveout do not necessarily support the conclusion that the carveout is a deliberate tax incentive and offer stronger support for the dissent’s view that “residence” requires a “genuine economic connection.” The article concludes with observations about the implications of this case for the Canadian GAAR and the PPT in tax treaties.

I. The Importance of ‘Purpose’

The notion of purpose is the “black box at the centre of most GAAR rules and GAAR disputes.”⁸ The same can be said of the PPT. It is challenging to apply the purpose test in domestic tax avoidance cases as well as treaty-shopping cases.

A. The Purpose Test

The purpose test under the GAAR and PPT has two aspects. The first is the determination of whether the primary purpose of the transaction is to obtain a tax benefit, which is the precondition for applying the GAAR or PPT. The second is the purpose of the provisions relied upon by the taxpayer in obtaining the tax benefit. This article considers only the latter — legislative purpose.

1. GAAR

The Canadian GAAR states that if a transaction is an avoidance transaction (a transaction entered into for the primary purpose of obtaining a tax benefit), the benefit will be denied, but only if the transaction is abusive. Canadian courts have interpreted abuse to involve a two-step inquiry.⁹ First, what is the object, spirit, and purpose of the provisions relied on for the tax benefit? This is established by conducting a “textual, contextual and purposive” construction of the provisions. Second, does the avoidance transaction frustrate the object, spirit, and purpose of the provisions? Because “spirit” is an elusive concept, the “object, spirit, and purpose” test is often reduced to a “purpose” or “rationale” test. As such, finding the legislative purpose of the relevant provision is the key.

By virtue of section 4.1 of the Income Tax Conventions Interpretation Act (ITCIA),¹⁰ the GAAR applies to Canadian tax treaties.

2. PPT

The PPT in article 7(1) of the Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting (MLI) is included in many Canadian tax treaties.¹¹ It states:

Notwithstanding any provisions of a Covered Tax Agreement, a benefit under the Covered Tax Agreement shall not be granted in respect of an item of income or capital if it is reasonable to conclude, having regard to all relevant facts and circumstances, that obtaining that benefit was one of the principal purposes of any arrangement or transaction that resulted directly or indirectly in that benefit, unless it is established that granting that benefit in these circumstances would be in accordance with the object and purpose of the relevant provisions of the Covered Tax Agreement. [Emphasis added.]

Unlike a limitation on benefits provision, which looks through legal constructs to determine residential connections by reference to specific tests, such as ownership and active trade or business tests, the PPT is a purpose-based rule, like the GAAR.¹²

B. Challenges in Finding Legislative Purpose

1. Domestic Law

Finding legislative purpose is inherently challenging because the text of the controversial provision is open to at least two possible interpretations. One, often textual, is relied upon by taxpayers in designing the avoidance transaction. The other is often advanced by the Minister of National Revenue (MNR) and goes beyond the text into the realm of general principles, policies, or even theories. Judges must find the “right” purpose as a matter of statutory interpretation to resolve the dispute at hand.

There is a tendency to turn a GAAR case into a reference case on the general meaning of fundamental concepts, such as “cost”¹³ or “capital,”¹⁴ which borders on legislative or policymaking, domains into which judges are generally hesitant to venture. As such, Canadian courts require the government to bear the burden of establishing the existence of abuse such that “it cannot be reasonably concluded that a tax benefit would be consistent with the object, spirit or purpose of the provisions relied upon by the taxpayer.” Otherwise, the benefit of the doubt goes to the taxpayer.¹⁵

Legislative purpose is difficult to establish because the government must use tangible evidence (such as the statute read as a whole,¹⁶ legislative scheme, or extrinsic documents) to reveal the intangible purpose. Having the Income Tax Act read as a whole is a fiction because it is unreasonable to expect anyone to meaningfully read the entire statute: It weighs over one kilogram in print and embodies the fiscal choices made over a century. It takes a great deal of effort and capacity on the part of lawyers to distill the materials into coherent legal analysis and arguments for judges to consider. It is therefore unclear what contextual analysis entails in a given case.

Judicial attitude toward the GAAR may also play a role.¹⁷ The GAAR jurisprudence, including the Supreme Court of Canada’s decision in Lipson,¹⁸ suggests that judges who favor tax certainty and taxpayers’ right to tax minimization or the Duke of Westminster principle¹⁹ tend to find a legislative purpose that confers the tax benefit sought by the taxpayer.

C. Treaty Interpretation

Finding the object and purpose of treaty provisions is a different exercise and can be challenging for several reasons. First, as an agreement between two countries, a treaty reflects the bargain of the parties and is an instrument of public international law. Treaty interpretation is guided by article 31 of the Vienna Convention on the Law of Treaties (VCLT)²⁰:

A treaty shall be interpreted in good faith 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.

