MECHANISMS TO FIGHT TREATY ABUSE: AN OVERVIEW
Professor of Tax Law at School of Law of University of Minho, Braga, Portugal. Head of Team and Coordinator of Glob – (Globalisation, Democracy and Power), research group of JusGov (Research Centre for Justice and Governance). Director of the LL.M.in European and Transglobal Business Law. PhD in Law, Universidade Nova de Lisboa; LL.M., London School of Economics and Political Science.
DOI: http://dx.doi.org/10.46801/2595-7155-rdtia-n7-13
INTRODUCTION
The purpose of this brief article is to address the topic of treaty abuse, which has gained increased importance in the framework of the OECD/G20 Base Erosion and Profit Shifting Project (BEPS).
The author will start by briefly delimiting the concept of treaty abuse. Subsequently, we will be addressing the classical solutions to prevent treaty abuse. In the third and final part of this article the author will revisit the classical solutions in light of the latest developments to determine what has changed.
THE CONCEPT OF TREATY ABUSE
“Treaty abuse” and “the abuse of tax treaties” are alternative terms that can be used interchangeably to address the same situation - wherein a taxpayer, though complying with the wording of a given tax treaty provision(s), attempts to obtain advantages that are beyond the rationale of that or those provision(s).
Treaty shopping is one of the most important forms of treaty abuse and one of the most important sources of concern. Although treaty shopping is often implicit, throughout this analysis, the author will interpret the concept of abuse in its broadest sense.
Treaty shopping refers to any situation where a taxpayer who is not a resident of a State tries to obtain the benefits granted by a tax treaty concluded by that state, manipulating, for that purpose, the residence as a connecting factor1, by setting up, for example, a conduit or letter company in that State. In doing so, many taxpayers manage to take advantage of a treaty network that otherwise would not be available for them and obtain important tax
advantages. Examples include the reduction of withholding taxes in the source state of income, and exemption in the residence state.
ABUSE PREVENTION BEFORE BEPS
This article will not seek to analyse all of the mechanisms to prevent abuse, but solely the ones that occur more often, either because they have been suggested by several commentators or have already been adopted by tax treaties, in particular the OECD Model Convention and its commentary. The author will refer to: (i) the possibility of sustaining the existence of a general anti-abuse principle of international law, (ii) the application of domestic anti-avoidance rules to tax treaties, also and (iii) treaty provisions that prevent the abuse of tax treaties.
The general anti-abuse principle of international law
Some commentators have discussed the existence of a general anti-abuse principle in international law also applicable to tax treaties. Article 26 (pacta sunt servanda) of the Vienna Convention on the Law of Treaties would serve as a basis for such a principle2. There is no unanimity, though, on whether there is a principle in international law prohibiting abuse.
Authors such as Klaus Vogel and Ward believe that Article 26 of the Vienna Convention and the principle of good faith it entails give support to the argument that there is a ‘substance over form’ doctrine based on international law, which enables tax authorities and Courts to deny treaty benefits to taxpayers who engage in operations void of economic rationale, that is, aimed solely at obtaining treaty tax benefits3.
Luc De Broe, referring specifically to Vogel’s reasoning, notes that the stated position fails to distinguish between the abuse of treaties by contracting states and abuse by taxpayers operating on a domestic level. He stresses, in this context, that individuals that derive rights from the treaties are not parties to the treaty. He considers, however, that Article 26 should be interpreted in conjunction with Article 31 of the Vienna Convention, which, in his opinion, asserts that good faith cannot be restricted to the performance of the treaty but also applies to its interpretation. By focusing on good faith in the interpretation of tax treaties - which should be honest, fair, and reasonable to give proper effect to their object and purpose and to the common intentions of contracting parties and its direct consequences - De Broe finds a way of overcoming the obstacle set forth by the fact individuals are not parties to the treaties4.
De Broe concludes his analysis in a pragmatic way, stating that “…one can conclude to the existence of an anti-abuse doctrine in tax matters as general principle of international law independently from whether the abuse of rights doctrine exists as a general principle of international public law5”. In other words, De Broe suggests an interpretation principle within Article 31, according to which treaty benefits should not be granted in abusive cases.
Such a principle of interpretation faces two major challenges. First of all, it is not accepted universally. Many states have enacted different sorts of domestic anti-abuse provisions (specific and general), or have included different specific abuse provisions in the treaties themselves (or neither), showing that they do not rely on a putative general principle of interpretation in the same way. Moreover, even if they accepted such a principle, disparities in its application, from country to country or from treaty to treaty, would undermine that principle.
