Determining the Beneficial Owner of Dividend Income Compared to Other Items of Income in International Taxation: nihil sub sōle novum¹

The main proponent for including the concept of BO to the OECD MTC at the forum of the OECD was the UK. It also was a trendsetter in including the concept of BO in tax treaties.³¹ In fact, the genesis of adding the concept of BO to the OECD Model in 1977 is closely related to the UK’s needs at the time, resolving the taxpayer’s problem of income allocation.³² This problem arose from the peculiar legal regulations in force in that country.

The UK regarded trustees and nominees as taxpayers, and therefore they were entitled to benefits under UK tax treaties in respect to income received from abroad, even though the income was received for the benefit of other people, known as the beneficiaries of the trust. This model of taxation developed in the UK due to the fact that the right to property was originally seen as an indivisible right, but in light of the principle of equity, the concept of the divisibility of the right to property developed. The property right was divided into two categories: legal and economic, which can accrue to different entities. In other words, under the principle of equity, the right to property may be simultaneously divided between two people in different ways. One person, who holds only the legal title to an asset, without being able to use it or benefit from it, is called the legal owner. This person is a trustee in a trust. The other person, who has the right to benefit from and control the asset, is called the beneficial owner.³³ This person is the beneficiary in the trust. If the beneficiaries of the trust were not UK tax residents, the UK trustees were not taxed on the income received on their behalf in the UK, while also retaining the rights to benefit from UK tax treaties.

The UK delegation to the OECD was concerned that such rules on income allocation to the beneficiaries of trusts could lead to an abuse of the UK tax treaties. If the beneficiaries were tax residents of a different state than the of the trustees of the trust, the payments paid to the trustees would be subject to double nontaxation. They would be exempt from WHT in the source state of the payments under the tax treaty between the UK and the source state, and not subject to tax in the UK, which would be the resident state of the recipient of the payments (the trustee), in accordance with the domestic rules mentioned above concerning the allocation of income for tax purposes. Hence, the resident of a third state could reduce WHT in a source state with which the UK has a tax treaty in force by interposing a UK resident nominee or trustee who would not be subject to UK tax on income from the source state.³⁴

In 1963, as a result of these concerns, the British delegation to the OECD expressed the need to add the subject-to-tax clause³⁵ or the concept of BO to Articles 10 and 12 of the OECD MTC,³⁶ in order to eliminate the risk of an abuse of tax treaties by non-residents through trustees and nominees.³⁷ In the view of the British delegation, the concept of BO was supposed to be a panacea on the unreasonable effects of the rule on income allocation under tax treaties, allowing for an exception from relying on those rules in certain situations. Somewhat surprisingly, all UK treaties concluded in the 1960s included a subject-to-tax, and later BO, requirement for dividends, interest, and royalties.³⁸ This shows that the UK problem did not actually exist, but the UK most likely desired to have the subject-to-tax clause or the concept of BO in the OECD Model “to save arguments when negotiating treaties.”³⁹

As a result of the UK’s pressure, during the 27th session of the OECD’s Fiscal Committee, from September 19 to 27, 1967, the committee created Working Party 27 (WP27) consisting of the delegations to the OECD from Luxembourg, France, and the Netherlands. Its task was to revise the rules on the taxation of interest and royalties in accordance with Articles 11 and 12 of the OECD in order to respond to the concerns of the British delegation. This issue in relation to Article 10 of the OECD MTC was the subject of WP23, which mainly concentrated on the differing corporate shareholder tax systems in the OECD’s member states and how they could be reconciled in a tax treaty framework. This topic was discussed later as a result of the work of WP1 in 1973.⁴⁰

In its first report from December 30, 1968, WP27 heavily opposed the British delegation’s suggestion to add a subject-to-tax clause to Articles 11 and 12 (and by analogy to Article 10) of the OECD Model, because it would be contrary to the principal purpose of tax treaties based on the OECD Model.⁴¹ With regard to the concept of BO, WP27 stated that “it is evident that relief in the country of source applies only if the recipient is actually a resident in the other contracting state. This is stated clearly in the text of paragraph 1. Moreover, determining who the true recipient is and their state of residence is “a matter of administration and inspection.”⁴² In addition, WP27 perceived the problem raised by the UK delegation with determining the “true recipient” in the resident state as strictly procedural by nature (administrative and evidentiary). In the same vein, that problem appeared to be perceived by the US in the course of negotiating the tax treaty with the UK.⁴³ Accordingly, WP27 did not see a problem with determining the beneficial owner of income from the perspective of an abuse of tax treaties. Likewise, it did not refer to any substantive antiabusive measures as a remedy for the UK problem. The first report of WP27, therefore, did not recommend adding the concept of BO, as it would be redundant (there is no need to add a concept already implicitly existing in tax treaties) and possibly complicating the application of tax treaties (an assumption that any addition of new terms to tax treaties must mean “something new”).⁴⁴ In its second report, on February 16, 1970, WP27 changed its mind with regard to the concept of BO, recommending its addition to Articles 11 and 12 of the OECD MTC, in order to exclude agents and nominees from treaty benefits under these provisions:

The second solution consists in taking into consideration the State of residence of the beneficial owner, the dividends, interest, or royalties and in disregarding the State of residence of the person having the receipt of such income, whether so doing in the name and on behalf of the beneficial owner, or in his own name but on behalf of the beneficial owner. There is no reason to think that the case of the person acting manifestly as an agent in the name and on behalf of the beneficial owner gives rise to any difficulties. It may be otherwise with the trustee acting legally in his own name out on behalf of the beneficial owner. It is probable that certain States have no difficulty from the legal standpoint in attributing to the beneficial owner income received through a trustee. Such States would, however, be in no way harmed by the insertion into the Model Convention of a text to ensure legal safety in other States for whom the interposition of a trustee could cause complications. The Working Party therefore recommends that there be written into the Model Convention a provision whereby the “beneficial owner text” would be applied.⁴⁵ (emphasis added by the author)

