BEPS Action 6: Introducing an Antiabuse Rule in Tax Treaties

The OECD’s action plan on base erosion and profit shifting provides for the following measures as action 6, under the heading “Prevent treaty abuse”:

Develop model treaty provisions and recommendations regarding the design of domestic rules to prevent the granting of treaty benefits in inappropriate circumstances. Work will also be done to clarify that tax treaties are not intended to be used to generate double non-taxation and to identify the tax policy considerations that, in general, countries should consider before deciding to enter into a tax treaty with another country. The work will be coordinated with the work on hybrids.

The OECD has proposed a series of rules to be introduced into income tax treaties in a public discussion draft titled “BEPS Action 6: Preventing the Granting of Treaty Benefits in Inappropriate Circumstances”1 (hereinafter “Action 6 discussion draft”). This is, however, merely a consultation draft, and the — overwhelmingly critical — statements already submitted to and published by the OECD may still prompt it to amend or even withdraw the proposals.2 Nevertheless, the formulations chosen in this public discussion draft indicate that the OECD will be focusing its efforts on commenting on the rules under discussion in more detail instead of reviewing the rules on their merits. The OECD is explicitly asking for proposals on how to design the OECD commentary on these rules.3

The public discussion draft contains the draft for Article X, “Entitlement to Benefits,” the first five paragraphs of which are described therein as “specific antiabuse rule aimed at treaty shopping” and again, according to the OECD, are based on rules already contained in several income tax treaties.4 The U.S. treaties are mentioned first, followed by those of Japan and India.5 These first five paragraphs of Article X are meant to establish a limitation on benefits provision.

Action 6 discussion draft also contains a proposal for Article X(6),6 “Rules aimed at arrangements one of the main purposes of which is to obtain treaty benefits”:

Notwithstanding the other provisions of this Convention, a benefit under this Convention shall not be granted in respect of an item of income if it is reasonable to conclude, having regard to all relevant facts and circumstances, that obtaining that benefit was one of the main 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 OECD proposal definitely reflects the current trend — many countries have recently introduced general antiavoidance rules or have seriously considered their introduction. The European Commission has explicitly recommended the creation of those clauses to its member states, yet it proposed a wording that has so far been met with little positive response.7 Draft directives, such as those on the introduction of a common consolidated corporate tax base (CCCTB)8 or a financial transaction tax,9 and proposals for the amendment of existing directives (such as that on the parent- subsidiary directive10) also contain formulations for those clauses. These proposals, however, vary not just in detail. Little international consensus exists on how these rules must be designed and whether they are useful at all.

Changes to existing income tax treaties always carry the risk of the reverse conclusion.11 If, therefore, a rule like the proposed Article X(6) is introduced in new income tax treaties and in existing treaties with future effect and is not simultaneously included in all treaties, the question will inevitably arise as to what relevance the principles enshrined in this rule have beyond their scope of application. The absence of such a rule in several treaties will lead some interpreters to believe that precisely these principles are not binding therein.

Although such an argumentum a contrario is methodically very problematic, particularly regarding  treaties,12 experience has shown that some courts will unavoidably draw exactly that conclusion. This will not change even if the OECD commentaries repeatedly and emphatically point out that those rules are only for clarification purposes. In the past, rules that brought about changes in substance have all too often been falsely declared as clarifications so as to provide tax administrations with better arguments in favor of their application with retroactive effect.13 Meanwhile, explicit references to the clarifying character of some rules are even further fueling the suspicion that their actual intention is to bring about substantive change. It is difficult to understand why a rule should prescribe something that is not even necessary. For all these reasons, the OECD should be cautious in proposing new rules because of the obvious risk that the substance of various treaties will begin drifting apart. New rules only make sense if their benefits clearly outweigh the misgivings about them. We will now examine in more detail whether there are more reasons in favor of the introduction of an Article X(6) into the OECD model treaty.

Scope of Application

The scope of application of a rule corresponding to Article X(6) to be introduced into a treaty is limited to the respective treaty. The rule refers to the “benefit under this Convention.” Although that rule cannot lead to the conclusion that something else definitely applies beyond its scope of application, tax benefits granted on the basis of domestic tax law could not be denied by invoking that rule. It is equally inconceivable that that rule could resonate with other treaties and provide a legal basis for denying benefits under other treaties concluded by one of the contracting states or even by completely different states. A rule in a bilateral treaty corresponding to Article X(6) cannot even cover other international law treaties concluded between the same contracting states — even if they have the same or similar personal or substantive scope of application.

Interestingly, the scope of application of the proposed rule in Article X(6) does not even extend to the entire treaty. According to that rule, the benefit will not be granted “in respect of an item of income.” Therefore, it does not cover, for instance, the taxes on capital covered by the substantive scope of application of the OECD model.

