Page 1
© Richard Vann 2016
Disclaimer: The material and opinions in this paper are those of the author and not those of The Tax Institute. The Tax Institute did not review the contents of this paper and does not have any view as to its accuracy. The material and opinions in the paper should not be used or treated as professional advice and readers should rely on their own enquiries in making any decisions concerning their own interests.
2016 FINANCIAL
SERVICES TAXATION
CONFERENCE
Financial Services in a Post-BEPS World
Written by:
Richard Vann
Challis Professor
Sydney Law School
Consultant,
Greenwoods &
Herbert Smith
Freehills
Presented by:
Richard Vann
17–19 February 2016
Surfers Paradise Marriott Resort & Spa, Gold Coast
Page 2
Richard Vann Financial Services in a Post-BEPS World
© Richard Vann 2016 2
CONTENTS
1 Overview ......................................................................................................................................... 3
2 BEPS Policy and the Corporate Tax ............................................................................................ 4
3 Hybrid Entities and Tax Treaties .................................................................................................. 8
3.1 BEPS treaty provisions ............................................................................................................. 8
3.2 Relationship to OECD Partnership Report and other treaty provisions .................................... 9
3.3 What does fiscally transparent mean? .................................................................................... 11
3.4 Other problems with fiscally transparent entities .................................................................... 14
3.4.1 Special conditions for treaty relief .................................................................................... 14
3.4.2 Procedure ........................................................................................................................ 15
4 Tax Treaty Abuse ......................................................................................................................... 17
4.1 Action 6 Report ....................................................................................................................... 17
4.2 Application of LOB and PPT Tests ......................................................................................... 18
4.3 PE and special regime abuse rules ........................................................................................ 20
5 Permanent Establishment ........................................................................................................... 22
5.1 BEPS on Permanent Establishment ....................................................................................... 22
5.1.1 Time thresholds ............................................................................................................... 23
5.1.2 Preparatory and Auxiliary Activities ................................................................................. 23
5.1.3 Agency PEs ..................................................................................................................... 24
5.2 At Sixes and Sevens: The Australian MAAL........................................................................... 26
5.2.1 Off on a frolic? ................................................................................................................. 26
5.2.2 Operation of MAAL .......................................................................................................... 27
Page 3
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 3
1 Overview
This paper will examine (some aspects of) the future state of the financial services industry in a post-
BEPS world. The G20/OECD base erosion and profit shifting (BEPS) project as is well known was
driven very much by the digital disruption to the tax system, not to mention the economy, caused by
new technology and ways of doing business. To date that disruption has not shown up so much in the
financial services industry in Australia as in other areas.
When the BEPS Final Reports were released on 5 October 2015, the new Australian Treasurer was
very quick to adopt more or less all of it and indicate that Australia was already well on its way to
implementation, or to put it another way, there was not much in the way of further implementation
required of Australia.1 Since then the government has been actively proceeding with implementation
of those BEPS measures not already in place (hybrids mentioned in the next paragraph, the bill on
country-by-country reporting, the new thoroughly-BEPSed treaty with Germany) and BEPS inspired
measures (such as the multinational anti-avoidance law).
Although the BEPS project has reached the conclusion of its development phase and is now in the
implementation phase, there is still much substantive work to be done in a number of areas, including
some which is of most interest to the finance industry. For example, the BEPS Report on Action 22 on
hybrids requires extensive work on domestic law implementation for which consultation is currently in
progress by the Board of Taxation.3 That Report leaves up in the air much that is of concern to
financial institutions, including deductible-frankable hybrids, transition and grandfathering, and
regulatory capital hybrids. It is not possible to predict where all this will lead both in Australia and
internationally at the moment but there are signs that countries are likely to struggle with domestic law
implementation of this particular Action, compared to the treaty aspect of hybrids which is much easier
to implement, though currently unclear on exactly what the treaty provision means.
Rather than simply catalogue what the G20/OECD decided and what has happened in Australia this
paper will be selective, looking at substantive issues (rather than transparency covered elsewhere
during the conference and procedural issues) where outcomes are fairly clear at the OECD level and
implementation very likely, but problems already apparent. It covers the treaty measures on hybrid
entities, abuse and permanent establishment, as well as comparing and contrasting the MAAL with
the BEPS outcomes and discussing the broad policy issue of what BEPS means for the corporate
income tax. The last issue is directly relevant to the tax reform debate in Australia and is taken up first.
Transfer pricing is not covered as it is also dealt with in another paper.
1 Treasurer, OECD report supports Australian Government action on multinational tax avoidance (6 October 2015), especially
the table at the end. 2 OECD/G20 Base Erosion and Profit Shifting Project, Neutralising the Effects of Hybrid Mismatch Arrangements ACTION 2:
2015 Final Report (2015). 3 Board of Taxation, Implementation of the OECD Anti-Hybrid Rules Consultation Paper (2015).
Page 4
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 4
2 BEPS Policy and the Corporate Tax
There has been a tax policy paradox surrounding the corporate tax in Australia: in the domestic tax
reform debate the corporate tax has been derided as one of the most inefficient taxes in Australia,
inhibiting growth and being largely borne by labour, not shareholders. In the international tax debate
surrounding BEPS the corporate tax has been seen as an essential tax and the means of taxing
foreign capital as well as ensuring the fairness of the tax system. The cry in the tax reform debate is
cut the corporate tax and replace it with other more efficient taxes; the cry in the international tax
debate is bolster the corporate tax and ensure it operates as intended. One of the unnoticed
achievements of the BEPS work has been to illuminate part of this policy paradox.
The illumination on this issue comes mainly from the Action 1 Report on the digital economy4 though
there is some similar analysis in the other report dealing directly with policy, Action 11 on measuring
BEPS.5 As with many often-proclaimed economic truths about taxation, the apparently contradictory
views of the corporate tax depend on the assumptions underlying the economic models. The Treasury
study in relation to tax reform which ranks the corporate tax as inefficient6 is based on perfect
markets: perfectly competitive markets without economic rents, perfectly capital mobility etc. It also
assumes that the supply of foreign capital is perfectly elastic, the supply of domestic capital is
perfectly inelastic, the elasticity of domestic labour demand with respect to capital stock relatively high
because capital and labour are modelled as gross complements – the rise in price of capital relative to
the wage reduces both capital and labour. These assumptions more or less inevitably lead to the
conclusions about the inefficiency of the tax and its incidence on labour.
The Digital Economy Report recognises that there are economic rents involved in the digital economy
which vary over time, ie not perfectly competitive markets. This is enough to change the outcome of
economic models significantly. The conclusions are (see Annex E Economic incidence of the options
to address the broader direct tax challenges):
• In the case of a perfectly competitive market for digital goods and services, the incidence of
the corporate income tax increase is likely to be borne by labour in the affected foreign
suppliers’ production country and consumers in market countries, depending on the
importance of the affected suppliers in the particular market and the availability of
replacement suppliers with similar cost structures and the availability of alternative goods and
services.
• If the market is imperfectly competitive, the corporate income tax increase is likely to be
borne principally by the equity owners of the affected foreign suppliers.
4 OECD/G20 Base Erosion and Profit Shifting Project, Addressing the Tax Challenges of the Digital Economy ACTION 1: 2015
Final Report (2015). 5 OECD/G20 Base Erosion and Profit Shifting Project, Measuring and Monitoring BEPS ACTION 11: 2015 Final Report (2015).
6 Treasury, UNDERSTANDING THE ECONOMY-WIDE EFFICIENCY AND INCIDENCE OF MAJOR AUSTRALIAN TAXES
(2015) Treasury Working Paper 2015-1.
Page 5
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 5
Always be wary of economists with strong tax messages and ask them about the underlying
assumptions!7 The Action 11 report recognises taxing economic rents as one of the justifications for
the corporate tax (see pp 112-116, 182).
Another aspect of the Action 11 report highlights the two main policies underlying substantive BEPS
proposals. The one that is quoted very commonly (and often taken to describe the BEPS project
overall) is as follows:8
No or low taxation is not per se a cause of concern, but it becomes so when it is associated
with practices that artificially segregate taxable income from the activities that generate it.
This sentence appears in a paragraph which identifies (not very clearly) not one but two policy drivers
as follows:
BEPS relates chiefly to instances where the interaction of different tax rules leads to double
non-taxation or less than single taxation. It also relates to arrangements that achieve no or
low taxation by shifting profits away from the jurisdictions where the activities creating those
profits take place. No or low taxation is not per se a cause of concern, but it becomes so
when it is associated with practices that artificially segregate taxable income from the
activities that generate it. In other words what creates tax policy concerns is that, due to gaps
in the interaction of different tax systems, and in some cases because of the application of
bilateral tax treaties, income from cross-border activities may go untaxed anywhere, or be
only unduly lowly taxed.