According to WP27, adding the concept of BO to Articles 11 and 12 of the OECD Model would not be a significant change for the vast majority of OECD member states, because these states, unlike the UK, did not apply rules to allocate income for tax purposes to a person that is not entitled to that income, such as a nominee or a trustee.⁴⁶ Thus, the addition of the concept of BO to Articles 11 and 12 of the OECD MTC was intended to solve a uniquely UK problem (allegedly) without detriment or complication to the application of tax treaties by the other states, but also without much need for such an amendment to the OECD Model from their perspective. Therefore, in the end, the other OECD member states did not object to the implementation of the British delegation’s idea.⁴⁷ As a result, in the third report of November 4, 1970, WP27 produced a redraft of Articles 11 and 12 with the BO condition and the commentary attached to only the first paragraph of these provisions.⁴⁸

The OECD MTC was published on April 14, 1977. Unlike its predecessor from 1963 (the Draft Model), the OECD Model from 1977 included the concept of BO in Articles 10, 11, and 12. Table 1 below sets out the wording that makes up the concept of BO under these provisions.

As a result of adding the concept of BO to the OECD Model in 1977, that concept was commonly added to tax treaties worldwide concluded or modified after 1977.

The OECD expressed an identical approach to understanding the concept of BO to dividend, interest, and royalty income from the very origin of that concept in MTC (1977). In paras 12, 8, and 4 of the commentary on Articles 10, 11, and 12 of the OECD Model (1977), respectively, we can read that:

the limitation of tax in the State of source is not available when an intermediary, such as an agent or nominee, is interposed between the beneficiary and the payer, unless the beneficial owner is a resident of the other Contracting State. States which wish to make this more explicit are free to do so during bilateral negotiations. (emphasis added by the author)

This implies that the concept of BO under the 1977 OECD MTC: (1) explicitly excluded only two types of intermediaries – agents and nominees – from the scope of beneficial owners that may benefit from Articles 10, 11, and 12; and (2) it had only a clarificatory function on the understanding of the payments of dividends, interest, or royalties from the source state to the resident state of the recipient under Articles 10, 11, and 12.⁴⁹ That is to say, the function of the concept of BO since its addition to the OECD MTC was to clarify the allocation of income to the taxpayer for the purposes of the proper application of tax treaties, and that the status of the beneficial owner should be assessed from the perspective of the tax law of the resident state of recipient of dividends, interest, or royalties.⁵⁰ There was no difference in that regard between the various items of income.

Accordingly, no single distinction regarding dividend income that would be of legal meaning and consequence in determining BO of that income arises from the 1977 OECD MTC and Commentary.

In the author’s opinion, the above approach of the OECD reflects the views of WP27 that the role of the concept of BO was to clarify the allocation of income to the taxpayer for the purposes of properly applying the provisions of the tax treaties, as could be interpreted from the wording of Articles 10(1), 11(1), and 12(1) in the OECD MTC already prior to 1977.⁵¹ Notably, neither the OECD MTC nor the Commentary, unlike the reports of WP21,⁵² referring to conduit or holding companies as intermediaries that escape the pool of beneficial owners under Articles 10, 11, and 12. This implies that the concept of BO added to Articles 10, 11, and 12 of the OECD Model in 1977 should not be equated with the anti-abusive, economic concept of BO that apparently could be deduced from the reports of WP21. It also is this author’s view that para. 10 of the Commentary to Article 1 of the 1977 OECD MTC, which implies that the concept of BO covers some of the situations of abusive tax avoidance under tax treaties,⁵³ was added to that MTC due to the conflation of the works of WP21 with those of WP27. They both used the same term “beneficial owner” for different purposes: WP21 for antiabusive and antifraud purposes from the perspective of a resident state, while WP27 for proper income allocation purposes from the perspective of a source state.⁵⁴

Paragraph 14(b) of the OECD’s report “Double Taxation Conventions and the Use of Conduit Companies”⁵⁵ (the Conduits Report) from 1986 places conduit entities in the group of agents and nominees excluded from the scope of beneficial owners. It clarifies the meaning of such an intermediary, defining a “conduit”⁵⁶ as “a person [who] enters into contracts or takes over obligations under which he has a similar function to those of a nominee or an agent,” “the formal owner of certain assets [who] has very narrow powers which render it a mere fiduciary or an administrator acting on account of the interested parties.” Consequently, a conduit is an intermediary obliged to act in respect of the received income on account of interested parties, just as an agent or a nominee would act.

These characteristics of a conduit were identical for all items of income covered by the concept of BO under Articles 10, 11, and 12. It further strengthens the observation that the determination of the concept is the same for dividend income as it is for interest and royalty income.

The author observes that that the OECD wanted to address the concerns of tax authorities of its member states via the Conduits Report, due the lack of antiabusive provisions in the OECD MTC at the time and the OECD’s position whereby the international principle pacta sunt servanda would prevent the contracting states from applying their domestic anti-abuse rules and doctrines to deny treaty benefits.⁵⁷ As a result, the Conduits Report brings more confusion to the understanding and policy goals of the concept of BO. The motivation of the OECD appears to be pleasing the tax authorities of its member states with more ambiguous, antiabusive characteristics of the concept of BO, as if it included several antiabusive solutions in its wording and structure,⁵⁸ or even as if it was a kind of GAAR with the embedded substance over form mechanism.⁵⁹ The OECD promoted such an inconsistent and vague approach to the concept of BO irrespective of the income to which this concept supposed to apply at the time.

