Featured Articles

Traveling Without a Destination: Post-BEPS Anti-Treaty-Shopping Rules and Non-CIV Funds in Canada and the U.S.

Posted on September 20, 2017 by Rémi Gagnon

Table of Contents

  1. I. Setting the Boundaries
    1. A. Defining Treaty Shopping
    2. B. Defining Non-CIV Funds
    3. C. Non-CIV Funds and Treaty Shopping
  2. II. Setting the Stage
    1. A. Canadian Tax Treaty Policy
    2. B. U.S. Tax Treaty Policy
  3. III. Treaty Shopping Meets Treaty Policy
    1. A. The United States
    2. B. Canada
  4. IV. The Way Forward
    1. A. The LOB Provision
    2. B. The PPT Provision
  5. V. Conclusion

“Would you tell me, please, which way I ought to go from here?” said Alice.

“That depends a good deal on where you want to get to,” said the Cat.

“I don’t much care where,” said Alice.

“Then it doesn’t matter which way you go,” said the Cat.

— Lewis Carroll, Alice’s Adventures in Wonderland

Treaty shopping has never been as high on the political agenda as it is today. At the same time, non-collective investment vehicle (non-CIV) funds, particularly private equity funds, have never been as prevalent as they are today.1 The interplay of these two phenomena has given birth to a remarkable tax conundrum, and there is no guarantee that non-CIV funds will come out on top. The purpose of this article is to analyze the development of anti-treaty-shopping measures in a post-base erosion and profit-shifting era, as illustrated by non-CIV funds.

In this article, I argue that the anti-treaty-shopping initiative shepherded by the OECD fails to properly take into account the diverging tax treaty policies of various states. As a result, the OECD’s one-size-fits-all proposal to curb treaty shopping is at times squarely incompatible with some of these national policies. This is demonstrated through a comparative analysis of the application of anti-treaty-shopping measures to non-CIV funds in Canada and the United States.

After addressing preliminary matters, I will examine Canadian and U.S. tax treaty policies to establish their importance. While Canada has an interest in both residence and source taxation, the United States has an interest principally in residence taxation. While Canada has a very large and uniform tax treaty network, the United States has a smaller and less uniform tax treaty network. As we shall see, the implication is that treaty shopping is harmful to the latter but less so to the former.

Next, I develop a tangible framework to explore the two main approaches to combatting treaty shopping. The United States, as the instigator of the limitation on benefits provisions, is well suited to an LOB analysis. Canada represents a perfect case for a principal purpose test (PPT) because in 2014 Canada proposed a domestic anti-treaty-shopping rule based on a similar test. Equally important, Canada also has a general antiavoidance rule. Again, as we shall see, both approaches would benefit from further refinements.

Finally, I lay out recommendations for improving the two main anti-treaty-shopping approaches proposed by the OECD.

While this article focuses on non-CIV funds, it is at its core a critique of the anti-treaty-shopping initiative. Thus, many of the propositions made in this article could apply in other contexts. Non-CIV funds are mainly discussed in the first and fourth parts of this article.

I. Setting the Boundaries

A. Defining Treaty Shopping

The OECD in its BEPS action 6 final report refers to treaty shopping as “arrangements through which a person who is not a resident of a contracting state . . . attempt[s] to obtain benefits that a tax treaty grants to a resident of that state.” It continues, “Treaty shopping cases typically involve persons who are residents of third states attempting to access indirectly the benefits of a treaty between two contracting states.”2 In effect, as H. David Rosenbloom asserts, treaty shopping essentially amounts to “borrowing” a tax treaty by forming an entity in a particular state.3 Despite these definitions, the expression “treaty shopping” has also been used to describe other tax treaty abuses and appears at times to lack a well-defined meaning.4 For this article, the overall concept of treaty shopping deals with the Canadian- or U.S.-inbound international tax system and treaty benefits — particularly, the reduced withholding tax rates — being extended to persons who are not entitled to them.

B. Defining Non-CIV Funds

The OECD defines CIV funds as those that are “widely-held, hold a diversified portfolio of securities and are subject to investor-protection regulation in the country in which they are established.”5 Thus, non-CIV funds are generally managed alternative investment funds that are not widely held. While the definition encompasses a wide range of funds, this article focuses mainly on private equity funds.

C. Non-CIV Funds and Treaty Shopping

Anti-treaty-shopping rules raise several issues for non-CIV funds. The difficulty arises from the passthrough nature of non-CIV funds, the use of blocker corporations, and the fact that some investors may reside in non-treaty states. Non-CIV funds are generally structured as partnerships, which are often treated by states as flowthroughs for tax purposes, meaning entitlements to treaty benefits apply through non-CIV funds to the ultimate investors. This is the case in Canada and the U.S. and will be assumed throughout this article.6 Further, non-CIV funds rely extensively on blocker corporations resident in treaty states. The stated reasons for using a blocker corporation may be purely commercial, purely tax-related, or a mix of both. These include accommodating the needs of co-investors, leveraging the investors’ capital to finance the investments, providing additional protection against liabilities associated with a distressed business, streamlining the process of claiming treaty benefits, avoiding the need to navigate the complex and inconsistent national rules concerning income derived through a passthrough entity, stripping earnings, or addressing specific tax concerns of nonresident or tax-exempt investors.7

II. Setting the Stage

A central contention of this article is that treaty-shopping discussions cannot ignore states’ specific tax treaty policies. Thus, the first step is to examine Canadian and U.S. policies.

A. Canadian Tax Treaty Policy

Canada, as a small open economy, has interests in both residence and source taxation. As Brian Ernewein, general director of the Tax Policy Branch of the Department of Finance (DOF), explains, Canadian tax treaty policy is “somewhere in the middle, although tending toward residence-based” taxation.8

Generally, Canada’s tax treaty policy aims to reduce its 25 percent domestic withholding tax rates on dividends, interest, and royalties to 5 percent in exchange for corresponding reductions from its treaty partners.9 At first blush, this seemingly puts Canada more toward the residence side on the source-residence tax spectrum. However, Canada’s historic refusal to lower its withholding tax rates below 15 percent, and its resource-based economy, indicate its interest as a source state.10 Also, the shrillness with which it responds to treaty shopping, emphasizing concerns about revenue loss, indicates some interest in source-based taxation.11

Indeed, the 2014 budget, released after a consultation paper outlining Canada’s treaty-shopping concerns,12 proposed a domestic anti-treaty-shopping rule based on a “one of the main purposes” test, even though the OECD was considering this very issue in its work on BEPS action 6 at the same time. The proposed rule was never made final because it conflicted with the treaty-based approach favored by BEPS action 6.13 On June 7, 2017, Canada signed the OECD’s Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting (MLI) to implement the BEPS action 6 recommendations on treaty shopping. That being said, curbing treaty shopping may also foster residence-based taxation. Canada’s response to treaty shopping arguably illustrates its interests as both a source and a residence state.

