Abstract Traditionally, tax authorities endeavour to resolve their tax treaty disputes among themselves, by amicable settlement through a mutual agreement procedure (commonly known as ‘MAP’ procedure), without involvement from any third parties—neither arbitrators nor mediators. In past years, due to globalization of countries’ economies and spread of tax treaty networks, the number of disputess, their complexity and revenue interest involved have gone up drastically, exceeding many authorities’ capacities, and resulting in MAP cases taking up increasingly more time, or remaining unresolved at all. It is generally expected that the recent OECD/G20 initiated ‘BEPS’ (short for: Base Erosion and Profit Shifting) measures against international tax avoidance will add further to this. Arbitration so far having been hardly tried in practice, the recent arbitration piece under the BEPS multilateral treaty (MLI) and EU Directive on dispute resolution in international tax matters, however, create new momentum. It is now up to tax authorities if they can accustom themselves to the use of arbitration as an ordinary, and in certain circumstances preferable tool for resolving their disputes. 1. USE OF THIRD PARTY ASSISTANCE 1.1 ‘Cold feet’ Binding arbitration in tax treaty disputes has so far failed to gain much popularity with tax authorities. The same holds, even, for any of the softer forms of non-binding third party assistance in dispute resolution like mediation or technical advice. This separates tax from virtually all other areas of law, both public and private, where arbitration and conciliation are almost daily bread and butter. This unpopularity is ill-founded and undeserved of course, and due mostly to tax authorities having cold feet—but it is hard and obstinate as a fact no less. For instance, the OECD made brief reference to mediation as a possibility in the 2008 update to its model tax convention,1 and a somewhat more elaborate reference in its 2007 MAP Manual (MEMAP),2 picking up on country practices emerging at the time at the level of domestic taxation. However, while the OECD advised that further work should be done to make mediation operative in the particular context of MAP, no such work was ever taken up anymore. Neither the Base Erosion and Profit Shifting (BEPS) Action 14 on more effective dispute resolution mechanisms3 nor the recent BEPS multilateral treaty (MLI)4 makes any provision for mediation. Likewise, the recently agreed EU Directive on double taxation dispute resolutiosn5 only mentions mediation as one of several alternative means tax authorities have to resolve their tax treaty disputes before becoming subject to mandatory arbitration—without giving any clue how a mediation process might work in practice. My TRIBUTE organization was the first, and as far as I am aware the only one to date, that offered a model agreement for professional MAP mediation, spelling out the subsequent steps in the process, duties of mediators, and reasonable expectations for disputing parties.6 It is said that the OECD would reason that any support for mediation would politically undermine its stand in favour of arbitration taken in Article 25(5) of its model tax convention. This suggests a perception of mediation and arbitration as being mutually exclusive—as if the choice should necessarily be between the two, as ‘either … or’. The same seems to hold for the United Nations, but then in reverse, where presently its Tax Committee pairs an interest in mediation with a rejection of arbitration.7 It remains to be seen if this sudden interest in mediation is truly genuine, and not merely a pretext to make the outer world believe that countries that reject arbitration might nevertheless be dedicated to resolving disputes—instead of leaving them unresolved to escape from any unfavorable treaty applications. The rejection by the Committee of a proposal to include an explicit provision on mediation in the United Nations model tax convention8 in any case does not look promising. Any mediator will confirm that mediation, while it may prevent arbitration from eventually becoming necessary, may, if it fails to produce a resolution by itself, be useful also as preparation for a proper and efficient arbitration process. In mediation, all the relevant facts can be sorted out and agreed until there is no unnecessary dispute left on those anymore, and it can be examined and agreed what essentially is the scope of the dispute and where and why the disputing parties’ positions differ. There may be tax authorities that think it would be a mere duplication of effort, and therefore a waste of time and cost, to go to mediation where arbitration is available. But arbitration is not meant for fleshing out factual circumstances of a case, and it would be, in contrary, a waste of their precious time if arbitrators had to. I believe arbitrators should refuse to even take a case on if it is such unfit. 1.2 Authorities’ accountability While the MAP arbitrations globally are said to have been only a few to date—there are no public records available—there is no mentioning of any MAP mediations ever having taken place at all. I am counting out any cases, as there may have been, where tax authorities consulted with an outsider expert on a mere informal basis, or where a taxpayer called on an outsider expert to facilitate its communication with authorities. It appears to me this is what tax authorities hold in common against both arbitration and mediation: that they simply do not wish to allow any third parties into their MAP procedures, who would then see and might criticize what and how authorities are doing. It should be reminded that hardly no authorities, if any at all, publicly disclose individual details on their MAP procedures—not even to their own parliaments! They may point in this respect at their broad and exclusive discretion given to them under tax treaties by their Governments and parliaments.9 In particular in rejecting arbitration, authorities are used to referring to their ‘sovereignty’ to decide themselves over their disputes. But no such sovereignty is at stake in mediation, as mediation is by its nature non-binding on disputing parties. I am afraid reality, rather, is that tax authorities simply do not wish to be accountable, not even for mediators who are only there to help disputing parties in managing to resolve by their own strength. Any mediator would be quick to point out that the subject matter of a dispute is not even the focus of a mediation, at least not ordinarily. Instead, mediation primarily targets at the working relations between disputing parties—how they communicate with each other, if they communicate at all, if they care to listen to each other and try to understand each other’s position and arguments, and if they are willing to critically assess their own position and arguments. 1.3 ‘Happy family’ But here comes in an inconvenient truth for tax authorities: the fiction that they would be convivial colleagues to each other—‘one big happy family’. In reality, and more often than not in these present days of global spread of the tax treaties network, there may be severe misunderstanding, controversy, or non-communication, due to differences in culture, or inequalities in expertise or political or economic power. And then, why should tax authorities behave like they were colleagues anyway? International taxation is a zero sum game, and authorities are each others’ competitors for tax revenue, or for investments yielding taxable income. The assumption of a large reserve of profits of multinational enterprises presently escaping from taxation, that underlies the BEPS project, may suggest there would still be plenty to tax left for every country. But it is those low or non-taxed profits—outlawed, so they seem—that are the ones tax authorities contest most hotly, and at the end of the day are most likely to suffer double or even multiple taxation. Tax authorities, the OECD included, have a habit of putting up a brave face, and pretend their world would be carefree of any such relational problems. Strikingly, the OECD model does not even use the term ‘dispute’, and the OECD Commentary hardly—as if disputes would not exist! But, again as any mediator will confirm, disputes are daily bread and butter, and relational issues a common cause of disputes or a common impediment for dispute resolution. There is no reason to believe this would be any different for tax authorities. 2. THE CASE FOR ARBITRATION 2.1 Historical neglect Dispute resolution is arguably the least developed part in the OECD and United Nations model tax conventions. This is the more striking if one recalls that the very first model tax convention, the 1928 League of Nations model,10 in line with other international conventions, did provide for a fully equipped dispute resolution mechanism, including such wide variety of instruments as amicable settlement, non-binding or binding technical advice, by a body to be appointed for this purpose by the Council of the League of Nations, binding arbitration, and even adjudication, by the Permanent Court of International Justice, the predecessor of the present International Court of Justice at the Peace Palace in The Hague. Under the subsequent 1943 Mexico and 1946 models, all of this was done away with, and replaced by a mutual agreement procedure pretty much as known today. It has been suggested as a reason for this that countries had lost their faith in the international institutions, due to their failure to prevent the World War.11 Another plaudible reason might be that countries in practice had experienced little need for such elaborate mechanism, given the very limited number of tax treaties concluded by that time, lagging far behind with original expectations back in 1928. The mutual agreement procedure—MAP—was a European invention of the 1930s and first emerged in the 1930 tax treaty between Italy and France,12 as a means for taxpayers to ‘protest’13 with the authorities against any double taxation occuring in their case. The Commentary to the 1946 model, rather misleadingly, referred in this respect to a ‘taxpayer’s right of appeal’, adding that it seemed ‘legitimate that a taxpayer should be able to obtain the protection in tax matters of one or the other State’. Yet, at the same time it was quick to clarify that this was ‘no judicial procedure but only a direct consultations between the authorities involved’. 