In 1999, the OECD’s Committee on Fiscal Affairs (CFA) released a report on “The Application of the OECD Model Tax Convention to Partnerships”⁶⁰ (the Partnerships Report) setting out solutions to the problem of applying tax treaties to partnerships.⁶¹ In this report, the OECD reinforced the idea that the source countries should take into account how an item of income arising in its jurisdiction (the resident state from the perspective of the source state applying the treaty) is treated in the jurisdiction of the taxpayer claiming the benefits under the treaty.⁶² Thus, according to the solutions provided by the Partnerships Report, an allocation of income from dividend, interest, or royalties to an entity under the tax law of its resident state is decisive in determining the beneficial owner of that income and the proper application of the treaty with the resident state by the source state of that income. The principles in the Partnerships Report confirm and even reinforce the original purpose and perception of the concept of BO in light of tax treaties, insofar as they require the source state to determine the beneficial owner based on the actual allocation of income to the income recipient by the resident state of that recipient. At the same time, the source state should disregard the effect of rules on the income allocation of the resident state whenever they generally allocate an income to an entity, but in certain circumstances they waive that allocation, meaning that the income flows through the entity located in the resident state to other entities.⁶³

Such an approach to determining the concept of BO was presented in the Partnerships Report without any distinction between dividend, interest, or royalty income under Articles 10, 11, and 12 OECD MTC. Thus, the determination of that concept remained the same for all items of income it covers. The merged approach under the concept of BO and the general approach under the Partnerships Report is now reflected in para. 13 of the Commentary to Article 1 of the 2017 OECD MTC, which clarifies that the source state needs to follow the rules of income allocation in the resident state, unless tax transparent partnership (albeit seen as the tax resident by the state of its commercial registry) receives the income on the account of other entities (also other than its partners), for example as their agent or nominee.

On January 28, 2003, the OECD adopted significant changes to the Commentary to the OECD MTC directed at preventing the abuse of tax treaties.⁶⁴ The changes were proposed in the Report on Restricting the Entitlement to Treaty Benefits.⁶⁵ They included, among other things, several recommendations from the Conduits Report⁶⁶ and the Report on Harmful Tax Competition⁶⁷ that allow tax treaties to be interpreted in an antiabusive way without changing the content of tax treaties.

The new wording of para 12 to Article 10 of the 2003 OECD MTC reads as follows:

The requirement of beneficial ownership was introduced in paragraph 2 of Article 10 to clarify the meaning of the words “paid … to a resident” as they are used in paragraph 1 of the Article. It makes plain that the state of source is not obliged to give up taxing rights over dividend income merely because that income was immediately received by a resident of a State with which the State of source had concluded a convention. The term “beneficial owner” is not used in a narrow technical sense, rather, it should be understood in its context and in light of the object and purposes of the Convention, including avoiding double taxation and the prevention of fiscal evasion and avoidance. (emphasis added by the author)

Similar amendments were added to paras 8 and 4 to Articles 11 and 12, respectively. All of them seem to be clarificatory by nature and thus could be applied to all tax treaties, i.e., also those negotiated and concluded before 2003. Their language is the same for all items of income without any distinguishable features attributed to dividend income. They affect the understanding of the concept of BO in the same ways by clarifying that this concept:

was added to the OECD Model to clarify the meaning of the words “paid … to a resident” in respect to dividends or interest paid to a resident under Article 10(1) and Article 11(1), and to clarify the meaning of the payments of royalties made to intermediaries under Article 12(2);

This Commentary, in the author’s view, confirms that the concept of BO was added to the OECD MTC in order to clarify the allocation of income to the taxpayer for the purposes of the proper application of tax treaties, and that the status of the beneficial owner should be assessed from the perspective of the tax law of the resident state of the recipient of dividends, interest, or royalties.⁶⁸

The next change in the Commentary to the 2003 OECD MTC in respect to the concept of BO was the addition of para 12.1 to Article 10, which reads as follows:

Where an item of income is received by a resident of a Contracting State acting in the capacity of agent or nominee, it would be inconsistent with the object and purpose of the Convention for the State of source to grant relief or exemption merely on account of the status of the immediate recipient of the income as a resident of the other Contracting State. The immediate recipient of the income in this situation qualifies as a resident, but no potential double taxation arises as a consequence of that status, since the recipient is not treated as the owner of the income for tax purposes in the State of residence. It would be equally inconsistent with the object and purpose of the Convention for the State of source to grant relief or exemption where a resident of a Contracting State, otherwise than through an agency or nominee relationship, simply acts as a conduit for another person who in fact receives the benefit of the income concerned. For these reasons, the report from the Committee on Fiscal Affairs entitled “Double Taxation Conventions and the Use of Conduit Companies” concludes that a conduit company cannot normally be regarded as the beneficial owner if, though the formal owner, it has, as a practical matter, very narrow powers which render it, in relation to the income concerned, a mere fiduciary or administrator acting on account of the interested parties. (emphasis added by the author)

Almost identical amendments were added to paras 10 and 4.1 to Articles 11 and 12, respectively. This indicates that no distinction should be given to determining the concept of BO to dividend income as compared to interest and royalty income in accordance with the 2003 Commentary to the OECD MTC.

It is worth pointing out that this change affects the understanding of the concept of BO in a considerably far-reaching way, in comparison to what could be interpreted from the previous versions of the Commentary to Article 10, 11, and 12 and the relevant OECD reports. It explicitly excludes conduit entities from the scope of beneficial owners. Despite the explicit reference in this respect to the Conduits Report, the manner in which the conduits were excluded in the above quoted part of the Commentary shows aberrations from the Conduits Report. Notably, in the last sentences of the quoted passage of the Commentary (paras 12.1, 10, and 4.1 to Articles 10, 11, and 12 of the 2003 OECD Model, respectively), reference is made to the conclusions of the Conduits Report, indicating that this report “concludes that a conduit company cannot normally be regarded as the beneficial owner if, though the formal owner, it has, as a practical matter, very narrow powers which render it, in relation to the income concerned, a mere fiduciary or administrator acting on account of the interested parties.” The emphasized phrase “as a practical matter” does not appear in the Conduits Report in relation to the conduit’s “very narrow powers which render it a mere fiduciary or an administrator acting on account of the interested parties.”⁶⁹ The OECD’s statement added to the Commentary is therefore incorrect and constitutes, at most, a distorted conclusion from the Conduits Report. It falsely suggests that the conclusions of the Conduit Report led to the need to examine conduits based on the existence of a very narrow entitlement to income received due to practical considerations, rather than due to legal or contractual obligations.