To paraphrase, one policy target is exploiting gaps and differences in tax systems’ rules to produce
low or no taxation and the other is separating the location of income from the location of value adding
activity to produce low or no taxation; if neither target is present then the fact that a country has low or
no corporate taxation is not a cause for concern. The policy division of the BEPS substantive work
partly reflects and partly hides this division with Actions 2-5 grouped under the rubric “Establishing
international coherence of corporate income taxation” (intended to cover the first policy) and Actions
6-10 under “Restoring the full effects and benefits of international standards” (intended to reflect the
second policy). The match is not complete, however. Only Action 2 on hybrids and possibly Action 3
on CFC regimes reflects the first policy directly.9 Action 4 on interest deductions and even more
Action 5 on harmful tax practices (so far as Action 5 is about patent boxes and the like as opposed to
transparency of rulings) more easily fit into the second policy concerning separation of income from
value adding activities.10
7 The government has just released more recent modelling which is said to show that the GST-income tax switch will produce
little or no efficiency gains and be very unfair with a great impact on those on lowest incomes even after compensation, ABC,
Tax reform: Treasury modelling presents taxing dilemma for Government, see http://www.abc.net.au/news/2016-02-
12/treasury-modelling-presents-taxing-dilemma-government/7161640 and http://www.abc.net.au/news/2015-12-09/eight-
options-for-tax-reform-explained/7013380. It will be interesting to know what the assumptions in this modelling are. 8 OECD, Action Plan on Base Erosion and Profit Shifting (2013) p 10.
9 OECD/G20 Base Erosion and Profit Shifting Project, Designing Effective Controlled Foreign Company Rules ACTION 3: 2015
Final Report (2015). 10
OECD/G20 Base Erosion and Profit Shifting Project, Limiting Base Erosion Involving Interest Deductions and Other Financial
Payments ACTION 4: 2015 Final Report (2015), OECD/G20 Base Erosion and Profit Shifting Project, Countering Harmful Tax
Practices More Effectively, Taking into Account Transparency and Substance ACTION 5: 2015 Final Report (2015).
Page 6
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 6
To establish coherence of the corporate tax it is necessary to know what the purpose of the corporate
tax is. Commonly two goals are stated. The first has to do with the realisation basis of the income tax
and the treatment of companies as separate taxpayers from their shareholders. If there were no
corporate tax, income tax could simply be avoided at both company and shareholder levels by the
company retaining all its income and the shareholders not selling their shares. The corporate tax is
thus necessary as a tax on retained income of companies. This is largely a domestic justification for
the tax in relation to resident shareholders.11
The second purpose is to effectively tax foreign
shareholders: it is much easier administratively to collect the tax from a company operating in a
country than from its foreign shareholders and reflecting this the international tax system has been set
up to significantly limit cross border tax on dividends, especially intragroup dividends and capital gains
on shares held by foreigners.
The corporate tax is also essentially a source based tax even though it is nominally levied on a
residence and source basis in most countries. This is shown most clearly by the participation
exemptions now common for foreign active income of resident companies, and the exclusion of active
income from CFC attribution. For active income the residence tax is levied on the ultimate
shareholders of the corporate group rather than the companies in the group. I have argued this
proposition at length elsewhere12
If it is accepted, then it has a number of consequences for the policy
of the BEPS project. It justifies taxing profits where value adding activities occur (as this is essentially
restoring international standards on source based taxation which is the objective of Actions 6-10). It
also justifies the work on interest deductions which is essentially focused on stripping income from
source countries even though applied generally to ensure it does not offend various non-
discrimination tax rules, particularly in the EU, and the work on patent boxes which effectively requires
that reduced taxes be given only for R&D activities in a country and not as a means of attracting
income without activity from other countries, even though written on an entity rather than location
basis for similar reasons.13
When it comes to Action 2 and Action 3, the underlying policy seems to be that the relevant income
which exploits gaps or differences between tax regimes should be taxed somewhere but it does not
matter much where. This is hardly a tax-policy principled approach though it may be justified by game
theory where players are better off through co-operation. For that reason these Actions in my view are
going to prove the hardest to implement, apart from their technical complexity. Further policy
differences over the purpose of CFC regimes explain why it was not possible to get consensus on the
Action leaving it as a best practice (ie optional) recommendation, rather than an obligatory standard or
an almost obligatory common approach in BEPS-speak. Some countries consider the purpose of CFC
rules is to prevent domestic to foreign stripping (round tripping), that is, undermining the source tax
base of the company’s residence country,14
whereas other countries consider that CFC regimes are
11 The corporate tax also operates domestically as a tax on distributed income. In Australia the imputation system turns the
corporate tax into a withholding tax on income distributed to resident shareholders. Other countries nowadays generally have
lower taxes on dividends paid to non-corporate shareholders in recognition of corporate tax having been paid. 12
Vann, Taxing International Business Income: Hard-Boiled Wonderland and the End of the World (2010) 2(3) World Tax
Journal 291-346. 13
The article in the previous note also argues that in the real world the residence of corporations effectively operates as a
sourcing principle in international taxation. In the case of the Action 5 the substantial activity test is applied to activities of the
entity in question, not its associated entities, which indirectly operates as location test. 14
Two recent examples in the Australian context are the Chevron case [2015] FCA 1092 and the Orica case [2015] FCA 1399
both of which involved an Australian parent company paying interest to a US subsidiary (though in the first case there was an
ultimate US parent). It is noteworthy that in neither did the ATO seek to apply the Australian CFC regime for reasons which are
Page 7
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 7
also intended to deal with foreign to foreign stripping, that is, tax in the residence country of a parent
company discourages stripping of income by the group from a source country into a CFC where it is
subject to no or low taxation. One difficulty with the latter policy of taxation based on residence rather
than source when there are chains of companies resident in various countries is determining which
residence country CFC regime should apply, which in itself demonstrates the frailty of corporate
residence as a genuine taxing nexus. The answer in ch 7 of the Action 3 report is all of them but with
foreign tax credits for unrelieved double taxation in the chain subject to the usual foreign tax credit
limit, which seems to mean tax at the highest rate of all countries in the chain.
The lack of firm policy principle in real corporate residence taxation is also evident in the report on
Action 2. Consider a debt-equity hybrid instrument within a corporate group (debt in the payer country,
equity in the payee country) which prior to the BEPS project produced a deductible/non-inclusion
(D/NI) outcome. There seem to be three recommendations for this case. First, the payee country
should modify its participation exemption to deny it when the payment is deductible in the other
country; then the payer country should deny the deduction if there is a hybrid mismatch, that is, D/NI
outcome; finally the payee country should eliminate the exemption if there is a hybrid mismatch which
the payer country has not eliminated (even though the hybrid outcome in the second and third steps
seems to result from the lack of action of the payee country under the first step).15
Compare this to
Action 4 on interest deductions where the focus is the payer (source) country.
This is not to say that there may not be cases where real corporate residence taxation is appropriate
– income truly without a source in a country such as from international transport or activities in terra
nullius (Antarctica, space) may be examples, but such taxation on the basis of failure to tax at source
due to gaps or differences in tax systems (foreign to foreign stripping and D/NI hybrids) but not in
other cases where the source country does not tax is clearly more debatable. It will be interesting to
compare country implementation of Action 2 and Action 3 with other BEPS Actions.
While the analysis in the Action 11 Report is clear that policy in the corporate tax is complex and
contestable and seeks seriously to come to grips with the policy and associated measurement and
modelling issues raised by BEPS, obviously it was not going to be critical of the project. Nonetheless
its nuanced approach to the corporate tax is much more sophisticated than and a healthy corrective to
the simple minded material we have been getting in the domestic tax reform debate on the corporate
tax.
more obvious in the Orica than in the Chevron case, and instead relied on transfer pricing rules in Chevron and Part IVA in
Orica. 15
Recommendations 2(1), 1(1)(a), 1(1)(b) respectively.
Page 8
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 8
3 Hybrid Entities and Tax Treaties
3.1 BEPS treaty provisions
In relation to the tax treaty treatment of hybrid entities the OECD/G20 Report on Action 2 ch 14 has
adopted the following Article 1(2), which is expected to be included in the BEPS multilateral treaty,
probably on an optional basis:
For the purposes of this Convention, income derived by or through an entity or arrangement
that is treated as wholly or partly fiscally transparent under the tax law of either Contracting
State shall be considered to be income of a resident of a Contracting State but only to the
extent that the income is treated, for purposes of taxation by that State, as the income of a
resident of that State.
It appears in article 1(2) of the Australia Germany 2015 treaty with the addition of “(including profits or
gains)” after the first reference to “income”, and a variant of it is in article 1(2) of the 2009 Australia
New Zealand treaty.
The provision is designed to ensure that hybrid entities are not used to get treaty benefits when the
income is not being taxed in the residence state just because the states characterise the entity
differently with respect to particular income for tax purposes (as transparent in one state and as
opaque in the other). As an example assume that interest income is derived from source state S by an
entity based in state E with its members resident in state R. S and E regard the entity as transparent
and therefore attribute the income to the members resident in state R. State R regards the entity as a
tax opaque company resident in state E and therefore does not tax the members resident in state R
on the income when it is derived by the entity (assuming ant CFC rules in state R are not activated).