Despite the uncertainty, the OECD formulated a guiding principle of treaty interpretation in the 2003 revision of the OECD commentary of the OECDE model convention, which mirrors the interpretation principle based on Article 31 of the Vienna Convention. In fact, as § 9.5. of the OECD Commentary on Article 1 states:
A guiding principle is that the benefits of a convention should not be available where a main purpose for entering into certain transactions or arrangements was to secure a more favourable tax position and obtaining that more favourable treatment in these circumstances would be contrary to the object and purpose of the relevant provisions.
Despite the fact that the addition of a guiding principle was indeed a remarkable progress, some problems persist.
First, this principle was only introduced in 2003. Therefore, even though in the view of the OECD the most recent version of the commentary applies to treaties concluded before that date (the dynamic approach), commentary is not binding and Courts often express different views.
Second, the fact that such guiding principle is not embodied in the OECD Model Convention itself, but simply incorporated in the commentary, can generate a problem that is very similar to the one that affects the interpretation principle derived from Article 31 of the Vienna Convention, which the author analysed before: namely, the fact that States have heterogeneous practices in regard to that principle.
Domestic anti-avoidance rules
Resorting to domestic anti-avoidance provisions and doctrines can also be a mechanism against treaty abuse.
The OECD Model Convention, in its 1977 commentary6, started by implying that domestic anti-abuse measures could not be applied to tax treaties. Later, in 1992, the revised commentary, although confirming indirectly the previous implied rule (prohibition of application of domestic anti-abuse rules), added new paragraphs7 stating that domestic anti-abuse rules should be specifically confirmed in tax treaties8, which represents a major change.
In 2003, the commentary is radically changed and seems to clarify the relationship between domestic anti-abuse provisions and tax treaties. One of the most significant changes was the deletion of the final part of §7 of the 1992 version of the commentary on Article 1, eliminating the rule that implied that domestic anti-abuse rules had to be specifically confirmed in tax treaties. Thus, as a general rule, there will be no conflict between such rules and the provisions of tax conventions9. This position is reaffirmed when, in the same commentary, it is stated that «the potential application of [domestic] general anti-abuse provisions does not mean that there is no need for the inclusion, in tax conventions, of specific provisions aimed at preventing particular forms of tax avoidance»10.
The relationship between domestic anti-abuse provisions and tax treaties because, even in light of the 2003 addition to the commentary, is not clear-cut. As a result, a very clear divide should be made between strictly internal anti-abuse provisions, which arguably can be referred to the allegedly general rule set out by the commentary – which asserts that there will be no conflict between domestic rules and the provisions of tax treaties – and those provisions having an international effect which should not be considered to be covered by that rule.
One has to acknowledge that it is very difficult to harmonise domestic anti-abuse rules that disallow the application of the treaty both with Article 27 of the Vienna Convention and Article 2(4) of the OECD Model Convention. Respecting these two articles is paramount to avoid treaty amendment through the back door, that is, by using anti-abuse domestic provisions that would apply automatically. Article 27 states that a party may not invoke the provisions of its internal law as justification for its failure to carry out a treaty. Article 2(4) of the OECD Model Convention asserts that “the competent authorities of the Contracting States shall notify each other of any significant changes that have been made
in their taxation laws”. This latest assertion is clarified by the commentary on that article in §8, which states that “Member Countries are encouraged to communicate other significant developments [besides significant changes made to its taxation11] as well such as new regulations…”12. Thus, the need to be consistent with either.
In this context it is interesting to note that countries such as Ireland, the Netherlands, Switzerland, and Luxembourg expressed reservations with respect to the 2003 commentary revisions - in the sense they do not accept that at least some anti-abuse rules are compatible with treaties13.
The arguments set out above lead to the conclusion that there seems to be no full agreement on whether domestic anti-avoidance rules and doctrines (judicially developed) apply to cross-border situations governed by tax treaties.
Treaty Provisions
In line with the OECD Model Convention very few treaties contain general anti-abuse provisions, specific anti-abuse provisions being much more prevalent14.
In this section the author will make a very brief reference to some of the provisions embodied in the OECD Model Convention under the heading of miscellaneous provisions and will subsequently summarize the concepts of beneficial owner and limitation on benefit clauses.
Miscellaneous Treaty Provisions
There are few provisions in the OECD Model Convention that aim at preventing treaty abuse. Examples of provisions of the OECD Model preventing treaty abuse are Article 4 (1) on the limitation of the condition of resident to taxpayers who are taxed on worldwide basis; Article 7 (2) and 9) (1) on profits adjustments in case of non arm’s length; Articles 11 (6) and 12 (4) on the special relationship between payor and payee; the beneficial owner concept of Articles 10, 11 and 12; Article 13 (4) on gains on shares in real estate companies; Article 17 (2) on the special artistes company15, amongst others.