A “benefit under this Convention” must refer to a benefit resulting from the treaty. Definitions such as those provisions based on article 3 of the OECD model alone will not suffice, since no legal consequences result. This also applies to the definition of residence in article 4 of the OECD model. Even the personal and substantive scope of application of the income tax treaty in those provisions based on articles 1 and 2 of the OECD model do not in themselves convey a benefit. In the end, such a benefit must result either from the distribution rules (articles 6 to 8, and 10 to 21) and usually in the source state, or from the methods article in the state of residence. It appears odd when the protection from discrimination derived from article 24 of the OECD model is seen as a benefit that can be obtained abusively. This is obviously how the authors of the draft see it.14 On the other hand, the proposed rule of Article X(6) can hardly affect article 28 of the OECD model, which refers to the fiscal privileges of members of diplomatic missions or consular posts. This rule merely makes it clear that the treaties based on the OECD model are not applicable. The privileges result from “the general rules of international law or under the provisions of special agreements” and not from a treaty.15

It is questionable whether the authorization to initiate a mutual agreement procedure and subsequently an arbitration procedure according to article 25 of the OECD model could also be denied by invoking the proposed rule of Article X(6). These procedural rules, however, will not in themselves contain a “benefit under this Convention,” since this procedure primarily serves to help the administrative authorities of the two states to exchange views on whether there is a risk of a taxpayer obtaining a benefit he is not entitled to.

The question whether the application of the rules on information exchange and mutual enforcement assistance can also qualify as a “benefit under this Convention” may prima facie appear ironic. In most cases, the implementation of mutual assistance does not bring a benefit to the affected taxpayers; it is often detrimental to them. On the other hand, it also gives administrative authorities the opportunity to obtain a better picture of the facts and can thus help taxpayers enforce their claims under the treaty. In this case, however, when considered in isolation no benefit results from the treaty anyway, but these procedural provisions serve the enforcement of other treaty provisions. When it comes to information that allows the application of domestic law, one can for this reason alone not speak of a “benefit under this Convention.” This also applies to the assistance in the collection of taxes referred to in article 27 of the OECD model. Occasionally, however, domestic rules (especially in EU member states) attach benefits to the condition that comprehensive mutual administrative and enforcement assistance applies to a specific state. In this case, however, a benefit resulting from domestic law is under consideration and not a treaty benefit, so that for this reason the proposed rule of Article X(6) cannot be effective.

The “benefit under this Convention,” stipulated as a requirement for the application of Article X(6), also raises other interpretation issues. Obviously, what is meant is that the tax situation would have to improve for the taxpayer as a result of the application of one or several treaty provisions as compared with the domestic legislation so that one can speak of a benefit. But does this refer to taxation in one of the contracting states, or is an overall consideration of the tax burdens in both contracting states required? If the arrangement chosen by the taxpayer is geared toward the application of a withholding tax reduction under the treaty, this would reduce the tax burden for the taxpayer in the source state but not necessarily the overall tax burden. This is because the lower withholding tax burden only has an impact on the distribution of the taxation rights between the two states if merely the state of residence benefits, which is obliged to credit a lower foreign withholding tax, thus reducing its forgone tax revenue. It is difficult to assume a “benefit” for the taxpayer in that situation. Although the higher tax remaining in the state of residence can increase the possibility of crediting other foreign taxes there, this benefit can be then at most attributed to the same treaty if the foreign tax originates from the same contracting state. When it comes to crediting taxes of other states, one cannot talk of a “benefit under this Convention.” But even if the obligation to credit taxes also results from unilateral rules in the absence of a treaty application, it is highly questionable whether there is a benefit at all as a result of the treaty.

The benefit cannot only result from the taxpayer attempting to be covered by a withholding tax reduction under the treaty, but by a more favorable one than he would otherwise be entitled to. He could therefore aim at being treated subject to the often more favorable rule of article 11 of the OECD model on interest than that of article 10 of the OECD model, or even subject to the dividend rule for parent-subsidiary situations than the portfolio investment dividend rule within article 10(2) of the OECD model. According to the public discussion draft, a “benefit under this Convention” can also exist when the benefit does not only result in comparison to the application of domestic law but also in comparison to another otherwise undisputedly applicable treaty provision.

It is more difficult to answer the question whether the proposed rule of Article X(6) will come into effect, even if the taxpayer wants to use creative measures to have its income fall under a different distribution rule with the intention of reducing his withholding tax because a higher — calculated from the gross amount — withholding tax would exceed the maximum tax credit in the state of residence and double taxation would therefore not be eliminated. In that case, the taxpayer is attempting to bring about the effect intended by the treaty (that is, avoiding double taxation). It would be odd if the tax administration of the source state denied this to him by invoking the antiabuse rule. The envisaged benefit would then probably be “in accordance with the object and purpose of the relevant provisions of this Convention.”

The overall tax situation, however, is not always the defining factor. Within the scope of application of the exemption method, a benefit may in itself result simply when the taxpayer is aiming to benefit from the legal consequences of a distribution rule more favorable to him in the source state. Similarly, the efforts of the tax- payer to fall under the exemption method instead of the credit method and thus completely rule out taxation in this country may be covered by Article X(6) in the state of residence.

Obviously, the introductory phrase “Notwithstanding the other provisions of this Convention” is meant to express that more specific antiabuse rules should not exclude the application of Article X(6). This is by no means self-evident, if one considers the opposite case law of the German Federal Tax Court on the relationship between the general antiabuse rule of section 42 of the Tax Code (Abgabenordnung) and the more specific antiabuse rules in German tax law.16 It is questionable whether it makes sense to precisely describe situations susceptible to creative setups in the treaty, only to extend the legal consequences provided therein to other situations beyond the scope of application of these specific rules through the general antiabuse rule. When, for instance, the immovable property clause of article 13(4) of the OECD model covers the alienation of shares “deriving more than 50 per cent of their value directly or indirectly from immovable property,” it is hard to understand why the general antiabuse rule should in some cases reduce the percentage to 50 percent or less.