A tax treaty differs from a domestic tax statute in terms of its objectives and nature. One of the main purposes of a tax treaty is to prevent double taxation, whereas the main objective of domestic taxation is raising revenue. To prevent double taxation, a treaty assigns taxing rights to the two countries by reference to the residence of taxpayers and source of income. In effect, a treaty limits the application of domestic law and reduces a contracting state’s tax revenue. From a taxpayer perspective, a treaty is relieving as it can reduce (but never increase) the tax liability that otherwise exists under domestic law. However, unlike tax incentive provisions in domestic law, treaty provisions confer tax benefits to taxpayers in order to enable the two tax systems to intersect without creating double taxation, as opposed to explicitly encouraging specific taxpayer behaviors. The approach to interpreting domestic tax incentive provisions²¹ may therefore be inappropriate to treaty interpretation.

A tax treaty also differs from domestic law in drafting style. Most Canadian treaties are based on the OECD model, and the provisions are drafted in more general language. In contrast, the ITA is uniquely Canadian as it reflects the fiscal choices made by Canadians. It is drafted in highly technical, detailed, and complex language, replete with exceptions, tax expenditures, and specific antiavoidance rules. It is amended frequently, sometimes many times a year. A tax treaty is seldom amended. Treaties should not be expected to be as technically detailed as domestic law, but judges may not be alert to this point.

Determining the purpose of treaty provisions involves different considerations, such as the contractual nature of treaties, the drafting style and relieving effect of treaties, the existence of the OECD model and commentaries that have no counterparts in domestic law, and the treaty’s reliance on domestic law to take effect. How to incorporate these considerations in applying the purpose test to treaty-shopping arrangements is challenging. This is evident in the Alta Energy case.

II. The Alta Energy Case

A. Overview

Alta Energy is known as a treaty-shopping case. It continues taxpayers’ winning streak in Canadian courts: MIL (Investments) (2007),²² Prévost Car (2009),²³ and Velcro Canada (2012).²⁴ Like MIL (Investments), Alta Energy deals with article 13(4) of the Canada-Luxembourg treaty and the relevance of the Canadian GAAR, but with an emphasis on the carveout. Article 13(4) states:

Gains derived by a resident of a Contracting State from the alienation of . . . shares . . . forming part of a substantial interest in the capital stock of a company the value of which shares is derived principally from immovable property situated in that other State . . . may be taxed in that other State. For the purposes of this paragraph, the term “immovable property” does not include property (other than rental property) in which the business of the company, partnership, trust or estate was carried on. [Emphasis added.]

Under article 13(4), if a resident of Luxembourg sells the shares of its Canadian subsidiary company that owns immovable property in Canada, the capital gains may be taxed in Canada under Canadian domestic law. However, article 13(4) does not apply to immovable property (other than rental property) in which the business of the Canadian company was carried on (the carveout, or business property exemption). The capital gains will be taxed only in Luxembourg under article 13(5).

The carveout applies if the shareholder of the Canadian company is a resident of Luxembourg and the gains are attributable to business property. The second condition was found by the Tax Court of Canada to have been met and that decision was not the subject of appeal.

The main issue before the Court is whether the taxpayer’s residence in Luxembourg was consistent with the purpose of article 13(4) or was abusive within the meaning of the GAAR. The taxpayer conceded that the treaty-shopping arrangement was an avoidance transaction because the primary purpose was to take advantage of the carveout clause. The key question is whether a Luxembourg resident must have some economic connection with Luxembourg. Six justices answered no and three answered yes.

B. The Facts

The facts of this case are not disputed.²⁵ In April 2011 Alta Energy Partners LLC, a Delaware LLC (Alta US), was created by a Texas-based oil and gas firm and a New York-based private equity firm in the form of a limited partnership to acquire and develop unconventional oil and natural gas properties in North America. Approximately 50 percent of the private equity investors were U.S. citizens or residents, including institutional investors.

In June 2011 Alta Canada was incorporated as a wholly-owned subsidiary of Alta US. From June 2011 to April 2012, Alta Canada obtained licenses granting limited term exclusive rights to drill for and recover oil and natural gas. These licenses do not grant legal title to the surface of the land. By March 2013 Alta Canada had acquired licenses and leases for 67,891 acres in the Duvernay shale (in the province of Alberta), most of which were acquired before the restructuring steps described below.

On April 19, 2012, the taxpayer company, Alta Energy Luxembourg SARL (Luxco) was incorporated under the laws of Luxembourg. Its sole shareholder was a partnership established in Canada, and the partners were the investors in Alta US. On the same day, Alta US transferred its common shares of Alta Canada to Luxco for a demand promissory note. Luxco had no employees or office of its own. It was a single-purpose holding company. Between 2012 and 2013 Alta Canada had six wells drilled and was a nonoperator in two additional wells.

In August 2013 Luxco sold the shares of Alta Canada to Chevron for a gain of C $380 million. Luxco did not conduct any other business or investment either before or after the sale. It claimed exemption of the capital gains from Canadian tax by virtue of the carveout. The MNR accepted that Luxco was a resident of Luxembourg²⁶ but invoked the GAAR to deny the treaty exemption.