Under this provision treaty benefits cannot be claimed in state S under the treaty between state S and
state R due to the words “but only” etc in the new provision. Conversely the provision is intended to
ensure that treaty benefits flow where otherwise they might be denied. Suppose in the example in the
previous paragraph state S regarded the entity as a tax opaque company but state E and state R
regarded it as tax transparent. Even though state S would attribute the income to the entity in state E,
state R would tax the members in state R. The provision in this case ensures that the members get
the benefit of the treaty between state S and state R. This double operation is made clear in the
OECD Commentary to the new provision para 26.6.
Both these recent Australian treaties lack one important additional element – the provision based on
the US saving clause which is recommended in the OECD/G20 Report on Action 616
pp 86-89 as
Article 1(3) of the OECD Model:
This Convention shall not affect the taxation, by a Contracting State, of its residents except
with respect to the benefits granted under paragraph 3 of Article 7, paragraph 2 of Article 9
and Articles 19, 20, 23 A [23 B], 24 and 25 and 28.
16 OECD/G20 Base Erosion and Profit Shifting Project Preventing the Granting of Treaty Benefits in Inappropriate
Circumstances ACTION 6: 2015 Final Report (2015).
Page 9
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 9
This has an important part to play in the treatment of hybrids (as is recognised in the Action 2 Report
p 143). By way of variation on the previous examples assume the entity is resident in state S which
regards it as a tax opaque company whereas state R regards it as transparent. The saving clause has
the result that state S is not obliged in this case to apply the treaty between state S and state R as
article 11 is not mentioned as an exception. While it can be argued whether this is the right result, the
OECD view is based on the premise that a state should not generally be prevented by treaty from
taxing its own residents (which also has important applications in relation to CFC regimes and the
like). In the case of the New Zealand treaty this issue is “solved” by an agreement between the
negotiators reflected in the EM to the act implementing the treaty para 2.25:17
During negotiations, the two delegations noted that:
‘It is understood that (this) paragraph shall not affect the taxation by a Contracting
State of its residents.’
Presumably a similar fix will be applied to the new German treaty (no draft EM has appeared to date)
but it is not clear why the saving clause was not adopted in the treaty.
3.2 Relationship to OECD Partnership Report and other treaty
provisions
Those familiar with the OECD 1999 Partnership Report18
will recognise that the examples and
solutions given in the previous section come from that report which derived them as a general matter
of treaty interpretation rather than as following from express provisions. The ATO has accepted the
conclusions in that sense but confined them to partnerships in TD 2011/25. Further the Partnership
Report applies on an all or nothing basis, the issue of whether an entity that is transparent for some
income but not for other income being left for future analysis (which has never been finalised or
published). Trusts fall into this category as well as not being partnerships and that is why the New
Zealand treaty adopted a specific provision to ensure broader coverage of the principles, see EM para
2.10. From an OECD perspective the fact that not all countries accepted the conclusions of the
Partnership Report provided an additional reason for inclusion of the provision, Action 2 Report p 139
para 435.
The New Zealand EM para 2.10 and the new OECD Commentary paras 26.4-26.7 are clear that the
treaty provision is intended to produce the same outcomes as the Partnership Report but as will
appear that is not completely obvious and in any event both the German and New Zealand versions
vary to some degree from the OECD language and there are other even more differently worded
provisions in Australia’s treaties dealing with partnerships and other transparent entities, which raises
the question to what extent the principles of the Report are displaced by express treaty provisions. In
fact the OECD in its 1999 Report warned against express treaty provisions creating problems rather
than solutions, which is repeated in the new Commentary para 26.4.
17 Explanatory Memorandum (EM) to International Tax Agreements Amendment Bill (No 2) ch 2.
18 OECD, ISSUES IN INTERNATIONAL TAXATION No. 6 The Application of the OECD Model Tax Convention to Partnerships
(1999).
Page 10
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 10
In particular Australia’s treaties with France (2006) and Japan (2008) have detailed provisions dealing
with the issue. The latter treaty is also designed to get the same results as the Partnership Report
(and the New Zealand treaty) though with much more detailed drafting as a comparison of the Japan
and New Zealand EMs will show (there is significant duplication of text and examples, despite the
quite different treaty language).19
The Australian treaty with France Article 29 has similarities with the
Japan treaty approach, though not as much as more recent French treaties which are almost identical
to the Japanese language, eg France United Kingdom (2008).20
Some features of the France treaty
are discussed under the next heading.
This interaction issue was raised but not discussed in the Resource Capital Fund III LP Case.21
Both
the 1953 and 1982 treaties between Australia and the United States contained provisions on
partnerships. The provision in 4(1) of the 1982 Australia United States treaty states:
(b) a person is a resident of the United States if the person is: …
(iii) any other person (except a corporation or unincorporated entity treated as a corporation
for United States tax purposes) resident in the United States for purposes of its tax, provided
that, in relation to any income derived by a partnership, an estate of a deceased individual or
a trust, such person shall not be treated as a resident of the United States except to the
extent that the income is subject to United States tax as the income of a resident, either in its
hands or in the hands of a partner or beneficiary, or, if that income is exempt from United
States tax, is exempt other than because such person, partner or beneficiary is not a United
States person according to United States law relating to United States tax.
The trial judge commented on this and the 1953 provision as follows:
Under art II(1)(g) of the 1953 Convention, a partnership created or organised in or under the
laws of the US (a US domestic partnership) was a resident of the US for the purposes of the
1953 Convention whether or not the partners were liable to US tax on the income of the
partnership. If the partners in the US domestic partnership were foreign companies (ie,
incorporated outside the United States) and the partnership income was not effectively
connected with the conduct of a trade or business carried on in the US, neither the
partnership (because it was fiscally transparent) nor the partners would be liable to US tax on
the partnership income, but because the partnership was a resident of the US for the
purposes of the 1953 Convention, it was entitled to the benefits of the 1953 Convention vis-à-
vis Australian source income. The negotiators of the [1982] Convention were conscious of this
anomaly when drafting art 4(1)(b)(iii); thus, under the Convention, even a US domestic
partnership will only be a resident of the US for the purposes of the Convention to the extent
that the income of the partnership is subject to US tax in the hands of a partner or, if that
income is exempt from US tax, is exempt other than because such partner is not a US person
according to US law relating to US tax.
19 Compare EM to International Tax Agreements Amendment Bill (No 1) 2008 paras 1.43-1.51 with New Zealand EM paras 2.8-
2.33. 20
Compare Article 4(5) in both treaties. 21
Resource Capital Fund III LP [2014] FCAFC 37, 16 ITLR 876, reversing [2013] FCA 363, 15 ITLR 814; leave to appeal to
High Court refused [2014] HCA Trans 235. For an analysis from which the comments in this paper on RCF are drawn, see
Vann, Hybrid entities in Australia: Resource Capital Fund case in Lang et al, eds, Tax Treaty Case Law 2015 (2016, IBFD
forthcoming).
Page 11
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 11
The problem with the 1953 provision was that it gave treaty benefits even if there were no US
taxpayer liable to tax in the United States. As demonstrated by the RCF case the 1982 provision had
the opposite vice of being too narrow in the sense of achieving a similar outcome to the Partnership
Report; it only operated to give treaty benefits to the LP directly if both the LP and the partners were
resident under US domestic tax law. As RCF was a Cayman Islands LP, the 1982 provision did not
apply to it.
Clearly it was possible to argue on the wording of the Australia United States 1982 treaty that its
express provision on partnerships excluded the operation of the Partnership Report principles to the
extent that they were more extensive. Neither party took this position and the courts seemed willing to
accept the view that the Partnership Report was relevant to the partners, though the judges of the Full
Federal Court expressly indicated that they were not deciding this issue. While the provision now
proposed by the OECD is designed to overcome this particular issue, the same issue will arise if and
to the extent that it is more limited than the principles in the Partnership Report.
3.3 What does fiscally transparent mean?
To come within the new OECD provision the entity must be “fiscally transparent” so it is critical to
know what this phrase means. The new Commentary is fairly brief:
26.10 The concept of “fiscally transparent” used in the paragraph refers to situations where,
under the domestic law of a Contracting State, the income (or part thereof) of the entity or
arrangement is not taxed at the level of the entity or the arrangement but at the level of the
persons who have an interest in that entity or arrangement. This will normally be the case
where the amount of tax payable on a share of the income of an entity or arrangement is
determined separately in relation to the personal characteristics of the person who is entitled
to that share so that the tax will depend on whether that person is taxable or not, on the other
income that the person has, on the personal allowances to which the person is entitled and on
the tax rate applicable to that person; also, the character and source, as well as the timing of
the realisation, of the income for tax purposes will not be affected by the fact that it has been
earned through the entity or arrangement. The fact that the income is computed at the level of
the entity or arrangement before the share is allocated to the person will not affect that result.
States wishing to clarify the definition of “fiscally transparent” in their bilateral conventions are
free to include a definition of that term based on the above explanations.