14 Since, an effective dispute resolution mechanism remained the subject of neglect—I cannot put it more mildly—for many decades, all through the 1963 and 1977 OECD models until the 2008 Update. The sole amendment made was the addition under the 1977 model15 of the possibility of a joint commission consisting of tax authorities’ representatives. This commission was a far cry from the technical advisory body under the 1928 League of Nations model, as that latter body was independent while the commission the OECD envisaged consisted of the authorities themselves—although the Commentary suggested that delegations might best be headed by a judge or superior official, apparently for reason of their supposed distance from the case in dispute. I am unaware of any tax authorities ever having installed such a joint commission. Even so, the Commentary to the 1963 model introduced some interesting suggestions, firstly that the OECD Committee on Fiscal Affairs might be called on to give an opinion on the correct understanding of a treaty provision,16 which might be regarded as some form of mediation. Also, there was mentioning of the possibility of asking the opinion of ‘certain persons’ acting as independent arbitrators, and even of a periodic list of such persons to be drawn up by the Committee on Fiscal Affairs from whom authorities could choose arbitrators.17 But the reference made to ‘advisory opinion’ makes it clear that this type of arbitration was not meant to be binding. Both references were deleted again by the 2008 update, apparently considered as having become obsolete upon the introduction of ‘true’ mandatory binding arbitration in Article 25(5). 2.2 Reluctant change The change back for arbitration only began in 1990, with the European countries signing their Arbitration Convention.18 It is not to say that this Convention had been a quick and easy process—far from it. Discussions on various proposals, including one for a EU Directive, had been dragging for some 20 years, with the countries of the North generally in favour and those of the South against. While the Northern countries were driven mainly by the interests of their many multinational companies, the Southern countries had strong concerns that their lesser administrative abilities would cause them in arbitrations to be mostly at the loosing end. Part of the deal struck, eventually, was a limitation of the scope of the Convention, covering only transfer pricing, which at that time was still considered a relatively harmless factual matter, and excluding any legal issues, such as permanent establishment determination. That the Convention continued to be controversial also after it had been signed, appears from the fact that in 1999 the European countries could reach agreement on an extension of the Convention’s duration only after the Convention had first expired. Also the USA had begun to experiment with arbitration as from the late 1980s, starting with its treaties with Canada (1986), Germany (1990) and The Netherlands (1992), at first only as a principle without any actual implementation. It was only through the 2008 update of the OECD model, which included the present provisions of Article 25(5) on mandatory binding arbitation, that the OECD for the first time paid serious attention to the issue of dispute resolution. It seems almost as if the OECD had been delaying until it could no longer. Even then, the OECD did very little to assist authorities in actually having an arbitration. Article 25(5) did itself not contain any rules on arbitral procedures, as to date it still does not. Reference to procedures was made only in a sample Memorandum of Understanding (MoU) included in the Commentary, for authorities to negotiate between each other should they wish so. Authorities thus were not put under any obligation to actually implement Article 25(5) and make arbitration operative. Moreover, the sample MoU in the Commentary had only a limited scope, suggesting no more than two types of arbitration, ‘independent opinion’ and ‘last best offer’ (or: ‘baseball’) arbitration, as if there were no other modalities available—which there are, in particular in commercial arbitration—and for no more reason than that those two types just happened to be have been occasionally practised, in Europe and between the USA and Canada. In addition, the OECD omitted relevant details on the factual operation of these two types and experiences gained in the cases where they were applied, thereby making it difficult to judge if they might be suited for use in other cases, with different issues at stake and under different factual circumstances. In particular for ‘last best offer’ arbitration this miss was painfully felt, as this is a variant that is by all means peculiar and rarely applied in any other areas of arbitration outside tax. Perhaps the OECD considered that arbitral panels once appointed would fill in any such operational details where needed. As if authorities, having hardly overwon their concerns about surrendering sovereignty over a dispute to arbitrators, would happily trust them, too, with such a delicate issue as the determination of the applicable procedure and its conditions for an effective and fair resolution. Or, as if between authorities—‘happy family’ as they are supposed to be—no controversy could ever arise over matters of due process, and no procedural rules would therefore be needed. The elaborate arbitration board operating guidelines the US proposed and agreed under a number of its tax treaties, including for instance with Germany, Canada, France, and the UK, however indicate otherwise. The United Nations, even more reluctantly, finally followed the OECD example by including mandatory binding arbitration in its 2011 model tax convention,19 but only as an optional clause, along with a sample MoU comparable to that of the OECD. 2.3 Voluntary arbitration Amid the 2008 revisions to the OECD Commentary, there suddenly emerged a casual reference to voluntary arbitration,20 as a possibility at any time and without a taxpayer claim requiring it, and without even the need for any special treaty provision besides the ordinary text of Article 25(3). This reads as a confirmation that voluntary arbitration had been possible all along since the 1963 model. Apparently the OECD had never bothered to mention this before, or perhaps it had not wanted to encourage authorities to follow this path. The guidance offered on the operation of voluntary arbitration remained sparse though, in fact merely referring to that on mandatory arbitration. Whereas TRIBUTE at various occasion, including in its submission for the MLI negotiation, emphasized the significance of voluntary arbitration and the need for more guidance in this respect, disappointingly no such elaboration was included in either BEPS Action 14 or the MLI. 2.4 Amicable settlement In sharp contrast to the rather meager and seemingly oblique facilitation of dispute resolution through arbitration was the extensive elaboration in the 2008 update on amicable settlement, borrowing heavily from the ‘MEMAP’ MAP manual released in 2004. It is not to say that all this attention for amicable settlement was unnecessary, since amicable settlement until then had been suffering from neglect just the same as arbitration. But it clearly revealed where the OECD had its preference, this being amicable settlement and not arbitration. Traditionally, tax authorities, and in particular those in the OECD community, place great trust indeed in their capacities to resolve disputes by amicable settlement. This may be assumed as the main reason why they chose to ignore arbitration for such a long time: they simply did not anticipate they would ever need it. This belief in amicable settement originates from the early days when the OECD model was developed, in the 1950–60s, at a time tax treaties were still very limited in number and only between countries of the same, Western, culture, and norm was that tax authorities were personally befriended with each other, and the resolution of any disputes, in the rare cases they occurred, a gentlemen’s affair. The weaknesses of the MAP procedure, however, have since shown all too often in practice, and have become the subject of almost ritual complaints from taxpayers: that there is no hard obligation for authorities to resolve, and no time limit on authorities’ negotiations, allowing them to drag on for years at length. In modern days, due to economic globalization and along with it the wide spread of the international tax treaties network, the pretence of amicable settlement as a universal panacea is simply outdated. Both the numbers of MAP cases and their average cycle time have gone up drastically over the last 10years or so. According to OECD statistics,21 by the end of 2016, which are the most recent data available, the total number of pending MAP cases in the countries participating in the ‘BEPS Inclusive Framework’22 had risen to over 7000, divided between some 4000 transfer pricing cases and over 3000 other cases, with an average duration per case of just under 23months, and specifically for transfer pricing cases, which are most elaborate, of some 33months. The latest EU statistics,23 dealing with transfer pricing cases only, reported by the end of 2015 a total number of pending MAP cases in the EU of some 1500, with an average duration per case reaching well over 30 months for the majority of countries. These figures, first of all, reflect a strong rise of disputes and inherently increased need for MAP resolution, which on its turn stems for a large part from the wider spread and intensified use of tax treaties. But they also, and no less importantly, tell that taxpayers have enough faith left in the functioning of MAP to request for it, instead of pursuing domestic court litigation. It should be noted in this respect that the fact that a country is involved in a large number of MAP cases also demonstrates a commitment of that country’s authorities to MAP resolution, whereas by contrast a low number of MAP cases might result from a country’s authorities’ lacking such commitment, and not necessarily from just a lack of disputes. The OECD is now making a serious attempt to patch up the MAP procedure, by committing its own member countries as well as non-member countries volunteering for its ‘BEPS Inclusive Framework’ to minimum standards for taxpayer access to MAP and by submitting their tax