Another relevant component of Canadian tax treaty policy is that Canada strives to achieve the same outcomes in its various tax treaties, thereby resulting in a generally uniform tax treaty network.14 The only exception is the tax treaty with the United States, in which, in light of the magnitude and proximity of the economic relations between the two countries, there are “different terms and a different level of detail.”15

Furthermore, since the 1972 tax reforms, Canada enjoys two systems of international taxation. The changes made to the Income Tax Act in 1972 were designed to use a carrot-and-stick method to encourage other states to negotiate tax treaties with Canada.16 The stick component encourages other states to negotiate tax treaties by taxing residents of non-treaty states at a higher rate; the carrot component encourages Canadian residents to invest in treaty states by lowering or removing Canadian taxes on the income derived from these states. At the heart of the stick component lies an increase of the withholding tax rate from 15 percent to 25 percent, even though Canada maintained a general tax treaty policy goal of withholding rates of 15 percent or less.17 The carrot component, for its part, is essentially rooted in an exemption for dividends paid out of active income to Canadian parent corporations by foreign subsidiaries resident in treaty states.18

In 2009 Brian Ernewein underscored that the exemption system had “for the last 30 years . . . driven [Canada’s] treaty policy to a considerable extent . . . [and] has certainly been a catalyst in Canada’s case for negotiations.”19 Accordingly, Canada now enjoys “the luxury of two international tax systems,”20 R. Alan Short, Canada’s former tax treaty negotiator and chief of international tax policy, explained in 1973.

B. U.S. Tax Treaty Policy

The United States has interest in residence-based taxation. While the United States is generally thought of as a debtor, capital-importing state,21 payment flows between related entities (dividends, interest, royalties) have historically favored the United States, with the result that, for tax treaty purposes, the United States views itself as a capital-exporting state.22 U.S. tax treaty policy therefore seeks to eliminate source taxation. This is underscored by the U.S. model tax treaty providing withholding tax rates as low as 5 percent on dividends and 0 percent on interest and royalties.23 The hallmark of this policy, as Rosenbloom observed, is that “the United States has always entered the negotiation of treaties with the view to achieving the best deal it could with each treaty partner; . . . with the idea of getting the lowest possible taxation at source that [it] can.”24 Since treaty partners are not always ready to reduce withholding tax to the same rate, let alone to 0 percent, the result is that U.S. tax treaties vary greatly.25 Accordingly, U.S. tax treaty policy, exemplified by an amalgam of tax treaties providing various and inconsistent withholding tax rates, can be contrasted with the Canadian tax treaty policy, which gives rise to a relatively uniform treaty network setting out more or less the same withholding tax rates.26As we shall see, this distinction has important ramifications in fashioning a principled response to treaty shopping.

III. Treaty Shopping Meets Treaty Policy

A principled solution to treaty shopping depends a great deal on a source state’s tax treaty policy. Indeed, only by taking this policy into account can the true mischief of treaty shopping, if it exists, be tackled. This can be illustrated by considering whether Canada and the United States, in light of their respective tax treaty policies, should respond to treaty-shopping arrangements.

A. The United States

The answer is clear for the United States: Treaty shopping must be curbed. This answer stems from the United States’ interest in residence taxation and its use of tax treaties as a means of reducing taxation at source, inevitably leading to a treaty network with inconsistent withholding tax rates.27 To guard against states feeling no need to negotiate their own treaties because treaty shopping allows their residents to take advantage of treaties negotiated by other states, the United States must safeguard the integrity of each treaty.28

As Rosenbloom observes, “An overriding policy of using treaties, insofar as possible, to reduce source-basis taxation might be viewed as leading naturally to a policy of limiting treaty benefits as a matter of course.”29 The action 6 report recognizes this, stating, “If . . . a state knows that its residents can indirectly access the benefits of treaties concluded by another state, it may have little interest in granting reciprocal benefits to residents of that other state through the conclusion of a tax treaty.” Thus, the report finds that it is apparent that treaty shopping “undermines tax sovereignty.”

This concern is best illustrated by the treaty negotiations between the United States and Germany. The United States was after Germany for years to conclude a treaty to reduce taxation at source, but Germany would not agree because at the time, German corporations were able to engage in treaty shopping and invest in the United States through the Netherlands.30 However, as soon as an LOB provision was added to the Netherlands-U.S. treaty, Germany came running, asking the United States to reduce taxation at source.31 The disincentive to negotiate tax treaties caused by treaty shopping could not have been more obvious, and this example illustrates why the United States has been including LOB provisions in its tax treaties for decades.

Ultimately, this is a tax revenue issue, but not in the sense usually referenced in treaty-shopping discussions. When the action 6 report asserts that nonresidents engaged in treaty-shopping arrangements deprive states of tax revenues “by claiming treaty benefits in situations where these benefits were not intended to be granted,” it refers to the source state’s withholding tax rates. As far as the United States is concerned, however, treaty shopping arguably has nothing to do with source taxation. In fact, as the U.S. model tax treaty shows, the United States is prepared to largely eliminate the withholding tax imposed on nonresidents.32 The risk of reduced U.S. withholding tax rates being abused by nonresidents engaged in treaty shopping is merely a distraction. Rather, what matters is the effect treaty shopping has on the United States’ interest in residence taxation.

Specifically, the revenue cost to the United States consists of the increased foreign tax credits granted to U.S. persons who must pay higher withholding tax rates to the treaty partner. Had the treaty partner been inclined to reduce its rates, the United States might have been able to negotiate a lower rate and reap additional tax revenues from U.S. payees as a result. Because it was convenient for its residents to engage in treaty shopping, the treaty partner was unwilling to change its withholding tax rates (or to negotiate any treaty for that matter) and the deal never happened. In other words, the United States could not care less that nonresidents are enjoying nominal withholding tax rates — which is, after all, the very purpose of treaty shopping — as long as U.S. persons benefit from the same rates.

More broadly, that this nuance is missing from the action 6 report is troublesome. How can one purport to address treaty shopping without a proper understanding of the actual damages it engenders? Is baldly proclaiming the source state’s loss of tax revenues as the heart of the problem, without any inquiry about a given state’s tax treaty policy, the right way to approach this issue? While this may not seem problematic for the United States, which has a tax treaty policy that similarly supports anti-treaty-shopping measures in any event, it does not follow that this is always the case. As I argue later, treaty-shopping discussions are also guilty of this rush to judgment regarding Canada.

B. Canada

Considering Canada’s interest in both source and residence taxation as well as the enthusiasm with which Canada has embarked on the journey to thwart treaty shopping, one might be inclined to think that Canada has a lot at stake in the treaty-shopping debate. Steps it has taken include amending GAAR to make it expressly applicable to tax treaties,33almost enacting a domestic anti-treaty-shopping rule in 2014, and recently signing the MLI. This conclusion, however, does not bear a moment’s examination in light of the Canadian tax treaty policy.

First, contrary to the United States, Canada does have a uniform tax treaty network.34 Thus, treaty shopping does not appear to be a threat since the reduced treaty rates, while maybe not all the same practically speaking, are the same in policy terms.35 Second, it is hard to see how treaty shopping, in light of the Canadian tax treaty policy leading to two distinct international tax systems — the treaty system and the non-treaty system — could truly be as objectionable as the consultation paper and the action 6 report contend. Indeed, as stated by David A. Ward:

There is probably a legitimate question whether the access to the more favourable system of tax is in fact abusive if the distinction between the two systems was created and maintained principally to assist Canada in negotiating and completing treaties with other countries and not because one system, the non-treaty system, is inherently a better tax policy system than the other system, the treaty tax system.36

Given the breadth of the Canadian tax treaty network, which includes 93 tax treaties,37 one can probably safely assert that the Canadian tax treaty prophecy has been fulfilled.