This language has two sources. The first part of the second sentence comes from the Partnership
Report para 40. The second part of the sentence is a shorthand reference to the language in the US
regulations under Internal Revenue Code s 894(c)22
which defines the meaning of fiscally transparent,
s 1.984-1(d)(3)(ii):
an entity is fiscally transparent under the laws of the entity's jurisdiction with respect to an
item of income to the extent that the laws of that jurisdiction require the interest holder in the
entity, wherever resident, to separately take into account on a current basis the interest
holder's respective share of the item of income paid to the entity, whether or not distributed to
22 26 CFR s 1.894–1(d)(3) and the examples in s 1.894-1(d)(5).
Page 12
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 12
the interest holder, and the character and source of the item in the hands of the interest
holder are determined as if such item were realized directly from the source from which
realized by the entity
The examples concerning trusts in particular state:
Example 4. Treatment of grantor trust. (i) Facts. Entity A is a trust organized under the laws of
Country X, which does not have an income tax treaty in effect with the United States. M, the
grantor and owner of A for U.S. income tax purposes, is a resident of Country Y, which has an
income tax treaty in effect with the United States. M is also treated as the grantor and owner
of the trust under the laws of Country Y. Thus, Country Y requires M to take into account all
items of A's income in the taxable year, whether or not distributed to M, and determines the
character of each item in M's hands as if such item was realized directly from the source from
which realized by A. Country X does not treat M as the owner of A and does not require M to
account for A's income on a current basis whether or not distributed to M. A receives interest
income from U.S. sources that is neither portfolio interest nor effectively connected with the
conduct of a trade or business in the United States.
(ii) Analysis. A is not fiscally transparent under the laws of Country X within the meaning of
paragraph (d)(3)(ii) of this section with respect to the U.S. source interest income, but A may
not claim treaty benefits because there is no U.S.-X income tax treaty. M, however, does
derive the income for purposes of the U.S.-Y income tax treaty because under the laws of
Country Y, A is fiscally transparent.
Example 5. Treatment of complex trust. (i) Facts. The facts are the same as in Example 4
except that M is treated as the owner of the trust only under U.S. tax law, after application of
section 672(f) [26 USCS § 672(f)], but not under the law of Country Y. Although the trust
document governing A does not require that A distribute any of its income on a current basis,
some distributions are made currently to M. There is no requirement under Country Y law that
M take into account A's income on a current basis whether or not distributed to him in that
year. Under the laws of Country Y, with respect to current distributions, the character of the
item of income in the hands of the interest holder is determined as if such item were realized
directly from the source from which realized by A. Accordingly, upon a current distribution of
interest income to M, the interest income retains its source as U.S. source income.
(ii) Analysis. M does not derive the U.S. source interest income because A is not fiscally
transparent under paragraph (d)(3)(ii) of this section with respect to the U.S. source interest
income under the laws of Country Y. Although the character of the interest in the hands of M
is determined as if realized directly from the source from which realized by A, under the laws
of Country Y, M is not required to take into account his share of A's interest income on a
current basis whether or not distributed. Accordingly, neither A nor M is entitled to claim treaty
benefits, since A is a resident of a non-treaty jurisdiction and M does not derive the U.S.
source interest income for purposes of the U.S.-Y income tax treaty.
A complex trust in US tax law includes a discretionary trust and is in contradistinction to a simple trust,
which, in the relevant year, is required by the terms of the trust to distribute all its trust-law income
currently other than by deductible charitable donation, and which in that year does not distribute other
Page 13
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 13
amounts.23
There is a serious possibility based on these examples that the term “fiscally transparent”
will not cover discretionary trusts. Unfortunately the EM to the Japan treaty does not even refer to
trusts in this context while the EM to the New Zealand treaty refers to trusts but not to discretionary
trusts explicitly, though it does say in para 2.14:
In general, paragraph 2 relates to particular items of income of entities that are fiscally
transparent under the laws of one or other country. Entities falling under this description in
Australia and New Zealand include certain partnerships and trusts. In the case of Australia it
includes partnerships subject to Division 5 of Part III of the Income Tax Assessment Act 1936
(ITAA 1936) (but not corporate limited partnerships subject to Division 5A of Part III), and
trusts which are subject to Division 6 of Part III where the beneficiary of the trust is presently
entitled to the income and assessable accordingly (but not a corporate unit trust or public
trading trust subject to Division 6B or 6C of Part III).
The French treaty in Article 29 actually includes the character, source and timing language in the
treaty itself in relation to third state partnerships and Australian partnerships [my emphasis]:
In the case of a partnership … which is treated in that third State as fiscally transparent … a
partner who is a resident of a Contracting State and whose share of the income, profits or
gains of the partnership is taxed in that Contracting State in all respects as though those
amounts had been derived directly by the partner, shall be entitled to the benefits of this
Convention with respect to their share of such amounts arising in the other Contracting State
as though the partner had derived such amounts directly, subject to the following conditions:
… the partner’s share of the income, profits or gains of the partnership is taxed in the same
manner, including the nature or source of those amounts and the time when those amounts
are taxed, as would have been the case if the amounts had been derived directly
In the case of a partnership or similar entity … which is treated in Australia as fiscally
transparent … a partner who is a resident of Australia and whose share of the income, profits
or gains of the partnership is taxed in Australia in all respects as though such amounts had
been derived by the partner directly …
This language could be read as imposing one, two or three separate conditions in the italicised parts
for third state partnerships and one or two for Australian partnerships. Though the EM recognises the
different language it does not really throw much light on whether it is “elegant” variation for the same
concept or is imposing different conditions.24
The Japan treaty avoids all of this terminology and uses
the much more neutral “derived through an entity … and treated as the income … of the beneficiaries,
members or participants”.
The other treaty which may provide some guidance is the US 2001 protocol which varies the normal
trust PE provision in the business profits article to read as follows [my emphasis]:
a resident of one of the Contracting States is beneficially entitled, whether directly or through
one or more interposed fiscally transparent entities, to a share of the business profits of an
23 26 USC s 651(a), 26 CFR s 1.651(a)-1.
24 EM to International Tax Agreements Amendment Bill (No 1) 2007 paras 1.274-1.286.
Page 14
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 14
enterprise carried on in the other Contracting State by the fiscally transparent entity (or, in the
case of a trust, by the trustee of the trust estate) [my emphasis]
The concluding words could be read as suggesting that trusts are not fiscally transparent entities,
though it seems much more likely that it is a recognition that Australia taxes the trustee rather than the
trust, compare Income Tax Assessment Act 1936 s 317 definition of trust, Income Tax Assessment
Act 1997 s 960-100. Neither the Australian nor US explanations provide assistance on this issue.
It seems likely that the narrowness or otherwise of the interpretation of fiscally transparent will depend
on how close a match between derivation of income directly and derivation through an entity is
required. There are differences in timing (generally the partner or beneficiary derives its share of the
income at the end of the income year rather than when the partnership or trust derived the income),
the treatment of losses (do not pass through trusts, and subject to additional tests when passing
through limited partnerships taxed on a transparent basis in Australia) and in several other respects. If
a strict approach is taken as suggested by the US regulations then it may be found that not many
Australian entities are fiscally transparent in that strict sense. That has not been the assumption of the
Treasury, ATO or private professionals to date but Australia may find itself hostage to the US
regulations through the medium of the OECD Commentary.
3.4 Other problems with fiscally transparent entities
The RCF case also raises two other issues in the area of hybrid entities: to what extent do hybrid
entity treaty provisions deal with various special conditions for treaty relief (a matter with which the
Partnership Report struggled, paras 79-92); and how are treaty benefits asserted procedurally when
the parties to a treaty regard different persons as the relevant taxpayer.
3.4.1 Special conditions for treaty relief
There is a variety of such conditions in treaties, particularly in the dividends article which generally
has beneficial ownership and a condition for source tax rates below the normal treaty default of 15%.
In the OECD Model the conditions are expressed as the “beneficial owner is a company (other than a
partnership) which holds directly” a certain proportion of the company paying the dividend. Australia’s
treaties show some variation in this condition, though the 2015 German treaty uses this quoted part of
the OECD language.
This issue was raised by the trial judge in RCF in discussing whether the Australia US treaty was to
be applied with respect to the LP or with respect to the partners in reference to Article 10(2)(a) (which
is similar to Article 10(2)(a) of the OECD Model though referring to 10% of voting power rather than
25% of capital and omitting the reference to a partnership). He said:25
The [US] Technical Explanation 2001, discussing art 10 of the Convention, provides that
‘[c]ompanies holding shares through fiscally transparent entities such as partnerships are
considered for purposes of this paragraph to hold their proportionate interest in the shares held
25 [2013] FCA 363, para. 73. The current Article 10 was substituted by a 2001 Protocol which is why the reference is to that
year. The 1982 definition of resident quoted above remained unchanged by the Protocol so far as relevant here.
Page 15
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 15
by the intermediate entity (Technical Explanation 2001, art 6). This is consistent with the US
position that ownership measurement is performed on a look-through basis ...