authorities to a peer review process. However worthwhile this effort undoubtedly is, since there are still many countries where even the merest guidance for taxpayers on where and how to request MAP is failing, it is nonetheless deemed to be insufficient. The OECD cannot possibly deal with the core issue, which is the lack of problem solving capacity of MAP, as this is an issue relating to tax authorities’ capacities, attitude, and mutual interaction in individual cases and as such is outside the OECD’s view and reach. It is said of some authorities that they in their MAP negotiations would regularly trade off cases against another, or even entire classes of cases—‘horse-trading’—in an attempt to increase the effectiveness of MAP. To trade off cases might seem a quick and easy, and therefore convenient way to resolve disputes. But it would arguably be in conflict with the good faith obligation under Article 25(3) of both OECD and United Nations model taks conventions to ‘endeavour’ a resolution, which must be understood to suggest that authorities undertake a serious effort to reach a principled solution that does justice to the particular merits and circumstances of an individual case. Also, it would be doubtful if authorities under Article 25(3) would be allowed to agree any other resolutions than based on a reasonable interpretation of treaty provisions. In the BEPS Action 14 Public Discussion Draft it was explicitely stated that each MAP case should be approached in a ‘fair and principled’ manner, and ‘not by reference to any balance of results in other cases’,24 but unfortunately none of this made it into the 2015 Final Report. Each MAP procedure thus will normally be a one-off. In any case, in order to resort a balanced outcome, such trade off would require a substantial volume of cases under negotiation, with equitable interests at both authorities’ sides, which in practice, especially in relations between countries of different size or at different stages of development, will often not be the case. 2.5 ‘Threat’ At first, the prevailing perception was that treaty provisions for mandatory binding arbitration in MAP cases would help to speed up amicable settlement—in fact, that this would be their sole purpose—as it was thought that authorities would be prepared to agree anything if only to escape from having to submit to arbitration. As some sharply put it: the objective of any arbitration would be that no arbitration should ever take place. One may wonder how amicable a settlement actually is if it can only be achieved under threat of arbitration. This perception suggests an image of arbitration as an embarrassment, or punishment even, for authorities having failed to resolve their dispute through settlement, instead of a useful addition to the toolkit of resolution instruments as arbitration really is. True is that the mere foresight of arbitration may have the sanitizing effect of driving authorities into abandoning a former uncompromizing attitude, or moderating a former extreme position to more a realistic level. Also true is that authorities, if they are attached to preserving their mutual relationship for the longer term, may find settlement preferable over pursuing arbitration. Settlement, however, may be unattractive for instance in cases with huge complexity, where there is a huge amount of revenue at stake, or where a balanced outcome is not a realistic expectation due to substantial mismatches in authorities’ capacity or expertise, countries’ economic or political power, or in cultural attitude. Tax authorities should not feel forced to settle at all cost, purely for the sake of resolving a dispute, by conceding on cases where they consider their own position as well founded and superior to that of foreign authorities. To give up a position authorities believe is strong purely for the sake of reaching a compromise is a heavy price to pay. It may be better then for authorities, at least for their own sense of justice, to maintain such position and defend it in an arbitration. Also, where a MAP procedure seems to promise little result from the outset, for instance in view of the principled nature of a dispute or a large revenue interest at stake, going to arbitration may be more efficient than pursuing an amicable settlement against all odds. There is no rule that obliges authorities to sit through the full cycle of two years before taxpayers may claim a mandatory arbitration. Authorities are free to submit a case to arbitration—a voluntary arbitration, this is—at any earlier time should they agree so, and without a taxpayer even requesting for it. 2.6 Necessity In any case, also this perception of arbitration as means to speed up MAP procedures has proven wrong in practice. Not only do the EU statistics as mentioned show that despite the application of the European Arbitration Convention the average cycle time of MAP transfer pricing cases has gone up only further, to the extent that hardly any cases get resolved anymore within the period of 24 months before arbitration becomes mandatory. Moreover, an increasingly large number of cases even remain unresolved at all, due to authorities foresaking their obligation to submit them to arbitration. According to the European Commission,25 by 2016 the number of cases qualifying for arbitration but not actually sent in for arbitration had risen to a stunning high of some 900. What allowed this clearly unwanted situation to develop, is the absence of a sanction on authorities forsaking their obligation under the Arbitration Convention to organize an arbitration process. Without any such sanction, arbitration as a threat to authorities permitting themselves a lax or uncooperative attitude in a MAP procedure remains a toothless affair. This excessive pending of arbitrations drove the Commission to propose the new EU Directive, providing for a default arbitral procedure to apply where authorities, for whatever reason—whether disagreement, inability, or unwillingness, as may not be excluded—fail to set up an arbitration by their own strength. Apparently, also the EU member countries considered themselves the functioning of the present Arbitration Convention to be such poor, that they agreed with the proposal without almost any discussion or delay. This may have come to some as a surprise, since the change from Convention to Directive implies a transfer of sovereignty from the EU member countries to the Commission, which had been the main reason why countries back in 1995 still preferred the format of a Convention. Following in the footsteps of the new EU Directive, also the BEPS MLI provides for a default arbitration procedure. This is a main improvement, too, from the present arbitration provisions under the OECD and United Nations model tax conventions, both of which like the European Arbitration Convention lack any sanction on authorities’ failures. The recent developments of the MLI and new EU Directive confirm that arbitration finally has gained—or from history perspective: regained—its place as an ordinary feature in tax treaties. This is not because arbitration would by now have become widely accepted. On the contrary, the majority of countries, unfortunately, still is not committed to arbitration, and it may take a long time before they are—if they ever will. That arbitration is only an option under the MLI, and not one of the Action 14 minimum standards, is because of this diversion, that also splitted the G20 group of countries that commissioned the BEPS project from the OECD. The very fact that the OECD succeeded in having arbitration included in the MLI, despite the lack of unanimous political support within the G20, confirms that in all objectivity tax treaties cannot do without arbitration anymore if countries are determined to have all their disputes resolved one way or another, including the severe ones that are of a predominantly principled nature or technically complex, or carry a large financial interest. 2.7 Advance pricing agreements In recent years, in the area of transfer pricing, taxpayers have increasingly been seeking refuge to Advance Pricing Agreements (APAs), in an attempt to escape from a troublesome and time-consuming MAP process. A unilateral APA this is, since an international—bi- or multilateral—APA is nothing other than a MAP. Tax treaties do not provide for any special mechanism for agreeing an international APA other than MAP. In many cases, unilateral APAs have proven a successful means to pursuade tax authorities to state their position, often even on a preliminary basis, thereby preventing MAP procedures from being necessary just for that purpose. An APA is often advocated as an instrument to prevent disputes, and preferable over a MAP that comes in play only once a dispute has arisen and may along with it bring high cost of litigation and strained relations with authorities. This is a false suggestion, however, for various reasons. Firstly, a dispute may arise just as well in the process of obtaining an APA. Engaging in an APA often is only a discretion of authorities and no any legal obligation, and authorities may refuse a proposal from a taxpayer at any time and for whatever reason, even if the proposed terms fully comply with the OECD Transfer Pricing Guidelines. A proper dispute resolution facility is essential for an effective, reliable APA programme, and the APA programmes available in most countries indeed offer one. Where an APA programme fails any dispute resolution facility, as is the case for instance in China, taxpayers are effectively left at the mercy of the authorities. The absence of any independent body to scrutinize and correct them enables authorities to misuse their discretionary power, for instance in order to press taxpayers into accepting conditions that are less favourable to them than the OECD Transfer Pricing Guidelines, or requirements for providing information that has no relevance for a proper pricing determination. Furthermore, a taxpayer does not have to uphold a MAP request until an actual dispute has arisen, but he may request for MAP already as soon as he perceives a risk of a treaty provision being incorrectly applied in his case. An early MAP request is even advisable, since the more discussion authorities have had with a taxpayer, the larger is the chance that they have dug their heels deep in the sand and are reluctant to change position anymore over a MAP procedure. With an unilateral