Multilateral Cooperation in Tax Law

S. 1341. Bilateralism vs Multilateralism

1.1. Overview

Globalization has changed the way of doing business, allowing taxpayers to exploit gaps and frictions that have always been present in existing bilateral treaties. The concept of an ambulatory OECD Model enables the OECD to better address tax issues arising from an economy that evolves over the years. Nonetheless, the OECD Model is not binding, and the effectiveness of its implementation in the tax treaty network is up to OECD member countries.

Hence, the MLI, was designed to modify existing bilateral tax treaties in a swift and coordinated manner, addressing BEPS concerns. Nonetheless, the research for a broad consensus on substantive rules is still perceived as a difficult challenge for multilateralism. Besides, states prefer negotiating tax treaties over a multilateral convention as tax treaties consider the bilateral specificities.

In light of the reasons mentioned above, in literature it is debated as to whether the shift from bilateralism to multilateralism rather implies a blending between these two overlapping forms. The following chapters aim to address the core definitions of bilateralism and multilateralism. This distinction should pave the way to a better understanding of the hybrid nature of the MLI as an instrument not entirely convincing on a multilateral basis.

1.2. Bilateralism

The dimension of bilateralism is underpinned by a double tax convention (DTC) network which, as of today, counts on the existence of 3,000 treaties. Some scholars explore the question of why states prefer to conclude bilateral treaties and point out that, above all, the respective bilateral specificities can be considered S. 135during negotiations. Precisely, the volume of trade, the investment, and either the size and the composition of the economies and their respective fiscal system might be relevant elements. Tax treaties consider existing trade relations and influence investment flows between both states. Thus, granting tax relief might represent the outcome of a negotiated balance of a tax treaty considering the respective demand of goods and services and the respective investment flows.

As economic relationships deeply vary among the states concerned, tax benefits could be granted to residents of some treaty partners while not to others. This creates the possibility of arbitrage between taxpayers seeking to take advantage from the best tax regime offered by different countries, creating situations of treaty shopping. Other kinds of distortions derive from triangular cases described in literature as something potentially able to produce tax avoidance. This might occur, for instance, when a person who is a dual resident of two contracting states receives income from a third state. In such a scenario, the source state in a dual resident triangular case may satisfy its treaty obligation by applying the tax treaty most favourable to the taxpayer. In literature, it is debated that bilateral treaties might not constitute a responsive and efficient mechanism to prevent situations stemming from the modern globalized world. Trade liberalization, the removal of currency controls, and technological developments have led to an increase in flows and capital between countries. In their quest to offer the best products at the best prices, companies have to choose where and how to invest. To achieve this result, it is natural for a company to invest where profitability is highest and tax is one of the factors of profitability. As multilateral treaties should contain common provisions applicable to all countries, some scholars have emphasized that international taxation should move toward a dimension of multilateralism.

S. 1361.3. Multilateralism

International relations scholars have conducted research aimed at defining the principles behind the definition of multilateralism. Multilateral treaties [hereinafter MTT] appeal to equal obligations and rights between parties. This is in contrast to bilateral concessions agreed through bilateral treaties. However, states are “sovereign”, and consensus on norms can easily be frustrated by reservations and declarations to multilateral instruments.

García Antón developed an in-depth study of multilateralism, emphasizing the relevance of customary international law. Following that, multilateralism is visible in the consensus on particular norms, accepted by the states, and consistently applied (state practice) out of a sense of legal obligation (opinion iuris). The author describes the pivotal role of the MTT UNCLOS III as an example of a multilateral treaty codifying customary international law in the Law of the Sea. This multilateral treaty is important in that it codifies certain principles that are, in fact, the expression of a multilateral consensus between parties. Precisely, this experience has shown that consensus in an MTT is achieved to the extent that reservations incompatible with the object and purpose of the MTT are prohibited. The Vienna Convention on the Law of Treaties [hereinafter VCLT] enshrined this idea in Article 19. Another important feature of multilateralism is the prohibition of subsequent unilateral or bilateral agreements unless they are aimed at restoring the original consensus reached on a multilateral basis in accordance with Article 41 VCLT. Finally, a binding dispute resolution mechanism to adjudicate disputes between the states is considered a further expression of multilateralism.