The same applies to the application requirements of an LOB rule defined in detail in the draft of Article X(1)-(5).

The specific antiabuse rules are thus running the risk of becoming irrelevant. By the same token, however, the opposite opinion may prevail in case law. The “in accordance with the object and purpose of the relevant provisions” requirement (discussed below) could in those cases deprive the general antiabuse rule of its scope of application. After all, the object and purpose of specific antiabuse rules is to clearly define the dividing line between those arrangements that convey the envisaged benefit and those that no longer do. In any event, the coexistence of general and specific antiabuse rules leads to additional ambiguities. The attempt to vest tax administrations with both general and specific antiabuse rules, thus providing them with particularly potent weapons, may have the reverse effect: In many years’ time, courts will perhaps establish that these antiabuse rules are largely irrelevant.

‘Main Purpose’

The essential application requirement for the proposed rule is the subjective criterion — obtaining a benefit must be one of the “main purposes” in order to trigger the legal consequences of Article X(6). The difficulty of that criterion is obvious because it is impossible to prove an intention. Although the Action 6 discussion draft euphemistically speaks of the need to carry out an “objective analysis of the aims and objects of all persons involved,”17 the declared objective is to draw conclusions on the intention of the acting individuals on the basis of that “objective analysis.” Subjective criteria can always be deduced on the basis of external facts, yet the truth remains that those motives are impossible to prove. Therefore, legislators abstain, when possible, from attaching fiscal consequences to the existence of such an intention. One of the rare and no less problematic exceptions in treaty law is article 19(1)(b)(ii) of  the

OECD model, according to which the state of residence has the taxation right for income from government when the services are rendered in that state of residence and the individual “did not become a resident of that State solely for the purpose of rendering the services.”

In those cases, the rules on the burden of proof usually determine the result. If, as part of its official duty of investigation, the tax authority must furnish proof that one of the main objectives of the taxpayer was to obtain the benefit, it is already fighting a losing battle. Alternatively, the taxpayer has no chance of fending off the accusation of abuse if it is up to him to furnish evidence that benefiting from one or several treaty provisions was not one of his primary motives.

Against this background, Article X(6) does not leave this procedural question up to the domestic law of the contracting state but regulates it itself. If it is “reasonable to conclude” that one of the main objectives of  the taxpayer was to obtain the benefits of the tax treaty, Article X(6) comes into effect. On the one hand, this rule puts the ball in the tax authority’s court, which must draw this conclusion and justify it as well. On the other hand, the requirements are not too demanding — it must be merely “reasonable” but not, for instance, compelling. Therefore, the tax authority does not need to produce full evidence thereof.18 The rule thus attempts to establish a balance between the interests of the authority and those of the taxpayer.

The bias in favor of the tax authority, however, is fairly obvious, a fact that was also critically commented in several statements submitted to the OECD.19 In practice, furnishing evidence of the motives will therefore not be relevant, but tax authorities will be tempted to presume intention simply because of the presence of a benefit. For this reason alone, the subjective criterion runs the risk of not gaining any significance in itself.

The explanations provided in the public discussion draft clearly indicate that the required subjective criterion must not necessarily lie with the taxpayer who would actually claim the benefit. This is implied by the words “directly or indirectly.”20If, therefore, a group company transfers an income source to a company of the group resident in another state, and this other company can benefit from a more favorable treaty, it should be possible to apply Article X(6), although the driving force behind the arrangement is the transferring company, which from the group’s overall perspective is probably targeting a tax benefit. This benefit, however, has an effect on the receiving company. Yet it could be disputed even in this case, because without an income source, this company would probably not have a tax burden to bear. Therefore, it would not require the application of the treaty at all. If one considers this case to be covered under Article X(6), then (even if this is not expressly dealt with in the public discussion draft) that rule would also apply within a family if the mother bestows her daughter resident in another state with income-generating assets, thus achieving a more favorable tax treatment of this income due to treaties. The fact that the transfer is legally effective under civil law and thus triggers consequences under corporate and inheritance law, as well as other consequences, seems to be immaterial. As a result, the scope of application of the rule can be stretched to such an extent that the actual or alleged intention of an individual (in this case, the mother) can trigger the denial of fiscal benefits to another individual (in this case, the daughter). Denying a taxpayer benefits that he would otherwise be entitled to because of the allegedly detrimental motive of another taxpayer is problematic.