C. The Majority’s Decision

Justice Suzanne Côté, writing for the majority, concludes that the MNR has not discharged her burden of proof about abuse and the GAAR does not apply. Her opening paragraphs set the tone for her analysis and reasoning:

The principles of predictability, certainty, and fairness and respect for the right of taxpayers to legitimate tax minimization are the bedrock of tax law. In the context of international tax treaties, respect for negotiated bargains between contracting states is fundamental to ensure tax certainty and predictability and to uphold the principle of pacta sunt servanda (parties to a treaty must keep their sides of the bargain and perform their obligations in good faith, Art. 26 of the [VCLT]), pursuant to which parties to a treaty must keep their sides of the bargain.

[The GAAR] acts [as] a legislative limit on tax certainty by barring abusive tax avoidance transactions, including those in which taxpayers seek to obtain treaty benefits that were never intended by the contracting states. . . . In the bilateral treaty context, there are two sovereign states whose intentions are relevant.

In Côté’s view, the MNR’s argument that the Luxembourg treaty was not intended to benefit residents without “sufficient substantive economic connections” to their state of residence and that Luxco’s treaty benefit should be denied under the GAAR is tantamount to asking the Court to use the GAAR to change the result of the treaty bargain by “fundamentally altering the criteria under which a person is entitled to the benefits of the Treaty, thus frustrating the certainty and predictability sought by the drafters”²⁷ (emphasis in original).

The criterion for accessing treaty benefit is “residence.” Under articles 1 and 4(1) of the treaty, residence requires the taxpayer to be “liable to tax,” not “in fact subject to taxation.”²⁸ A formal test, such as the “place of incorporation” or “legal seat,” as opposed to the “real location of a corporation’s economic activities” is used by many countries, including Canada.²⁹ The fact that article 28(3) of the treaty excludes some “holding companies” from the treaty implies that not every company with limited economic ties to Luxembourg is not a resident of that country. The “spirit” of articles 1 and 4 was not to limit access to the benefits of the treaty to corporations with “sufficient substantive economic connections” to their country of residence.

Côté finds the carveout to be a clear departure from the economic allegiance theory and that its purpose is to encourage investment in Canada. As a tax incentive, the carveout “was never intended to be limited to companies with ‘sufficient substantive economic connections’ to Luxembourg.”³⁰ Since Canadian treaty negotiators understood that Luxembourg was a low-tax jurisdiction and decided to trade tax revenue for jobs and economic opportunities by agreeing to the carveout, the GAAR cannot be used “to judicially amend or renegotiate a treaty.”³¹

Côté also draws negative reference from the absence of an anti-treaty-shopping rule in article 13. She says Canada and Luxembourg chose to restrict the beneficial ownership rule to articles 10, 11, and 12, and to not include a subject-to-tax provision in article 13. Had the parties truly intended to deprive conduit corporations of the benefits of the carveout, they would have made the carveout subject to a purpose-based antiavoidance rule (such as the rule in Canada’s treaties with Nigeria, Ukraine, Kazakhstan, Uzbekistan, and Peru)³²:

The absence of any such anti-avoidance measure that would have limited access to the carve-out in a treaty with a country known for not taxing capital gains leads me to believe that Canada weighed the pros and cons and concluded that its national interest in attracting foreign investors, using Luxembourg as a conduit to take advantage of the carve-out, outweighed its interest in collecting more tax revenues on such capital gains. . . . This choice must also have been motivated by the fact that Canada was not keen on going its own way at a time when the international community was not yet as serious about curtailing treaty shopping as it was during the years leading to the signature and ratification of the [MLI].³³

Côté also takes notice of some broad background evidence, such as the facts that:

• Canada is a capital importing country;

• “harsh source taxes chase away foreign investors, whereas tax breaks attract them”;

• Luxembourg is a known tax haven; and

• treaty shopping was not an unforeseen tax strategy at the time of the Luxembourg treaty.³⁴

Given the holding that the object, spirit, and purpose of the carveout is to foster international investment,³⁵ the treaty provisions “operated as they were intended to operate.”³⁶ Therefore, there was no abuse.

D. The Dissent

Justices Malcolm Rowe and Sheilah L. Martin wrote the dissent opinion and were joined by Chief Justice Richard Wagner. They agree with the MNR and held that the treaty-shopping arrangement is abusive under the GAAR. They state:

Multinational companies exploiting gaps and mismatches in international tax rules erode domestic tax bases and cost countries an estimated US$100 to US$240 billion in lost revenue annually. . . .

In introducing the GAAR . . . Parliament made a policy choice by which it intended to fight harmful tax avoidance schemes that cross the line of legitimate tax planning and venture into the realm of abusive tax avoidance.

Courts now have the responsibility to give proper effect to the intention of Parliament and ensure the GAAR plays a meaningful role in controlling avoidance transactions that technically comply with the provisions of a tax treaty but frustrate their underlying rationale. The interpretation exercise that is mandated in a GAAR analysis thus vests upon courts the unusual duty to look beyond the words of the applicable provisions to determine whether the transactions in question frustrate the underlying rationale of those provisions. [Emphasis in original.]