To the extent that the 2003 Commentary appears to require determining the very narrow powers of an intermediary under factual rather than legal or contractual obligation, it brings the concept of BO close to nonsensical.⁷⁰ Scholars have rightly observed that the concept of BO has never been and should not be based on the examination of factual obligations,⁷¹ but should remain limited to the examination of legal or contractual obligations.⁷² From this policy point of view, a beneficial owner of income from dividend, interest, and royalties is only the ultimate economic owner of that income. This understanding of the concept of BO has never been intended under tax treaties, nor in the Conduits Report.⁷³ At most, it was used for seemingly similar policy purposes only in the works of WP21 in the context of preventing fraud such as tax evasion in the broader spectrum of cross-border issues considered from the perspectives of resident countries.⁷⁴ As history shows, however, this concept of BO has never been added to the OECD MTC.⁷⁵

The 2014 update of the Commentary to the OECD MTC seems to significantly restrict its scope of application by carving out from the pool of beneficial owners only entities engaged with mutually interdependent payments of dividends, interest, and royalties. Paras 12.4, 10.2, and 4.3 of the 2014 Commentary to Articles 10, 11, and 12, respectively, read as follows:⁷⁶

In these various examples (agent, nominee, conduit company acting as a fiduciary or administrator), the direct recipient of the dividend is not the “beneficial owner” because that recipient’s right to use and enjoy the dividend is constrained by a contractual or legal obligation to pass on the payment received to another person. Such an obligation 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. This type of obligation would not include contractual or legal obligations that are not dependent on the receipt of the payment by the direct recipient such as an obligation that is not dependent on the receipt of the payment and which the direct recipient has as a debtor or as a party to financial transactions, or typical distribution obligations of pension schemes and of collective investment vehicles entitled to treaty benefits under the principles of paragraphs 22 to 48 of the Commentary on Article 1. Where the recipient of a dividend does 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, the recipient is the “beneficial owner” of that dividend. It should also be noted that Article 10 [11 and 12] refer to the beneficial owner of a dividend [interest or royalties] as opposed to the owner of the shares [the debt claim with respect to which the interest is paid or the right or property in respect of which the royalties are paid], which may be different in some cases. (emphasis added by the author)

Again, there is no distinction between the way the concept of BO operates towards dividend income as compared to interest and royalty income.

In respect to all items of income covered by the concept of BO, the Commentary focuses on establishing legal or contractual restrictions on the use and enjoyment of the specific stream of income received. If such restrictions result from a legal or contractual obligation on the part of the recipient of the income to transfer that income to another entity, then the recipient does not seem to be a BO of that income. However, not every contractual or legal obligation to pass on the payment received to another person is enough to exclude a person from the scope of the BO of that income, but only some, i.e., only if it is dependent on the direct recipient receiving that income. Only if the contractual or legal obligation means that the obligation to pass on the income to another person is conditional (dependent) on the receipt of that income does the obligation taint the relationship between the person paying the income and the direct recipient of the income. The tainted relationship in respect to receiving and transferring income arising from the mentioned obligations deprives the direct recipient of that income of BO over it.⁷⁷ Such obligations can be called “qualified obligations.”⁷⁸ Recipients of the income remain BOs over it, even if they transfer that income to another person, just as long as that transfer does not arise from a qualified obligation. Anything less than a qualified obligation does not deprive the recipient of BO.⁷⁹

However, even when a qualified obligation occurs, it will not always exclude the status of BO with the effect of the denial of benefits under Articles 10, 11, and 12 of the OECD MTC. Indeed, it can be argued that there must be some abuse of the treaty protection in order for the treaty benefits to be denied.⁸⁰ To this end, in this author’s view, it must also be demonstrated, based on the relevant circumstances, that the sole, or at least main intention of the establishment and operation of the structure with the intermediary, is to obtain benefits under Articles 10, 11, or 12 in a manner contrary to the purpose of these provisions. A denial of treaty benefits in respect of such payments of income would be justified under Article 31 VCLT, insofar as it precludes the abuse of Articles 10, 11, or 12 of the tax treaty between the source state of the income received by the intermediary and the resident state of the intermediary by the person that ultimately receives that income from the intermediary.

In 2017, the updated Commentary almost did not touch on the concept of BO at all, apart from minor changes that were dictated by the far-reaching antiabusive changes to the OECD MTC as a result of the implementation of the BEPS Action 6, most importantly the addition of Article 29, which includes the LOB clause in paras 1–8 and the PPT in para. 9. Consequently, paras 12.5, 10.3, and 4.4 of the Commentary to Articles 10, 11, and 12⁸¹ respectively, refer to the new Article 29 (the LOB clause and the PPT) and the antiabusive principles put forward in the Commentary to Article 1 (‘Improper use of tax treaties’).⁸² In addition, these passages of the Commentary in 2017 indicated that the concept of BO does not deal with “other cases of abuse,” rather than, as the 2014 version of the Commentary explained, with “other cases of treaty shopping.” Again, this change was not made to add or subtract any meaning from the concept of BO per se, as it was driven solely by the fact that in 2014 the PPT had not been added to the OECD MTC. Since this rule aims to address all forms of treaty abuse, not just abusive treaty shopping,⁸³ it was logical to replace the seemingly narrower phrase “other cases of treaty shopping” with the broader “other cases of abuse.”

The biggest impact of the 2017 update of the OECD MTC and its Commentary in respect of the concept of BO stems from the inclusion of the PPT, i.e., when comparing the scope of the PPT with the concept of BO, one cannot shake off the impression that the PPT has almost completely reduced any antiabusive role of that concept. Indeed, the Commentary to the PPT (Article 29(9) of the 2017 OECD MTC) indicates that the PPT covers “limitations on the taxing rights of a Contracting State in respect to dividends, interest, or royalties arising in that state, and paid to a resident of the other state (who is the beneficial owner) under Article 10, 11, or 12.”⁸⁴ This suggests that the PPT, rather than the concept of BO, addresses all conduit cases that lead to abuses of tax treaties.⁸⁵ The OECD’s examples of the application of the PPT further enhance this observation.⁸⁶

Still, the OECD did not provide any wording in the 2017 MTC and the Commentary that would imply that the concept of BO should be determined differently in respect to dividend income in comparison to interest and royalty income.

