Recents in Beach

header ads

FIT FOR PURPOSE? Reconceptualising arbitration clauses in international tax treaties

Iqra Bawany*




Before the early 2000s, arbitration in tax treaties was relatively uncommon. However, this trend was reversed in 2008 when the Organisation of Economic Cooperation and Development (OECD) introduced a dispute resolution clause in the OECD Model Convention for Tax Treaties. This move was followed by the UN’s inclusion of an arbitration clause, in 2011, in the UN Model Convention for Tax Treaties. It must be noted from the outset that the OECD model is typically targeted towards bilateral tax treaties signed between developed OECD member countries. On the other hand, the UN Model targets taxation treaties between developed and developing countries and was designed to accommodate the needs of developing countries which are typically disadvantaged by tax treaty negotiation.[1] Nonetheless, the OECD Model is far more popular with both developed and developing countries and has been expanded in scope to accommodate the popularity of the convention as demonstrated by the OECD’s introduction of a more inclusive framework for its recent Base Erosion and Profit Shifting (BEPS) project.


Since the rapid expansion of the OECD Model Convention to over 135 countries, tax treaty disputes have increased significantly as a result of various measures that have been introduced to address the issues of tax avoidance and treaty abuse.[2] Moreover, as international competition for tax revenues intensifies between states, so do issues of double taxation. This combined with the digitisation of the global economy, which has brought the question of the taxation of multinational enterprises (MNEs) to the fore, has made tax treaty disputes more important than ever.  As MNEs engage in more thorough evaluations of jurisdictions for their activity and have unparalleled flexibility to move jurisdictions at their convenience, tax considerations have become increasingly important in driving business decisions. Therefore, there was a desperate need to create a ‘dependable, independent and treaty-based mandatory dispute resolution [system] as a cornerstone of BEPS’ which was the ‘only realistic antidote to predictable tensions.’[3] This necessitated the introduced of an arbitration clause in the respective taxation conventions, especially given that, arbitration has natural advantages as an alternative dispute resolution mechanism for international tax treaties, given its ‘quasi-judicial method’[4]. However, the OECD and UN’s introduction of arbitration clauses has been less than satisfactory and is not nearly robust or effective enough to deal with the actual issues arising out of tax treaty disputes.


Therefore, this essay is split into three sections. The first is a brief overview of the development of methods of dispute resolution such as the mutual agreement procedure (MAP) to establish why it was necessary to introduce arbitration clauses into the OECD and UN Model Conventions. The second section focuses on the OECD and UN Model Conventions arbitration clauses. The aim of this section is to illuminate why arbitration is still underutilised in the case of tax treaty disputes and the issues that have arisen from these arbitration clauses. Lastly, this essay will look towards the future and how arbitration can be reconceptualized in the context of tax treaty disputes to increase its utilisation. It must be noted though that since the introduction of arbitration in OECD and UN Models is so recent, information about the actual experience of countries with arbitration is ‘scarce and at best anecdotal’. [5]




Between the First and Second World Wars, there was a trickle of bilateral double taxation agreements under the guise of the League of Nations 1924 Model Double Taxation Agreement. While at the time it was not felt that there was a need to include a dispute resolution mechanism, the Agreement did include a technical committee to decide on double taxation disputes. [6]


In 1943, Mexico Draft Convention was established and followed by the 1946 London Draft Convention, which both led to an explosion of bilateral double taxation agreements, with over 70 treaties being signed between 1946 and 1956.[7] As a result, rules for a mutual agreement procedure (MAP) started to emerge within following draft conventions in order to tackle the international problems arising from double taxation treaties such as difficulties in the interpretation of the provisions of a convention or cases of double taxation not covered by the guiding principles of the chosen convention.[8] In essence, MAP is a mutual negotiation-based dispute resolution mechanism between the two treaty partners of a Double Tax Agreement (DTA).[9] While the form of MAP can vary, it typically involves meetings between the relevant, competent tax authorities to consult to resolve disputes regarding the application of the DTA and these authorities may appoint a commission of representatives to oversee a case.