APA, it remains up to the taxpayer to convince foreign tax authorities to adopt the same transfer price or pricing policy, as the case may be. It might indeed be that those foreign authorities make no problem of adopting a corresponding position, either because they agree without any discussion, or, more simply, because the entire issue remains unnoticed by them—the latter being a likely possibility in countries with a less skilled tax administration. But if foreign authorities do raise a discussion, it is for the taxpayer to negotiate an agreement that convenes the interests of all authorities involved. It is arguable that authorities by leaving it to a taxpayer to negotiate a corresponding agreement, instead of communicating directly with each other for the purpose, forsake their treaty good faith obligation under Article 25(3) of the OECD and United Nations model tax conventions to endeavour a solution by mutual agreement. If a taxpayer fails to agree a corresponding arrangement with foreign authorities, he is forced to request a MAP after all—if he cares enough about reaching a convergent treatment at treaty level. The route of requesting an APA is therefore no guarantee for preventing a dispute, nor even a MAP. But an APA can be the subject of MAP arbitration, too, like any disputed topic, and in that way taxpayers would in any case be guaranteed of an internationally convergent resolution. The US 2016 model tax convention26 secures this through an express provision. The MLI does not include any such provision, nor do the European Arbitration Convention and the new EU Directive, apparently because it was considered self-evident that APAs might be fit for arbitration and a provision to this effect unnecessary. 2.8 Sovereignty Countries that refuse to commit to arbitration—and this is sadly still the majority—often claim sovereignty reasons. Occasionally, a country’s constitutional or other compelling law may indeed prevent tax authorities to submit to any jurisdictions other than their own, or to enforce the judgments of such other jurisdictions, and such impediment must of course be respected. The large frequency and ease with which non-committed countries are used to alleging sovereignty objections, however, makes the argument sound suspected. Tax authorities are wrong if they take it as their sovereign right to decide by themselves whether or not they want to cooperate in resolving a particular dispute. It may be that the obligation Article 25(3) of the OECD and United Nations models imposes on authorities is to ‘endeavour to resolve’ a dispute only, and not to actually resolve it, but this is not meant to imply a right for authorities to veto any treaty interpretations or applications they may consider inconvenient from their unilateral point of view. The term ‘endeavour’, instead, holds a good faith obligation for authorities to resolve any dispute or difficulty, and by any means, which in principle also includes arbitration. Also, it is for this purpose that Article 25(3) gives authorities the wide discretion they have, extending even to the elimination of double taxation in cases not expressly provided for. The reason why there is no hard obligation for tax authorities to actually resolve is, as the United Nations MAP Guide notes,27 that under a country’s domestic law there may be other authorities such as the Ministry of Foreign Affairs who should be involved as well in processes of interpreting international treaties and agreements. 2.9 Developing countries A striking example of how persistent this fiction of an ‘implied veto right’28 still is, in particular in the developing world, presented ActionAid, one of the largest international development organizations, in its reaction on the MLI.29 ActionAid called on developing countries not to opt for mandatory binding arbitration, since, as it put it, ‘the threat of arbitration risks discouraging countries from fully using their taxing rights’. In fact, arbitration should decide if, and how much, taxing rights countries actually have. By the same token, ActionAid recommended developing countries even against opting for the provision on corresponding adjustments of Article 9(2) of the OECD model tax convention, which is an undisputed part of the United Nations model convention as well, on grounds that ‘else they have to accept the OECD transfer pricing standards’. This would turn back the clock to the days of the OECD 1963 model, when the absence of Article 9(2) caused confusion among countries, and diverse interpretations, whether or not transfer pricing issues or other issues of economic double taxation might be the subject of a MAP procedure. ActionAid appears to ignore the importance of legal certainty for foreign investors in planning their investments and the significant contributions the provision of Article 9(2) and the availability of mandatory binding arbitration mean in this respect. It is for sure that developing countries when adopting these ActionAid recommendations would run a severe risk of loosing a vast portion of foreign investment. Reality, in contrast, is that developing countries would arguably be the ones to benefit most from arbitration. Their lack of capacity and economic or political power usually make it difficult for them to cope with large foreign companies or foreign authorities. Arbitration may often be the only real chance they have to obtain a neutral and fair judgment for their disputes, which otherwise they might not even dare to engage in. Anita Kapur, a former chairperson of India’s Central Board of Direct Taxes, once suggested in defence of India’s rejection of arbitration that, where authorities had been unable to find a fair solution for their dispute by mutual agreement, she could not possibly imagine how arbitrators might.30 But this is exactly why arbitration is needed, because authorities due to an imbalance between their capacities or powers are unable to have a fair discussion in the first place. India may have itself a sufficiently strong position at the international front to be able to afford this view, but this is certainly not the case for most other countries in the developing world. Developing countries may feel encouraged by statistics on investment arbitrations showing that it is them that win most of those arbitrations. 2.10 ‘Law making’ There is no ground for countries to reject arbitration out of concern over any ‘law making’ by arbitral tribunals, that might compromise the sovereignty in this respect of their national parliaments. The scope of arbitration is by its nature confined to the interpretation or application of existing treaty terms, that already have parliamentary approval. Also, all jurisdiction arbitrators have is what authorities transfer to them, and cannot possibly be larger than the discretion authorities are themselves attributed with under a treaty. Moreover, authorities can limit the arbitral jurisdiction further, by agreeing restrictive terms of reference or specifying legal sources arbitrators are only allowed to use in making their decision. Neither can sovereignty as an argument against arbitration be explained for by the circumstance that arbitration may be forced on authorities against their will, by a taxpayer claim, given that non-committed authorities are equally unwilling to engage in arbitration voluntarily, on their own initiative—at least, I am unknown with any voluntary MAP arbitration ever having taken place. As a matter of fact, there appears no real difference between non-committed and committed authorities when it comes to their attitude toward voluntary arbitration—they are both just as reluctant. This has given reason for some to argue that arbitration can only work if it is mandatory. 2.11 ‘Ad hoc’ Sovereignty as an argument against arbitration should be distinguished from the sovereignty authorities, as the owners of their disputes, may rightfully claim in appointing the arbitrators in an actual arbitration case and determining the applicable mode of arbitration, including the type of arbitration and procedural rules. This type of sovereignty is preserved at its fullest by ‘ad hoc’ arbitration, where authorities mutually agree the appointment of arbitrators and mode of arbitration, either case-by-case or more generically. This is why ‘ad hoc’ arbitration is the standard under the OECD and United Nations model tax conventions, as well as the European Arbitration Convention and the new EU Directive. The only case of a tax treaty to date I am ware of that applies ‘institutionalized’ instead of ‘ad hoc’ arbitration is the 2000 treaty between Germany and Austria. Any arbitrations under that treaty—one, so far31—are referred to the European Court of Justice, applying the Court’s ordinary procedural rules, and with the Court’s judges acting as arbitrators, without authorities having any say in their appointment, thus making the process rather look like adjudication than arbitration. This choice for the the European Court is the more extraordinary, since the Court has no experience in either arbitration nor tax treaty application. It is not by chance that no role has been reserved for the European Court under the new EU Directive. 