An examination of legal developments in the European Union will reveal that there are different manifestations of the subjective criterion. In its judgment in the Halifax case, the European Court of Justice demanded that the “essential aim” of the transactions concerned must be to obtain a tax advantage for an abusive practice to be found to exist.21 This definition was seized upon by the European Commission when it proposed the introduction of a general antiabuse rule in its recommendation for the fight against aggressive tax planning.22Subsequently, it used the same definition in its proposals for the adoption of a directive on the financial transaction tax, and for the amendment to the parent-subsidiary directive.23 The ECJ, however, did not remain consistent in its terminology and spoke of a “principal aim” in its judgment in the Part Service case.24 Even the European Commission changed its terminology; in article 80 of its proposal for a directive on the CCCTB, the commission demanded that transactions must be “carried out for the sole purpose of avoiding taxation.”25The European Parliament voted in favor of withdrawing these requirements and replacing “the sole purpose” with “the main purpose.”26 The OECD proposals make it easier for tax administrations to assume an abuse. It is by no means required that the “sole purpose” of the arrangement must consist in obtaining a tax benefit. It does not even have to be an “essential,” “principal,” or even the “main purpose.” Instead, it suffices if one of the main purposes of a transaction is to obtain a benefit. Therefore, the rule assumes that not merely one main purpose but two or even several main purposes can exist. So even if the taxpayer succeeds in providing proof that the arrangement he chose is also motivated by reasons other than fiscal ones, the tax authority can retort that, for the antiabuse rule to apply, it suffices if he also targeted the tax benefit. Even if the taxpayer can successfully present his nonfiscal motive as the main purpose, it may still not be enough since according to the rule, several main purposes can exist for an arrangement. The accusation of abuse applies even if the tax- payer, apart from one or even several nonfiscal main purposes, also targeted the main purpose of obtaining a treaty benefit. It remains unclear what criteria must apply in distinguishing between main purposes and secondary purposes on the one hand, and between different main purposes on the other. This distinction, however, is paramount for the application of Article X(6) and thus for the foreseen legal consequences.

The criteria discussed in the EU not only are more balanced and comparatively easier to handle, but also have an additional advantage that should not be underestimated. Their interpretation, should they be implemented, is determined in a binding manner by the ECJ, a court that is vested with the monopoly on the interpretation of EU legislation and is decoupled from the fiscal interests of the individual member states. In the long run, this at least guarantees a common understanding of the subjective criterion. Treaty practice, on the other hand, can drift apart. National courts may reach completely different judgments, and they are often unable to free themselves from the fiscal interests of their state. In dubio pro patria (when in doubt, for his country) seems to be a principle that often determines the case law on treaties. Even awards of bilateral arbitration boards do not have a binding effect going beyond the specific case. It is not even guaranteed that they will be taken into account in the interpretation of identical rules in another treaty. The result can be double taxation. The vaguer a rule is, the greater the risk that national courts will reach different results in their interpretation.

Therefore, it remains to be seen whether a rule like Article X(6), when understood as a separate element with an independent normative significance, can survive in constitutional legal systems characterized by the rule of law.27 It does not come as a surprise that the configuration of the subjective criterion was severely criticized in many of the statements submitted to the OECD.28 It is hard to ignore the conflict with the legality principle, which usually requires that rules are clear and their application is predictable and not left to the discretion of the tax authorities and the courts. Similarly, in many states constitutional provisions require the equal treatment of comparable cases. It remains to be seen whether the unequal treatment of identical arrangements, which differ from each other only in that the taxpayer has in some cases managed to present the envisaged benefit only as a secondary purpose, will suffice to meet these requirements, and it will have to be separately debated in each state and will probably have to be answered in a different manner against the backdrop of  different constitutional frameworks.

The OECD draft is attempting to defuse the criticism by repeatedly pointing out in its explanations how meticulously the tax administrations will have to consider the facts and circumstances in the application of these rules. These explanations also attempt to relax these criteria and literally relativize their  significance.

This is also reflected by the fact that the OECD invites interested parties to submit further proposals for the commentary on this rule. Some of the authors of these statements emphasize the need for clarifying explanations and examples in the OECD commentary.29 They obviously nourish hope that the rule can be mitigated by the OECD commentary. Yet these experts seem to overlook that, at the end of the day, the tax administrations and the courts will be applying the rule, not the OECD commentary.

Regardless of whether one attaches only secondary relevance to the OECD commentary in the interpretation of treaty rules according to article 32 of the Vienna Convention on the Law of Treaties (VCLT), or even sees it covered under article 31(2) or (4) of the VCLT,30 its importance will never match that of a treaty provision. An opinion in the OECD commentary that relativizes, and thus contradicts, the rule itself is most likely to be ignored altogether, and rightly so. Those who regard the subjective criterion stipulated in Article X(6) as unsuitable must oppose the rule itself.

‘Object and Purpose’

Article X(6) also stipulates another requirement, formulated as an exception: If the subjective requirement applies, the envisaged benefit will not be granted “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.” Different procedural standards become obvious here. While for the subjective criterion it suffices if it is “reasonable to conclude,” an exception can only apply if the requirements for it are “established.” Here, the bias in favor of the tax authorities is downright palpable. Many of the statements submitted to the OECD even expressed the concern that the burden of proof would rest exclusively with the taxpayer.31

The fact that this criterion was formulated as an exception is irrelevant. Although literature and unfortunately sometimes even the ECJ case law states that exception rules are to be narrowly interpreted,32today this position is evidently obsolete and methodically untenable.33 Article X(6) underscores that the question of whether a requirement is formulated as an exception depends on legal drafting and thus on coincidences in the legislative procedure. There are no reasons to ascribe a different tenor to the rule if it were formulated so that its application would also require, in addition to the taxpayer’s motive, that the granting of a benefit must oppose the object and purpose of the relevant treaty provisions. Besides, Article X(6) is designed as an exception to the otherwise granted treaty benefits; therefore, the reference to the object and purpose of the rule would be an exception from the exception and, as a result, not an exception at all but a confirmation of the basic rule. Those who take the view that exception rules should be “narrowly” interpreted would have to interpret this requirement even more widely. All of this demonstrates how ill-founded the theory of the narrow interpretation of exception rules is.