In the author’s view, the analysis in sections 2.1–2.7 demonstrates that the concept of BO appears to be an income allocation rule for the purposes of all relevant items of income. It is worth explaining what the author means by considering the concept of BO as an income allocation rule rather than an antiabusive rule for the purposes of all items of relevant income under tax treaties. In general, it means that the concept of BO primarily ensures the proper operation of Articles 10, 11, and 12 of the OECD MTC by following the actual (rather than simulated) allocation of income.

An actual (effective/real) allocation of income is a heightened tax-specific notion of when income belongs to a person (natural or legal), i.e., it is about connecting income to a taxable person rather than allocating the taxing jurisdiction. The allocation of income is triggered solely by the domestic tax law of contracting states, while the concept of BO is a threshold factor to establish whether the limit on WHT under a tax treaty in the source state should apply or not. This threshold factor is met when income from the source state is actually allocated to the resident state. Consequently, the concept of BO requires a careful analysis of facts that trigger an actual allocation of income under the domestic law of the contracting state in conjunction with a precise legal analysis of that law and the relevant tax treaty provisions (usually: equivalents of Articles 10, 11, and 12 of the OECD MTC under a treaty between the source state and the resident state) and EU directives. If real facts trigger the allocation of income under the tax law of the resident state, then the allocation is actual; if not, then the allocation is simulated (not real/ineffective). An allocation of income is not real, for example, when the transaction described in the documentation between companies does not take place in reality. This is the case when the documentation says that payments are transferred from one company (resident in a source state, e.g., Italy) to a second one (resident in the resident state, e.g., the Netherlands), but in reality, the payments bypass the second company and are transferred directly to a third company (not resident in the resident state, e.g., the United States).⁸⁷

Even though in many countries such a simulated allocation of income can be ignored for tax purposes as a sham transaction (i.e., fraud under general law), in a cross-border scenario it may be very complicated for the tax authorities of the source state to establish whether or not the allocation of income to a resident of the resident state is real. It is also necessary extremely fact sensitive and requires a near-flawless exchange of tax information between the source state and the resident state.⁸⁸ The concept of BO aims to ensure that the tax authorities of the source state will perform that arduous task in a proper way. Otherwise, a tax treaty with the resident state may be applied in an improper way. This shows that the notion of the “real allocation of income” refers to a general legal reality, i.e., to the allocation of income based on real facts. The tax reality is, in turn, valid when facts are legally valid and factually established (when they are real) for all general law purposes, but for tax law alone, they are to be disregarded. They can be disregarded, for instance, under GAARs, the PPT, or other anti-tax avoidance measures such as substance-over-form or economic substance doctrines.⁸⁹

Accordingly, the claim of this article is that the concept of BO does not operate to disregard the allocation of income by the resident state as followed by the source state for the purposes of tax treaties and EU directives in accordance the anti-tax avoidance measures mentioned above. This is the role of those measures, and the concept of BO does not have premises of tax avoidance, in contrast to the mentioned measures, which could lead to ignorance about the allocation of income in line with the letter of the law. Consequently, the concept of BO is a search for legal substance at the level of the entity (often: an intermediate entity) in the resident state that receives income from the source state rather than the economic substance.⁹⁰ As Frederik Zimmer clarified, the legal substance “most often refers to the characterization that emerges from a close study of the rights and obligations in a legal relation,” and its main function “is to point out that sham or simulation transactions and wrong legal characterizations by the taxpayer will be disregarded for tax purposes.”⁹¹

This is also the claim of this article, namely that the determination of facts towards an allocation of income to the intermediary entity should be assessed mainly from the perspective of the resident state of that entity, which is usually a direct recipient of the income from the source state. To this end, the source state should enter into an exchange of tax information with the resident state in order to gather the relevant information for an appropriate application of a tax treaty with the resident state. If a resident state refuses to enter into such an exchange of tax information agreement, or it makes inappropriate claims in order to pursue its own interests under such agreement, making it impossible for the source state to obtain genuine tax information allowing it to determine the status of BO, then the source state should be entitled to deny tax benefits stemming from Articles 10, 11, and 12 and the IRD and the PSD.⁹² This is justified by analogy to the reasoning of the CJEU under EU law aiming to ensure effective fiscal supervision.⁹³

There are three means of interpretation that support such an approach.

First, whenever there is a conflict of income allocation between the resident state of a payment’s recipient and the source state of the payment, paras 53–54 of the Partnership Report indicate that the source state should take into account how an item of income arising in its jurisdiction is treated in the jurisdiction of the taxpayer claiming the benefits under the treaty, i.e., in the resident state of the income recipient. These principles are now reflected in the 2017 OECD MTC and its Commentary and now apply to all entities, not only to partnerships.⁹⁴ The Commentary highlights that, in the case of fiscally transparent partnerships acting as agents or nominees, the source state is not required to follow the income allocation of the resident state to determine BOs, which is an obvious example of a lack of BO.

Second, the purposive interpretation of Articles 10– 12 of the OECD MTC pulls towards the elimination of double taxation without creating opportunities for double non-taxation or reduced taxation of dividend, interest, and royalty income stemming from abusive tax avoidance.⁹⁵ Following only the allocation of income by the source state may create opportunities for nontaxation or reduced taxation through tax avoidance, including through treaty-shopping arrangements aimed at obtaining relief provided in a tax treaty for the indirect benefit of residents of third states. For instance, the source state of dividend income allocates that income to a foreign company, while the resident state of that company does not. Hence, there is no taxation of that income in the resident state. The source state does not tax that income as well because the tax treaty with the resident state exempts the dividend income in that situation. This results in the double non-taxation of dividend income. The elimination of double taxation by applying the WHT exemption in the source state under the tax treaty with the resident state is not reasonable because it takes a place in a situation in which double taxation is not possible, thereby creating opportunities for double non-taxation. By contrast, if the income allocation of the resident state is followed instead of the source state to determine the BO, the identified risk of tax avoidance is eliminated—the source state provides relief under the tax treaty only if the resident state allocates income paid by a resident of the source state. At the same time, this approach does not create any risk of double taxation—if the resident state allocates income, then the source state provides relief under the tax treaty. Hence, the presented interpretative approach fulfils the main operative purpose of Article 10 without creating opportunities for tax avoidance. The same is valid for Articles 11 and 12 (interest and royalties).