Beneficial owner
The concept of ‘beneficial owner’16 originated in common law countries and, although it had already appeared in several treaties concluded before 1977, it was only introduced in the OECD Model Convention in that year. The purpose of that concept was to deal with treaty shopping situations where income was paid to an intermediary resident (such as an agent nominee or conduit), who was not treated as the owner of that income for tax purposes.
When the concept was introduced in the OECD Model Convention it was meant to clarify the term ‘paid to’ embodied in Articles 10, 11 and 12 of that Model Convention. However, in the 1997 Model Convention version of the commentary the term was not defined clearly in a clear way. The commentary only defined the concept in a negative way, saying that benefits were not available when an intermediary, such as an agent, or nominee was interposed between the beneficiary and the payer.
In 2003, the commentary was again changed and took on some of the ideas conveyed in the 1986 Conduit Companies Report, making further clarifications to the concept17. The concept became broader, encompassing, together with agents and nominees, entities that simply act as a conduit for another person who in fact receives the benefit of the income concerned18.
Another important 2003 addition to the commentary on Article 10 – perhaps the most significant – was §12, which states:
the term ‘beneficial owner’ is not used in a narrow technical sense, rather, it should be understood in its context in light of the object and purpose of the Convention, including avoiding double taxation and the prevention of fiscal evasion and avoidance.
Ambiguity, however, persisted in the sense that views regarding the formulation of the concept kept on being diverse.
For some authors19, based on Article 3 (2) of the OECD Model Convention, the concept should be interpreted according to the domestic law meaning of beneficial owner.
For others, there should be an independent interpretation based strictly on international tax law20.
Two other alternatives have also been discussed by commentators and followed by inconsistent court decisions21:
First, a strictly legal one, according to which the only relevant substance of an operation is the one that results from the parties’ legal rights and obligations. This notion implies that the beneficial owner concept does not apply to situations where the intermediary transfers the income to which he is legally entitled to others. It also seeks to accommodate the need for going beyond a ‘narrow technical sense’ of the concept by looking to the purpose of the operation, but does not detach itself from the strict legal sense.
Second, there is a more far-reaching approach that considers the underlying economic reality of the transaction rather than just the legal reality. According to this view, the concept of beneficial owner also applies to those situations where a certain taxpayer is involved in an operation void of economic substance.
The 2014 commentary revision allowed for the clarification of some issues, but left others immersed in uncertainty. In relation to whether the concept of beneficial ownership should or should not be interpreted according to the domestic law meaning, § 12.1 of the commentary to Article 10 of the OECD Model Convention closes the issue by stating clearly that:
the term ‘beneficial owner’ was added to address potential difficulties arising from the use of the words ‘paid to …a resident’ in paragraph 1, it was intended to be interpreted in this context and not to refer to any technical meaning that it could have had under the domestic law of a specific country…The term ‘beneficial owner is therefore not used in a narrow technical sense (such as the meaning that is has under the trust law of many common law countries…)
In relation to the choice between a strictly legal or economic interpretation, uncertainty persists. On the one hand, it seems that a substance over form interpretation based on the economic reality should be also considered when the 2014 version of the commentary to Article 10 states in § 12.4 that:
Such an obligation [contractual obligation the recipient has to pass on the payment received to another person which constrains its right of enjoying the dividend and
therefore being considered a beneficial owner] will normally derive from relevant legal documents but may also be found to exist on the basis of facts and circumstances showing that, in substance, the recipient clearly does not have the right to use and enjoy the dividend unconstrained by a contractual or legal obligation to pass on the payment received to another person.
Even though the commentary sends conflicting signs22, making a clear distinction between the legal and economic approach is a daunting task, since legal reality and economic reality are intertwined. Nevertheless, it is important to limit the economic approach, since it often relies on subjective facts, which are extremely difficult to prove.
Limitation on benefit provisions
According to Limitation on Benefits (LOB)23 provisions, treaty benefits related to investment income (dividends, interest and royalties) are only granted if the recipient of that income is a bona fide resident of the other state, that is, a person who did not become resident with the main intention of claiming treaty benefits. Thus, it also disqualifies those taxpayers who, although formally resident in a certain country, have income that is not sufficiently subject to taxation in that country to benefit from the extension of the treaty reductions of source-based taxation.