I find the fear expressed in some statements that the emphasis on an interpretation “in accordance with the object and purpose of the relevant provisions” in Article X(6) would give tax administrations the opportunity to apply the law as they please to be unjustified. For an experienced lawyer, it is evident that the object and purpose of legal provisions must always be taken into account in their interpretation. Any interpretation that contents itself with the mere wording of a rule would indeed not correspond to the state of the art of legal method. Even those holding the view that the wording of a rule is not just the beginning of the interpretation but also defines its limits will have to realize that the wording alone is most rarely so unambiguous so as to allow for the interpretation of a rule while ignoring its object and purpose. Therefore, the interpretation of all rules, both favorable and unfavorable, must always consider their object and purpose, not only when it is assumed that one of the main motives of the taxpayer was to obtain any benefit. For international law treaties, article 31 of the VCLT confirms the universal significance of focusing on the object and purpose of a rule in its interpretation.

Especially against this background, however, the question arises regarding why a separate rule is required that expressly underscores the relevance of the objective in the interpretation of treaty provisions while it also seems to attach this to the requirement that the taxpayer is attempting to obtain a benefit from the chosen arrangement. A conceivable interpretation would be to draw the reverse conclusion from it and assume for other cases — not explicitly covered by Article X(6)— that the object and purpose of the rules must be ignored in their interpretation, and that the interpreter must restrict himself to the mere wording, considering historic developments and context at most. Article X(6) would then have to be understood as a positivist interpretation rule, sending us back to the stone age of legal method, and as a deviation from the interpretation rule of article 31 of the VCLT, in which object and purpose of the rules are particularly emphasized for the interpretation of international treaties. For several reasons, however, that understanding of Article X(6) is anything but evident.

First, the explanations in the public discussion draft do not contain any clues indicating that the main meaning of this rule is to be found beyond its actual scope of application and that the interpretation of treaty law as a whole should change dramatically. Second, ignoring the object and purpose in its interpretation is hardly possible, since the individual aspects of the interpretation process, which aims at establishing the meaning as a whole, are inextricably linked with each other and cannot be split and divided at will. Third, one can barely assume that the intention of Article X(6) is to provide the benefit to the taxpayer when the authorities are unable to “reasonably conclude” that one of the main purposes of a transaction was for a benefit under a tax treaty to be obtained, even if this is not consistent with the object and purpose of the rule that governs the benefit.

Therefore, a different interpretation of this rule is more evident, which can a priori eliminate otherwise tangible concerns against it. Article X(6) stresses that a benefit under the model can only be granted if it is in accordance with the object and purpose of the model. Even if Article X(6) particularly emphasizes the case in which one of the main purposes of the taxpayer is geared toward obtaining this benefit, this rule does not exclude that in all other cases, too, benefits are only granted if this is in accordance with the object and purpose of these rules. Therefore, the interpretation of a treaty always depends on the object and purpose of the rule. Against this background, however, the motive of the taxpayer, which was also mentioned as a requirement in Article X(6), is rendered insignificant — which motives are behind the taxpayer’s actions is irrelevant. Article X(6) does not have an independent legal significance but merely underlines the already evident need for interpretation to be based on the object and purpose of the rules. Therefore, the rule is a mere hint for the interpretation and totally expendable. It is never possible to interpret rules while ignoring their object and purpose.

Legal Consequences

Those who disagree with the above opinion and regard Article X(6) as a separate exception that justifies the denial of benefits under a treaty must also address the legal consequences of this rule. The wording of the rule is concise: The benefit “shall not be granted.” The consequence that the envisaged benefit should not be granted, however, does not yet suffice to get us out of the woods. The question here is which treaty provision should be applied instead: none or a different one, and, if so, which one?

If, for instance, a shareholder capitalizes his company resident in the other contracting state with debt instead of equity so that the consideration falls under the treaty provision based on article 11 (and not article 10) of  the OECD model and so as not to pay any   taxes instead of the 15 percent withholding tax provided for in article 10 of the OECD model — because the specific treaty allocates the right to tax interest exclusively to the state of residence — the application of Article X(6) will result in the non-application of the article on interest. Does this, however, mean that none of the distribution rules of the treaty apply, and that the source state collects the withholding tax provided under its domestic law, so that the treaty is no longer relevant? Or is the rule based on article 11 of the OECD model to be ignored and the catch-all provision of article 21 of the OECD model to be applied instead? As a rule, however, this would not alter the result, since under this provision the source state does not have a taxation right either. Or is the legal consequence of the provision based on article 10 of the OECD model relevant instead, and the taxation right should, for instance, be reduced to 15 percent? According to the meaning of the rule, the option last mentioned is the more obvious. Yet Article X(6) only refers to the rule that should not be applied, without determining the application of the “suitable” rule. The reduction to 15 percent is only justifiable if one only considers the envisaged reduction of the withholding tax from 15 percent to 0 percent as the “benefit” to be denied, but not the application of the article on interest.

This also raises the question regarding what this means for the state of residence. Article X(6) does not explicitly address only one of the two contracting states, so that the provision could be applied in both states. It does, however, refer to the benefit that the taxpayer, based on the chosen arrangement, only envisaged in the source state, at least in the underlying example, and which only there was denied to him. If, however, the consequence thereof is that the state of residence continues to apply the interest article of the treaty and, as a result, considers itself as being exclusively entitled to tax the income (without crediting a tax collected in the other contracting state) this will lead to double taxation. It is questionable whether such a result is conclusive. It may seem fair at first that those who are intent upon exploiting the benefits under a treaty should accept that they may not obtain the envisaged benefit but also not become exposed to double taxation. The wording of Article X(6), however, contains no indication that those penalizing consequences are to be drawn from this rule. Income tax treaties do not contain criminal law provisions. Those who, like the OECD Committee on Fiscal Affairs in its partnership report and subsequently in the commentary on the OECD model, solve qualification conflicts between the source state and the state of residence by stipulating that the residence state must avoid double taxation34 should also plead this case here. When withholding tax is collected in the source state “under this Convention,” this could prompt the state of residence to credit this withholding tax or to exempt the income. The only objection against this would be that this qualification conflict is based not on diverging national regulations but on the application of different treaty provisions in the two states.35 The wording of article 23A and B does not, however, make a distinction according to whether the other state regards itself as en titled to taxation “under this Convention” because of this, since this qualification conflict originates in domestic law.36

Those who consider the legal consequences provided for dividends appropriate and therefore advocate a withholding tax not exceeding 15 percent in the source state by invoking Article X(6) could also be in favor of the application of the same statutory provision in the state of residence and demand the crediting of the withholding tax there under the treaty provisions based on articles 10 and 23 of the OECD model. This would eliminate the otherwise imminent risk of double taxation. From the point of view of the state of residence, however, this would raise the question as to why this state must waive tax revenue because the taxpayer aimed at a tax benefit not intended for him and therefore fell under the scope of Article X(6). One could argue that if the arrangement had been appropriate beforehand, the shareholder would have financed his company with equity in the first place and would have received dividends accordingly. In this case, the state of residence would have to credit a withholding tax not exceeding 15 percent in accordance with the treaty provisions based on articles 10 and 23 of the OECD model.

The problem is even more complicated in situations involving three states. The following case could illustrate this: A taxpayer resident in state A holds shares in a company resident in state C and transfers his mostly debt-financed participation in the company to a subsidiary resident in state B, due to the lower withholding tax rates admissible in the state B treaty. The allegation is that this “main purpose” will cause the rule otherwise provided for in the treaty between C and B not to be applied, according to which the withholding tax must be reduced to 10 percent. If the treaty between A and C provides for a maximum of 15 percent withholding tax for dividends, does this mean that the denial of the envisaged benefit under the treaty between B and C will result in state C having to apply the article on dividends from the treaty between A and C instead, or that at least the legal consequence stipulated therein becomes relevant and the withholding tax is limited to 15 percent?

The underlying and more far-reaching question is whether Article X(6) ultimately feigns that the dividends from the company resident in state C are paid directly to the partner resident in state A, and which consequences this will possibly have for the other states and taxpayers involved. Must state A therefore credit the withholding tax of state C admissibly collected according to the treaty between A and C to the income tax collected on the level of the partner, although the dividends originate from the company in state B and thus not from state C at all? Under which treaty is  state A obliged to do so? Or does the obligation of  state B to credit the withholding tax paid in state C according to the treaty between B and C also remain in effect? And if this obligation remains in effect, must the amount to be credited remain unlimited (that is, must the tax be credited regardless of the amount collected in state C according to its domestic law)? Or must the tax be credited in accordance with the dividend article of the treaty between B and C, which, however, is not applicable to state C due to Article X(6)? Or must state B credit the highest withholding tax admissible in state C according to the treaty between A and C?

These questions are only touched upon here. But even if they were addressed in greater detail, the answers would not be more gratifying. The dilemma remains the same if one understands Article X(6) as a taxation rule with its own requirements and legal consequences. If one regards this rule only as an order to abolish specific treaty benefits in one state, this rule leads to double taxation and has penalizing effects.

Whoever refuses to accept these consequences must derive from Article X(6) a fiction that goes beyond the arrangement considered appropriate. One would then be facing similar difficulties to those often encountered in the application of GAARs: Those who interpret a legal provision as an entitlement to the fiction of facts must answer the question of how far this fiction reaches and when it will be replaced again by the real facts.37 Reestablishing the actual facts of taxation will lead to seemingly arbitrary and thus unsatisfactory consequences.

Concluding Summary

The analysis carried out here shows that upon closer examination, the proposed antiabuse rule turns out to have no legal relevance. Yet the mere existence of this rule will cause uncertainty, and individual tax administrations and courts will not be deterred from using it as a basis for the denial of treaty benefits. If such a rule is actually introduced into income tax treaties, its interpretation and application will be dealt with intensively both in practice and in theory. The criticism contained in the statements submitted to the OECD is directed against the design of the rule, unilaterally geared toward the interests of tax administrations, but not against such a rule altogether.38 The difficulties noted herein, however, are not only attributable to the details of the rule. Similar and equally justified criticism would also be directed against any other tax rule that makes combating tax avoidance subject to a combination of  objective and subjective requirements.