Further, it is arguable that Canada entered into tax treaties with some states, most notably the Netherlands and Luxembourg, not specifically to facilitate bilateral trade but rather to encourage nonresidents not residing in a treaty state to invest in Canada through these states.38 Treaty shopping is not a new tax-planning technique, and it is not a secret that those states promote themselves as intermediaries.39 The Tax Court of Canada, in the treaty-shopping case of MIL (Investments) S.A. v. Canada,40 underscored this, stating, “Prior to negotiating the [Canada-Luxembourg tax treaty], Canada undoubtedly had knowledge of Luxembourg’s” tax regime. Also, the Canada-Netherlands and Canada-Luxembourg tax treaties were entered into in 1986 and 1999, respectively,41 long after the United States started including LOB provisions in its treaties to prevent treaty shopping.

More to the point, when the Canada-Luxembourg treaty was signed, the Canada-U.S. tax treaty already included a LOB provision, which was added through the third protocol in 1995.42 Clearly, Canada was more than aware of the treaty-shopping phenomenon.43 This appears to lead commentators to conclude that “it was commonly understood that the Netherlands and Luxembourg would serve as conduit countries.”44 This is consistent with the Tax Court of Canada’s suggestion in MIL that the Canada-Luxembourg tax treaty was entered into out of “the desire to encourage foreign investment in Canadian property.”45

So can Canada really claim to be surprised that nonresidents are using the Netherlands and Luxembourg as springboards for treaty-shopping purposes? Canada may now think that those and other tax treaties have turned out to be bad deals. However, that can only be the case if the high Canadian domestic withholding tax rates are positive policy-wise. Coupled with the exemption system of a residence state, they are not.46

Actually, in this case treaty shopping appears more consistent with the spirit of international tax treaty policy than efforts to curb it.47 To illustrate, consider an example in which the residence state’s domestic tax rate is at least as high as the source state’s domestic withholding tax rate.48 As long as the residence state grants credit for foreign taxes paid by its residents, treaty shopping simply cannot occur. The residence state’s residents are indifferent to the source state’s tax rate, since they can claim a foreign tax credit against their residence state tax liability. In fact, if every state’s domestic tax rate on its residents exceeds that on nonresidents, then all nonresidents are treated identically by every source state. The treaty is simply the method by which states allocate taxing rights, without any economic effect on nonresidents. This results in a simple and seamless international tax treaty policy exempt from abuses.

But notice how this elegant paradigm breaks down once the residence state adopts an exemption system. Nonresidents are no longer indifferent to the source state’s withholding tax rate, because the residence state no longer provides a refund through a foreign tax credit. For nonresidents, treaty shopping now becomes useful.49 Interestingly, by leveling the source states’ withholding tax rates, treaty shopping restores the status quo ante and nonresidents are thereby treated identically by every source state once again. Taking this perspective, treaty shopping is not an abuse of tax treaty policy — instead, exemption systems are an abuse of tax treaty policy, which treaty shopping seeks to correct.50

It is somewhat ironic to see Canada, which through its exemption system contributed to the very disconnect that causes treaty shopping, scrambling to find a way to tame the tax beast it has contributed to unleashing. Viewed this way, only the objections of states with a worldwide tax regime, such as the United States, appear legitimate.

What should Canada do? The best course of action would be to enact its treaty withholding tax rate policy, which is now 5 percent, directly into its domestic tax system.51 This would essentially amount to a neutral approach. If the withholding tax rates were lowered directly by the ITA, no treaty shopping would occur since every nonresident, whether from a treaty state or non-treaty state, would be entitled to the same 5 percent withholding tax rate. In essence the result would be akin to the environment that prevailed before the 1972 tax reform, when Canada’s treaty rate policy was equivalent to its domestic withholding tax rate of 15 percent.

At the time, lawmakers made the following comment about this long-standing Canadian approach: “While this neutral approach is commendable, it leaves Canadian negotiators at a distinct disadvantage at the treaty table, when trying to obtain treaty advantage for Canadian businessmen investing abroad.”52 If the principal justification for higher domestic withholding tax rates is solely about bargaining power in tax treaty negotiations, then, given that Canada has 93 tax treaties in force under which treaty advantages for Canadian residents have already been obtained, there is arguably no longer any meaningful explanation for Canada not to include its treaty withholding tax rates directly in its domestic law.

In any event, Canada already has the tool it needs in its domestic law to address treaty shopping: GAAR. Even though the Canada Revenue Agency’s argument based on GAAR has not yet succeeded in court, GAAR is still a relevant tool to address treaty shopping. In fact, the Supreme Court of Canada has not had a chance to present its view on whether GAAR applies to treaty-shopping arrangements. Both the Tax Court of Canada and the Federal Court of Appeal held in obiter that GAAR did not apply in Fundy Settlement v. Canada 53because they found no abuse of the Barbados-Canada tax treaty. However, the Supreme Court specifically stated, regarding the application of GAAR, that it “should not be understood as endorsing the reasons of the Federal Court of Appeal on those matters.” Thus, the Court seems to have left the door open for the application of GAAR to tax treaties.

This is exactly what the CRA is trying to do, as illustrated in the context of non-CIV funds by its ongoing litigation with Alta Energy Luxembourg SARL (Alta Luxembourg).54 Although not yet heard by the Tax Court of Canada, this case highlights the CRA’s intention to use GAAR to combat treaty-shopping arrangements.55 In this context, one wonders whether it was necessary for Canada to sign the MLI, at least when it comes to anti-treaty-shopping measures, if it thinks it can successfully police treaty shopping on the basis of its domestic-law-based GAAR regardless. It may still be worthwhile, if only to make GAAR’s application to tax treaties even more explicit.

In summary, discussions about treaty shopping cannot ignore the tax treaty policy of the source state. This is made evident by the contrast between the United States’ and Canada’s respective tax treaty policies and the purported mischiefs wrought by treaty shopping in each nation. When tax treaty policy is considered, it seems clear that addressing treaty-shopping arrangements is of critical importance to regimes like the United States’, which, unable to largely eliminate withholding taxes in a uniform fashion, give rise to varied rates.

The same, however, cannot be said of the latter, more uniform, regime. Yet the action 6 report fails to acknowledge this basic distinction, instead simply assuming that treaty shopping is an evil in need of remedy. In so doing, the OECD is guilty of a rush to judgment, assuming the problem is the same across the board, identifying a one-size-fits-all fix, and only inviting member states to fine-tune that remedy.

The more practical question, however, is: What do we do now? While a different, more refined approach would have been preferable, the OECD and member states are moving forward with the action 6 report and states have begun committing to the MLI. The next section formulates recommendations as to how Canada and the United States should proceed.

IV. The Way Forward

The OECD essentially proposes two tools for addressing treaty shopping: a PPT provision and an LOB provision. The action 6 report establishes the OECD’s minimum standard for fighting treaty shopping, requiring states to include in their treaties:

  • a combined approach of LOB and PPT provisions;

  • a PPT provision alone; or

  • an LOB provision supplemented by a mechanism to address conduit financing arrangements not already dealt with in tax treaties.

The minimum standard was then refined in the MLI. The major development from the MLI is that the PPT provision is now presented as the default option.56 States can supplement the PPT provision by applying a simplified LOB provision.