S. 1372. History and Purpose of the MLI different types of provisions

2.1. History of the MLI

Globalization, the 2008 financial crisis, as well as political and economic changes have made it difficult for countries to raise sustainable revenues for their economies. Globalization has changed the way of doing business by allowing taxpayers to exploit gaps and frictions that have existed in existing bilateral treaties. Precisely, the mobility of production, capital, the digital economy, and intangibles require some amendments.

At the same time, amendments to the OECD Model and its Commentaries in relation to the notion of royalties and permanent establishments “show the tension” between source and residence taxation. However, these new challenges have resulted in global revenue losses requiring countries to foster a climate of cooperation and harmonization partly to prevent other countries from taking unilateral measures to protect their tax base. Within this context, multilateral treaties are supposed to “set the framework” for a shift from bilateralism to multilateralism. Precisely, the Multilateral Convention on Mutual Administrative Assistance in Tax Matters aims at strengthening tax transparency.

Further, the international framework has witnessed the drafting of the BEPS Multilateral Convention to Implement Tax Treaty Related Measures in order to tackle aggressive tax planning as well as base erosion and profit shifting [hereinafter MLI]. The MLI, transposing the 15 points agreed by the OECD, follows from the release of the Report ‘Addressing Base Erosion and Profit Shifting’ in February 2013. It was negotiated by more than 100 states and non-state jurisdictions and, as of today, 100 jurisdictions have signed it. The project was agreed upon in S. 138the 35 OECD member countries and eight countries that are members of the G20 but not OECD members, that is to say, Argentina, Brazil, China, India, Indonesia, Russia, Saudi Arabia, and South Africa. Questions about the legitimacy of the implementation of the MLI beyond the group of 44 decision makers have been raised.

Despite the political forum involved in the Inclusive Framework, the role on the decision-making process of the developing country was very limited. The developing countries are still demanding more source taxation because of their territorial taxation system, but the global economic inequalities follow from the benefits principle. This principle states that active income, i.e. business income, should be taxed primarily at source whereas passive income, i.e. investment income, should be taxed primarily in the residence state.

On 5 October 2015, the final package of reports on the OECD/G20 BEPS initiative was released. Finally, the signing ceremony of the OECD’s “Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting” (the MLI) took place in Paris at the beginning of June 2017.

2.2. Purpose of the MLI

2.2.1. Swift implementation of tax treaty measures

The current network of tax treaties is not well-synchronized with the model tax conventions since some treaties are years behind the models on which they are based. The OECD amended its model, on average, every 2.5 years. In fact, this ensures that its model addresses the new tax issues arising with the evolution of global economy.

From the other side, the OECD Model is not binding, therefore, updating tax treaties depends on the willingness of Member States to follow OECD Model updates. Conversely, the MLI is a binding instrument that will lead to the implementation of BEPS-agreed treaty measures within a short period.

S. 139Hence, tax treaties will be modified in a synchronized way with respect to BEPS issues without the need to renegotiate tax treaties bilaterally. Below are the positions of some tax scholars who emphasize the effectiveness of a multilateral solution for the transposition of BEPS measures.

Bravo argues that a multilateral instrument could avoid the cumbersome and time-consuming renegotiation of tax treaties. Indeed, when parties renegotiate tax treaties among themselves, allegedly, they will take several negotiation rounds to discuss demands as well counter-demands or further amendments.

Broekhuijsen and Vergouwen agree in recommending the use of multilateral solutions to the OECD in order to ensure rapid implementation of updates to the tax treaty network. According to their analysis, updating tax treaties requires an average of 18 years of negotiation because of the subsequent amendments.

Avery Jones and Baker point out that the negotiation of bilateral treaties could fuel the risk that Member States may wish to open up some further issues during the negotiations. The OECD has so far followed the approach of amending the OECD Model and the Commentaries adopted alongside the OECD Model. Even though courts have been willing to have recourse to the Commentaries, the Commentaries cannot change the wording of a specific DTC. Thus, the ideal approach to amending the wording of articles embodied in a large number of tax treaties is represented by adopting a multilateral agreement.