From Australia’s perspective RCF was a company which directly held more that 10% of the voting
shares in the Australian mining company that it sold. If RCF had been a US resident partnership and
had received a dividend, it is strongly arguable from the Australia US treaty provision on partnerships
quoted above, that RCF was the person claiming treaty benefits under Article 10 and hence it is
arguable that it should have been entitled to the reduced rate of tax under Article 10(2)(a). The US
Technical Explanation language quoted by the judge is common in US Technical Explanations
generally and is drawn from the Technical Explanation to the US Model 2006. There, however, it is
justified by reference to the fiscally transparent entity provision in that Model from which the OECD
proposed provision is derived, whereas in the RCF case it had to be derived from general principles
(as the equivalent partnership provision in the treaty did not apply since RCF was not a US
partnership) and arguably would not have produced the same result if RCF had been a US
partnership.26
This discussion is not intended to provide conclusions on these issues but simply to point out that the
proposed OECD provision will raise as many questions as it answers, and currently many of the
questions are not dealt with explicitly or implicitly in the proposed Commentary to accompany the new
provision.
3.4.2 Procedure
Australian tax law provides relatively strict rules on who may challenge action by the ATO; for a tax
assessment, generally only the taxpayer may appeal against an assessment made in relation to the
taxpayer.27
Hence if the partners in RCF wish to raise their treaty rights under Partnership Report
principles in relation to an Australian tax assessment on RCF, the procedural question is how would
the issue come before the court. Probably RCF could raise the issue, but it may not be willing to do so
and there is no immediately obvious way under tax law for the partners to force it to do so. For
example, if RCF only had one small US investor and mainly investors from non-treaty countries, it
may not be interested in appealing on the treaty issue for the benefit of that investor alone.
This procedural issue may have influenced the reasoning of the judges, although it only came to be
clearly articulated in the application for leave to appeal to the High Court of Australia. If it is the case
that the partners have no standing to assert their treaty rights under normal domestic appeal
procedures because RCF is the taxpayer, then one has some sympathy with the view of the trial
judge that domestic law should be required to assess the partners directly and not RCF. Similarly, the
26 Further issues would also arise under Article 10(2)(a) of the OECD Model regarding the meaning of “holds directly” (which
also appears in the Australia US treaty) and the exclusion of partnerships (which does not). The protocol to the 2015 German
treaty para 3 provides, “It is understood that where dividends derived by or through a fiscally transparent entity or arrangement
are treated, for the purposes of taxation by a Contracting State, as the income, profits or gains of a resident of that State, Article
10 shall apply as if that resident had derived the dividends directly.” It is understood that this is necessary because otherwise
the words “holds directly” would not be satisfied in the German view. The protocol language adopted goes further than the ATO
has to date on this point, see TD 2014/13 accepting that shares held by custodians or nominees count for this purpose. 27
Income Tax Assessment Act 1936, s. 175A(1). Courts in recent times have sought to overcome procedural hurdles, but it is
not automatic that a way will be found.
Page 16
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 16
equivocation of the Full Federal Court on whether the partners could raise Partnership Report
principles to avail themselves of treaty benefits may reflect concerns about procedure.
In the application for leave to appeal, counsel for the taxpayer put this matter clearly:28
... because of the way in which the assessment has been raised on the collective relationship,
the partners are unable to test properly their individual liability to tax.
This concern was not directly addressed by the ATO’s counsel though there was some general
discussion of the partners challenging the assessment.29
The only clear avenue identified was the
mutual agreement procedure and counsel for the taxpayer not unreasonably maintained:30
we would respectfully say that that is not an appropriate solution. It is a matter for individual
negotiation in individual cases between the Australian Tax Office and the US Internal Revenue
Service, and that is inappropriate as a principal solution to the issue.
Counsel for the ATO made reference to the practical material in TD 2011/25 but this simply tells the
partners (and LP) what information the ATO needs and undertakes that refunds will be made if
necessary; it provides nothing on the legal means of enforcing the partner’s treaty rights.
28 [2014] HCA Trans 235, p. 14.
29 Supra n. 4, [2014] HCA Trans 235, pp. 12-13.
30 Supra n. 4, [2014] HCA Trans 235, p. 8.
Page 17
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 17
4 Tax Treaty Abuse
It is clear from the Australia Switzerland 2013 tax treaty protocol para 1, the Australian Germany 2015
treaty Article 23(2) and the MAAL discussed below that Australia will pursue the principal purpose test
(PPT) approach to treaty abuse among the menu that has been provided by the OECD/G20 Report
on Treaty Abuse.31
The German treaty Article 23(3) also indicates that preservation of domestic anti-
avoidance rules from being overridden by treaties may also be part of the Australian package.
Australia started including more limited PPT provisions in its treaties from 2003 with the Australia
United Kingdom treaty and has consistently done so since. That treaty also included the general
preservation of domestic anti-avoidance rules but that provision has not been used consistently, often
being limited to the non-discrimination article. In any event it is clear under International Tax
Agreements Act 1953 s 4(2) that Part IVA will always trump a treaty when it is successfully applied.
The Treasurer has indicated that Action 6 will be adopted in future treaty negotiations without
specifying exactly what that means, though the German treaty is probably a good guide.
In this section the BEPS Action 6 Report is discussed and then some factual scenarios are analysed
to see what it all means.
4.1 Action 6 Report
Treaty abuse has been one of the most contentious areas in the BEPS project and the OECD Report
also had to wrestle with a number of delaying factors:
The US released proposed changes to its Model in the anti-abuse area in May 2015 with
comments due by September (after the OECD work was due to be finished – perhaps a
strategic delaying tactic by the US); the final version was released on 18 February 2016 just
after presentation of this paper;
A simplified limitation of benefits (LOB) rule was proposed in 2015 for states wishing to
combine the LOB rule with a principal purpose test (PPT) rule, which is one of the options on
the OECD abuse menu; this suggestion emerged too late for the development of
Commentary on the new version;
Work on the application of the LOB rule to a wide variety of investment funds was more
complex than anticipated; and
The variety of constitutional and institutional issues, domestic anti-abuse rules and
administrative capacity complicates the choice of rules for many states.
These complexities mean that there is still much substantive work to do in this area (though with a
short deadline of the first quarter of 2016) and an even greater degree of optionality over details than
announced in 2014. Nonetheless the 2014 new minimum standard remains in place and treaty abuse
rules are slated for inclusion in the multilateral treaty. The standard is inclusion of a preamble to the
effect that treaties are not meant to produce unintended double non-taxation and one of:
31 OECD/G20 Base Erosion and Profit Shifting Project, Preventing the Granting of Treaty Benefits in Inappropriate
circumstances ACTION 6: 2015 Final Report (2015).
Page 18
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 18
Detailed LOB treaty rule plus anti-conduit rule in treaty or domestic law;
PPT treaty rule; or
Simplified LOB and PPT treaty rules.
The 2014 standard has been watered down in some respects:
States are only required to adopt it if requested by other states to do so during negotiations (if
not parties to the multilateral treaty);
It applies to new and existing treaties but there is no fixed timeframe to include it in existing
treaties; and
States are not obliged to apply the rule when it appears in their treaties if they have no
objection to treaty shopping as the state of source of income.
In terms of the details, the LOB material has become so complex that it may well have the effect of
scaring countries off this option, except of course the US which seems to revel in the detail (the new
2016 US Model version is considerably longer than the 2006 version). There are a number reasons
for the complexity. The existence of two versions of the LOB rule has made its presentation
unreadable. The several parts of the two rules are scattered in the Commentary so that if readers
wish to see the complete version of either rule in one place they have to cut and paste the document.
The material on the application of the LOB to just widely held regulated collective investment vehicles
runs to four pages and contains several variant provisions designed to cover off the existing treaty
variants in the OECD Commentary on Article 1 for treating CIVs as treaty entitled residents;32
while
the material for other funds including pension funds is only in embryo at the moment. And the fact that
the LOB has to be combined with the PPT or an anti-conduit rule then makes the Commentary on the
PPT confusing as it tries to cover off both possibilities, including apparently duplicated examples.
Drafting the options for the anti-abuse rule in the multilateral treaty is going to be a challenge. It may,
however, distract states from pursuing some of the detailed anti-abuse rules also proposed (but not
part of the minimum standard), such as period testing for benefits under some treaty rules (dividends
and capital gains), a PE abuse rule and adoption of some of the 2015 proposed US anti-abuse rules
which made it into the final Action 6 report – but as suggestions for future consideration! Some of the
more familiar of these specific provisions are likely to receive more attention, such as the proposal for
the change to the dual resident tie-breaker for companies referred to in relation to in the Report on
Action 2 and the saving rule which also supports the Action 2 proposals (see above).
4.2 Application of LOB and PPT Tests
This section considers briefly three scenarios and how the OECD full LOB (that is, the US style
version) and the PPT may apply to them.
Scenario 1: Bad Inc is a former listed company resident in Utopia which is now insolvent and
in liquidation, and was formerly an investment bank; the liquidator has sold its business
assets in Utopia and subsequently sold the business assets of Badsub Ltd, a 100%
subsidiary of Bad resident in Australis which is also in liquidation; after the sale of its assets
32 OECD, Model Tax Convention on Income and on Capital (2014 Condensed Version) pp 50-60.
Page 19
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 19
Badsub’s funds are paid out to Bad by way of a liquidation distribution which is treated partly
as a dividend and partly as a capital gain in Australis.