In other words, this provision aims to set aside subjectivity when determining whether taxpayer has engaged in treaty shopping. This is achieved by setting forward a series of objective tests aimed at determining whether the taxpayer has a real business purpose, or whether the connection it has to the other state is strong enough to allow treaty benefits, even in the absence of a business purpose.
Although these provisions have advantages over vague poorly defined treaty concepts, such as ‘beneficial owner’, and reduce uncertainty, they are extremely technical and complex and therefore difficult to negotiate and administer by tax authorities. This reason explains their absence from most treaties, even after 2003, when, through the new version of the OECD Model Convention commentary to Article 1, an express reference to LOB, in § 20, was included. Only the treaties conclude by the United States and very few other countries (Japan and India) have included such provisions.
TREATY ABUSE AFTER CHANGES TRIGGERED BY OECD/G20 BASE EROSION AND PROFIT SHIFTING ACTION PLAN
Action 6 of the BEPS Action Plan identifies abuse, in particular treaty shopping, as one of the most important concerns. The report on this action was published in October of 201524 and contains several proposals that relate directly to the issues addressed in each of the previous sections. The report consists of three parts: A) Treaty provisions and/or domestic rules to prevent the granting of treaty benefits in inappropriate circumstances; B) Clarification that tax treaties are not intended to be used to generate double non-taxation; and C) Tax policy considerations that, in general, countries should consider before deciding to enter into a tax treaty with another country.
We will focus on Part A that contains several proposals that relate directly to the content of each of the previous subsections and brought significant modifications to the OECD Model Conventions, as can be noted by analysing the 2017 version of the Convention. Instead of separately addressing Part A of the report, we will revisit the theme of each of the sections and try to determine if some of the problems identified in connection with each of those subsections are addressed and eventually solved in the context of Action 6: 2015 Final Report.
We will start by looking at what I call a general anti-abuse principle of international law. Subsequently, domestic anti-avoidance rules will be addressed and finally treaty provisions will be considered.
General anti-abuse principle of international law after BEPS Action 6 embodied in the 2017 version of the OECD Model Convention25
It is precisely in the context of the general anti-abuse principle of international law that the progress is greater. As the reader may remember, one of the main drawbacks of the general anti-abuse principle was the fact that it was not embodied in the text of the model convention but simply in the commentary (it was a principle, not a rule). Subsequently to Action 6 Final report, a new article and its respective commentary was suggested to be included in the model convention. The new Article 29 (9) reads as follows:
Notwithstanding the other provisions of this Convention, a benefit under this Convention 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 this Convention.
The rule embodied in Article 29 (9) is addressed under BEPS action as a Principle Purpose Test (PPT)26. This provision is perfectly in line with the guiding principle mentioned before27 in the sense that it also includes the option for the “principal purposes” rather than “single purpose”, which, in the same way as the guiding principle does, allows tax authorities an easier application of the provision to disallow tax benefits.
Regardless of the fact that it does not bring any independent legal significance beyond what was already consubstantiated in the commentary and could also be derived from Article 31 (1) of the Vienna Convention, the addition of these new provisions has its merits.
First of all, for the sake of uniformity in the application of tax treaties, it is better to have a clear provision that states may want to include in the treaty rather than principles that are not always recognised or applied the same way, such as the guiding principle and the one that could be derived from Article 31 (1) of the Vienna Convention.
Second, these provisions will act as an alternative mechanism to fight tax abuse when others fail to do so. This was already recognised in the commentary28, as referred to above, in relation to the ‘beneficial owner’ test. Regarding LOB, the Action 6 final report also acknowledges that, considering that each rule has strengths and weaknesses, a combination of LOB and the Principle Purpose test is necessary, mainly when abuse is not covered by the former29.
Finally, the provision itself has a detailed and extensive commentary30 which would certainly be of use.
In spite of the advantages identified, one has to acknowledge that the PPT provision is no more than a general anti-abuse provision (GAAR) with a wide scope and is therefore invariably, in the author’s opinion, extremely difficult to put in practise. That provision seems to me more of a hint on how taxpayers should behave and a political sign that governments are willing to fight abuse rather than an effective anti-abuse mechanism. Admittedly, this position is biased and influenced by personal observation of GAARS working in practice, but may nevertheless represent the correct view.
