How well are mutual agreement procedures working?
Edward Morris of Deloitte UK and Sobhan Kar of Deloitte India question whether the adage that mutual agreement procedures work, but too slowly, is still correct.
Among various tax dispute prevention and/or resolution mechanisms, the mutual agreement procedure (MAP) occupies centre stage in terms of the number of cases, the network of engagements among competent authorities, and the successful resolution of tax disputes. MAPs remain the only legitimate way of eliminating double taxation.
The Action 14 report of the BEPS project sought to modernise MAPs significantly, with a view to increasing their efficiency and effectiveness, introducing a number of recommendations which constitute a minimum standard for MAPs. This comprised 21 elements and 12 best practices.
In addition to the minimum standard, peer reviews of the performance by competent authorities and annual publication of a MAP statistics report were key additional elements of the new regime. In January 2023, the OECD/G20 Inclusive Framework on BEPS announced that, starting in 2024, there would be additional data points that would be captured in the MAP reporting framework and the publication of a separate annual advance pricing agreement (APA) statistics report.
All this transparency is, of course, most welcome. The current MAP reporting framework builds on the work done at the OECD more than a decade ago – statistical reporting of MAPs in some form has been around since 2006 (brought in alongside the OECD work on a Manual on Effective Mutual Agreement Procedures and the arbitration article in the Model Treaty). This meant that stakeholders, including the OECD, the tax administrations and competent authorities concerned, and the multinational enterprises (MNEs) making MAP cases have a plethora of statistical data available. This enables judgements to be made over how well MAP is working and helps to identify areas for improvement.
The reporting framework envisages the following MAP statistics:
Segregating the inventory of cases into two periods – MAP applications made prior to January 1 2016 (pre-2016 cases) and on or after that date (post-2015 cases).
Segregating the inventory of cases into two types:
Attribution (profit attribution to permanent establishments) and allocation (transfer pricing, or TP) cases; and
Giving details regarding the MAP cases resolved, including the category breakdown of resolutions or closures.
Agreeing common ‘start’ and ‘end’ dates with treaty partners so as to ensure reporting of the same average period for MAP resolutions by both countries.
Agreeing and reconciling MAP inventory with treaty partners every year so as to report a common opening balance of cases, the addition of new cases, the resolution of cases, and a closing balance of cases.
This data allows stakeholders to review not only the time taken to conclude MAPs but also to see how they are being concluded. It is commonly said by many observers that MAPs work but could work faster. This article next examines the experience of one country with extensive MAP experience – India.
The Indian experience
The reporting of MAP statistics and peer review of the performance of its competent authorities involved significant change for India. India adapted from a jurisdiction that was initially resistant to a peer review mechanism, primarily because of the initial idea of including taxpayers directly in the peer review process, to one that undertook a Stage 1 peer review within the MAP forum in March 2019.
Similarly, from a jurisdiction that seldom made its MAP inventory public, and which counted its inventory based on the total number of assessment years in dispute, to one that furnished timely and accurate reports under a completely new and different reporting framework, India has fully integrated itself within the OECD reporting requirements.
While India is yet to achieve the gold standard of resolving cases in 24 months on average, its endeavour to do so is evident from the gradual reduction in the time taken to close post-2015 MAP cases over the past few years, even though COVID impacted that trend adversely.
Looking at the 2021 MAP statistics report, India took 39 months on average to close TP MAP cases and 32 months on average to close other MAP cases. Both compare unfavourably with the global average of 32 months and 21 months, respectively.
However, India closed more MAP cases in 2021 than were added to the inventory for post-2015 cases. This is a positive development, as it reduces the size of the MAP inventory. In the Indian context, it is easily discernible that the success in resolving TP MAP cases has been greater than what has been achieved for other MAP cases.
One reason for this TP success was the framework agreement entered into by the Indian and US competent authorities in January 2015. This has resulted in quick and efficient settlement of many TP MAP cases for certain categories of taxpayers. As India has the highest MAP caseload with the US, such quick resolutions under the framework have been an important milestone for India’s successful implementation of MAPs, alongside the development of working relationships with other treaty partners, including the UK. The OECD statistics are available at a country-to-country level so it is possible to identify the effective success of MAPs between two countries.
India also has a good track record if we look at the proportion of cases that are resolved with an outcome that eliminates double taxation completely. This is especially true in respect of TP MAP cases. In 2019, India resolved 86% of TP MAP cases with complete elimination of double taxation.
Although this statistic fell significantly in 2020 and 2021, this was due to domestic developments and not the effectiveness or otherwise of the MAP process. Under the requirements of the BEPS Action 14 mandate, each jurisdiction had to publish its MAP guidance.
Impact of the Income Tax Appellate Tribunal
Until it published its MAP guidance in 2020, India had never documented its MAP practices and the likely approaches to be taken in certain situations. Part of this MAP guidance laid out an approach by the Indian competent authority in circumstances that materially affected how MAP results were to be treated, concerning MAP cases where the Income Tax Appellate Tribunal (ITAT) passes an order deciding the dispute.
The MAP guidance clearly stated that in such situations, the Indian competent authority would not negotiate the MAP case any further and would close the MAP case by categorising the outcome as a resolution through domestic remedy. This was unexpected and reduced the efficacy of the MAP, because the ITAT decisions are not final decisions, and the dispute could be carried by either party to the higher constitutional courts. Thus, dispute resolution and tax certainty through the MAP article were impacted. This development was the cause for the fall in 2020 and 2021 of MAP cases being resolved with full elimination of double taxation – many MAP cases in the Indian inventory were those in which the ITAT had already issued an order.
Hence, during 2020 and 2021, the Indian competent authorities wrote to their counterparts to say that they could not negotiate these cases any further and requested the latter to close such cases as having been resolved under domestic remedy. Accordingly, this category of outcome took the top spot in respect of MAP resolutions and the category of full elimination of double taxation was relegated to the second spot. For example, in 2021, 45% of TP MAP cases and 70% of other MAP cases were categorised by India as resolved under domestic remedy. On the other hand, resolutions with complete elimination of double taxation slid down to 28% and nil, respectively.
This development of closing MAP cases pursuant to ITAT orders also created a picture of India resolving many MAP cases in 2020 and 2021, whereas the reality was that in many cases there were no negotiated resolutions but just closures in view of such orders.
From one point of view, while the issue relevant to a MAP is still within the domestic court process, there is no actual double taxation, only the possibility of it. Ultimately, it will be the ability to access a MAP procedure after litigation, and whether the MAP can override a domestic court decision, that will determine whether MAPs are working effectively to eliminate double taxation.
Nevertheless, the BEPS Action 14 report has had a positive impact on the alternative dispute resolution environment in India, with the Indian competent authorities making efforts to improve the efficiency and effectiveness of the MAP along with their treaty partners. Together with a successful APA programme, MAP initiatives have instilled more confidence among taxpayers on India’s commitment to dispute prevention/resolution and tax certainty.
The OECD originally identified three broad areas where MAPs needed improvements, in relation to:
Despite the increase in the number of cases and longer resolution times, the overall takeaway from the Indian MAP statistics is that MAP processes are improving, with the clear implication that improvements are therefore being made in the three areas identified by the OECD. The ITAT-related development shows that issues around access for cases in litigation remain a relevant area for further consideration.
The global picture
Turning back to the global statistics, what matters to many MNEs involved in MAPs is threefold: not only whether double taxation is eliminated, but when (i.e., the time the MAP takes) and, of course, how (i.e., a principled resolution in accordance with the arm’s-length principle and the facts of the case).
Examining each of these considerations in turn, the statistics show that, for TP MAPs, the majority of cases end with the complete elimination of double taxation. This is because the competent authorities agree completely to tax and relieve the same amount of TP adjustment, or for some cases a resolution is achieved via a domestic route that essentially achieves the same result (i.e., that the MNE is taxed only once on the same profit).
In a number of cases, the elimination of double taxation is only partial, with the implication that some amount is doubly taxed. These are, presumably, cases where the competent authorities can only go so far in reaching agreement and agree to disagree on the balance of the adjustment. While partial agreement is better than no agreement, partial agreement is clearly worse than full agreement. One suggestion could be to submit the remaining area of disagreement to more senior review or a faster, more limited form of arbitration.
But how about increased times to resolution? Some observers might say the time taken is of secondary importance to achieving the result, but time is an understandable factor since uncertainty remains until a final solution is achieved. With more cases going into MAPs (and this, given the OECD’s concern about taxpayers being unable to access MAPs, should be seen as a positive, albeit a situation derived from more TP audits to begin with) but with the same level of competent authority resource available, cases will take longer.
Of course, the statistics present only an average – some cases will take longer to resolve than others based on complexity, for instance, while simpler cases are likely to be faster. A welcome development is the ability of competent authorities to communicate directly via video call, since this can easily be arranged for individual cases, rather than face-to-face meetings, which, by their nature, typically require several cases to be at a similar state of potential resolution before resources can be committed to a physical meeting. However, many competent authorities would point out the importance of building a trusted relationship with their counterparts and there are clearly occasions where this can be done better face to face.
A trusted relationship as part of an efficient working relationship will clearly help to resolve cases quicker. The role of the competent authority is not to rework the case, for instance, but to eliminate the double taxation in line with the arm’s-length principle. A trusted partnership will enable this and engender principled resolutions, in themselves helping to build trust for the benefit of future cases.
As is always the case, statistics are capable of being interpreted in a number of ways and will not reveal absolute truths that apply in every MAP case. But taken as a whole, the OECD statistics show that MAPs are working for many MNEs.
Over time, more cases will inevitably proceed to arbitration; those cases that do so will increase the average time to completion of the overall case numbers, unless, for instance, arbitration happens earlier in the process. It could be a useful development, for example, if competent authorities could consider moving certain cases into the arbitration phase of the MAP before the usual two-year deadline where it was agreed by stakeholders (including the persons making the MAP case) that this was appropriate. It must, on occasion, become obvious that a case needs arbitration to reach a resolution well before the two-year period envisaged comes to an end.
As more cases enter MAPs, it can be expected that times to resolution may increase. Many observers would argue that, however regrettable a development this might be, what matters most is the resolution of cases, not how long that resolution might take.
One thing is overwhelmingly clear: statistics mean transparency and enable the identification of areas for improvement. The time competent authorities take to compile statistics is well worthwhile.