2.2.2. Modification of the application of numerous bilateral tax treaties

The primary objective of the MLI is to transpose the BEPS agreed measures thereby accordingly modifying the tax treaties. The MLI is not going to terminate the pre-existing network of bilateral treaties, but it rather aims to achieve a concurrent application of its provisions with the bilateral treaties. The Explanatory Statement purports that the Convention will not act as an amending protoS. 140col, but it will rather be applied alongside existing tax treaties. This requires clarifying the existing relationship between the MLI and the covered tax agreements [CTA hereinafter]. Some tax scholars have addressed the question, emphasizing the content of Article 2, paragraph 2 of the MLI. This article covers the undefined terms of the Multilateral Convention suggesting to refer back to the CTA “unless the context otherwise requires”. Hence, the provision creates an interesting “circularity” because the MLI is supposed to amend the covered tax agreements but, on the other hand, the MLI contains very few definitions of terms.

Lang convincingly argues that an autonomous interpretation of the context of the MLI has to be preferred over the straightforward recourse to the covered tax agreements. However, given the connection between the MLI and OECD Model, an autonomous interpretation of the MLI is “extremely rare”. This argument leads tax scholars to spell out the "bilateral mindset" embedded in the MLI as a multilateral instrument mixed with bilateralism.

A second relevant argument concerns Article 30 that allows the parties to amend the CTA bilaterally after the modifications effected by the MLI. Blum and Szudoczky stress the importance of the provision as an expression of “minimalist multilateralism” of the MLI. Indeed, this article permits keeping the covered tax treaties unaffected by the provisions of the MLI. Even more, this provision does not even carve out the minimum standards from its scope.

Hattingh conducted an analysis aimed at estimating the effectiveness of the MLI in changing the tax treaty network. It follows that the MLI will effectively change less than 100 or even 50 out of 3,000 tax treaties in force, particularly regarding non-binding measures. Also, some countries (e.g., Germany, Mauritius, and Switzerland) have agreed to renegotiate bilaterally tax treaties in order to implement the BEPS minimum standards.

S. 1412.2.3. Reduction of the opportunity for tax avoidance

As discussed above, it is possible to infer the presence of consensus in a multilateral treaty, insofar as reservations incompatible with the object and purpose of the MTT are prohibited. According to the Explanatory Statement, “the object and purpose of the Convention is to implement tax treaty-related BEPS measures”. Among the actions recommending changes to tax treaties, only the provisions implementing minimum standards are binding on the parties. The minimum standards’ provisions have gained legal effect due to the legal quality of the MLI as a multilateral contract. It was agreed in the Explanatory Statement that a number of provisions are minimum standards meaning that countries have agreed to implement these standards.

The Final Report on Action 6 represents one of the BEPS minimum standards, and it is aimed at preventing treaty abuse. Regarding the reservation in Article 7 to prevent abuse of the Treaty, the following comments are made. According to Article 7(15) MLI, parties may reserve the right to not apply the principle purpose test as the default option on the grounds of combining a detailed limitation-on-benefits provision and either rule aimed at addressing conduit financing structures. In this scenario, both parties are compelled to reach a mutually satisfactory solution that meets the minimum standard.

García Antón pointed out that this provision is clear proof that, despite having a minimum standard, there is no consensus yet on anti-avoidance. Much more importantly, there are grounds to say that this provision discloses the bilateral nature of the MLI as Article 7(15) MLI relies on a bilateral negotiation between contracting states instead of a consensual multilateral solution.

Further, multilateral consensus exists if the subsequent bilateral agreements are prohibited unless they are aimed to restore the consensus achieved on a multilatS. 142eral basis. Referring to Article 30 of the MLI, Bravo has pointed out that this provision safeguards the BEPS minimum standard as these latter represent the purpose and object of the MLI. This argument is not supported by other academics. They argue that the provision allows parties to adopt the MLI without being in breach of any obligation if they maintain tax treaties without implementing the minimum standards.