At first glance, the OECD proposal to introduce such a general antiabuse rule in the treaty is a convenient solution for all sides. In this manner, the OECD would have implemented part of its action plan and introduced measures. That quick success can be sold in the media as another tool in the fight against tax evaders. Tax administrations would receive an additional instrument to deny tax benefits that seem suspicious to them. Understandably, companies reluctantly resist such a rule, which is often directed primarily against its proposed design.39 Many obviously presume that a careful preparation will allow them to dispel the impression that an envisaged tax benefit had priority in the planning of a transaction. The resulting additional effort required for tax planning considerations will, in turn, benefit consultants the most. If they succeed in dispelling the suspicion that obtaining a benefit was a main purpose, companies and consultants can rely on the mere wording of the treaty provisions for their tax planning considerations and ignore the object and purpose of the rules that could pose an obstacle to their envisaged success.

It is precisely for this reason that such a rule is not merely superfluous but also detrimental to legal culture.40 When it is not certain whether an arrangement chosen by a taxpayer is covered by a treaty provision, practitioners will not ask for the object and purpose of the treaty provision but will apply a vague antiabuse provision instead, and the facts will be assessed based on the practitioner’s legal instinct and not from the law. Alternatively, tax administrations and courts will be tempted to restrict themselves to the alleged clear wording and to ignore the object and purpose of the legal provisions. The application of law, however, will eventually be impoverished if it limits itself to the “primitive positivism of the naked word.”41 The late German Supreme Court judge Ludwig Schmidt pointed out in section 42 of the German Composition Code (Abgabenordnung) that a good lawyer does not require an antiabuse rule,42 since he will apply the teleological interpretation. A weak lawyer, on the other hand, will clutch at the straws that general antiabuse rules seemingly offer and will hope to avoid the often painstaking and demanding analysis of the object and purpose of the rule by resorting to a rule that allows him to replace the interpretation of the law with his subjective sense of justice. In his closing arguments in Cartesio, former Advocate General Luís Miguel Poiares Maduro described in reference to Gutteridge the abuse of rights principle as “a drug which at first appears to be innocuous, but may be followed by very disagreeable after effects.”43 The OECD should keep its hands off it!44

Hope can be found in the position of the United States, which has previously managed to ensure that the BEPS proposals do not run contrary to its interests. The LOB rule found in Article X(1)-(5) of the draft, which largely corresponds to the rule in U.S. treaties and was intentionally not included in other treaties by many countries, testifies to this. In this manner, the OECD model is adjusted to the U.S. model. In contrast, rules similar to Article X(6) have so far been rejected by the United States. The U.S. Senate has even made the approval of the ratification of already negotiated treaties conditional on the deletion of a “main purpose test” contained in the treaty.45 The reservations expressed at the time against such a rule were clearly articulated by the Senate committee:

The new main purpose tests in the proposed treaty are subjective, vague and add uncertainty to the treaty. It is unclear how the provisions are to be applied. In addition, the provisions lack conformity with other U.S. tax treaties. This uncertainty could create difficulties for legitimate business transactions, and can hinder a taxpayer’s ability to rely on the treaty.46

These arguments are as convincing now as they were at the time. If, however, the OECD cannot be deterred from introducing a general antiabuse rule similar to Article X(6) into the model, it will not only impair the quality of the OECD model but will also contribute to its loss of significance because not all states will be able to accept such a rule in their treaties.  

(Editor: Zhang Chi)

Personal details:

Michael Lang is head of the Institute for Austrian and International Tax Law at WU (Vienna University of Economics and Business) and director of the postgraduate LLM Pro- gram in International Tax Law and of the Doctoral Program in International Business Taxation at that university.

Footnotes: 

[1]. OECD, Action 6 discussion draft, at 5.

[2]. See the numerous and partly detailed statements in OECD, “Comments Received on Public Discussion Draft ‘BEPS Action 6: Preventing the Granting of Treaty Benefits in Inappropriate Circumstances’’’ (2014).

[3]. OECD, Action 6 discussion draft, at 10.

[4]. OECD, Action 6 discussion draft, at 5.

[5]. See id.

[6]. See OECD, Action 6 discussion draft, at 10.

[7]. See Michael Lang, ‘‘Direkte Steuern und EU-Steuerpolitik — Wo bleiben die Visionen?’’ IStR 2013, 365 (367f).

[8]. Proposal for a Council Directive on a Common Consolidated Corporate Tax Base (CCCTB) COM(2011) 121/4, at 48.

[9]. Proposal for a Council Directive implementing enhanced cooperation in the area of financial transaction tax COM(2013) 71, at 30.

[10]. Proposal for a Council Directive amending Directive 2011/ 96/EU on the common system of taxation applicable in the case of parent companies and subsidiaries of different Member States, COM(2013) 0814, at 10.

[11]. Regarding the problem of the reverse conclusion after amendment of article 3(2) of the OECD model, see Lang,  ‘‘Tendenzen in der Rechtsprechung des österreichischen Verwaltungsgerichtshofs zu den Doppelbesteuerungsabkommen,’’ IFF 2012, 26 (32).

[12]. See Lang, IFF 2012, 26 (32 et seq.).

[13]. See Lang, ‘‘Her Majesty the Queen v. Peter Sommerer: Abkommensrechtliche Fragen der Zurechnung von Einkünften,’’ in: Fitz, Kalss, Kautz, Kucsko, Lukas, and Torggler (publisher), Fest- schrift für Hellwig Torggler (2013) 713 (722f).

[14]. See remarks in OECD, Action 6 discussion draft, at 11.

[15]. See Daniel S. Smit, ‘‘General Report,’’ in: Lang, Pistone, Schuch, Staringer, and Storck (eds.), Tax Rules in Non-Tax Agreements (2012), 1(7).

[16]. See Federal Fiscal Court, Dec. 15, 1999, I R 29/97; Jan. 19, 2000, I R 94/97; Mar. 20, 2002, I R 63/99; May 31, 2005, I R

74/04.

[17]. OECD, Action 6 discussion draft, at 12.

[18]. Similarly, id.

[19]. See supra note 2.

[20]. See OECD, Action 6 discussion draft, at 6.

[21]. Halifax (C-255/02), Feb. 21, 2006, I-1609, para. 75.

[22]. Critical comments on this by Lang, ‘‘‘Aggressive Steuerplanung’ — eine Analyse der Empfehlung der Europäischen Kommission,’’ SWI 2013, 62 (62ff).

[23]. See supra notes 9 and 10.

[24]. Part Service (C-425/06), Feb. 21, 2008, I-897, para. 45.

[25]. See supra note 8.

[26]. European Parliament legislative resolution of Apr. 19, 2012, on the proposal for a Council directive on a Common Consolidated Corporate Tax Base (CCCTB) COM(2011) 0121, C7- 0092/2011,  2011/0058(CNS).

[27]. On the debate in Austria, see Christoph Ritz, BAO- Kommentar 5 (2014), section 22 Rz 5.

[28]. See supra note 2.

[29]. Id.

[30]. See Lang and Florian Brugger, ‘‘The Role of the OECD Commentary in Tax Treaty Interpretation’’ (2008), 23 Austl. Tax F. 95 (95ff).

[31]. See supra note 2.

[32]. See Oude Luttikhuis (C-399/93), Dec. 12, 1995, I-4515, para. 23; Commission v. France (C-384/01), May 8, 2003, I-4395, para. 28; Commission v. France (C-94/09), May 6, 2010, I-4261, para. 29; Commission v. France (C-492/08), June 17, 2010, I-5471, para. 35.

[33]. See Hans Georg Ruppe, Die Ausnahmebestimmung des Einkommensteuergesetzes (1971), at 28; Gerold Stoll, Das Steuerschuldverhältnis in seiner grundlegenden Bedeutung für die steuerliche Rechtsfindung(1972), at 104; Lang, Doppelbesteuerungsabkommen und innerstaatliches Recht (1992), at 75f.

[34]. See OECD, ‘‘The Application of the OECD Model Tax Convention to Partnerships,’’ in: OECD, Issues in International Taxation No. 6 (1999), para. 25; critical comments by Lang, The Application of the OECD Model Tax Convention to Partnerships (2000),37ff.

[35]. See Helmut Loukota, ‘‘Der Einfluss des Ertragsteuerrechtes auf die Auslegung von Doppelbesteuerungsabkommen,’’ in: Beiser, Kirchmayr, Mayr, and Zorn (publisher), Ertragsteuern in Wissenschaft und Praxis (2007), 263 (280f).

[36]. Critical comments by Lang, ‘‘Personengesellschaften und Doppelbesteuerungsabkommen,’’ in: Bertl, Eberhartinger, Egger, Kalss, Lang, Nowotny, Riegler, Schuch, and Staringer (publisher) Personengesellschaften im Unternehmens- und Steuerrecht (2013), 225 (239ff).

[37]. For more on this problem, see Lang and Christian Massoner, ‘‘Die Grenzen steuerlicher Gestaltung in der österreichischen Rechtsprechung,’’ in: Lang, Schuch, and Staringer (publisher), Die Grenzen steuerlicher Gestaltung in der Österreichischen Rechtsprechung (2009), 15 (32ff).

[38]. See supra note 2.

[39]. Id.

[40]. Already commented on by Lang, SWI 2013, 67 (on the European Commission’s recommendation for the introduction of abuse rules in the EU member states).

[41]. Walter Antoniolli, ‘‘Gleichheit vor dem Gesetz,’’ ÖJZ 1956, 646 (647).

[42]. See Wolfgang Schön, ‘‘Ludwig Schmidt (1928-2011),’’ FR

2011, 1125 (1125  f).

 [43]. Opinion of AG Maduro of May 22, 2008, Cartesio oktató és Szolgáltató bt (C-210/06), [2008], I-9641, para. 55 with reference to H.C. Gutteridge, ‘‘Abuse of Rights,’’ 5 Cambridge L.J., 22, 44,1933-1935.

[44]. On the recommendation of  the EU, see Lang, SWI 2013, 68.

[45]. Kristen A. Parillo, ‘‘Italy-U.S. Tax Treaty Enters Into Force,’’ Tax Notes Int’l, Jan. 4, 2010, p. 37.

[46]. Senate Executive Report 106-8 (1999), 4.