Third, if the source state’s income allocation would be conclusive for tax treaty purposes (i.e., determining the BO), that would be at odds with the principle of interpretation in good faith and the common intention approach to determine the concept of BO. The approach to determining the concept of BO that is most compatible with the principle of interpretation in good faith and common intention would be to do so by considering the allocation of income by the source state and the resident state, rather than only the former or the latter. However, as the two interpretative means show (see the above paragraphs), following the income allocation by the resident state is better suited to fulfil both purposes of Articles 10–12 of the OECD MTC, i.e., to eliminate double taxation without creating opportunities to avoid taxation. This also is a better way of respecting the principles of interpretation in good faith and common intention and resolving the conflicts of income attribution for tax treaty purposes.

From the perspective of the resident state, income may be allocated under ordinary rules allocating income to its residents or via extraordinary rules that have an anti-tax avoidance function, e.g., under GAARs, the PPT, substance-over-form, or economic substance doctrines. If the income fails to be allocated to the entity in the resident state, either under the former or the latter rules, then the source state should respect that situation also for the purposes of applying tax treaties.⁹⁶ The source state is free to apply its own domestic anti-tax avoidance measures to disregard the tax consequences of the allocation of income by the resident state for purposes of tax treaties and EU directives. However, this cannot be done via the concept of BO, which is not an anti-tax avoidance rule (measure) and cannot be applied as a handy substitute for any of them.

To finalize that the approach in order to properly determine the facts in connection with the concept of BO, it is worth referring to Zimmer’s observation whereby “there is a difference of principle between establishing the facts on the basis of rules of evidence and deciding whether the legal conditions for declaring tax avoidance are fulfilled.”⁹⁷ The premise of this article is that the source state must take into account the allocation of income by the resident state, and therefore a holistic approach is needed. It means that the facts on the basis of rules of evidence and the legal conditions for determining tax avoidance, as applied by the resident state, are of relevance when applying the concept of BO by the source state, insofar as they allow the source state to apply a tax treaty or the IRD or the PSD with the resident state in line with the actual allocation of income from the source state to a resident in the resident state.

The concept of BO is explicitly included in Article 1(1) of the IRD, which determines the subjective and objective scope of its application:

Interest or royalty payments arising in a Member State shall be exempt from any taxes imposed on those payments in that State, whether by deduction at source or by assessment, provided that the beneficial owner of the interest or royalties is a company of another Member State or a permanent establishment situated in another Member State of a company of a Member State. (emphasis added by the author)

Article 1(4) of the IRD describes the role that has to be attributed to the recipient of interest or royalties to be deemed as their beneficial owner:⁹⁸

A company of a Member State shall be treated as the beneficial owner of interest or royalties only if it receives those payments for its own benefit and not as an intermediary, such as an agent, trustee or authorised signatory, for some other person. (emphasis added by the author)

By comparison, the PSD does not use the term “beneficial owner” or “beneficially owned” at all. One may therefore conclude that it is impossible to compare the determination of the concept of BO of dividend income to that of interest and royalty income under the EU, because the concept does not exist in respect of dividend income.

Nevertheless, in the author’s view, the concept of BO can be interpreted from the wording of the PSD, in particular its Article 5: “Profits which a subsidiary distributes to its parent company shall be exempt from withholding tax.” This is subject to a critical caveat: the concept of BO can be interpreted from the PSD only if its function, as stemming from that interpretation, is to ensure that the PSD applies only to the actual (rather than simulated) allocation of dividend income by its payment (distribution) from a subsidiary to its parent company.⁹⁹ In the author’s view, the same observation on the function of the concept of BO applies to the IRD and to Articles 10, 11, and 12 of the OECD MTC. The point to be clarified here is that the concept of BO, understood as income allocation rule, is to be determined identically for dividend income as for interest and royalty income, not only under Articles 10,11, and 12 of the OECD MTC, but also under the IRD and the PSD.¹⁰⁰

The conclusions about the lack of an anti-abuse function in the concept of BO may be inferred from the materials of the EU Commission on the IRD, in the context of the wording and purposes of that directive. In the Proposal to the IRD from 1998, the Commission pointed out that the proposed directive “includes provisions to ensure that Member States are not precluded from taking steps to combat fraud or abuse.”¹⁰¹ However, there is no indication that the concept of BO come under those antiabusive provisions. Although the EU Commission later indicated, in 2009 in its Report on the application of the IRD that the concept of BO “aims at ensuring that relief under the [d]irective is not wrongfully obtained through the artificial interposition of an intermediary” (“artificial conduit arrangements”),¹⁰² this indication has never been supported by the wording of the IRD, which distinguishes between the concept of BO in Article 1(1) and (4) and the antiabusive rule in Article 5.¹⁰³ The concept of BO must be interpreted literally and strictly, rather than broadly and purposively, in order to properly safeguard the achievement of the principal purpose of the IRD, which is to eliminate the double juridical taxation of interest and royalty payments from subsidiaries to their parent companies in the EU.¹⁰⁴ A strict literal interpretation of the concept of BO cannot appropriately target conduits that hide companies not meeting the conditions to benefit from the IRD because they have, for instance, tax residence outside the EU, meaning that direct payments to them would trigger higher WHT.¹⁰⁵ To effectively target such arrangements, and thus to prevent the abuse of the IRD, it appears more appropriate to rely on the transpositions of Article 5 of the IRD (and Article 1(2) of the PSD) by EU member states to their domestic tax laws, instead of a broad and purposive interpretation of the concept of BO.

This section constitutes the author’s observations on selected tax jurisprudence on the concept of BO, in order to supplement the ongoing analysis with the law in action.¹⁰⁶ The overall conclusion is that the tax jurisprudence approaches in the EU, the US, and Canada to determining the concept of BO remain the same with respect to dividend income compared to interest and royalties.