Rules for MAPs started to become more and more common in order to authorise the competent authorities to enter dispute resolution negotiations with regards to individual cases that were not caught by the convention.  The OECD’s Draft Convention on Income and on Capital formalised these rules in 1963 therefore making the MAP the foremost mechanism for resolving tax treaty disputes. Subsequently, the OECD’s Committee on Fiscal Affairs set about revising its dispute resolution mechanisms from 1971 onwards. Finally, in 1977, the OECD adopted its Model Convention for Tax Treaties which codified the MAP in Article 25 of the convention. This convention is the basis for over 3000 tax treaties and serves as the framework for the negotiation of bilateral tax treaties.[10] The importance of the MAP ‘stems from the fact that it provides taxpayers with an alternative to tax litigation, which can be cumbersome and uncertain.’[11] Therefore the MAP is itself an alternative dispute resolution mechanism.


The problems with the MAP are multifaceted. The procedure itself is significantly limited as it only ‘contemplated the appointment of commissions to serve as consultative bodies for each particular dispute’ and did not provide a binding legal authority or instrument to resolve them.[12] Moreover, the Convention lacked a clear procedure or mechanism for dispute resolution, and nor did it provide clarity on how taxpayers were able to utilise the procedure to bring forward disputes. Even after its revision in the 1977 OECD Model Convention, MAP was not a strong solution for the dispute resolution problems in international tax disputes. The revised procedure built off the original MAP mechanism but only entertained the possibility of a joint commission for the purposes of oral negotiations via representatives. The OECD does not specify how a competent authority could or should be formed but only protects taxpayers’ rights to make their case in front of such an authority should it emerge.[13]


Furthermore, more and more tax treaties have been concluded since the turn of the century particularly by developing countries who are keen to attract business from MNEs. As more Foreign Direct Investment (FDI) flows into emerging economies, there have been increased numbers of disputes where DTAs are applicable. However, the tax administrations of developing countries often lack the experience and resources needed to deal with the negotiation-based MAP or have been denied access to both MAP and arbitration.[14] The underutilisation and improper functioning of the MAP mechanism has directly necessitated the broadening of the arbitration options available under the OECD and UN model Tax Conventions. The known weakness of the MAP, including the inability of the competent authorities to effectively conclude agreements, have made the expansion of arbitration clauses in the conventions much more appealing. However, the form and content of these clauses have failed to deliver a more robust dispute resolution mechanism in tax treaties.






The provisions for dispute resolution in the OECD lie in article 25 of the OECD Model Convention on Taxation Treaties. Under subclause 1 of the OECD Model, taxpayers can initiate proceedings with the competent authority in their residence state, if they believe that taxation is not or will not be in accordance with the tax treaty. This constitutes the first stage of the MAP. The second stage of the procedures takes place between said competent authority in the residence state and the contracting authority in another state to negotiate, but not necessarily reach an agreement on the issue. Sub clauses 2 and 3 lay out the specifics of the MAP, including taxpayers’ rights and time limits for action (3 years from the notification of action). Article 25(5) which was adopted in 2008 introduced compulsory ad-hoc arbitration as an option for the taxpayer. The clause provides that a taxpayer can request that any issues which remain unresolved by the MAP can be resolved through arbitration if the competent authorities are unable to resolve them within 2 years.[15] There is no opportunity for arbitration on issues where a decision has already been rendered by the court or administrative tribunal of either State.[16] The details and form of the arbitration process are left to the discretion of the competent authorities. Unless the person directly affected by the case does not agree to the mutual agreement that implements the arbitration decision, the arbitration decision is binding.




The UN Model has two main provisions dealing with MAP, Article 25 A and B, only one of which provides for arbitration. This article as a whole mostly draws on the text of the OECD model. For example, Article 25, paragraphs 1 to 3 of the UN Convention are almost identical as the OECD convention. As per Article 25 B sub clause (5), unresolved issues in a MAP case that have not been resolved within 3 years (compared to 2 years under the OECD) can be submitted for arbitration but only by one of the competent authorities.[17] This is the main point of divergence between the two models, in that under the UN convention, the taxpayer is not entitled to request that the issue be pushed through to arbitration but is only entitled to be notified of such a request by the competent authority.