Given the GAAR can only find application where a taxpayer has complied with the strict requirements of a provision, absolute certainty cannot be achieved, nor was it intended. This is a legislative choice that Parliament made in order to strike a necessary balance between the uncertainty inherent in the GAAR and the fairness of the Canadian tax system as a whole achieved by defeating abusive tax avoidance schemes.³⁷

They agree with the majority that the carveout is a tax incentive, but its purpose and rationale is to limit the tax benefit to companies with “genuine economic connections” with Luxembourg.

The main basis for their finding is the economic allegiance theory. Rowe and Martin state that this theory explains why articles 1 and 4 limit the beneficiaries of the treaty to the residents of either state and why active income is generally taxable in the source country while passive income is taxed primarily in the residence country. While capital gains are not clearly active or passive income, article 13 generally reflects the economic allegiance theory.³⁸ The carveout clause allocates the taxing right to the residence state to reflect the fact that gains from immovable property in which business is carried on are driven by business activity and have closer economic ties with the residence state.³⁹ It follows the same logic of the treaty. To benefit from the carveout, the taxpayer must have economic ties with the residence country.

In the view of Rowe and Martin, the lack of a specific anticonduit rule in article 13 “sheds little light on their underlying rationale,”⁴⁰ and the implied exclusion principle has already been clearly rejected by the Canadian Supreme Court in Copthorne on the ground that it would be “a full response in all GAAR cases, because the actions of a taxpayer will always be permitted by the text of the Act.”⁴¹ They believe that Canada has not deliberately decided to extend the tax incentive in the Luxembourg treaty to investors in third countries:

Our colleague’s reasons assume that the federal government deliberately set out, in the exercise of its treaty making authority, to create the conditions for unlimited tax avoidance by means of schemes such as that in which Alta Luxembourg was used. To state such a proposition is to expose its absurdity, yet our colleague seeks to legitimize such blatantly abusive tax avoidance based on the view that Canada should have negotiated different treaty terms. The focus on what else could hypothetically have been agreed to is misplaced. It involves ex ante speculation about how the treaty parties ought to have proceeded based on alternatives said to have been available to them. However, such an argument gives primacy to what is not there. We are of the view that Parliament was entitled to rely on the GAAR to address abusive uses of the Treaty rather than negotiate the inclusion of a specific rule. The focus should be on what was actually agreed upon and whether the underlying rationale of the relevant provisions was frustrated by the avoidance transactions undertaken. In the give and take of treaty negotiation, Canada certainly did not give up the GAAR.⁴²

Given that the purpose and rationale of the carveout is to encourage investments by companies with genuine economic ties to Luxembourg, the avoidance transactions are “disconnected from the economic objectives underlying the bargain”⁴³ and thus abusive. Luxco’s presence in Luxembourg “is not genuine,” is “mere gossamer,” “was manufactured out of whole cloth,”⁴⁴ and was used to liquidate an investment in Canada without paying Canadian tax.

E. Some Takeaways

There are some important takeaways from the Alta Energy decision. Despite the 6-3 split in the decision, all nine justices find that:

• economic allegiance theory underlies the general allocation of taxing rights;

• the carveout is a tax incentive; and

• treaty shopping is not inherently abusive under the GAAR.

The majority and dissent draw inferences from broad extrinsic evidence about the intention of treaty negotiators but came to different conclusions. They differ about whether any economic connections with Luxembourg are required to benefit from the tax incentive. That, of course, is the key issue in treaty-shopping cases.

Côté states that treaty shopping may be considered immoral,⁴⁵ but the application of the GAAR is not premised on a “value judgement of what is right or wrong.”⁴⁶ Taxpayers are entitled to enter into transactions to minimize tax, and the government can decide what is right or wrong and translate these decisions into legislation that courts can apply. “The courts’ role is limited to determining whether a transaction abuses the object, spirit, and purpose of the specific provisions relied on by the taxpayer. It is not to rewrite tax statutes and tax treaties to prevent treaty shopping when these instruments do not clearly do so.”⁴⁷

Rowe and Martin agree with the majority on the morality point. However, they draw a line between acceptable and abusive treaty-shopping. A treaty-shopping arrangement is abusive when there is an absence of any “genuine economic connection with the state of residence.” When contracting parties allocate taxing rights to the state of residence on the basis of economic allegiance, this abusive type of treaty shopping “upsets the balance and reciprocity of the tax treaty,” and undermines the rationale of the treaty provisions.⁴⁸ They do not say, however, what constitutes “genuine economic connection” in situations in which the conduit is more than a “mere gossamer.”