3. NEW MOMENTUM: THE BEPS MLI AND EU DIRECTIVE 3.1 Default rule The arbitration piece of the MLI32 is cause for a little celebration. Not only has its multilateral approach the potential of accelerating the number of agreed tax treaty clauses on mandatory binding arbitration beyond the traditional, more laboursome bilateral approach, which so far resorted only a meagre number of some 200 clauses, out of the total of over 5000 tax treaties in existence. Moreover, while the MLI still leaves it up at first instance for tax authorities themselves to agree an arbitral procedure and set up an arbitration process, its provision for a default rule where authorities fail to do so, ensures that countries’ commitment to arbitration cannot remain mere lip service. Countries having signed up to the MLI arbitration piece may be expected to make it part also of their future tax treaty policies, thus extending its effect beyond an amendment of only existing treaties. The inclusion of an effective default rule was the main target also of the input the TRIBUTE initiative provided for the MLI negotation.33 One would logically expect the OECD to take the MLI default rule aboard Article 25(5) of its model convention. However, the 2017 update of the model34 includes it only in a new sample MoU, as a mere suggestion, thereby still allowing countries to leave arbitration inoperative if they wish. The new EU Directive should be valued mostly for the pressure it puts on tax authorities to achieve a timely resolution by their own strength, rather than for the convenience of the arbitral procedure it provides as default. For purposes of such earlier solution the Directive expressly recommends a wide range of alternative instruments, including mediation, conciliation, or arbitration through the European Arbitration Convention, which continues to apply, or, where appropriate, arbitration through bilateral tax treaties or the MLI. Indeed, with a backlog of pending disputes that large as the number of 900 the European Commission recorded, countries cannot really afford to leave any instruments unused. The default procedure under the Directive is by all means elaborate and complex, and thereby little attractive for actual use. It even seems as if the default was deliberately made unattractive, as a further encouragement for authorities to use alternative, easier means. This holds in particular for the rights the default provides to taxpayers to claim appointment of arbitrators with local courts, as well as a hearing for presenting their views on a case. Both these rights, unprecedented as they are under any of the tax treaty arbitration rules existing to date, are no less than a revolution. The Directive makes reference in this respect to the fundamental human right of fair trial. This particular reference may not be entirely appropriate, given, firstly, that while it is taxpayer money that is at stake, taxpayers are themselves not one of the disputing parties. Secondly, arbitration for its voluntary nature cannot be compared with a court procedure—although arbitrators may be expected to observe that arbitral proceedings are due and fair. Even so, by the introduction of these taxpayer rights the Directive may have set the tone for a future development of tax treaty arbitration, from the present exclusive domain of tax authorities into one where, eventually, authorities have to share with taxpayers as equal parties. This would be a recognition for those many from business and advisory communities who for a long time already are arguing for more participation by taxpayers in tax treaty dispute resolution processes. Indeed, the 2017 update of the OECD model tax convention, that was released shortly after the EU Council agreed the new EU Directive, took a first step down this route, too, by including in its sample MoU a right for taxpayers to claim the appointment of arbitrators with the appointing authority if authorities would themselves fail to appoint.35 3.2 More treaty clauses At the start of the MLI negotiation, at the end of 2015, the OECD boasted some 20 countries having publicly stated their support for mandatory binding arbitration, covering over 90 per cent of all MAP procedures within the OECD territory.36 By the official signing of the MLI, in June 2017, this number had gone up to 25 countries, with Singapore as a useful new addition—the only country to date in the South-Asian region. The USA dropping out at the very last moment was a significant and painful loss, however. If it had not been for political pressure from the previous US administration, that had publicly declared, in Congressional hearings on BEPS, its ambition for an arbitration clause to be the sole reason for joining the MLI negotiation, it would have been doubtful if arbitration were included in the MLI in the first place. It may be hoped that it is only a temporary inability that is bothering the USA, and that in time the present or a subsequent US administration will sign up after all. The true success of the MLI arbitration piece will in any case only show over time, as it becomes clear how many other countries decide to follow the example of this leading group of countries. In due course, once leading countries have gained more and positive experience in actual arbitration cases, it may be expected that other countries that at present are still undecided or only occasionally face tax treaty disputes and therefore have little interest—as holds, for example, for many newly emerging economies in Asia and Latin America—become convinced and join the MLI arbitration piece, too. This is of course subject to countries being able to free themselves from any confidentiality restrictions that prevent them from sharing their experiences. There are some countries that, despite being becommitted to tax treaty arbitration, nevertheless have not signed the MLI arbitration piece, apparently because of a preference to pursue arbitration through different means or on other terms than those of the MLI. One example is most of the Eastern European countries, that are committed to arbitration under the European Arbitration Convention and under the new EU Directive, but are not, or not yet, prepared to commit themselves in relation with any other than EU countries. Another example is OECD member countries Mexico and Chili, that both agree arbitration clauses on a regular basis under their bilateral tax treaties. The OECD clearly intends to give further proliferation of tax treaty arbitration a firm push by its policy announcement that in future countries must make an explicit reservation in the Commentary to the OECD model tax treaty if they do not accept arbitration.37 This will raise the threshold for countries wanting to reject arbitration without a solid reason to support it. ‘Naming and shaming’—for this is what the OECD is seeking to do here—is a harsh strategy to use, but I agree it is justified considering how vital the instrumentality of dispute resolution is to a proper functioning of tax treaties. That this strategy of the OECD works, shows the 2017 update of the OECD model, where the number of OECD member countries having made a reservation against arbitration38 is significant less than the number of member countries not having signed up for arbitration under the MLI. In my view, committed countries should ideally not conclude tax treaties anymore with countries that refuse to submit to arbitration, whatever their reasons may be. The impact of the arbitration piece of the MLI may in practice be somewhat limited by the reservations countries are allowed to make, and how the reservations various countries have made match against each other. One obvious reservation to make is the exclusion of tax treaties that already carry satisfactory arbitration clauses. The exclusion of old cases, another permitted reservation, may be attractive for countries that in the past made it their habit to maintain feeble positions and by consequence should have more than normal concern over loosing cases, but less so for countries that face substantial backlogs of cases. By its retroactive application the MLI distinguishes itself from the new EU Directive, that will only apply to new cases. One reservation under the MLI I hope countries are wise enough not to make is that in favour of ‘reasoned opinion’ arbitration instead of ‘final offer’ (or: ‘baseball’) arbitration as the default type of arbitration. That reservation, if countered by another country’s objection, may lead to the arbitration piece in its entirety becoming void in relation to such other country—thereby frustrating the very objective why countries signed up to the arbitration piece in the first place. It makes more sense then to negotiate an agreement, best case by case, for a type of arbitration that has the consent of both countries. 3.3 More disputes It is generally recognized, among tax authorities and business and advisory communities alike, that the number of tax treaty disputes are bound to go further up dramatically over the next years. This comes on top of the autonomous growth that OECD and EU MAP statistics reported has been going on for many years in a row already. The expected additional increase of disputes is mainly due to BEPS itself. BEPS, after all, was designed to alert authorities, in particular from less developed countries, for possible international tax avoidance schemes, and provide them with tools to challenge such schemes. It is unquestionable that lesser developed countries should arm themselves better against international tax avoidance schemes and the inherent loss of revenue, and that the BEPS project is a substantial contribution to this end. From this particular angle, I believe it should even be called a success if tax authorities wage more disputes. Furthermore, there is a range of refined or wholly new tax treaty rules introduced under BEPS, in such sensitive, high-risk areas as transfer pricing and profit allocation, determination of permanent establishments, and anti-tax avoidance and anti-treaty abuse. These new rules allow wider scope for source country taxation that no doubt countries will actively seek to explore. Their application is a further potential for frequent and serious disputes. I am afraid the increase of disputes will not be a mere temporary effect, until the new rules have settled in practice. I remind that over the years the OECD has constantly been refining its Guidelines on transfer pricing, in an attempt to create more clarity and unity. But in the process the Guidelines also became more and more complex—too complex, so it seems. As a result, many tax authorities today, in OECD member countries no less than in non-member countries, when imposing assessments frequently misapply the Guidelines or even completely ignore them, whether for lack of understanding or deliberately. Instead, there is a preference noted among authorities for methods of their own invention—allowing for more tax, and satisfying political and public pressures at their domestic front. I call this the paradox of transfer pricing: while the rules were originally meant to reconcile countries’ various domestic practices, what they in fact achieved was only to raise more controversies. As a reaction, the United Nations Transfer Pricing Manual now offers simplified rules that should better match with developing countries’ limited abilities. The inherent risk of any such simplification process is of course that what may be presented as a mere simplification, in fact works as a substantive departure from the original concept. This dispute raising effect of BEPS pairs with the introduction of CbC (country by country) reporting, UBO (ultimate beneficial owner) registration, and the new OECD CRS (common reporting standard) format for international automatic exchange of information. The latter developments will generate many new flows of bulk information, that, fragmentary or unverified as they often will be, may easily lead receiving tax authorities into misunderstandings or misuses, and on account of that be cause for additional disputes. At the same time, increased transparency demands, exchanges of rulings, threat of EU State Aid claims, make it less viable for taxpayers, and for authorities as well, to engage in any settlements, either at pre- or post-dispute level. Court litigation may not be considered reliable or effective, for a lack of capacity, expertise, or independence on the part of the courts—there are court cases known in some jurisdictions to have lasted over 30 years!