Moreover, the decision of the arbitration is not necessarily binding under the UN convention as the competent authorities have 6 months after the arbitration decision to reach a different resolution.[18] From these provisions it is clear that the UN Model Convention is driven primarily by the competent authority and not the taxpayer, which is rather unusual as the person who brings the dispute is not given a mandate over how the issue is ultimately resolved. Lastly, under the UN Model Convention, arbitration is not truly compulsory but an alternative option to article 25 A which reinforces the MAP and aims to resolve disputes through negotiation. Subsequently there is no right to arbitration and the relevant competent authorities ‘can agree to prevent an unresolved issue from proceeding to arbitration.’[19] The explanatory text of the Convention offers up a reasoning behind this, stating that ‘under voluntary arbitration countries preserve great flexibility as to the issues that will be subjected to arbitration and may restrict the potential number of cases that could proceed to arbitration and reduce the potential costs of arbitration.”[20]




There are clear advantages to having robust arbitration clauses in DTAs that need not be understated. For example, the mere inclusion of an arbitration clause in a tax treaty might affect relationships between contracting parties in a positive way.[21] Moreover, arbitration clauses provide an instrument through which unresolved disputes under MAP can come to a close. Without the inclusion of such a provision, tax treaty disputes often have no consequences for competent authorities and therefore, no incentive to resolve cases. If these authorities are subject to mandatory arbitration, they are more incentivised to find a domestic resolution to the solution for fear of surrendering their tax sovereignty to a third party.[22] Significantly though, arbitration may relieve political pressure for competent authorities who may be expected to arrive at conclusions that are beneficial to their government, regardless of technical merits of the case.[23]


Moreover, the chance of enforcing an arbitral award is higher than obtaining enforcement based on the outcome of MAP negotiations between two tax authorities.[24] For example, under the New York Convention, courts are required to enforce an arbitral award, which they cannot do for negotiation based agreements.[25] Indeed, this was the primary driving force behind the inclusion of a compulsory arbitration clause in the OECD Model in the first instance, as there was a growing concern that MAP was not providing consistent results in a growing number of complex disputes fuelled by cross border trade. [26]




From the outset, one of the primary reasons that the current arbitration mechanisms have fallen short is that these conventions do not recognise arbitration as its own sufficient mechanism for alternative dispute resolution. The model conventions continue to tack arbitration on to the MAP and see it as ‘an extension of the MAP process’ rather than a standalone, independent mechanism that could effectively resolve cross border tax disputes.[27] Arbitration is not available independently of the MAP and tends to be a last resort where the competent authorities of the parties to a DTA have not been able to reach an agreement on the issues within a case that has gone through the MAP already. Consequently, arbitration is a part of MAP and subject to any and all limitations in the procedure itself. For example, the binding nature of the arbitral award is conditional upon the acceptance of the MAP by the taxpayer. Therefore, a taxpayer bringing forward a tax dispute through the MAP in the OECD convention could reject the result of a MAP based arbitration agreement.[28] Similarly, in the UN Model Convention, far too much emphasis is placed on the competent authority instead of letting the taxpayer drive the dispute forward. Therefore, the wishes of the competent authorities can easily override those of a taxpayer and subsequently authorities can also overrule an arbitration decision that they disagree with.


To some extent, this position in the UN Model Convention can be justified since developing countries predominantly use the UN Convention.  Due to the lack of expertise of tax authorities in developing countries, arbitration would be unfair when the dispute occurs with a country that has significantly more expertise.[29] Arbitrators cannot be expected to make up for this lack of expertise and nor can they be burdened with the safeguarding of a state's tax policy.[30] There also concerns as to the qualifications, neutrality and independence of potential arbitrators, especially since neither arbitration clauses sets out a standard for qualifications nor the criteria for finding an arbitrator. Moreover, there are also concerns, from non-OECD countries in particular, about the high costs of arbitration, especially to the tax authority itself, which can be just as high as traditional dispute resolution (i.e., domestic litigation). Furthermore, there are ex ante costs of negotiations and mediations in MAP before the pursuit of ad hoc arbitration which            `make it cost-prohibitive for developing countries to pursue arbitration at the final stage, having already spent money on MAP. Consequently, tax authorities from weaker economies ‘stand to lose their advantages when entering into ad hoc arbitration’ in international tax matters.[31]