III. The Purpose of Treaty Provisions

A. The Alta Energy Approach Is Problematic

The Court’s ways of finding the purpose of the carveout in article 13(4) are problematic for three main reasons. First, the Court fails to pay sufficient attention to the differences between domestic tax statutes and tax treaties. The majority apply the implied exclusion doctrine as if the treaty were written in the same way as the ITA. The whole court characterizes the carveout as a tax incentive provision because of its apparent departure from the OECD model and its effect. This shows a misunderstanding of the role of the OECD model. As explained in more detail below, such characterization is not supported by the text and context of article 13(4). On the contrary, the carveout is arguably more consistent with the “normative benchmark” of the treaty in distributing taxing rights over income from property and capital gains from alienation of property that is effectively connected with a business.

Second, the Court seems to conflate intention of treaty negotiators and the purpose of treaty provisions. As discussed in more detail below, this is problematic. In applying the GAAR to domestic law provisions, the Court has emphasized legislative purpose or rationale, as opposed to legislators’ intention.⁴⁹ Intention denotes a state of mind of a large and diverse group of legislators at the time of enactment of legislation. Purpose or rationale (which can be broadly understood to include policy) of statutory provisions may not depend on any one person’s intentions. The same can be said of tax treaties.

Finally, the Court infers the intention of Canada and Luxembourg from broad extrinsic materials, including scholarly and professional commentaries. The majority also considers the implied exclusion principle and the fact that Canada was interested in attracting investment. The dissent considers the economic allegiance theory and general treaty logic in allocating taxing rights. Insufficient consideration was given to the treaty’s text and context.

B. The Implied Exclusion Doctrine

Under the implied exclusion doctrine, an argument is made that if the contracting states had meant to address an avoidance plan, they would have referred to that plan expressly, and the failure to do so is presumably intentional. The majority adopts this doctrine in finding that the absence of a specific anticonduit rule in article 13(4) confirms the view that Canada’s primary objective was to attract foreign investment.

The majority’s view is incorrect. As pointed out by the dissent, the implied exclusion doctrine was rejected by the Court in Copthorne as it would gut the GAAR. Applying this doctrine to treaty interpretation ignores the fact that treaties are not written in a detailed, airtight manner because they are meant to coordinate the intersection of two tax systems as opposed to specifying the details on determination of tax liability for taxpayers. More importantly, section 4.1 of the ITCIA is intended to reduce the need for specific antiavoidance rules in tax treaties.

C. Purpose Differs from Intention

The purpose test in the GAAR focuses on the purpose of the provisions in domestic law or treaties, not the state of mind or intention of the lawmaker or treaty negotiator. As explained by Lord Burrows, purposive statutory interpretation “may be said to be analogous to identifying the principle behind a common law precedent and that, too, is not dependent on trying to identify any person’s (i.e. judge’s) intention.”⁵⁰ The legislator’s intention is practically irrelevant. The focus should be on the statute’s words, context, and purpose:

A serious objection to any reference to legislative intention is that it is advocating an approach that favours the law’s ossification by inappropriately freezing the law in the past. We would not accept such an approach for the common law and there is no good reason why we should regard it as acceptable when interpreting legislation.⁵¹ [Emphasis added.]

If legislative purpose is not dependent on legislative intention in constructing domestic statutes, should the same approach apply to treaty interpretation?

It can be argued that intention may be more relevant in treaty interpretation because a treaty is a contract between countries, and the intention of contracting parties is important in contractual interpretation. Also, the intention of the parties is considered part of the “context of the treaty” within the meaning of article 3(2) of the OECD model.⁵²

On the other hand, concerns with imputing legislative purpose from legislators’ intention in statutory interpretation also apply to treaty interpretation. The meaning and purpose would be frozen at the time of conclusion of the treaty, which is contradictory to the general principle of ambulatory interpretation of tax treaties.⁵³ Further, inferences may be drawn from different extrinsic materials. For example, the majority in Alta Energy considers the absence of a specific anticonduit rule in the treaty, while the dissent regards the implied exclusion principle irrelevant and relies on the economic allegiance theory as underpinning the logic of the treaty in allocating taxing rights to the residence or source country.

Perhaps of greatest importance is the fact that, in Canada, the GAAR applies to treaty provisions because of two domestic statutes — the ITA and the ITCIA. The GAAR explicitly refers to “an abuse having regard to those provisions” and makes no reference to the possible state of mind (or intention) of individuals at the time the provisions were drafted.

The danger in the majority’s approach is to downplay or even neutralize the effect of using the GAAR provision to replace a proliferation of specific antiavoidance rules in both domestic statutes and treaties. This approach may also result in attributing bad faith on the part of treaty negotiators toward third countries, by effectively making the carveout available to persons in a third country (in the case of Alta Energy, to residents of the United States) in a way that undermines the treaty bargain between Canada and that third country.

The danger in the dissent’s reasoning is imputing knowledge about the economic allegiance theory that may not be considered by treaty negotiators. To begin with, the influence of this theory on the model convention is unclear. The theory was formulated in a 1923 report by four economists who were commissioned by the League of Nations to study the issue of double taxation.⁵⁴ The economists’ considerations and final recommendations provided some intellectual basis but “were ultimately tempered by practical considerations”⁵⁵ in developing the original model conventions in the 1920s. Secondly, this theory is not mentioned in the text of original models, the first OECD model published in 1963, or the model in effect when the Canada-Luxembourg treaty was negotiated. It is also absent in the commentaries on the OECD model. It is the model convention that was relevant in treaty negotiations. In general practice, treaty negotiations regard traditional treaty norms as most important, and only a small part of tax treaties is originally drafted during negotiations.⁵⁶ No Canadian treaty mentions the economic allegiance theory.