—or a court procedure may not be in place at all. In any case can a domestic court decision not be expected to establish an internationally convergent resolution as an effective resolution of a tax treaty issue requires. 3.4 More pending MAP cases Since BEPS does not provide for any alternative coordinating mechanisms, there is only the traditional MAP instrument available to establish a convergent treaty application by all countries involved. As said, the numbers of MAP cases taking up excessive time to resolve, beyond the 2-year time limit set under the OECD model and EU Arbitration Convention as a marker for mandatory arbitration, have risen sharply over the past years. In particular in the complex area of transfer pricing, most European countries nowadays do not even succeed anymore in completing any MAP cases within a 2-years period. From both the EU39 and US40 APA statistics, a similar trend can be noted for international—bilateral or multilateral—APAs. The most recent EU data, over 2015, reported an average period of well over 35 months for international APAs between EU members to complete—without further distinction between new APAs or renewals of existing APAs. The latest US figures, over 2016, report for new international APAs an even higher average period of some 50 months, and for renewals of just under 35 months. Next to various causes relating to the MAP process itself, as I explained earlier, lack of capacity on the part of tax authorities should be regarded as the main external cause for this. Capacity as an issue is usually related with developing countries, but when it comes to dispute resolution developed countries are not spared from it either. It is said that capacity would even be the main target of the OECD MAP Forum, where the participants are almost only experienced authorities from developed countries. The MAP Forum not having published anything so far on its activities, this is unconfirmed of course, but if it were indeed true, then this would not come as a surprise. In many countries, the capacity of their MAP competent authority function has not kept in tune with the more agressive stand adopted by their local tax auditors and the increase of controversies resulting from that. On the contrary, countries seem to have even been economizing on their competent authority capacity over the past years of financial crisis, to meet government budget cuts. The capacity issue is also the reason why I do not believe that extending the term for amicable settlement from 2 to 3 years, as both the MLI and new EU Directive offer as an option, will help much to prevent cases from ending up in arbitration. It does not make much difference how much time authorities have for completing a MAP case if they lack the resources to deal with it in the first place. MAP guidance and training as the United Nations now is planning to offer to developing countries41 is indeed necessary, but its effect will inevitably remain limited and noticeable only on longer term. Just as only few developed countries will be able to afford reinforceing their MAP capacity, if they can manage to find experts who are sufficiently qualified for this specialist work to begin with – there is a good argument that MAP skills can only be properly acquired by training on the job. To date, Germany is the only country I am aware of that announced it would reinforce its competent authority function in response to the BEPS Action 14 minimum standards. But then, Germany is number two on the OECD list of countries having most pending MAP cases, and number one by far on the European list, so it was about time as well. Even so, all this patching up, whatever may come from it in practice, is bound to remain insufficient. The capacity issue is simply too large, widespread and permanent for that, and the more so when taking into account the strong growth of disputes that is to be expected. Hence there is due to be a substantial rise of MAP cases, in near future already, that require third party facilitation, either technical advice, mediation, or binding or non-binding arbitration, in order to achieve an effective, fair, and expert resolution. This is not a matter of policy, but a sheer necessity. 4. HOW TO MAKE MAP ARBITRATION WORK 4.1 Procedural agreements With the substantial increase of disputes as there is to be expected over the coming years, due to BEPS and other measures, there will be serious need for MAP arbitration—that is, if authorities are true to their commitment to seek a resolution for their disputes, and, moreover, a resolution that is fair and expert. While the MLI arbitration piece and the new EU Directive are designed to respond to such need, there still remains additional work for tax authorities to do to be able to use the potential the MLI and Directive offer in an effective and prudent manner, and in the appropriate cases. To make arbitration operative in practice, at the level of in individual cases, there are a number of procedural issues that tax authorities still need to agree. Authorities could of course rely on the default rule of either MLI or EU Directive, but that may not be their first preference. The Explanatory Statement to the MLI suggests that committed authorities jointly draft a sample memorandum of understanding for this purpose. The alternative are bilateral administrative agreements under individual tax treaties, as some authorities have entered into it already, among which those of the USA, Canada, the UK, and Germany.42 The OECD has included a revised version of its sample MoU in the recently released 2017 update of its model convention. However, their choice for ‘baseball arbitration’ as the main type of arbitration, as well as that for OECD officials to act as appointing authority, failing authorities’ agreement on the appointment of arbitrators, may not be to every authorities’ liking, both outside and inside the OECD area. 4.2 Type of arbitration One of the most important issues for authorities to deal with is indeed the type of arbitration procedure. The European Arbitration Convention and the new EU Directive alike apply conventional ‘reasoned opinion’ arbitration. By contrast, the MLI, following the example of tax treaties the US has concluded with Canada as well as some other countries, aims for ‘baseball’ (or: ‘last best offer’ or ‘final offer’) arbitration, that allows arbitrators in deciding a case only to choose for either of the positions defended by the disputing authorities, and not to state their own views. ‘Baseball’ arbitration is a short procedure. It is entirely in writing, with only limited volumes of position papers and other supporting documents allowed. There are therefore no hearings, and no witness testimonies or examinations—to avoid any time-consuming lawyering from happening. The ‘baseball’ approach should exclude any easy splitting down the middle solutions, which authorities might just as well have agreed among themselves. It is often argued as another feature of the ‘baseball’ approach that it forces authorities to mitigate their positions such that they are reasonable enough to win arbitrators’ sympathy. The latter, however, is not exclusive for the ‘baseball’ variant, since also under ‘reasoned opinion’ arbitration it is unlikely that excessive positions would ever succeed to make a positive impression on any arbitrators. The short form of ‘baseball’ arbitration is generally regarded for its greater efficiency. The opponents of the ‘baseball’ approach usually argue that it would be an inferior procedure precisely because of its short form, which, so they suggest, would not allow a careful consideration of a dispute. Whether this holds much truth, is difficult to judge, since to date there has been only so little practical experience with ‘baseball’ arbitration—just the handful of cases between the USA and Canada—and none of which accessible to any others besides the disputing authorities, neither governments nor academics. The debate over the type of arbitration is less of a ‘clash of schools’ than may seem at the surface. The main driver behind the ‘baseball’ approach is a wish for a quick and expert resolution of a case. The predominant consideration supporting the ‘reasoned opinion’ approach is one of due process. Yet, the efficiency argument should matter also to EU member countries such as Germany, France, Italy, or Belgium, that according to the EU and OECD statistics are faced with large numbers of pending MAP cases which are likely to bound for arbitration in near future. It is difficult to imagine how the tens or more of arbitration cases as there inevitably are going to be each year might be processed within reasonable time other than on the basis of some short form type of arbitration. The new EU Directive recommends ‘baseball’ arbitration, next to various other instruments, as a means for authorities to achieve a timely resolution for a dispute and prevent the Directive’s default rule from having to apply—and for good reason. Time effectiveness of arbitration is a concern tax authorities and business commonly share. Litigation in hearings usually being the most time-consuming part of arbitration processes, its deletion under the ‘baseball’ approach therefore makes some sense. The default procedure under the new EU Directive applies itself strict timelines, too, contrary to the prior European Arbitration Convention that held none and as result gave cause for complaints over too lengthy procedures. Effectiveness is also the main consideration underlying the joint fast track arbitration model of TRIBUTE and the Permanent Court of