Furthermore, there is strong opposition in general towards arbitration clauses, but particularly by developing countries. This is evidenced by the fact that out of nearly 3500 DTAs signed between 2008 and 2014, only 178 of them included some form of an arbitration clause. Of these 178, 20 of these were not arbitration clauses per se but could be considered ‘semi-arbitration’ clauses.[32] Such clauses have little practical relevance but can be useful where the two states have opposingly views on arbitration and can be encouraged to engage in the dispute resolution through the inclusion of a ‘most-favoured-nation’ clause.[33]  Of the remaining 158 that did include a compulsory arbitration clause, 82 were between an OECD Member and a non-OECD member. Only 13 were concluded between two non-OECD member countries.[34]  By way of comparison, there were the 63 DTAs signed between two OECD member states that followed the OECD arbitration clause model but only nine that did not. [35] The figures therefore paint a very grim picture about the acceptance of arbitration among developing nations. Moreover, despite the high number of DTAs with arbitration clauses between OECD member states, the reality is that a little general consensus on arbitration and instead limited number of states have widely accepted and included such clauses in their DTAs. These states include the Netherlands, with 38 such agreements, Canada with 20, Switzerland with 17 and Italy and the UK with 16 each.[36]




The increased mobility of economic activity, especially in the current digital age, has led to MNEs evaluating their tax liabilities more thoroughly when determining potential investment locations. Therefore, improving the effectiveness of tax treaty arbitration is an essential cornerstone for economic growth. The introduction of a mandatory ad hoc arbitration clause in the OECD model and an optional clause in the UN convention was the first step in making alternative dispute resolution mechanisms to MAP available and expanding avenues to justice for taxpayers. The BEPS 1.0 plan included an action plan to making dispute resolution mechanism more effective (Action 14) highlighting the importance of making tax treaties as effective as possible at the point of implementation in the increasingly competitive global economy.[37] However, these measures need to be followed up by a more robust expansion of alternative dispute resolution mechanisms, especially since the OECD has doubled down on its BEPS framework in 2020 with the introduction of its second action plan and expanded to include developing countries in the inclusive framework.[38] The OECD’s BEPS 2.0 plan also includes a multilateral instrument that would amend all existing DTAs in line with the OECD’s recommendations to tackle tax avoidance and profit shifting, without having to renegotiate individual treaties. The corollary of this, however, is that since the multilateral instrument will affect more than 1,600 tax treaties, there will inevitably be a number of tax treaty disputes that arise.[39] In order to effectively and efficiently deal with this fallout, the conventions must make arbitration a viable alternative to MAP instead of a last resort. In doing so parties to DTAs and taxpayers will both benefit from the full range of the advantages of arbitration rather than treating it as ‘a perfunctory instrument used only to force the competent authorities to find a solution for disputes.’[40]

This would mean systematically tackling developing countries’ concerns about the arbitration process, including introducing a minimum standard and set of best practices for arbitrators so that concerns about neutrality and impartiality can be overcome. Another solution would be to replace ad hoc arbitration with institutional arbitration where parties can choose their set of arbitration rules in advance. Since institutional arbitration permits the parties to use rules that are already in place, this will bring an element of certainty and predictability to the proceedings in terms of outcome as well as costs and resources. This will also allow arbitrators, who are already specialists in the rules and procedures of that arbitral institution, to function more effectively while alleviating the burden of finding an arbitrator in the first place.[41] Institutional arbitration could then easily take centre stage, especially for taxpayers, without forcing parties to navigate through the MAP first. Moreover, institutional arbitration would bring a level of uniformity in tax treaty interpretation to arbitration proceedings that does not currently exist under the OECD and UN conventions.[42] Against the backdrop of the current ad hoc arbitration procedure, the logistical and enforceability advantages of institutional arbitration are substantial. More radically even, perhaps it is possible for a specific set of rules to be developed by the OECD/UN to deal with issues arising out of tax arbitration and tax treaty interpretation (much like UNCITRAL).[43] Lastly, making arbitration a standalone dispute resolution option, and thereby more common, will allay states’ fears about engaging in arbitration and ‘disarm those states who might misleadingly resort to their tax sovereignty concerns’ as a way of avoiding a fair dispute resolution process.[44]