2.2.4. Solution to close the gaps in existing international tax rules

A holistic approach requires verifying whether states agreed a binding dispute resolution mechanism. As mentioned above, this further represents an element of a multilateral consensus among states. The MTT on the Law of the Sea obligated states that became party to the convention in order to choose between a dispute settlement mechanism. Otherwise, states would be free to act in a way that “run counter” the agreed terms and conditions of the conventions.

The MLI’s response, in this regard, is represented by the norms on arbitration entailed in Part VI in which the content is optional. Indeed, on 18 December 2014, a discussion draft on Action 14 was published. This document showed that there was no consensus among states to move towards the adoption of mandatory and binding arbitration for all tax treaties.

Pistone and Čičin-Šain argue that “a giant leap” would have been made in making mandatory binding arbitration mandatory. Tax arbitration represents the only quasi-judicial procedure able of settling a cross-border tax dispute with the involvement of a third party. The authors therefore suggest that, in a future review, the whole of Part VI of the MLI should be converted into a minimum standard.

S. 143On the other hand, the intensification of cross-border activities with an increase in the number of tax disputes and taxpayers craving clarification on the application of tax treaties led to the improvement of mutual agreement procedure (MAP) provisions. The provisions regarding the MAP are part of the minimum standard and are embodied in Article 16 (1) MLI. This norm reproduces the content of Article 25 of the OECD Model Tax Convention with one exception. Article 16(1) of the MLI states that a MAP request may be addressed to either of the competent authorities of the relevant jurisdictions. Instead, Article 25 of the OECD Model states that a MAP request may be addressed to the competent authority of the taxpayer's jurisdiction.

Article 16 (5) MLI allows parties to opt out on the grounds that they intend to meet the minimum standard for improving dispute resolution under the OECD/G20 BEPS package. Precisely, a peer review process made up of two stages ensures that countries have implemented the minimum standard.

García Antón stressed that few states signed up for the arbitration provisions in Part VI of MLI. This reveals a “mistrust” of a separate, impartial, non-national arbitration body to adjudicate international tax disputes. The minimum standard of Article 16 of MLI encourages both authorities to solve the case. Nonetheless, the provision does not prevent domestic interference in the resolution of international tax disputes. Only by obligating signatory states to choose an independent dispute settlement mechanism can a multilateral consensus be achieved to the extent that states have agreed on the terms of the treaty and its application.

3. Different types of MLI provisions

3.1. Overview

The provisions of the MLI can be divided into three categories. The first are the provisions forming part of what is known as the minimum standard, and, secS. 144ondly, the provisions that apply upon ratification of the MLI provided that “a party” made no reservation. Thirdly and lastly are the provisions that do not apply unless parties specifically choose to apply them. This chapter first discusses the provisions that are part of the minimum standard. It then examines the other two types of provisions.

3.2. Provisions required in order to meet a minimum standard

The provisions implementing the minimum standards are Article 7 on the prevention of abuse of the treaty and Article 14 which aims to implement the provisions on the mutual agreement procedure. Article 6 concerns the text of the preamble, and Article 17 of the MLI concerns the corresponding adjustments in relation to Article 9, paragraph 2 of the OECD Model.

The principle purpose test has been widely adopted by the signatory countries. The wide acceptance of this provision has led some academics to argue that it could become international customary law on a large scale.

Mosquera Valderrama writes that the provision’s widespread adoption demonstrates that it is a state practice that is applied “uniformly, extensively, representatively, and consistently by countries”. Avi-Yonah argues that the PPT will probably become a CIL as most of the world accepts it.

It has already been observed that there is no multilateral consensus on anti-avoidance rules. The following analysis aims to examine whether the widespread acceptance of the PPT could constitute customary international law (CIL) and enhance the coordination of international tax rules.