Arbitration. At the same time, the US cannot be said not to bother or only a little about the condition of due process in ‘baseball’ arbitrations, in view of the elaborate operational guidance for arbitral boards it has agreed at administrative level with a number of its treaty partners. It would be a mistake to think that ‘baseball’ arbitration by reason of its short form could do without procedural rules. By contrast, this type of arbitration seems particular sensitive for its tight conditions on volumes of submissions and on timelines, which require strict observation. It has often been suggested that ‘baseball’ arbitration would be an informal procedure and that this would be one of its merits, but this informality is only relative, by comparison to a court procedure. Arbitral decisions are vulnerable to claims for annulment by domestic courts if conditions of due process have not been properly met. This appears to hold for MAP arbitrations as well, since the possibility of annulment by domestic courts is not expressly excluded under neither the MLI, the European Arbitration Convention or the new EU Directive, nor, as I am aware, under any of the present bilateral tax treaties. Whether this was deliberately meant so, or the possibility of annulment was never anticipated to occur, remains guessing. Due process concerns might also be addressed by having an experienced external arbitration institute administer the arbitral process, that could alert arbitrators on any procedural issues and advice how to best deal with them. Another alternative might be to have as presiding arbitrator a generalist arbitrator, or a judge as the new EU Directive suggests, who can be trusted to have a good understanding of procedural issues and how they may be prevented or resolved. However, that would go at the expense of an arbitral panel’s expertise in the substance matter of a dispute. Perhaps these two opposing approaches of ‘baseball’ and ‘reasoned opinion’ arbitration might be conciliated with each other to some extent, for instance by allowing a short hearing, via video or telephone conferencing, to be part of the ‘baseball’ procedure. Other possibilities might be under the ‘baseball’ approach to allow arbitrators adding short reasons to their decisions, or, at the other end, under the ‘reasoned opinion’ approach to impose a fee cap on arbitrators, preventing them from submitting too lengthy considerations. One might imagine that authorities having lost an arbitration like to know, at least in summary, why they lost it. 4.3 Taxpayer participation Related to the issue of the applicable type of arbitration is the issue of taxpayer participation in the arbitral process, in particular through a hearing. The new EU Directive provides a right for taxpayers to claim that a hearing be held, with reference to the legal principle of ‘fair trial’. The MLI, by contrast, leaves taxpayer participation entirely unruled—it is not ruled in, but not ruled out either. The OECD sample MoU provides for a role for taxpayers in arbitral proceedings as a witness, giving testimony in writing or orally in a hearing, but only under the ‘reasoned opinion’ approach. ‘Baseball’ arbitration at present does not provide for any taxpayer participation at all. Taxpayer participation apparently is regarded as only to delay the arbitral process, just where the speediness of the process is a main attraction of the ‘baseball’ approach. If taxpayers have been allowed ample opportunity to present their views on a case to either authorities at the preceeding stage of MAP negotiations, as the OECD Commentary suggests they should be, then there may indeed remain little for them to usefully add anymore in arbitration. But it seems much the same would hold for tax authorities as well, to this effect that they may achieve substantial time gain in arbitration through a good preparation at prior MAP and pre-MAP stages, by clarifying and agreeing facts, and identifying and narrowing the scope of their dispute as much as they possibly can. Some have suggested that taxpayers should be given the opportunity to submit their ‘final offer’, too, to the arbitrators. This might make sense in particular if the final offers of either of the authorities substantially differ from their positions in the MAP negotiations, thereby creating a new momentum also for taxpayers. Even if taxpayers would themselves feel they have nothing to add anymore, their participation in arbitral proceedings, whether in a hearing or in writing, may still be useful from the viewpoint of arbitrators. Taxpayers in a capacity as witness can be a useful and reliable source for arbitrators to direct any questions for further information to as they still may have, since taxpayers, unlike tax authorities, are familiar with relevant circumstances in all jurisdictions involved, and are unbiased to the extent that all they want is to have a resolution, but not necessarily a resolution in favor of one or the other authorities’ position. Such witnesses as taxpayers, who are there not to support either authorities but only to inform arbitrators, are customarily named ‘amicus curiae’—literally, an arbitrators’ friend. I believe this function is the reason indeed why authorities should accept taxpayers participating in an arbitration. 4.4 Appointment of arbitrators There are other issues besides the type of arbitration where differences in views may prove more difficult to bridge in practice. The first of those issues deals with requirements for appointment as arbitrator. The MLI stipulates as sole condition that arbitrators must have expertise or experience in the field of international taxation, and so does the OECD sample MoU. This stands to reason, considering that it is the assumption that arbitrators are better skilled in the substance matters of disputes that makes taxpayers request for arbitration, instead of pursuing a court procedure. By contrast, the new EU Directive applies as leading principle, subject to countries bilaterally agreeing differently, that the presiding arbitrator be a judge—probably, as I suggested earlier, for reason of securing proper observance of the ‘fair trial’ principle. I note that in many EU countries judges, at least those who are full time, are not allowed to engage in any activities on the side, including arbitration, so this requirement if adopted may be problematic to operate in practice. Next to qualifications, also the neutrality of arbitrators appears a likely subject of controversy. The MLI and OECD sample MoU demand that arbitrators be independent from the disputing authorities. By contrast, the new EU Directive, copying the prior European Arbitration Convention on this, reserves seats in arbitral panels for officials of disputing tax authorities. The latter should serve as ensurance to tax authorities that their views are properly heard by arbitral panels, but at the same it is at variance with the general view that arbitrators must be neutral and as such continues to attract severe criticisms. The new EU Directive excludes as arbitrators all tax professionals whose firms have dealings with taxpayers in a disputed case, or have had in recent past—so not, like the MLI does, just those professionals who were themselves involved with the disputed case. This may eventually rule out for instance those working for either of the ‘Big Four’. Furthermore, while for instance the USA, under the operational guidances it has agreed with its treaty partners, is used to excluding as arbitrators any fellow nationals of either disputing authorities, the new EU Directive, on the contrary, stipulates that arbitrators must have one of the EU nationalities. Exclusions and qualification requirements like these may make it difficult to find arbitrators with personal details matching a particular arbitration case, and particularly so when numbers of cases go up. Lists of available arbitrators, as are common in other aeras of arbitration, specifying for nationality, professional background, areas of expertise and other relevant details, may assist in this search. TRIBUTE avails such a list, which, moreover, is public and transparant. It would be most welcome if the OECD or United Nations or arbitration institutes were to develop similar lists as well. The lists employed for appointing presiding arbitrators under the European Arbitration Convention or some of the US bilateral tax treaties are something of a start, but as these are ‘closed’ lists, meaning that they exclude the appointment of any other persons than those on the list, and contain only small numbers of persons, the choice they offer may be too limited in practice. A final matter concerns who is going to ‘police’ if arbitrators truly meet applicable requirements. Some verification is necessary in any case, provided tax authorities do not wish to rely entirely on what arbitrators declare themselves. A cumbersome and sensitive duty as this may be, authorities may prefer to leave it to an external and experienced arbitration institute, rather than trying it themselves. 4.5 Publication Another likely controversial issue in practice is that of publication of arbitral decisions. The MLI and OECD sample MoU stipulate that arbitral decisions should have no precedential value, implying no publication. This is in conformity with current practice, in all areas of arbitration, that the majority of cases never gets published. Likewise, none of the arbitrations under the European Arbitration Convention and the bilateral tax treaty between the United States and Canada was officially published—their existence is only known from hearsay. The new EU Directive, however, attempts to push for publication by adopting it as its main principle, subject to authorities agreeing otherwise. Publication indeed contributes to greater legal certainty for taxpayers, and transparency of the instrument of arbitration and its uses for the wider public. Recently, investment arbitration in particular has come under severe public pressure for its lack of transparency, suggesting that arbitral decisions would too often favour multinational businesses at the expense of public interests. Such public hostility could easily become directed against MAP arbitration as well, and if this would happen, it might seriously damage the chances for MAP arbitration to further develop. But the risk of precedential value that publication holds at the same time impacts the freedom for tax authorities to agree different solutions in similar cases if they feel that suits them better for reaching an amicable settlement. It is difficult to foresee how this dilemma might be resolved other than by authorities eventually yielding to publication. In any case, the legal protection that many countries under their domestic laws provide for the confidentiality of fiscal data of individual taxpayers will as a rule necessitate that any publication can only be in anonymized form. As far as taxpayer interests are at stake, those would seem sufficiently addressed by a mere anonymization of decisions, as is customary practice in many countries for court decisions. 