V.             CONCLUSION


In conclusion, the increasingly interconnected and globalised economy has resulted in the immense proliferation of tax treaty disputes. As MNEs engage in more thorough evaluations of jurisdictions for their economic activity, tax considerations have become increasingly important in driving such decisions. As a result, there is a desperate need to reconceptualise alternative dispute resolution mechanisms in tax treaties to create an efficient, cost effective and timely method for resolving tax disputes. However, the previous mutual agreement procedure in both the OECD and UN Model Convention on Tax treaties, the two most popular conventions underlying DTAs in the last half century, has been below par. The MAP has several known weaknesses that make it a poor mechanism for resolving tax disputes. This includes the lack of clearly outlined procedures for dispute resolution beyond protecting the right to oral representations between competent authorities and taxpayers. Moreover, MAP has also resulted in the inability of the competent authorities to effectively conclude agreements without any consequences.


The failure of the MAP has driven the inclusion of arbitration clauses in the DTAs but with little success. The form and content of these arbitration clauses have failed to deliver a more robust dispute resolution system in tax treaties. This is driven partly by the treatment of arbitration as a last resort mechanism when MAP has failed, instead of a stand-alone dispute resolution method. The ad hoc nature of the clauses has also buoyed fears about the choice, neutrality and impartiality of arbitrators as well as costs. In particular for developing countries that lack the expertise for MAP, pursuing arbitration is even less likely. While tax sovereignty also plays a role in states unwillingness to engage in arbitration, this is less likely to be the case when arbitration is a full standalone option. With regards to the UN Convention in particular, arbitration is not even a backup but an optional alternative and so, the competent authority is able to dominate the proceedings. Therefore, the wishes of the competent authorities can easily override those of a taxpayer, making arbitration an unattractive option for taxpayers in certain jurisdictions.


This essay advocates for more robust and formal arbitration procedures that would systematically tackle states’ fears about arbitration and make it more attractive as a method for resolving tax treaty disputes. These recommendations include the introduction of a minimum standard and set of best practices for arbitrators so that concerns about neutrality and impartiality can be overcome easily. A vast majority of these concerns can also be dealt with by simply replacing ad hoc arbitration with institutional arbitration thereby reducing the uncertainty and unpredictability of proceedings, while simultaneously bringing down costs. The logistical and enforceability advantages of institutional arbitration are substantial in this way. A more radical solution would be the development of a specific set of rules for tax treaty disputes by either the OECD or UN, however this is unlikely to be enacted any time soon. While we are unlikely to see changes to either model conventions in the near future, there is potential to massively expand the scope of arbitration in tax treaty in a way that is beneficial to everyone.


* University of Cambridge. E-mail:




 Primary Sources


United Nations Model Double Taxation Convention between developed and developing countries (originally adopted in 1988, updated in 2017), accessed 22nd January 2021, 

OECD, Model Tax Convention on Income and on Capital: Condensed Version 2017, OECD Publishing, Paris. Accessed 22nd January 2021 

OECD (2015), Making Dispute Resolution Mechanisms More Effective, Action 14 - 2015 Final Report, OECD/G20 Base Erosion and Profit Shifting Project, OECD Publishing, Paris, 

Convention on the Recognition and Enforcement of Foreign Arbitral Awards [New York Convention] (adopted 7th June 1959)


Secondary Sources


Chetcuti, Jean-Phillippe ‘Tax Dispute Resolution: Arbitration in International Tax Dispute Resolution.’ (2001), accessed 22nd January 2021 

Dagan, Tsilly International Tax Policy: Between Competition and Cooperation (Cambridge University Press 2017) 

Implementation of the Multilateral Convention” Deloitte, Accessed 17th April 2021 

Lang, Michael, and Jeffrey Owens. 2015. International Arbitration in Tax Matters. WU Institute for Austrian and International Tax Law European and International Tax Law and Policy. Amsterdam, The Netherlands: IBFD. 

Pitt, H.M. Arbitration under the OECD Model Convention: Follow-up under Double Tax Conventions: An Evaluation, Intertax 2014 

Salehifar, Alireza   'Rethinking the Role of Arbitration in International Tax Treaties', (2020), 37, Journal of International Arbitration, Issue 1, pp. 87-130, accessed 22 January 2021. 


[1] Tsilly Dagan, International Tax Policy: Between Competition and Cooperation (Cambridge University Press 2017), 148

[2] Jean- Pierre Lieb, “Introduction: Taking the Debate Forward” in Michael Lang & Jeffery Owens (eds) International Arbitration in Tax Matters (IBFD 2015)

[3] Ibid, s1.1.

[4]  Jasmin Kollmann and Laura Turcan, “Overview of the Existing Mechanisms to Resolve Disputes and Their Challenges” in Michael Lang & Jeffery Owens (eds) International Arbitration in Tax Matters (IBFD 2015)

[5] Brian Arnold, ‘The Scope of Arbitration under Tax Treaties’ in Michael Lang & Jeffery Owens (eds) International Arbitration in Tax Matters (IBFD 2015)

[6]  Jean-Phillippe Chetcuti, ‘Tax Dispute Resolution: Arbitration in International Tax Dispute Resolution.’ (2001), accessed 22nd January 2021

[7] Dagan (n 1), 146

[8] Ibid.

[9]Alireza Salehifar, 'Rethinking the Role of Arbitration in International Tax Treaties', (2020), 37, Journal of International Arbitration, Issue 1, p87 accessed 22 January 2021.

 [10] Dagan (n 1), 147

[11] Kollmann and Turcan (n 4)

[12] Chetcuti (n 6)

[13] Ibid.

[14] Kollmann and Turcan (n 4)

[15] OECD, Model Tax Convention on Income and on Capital: Condensed Version 2017, OECD Publishing, Article 25(5) Paris .

[16] Ibid.

[17] United Nations Model Double Taxation Convention between developed and developing countries (originally adopted in 1988, updated in 2017), accessed 22nd January 2021

[18] Arnold (n 5), p113

[19] Ibid, 114.

[20] UN Model Commentary (n 16) art. 25, para 14

[21] Kollmann and Turcan (n 4)

[22] Ibid.

[23] A Salehifar (n 9), p 91

[24] Ibid, 90

[25] New York Convention, Art. III (7th June 1959)

[26] H.M. Pitt, Arbitration under the OECD Model Convention: Follow-up under Double Tax Conventions: An Evaluation, Intertax 2014, p.466

[27] A Salehifar (n 9), p88

[28] Kollmann and Turcan (n 4), 112

[29] UN Model Commentary 2011 (n 16), art 25, para 4

[30] Ibid.

[31] A Salehifar (n 9), p 93

[32] Kollmann and Turcan (n 4), 47

[33] Ibid

[34] Pitt (n 25), p 448

[35] Kollmann and Turcan (n 4), 48

[36] Kollmann and Turcan (n 4), 47

[37] OECD (2015), Making Dispute Resolution Mechanisms More Effective, Action 14 - 2015 Final Report, OECD/G20 Base Erosion and Profit Shifting Project, OECD Publishing, Paris,

[38] See also OECD Statement by the OECD/G20 Inclusive Framework on BEPS on the Two-Pillar Approach to Address the Tax Challenges Arising from the Digitalisation of the Economy,

[39] “Implementation of the Multilateral Convention” Deloitte, Accessed 17th April 2021

[40] A Salehifar (n 9), 94

[41] Ibid, 98

[42] Ibid, 99

[43] Ibid, 100

[44] Ibid, 129

Publicar un comentario

0 Comentarios