Judicial interpretation is about “discovering” the meaning and purpose of the law, not being a “mask for judges to hide their true reasoning by dressing a decision up as effecting Parliamentary intention.”⁵⁷ The same can be said about treaty interpretation:

In deciding on the best interpretation of a statute, the courts need to rely on the more concretised ideas that revolve around the words, context and purpose of the statute. Reliance on the “high-level” idea of Parliamentary intention is unhelpful, at best, and has the tendency to mask the true reasoning and power of the courts.⁵⁸

D. Let the Treaty Speak

1. Text and Context of the Treaty

If the Canada-Luxembourg treaty were to speak for itself, what would it say about the purpose of the carveout in article 13(4)?

To answer this question, one must look beyond the words in the carveout and consider article 13 as a whole, other articles directly related to article 13 in regard to distributing taxing rights between the two countries, as well as the title, preamble and other provisions of the treaty.

Under articles 31 and 32 of the VCLT, one can also consider any agreements between Canada and Luxembourg in connection with the conclusion of the treaty, any instrument related to the treaty, subsequent practice in the application of the treaty, and supplementary means of interpretation, such as preparatory work of the treaty and the circumstances of its conclusion.

Under Canadian case law,⁵⁹ OECD commentaries on the model convention have “high persuasive value” in treaty interpretation. The following discussions focus on the Canada-Luxembourg treaty per se.

2. The Carveout and Capital Gains Taxation

Article 13, read as a whole, suggests that, at its core, the carveout is part of the general scheme of the treaty in assigning taxing rights, as opposed to a deliberate tax incentive.

The plain text of article 13(4) shows the carveout as a departure from the paragraph, but the context of the provision shows that the carveout is more of an alignment with the general logic of allocating taxing rights:

• article 13(1) allocates taxing right to the source country in which immovable property is situated, revealing a principle of situs taxation;

• article 13(2) allocates taxing right to the source country in which a permanent establishment is located, revealing a principle of PE taxation;

• article 13(3) allocates taxing right to the residence country of an international shipping or airline company;

• article 13(4) allocates taxing right to the source country in which immovable property is situated when such property is held indirectly through a corporation, partnership, or trust with the exception of business property, which is consistent with the principle of situs taxation; and

• article 13(5) leaves all residual gains, including gains from alienation of shares not governed by article 13(4), to be taxed in the residence country.⁶⁰

Unlike other distributive provisions (for example, articles 6-8 and 10-12), article 13 has no independent scheme for allocating tax rights; it merely mirrors the scheme in other distributive provisions, depending on the use of the property that is alienated.⁶¹

Article 13(4) is a specific antiavoidance rule. A similar provision was first added to the U.N. model convention to protect the interest of source countries.⁶² It backstops the principle of situs taxation (taxation in the country in which immovable property is located), which is enshrined in article 6 (income from immovable property) and article 13(1) (gains from alienation of immovable property). Article 13(4) looks through the corporate fiction that separates, in law, the immovable property and its economic owner (shareholder) to protect the integrity of the principle of situs taxation.

On its face, the carveout is aligned with article 13(5) by assigning the taxing right to the residence country of the shareholder — gains from the alienation of any property not expressly referred to in articles 13(1)-(4) are taxable in the country of residence of the taxpayer. As a result, article 13(5) applies to the alienation of shares that do not derive their value principally from immoveable property and to shares that do, if a business is carried on in the property.

In effect, the carveout treats immovable property used in a business differently from immovable property in general. As explained below, this is consistent with the general scheme of business taxation.

3. The Carveout and Business Taxation

The carveout clause in article 13(4) singles out immovable property in which business is carried on. As noted by Rowe and Martin,⁶³ the capital gains from the alienation of shares of Alta Canada are derived, at least in part, from the business activity of Alta Canada. The purpose of the carveout may be gleaned from the treaty’s scheme for taxing business profit.

Articles 7, 5, 10, 11, and 12 reveal a scheme for taxing business activities and its prevalence over the principle of taxing passive income (or income from property):

• Article 7 assigns taxing rights over business profits to the source country in which the PE is located and through which business is carried on, revealing the principle of PE taxation. It imposes no limitation on tax rate.

• Article 5 defines PE to include a fixed place of business, an oil or gas well, or any other place relating to the exploration for or the exploitation of natural resources. An immovable property can thus be a PE.

• Article 10 assigns primary but limited (5 percent or 15 percent) taxing rights over dividends to the source country and residual rights to the residence country. However, article 10(7) carves out dividends that are in the nature of business profits or are effectively connected with a PE so that those dividends are taxable under article 7. In other words, when the income takes the form of dividends but is, in effect, derived from business, it is taxed as business income. Article 10(7) reflects a hierarchy in assigning taxing rights — business taxation trumps passive income taxation rules and the principle of PE taxation prevails.

• Article 11 assigns primary but limited (10 percent) taxing rights on interest to the source country and residual rights to the residence country. Article 11(6) is similar to article 10(7).

• Article 12 assigns primary but limited (10 percent) taxing rights on royalties to the source country and residual rights to the residence country. Article 12(5) is similar to article 10(7).

In the case of capital gains from the alienation of shares of a corporation carrying on a business, the taxing right lies with the country of residence of the shareholder, unless the shares derive their value principally from immoveable property situated in the source country that is not used in a business.

As a result, if a resident of Luxembourg carries on business through a PE in Canada, capital gains derived from the disposition of immoveable property are taxable by Canada under article 13(1) or (2). On the other hand, if a Canadian corporation is used to carry on the business, the taxing right will generally lie with the country of the shareholder’s residence under article 13(5). Such asymmetry reflects the basic architecture of the treaty that recognizes each corporation as a separate legal entity.

Why does the treaty assign taxing rights to a capital gain on an alienation of shares to the country of shareholder’s residence when a corporate form is used to carry on business in an immoveable property? The Court in Alta Energy says this is a deliberate tax incentive. If so, is article 13(5) that allocates taxing rights to the country of residence on the alienation of shares without immoveable property another deliberative incentive?

A better explanation may simply be that the look-through rule in article 13(4) and the carveout ensure that source-country taxation on capital gains from immoveable property used in a rental or other “passive” activity cannot be avoided by carrying out the activity in incorporated form. Similar concerns about tax avoidance do not apply to incorporated active business.

In addition to showing that the carveout reflects the general logic of the treaty, the above treaty provisions show the errors in the following statements of the Court:

• The residence state “has the primary right to tax passive income (e.g. interest, dividends, and capital gains), and the source state has only residual rights.”⁶⁴ Articles 10, 11, 12, and 13 say that the source state has the primary right to tax dividends, interest and royalties, capital gains from immovable property, and capital gains from movable business property.

• “As a general rule, residence is taken to indicate the state to which economic ties are closest.”⁶⁵ This generalization is not true for income that is taxable primarily in the source country under the PE taxation principle and situs of immovable property principle.

• The carveout of article 13(4) “allocates the taxing right to the state of residence for its residents’ capital gains from immovable property when it is driven by business activity, to reflect what Canada and Luxembourg considered to be closer economic ties with the residence state.”⁶⁶ This statement confuses the general rationale for using residence as a basis of assigning taxing rights and the business taxation principle.

4. Residence’s Treaty Meaning

Does residence in Luxembourg under the treaty mean what Luxembourg domestic law says (formal test by reference to legal seat), or does the context of the treaty require some economic connection? Why is residence chosen to define the scope of the treaty?

The answer can be gleaned from the logic underpinning the following provisions:

• Article 1 says that only residents are covered by the treaty.

• Article 4 refers to domestic law for the definition of residence.

• Article 28(3) says the treaty “shall not apply to holding companies within the meaning of the special Luxembourg laws (currently the Act of July 31, 1929, and the Grand Duchy Order of December 17, 1938) or any other similar law enacted in Luxembourg after the signature of the Convention, nor to companies subjected to similar fiscal laws in Luxembourg.” These holding companies are subject to some restrictions in their activities but are exempt from tax. As such, they are different from general corporations.⁶⁷

• Article 3(2) is a general interpretation rule that says, unless the context otherwise requires, the meaning of treaty terms that are not defined in the treaty shall take their meaning from domestic law. The context of the treaty may require modifications to the meaning under domestic law to achieve the objectives of the treaty. In cases such as Alta Energy, article 3(2) may permit the meaning of residence to be more than a mere formal presence required by Luxembourg domestic law to better achieve the objectives of the treaty.

The context of the treaty includes the preamble, title, and other provisions as well as preparatory work and subsequent practice.⁶⁸ Under the ambulatory approach to treaty interpretation, the “treaty is always speaking” in the sense that the meaning of treaty terms is not frozen at the time of concluding the treaty but is amended from time to time, often through changes in domestic laws.

The title and preamble of the Canada-Luxembourg treaty mention the “avoidance of double taxation.” To avoid double taxation, the treaty assigns taxing rights by reference to residence of taxpayers and source of income, imposes limitations on source taxation, and requires the residence country to recognize source-country taxation and provide double tax relief through exemption or foreign tax credit. As such, residence is a basis for assessing tax liability. Under the domestic law of both countries, residence-based taxation is broader than source-based taxation. In the absence of double taxation, the raison d’être of the treaty is missing.

The beneficial owner test in articles 10, 11, and 12 and the exclusion of Luxembourg holding companies in article 28(3) support the argument that the treaty is not applicable to a “formal” resident of a treaty country that is not taxed in the residence country. Nominal owners of income pass on their income to beneficial owners and do not pay tax on the income. The special holding companies in Luxembourg are tax-exempt entities under Luxembourg law. Therefore, the treaty does not benefit these formal residents. There is an implicit anticonduit principle emanating from these provisions.