It is difficult to see any abuse here but equally it is difficult to satisfy the LOB. Former listing is not
enough for the listing test for a qualified person. The order of sale (parent first, subsidiary second)
probably means the activity exception (which unlike the qualified person tests covering all items of
income applies on an item of income basis) is not satisfied and even if it is, whether it will apply to
these items of income. The competent authority discretion has almost invariably not been exercised
by the US and administrative law challenges to the US instransigence on the discretion are not having
great success.33
It seems unlikely that the PPT would be applied in this case given the lack of abuse.
Scenario 2: Resources Inc is listed and resident in Utopia and via a 100% subsidiary Eursub
BV resident in Europa holds a 100% interest in Resourcesub Ltd resident in Australis; under
the treaties between Utopia and Europa, Europa and Australis and between Utopia and
Australis, royalties are not taxable at source but otherwise there would be a 30% withholding
tax in Australis; Resourcesub pays a royalty for the use of a patent to Eursub which recently
acquired the patent from Resources; because of a patent box regime, the royalty is not
taxable in Europa, even though all the research for the patent was done in Utopia.
This scenario raises the problem that even where there is a listed company in the structure, it has to
be resident and listed in the country whose treaty is being applied (Europa Australis here) or a treaty
country subsidiary of such a listed company. The equivalent beneficiary addition in recent US treaties
which is included in the OECD version solves this problem but has been controversial as patent boxes
can create incentives to move IP as in this example. The OECD Action 6 Report p 43 leaves it open to
countries to limit the equivalent beneficiary test to non-deductible dividends – how this will play out in
the multilateral instrument is unclear at this stage, and is likely to depend in part on how successful
Action 5 on patent boxes is. The PPT may well be applicable to this example, given the existence of
the patent box in Europa and the shift of the IP into the patent box regime.
Scenario 3: Car plc is listed and resident in Europa and wishes to lend $10m to its subsidiary
Carsub which is resident in Australis; there is no tax treaty between Europa and Australis but
Europa has a tax treaty with Utopia which also has a tax treaty with Australis containing an
exemption for tax on interest at source if the lender is a bank resident in Utopia; Car deposits
$10m at interest with Bank Inc which is resident and listed in Utopia and then Bank lends
$10m to Carsub; the treaty between Europa and Utopia has a zero withholding tax rate on all
interest.
Despite the fact that this example prima facie seems to involve a back-to-back loan being used to
obtain double treaty benefits, it will not be caught by the LOB in either treaty since Bank satisfies the
listing test in Utopia and Car in Europa. Of course if the financial institution exemption in the Australis
Utopia treaty followed the Australian approach there would be no zero rate under that treaty because
of the back-to-back transaction exception but many treaties with such a financial institutions rule do
not include the back-to-back test. This example shows why the Action 6 minimum standard requires
countries adopting the LOB and rejecting the PPT (like the US) to have an anti-conduit rule either in
their treaties or domestic law (as is the case for the US with its conduit financing regulations). The
33 Starr International (United States District Court for the District of Columbia) 2 Feb 2016, partly reversing earlier decision,
Starr International 18 Sept 2015 https://casetext.com/#!/case/starr-intl-co-v-united-states-1.
Page 20
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 20
PPT test seems likely to apply in this case, which is why countries adopting the PPT are not required
under the minimum standard to have anti-conduit rules.
Based on these scenarios the PPT may seem to be superior to the LOB. The US objection to the PPT
is that it confers too much discretion because of its lack of objective tests (despite the OECD
Commentary picking up the somewhat incongruous Australian Part IVA language on “objective
purpose”.) Australian advisers with experience of Part IVA are likely to sympathise with the US
objections, but the weight of international government opinion seems to be in favour of the PPT.
4.3 PE and special regime abuse rules
The focus on the Action 6 report has mainly related to the LOB and PPT rules because they are so
broadly applicable. It also, however, contains some other more limited rules that financial institutions
in particular need to have regard to. The potential operation of these rules is demonstrated in the
following scenarios, though no detailed analysis is provided because this work is still ongoing (being
dependent on the outcome of the US revisions to its Model which have just been finalised), Action 6
Report pp 75-78, 96-98 on PEs, special tax regimes and future changes to domestic tax law.
Scenario 4: Under the treaty between Australis and Europa interest on loans made by a bank
resident in one state to a borrower resident in the other state are exempt from source tax on
interest; a bank resident in Europa makes a loan to a company resident in Australis; the loan
is effected by the bank’s branch in Asea. Asea does not tax the interest received because of
an offshore banking regime there, and Europa does not tax the interest received because
under the treaty between Europa and Asea the income of the branch in Asea is exempt in
Europa;
This is the kind of scenario that the PE abuse rule will deal with, but there is an exception for banks,
insurance companies and securities dealers carrying on an active business of that kind through the
PE. Thus pure booking operations are likely to be caught but normal branch operations of such
businesses will still be entitled to normal treaty benefits. Note that the Australian OBU regime would
not satisfy the 60% of the normal tax rate test (assuming Asea had a similar regime) and it would be
necessary to rely on the banking exception in the provision.
Scenario 5: Under the treaty between Australis and Europa interest on loans made by a bank
resident in one state to a borrower resident in the other state are exempt from source tax on
interest; a bank resident in Europa makes a loan to a company resident in Australis; the funds
for the loan were raised by the bank from outside Europa and Europa does not tax the
interest received by the bank because of an offshore banking regime there;
This case would be potentially caught by any special regimes measure in the treaty. It will be a matter
for negotiation between states to clarify what regimes are caught and what regimes are not. The
OECD has previously cleared the Australian OBU regime as not harmful under Action 5 (though
without applying the substantial activity test that is now the centre of the new regime for harmful tax
practices).
Scenario 6: Would it make a difference in scenarios 4 and 5 if the offshore banking regimes
were introduced after the treaty between Australis and Europa had been signed?
Page 21
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 21
The US inspired proposal that the OECD is looking at for new tax regimes provides:
If at any time after the signing of this Convention, either Contracting State provides an exemption
from taxation to resident companies for substantially all foreign source income (including interest
and royalties), the provisions of Articles 10 (Dividends), 11 (Interest), 12 (Royalties) and 21 (Other
Income) may cease to have effect pursuant to paragraph 3 of this Article for payments to
companies resident of either Contracting State.
To activate this provision all that is required is giving a notice through diplomatic channels to the
treaty partner which then comes into effect six months after the notice and also triggers an obligation
to negotiate on the issue. This seems to be so even if the treaty includes the provisions mentioned in
relation to scenarios 4 and 5 and they do not apply to the new regime, for example, because the
taxpayer is a bank in the case of the PE abuse rule.
It is to be noted that the US proposals have been more targeted in the final version of the 2016 US
Model as a response to comments.
Page 22
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 22
5 Permanent Establishment
There are conflicting tendencies evident in the OECD work in the permanent establishment (PE) and
transfer pricing areas which are central to the problems that the BEPS project is seeking to solve. At
least in the private sector (and putting aside some Australian peculiarities) there are several
assumptions about international tax rules that are widely shared (subject to any treaty rules expressly
to the contrary):
Each company is to be treated separately in the application of residence and PE rules;
Each location and potential agent is to be treated separately in the application of PE rules,
and agency is to be interpreted strictly in a legal sense of principal and agent;
It is necessary to find a transaction for a company or dealing for a PE which is to be the
subject of a transfer pricing analysis;
Recharacterisation of that transaction or dealing is only possible in very limited
circumstances; and
Absent recharacterisation, transfer pricing of transactions or dealings is to be confined to what
happened and to price.
The OECD work seems to be challenging these approaches and moving away from a formal legal
analysis to a substance economic analysis, which at the extreme could be regarded as saying or at
least inclining towards a position that once the residence or PE threshold is passed, all activities of
the multinational in a country related to the threshold passing presence (or maybe even unrelated to
it) is taxable and has to be priced based on a purely economic assessment of what is done and not
the legal relations involved. But the push-back from the private sector, and indeed several
governments of major OECD countries, especially the US, means that in almost every case we end
up in a halfway house, though not exactly sure where it is.
As indicated earlier, transfer pricing will not be covered in this paper but the PE work will. That work
will be contrasted with the MAAL which is Australia’s answer to part of the problems being dealt with
in the OECD PE work.
5.1 BEPS on Permanent Establishment
The BEPS Action 7 PE Report34
contains important, if limited, changes to the PE definition and is
another Action which will be included in the proposed multilateral treaty with at least part of it
compulsory it seems for signatories (the agency and anti-fragmentation rules). The Treasurer has
indicated that Australia’s treaty practice is already in line with the OECD approach. Although it is true
that Australia’s treaty language has been moving in the last dozen years in a similar direction as the
OECD, it is doubtful that the outcome is exactly the same, and certainly the Australia Germany treaty
moves straight to the BEPS language rather than continuing the recent Australian treaty language. As
with treaty abuse there is still much work to do in elaboration of the changes that have been agreed,
34 OECD/G20 Base Erosion and Profit Shifting Project, Preventing the Artificial Avoidance of Permanent Establishment Status
ACTION 7: 2015 Final Report (2015).
Page 23
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 23
especially further expansion of the Commentary on the changes. We may get some insight into that
expansion if the draft EM to the Germany treaty is released before any further BEPS drafts. In this
section the changes will again be illustrated with some scenarios.
As already had been signalled earlier in 2015, the work on attribution of profits to PEs that is part of
Action 7 has not been done and is left over to 2016 along with a lot of other transfer pricing work.
5.1.1 Time thresholds
The OECD has concluded that splitting time thresholds in treaties through the use of different
companies, which is only relevant to construction sites in the OECD Model but also applies to
substantial equipment and natural resources PEs in Australia’s treaties, can be handled by the PPT
rule discussed above. The Commentary, however, contains a specific provision that can be used for
this purpose if states which use the PPT nonetheless prefer an express rule on the problem. Australia
is in this category. It has been including such a provision in its treaties for some time now and it also
appears in the Australia Germany 2015 treaty. The Commentary notes that such a provision “should”
be used by states which do not accept the PPT rule such as presumably the US. This is illustrated by
the following scenario.
Scenario 7: Projects plc resident in Europa has recently won a tender to carry out a 15 month
construction project in Australis for Manufacturer Ltd. With the consent of Manufacturer the
project has been broken into three periods of 5 months each. Projectsub1, Projectsub2 and
Projectsub3, which are subsidiaries of Projects and resident in Europa contract with
Manufacturer to carry out one 5 month construction contract each.
5.1.2 Preparatory and Auxiliary Activities
While the amendment to make the exceptions to the PE definition only apply if they are all
“preparatory or auxiliary” is contained in the Action 7 Report, that change is no longer compulsory for
states to follow. This change is designed for the following kind of scenario.
Scenario 8: Digital Inc resident in Utopia sells books, music, films and TV programs over the
internet through a website on its servers located in Utopia. Customers in Australis have the
choice of downloading electronic versions of their purchases from Digital’s servers in Utopia
or having a hardcopy book, CD or DVD owned by Digital delivered from a warehouse where
Dsub, a subsidiary of Digital resident in Australis, stores the goods in Australis on behalf of
Digital.
Amazon has already indicated at least in Europe that it will ensure that it has a PE in such cases thus
exposing some of the sale profit for hard copy products to tax on the seller (as opposed to tax on the
small amount of profit arising in the warehousing subsidiary). It should be noted that it is not inevitable
that there will be a PE in the scenario after the BEPS change which deals with the fact that the current
warehousing exception is not qualified by the preparatory/auxiliary language. For the subsidiary’s
premises to constitute a PE of the parent, it is necessary that the warehouse be “at the disposal” of
the parent, and it would seem fairly easy to avoid that outcome. The subsidiary would not constitute a
PE under the BEPS agency rule as it is not involved in the process leading to the sale at all, just the
delivery after the sale. It seems to be conceded that local warehousing is core to the speedy delivery
Page 24
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 24
which is a critical part of the marketing strategy at least for some products. Where products are
warehoused and shipped to the consumer from offshore, there will not be exposure to a PE from the
warehousing activity.
The minimum necessary for states to do in this area is adopt the “anti-fragmentation” rule which
aggregates all activities in a state carried on by an enterprise and closely related parties constituting
“complementary functions that are part of a cohesive business operation” when determining whether
activities in a state are preparatory or auxiliary. The OECD Report adopts the wider version of this
provision out of the two possibilities originally floated which does not require that there otherwise be
any PE in the state. Scenario 9 demonstrates the kind of situation at which it is directed.
Scenario 9: Clothes SA resident in Europa sells fashion garments over the internet from its
servers in Europa to customers in Australis. For this purpose it has several separate offices in
Australis, each carrying out a separate function. One office organizes advertising of the
garments in various forms of local media, another office warehouses and delivers the most
popular lines of garments, and another office deals with aftersales service.
Here it could be the case that the aftersales service office constitutes a PE already, Action 7 Report p
31, and the warehouse may now constitute a PE as indicated in the previous scenario but whether or
not that is the case, the operation of the preparatory/auxiliary exception is now determined taking
account of all the activities in the state through fixed places of the enterprises and its closely related
enterprises. One of the fixed places will still need to be involved in carrying on the mainstream
business of the enterprise in order to qualify as a PE under Article 5(1) but that is the case here,
compare a warehouse that simply stores records of an enterprise as part of a back-office function.35
To the extent that there are closely related enterprises with places of business, each will have its own
PE and be taxed separately. Further what profits are attracted to the PE(s) in such cases will be a
matter for debate depending on how important the particular functions they perform are in the overall
sales activity of the enterprise (and perhaps how much is being taxed to any local subsidiaries
performing such functions). To the extent that the local operations are not essential to the sales, it
may be possible to relocate them offshore as a way of removing the PE or reducing the profit taxable
to any PE.
5.1.3 Agency PEs
There are some changes in the Final Report from the most recent (May 2015) OECD discussion
paper in this area. First, the wording of the revised agency PE rule has been settled as a person who
“habitually concludes contracts, or habitually plays the principal role leading to the conclusion of
contracts that are routinely concluded without material modification by the enterprise”, leading to a
contract in the name of the foreign enterprise or provision of goods or services by that enterprise
(even if it is not a party to the contract – a provision designed to catch commissionnaire structures). It
is confirmed that low risk distributors are not caught by the new provision – that issue is left to the
transfer pricing changes arising from BEPS. Whether this will prove a large loophole in the BEPS
outcomes remains to be seen. Apart from the commissionnaire situation, this change is the one that
will potentially catch Google style operations as in the following scenario.
35 Compare Unisys [2002] NSWSC 1115.
Page 25
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 25
Scenario 10: Searchengine Inc resident in Utopia has a de facto monopoly over internet
searching by people in Australis. It advertises on the search engine website and for this
purpose, its subsidiary in Australis canvasses potential local advertisers, provides them with
recommendations for placement and timing of advertisements and with information on pricing.
If a person wishes to place an advertisement on the website it is advised by the staff of the
subsidiary to go online to Searchengine’s client website and contract for the placement of the
advertising.
The question for Google-type structures is whether the subsidiary through its staff “habitually plays the
principal role leading to the conclusion of contracts that are routinely concluded without material
modification by the enterprise.” It is possible to imagine a reallocation of functions that may assist in a
negative answer to this question though for companies it is now as much a reputational question as a
technical one. It is interesting to compare this with an insurance scenario.
Scenario 11: Life Ltd resident in Asea has a network of 100 insurance agents in Australis who
canvass for customers and obtain applications for life insurance. The agents are not
employees of Life but the only work they do is for Life. The applications are filled in online by
customers with the assistance of the agents. The applications are risk assessed in the head
office of Life in Asea and accepted or rejected depending on the outcome of the risk
assessment, with 5% of applications being rejected in this way (in line with industry norms for
rejection of application in the risk assessment process).
It is difficult to say in such cases that contracts are routinely concluded without material modification
by the enterprise, and the case law in Canada is clear that no PE arises under current law in such
cases.36
There are no changes to deal with insurance enterprises as originally discussed by the
OECD. These are regarded as adequately covered by the changes to the agency PE rule. Australia
continues in the German treaty to exclude insurance business from the operation of the treaty, leaving
the matter to domestic law, though it is noticeable that life insurance is excluded from the exception
(as is also the case in the Switzerland 2013 treaty, perhaps indicating a change in Australian policy in
this regard).
The other change is to prevent the independent agent exception applying if the agent and principal
are closely related and the agent acts exclusively or almost exclusively for closely related enterprises.
This is the third example in the BEPS Action 7 redraft where there is an incursion into the general
position under treaties that the tax position of a company is determined on a stand-alone basis and
not by reference to activities of related companies. In each case, however, the change is peripheral to
the operation of the PE definition (time thresholds, preparatory/auxiliary exception, independent agent
exception). From a policy perspective the issue is why is the same approach not taken in the basic PE
inclusive definitions. It may be that the anti-fragmentation provision will often have a similar effect and
in that sense it may turn out to be the most important change to the definition.
36 American Income Life [2008] TCC 306, 11(1) ITLR 52, Knights of Columbus 2008 TCC 307, 10(5) ITLR 827.
Page 26
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 26
5.2 At Sixes and Sevens: The Australian MAAL
5.2.1 Off on a frolic?
Australia’s multinational anti-avoidance law (commonly called the MAAL and/or the Google tax) has
received criticism not only from the private sector but also from foreign government officials for
jumping the gun or going on a private frolic, and is even the subject of a recorded spoof from a
Canadian singing in an appalling Australian accent based on Waltzing MAALtilda:
Once a fiscal swagman found a tax-free billabong.