The selection of US, Canadian, and CJEU tax jurisprudence is dictated by their relevance and importance to the subject matter of this article. The judgment of the United States Tax Court (USTC) of August 5, 1971 in the Aiken Industries case is one of the most important and at the same time one of the first judgments concerning the concept of BO in international jurisprudence.¹⁰⁷ It predates the inclusion of the concept of BO to the OECD MTC in 1977, so it sheds a light on its development, constituting a vital contextual (historical) source about its understanding in tax treaty practice of the then and now the largest economy in the world. The judgments of the Tax Court of Canada (TCC) and the Federal Court of Appeal (FCA) on the landmark Prévost case,¹⁰⁸ in turn, are seen as effective judicial transplants of the OECD documentation by citing and faithfully following them in light of the general rule of treaty interpretation.¹⁰⁹ They decoded the meaning of the beneficial owner of dividends in a narrow legal way, closely in line with its perception as a rule of income allocation. This tax jurisprudence contributed to the uniform, international fiscal meaning of the concept of BO. Finally, the CJEU judgments in the Danish BO cases are of particular relevance as they constitute the only case law of the CJEU that touches upon the concept of BO under EU law. Despite the lack of formal status of CJEU judgments as universally binding EU law (there is no stare decisis principle),¹¹⁰ its judgments have the status of de facto precedents.¹¹¹ In addition to these doctrines, the case law of the CJEU is also a source of general principles of EU law, such as the general principle of a prohibition on the abuse of rights under EU law, which are considered as additional sources of EU primary law.¹¹²

Any overview of the tax jurisprudence on the concept of BO should start with the CJEU case law in the Danish BO cases. Although the CJEU only explored the meaning of the concept of BO under the IRD, due to the fact that only that directive uses the term “beneficial owner,” it did refer to that concept also while examining the PSD, even though that directive does not use the term “beneficial owner” or “beneficially owned” at all.

In the author’s view, the CJEU did not consider the interpretation of the concept of BO as a standalone and exhaustive interpretation of that concept under the IRD, but merely as an interpretative steppingstone to a denial of WHT relief under the IRD via the general principle of the prohibition on the abuse of rights under EU law, which was necessary to deny this tax benefit in the peculiar legal and factual circumstances of the Danish BO cases. In particular, in its earlier case law the CJEU stated that the Danish tax authorities cannot rely on the domestic antiabusive judicial doctrine to deny benefits under EU secondary law.¹¹³ Thus, they later used the concept of BO as a handy replacement for a statutory antiabusive rule to prevent abuse of EU secondary law, i.e., the IRD and the PSD in the Danish BO cases.

This, in conjunction with the highly complex tax structures of the taxpayers in the Danish BO cases, sheds light on the CJEU’s reasoning: it tangled up the concept of BO with the concept of an abuse of rights under the IRD and the PSD to allow the Danish tax authorities to prevent an abuse of the IRD and the PSD without a clear-cut statutory legal basis for doing so. It did not appear to be fully consistent with its previous case law (Kofoed case).¹¹⁴ The mentioned entanglement appears to be simply a by-product of bridging the CJEU’s reasoning on that concept with the mentioned general principle of EU law. However, to prevent an abuse of EU secondary law, the CJEU relied on the general principle of prohibiting an abuse of rights under EU law. The concept of BO was not decisive to this end. In fact, the CJEU has never stated that the lack of the status of BO is a self-standing and conclusive factor of an abuse of rights under the IRD and PSD. Nor has the CJEU actually classified the concept of BO as antiabusive rule. The confusing reasoning of the CJEU in respect of the concept of BO begs for a corrective interpretation of the CJEU’s judgments in the Danish BO cases in accordance with Article 31(1) VCLT and EU primary law in order to disentangle the concept of BO from the concept of abuse. This could be done by adopting a narrow and legal approach to the concept of BO, leaving a broad antiabusive effect to GAARs and PPTs, as applied in line with the general principle of the prohibition on an abuse of rights under EU law. Irrespective of whether it concerns dividend, interest, or royalty income, the concept of BO under EU law should not be seen as an antiabusive rule in light of the CJEU’s judgments in the Danish BO cases.

Turning to the national tax jurisprudence, the USTC’s judgment in the Aiken Industries case¹¹⁵ seems to enhance a non-abusive, legal approach to the concept of BO insofar as the court focused on identifying Industrias’s legal obligation to transfer the interest received from Aiken (formerly: MPI) in full to ECL. This approach brings the USTC’s understanding of the concept of BO close to its autonomous international fiscal meaning, which emphasizes the identification of a qualified legal or contractual obligation that makes the receipt of payments conditional on their further payment to another entity.¹¹⁶ The USTC further referred to the principle of “dominion and control” over the income, only to conclude that the treaty term “received by” means a corporation of a CS receiving income “as its own, and not with the obligation to transmit it to another [corporation].” As a result, the USTC did not recognize the BO of a corporation that was simply “a collection agent” with respect to the income it received, or, to put it differently, “merely a conduit for the passage [of income]” from the source state to the resident state of the ultimate recipient of that income. Such a company could not be deemed as having received the income as its own.

This reasoning, in this author’s view, merged the concept of BO with the rule of income allocation, thereby allowing the scope of beneficial owners to exclude only entities such as agents or nominees, including functional agents or nominees such as conduits that do not receive income on its own, but on account of other entities. Thus, the judgment in the Aiken Industries case does not confirm that the concept of BO is an antiabusive rule. The US tax jurisprudence since the Aiken Industries case enhances that observation.