The establishment of an international general anti-avoidance rule (GAAR) with exactly the same wording through the network of tax treaties offers such a compelling opportunity. However, despite the quantitative argument, in order to S. 145consider the PPT as a customary rule, it is necessary to assess the existence of a practice and whether it is general. Some elements make it “doubtful” that a uniform and consistent practice will arise. First of all, the definition of a PPT is not a self-executing norm but rather strongly dependent on a factual analysis but also on a typology of rules tackling tax avoidance. Indeed, typology and scope of the rules aimed at preventing tax avoidance vary from country to country. Further, putting a GAAR into practice requires an analysis of the facts at stake as well as an interpretation of the GAAR that is fiercely influenced by the domestic legal cultural of abuse. In a nutshell, the interpretative task examining the elements of the GAAR would vary among the concerned jurisdictions.

Further, according to tax academics, the interpretation of subjective and objective elements of the PPTjeopardizes the achievements of widespread practice. Thirdly, the burden of proof has also sparked criticism among academics. Indeed, the PPT provision overburdens the taxpayer, who has to interpret the object and purpose of a provision in a treaty, which is more appropriate for the tax authorities. Further, it may conflict with the burden of proof according to the domestic GAAR.

Elliffe supports the argument that drafting the TPP has failed to achieve a common international understanding. This is due to the fact that there are no defined terms in Article 7 and, therefore, Article 2(2) MLI refers us back to the tax treaties “unless the context otherwise requires”. From the other side, a common context is not feasible as the absence of similar provisions in the DTCs implies the establishment of a new treaty-based GAAR.

Concluding, the above reasonings do not support the argument that the PPT should strengthen the coordination of international tax rules.

3.3. Provisions that exceed the minimum standard

Another set of provisions are those that exceed minimum standards whereby parties are free to reserve their rights to the application of such provisions even S. 146in their entirety. Specifically, Article 3 on transparent entities and the provision on dual resident entity in Article 4 fall into this category. Further is Article 8 MLI on the Dividend Transfer Transactions, the proper treatment of income attributed to the permanent establishment situated in third jurisdictions (i.e. Article 10), and the rules against the artificial avoidance of permanent establishment status (i.e. Article 12-15 MLI). The following analysis will cover some of these provisions.

Article 3(1) of the MLI is intended to tackle the granting of treaty benefits to persons whose residence status varies among the states concerned. The provision adheres to the findings of the OECD Partnership Report and reproduces the content of Article 1(2) OECD Model Convention.

Pistone and Čičin-Šain observed that only 30 jurisdictions have not opted out of this provision. The authors argue that the adoption or rejection of Article 3 depends on whether the jurisdiction in question has transparent entities. The importance attached to the OECD Partnership Report is another element to be considered.

Article 4 of the MLI deals with the determination of the tax residence of dual resident companies. It codifies the switch from the concept of place of effective management as tie-breaker rule to a mutual agreement procedure. Many countries reserved the right to not apply Article 4 of the MLI to their DTCs. One of the reasons for this is that the countries' reports have demonstrated the potential costs, outlining a reallocation of human and financial resources within the tax administration. Moreover, this reallocation may not even be necessary to meet this need. Other reasons for the limited adoption of this rule are related to the fact that there are more appropriate provisions to tackle tax avoidance (e.g. Article 7).

The MLI contains three articles (Articles 12-14) restricting strategies to avoid establishing a permanent establishment. Below is one of them.

S. 147Article 12 addresses the commissionaire arrangements and other similar structures eroding the taxable base of the source state. Under the provision, a person that acts on behalf of a foreign enterprise in that jurisdiction will constitute a PE in that jurisdiction if the person “habitually concludes contracts, or habitually plays the principal role leading to the conclusion of contracts that are routinely concluded without material modification by the enterprise”. It should be noted that the definition of dependent agent already existed in the pre-BEPS era so, in some jurisdictions, it is not necessarily derived from the MLI. In Italy, for example, the Supreme Court in the Philip Morris case broadened the interpretation of agency PE.