The Canadian GAAR is part of the treaty context because Parliament clearly says so in section 4.1 of the ITCIA. Instead of adding specific anti-treaty-abuse rules, Parliament decided to prevent tax avoidance through the GAAR. Therefore, the GAAR can be understood to extend the anticonduit principle implicit in articles 10, 11, 12, and 28(3) to other provisions, such as article 13.

The intention of treaty negotiators forms part of the context of the treaty within the meaning of article 3(2). Treaty negotiators can be presumed to know their own country’s tax law, including the GAAR, and the other country’s domestic tax law. They can also be presumed to act in good faith while seeking to maximize the economic benefit of the treaty for their countries. As between the intentions found by the majority and dissent, the dissent’s finding is more consistent with what the treaty provisions imply.

Given the acceptance of the majority and dissent that economic allegiance underlies the allocation of taxing rights in the treaty, including article 13, and the fact that residence is one of the two expressions of economic allegiance, it is only logical to expect residence to include some economic connection. It is debatable about what type or level of economic connection is required, but categorically accepting formal or legal connection defies the logic of economic allegiance.

E. A Note on Litigation and Judicial Interpretation

The above contextual analysis of treaty provisions is largely missing in the Alta Energy decision. It is difficult to know whether any of the justices would accept this analysis if it were presented to them; judges generally consider only arguments advanced by the parties.

Litigation strategies may also affect the case outcome. In Alta Energy, the MNR accepted that Luxco was a resident of Luxembourg and was technically entitled to the treaty benefit. This reflects a feature of Canadian tax jurisprudence that favors a form-over-substance approach to constructing taxpayer’s transactions. To deny that benefit under the GAAR requires the MNR to prove that the abuse is clear. Had the MNR challenged the residence status from the inception of the dispute resolution process and argued that residence requires some economic connections, it may or may not have affected the outcome of the case, although the focus would not be on the purpose of the carveout, but the meaning of residence.

The majority holds the common-law principles, including respect for the right of taxpayers to legitimate tax minimization (or the Duke of Westminster principle), as the “bedrock of tax law”⁶⁹ and relies on them to limit the application of the GAAR. Such degree of adherence to common-law principles is somewhat surprising as the Court’s own jurisprudence holds that the GAAR was enacted to attenuate the Duke of Westminster principle.⁷⁰

IV. Observations

A. Implications for the GAAR

The Alta Energy case will likely make it more difficult for the GAAR to apply in general. The implied exclusion principle means that any tax avoidance transaction that is not explicitly addressed by a provision is presumed to be intended, and thus not abusive. Limiting the GAAR to preventing only “unforeseen” avoidance transactions narrows its scope.

This case also exposes the deficiency in the judicial interpretation of the GAAR by distinguishing between “general statutory interpretation” by focusing on the meaning of the provisions and purposive interpretation in the context of a GAAR analysis. The MNR presumably understood “residence” to have a technical meaning under Luxembourg domestic law and considered the treaty-shopping arrangements “valid on their face” so that GAAR must be used. Had a contextual and purposive interpretation been considered, the MNR may have pleaded differently.

To prevent treaty abuse, Parliament could consider amending the ITCIA by adding a provision to define “residence in a treaty country” to have some meaningful economic connections or incorporating a more evidence-based LOB as part of Canadian treaties.

B. Implications for the PPT

The PPT or similar purpose-based antiabuse rules in Canada’s tax treaties will likely be affected by Alta Energy. However, the inclusion of the following revised preamble by virtue of article 6 of the MLI may lessen the impact:

Intending to eliminate double taxation with respect to the taxes covered by this agreement without creating opportunities for non-taxation or reduced taxation through tax evasion or avoidance (including through treaty-shopping arrangements aimed at obtaining reliefs provided in this agreement for the indirect benefit of residents of third jurisdictions).

Presumably, the intention to prevent treaty-shopping arrangements would influence the finding of the object and purpose of provisions.

Further, since economic allegiance is accepted by the Court as the foundation for allocating taxing rights, lacking economic allegiance may be an important factor to consider in future cases. It may be possible to distinguish a future treaty abuse case from Alta Energy when the provision in question is not viewed as a tax incentive, although a court may find any deviation from the OECD model as a tax incentive of some sort.

C. Treaty Interpretation

The Alta Energy judgement illustrates some of the challenges in applying the purpose test in treaty-shopping cases. Even though it may have a unique Canadian flavor it should offer some food for thought about treaty interpretation in general and applying the purpose test in particular. The analysis and reasoning of the Court show the problems of assuming negotiators’ intention from broad extrinsic evidence, conflating intention with purpose and lacking sufficient consideration of the differences between treaties and domestic laws. To improve certainty and predictability, the finding of purpose should be more guided by the actual bargain reflected by the treaty itself.