On June 7, Canada joined 67 other jurisdictions and signed the MLI. As detailed in a backgrounder released on the day of the signing ceremony,57 Canada opted for the default PPT. This is not surprising. Not only was the proposed domestic anti-treaty-shopping rule in the 2014 budget based on a PPT provision, all of Canada’s most recent tax treaties include a PPT provision.58

However, the backgrounder also notes that Canada will negotiate a bilateral, detailed LOB when appropriate. The history of the Canada-U.S. tax treaty LOB provision suggested the likelihood that Canada would opt for a dual approach. When it first took effect in 1995 with the signing of the third protocol, the LOB provision applied only to the United States as Canada felt that GAAR was sufficient to address treaty shopping. This state of affairs, however, was revised with the fifth protocol in 2007, when Canada had a change of heart and the LOB provision was made reciprocal.59 Given the dual approach, this section analyzes both types of provisions regarding Canada.

The United States, for its part, has clearly stated that it has no interest in a PPT provision,60and it has not signed the MLI.61 As the primary force behind LOB provisions, it is not surprising to see the United States going all-in on the LOB approach, taking the position that it “adequately covers the purposes contemplated for the [MLI] by countries worried about treaty shopping.”62 For this reason, the recommendations regarding the United States in this section address only the LOB provision.

A. The LOB Provision

The crux of the analysis of a non-CIV fund under the LOB provision turns on the derivative benefit test. Consider a private equity fund organized as a partnership formed in the Cayman Islands. The investors in the fund are state-owned entities, pension funds, university endowments, and taxable persons. Some investors are from treaty states and others are from non-treaty states. The private equity fund interposes a Luxembourg blocker corporation to acquire, finance, and hold a Canadian or U.S. target. All payments from the target to the Luxembourg blocker corporation are transferred to the private equity fund and ultimately to the investors. In this case, the question is whether payments from the Canadian or U.S. target are entitled to reduced withholding tax rates under the Canada-Luxembourg or Luxembourg-U.S. tax treaties. The Luxembourg blocker corporation usually fails to satisfy most tests under the LOB provision, leaving the derivative benefit test as its last chance to obtain treaty benefits. This section will therefore discuss only the derivative benefit test.63

In general terms, the derivative benefit test in the U.S. model tax treaty provides that a corporation resident in a treaty state is entitled to treaty benefits if:

  • it is at least 95 percent owned, directly or indirectly, by seven or fewer persons who would be entitled to similar treaty benefits were they to receive the payment directly (equivalent beneficiaries); and

  • less than 50 percent of the corporation’s income is paid, directly or indirectly, to persons who are not equivalent beneficiaries in the form of deductible payment in the corporation’s state (base erosion test).

This is similar to the detailed LOB provision set out in the action 6 report. This derivative benefit test, however, is problematic for non-CIV funds.

First, the 95 percent test essentially precludes non-treaty investors from investing in non-CIV funds. While this makes sense if treaty benefits are either granted or denied altogether, it is not clear what policy is served by this all-or-nothing approach. Second, the prohibition on more than seven equivalent beneficiaries will often preclude the application of tax treaties to non-CIV funds, especially in the case of widely held funds.64 For example, if the above-mentioned private equity fund had eight investors, the Luxembourg blocker corporation would not be entitled to the reduced treaty rate on payments from the Canadian or U.S target under this version of the derivative benefit test, even if all of the investors are equivalent beneficiaries. However, there does not appear to be any significance — economic or otherwise — for a bright-line limit of seven equivalent beneficiaries.65 This problem is particularly acute in the case of pooled investment vehicles such as non-CIV funds that, by definition, may well have more than seven equivalent beneficiaries. What, in this context, should the United States do? Fortunately, the OECD provides part of the answer in its treatment of the LOB provision in the MLI.

The MLI sets out a simplified LOB provision, which in turn provides a simplified derivative benefit test. This version of the derivative benefit test reduces the ownership threshold to 75 percent and removes both the seven or fewer equivalent beneficiaries requirement and the base erosion test. Thus, a resident of a treaty state is entitled to treaty benefits under the MLI’s derivative benefits test if it is more than 75 percent owned, directly or indirectly, by equivalent beneficiaries.66

This is a welcome development. Indeed, as noted by the OECD, “The derivative benefits provision included in [the MLI], which does not include any limit on the number of equivalent beneficiaries, any base-erosion test or any requirement related to intermediate owners, mean[s] that few non-CIV funds would fail to qualify for treaty benefits under that rule.”67 Contrary to the OECD’s view, however, the refinement of the derivative benefit test is not complete.

The remaining issue with the simplified derivative benefit test is that determining whether non-CIV funds qualify for treaty benefits need not be an all-or-nothing proposition. As the test stands, if non-CIV funds fail the test, then none of their income qualifies for treaty benefits.68 Again, consider the example of the private equity fund investing through a Luxembourg blocker corporation. If only 70 percent of the private equity fund’s investors are equivalent beneficiaries, then the reduced treaty rate does not apply to any payments made by the Canadian or U.S. target to the Luxembourg blocker corporation. This seems to be a strange result considering that 70 percent of investors would have been entitled to treaty benefits if they invested directly in the Canadian or U.S. target.69 In fact, this outcome is even more unfortunate in the context of non-CIV funds — especially in the private equity field — since investors are usually sophisticated and have the means to invest directly. Choosing to invest in non-CIV funds should not come at the cost of treaty benefits.

Instead, if the 75 percent threshold of the simplified derivative benefit test is not met, then non-CIV funds should at least be able to claim treaty benefits for the proportion of interests in their fund that is owned by equivalent beneficiaries.70 Under this proportionate benefit approach, non-CIV funds could claim a blended withholding tax rate based on the rates under the treaties between the source state and the investors’ residence states.71 Using the private equity fund example, if, out of the 70 percent of investors entitled to treaty benefits, half would have been entitled to a 10 percent rate while the other half would have been exempted, then payments to the Luxembourg blocker corporation should be subject to a withholding tax rate of 11 percent if made by the Canadian target (30% × 25% + 35%  × 10% + 35%  × 0%), or 12.5 percent if made by the U.S. target (30%  × 30% + 35%  × 10% + 35% × 0%). In both cases, the blended rate takes into account the 30 percent of investors that would remain subject to the Canadian or U.S. unreduced domestic withholding tax of, respectively, 25 percent and 30 percent.

Although source states might be concerned that treaty benefits granted to non-CIV funds under this proportionate benefit approach would indirectly inure to investors who are not equivalent beneficiaries, this appears unlikely for commercial reasons. Investors who qualify as equivalent beneficiaries are equally interested in ensuring that each investor receives only the benefits to which it is entitled.72 This could be enforced contractually through the constitutive documents of non-CIV funds or through partnership agreements.73Market forces would thus safeguard the integrity of the proportionate benefit approach. Furthermore, identifying the investors should not be a hurdle considering that non-CIV funds already identify and gather substantial information about their investors for the Foreign Account Tax Compliance Act and common reporting standards.74 In any event, this approach would encourage non-CIV funds to maintain verifiable records of their investors and, therefore, could be seen as the regulatory burden non-CIV funds must bear to avail themselves of the proportional benefit approach.