Domestic anti-avoidance rules after BEPS Action 6
The Action 6 Final Report made domestic anti-avoidance rules one of its priorities. In connection with that it includes changes to the OECD Model Convention for the purpose of ensuring that tax treaties do not prevent the application of domestic anti-abuse rules to situations, such as thin capitalisation, dual residence strategies, transfer pricing, and so forth. Many other Actions contained in the Action plan deal with those situations: in particular, Action 2 (Neutralise the effects of hybrid mismatch arrangements); Action 3 (Strengthen CFC rules); Action 4 (Limit base erosion via interest deductions and other financial payments) and Actions 8, 9 and 10 dealing with Transfer Pricing. Therefore, some conciliation with the OECDE Model Convention, to the extent that some of its Articles prevent its application, is needed. The main proposal in this regard is to include a revised version of the section “Improper use of the Convention”, currently found in the commentary on Article 1, to enhance the articulation of the relationship between domestic anti-abuse rules and tax treaties31.
A strategy to fight abuse based on domestic provisions will be complex and uncertain in terms of implementation, especially because, in the context of BEPS, those provisions will gain sophistication which will magnify disproportionally the already existing problems. Hence, the strategy should be less ambitious – many difficulties are to be expected32.
Treaty provisions and BEPS action 6
The most apparent change in terms of new treaty anti-abuse provisions is not only the already analysed PPT (Principal purpose test) or GAAR (General anti-abuse rule) but also the brand-new detailed LOB provision and related commentary that, following the Action
6 Final Report, became part of the OECD Model Convention. This LOB provision corresponds to the first six paragraphs of Article 29 of the 2017 version of the OCDE Model Convention, which also encompasses the PPT or GAAR in § 9. The combination of those two provisions in the same Article is further evidence that the said report considers that both mechanisms have weaknesses and strengths and therefore should work together. The frequent acknowledgment that the objective tests upon which the LOB is based are not effective33 enough as far as determining taxpayers’ intentions34 may also justify that combination. In line with the PPT that materialised ideas which had already appeared in
the commentary, the LOB follows the same course – it is also being upgraded from the commentary to the body of the OECD Model Convention.
The Action 6 Final report also considers the need of creating targeted specific anti-abuse. In connection with that, the said report gives examples of situations with respect to which specific treaty anti-abuses rules may be helpful and proposals for changes intended to address some of these situations in both the OECD Model Convention and Commentary35.
Those situations are the following: (1) certain dividend transfer transactions that are intended to lower artificially withholding taxes payable on dividends; (2) transactions that circumvent the application of the treaty rule that allows source taxation of shares of companies that derive their value primarily from immovable property; (3) situations where an entity is resident of two Contracting States; and (4) situations where the state of residence exempts the income of permanent establishments situated in third states and where shares, debt-claims, rights or property are transferred to permanent establishments set up in countries that do not tax such income or offer preferential treatment to that income.
As I have argued, favouring specific anti-abuse rules as a mechanism of fighting abuse is always good. Thus, I consider that the effort made by Action 6 Final Report to propose new specific anti-abuse rules and to refine the existing ones is highly desirable and positive.
Even if the author considers that the efforts made in relation to treaty provisions are highly positive, very little attention was paid to the concept of ‘beneficial owner’, therefore missing an excellent opportunity to make the clarifications that the concept still lacks. Silence has the potential to produce even more uncertainty. Here is one possible scenario: based on the fact that there are now two extra provisions to tackle treaty shopping in a more specific way – the new GAAR and LOB – combined with the OECD Model Convention commentary on Article 10, § 12.5, some commentators, states or judges, may construe the concept of ‘beneficial owner’ merely as a condition, like residence, for entitlement to treaty benefits. According to this view, the beneficial owner concept would not be an anti-abuse mechanism anymore. Hence, the concept could be taken in a strictly technical sense and easily equated to a subject-to-tax concept. This outcome, according to which the anti-abuse dimension of the concept of ‘beneficial ownership’ would be severed, is not desirable. The more means exist to fight abuse, the better, as it should follow from the BEPS spirit. The role played by other anti-abuse mechanisms should be considered as simply supplementary, not alternative.
CONCLUSION:
The addition of Article29 (Entitlement to benefits) which encompasses the PPT and an LOB provision is overall positive. The fact that they emerged from the commentary, whose relevance is still controversial, and became treaty provisions is, indeed, a substantial improvement. The author also praises the fact that more attention is devoted to the specific anti-abuse provisions, which explicitly recognises, in line with what the author argued before, that “targeted specific anti.-abuse rules generally proved greater certainty for both taxpayers and tax administration”36.
The author disapproves, however, of the missed opportunity to, at once, clarify the concept of beneficial owner which, combined with the mixed signs sent by the addition of the new GAAR and the LOB provision, bring a new level of uncertainty.