4.6 Institutionalization The new EU Directive also provides the most interesting suggestion for tax authorities to bilaterally agree arbitration by a permanent arbitral tribunal, instead of arbitration by ‘ad hoc’ panels formed anew for each single case, which are currently norm. Having a permanent tribunal available would certainly be a relief for those authorities, which I believe is the majority, that for reason of their lack of knowledge or capacity are uncomfortable with having to set up themselves a proper arbitration process. Moreover, in light of the sharp increase of numbers of MAP arbitration cases as there is reason to expect particularly in the EU, it would be difficult to continue dealing with arbitration on an ‘ad hoc’ basis anyway. A permanent tribunal would take arbitration to a higher, more professional level, and in that way might lead to greater trust and wider spread acceptation of arbitration, most importantly among tax authorities themselves, and next to that also among the wider public. This suggestion in the new EU Directive coincides with the effort the European Commission has been making for some years already, and for similar reasons, to replace the commonly practised ‘ad hoc’ tribunals under investment arbitrations by a new permanent tribunal, the ‘multilateral investment court’. Tax authorities traditionally prefer ‘ad hoc’ arbitration, out of concern that institutionalization by a permanent tribunal might deny them any discretion to choose their own arbitrators and applicable procedural rules. This is indeed what institutionalization at its purest would hold. However, this does not necessarily have to be, since there is a milder form of institutionalization available that provides only for a transfer of case administration, which to most tax authorities would only be a welcome relief, while leaving the appointment of arbitrators and applicable procedures for authorities as under the ‘ad hoc’ approach. It would appear that the Permanent Court of Arbitration at the Peace Palace in The Hague is the best placed venue for such intermediate solution. That Court has a unique and long-standing expertise in facilitating arbitration especially in interstate disputes, and as an intergovernmental organization is perfectly neutral to authorities interests. Governments, funding the Court by their paid membership, may even expect their tax authorities to take their disputes to the Court, although I am not aware of any actual obligation in this respect. The cost involved with such external case administration would in any case not be much of an impediment for authorities generally, modest as it usually is, no more than a small percentage of the sum of arbitrators’ fees and expenses. Under the joint arbitration model of TRIBUTE and the Permanent Court of Arbitration, too, it is the Court that provides the case administration. 4.7 Case sensitivity It is clear there still is substantial work for tax authorities to do before MAP arbitration can really become operative. That said, all of the pending choices are customary to any arbitration, whether commercial, investment or trade, and thus also MAP arbitration. There is, therefore, nothing special or extraordinary about tax authorities having to make these choices. Neither do any of these choices affect the prior political commitment to MAP arbitration countries have expressed. However, due to their inexperience with arbitration, it may take tax authorities considerable time and internal, and hopefully also public, consultations before they are able to decide their choices. The decision process may be less burdensome if choices are made not on a generic basis, for all future arbitrations alike, but are reserved for an actual case to arise, where it is much clearer which impact various choices have, and the possibility is left for authorities to choose differently in subsequent cases. 4.8 Dispute management In the end, what matters most is that tax authorities accustom themselves to the thought of arbitration not as a threat or penalty (the ‘nuclear option’), but as a convenient solution they are free to choose, or not—as an option for them to weigh against other options they have available in deciding how to best deal with a particular dispute. Arbitration, after all, is designed to be used—not to be used. Whether a given dispute may best be prevented, simply by resting the case, or resolved through amicable settlement, or litigated before a local court, or submitted to an independent third party, either for non-binding mediation or binding arbitration, is something that authorities of each individual country must decide by themselves. Authorities may well decide differently, depending on such factors as their own expertise, capacity, available financial means, or on such factual circumstances of a case as its fundamental importance or budgetary interest involved. Tax authorities would do good to make their decisions on the basis of a clear and comprehensive, predetermined strategy for dispute management, addressing which issues they find worthy of disputing on, and in which manner they prefer such dispute to be resolved. Footnotes 1 OECD, Model Income Tax Convention on Income and Capital, Commentary on art 25, paras 86 and 87. 2 OECD, Manual on Effective Mutual Agreement Procedures, para 3.5.2 3 OECD, Making Dispute Resolution Mechanisms More Effective, Action 14 – 2015 Final Report. 4 Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting, Paris, 7 June 2017. 5 Council Directive (EU) 2017/1852 of 10 October 2017 on tax dispute resolution mechanisms in the European Union. 6 As part of the TRIBUTE submission of 30 June 2016 in response to the OECD public discussion draft of 31 May 2016 for the negotiation of the BEPS MLI, and published by the OECD on 30 June 2016 as part of the set of public comments received. 7 United Nations Committee of Experts on International Cooperation in Tax Matters, Fourteenth Session, New York, 3–6 April 2017, Coordinator’s Report on Work of the Subcommittee on the Mutual Agreement Procedure—Dispute Avoidance and Resolution (E/C.18/2017/CRP.4). 8 Report on the Fourteenth Session (E/C.18/2017/3), p 19, paras 90–91. 9 art 25(3) of the OECD model tax convention allows authorities to even consult for the elimination of double taxation in cases not provided for in the convention. 10 art 14. 11 Zvi Altman, Dispute Resolution under Tax Treaties, IBFD (IBFD 2005) 57. 12 ibid p 14, reports an even earlier 1922 multilateral treaty between Italy, Hungary, Poland, Romania, and Yugoslavia, but that treaty appears never to have taken force. 13 League of Nations Fiscal Committee, 1935 Report. 14 Commentary, Ad art XVII. 15 art 25(4). 16 Commentary on art 25, para 44. 17 Para 45. 18 Convention of 23 July 1990 (90/463/EEC) on the elimination of double taxation in connection with the adjustment of profits of associated enterprises, as amended by Protocol of 25 May 1999 (1999/C 202/01). 19 art 25, alternative B. 20 Para 69 of the Commentary on art 25. 21 OECD, Mutual Agreement Procedure Statistics for 2016, released on 27 November 2017. 22 These include all OECD member countries, as well as a large number of non-member countries, among which, for example China and India. 23 EU Joint Transfer Pricing Forum, Statistics on Pending Mutual Agreement Procedures (MAPs) under the Arbitration Convention at the end of 2015, released on 20 October 2016. 24 Public Discussion Draft of 18 December 2014, para 38. 25 Opinion of the European Economic and Social Committee (COM(2016)686 final – 2016/0338(CNS)), para 3.3. 26 art 25(3)(d). 27 Paras 53–54. 28 I copied this term from Michael Lennard, ‘Transfer Pricing Arbitration as an Option for Developing Countries’ (2014) 42(3) INTERTAX 179–88. 29 Briefing of 6 June 2017. 30 According to a report in The Economic Times of 13 October 2015, of an Assocham India conference in New Delhi that same day. 31 ECJ, decision of 12 September 2017, case C-648/15. 32 Part VI, arts 18–26. 33 TRIBUTE submission of 30 June 2016. 34 Released on 21 November 2017. 35 Para 3(3) of the new MoU refers to an appointment of one or more arbitrators by the OECD as appointing authority ‘after receiving a request to that effect from the person who made the request for arbitration’, which is the taxpayer or taxpayers involved in the case. Apparently this reference was such last-minute addition that there was no time left for any explanation to it in the Commentary on the new MoU. 36 BEPS Action 14-2015 Final Report, p 10. 37 BEPS Action 14-2015 Final Report, paras 22–23. 38 Under the new paras 97 and 102 of the Commentary on art 25, of the OECD member countries only Chile, Denmark, Hungary, Israel, Korea, Mexico, and Turkey have reserved a right not to include the specific provisions of art 25(5) in their tax treaties. 39 EU Joint Transfer Pricing Forum, Statistics on APAs in the EU at the End of 2015, released on 20 October 2016. 40 IRS, Announcement and Report Concerning Advance Pricing Agreements (‘2016 APMA Statutory Report’), 27 March 2017. 41 Committee of Experts on International Cooperation in Tax Matters, Report on the fourteenth session (3–6 April 2017) pp 17–19. 42 Murray Clayson, Elisabeth Snodgrass and Calum Young, ‘The Changing Face of International Tax Arbitration’ [March 2017] Tax Planning International Review 2–13. © The Author(s) 2018. Published by Oxford University Press on behalf of the London Court of International Arbitration. All rights reserved. For permissions, please email: email@example.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)
Arbitration International – Oxford University Press
Published: Mar 14, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
All the latest content is available, no embargo periods.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud