The OECD made its end-November 2016 deadline to release the text of the multilateral treaty to give effect to the BEPS Actions which involve changes to tax treaties, see here. The 49 page treaty text, which is commonly referred to as the multilateral instrument or MLI, and 85 page explanatory statement (ES) contain substantive changes to existing tax treaties in a relatively small subset of provisions.
The objective of the MLI is to have a single instrument which a country can sign to update its suite of treaties by a single stroke – that is, without having to re-negotiate each treaty individually. So, if Australia (or any other country) signs the MLI, potentially all its existing treaties could be amended in one place. Hence this one instrument has to be flexible enough to effect amendments to over 3,000 treaties based on different treaty models, some containing the amending provisions already (eg, an arbitration clause), of varying scope and age, some with protocols, in a variety of languages, and between countries that have different views on just how much and which parts of the BEPS agenda they want to implement. That drafting challenge explains a lot about why the instrument is so obtuse.
The changes which the MLI would make are hedged around with elections, options and the possibility of reservations, which is why the text of the MLI manages to be double the length of a typical tax treaty, while perhaps leaving readers wondering exactly what it all means. (Before grappling with the details of these complexities, it is helpful to read the summary in ES pages 3-7.)
Treasury released a Consultation Paper (CP) on 19 December 2016 setting out how it is proposed that Australia react to the menu of choices on offer in the MLI with submissions due by 13 January 2017. It will be interesting to see how many submissions were received given this timing. This Tax Brief discusses how and when the MLI will operate, the current likely Australian position on it and the potential impact of the new US President.
Australia has a variety of treaty types affecting income tax (often along with other taxes) – comprehensive income tax treaties, airline agreements, tax provisions of the Timor Sea treaty, limited agreements with tax havens etc. Australia and other countries will simply nominate particular existing treaties “for the avoidance of double taxation with respect to taxes on income (whether or not other taxes are also covered)” that the MLI will cover. According to the ES our airline agreements with China and Greece are not intended to be included – can the other less traditional forms be nominated?
The nominated treaties will only be effectively covered, however, if the other parties to the treaties also nominate them, which we will only finally know when they deposit their instrument of ratification (or later add them to their nominated list).
Countries may leave out treaties from coverage and it is possible that this will happen with recent bilateral treaties which adopt many of the changes proposed by the BEPS project. In Australia’s case the CP indicates the Australia-Germany 2015 treaty which recently entered into force as possibly in this category. And future treaties, eg, Israel, will probably not be nominated, even though they can be added to the MLI by notification, as they will presumably incorporate the BEPS changes on which both countries are agreed.
Provisions potentially affected by the MLI
The parts of Australia’s tax treaties which will be affected by the MLI involve:
- Hybrid mismatches: fiscally transparent entities, dual resident entities and relief of double taxation.
- Treaty abuse: preamble, principal purpose test (PPT), “simplified” limitation of benefits (LOB) test, dividend transfers, land rich companies, third state permanent establishments (PEs), saving clause.
- PE definition: agency test, minor activities exclusion, splitting time thresholds.
- Improving dispute resolution: mutual agreement procedure, corresponding adjustments.
- Arbitration (with much greater detail compared to the OECD Model arbitration provision).
These matters were the subject of Actions 2, 6, 7 and 14 of the BEPS Agenda.
The new substantive treaty texts that substitute for, or modify, existing provisions are not written in usual tax treaty style and vary from the proposed provisions put forward in the BEPS 2015 Final Reports for a number of reasons such as:
- Different terms such as Contracting Jurisdiction rather than Contracting State (for international law and drafting reasons).
- Inability to cross reference specific provisions of existing treaties leading to long circumlocutions to describe the content being replaced or modified (as the MLI sits alongside all the amended treaties with their different numberings and textual variations in affected provisions).
- In some cases the 2015 provisions required further work which has now occurred with the result that some of the provisions to be inserted in the OECD Model in 2017 (and already in the multilateral treaty) have changed; in this category are the “simplified” LOB rule, the third country PE abuse rule and the PE time threshold splitting rule (MLI Articles 7, 10, 14, ES paras 104, 142, 183).
Reservations, options and elections
Most importantly the content of the MLI is all effectively elective to a greater or lesser degree.
The only bottom line change required is to adopt the new treaty preamble referring to double non-taxation, but even then it need not be applied to existing bilateral treaties which have the same language (such as the Australia-Germany 2015 treaty). While it is mandatory for countries to adopt other BEPS minimum standards, they all have options within and outside the MLI, including using bilateral treaty, domestic law or administrative practice changes. For example, the MLI does not have a US style “detailed” LOB provision as the US is allowed to satisfy the standard through bilateral treaties (and it seems there was not the time or willingness to draft an MLI provision just for the US).
The minimum standards only cover some of the BEPS suggested treaty changes (three articles in the MLI) but the MLI goes far beyond those standards, which adds further to the available options.
So countries can one way or another “reserve” on virtually everything. Generally speaking, particular changes in the multilateral treaty will not apply to a nominated bilateral treaty on the list of both countries unless the reservations etc of each country match each other in relation to that change. The significance of using reservations as the main mechanism for setting a country’s MLI position is that the country is bound by the treaty change even though it does not have to apply it. If a country later withdraws a reservation, then the MLI provision affected by the reservation starts to apply going forward, without regard to the earlier reservation for all covered bilateral treaties where other countries have taken a similar position. To put it another way, over time the MLI can have its coverage for nominated treaties broadened but never narrowed (except in relation to arbitration).
Reservations are different to options (which give countries a menu of choices in some cases such as PEs) and the election to opt into compulsory arbitration which is the subject of much more extensive provisions in the MLI than the usual bilateral treaty form. For countries electing in (including Australia) arbitration is the biggest change in the whole treaty.
The flexibility of the MLI at almost every step could lead to strategic behaviour by countries to preserve what they regard as favourable treaty positions with respect to one or more other countries. The ES refers implicitly at a number of points to the general treaty obligation of good faith and seeks to limit such behaviour, eg, parties to the MLI will seek to identify the maximum number of treaties to be subject to the MLI and include them in their lists of affected treaties unless there are legitimate reasons to exclude them such as being recently negotiated and already including BEPS measures. On the other hand for MLI provisions that do not relate to mandatory new standards, countries are not obliged to sign up, or if they do, they have only to act consistently towards other countries in making reservations.
The ES will have a very high interpretive status in international law – to almost be binding, which is most significant in relation to arbitration. The ES otherwise, however, is largely concerned with procedural areas. The meaning of the substantive changes is still being left to the OECD Model and its Commentaries, so the usual interpretation issues about the status of OECD material – which version applies etc – will continue into the MLI. As the BEPS version of the Model and Commentaries is not due until later in 2017 and as already noted will reflect some changes from the 2015 BEPS Final Reports, the status of the new Commentaries as opposed to the versions in the 2015 Final Reports may become a live issue in future.
(In passing it may be noted that the BEPS project’s obsession with deadlines seems to have relaxed in 2016. A number of promised 2016 products, including the full revised version of the Transfer Pricing Guidelines did not appear and the latest OECD webcast suggests that the 2015 Final Reports on transfer pricing may change further and that there will be more delays in finalising that work. At the end of 2016, the OECD published an updated version of the Action 4 report on interest deductions, which includes the material released in draft earlier in that year on the banking and insurance sectors and the group ratio rule.)
It is also made clear in the MLI that countries can, by subsequent bilateral protocols or treaties, change the text of the MLI – the last in time prevails. Equally the MLI itself can be the subject of protocols but these will bind a particular country only if it separately signs on to the protocols.
For some countries, though not Australia, there is also an awkward language issue. The MLI only exists in English and French as official languages. The bilateral treaties being amended may not exist in either English or French. Translators in countries with such treaties will be kept busy.
Indeed it is likely that some countries will produce unofficial consolidated versions of bilateral treaties as amended by the MLI but because of the difference in drafting style it will be a challenging exercise to get the text right. The CP indicates that as with existing bilateral treaties and subsequent protocols, Australia will not consolidate the MLI amendments into official tax treaty texts. That exercise will apparently continue to fall to private publishers.
At least countries are being required in making their reservations etc to identify the affected bilateral provisions in detail, under notification clauses in the MLI, so there will be some official starting point.
Similarly the MLI has compatibility clauses describing how a particular change in the MLI which is adopted by a country interacts with its bilateral treaties which are affected by the MLI.
The OECD as Depositary for the MLI is currently developing software to assist in identifying which bilateral treaties will be amended in which respects by the MLI, based on notifications by countries, which will start to flow in mid-2017.
Countries which sign the MLI may in the future withdraw from it. If that occurs changes made to existing bilateral treaties through the MLI will not be affected and it will still be necessary to either terminate or renegotiate the bilateral treaty for a country to rid itself of particular MLI treaty obligations.
Timing – when will the big bang occur Down Under?
The MLI is already open for signature – from 31 December 2016 – and will not start to operate until five ratifications have been deposited with the OECD. In Australia’s case ratification will require the usual treaty review process and for a bill to pass through Parliament giving effect to whatever we sign up to. The CP indicates a target start date for the MLI in Australia of 1 January 2019, assuming sufficient ratifications by then, which in the light of the timing indicated in the MLI for coming into force and effect means passage of enabling legislation by mid-2018.
In the meantime Australia has to draw up a list of treaties it wants to be amended, sort through the many possible reservations, options and elections to decide its position on each one and then as a result identify all the affected provisions in the nominated treaties as required to be notified to the Depositary as part of signing and ratifying the MLI. The CP was the means to get feedback for this process. It is understood that countries prepared provisional lists for the last meeting in November 2016 of the group negotiating the MLI. The OECD has indicated that there is to be a “speed dating” meeting of treaty negotiators in February 2017 to allow countries to get together on both a bilateral and multilateral basis so they have a sense of what other countries are doing, thus allowing each country to firm up its own positions. Hence the very short consultation timeframe in the CP.
It is unlikely there will be any signatures on the MLI before the proposed signing ceremony in Paris in the week of 5 June 2017 back-to-back with the OECD Ministerial Council meeting when a sufficient number of high-profile politicians will be on hand to do the honours. The MLI provides for provisional notification of all reservations etc by countries at the time of signature and final notification at the time of depositing instruments of ratification.
What is Australia proposing to do?
Given Australia’s gung-ho attitude to BEPS to date, it is no surprise that Australia is proposing in the CP to embrace as much of the MLI as possible and to make only limited use of reservations. We have attached as an annex a modified version of the table from the CP summarising the proposed approach. Here we comment on some of the less obvious issues to which taxpayers (and in some cases tax administrators) need to direct their attention.
The fiscally transparent entity provision in Article 3 of the MLI, which Australia proposes to adopt, already appears in Australia’s treaties with New Zealand and Germany (and in a longer version in the France and Japan treaties). The main concern about this provision is that the accompanying OECD Commentary leaves it unclear whether the rule will apply to discretionary trusts (because it copies some language from US regulations in which the examples make clear that the US regulations do not apply to discretionary trusts). The ES para 45 makes clear that this provision will not displace specific provisions in treaties in relation to PEs of trusts being deemed to be PEs of beneficiaries (the main user of which internationally is Australia).
Australia proposes to sign up to the saving clause currently found in Australia’s treaty with the US, which Australian professionals tend to ignore as being a US oddity. The saving clause provides that treaties do not affect the residence country’s taxation of its own residents, with some specific exceptions such as the residence country’s treaty obligation to relieve international double taxation. It will have three particular impacts of note.
First, it will achieve the outcome that Australia is not prevented from taxing if the relevant entity is an Australian resident and the income Australian source, which the fiscally transparent entity provision may otherwise prevent. Oddly Australia has not included this protection in its German and New Zealand treaties but achieved a similar result by a side agreement between the treaty negotiators. Secondly, it will mean that the normal treaty transfer pricing provision typically found in Article 9(1) never applies (as it can only apply to residents and is not carved out of the saving clause) so that the ATO argument that tax treaties give an independent power to adjust will be hypothetical and the status of the Transfer Pricing Guidelines as the international standard will have to be ensured through domestic law (as has recently occurred in Australia but not in many other countries). Thirdly, it will ensure that tax treaties do not override Australia’s CFC regime, a result already arguably reached by Australian case law.
Not surprisingly given recent treaties, the Multinational Anti-Avoidance Law (MAAL) and proposed Diverted Profits Tax (DPT), Australia is proposing to sign up to the PPT as its treaty anti-abuse rule and not to add on the simplified LOB. It should be noted that the OECD has already published this year a discussion draft as part of developing further OECD Model Commentary on the PPT, specifically in relation to non-CIV funds, which raises the issue of what status that will have for the MLI given it postdates the finalisation of the MLI.
Australia proposes to sign up to all the PE definition changes. The OECD BEPS follow-up work on attribution of profits to PEs was not completed in 2016 as promised and we understand may never see the light of day. It is now going on four years since the Board of Taxation finalised its report on PE attribution and the government has not announced its response. So PEs will now be more broadly defined but we still will not know how to tax them – either from the OECD or from the Australian government side. To further complicate the policy picture, the third country PE anti-abuse rule is one of the few BEPS treaty related measures in the MLI that Australia is proposing to reserve on, which is not to suggest that such a reservation is inappropriate.
In relation to arbitration, Australia proposes to sign on – and then sign off if Part IVA is involved, thus effectively trying to insulate MAAL and DPT assessments from being subject to arbitration processes. Given that the DPT in particular is likely to produce double taxation (by denying the availability of FITOs in relation to a DPT assessment), this approach conflicts with the fundamental approach of tax treaties of preventing double taxation. One suspects that Australian officials have not thought through the strategic options available to taxpayers to circumvent this obviously unfair result.
And what will President Trump have to say about all of this?
There has been much twittering among the commentariat about whether President Trump will trash BEPS like the Trans Pacific Partnership – indeed he does not even need to sign an executive order in this case, just not sign the MLI and not pursue various other BEPS proposals (apart from County-by-Country reporting and arbitration). But that was a possibility whoever became US president, and the BEPS package has been designed in various ways to cater for that possibility.
The much more important issue is whether the new President will effectively ditch the whole tax treaty network and international taxation as we know it, by replacing the US corporate income tax with a destination based cash flow tax. This proposal is being seriously considered on the Republican side of Congress.
It may be doubted whether such a tax is an income tax and if not tax treaties would probably not apply to it. If it is an income tax, treaties would prevent the US levying the tax on non-US residents in the absence of a PE or subsidiary there making the sale. If the rest of the world continues with existing income taxes, there will be international double taxation (taxation in both production and consumption countries for imports into consumption countries) and double non-taxation (for imports into production countries) of the like never seen before.
If the rest of the world followed the US, the result would be an enormous cut in corporate tax revenues in Australia, Canada, South Africa and nearly all developing countries (as most revenue arising from natural resources produced by a country would no longer be taxed there as it is virtually all exported) and transfer the revenue to wealthy consumption countries, ie, Europe, US and Japan. These defects of the proposal are well known and should mean that the idea is an economic theory without practical relevance.
But one can never tell with the US in these unpredictable days.
Click here to view the Annex to this Tax Brief.