A proportionate benefit approach is consistent with anti-treaty-shopping measures and serves the same function as the derivative benefit test. If the use of an intermediate entity — the blocker corporation — by a non-CIV fund has not produced greater benefits than would otherwise be available to the fund’s investors, there is arguably no treaty shopping and no reason for concern.75 Since investors not entitled to treaty benefits have nothing to gain from the interposition of a blocker corporation in a treaty state, Canada and, more particularly, the United States need not fear that granting proportionate treaty benefits will create a disincentive for the investors’ state of residence to negotiate a tax treaty.76 Actually, by mirroring the tax treatment non-CIV funds’ investors would be entitled to through a direct investment, the proportionate benefit approach appears to be even more in line with the spirit of anti-treaty-shopping rules than the simplified derivative benefit test, which essentially allows up to 25 percent of investors to engage in treaty shopping.

Moreover, the proportionate benefit approach is consistent with the general precepts underlying the treaty-shopping discussion. First, by removing the blocker corporation from the analysis, the proportionate benefit approach (like the derivative benefit test) fully aligns with the notion suggested by the New York State Bar Association Tax Section that the “United States does not generally view treaty abuse as an issue of ‘substance,’ in the sense of an entity having a nexus with and reason to be in a particular country.”77 Second, the OECD generally assumes that the grant or denial of treaty benefits to an entity should depend on where its ultimate investors are resident.78 Specifically, the action 6 report places treaty shopping under the rubric of “Cases where a person tries to circumvent limitations provided by the treaty itself.”79 Having accorded determinative importance to the place where the ultimate investors of an entity resident in a treaty state are themselves resident, the OECD should follow the logical and equitable consequences of that approach, which entails “being open to creative approaches that would allow entities to claim treaty benefits to the extent that their ultimate investors would have qualified for the same or similar benefits.”80 This is precisely what the proportionate benefit approach seeks to achieve.

Therefore, if Canada and the United States choose the LOB route, they should do so by adopting an MLI-type simplified derivative benefit test coupled with the proportionate benefit approach.

B. The PPT Provision

The PPT provision of article 7(1) MLI stipulates that:

a benefit under [a tax treaty] 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 [tax treaty].

Based on comments from the OECD on the PPT generally and on its application to non-CIV funds, the goal of the PPT provision in this context appears to be twofold: preventing treaty shopping and achieving an acceptable level of certainty.81

1. Preventing Treaty Shopping

The scope of the PPT provision is broad and subjective. Accordingly, it would give tax authorities the ability to deny treaty benefits for practically any cross-border investment.82The breadth of the rule itself gives insight into the OECD’s concern — a well-founded one — about specific antiavoidance rules that can sometimes look more like detailed roadmaps for tax planners than effective tools for combatting undesired tax planning. Thus, it is not surprising to observe the growing deliberate enactment of broad and vague rules. The PPT provision clearly fits this trend. Indeed, it is clearly beneficial for the Canadian tax authorities to enjoy what commentators describe as “discretion to determine on an ad hocbasis what is improper treaty shopping and what is legitimate tax planning.”83 There is no doubt that the PPT provision would prevent treaty shopping. The question, however, is whether a broad PPT provision provides sufficient certainty to taxpayers.

2. Achieving an Acceptable Level of Certainty

When considering the level of certainty that the PPT provision would provide, it is helpful to consider rulings from Canadian courts involving domestic laws that use similar principal purpose tests.

First, the intent contemplated by the PPT provision has to be only one of the principal purposes for undertaking the transaction, not the principal purpose. Interpreting a similar test, the Federal Court of Appeal in Groupe Honco Inc. v. Canada 84 confirmed that “the phrase ‘one of the main purposes’ is unambiguous and implies that a taxpayer may have more than one main motive.” Determining whether the intention threshold is satisfied to trigger the application of the PPT provision appears to be a potentially onerous task. This undermines the level of certainty provided to non-CIV funds.

Second, experience with section 94.1 of Canada’s ITA may help reveal the likely implications of the PPT provision. Both are vague, broad antiavoidance provisions based on a “one of the main purposes”85 test. Interestingly, despite its vagueness, section 94.1 has generated very few disputes. While section 94.1 has been in force since 1985, Walton v. Canada 86 and Gerbro Holdings Company v. Canada 87 are the only two reported cases. At first blush, the section 94.1 experience could suggest that the PPT provision might not generate much uncertainty after all. But there is more to consider.

The problem with the principal purpose test of the PPT provision is that there is no indication about when a purpose becomes a main purpose. For instance, in Gerbro the Tax Court of Canada held that tax was one of the secondary reasons rather than one of the mainreasons for the investment. However, the court did not provide clear guidelines for distinguishing the two, emphasizing that “the line between a main reason and a secondary reason is difficult to draw . . . [and] is entirely factual.” The lack of guidance establishing the tipping point is not a problem limited to Canada. The U.S. conduit financing regulations, which specifically address treaty-shopping arrangements, are silent on this point.88Likewise, the two illustrating examples in the conduit financing regulations merely assume a principal purpose without discussion.89 It follows that, because of the highly fact-sensitive nature of the PPT provision, disputes between non-CIV funds and the CRA will arise. Viewed this way, the lack of reported cases under section 94.1 could be the result of taxpayers opting to settle with the CRA given the uncertainty surrounding the “one of the main purposes” test rather than because section 94.1 is actually workable.90 Ultimately, there will be uncertainty about the scope of the PPT provision until Canadian courts rule on it. But that could take a long time if non-CIV funds always end up settling with the CRA, thereby compounding the lack of judicial guidance. What a vicious tax uncertainty cycle this would make!

The OECD is not oblivious to those concerns. It proposes including three safe-harbor examples specific to non-CIV funds with the commentary to the PPT provision.91 Generally, the non-CIV funds industry has responded that, despite the examples, the line is still unclear as to when a non-CIV fund will be considered to have had access to treaty benefits as one of its principal purposes.92 That the PPT provision remains fraught with uncertainty notwithstanding the safe-harbor examples should not come as a surprise. This is not, however, mainly because of the OECD’s interest in retaining discretion regarding what constitutes objectionable treaty shopping and what constitutes legitimate tax planning. Rather, the OECD is victim of the same helplessness that precluded the Tax Court of Canada in Gerbro and the U.S. Treasury about the conduit financing regulations from satisfactorily defining the “one of the principal purposes” test — it is squarely impossible to do so.93

The inescapable conclusion is that the PPT provision would not achieve an acceptable level of certainty.

My concern is that the PPT provision would induce a chilling effect on non-CIV funds, leading investors to opt out of particular investments in Canada. This was judged to be the case with the proposed domestic anti-treaty-shopping rule in 2014 and will certainly be the case with the PPT provision.94 This outcome would be patently contrary to the overarching objective of tax treaties, which Canada acknowledged in Budget 2014 is “to encourage trade and investment.” Protecting the Canadian tax base and the integrity of the tax treaty network is laudable, but in so doing Canada would be well advised to remember the fundamental objectives of tax treaties and refrain, out of an unbounded desire to curb treaty shopping, from throwing out the baby with the bathwater. That being said, while I believe that it would be preferable for Canada to drop the anti-treaty-shopping initiative altogether, all is not lost for the PPT provision.

3. Saving the PPT Provision

There is a simple solution that would provide more certainty while retaining the flexibility of the PPT provision: a policy abuse test. In effect, this would turn the PPT provision into a Canadian GAAR-like provision.

In Canada three requirements must be met for GAAR to apply.95 First, there must be a tax benefit.96 Any reduction of tax qualifies as a tax benefit.97 Second, there must be an avoidance transaction.98 An avoidance transaction is any transaction that results, directly or indirectly, in a tax benefit, unless the transaction may reasonably be considered to have been undertaken primarily for nontax purposes. Just like the PPT provision, the avoidance transaction requirement catches only arrangements for which tax is a main purpose.99Canada recognizes as much, with the DOF stating in a 2013 consultation paper that “the factual determination required under a main purpose test is similar to that required to make an ‘avoidance transaction’ determination under the GAAR.”100 Contrary to the PPT provision, however, GAAR does have a third — and crucial — requirement: There must be an abuse of the underlying policy of a provision of the ITA.101

Abusive tax avoidance exists when a transaction achieves an outcome that a statutory provision was intended to prevent, thereby defeating the underlying legislative rationale of the provision.102 For a transaction resulting in a tax benefit to be denied under GAAR, not only must the transaction have tax as a main purpose, but it must also abuse the underlying policy of a provision of the ITA. The importance of the abuse test cannot be overstated; it is the abuse test that allows GAAR to achieve an acceptable level of certainty. This was acknowledged by the Supreme Court in Copthorne Holdings Ltd. v. Canada 103 when it held that while GAAR “has created an unavoidable level of uncertainty for taxpayers . . . [t]his uncertainty underlines the obligation of the [CRA] who wishes to overcome the countervailing obligations of consistency and predictability to demonstrate clearly the abuse.” Similarly, by adding an abuse test to the PPT provision, its application would be restricted to abusive treaty-shopping arrangements and certainty would correspondingly be increased.

Importantly, the abuse must target a specific tax policy. In the treaty-shopping context, this means that a treaty-shopping arrangement would have to abuse the source state’s specific treaty policy. The potential effect of supplementing the PPT provision with an abuse test is best illustrated by the treaty-shopping case Prévost Car Inc. v. Canada.104 In Prévost Car, a Canadian corporation distributed dividends to its Dutch parent corporation, which in turn distributed all dividends received to its own parents, a Swedish corporation and a U.K. corporation. At the time, the Canada-Netherlands tax treaty provided for a 5 percent withholding tax rate on this type of dividend, while the Canada-Sweden tax treaty provided a 15 percent rate and the Canada-U.K. tax treaty provided a 10 percent rate. The court held that the Canada-Netherlands tax treaty applied because the Dutch corporation was the beneficial owner of the dividends and not, contrary to the CRA’s position, merely acting as conduit for its parent corporations.

Prévost Car may be relevant to non-CIV funds because it describes a structure that, in abstract terms, is not dissimilar to a typical investment structure under which a non-CIV fund invests in Canada through a treaty-based holding corporation.105 The private equity fund example presented above (subsection IV(A)) as well as the ongoing dispute between Alta Luxembourg and the CRA (discussed in subsection III(B)) exemplify this point.

How would Prévost Car turn out under a PPT provision? Because it was acknowledged before the Tax Court of Canada that tax was a relevant consideration for choosing a Dutch holding corporation,106 it appears that the standard PPT provision would deny the benefits under the Canada-Netherlands tax treaty. In fact, Canada reached this very conclusion under the domestic anti-treaty-shopping rule proposed in the 2014 budget. However, that would not be a sound result as far as Canadian tax treaty policy is concerned.

Prévost Car took place in the mid-1990s, just after Canada had changed its tax treaty policy and reduced intercorporate dividend tax rates to 5 percent. While the Canada-Netherlands tax treaty provided the 5 percent rate, the Canada-Sweden tax treaty and Canada-U.K. tax treaty had not yet been renegotiated. Later, the 5 percent rate was also introduced in these tax treaties.107 Viewed this way, the structure in Prévost Car merely amounted to self-help and, even to the extent that it could be considered treaty shopping, was not abusive per se.108

What would happen in Prévost Car if the PPT provision were supplemented by an abuse test? As previously stated, it appears that one of the principal purposes for using the Dutch corporation was to obtain the reduced withholding tax rate under the Canada-Netherlands treaty. But the addition of an abuse test would nevertheless allow the Canada-Netherlands tax treaty to apply because there was no abuse of Canadian tax treaty policy. In fact, the DOF’s position that the structure in Prévost Car did not abuse Canadian tax treaty policy likely explains why the CRA did not rely on GAAR.109

The problem with the PPT provision is that the OECD does not really want the proposed rule to apply to any tax-motivated transaction. Indeed, the action 6 report explicitly provides that, “as regards the broader question of the treaty entitlement of non-CIV funds, the OECD recognises the economic importance of these funds and the need to ensure that treaty benefits be granted where appropriate.”110 This is exactly what an abuse test seeks to achieve. Furthermore, the abuse test would provide a more objective basis for testing entitlement to treaty benefits than the vague discretion given to the tax authorities under the PPT provision.

Supplementing the PPT provision with an abuse test provides an elegant solution to the quagmire of uncertainty in which the PPT provision now appears perpetually trapped. While this approach would remain more subjective than the LOB provision, non-CIV funds would at least be provided with an acceptable level of certainty. Importantly, this approach also highlights the fact that a state’s tax treaty policy matters a great deal in treaty-shopping discussions. Accordingly, since Canada is moving forward with the PPT provision, it should do so by supplementing it with an abuse test.


V. Conclusion

Treaty shopping isn’t a simple issue. Neither is responding to treaty shopping. It involves a variety of considerations ranging from protecting source states’ tax base to ensuring the effectiveness of tax treaties in eliminating tax-induced gratuitous impediments to international commerce. The central proposition of this article is that in crafting an appropriate response to treaty shopping, the specific tax treaty policy of a state matters. It matters so much that when properly accounted for, it appears that a state like Canada should not even be seriously concerned about treaty shopping.

The OECD has embarked on its treaty-shopping crusade without even considering the first logical step that the endeavor demands. By turning a blind eye to source states’ tax treaty policies, the OECD’s initiative was doomed from the outset. The one-size-fits-all solution imbedded in the MLI, which may very well work for some states but definitely not for all, is the consecration of this flawed process. It may be judicious to introduce internationally coordinated anti-treaty-shopping measures, but to do so while asking stakeholders like non-CIV funds to fine-tune the predetermined fix is not the solution. Just like Alice in Wonderland, who did not much care which way to go as long as she got somewhere, it appears that the OECD did not much care where its treaty-shopping project would go as long as it went somewhere. The problem, however, is that living in Tax Wonderland hardly makes for sound international tax treaty policy.


1 “Briefing: Private Equity — The Barbarian Establishment,” The Economist, Oct. 22, 2016 (“It is not just that old private-equity firms persist; new ones continue to spring up at a remarkable rate. . . . [T]here were 24 private-equity firms in 1980. In 2015 there were 6,628, of which 620 were founded that year.”).

2 OECD, “BEPS Action 6 Final Report: Preventing the Granting of Treaty Benefits in Inappropriate Circumstances,” 17 (2015).

3 H. David Rosenbloom, “Derivative Benefits: Emerging US Treaty Policy,” 2 Intertax 83 (1994).

4 See, e.g., action 6 report, supra note 2, at 99 n. 2; and Robert Couzin, “A Few Thoughts on Treaty Shopping,” 61 Canadian Tax J. 671, 672 (2013).

5 OECD, “The Granting of Treaty Benefits With Respect to the Income of Collective Investment Vehicles,” 3 (Apr. 23, 2010); and OECD, “Public Discussion Draft: Treaty Entitlement of Non-CIV Funds,” 3 (Mar. 24, 2016).

6 Income Tax Act (Canada), section 96; and IRC section 701.

7 Comments of Canadian Pension Funds in OECD, “Comments Received on Public Discussion Draft: Treaty Entitlement of Non-CIV Funds,” 162 (Apr. 22, 2016); Comments of Cleary Gottlieb Steen & Hamilton LLP in OECD, “Comments Received on Public Discussion Draft: Treaty Entitlement of Non-CIV Funds,” id.; and Michael N. Kandev, “Canada Intent on Stoppin’ the Shoppin’ and More,” Tax Notes Int’l, Mar. 31, 2014, p. 1201.

8 Comments of Brian Ernewein in Patricia Brown et al., “Insiders’ View of Treaty Negotiations,” in Report of Proceedings of the Sixty-First Tax Conference of the Canadian Tax Foundation 25:1, 25:6 (2010).

9 David A. Ward, “Access to Tax Treaty Benefits: Research Report Prepared for the Advisory Panel on Canada’s System of International Taxation,” 45-48 (Sept. 2008).

10 Charles I. Kingson, “The Coherence of International Taxation,” 81(6) Colum. L. Rev. 1151, 1157 (Oct. 1981); and Phil Jolie, “Further Thoughts on Non-Resident Taxation,” Canadian Tax Foundation Open Forum Blog (Apr. 14, 2014).

11 Jolie, supra note 10.

12 Canadian Department of Finance, “Consultation Paper on Treaty Shopping: The Problem and Possible Solutions” (Aug. 12, 2013).

13  For the OECD’s work after the 2014 budget, see OECD, “Public Discussion Draft: BEPS Action 6: Preventing the Granting of Treaty Benefits in Inappropriate Circumstances” (Mar. 14, 2014); and OECD, “Preventing the Granting of Treaty Benefits in Inappropriate Circumstances: Action 6: 2014 Deliverable” (Sept. 16, 2014). On August 29, 2014, the Canadian DOF announced that the government would “await further work by the Organisation for Economic Co-operation and Development and the Group of 20 (G-20) in relation to their Base Erosion and Profit Shifting initiative” before implementing the proposed rule set out in the 2014 budget. Canadian DOF release (Aug. 29, 2014). The proposed rule was officially abandoned in the 2016 budgetSee also Government of Canada, “Budget 2016: Tax Measures: Supplementary Information,” 48 (Mar. 22, 2016).

14 Comments of Ernewein, supra note 8, at 25:9.

15 Id.

16 Ward, supra note 9, at 1.

17 Canadian Standing Committee on Finance, Trade and Economic Affairs, Eighteenth Report of the Standing Committee on Finance, Trade and Economic Affairs Respecting the White Paper on Tax Reform 84-85 (Oct. 1970).

18 This system now also includes states with which Canada has entered into a tax information exchange agreement. ITA section 113; and section 5907(11) of the Income Tax Regulations (Canada) (defining “designated treaty country”).

19 Comments of Ernewein, supra note 8, at 25:10.

20 R. Alan Short, “Purposes and Effects of Income Tax Treaties,” in Report of Proceedings of the Twenty-Fourth Tax Conference of the Canadian Tax Foundation 100, 103 (1973).

21 Indeed, as Donald J. Trump asserted in 2016, “Our trade deficit in goods reached nearly — think of this, think of this — our trade deficit is $800 billion last year alone.” See Ana Swanson, “Trump Criticizes the Trade Deficit, but Leaves Out an Important Figure,” The Washington Post, Feb. 28, 2017; and Neil Irwin, “What Donald Trump Doesn’t Understand About the Trade Deficit,” The New York Times, July 21, 2016.

22 Comments of Brown in Brown et al., “Insiders’ View,” supra note 8, at 25:2.

23 U.S. model tax treaty (2016), articles 10-12. While the lowest withholding tax rate on dividends under the U.S. model tax treaty is 5 percent, the United States has granted a 0 percent rate to several of its major treaty partners. See, e.g., U.K.-U.S. tax treaty, article 10(3); Germany-U.S. tax treaty, article 10(3); and Netherlands-U.S. tax treaty, article 10(3).

24 Comments of Rosenbloom in Brown et al., “Insiders’ View,” supra note 8, at 25:7-25:8.

25 Id. at 25:8.

26 Id. at 25:7-25:8.

27 Rosenbloom, “Tax Treaties Abuse: Policies and Issues,” 15 Law & Pol’y Int’l Bus. 763, 776 (1983).

28 Comments of Rosenbloom, supra note 24, at 25:8.

29 Rosenbloom, supra note 27, at 776.

30 Comments of Rosenbloom, supra note 24, at 25:8.

31 Id.

32 See supra note 23 and accompanying text.

33 ITA section 245 and section 4.1 of the Income Tax Conventions Interpretation Act were amended in 2005 to expressly provide that GAAR applies to tax treaty benefits.

34 Rosenbloom, supra note 27, at 776 n.36.

35 See infra note 108 and accompanying text.

36 Ward, supra note 9, at 28.

37 See Canadian DOF, “Tax Treaties: In Force” (accessed May 16, 2017).

38 Jack Bernstein, “Canada’s Position on the Multilateral Instrument and Its Attack on Treaty Shopping,” Tax Notes Int’l, Feb. 13, 2017, p. 649.

39 Jolie, supra note 10.

40 MIL (Investments) S.A. v. Canada, 2006 TCC 460, para. 73, aff’d, 2007 FCA 236.

41 See Canadian DOF, supra note 37.

42 At the time, the LOB provision was not reciprocal and only became so with the fifth protocol in 2007. See infra text associated with note 59.

43 Kandev, “Treaty-Shopping Consultation,” 21(5) Can. Tax Highlights 4, 6 (May 2013).

44 Richard Tremblay, Patrick Marley, and Diana Yeung, “Canada Considers New Anti-Treaty-Shopping Rule,” 62 Tax Mgm’t Int’l J. 621 (2013).

45 MIL (Investments) S.A. v. Canada, 2006 TCC 460, at para. 74.

46 Jolie, supra note 10.

47 Id.

48 Id.

49 See Kingson, supra note 10, at 1153.

50 Jolie, supra note 10.

51 See John F. Avery Jones, “The David R. Tillinghast Lecture: Are Tax Treaties Necessary?” 53 Tax L. Rev. 1, 4 (Fall 1999).

52 Canadian Standing Committee on Finance, Trade and Economic Affairs, supra note 17, at 85.

53 Fundy Settlement v. Canada, 2012 SCC 14.

54 See Bernstein, supra note 38, at 650.

55 Id.

56 Article 7(1)-(5) MLI; and OECD, “Explanatory Statement to the Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting,” para. 90 (Nov. 24, 2016).

57 Canadian DOF, “Backgrounder: Impact of Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting” (2017).

58 See, e.g., Canada-Israel tax treaty, articles 10-13 (signed Sept. 21, 2016); and Canada-Taiwan tax treaty, articles 10-12 (signed Jan. 15, 2016).

59 See Comments of Ernewein, supra note 8, at 25:18.

60 Lee A. Sheppard and Stephanie Soong Johnston, “U.S. ‘Extremely Disappointed’ in DPT and BEPS Output, Stack Says,” Tax Notes Int’l, June 15, 2015, p. 1005; and Tax Section of the New York State Bar Association, “Report No. 1331 Relating to the Proposed Revision to the Limitation on Benefits Article of the U.S. Model Convention,” 19 (Nov. 16, 2015).

61 This is particularly unsurprising given the current U.S. administration’s openly stated aversion for multilateral treaties in general. See, e.g., Daniel W. Drezner, “Why the Trump Administration Hates Multilateral Trade Agreements the Most,” The Washington Post, Mar. 9, 2017; and Max Fisher, “Trump Prepares Orders Aiming at Global Funding and Treaties,” The New York Times, Jan. 25, 2017.

62 Sheppard and Johnston, supra note 60.

63 This is consistent with the fact that the work undertaken by the OECD regarding the application of the LOB provision to non-CIV funds has also focused on the derivative benefit test. See “Public Discussion Draft,” supra note 5; and “Comments Received on Public Discussion Draft,” supra note 7.

64 Comments of BlackRock in “Comments Received on Public Discussion Draft: Treaty Entitlement of Non-CIV Funds,” supra note 7, at 101.

65 Tax Section of the New York State Bar Association, supra note 60, at 36-37.

66 Article 7(11), (13)(c) MLI.

67 OECD, “BEPS Action 6 Discussion Draft on Non-CIV Funds Examples” (Jan. 6, 2017) (hereafter “Public Discussion Draft (2017)”).

68 Comments of Cleary Gottlieb Steen & Hamilton LLP, supra note 7, at 182.

69 Comments of Canadian Pension Funds, supra note 7, at 164; Comments of Cleary Gottlieb Steen & Hamilton LLP, supra note 7.

70 Id.

71 Comments of Canadian Pension Funds, supra note 7.

72 Id.

73 Id.

74 See, e.g., Comments of Canadian Pension Funds, supra note 7, at 166; Comments of Cleary Gottlieb Steen & Hamilton LLP, supra note 7, at 186.

75 See Rosenbloom, supra note 3, at 84.

76 Id.

77 Tax Section of the New York State Bar Association, supra note 60, at 20.

78 Comments of Osler, Hoskin & Harcourt LLP, in “Comments Received on Public Discussion Draft: Treaty Entitlement of Non-CIV Funds,” supra note 7, at 421.

79 Action 6 report, supra note 2, at 17.

80 Comments of Osler, Hoskin & Harcourt LLP, supra note 78, at 421.

81 See, e.g., action 6 report, supra note 2; “Public Discussion Draft,” supra note 5, at 3; “Public Discussion Draft (2017),” supra note 67.

82 Comments of Osler, Hoskin & Harcourt LLP in OECD, “Comments Received on Public Discussion Draft: BEPS Action 6 — Examples on Treaty Entitlement of Non-CIV Funds Examples,” 180 (Feb. 3, 2017) (hereafter “Comments Received on Public Discussion Draft (2017)”).

83 Reuven S. Avi-Yonah and Christina H.J.I. Panayi, “Rethinking Treaty-Shopping: Lessons for the European Union,” U. of Michigan Public Law Working Paper No. 182 (Jan. 5, 2010), at 5.

84 Groupe Honco Inc. v. Canada, 2013 FCA 128.

85 Technically, section 94.1 uses the word “reasons” rather than “purposes.” There is, however, no meaningful distinction between the two for purposes of this article.

86 Walton v. Canada, [1999] 1 C.T.C. 2105.

87  Gerbro Holdings Company v. Canada , 2016 TCC 173.

88 U.S. Treas. reg. section 1.881-3(b).

89 U.S. Treas. reg. section 1.881-3(e), Ex. 12, 13.

90 See, e.g.Sohmer 2012-4058(IT)G (consent to judgment allowed prior to hearing).

91 “Public Discussion Draft (2017),” supra note 67.

92 See, e.g., Comments of BlackRock in “Comments Received on Public Discussion Draft (2017),” supra note 82, at 30; Comments of Canadian Pension Funds in “Comments Received on Public Discussion Draft (2017),” supra note 82, at 45; and Comments of Osler, Hoskin & Harcourt LLP (2017), supra note 82, at 178.

93 One is reminded of the U.S. tax law’s distinction between debt and equity and how fraught trying to distinguish between the two can be. For instance, the bifurcation rule in the proposed section 385 regulations did not make its way into the final version out of a concern that it would not be possible to distinguish between debt and equity components of some financial instruments. See Robert Stack, “The Proposed Section 385 Regulations: Brilliant Tax Policy or Tax Policy Misstep?” remarks at IFA/NYU Conference (Oct. 13, 2016) (“You will never be able to figure out the difference between debt and equity no matter how long you try because there is no one answer”). Unfortunately Canada and, more generally, the OECD do not appear prone to exercise the same wisdom. See also Nathan Boidman, “One of the Main Purposes’ Test,” 22 Canadian Tax Highlights 9, 10 (May 2014).

94 See, e.g., Mary Teresa Bitti, “Treaty Shopping Crackdown Chills Foreign Investment,” Financial Post, Apr. 30, 2014; and Theophilos Argitis, “Canada Puts Treaty-Shopping Tax Plan on Hold,” Bloomberg, Sept. 2, 2014.

95 ITA section 245.

96 ITA section 245(1).

97  Canada Trustco Mortgage Co. v. Canada , 2005 SCC 54, para. 19.

98 ITA section 245(3).

99 Canada Trustco Mortgage, 2005 SCC 54, at para. 21.

100 Canadian DOF, Consultation Paper, supra note 12, at para. 7.1.

101 ITA section 245(4).

102  Copthorne Holdings Ltd. v. Canada , 2011 SCC 63, para. 72.

103 Id.

104  Prévost Car Inc. v. Canada , 2009 FCA 57.

105 David Davachi et al., “Implications for Private Equity Funds of Canada’s Budget 2014 Proposed Anti-Treaty-Shopping Measures,” Osler, Hoskin & Harcourt LLP release, Feb. 13, 2014, at 3.

106 Prévost Car Inc. v. Canada, 2008 TCC 231, para. 9, aff’d, 2009 FCA 57.

107 See Canada-Sweden tax treaty, article 10(2)(a) (signed Aug. 27, 1996); and Canada-U.K. tax treaty, article 10(2)(a) (signed Sept. 8, 1978 — the relevant amendment was made through the protocol signed May 7, 2003).

108 Kandev, supra note 7, at 1204.

109 Specifically, a DOF official was quoted as follows:

I agree that less Canadian tax was paid by routing the payments through a country with a lower dividend withholding rate. In a perfect world we would have negotiated and signed new treaties with European countries to provide them with the 5 percent rate all on the same date. Under that scenario, Canada would have been happy in policy terms to collect 5 percent on payments to the UK and Sweden. Given that we were negotiating with Sweden and UK at this time and our apparent willingness to provide the 5 percent rate as a matter of policy in any new treaty, I think it will be difficult for a court to smell the nastiness of this scheme by two multinationals resident in treaty countries, to avail themselves of the policy rate. [Emphasis in original.]

See Ward, supra note 9, at 47.

110 Action 6 report, supra note 2, at 16.