3.4. Provisions that do not apply unless specifically chosen by parties

The provisions that will apply to the DTCs, as the parties may choose to apply, are the following. The options for the applications of methods for the elimination of double taxation under Article 5 of the MLI, the options of the anti-fragmentation rules against the artificial avoidance of the permanent establishment status (i.e. Article 13 MLI), the simplified limitation on benefits (LOB clause), and, finally, agreement to arbitrate tax disputes not resolved through a MAP.

Article 5 of the MLI allows parties to choose between one of the three options or to opt out from the provision. Option A provides for a change of method in cases when differences in qualification result in double non-taxation. Option B provides a method switch in cases of divergent tax treatment of dividends in two jurisdictions resulting in double non-taxation. Finally, Option C provides a general switch to the credit method. This article contains a degree of flexibility. The exemption method is one of the most widely used methods in the network of tax treaties. It is reasonable to assume that states do not want to give up the exemption method for reasons of tax attractiveness. If, for example, the other party concludes treaties with neighbouring countries that only provide the exemption method, states may fear losing their tax attractiveness.

S. 148Article 13 MLI is intended to limit the application of the exemptions listed in Article 5(4) OECD Model to the extent that these activities are, in fact, preparatory or ancillary. Some countries opposed the view that the activities listed therein are in themselves auxiliary or preparatory. Hence, the article is composed of two options. Option B maintains the exception provided in covered tax treaties irrespective of whether the activity is preparatory. Under Option A, the exception applies only if the activity is preparatory or auxiliary.

4. The flexibility of the MLI

4.1. Overview

In attracting countries to sign up, the MLI had to maintain a high degree of flexibility. This flexibility is achieved in a number of ways, including a mechanism under Articles 1 and 2 of the MLI. According to Article 1, the MLI modifies all covered tax agreements. Article 2 defines a covered tax agreement as a double tax treaty with respect to which each Party has made a notification to a Depositary administered by the OECD. Thus, flexibility is achieved to the extent that a party may decide to exclude the application of the MLI by not notifying its tax treaty to the Depositary.

Another element of flexibility lies in the fact that the minimum standard can be met in different ways. Finally, a high degree of flexibility has been introduced in Article 28 MLI that explicitly lists the provisions that allow opting out of the multilateral instrument. In the following, the reservations are dealt with first.

4.2. Reservation power

As discussed in general terms in the first chapter, reservations could hinder the achievement of a multilateral solution. Therefore, it is important to consider the nature of the reservations contained in the multilateral instrument. A reservation is a unilateral declaration by which a state intends to exclude or modify the legal effect of certain provisions of the treaty. Reservations should maintain the flexibility of the instrument, allowing parties to opt out without compromising the core objectives of the multilateral instrument.

S. 149In his research, Garcia Antón pointed out that the reservations of the MLI do not fully match the reservations provided by the Vienna Convention of Law of Treaties. This is the result of the fact that there is no multilateral consensus between the parties. Hence, as the bilateral mindset of the instrument persists, a reservation “as a platform to reach mutual solutions” is entailed in the Article 7(15). This provision allows parties to opt out of the PPT provided that a mutually satisfactory solution that meets the BEPS minimum standard is reached.

In addition, Article 5(9) MLI echoes the same theme in that it allows a contracting state to not permit the other state to opt for Option C, specifically the credit method, instead of the exemption method. In this respect, the Explanatory Memorandum allows the parties to negotiate the above restriction through Option C on the basis of bilateral negotiations. It would seem that bilateralism above all supersedes multilateral consensus according to these latest provisions.

On the other hand, the multilateral nature of the instrument is supported by the symmetrical effects of the other reservations as well as by the prohibition of late reservations enshrined in Article 28(9) MLI. Based on reservation power, the MLI turns out to be a unique hybrid in which bilateralism and multilateralism mix and persist.

4.3. Opt-in and opt-out clauses

Another tool that ensures the overall flexibility of the multilateral instrument is the possibility of opting in or out of certain provisions. The distinction between reservations and opt-out clauses is debated in the literature. Although the OECD distinguishes between reservations and opt-outs, there is consensus that both mechanisms serve the same purpose. Opt-in clauses refer to provisions that are not mandatory, therefore, they are not automatically applicable when entering into the MLI.

S. 150By opting-in, the parties undertake additional obligations. These provisions can be construed in a way that allow a choice among different alternatives, including the possibility of applying none of them. Opt-in clauses are unilateral declarations intended to extend the obligations of a state. This is contrary to opt-out clauses that reduce the scope of the multilateral instrument.

The most important part affected by the application of this provision is Part VI of the MLI concerning the mandatory binding arbitration procedure. The implications of these provisions for the implementation of the MLI have already been discussed in the previous section.

4.4. Compatibility clauses

In drafting compatibility clauses, the MLI has adopted the approach of using one compatibility clause for each provision at stake. The issue of compatibility clauses also concerns the nature of the instrument and its relationship with the CTA. Indeed, according to the MLI’s Explanatory Memorandum, the instrument must be interpreted on a stand-alone basis.

However, a closer examination of the compatibility clauses of the MLI raises some doubts as to whether the intended separation could be achieved. Specifically, the MLI contains four types of compatibility clauses designed to adjust the relationship between the MLI and covered tax treaties. Thus, the MLI could replace an existing provision or be applied in the absence of existing provisions following the format drafted accordingly.

The MLI is a multilateral treaty aimed at amending existing tax treaties so it is not a stand-alone international agreement, and its scope and application will depend on the tax treaties covered.

S. 1515. Conclusions

The overall assessment of the MLI reveals the unique hybrid nature of this instrument where bilateralism and multilateralism persist and collide. It results in a mysterious “chiaroscuro effect” in which states have agreed to implement the four minimum standards, but apparently no multilateral consensus is achieved; rather, a bilateral mindset can be inferred from many elements. The MLI is a multilateral instrument that aims to amend tax treaties but, in fact, it exists on the border between bilateralism and multilateralism. The arguments in support of this conclusion can be summarized as follows.

Throughout history, states have experienced three different models of international governance. Bilateral treaties are expected to protect foreign investments. Multilateral treaties blossomed in fields such as human rights, the Law of the Sea, and climate change. As bilateral treaties are expected to protect foreign investments, states are very reluctant to adopt multilateral instruments of governance instead.Thirdly, regionalism constitutes a “thick multilateralism”, meaning “a self-conscious effort to construct regional identities by the use of multilateral identities and organisations”.

With regard to the multilateral nature of the MLI, it should be noted that the MLI currently has 100 signatories and parties. Further, the multilateral nature of the MLI transpires from elements such as the prohibition of late reservations that follow the logic of protecting the consensus that is reached. Moreover, the multilateral nature of the MLI is reflected in the symmetrical effects of the reservations in Article 28(3) of the MLI.

Otherwise, the MLI is not a multilateral instrument because there does not seem to be a core multilateral consensus among the parties. The bilateral logic seems to prevail. This is illustrated by the fact that some of the reservations contained in S. 152the instrument rely on bilateral consensus. In the above terms, the reservation entailed in Article 7(15) MLI is a “platform to reach mutual solutions”.

Further evidence can be inferred by considering the role played by Article 30 MLI which allows the parties to amend the CTA bilaterally after the changes introduced by the MLI. This provision does not even carve out the minimum standards. Finally, as far as the definition of terms is concerned, bilateralism is embedded in the MLI.

The idea of advocating for a shift from bilateralism to multilateralism is misleading insofar as it does not consider that the OECD multilateral framework is premised on the benefits principle. Neither the BEPS Project nor global transparency in exchange of information aim to combat global inequalities and poverty. Rather, international relations scholars continue to promote the idea, quoting Weiler’s metaphor, that multilateralism and bilateralism overlap like geological strata, disregarding the idea of a shifting.

This is clear in the sense that, although tax treaty negotiations remain bilateral, the OECD Model, which is regularly updated, serves as a basis for negotiation and is a multilateral treaty. Moreover, BEPS Action 15 does not abandon the “bilateral mindset”. The Confédération Fiscale Européenne argued that the multilateral instrument would modify but not replace existing tax treaties thereby leaving the bilateral nature of tax treaties untouched.