The courts did not rely on the concept of BO to prevent abusive tax avoidance. Instead, they did so via the application of the antiabusive step transaction doctrine, which was the jurisprudential prototype of the US concept of economic substance.¹¹⁷ These judgments confirm that the economic substance test derives from and is valid for the application of the antiabusive judicial doctrine. It did not come from nor is it relevant to the concept of BO. This case law also led the US legislator to recognize that the concept of BO does not have any meaningful role in combating abusive treaty shopping. At the domestic level, this role was taken over by specific anti-conduit provisions aimed against back-to-back financial arrangements and transactions that could be seen as a direct response to the case law discussed above.¹¹⁸ At the tax treaty level, the antitreaty shopping role was accorded to the comprehensive and detailed LOB clause, which has been added to US tax treaties on a regular basis since 1981, when this clause found its way into Article 22 of the US Model Convention.¹¹⁹

In the Prévost case, even more clearly than the USTC in Aiken Industries case, the TCC and the FCA interpreted the concept of BO in accordance with its autonomous, international fiscal meaning. Such an interpretative approach to the concept of BO resulted in its effective judicial transplant from the OECD international materials into the Canadian tax treaty soil.¹²⁰ According to the Canadian tax jurisprudence, “a non-resident holding a company receiving a dividend will be considered to be the beneficial owner” unless it “has absolutely no discretion as to the use of the funds received.”¹²¹ By comparison, the OECD much less clearly says that “a conduit cannot normally be regarded as the beneficial owner if, though the formal owner, it has, as a practical matter, very narrow powers rendering it, in relation to the income concerned, a mere fiduciary or administrator acting on account of the interested parties.”¹²² Consequently, even if one may say that the courts in the Prévost case to some extent deviated from the meaning of the concept of BO as presented by the OECD materials,¹²³ they did so with a very positive effect for the stability and predictability of access to benefits under tax treaties. Their strict and legal understanding of the concept of BO under the Dutch–Canadian tax treaty appeared to be more in line with the canons of interpretation relevant to that concept under the VCLT than the OECD’s clarifications. In particular, in addition to the reference to the OECD materials, the TCC and the FCA referred to the understanding of the concept of BO from the Canadian and Dutch perspectives, which was a manifestation of a good faith interpretation seeking to ascertain and give effect to the common intention of both contracting states of the tax treaty.¹²⁴

Alternatively, it could be said that the meaning of the concept of BO in the Prévost case was in line with its international fiscal meaning sensu stricto, i.e., following from its strict context—the words “paid to… a resident” point to a rule of income allocation. Indeed, the TCC and the FCA considered it decisive for the determination of the status of beneficial owner that, under Dutch tax and commercial law, PHB.V. was the taxpayer on and the owner of the income from the dividends paid by Prévost. Such an interpretive approach coincides with the autonomous international meaning of the concept of BO that may be deduced from the Partnerships Report, incorporated into the 2003 version of the Commentary and retained in its current version from 2017. That is to say, the concept of BO is first and foremost a rule clarifying the allocation of income to the taxpayer for the purposes of Articles 10, 11, and 12, which should be considered from the perspective of the resident state of the recipient, in that case the Netherlands.¹²⁵ In that sense, the Canadian courts simply followed a stricter and more legal side of the OECD’s clarifications on the concept of BO, rather than its quasi -antiabusive side cited above. As a result, the judgments in the Prévost case remove from the circle of beneficial owners only those conduits that have absolutely no freedom to dispose of the payments they receive, or where the payments automatically flow through them.¹²⁶ Such situations involve sham arrangements or transactions in which the income is not actually allocated to a conduit at all, and therefore no treaty between the source state of the income and the resident state of the conduit applies.¹²⁷ Moreover, this very hesitant approach by the TCC and the FCA towards piercing the corporate veil under the concept of BO was supported by the Conduits Report, which says that such an approach is “incompatible with the principle of the legal status of corporate bodies, as recognized in the legal systems of all OECD member countries, and except in cases of abuse.”¹²⁸

This shows that the TCC and the FCA completely rejected a broad antiabusive interpretation of the concept of BO, and thus the legitimacy of examining the substance of the company receiving payments to determine its status as a beneficial owner.¹²⁹ Apart from these advantages of a purely legal nature, such an approach to the concept of BO, namely the narrow, strict and legal approach to the concept of BO in the Prévost case, fits neatly in with the current OECD global tax policy, according to which it is the PPT and/or the LOB clause that are the dominant antitreaty abusive rules, while the concept of BO has a marginal role in this respect at most.¹³⁰

The significant difference in reasoning of the national courts and the CJEU analyzed above appears to stem from the extreme malleability of the concept of BO and its ambiguity. This, however, was of no importance for the courts determining the concept of BO under dividend income, compared to interest and royalty income. The courts did not give any legal or factual significance to it. What mattered, however, were the facts and circumstances indicating tax avoidance. Although in such circumstances the Canadian and the US courts managed to resist interpreting and applying the concept of BO in an antiabusive economic way, the CJEU case law in the Danish BO cases failed to resist this temptation.

This is an example of the judicial pragmatism in legal interpretation of the concept of BO:¹³¹ the courts choosing to follow an interpretative approach under each and every theme as it suits their vision of the concept of BO in a particular case and moving in the direction of the outcome they want to achieve. This is an example of the reversed approach to legal interpretation: First deciding the status of the beneficial owner, and then looking for the interpretive path that will lead to their pre-determined decision, or at least what seems to fit it best.¹³² Their decisions typically result from arguments that consider empirically identifiable factual consequences with which the courts agree or disagree. The disagreements stem from the consequences in the form of abusive treaty shopping or directive shopping, and the agreements from the lack thereof. In the former cases, the courts feel justified to rely on the broad, economic, and antiabusive approach to the concept of BO. In such cases, the courts use the OECD materials as a pretext to create biased meanings of the international concept of BO against abusive tax avoidance. Regrettably, such interpretative pragmatism to the concept of BO has been followed not only by the CJEU, but also by many other courts.¹³³ This interpretative approach cannot be commendable because it reads in that concept abusive premises that never occurred in its wording under tax treaties or the EU directives, i.e., the tax avoidance intention and the lack of economic substance. It pours fuel on the fire of the divergent and very worrisome understanding of the concept of BO that is loosely linked with the implicit subjective and blurred tax avoidance intentions of the taxpayer in combination with the undefined concept of economic substance.¹³⁴