Under the shade of a dodgy PE.
And he sang as he saved and waited till his scheme was foiled:
"I found a-waltzing Matilda, PE!" etc
Australia of course is not alone and much of the criticism is over the top. The UK Diverted Profits Tax
(DPT) preceded us, and now the whole of Europe is getting into the act with a 37 page proposal for a
Council Directive released by the European Commission at the end of January 2016 summarised as
follows:37
The schemes targeted by this Directive involve situations where taxpayers act against the
actual purpose of the law, taking advantage of disparities between national tax systems, to
reduce their tax bill. Taxpayers may benefit from low tax rates or double deductions or ensure
that their income remains untaxed by making it deductible in one jurisdiction whilst this is not
included in the tax base across the border either. The outcome of such situations distorts
business decisions in the internal market and unless it is effectively tackled, could create an
environment of unfair tax competition. Having the aim of combating tax avoidance practices
which directly affect the functioning of the internal market, this Directive lays down anti- tax
avoidance rules in six specific fields: deductibility of interest; exit taxation; a switch-over
clause; a general anti-abuse rule (GAAR); controlled foreign company (CFC) rules; and a
framework to tackle hybrid mismatches.
Australia is being relatively modest! At first glance the MAAL appears narrower than the DPT, which
might be called the Google and Starbucks tax and covers base stripping through royalties as well as
avoided PEs, but the MAAL does deal with royalty payments indirectly to the extent the effective
deeming of a PE activates the royalty withholding tax for payments by the foreign enterprise based on
Australian sales, see the Explanatory Memorandum to the Tax Laws Amendment (Combating
Multinational Tax Avoidance) Bill 2015, Example 3.9. Indeed it would not surprise if more tax is
collected on the withholding front than under the corporate tax though restructuring may attempt to
deal with that exposure.
There is a trend apparent here of tackling BEPS substantive Actions through GAARs rather than
directly. Within BEPS it is only Action 6 which relies on the GAAR approach, and then mainly in the
PPT, not so much the LOB. Otherwise the BEPS Actions propose detailed rule changes which are not
37 European Commission, Proposal for a COUNCIL DIRECTIVE laying down rules against tax avoidance practices that directly
affect the functioning of the internal market, Brussels, 28.1.2016, COM(2016) 26 final, 2016/0011 (CNS) p 3. It should be noted
that the Commission often makes proposals that go nowhere or take many years to come to fruition.
Page 27
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 27
dependent on purpose. To the extent that the issues being tackled in these measures relate to tax
treaties, the obvious reason is to effectively amend treaties through domestic law in a way which
seeks to attract OECD Commentary expressing the view that GAARs are not overridden by treaties.
(Within Europe there are no doubt a host of EU law reasons as well). That Commentary only dates
back to 2003 and is proposed to be buttressed by BEPS so in some countries there will be debates to
what extent this approach is effective. In Australian domestic tax law the overriding effect of Part IVA
is clear.
In terms of the BEPS Action Plan, the MAAL is a combination of Action 6 on treaty abuse and Action
7 on PEs discussed above. One obvious question is why have we mixed up Action 6 and Action 7 in
this way? One possibly attractive feature of the MAAL approach for a government is that it is a one
way street – it deals with avoidance of PEs on the inbound side (thereby potentially increasing
Australian taxation of foreign multinationals) but does not create foreign PEs of Australian
multinationals potentially increasing foreign tax at the expense of Australian tax. Of course if the other
country to a treaty introduces a MAAL then symmetry will be restored. When a treaty is renegotiated,
or (depending on its structure) the multilateral instrument is signed, symmetry is more or less
automatic.
Once proposed substantive treaty changes are implemented as in the case of the Australia Germany
2015 treaty, the scope for the operation of the MAAL given Part IVA’s last resort status in domestic
law is not so clear. It is interesting to consider the operation of the MAAL in the case of the scenarios
outlined in relation to the BEPS PE work above. So far as time thresholds, the preparatory/auxiliary
exception and the independent agent’s exception is concerned, it may be that the MAAL will then add
little. The real question is in relation to the agency PE. If, for example a multinational restructures its
Australian operations so that it falls outside the new treaty agency PE rule, can the MAAL still be
activated? Does the MAAL apply to insurance operations of the kind outlined above?
Again The MAAL’s less obvious application to withholding tax may be an issue even after treaties are
amended. It is the domestic PE definition that applies to the limb of the interest and royalty
withholding tax for payments made by an Australian PE of a non-resident (Income Tax Assessment
Act 1936 s 128B(2)(b)(ii), 128B(2B)(b)(ii), compare the treaty definition that applies to the exclusion
from withholding tax on royalties for royalties received by an Australian PE, International Tax
Agreements Act s 17A(4). To the extent that the treaty PE definition is wider than the domestic
definition as a result of the BEPS PE changes interest or royalties may be covered by the interest or
royalties article of a treaty but not be caught by Australian withholding tax provisions (though perhaps
leading to tax by assessment of the non-resident).
5.2.2 Operation of MAAL
The MAAL operates in relation to a foreign entity which is a significant global entity and obtains a tax
benefit involving the requisite principal purpose (borrowed from the PPT) if:
(i) a foreign entity makes a supply to an Australian customer of the foreign entity; and
(ii) activities are undertaken in Australia directly in connection with the supply; and
Page 28
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 28
(iii) some or all of those activities are undertaken by an Australian entity who, or are
undertaken at or through an Australian permanent establishment of an entity who, is an
associate of or is commercially dependent on the foreign entity; and
(iv) the foreign entity derives ordinary income, or statutory income, from the supply; and
(v) some or all of that income is not attributable to an Australian permanent establishment of
the foreign entity.
It is interesting briefly to note changes in the MAAL from the Exposure Draft, some of which broaden
its operation, some of which narrow its operation and one at least of which has an unclear effect. In
relation to a significant global entity, the threshold is now expressed in terms of “income” rather than
“revenue”. While an accounting meaning seems intended the reason for the change is not explained
and the impact of the change is unclear. AASB 118 p 7 under Objectives summarises the concepts in
this way:
Income is defined in the Framework for the Preparation and Presentation of Financial
Statements as increases in economic benefits during the reporting period in the form of
inflows or enhancements of assets or decreases of liabilities that result in increases in equity,
other than those relating to contributions from equity participants. Income encompasses both
revenue and gains. Revenue is income that arises in the course of ordinary activities of an
entity and is referred to by a variety of different names including sales, fees, interest,
dividends and royalties.
In terms of narrowing the scope of the MAAL the following can be noticed:
Supply still has its GST meaning but now excludes supplies of debt and equity interests which
narrows the concept and protects financial transactions to that extent from the operation of
the MAAL.
Activities in Australia must now be “directly” in connection with the supply.
There is now no reference to other domestic tax avoidance (such as stamp duty)
In terms of broadening the MAAL, the following are relevant:
“Australian customer” is used instead of “Australian resident” to describe the recipient of the
supply and defined to include anyone “in Australia” as well as an “Australian entity” in the
CFC sense.
The exclusion to identify non-group customers is now “not a member of the global group”
versus “not an associate”.
There is now no “design” test in the MAAL.
Similarly for no low or no tax test.
Deferral of foreign tax is now caught.
Various comments flow particularly from the Explanatory Memorandum. Firstly, Apple-type structures
with domestic subsidiary distributors are not covered, see Example 3.1, just as the expanded PE
definition in Action 6 does not include stripped risk distributors. The language of “directly connected”
and “in Australia” has a GST connotation flowing from the rules on GST-free exports but this meaning
is not picked up. Similarly “commercially dependent” seems likely in the context of agency PEs to take
a tax treaty meaning but the EM says it has its ordinary meaning (whatever that is).
Page 29
Richard Vann Session C: Financial Services in a Post-BEPS World
© Richard Vann 2016 29
There are some interesting wrinkles concerning the interaction of activities and PEs. First, activities in
Australia covered by the MAAL do not include work by fly-in-fly-out personnel if there no PE in
Australia or resident to which that activity is connected. Conversely, a seller can have an actual PE in
Australia but the MAAL can still apply to the extent to which there is income not attributable to the PE
related to activity in Australia. The additional factors in s 177DA(2) to determine purpose which apply
to the MAAL and not to Part IVA effectively reintroduce as factors tests in the ED that were removed
from the Bill. Although there is no design test, contrivance in the allocation or division of activities will
be a factor pointing towards application of the MAAL. Similarly the reference to the result under
foreign tax law is likely to reintroduce the no or low tax test.
For the application of the MAAL in terms of a tax benefit, the “might reasonably be expected” branch
of the concept will be the likely approach in most cases. The EM para 3.92 says:
In [the] typical case, the notional permanent establishment could include all of the activities of
the Australian-based entity that is described in subparagraph 177DA(1)(a)(iii) as well as the
functions, assets and risks of the foreign entity associated with formally concluding the
contracts. Another alternative may be that the notional permanent establishment includes all
of the activities that are undertaken by the foreign entity, the Australian-based entity or
another entity in connection with the supply.
We have recently learned from the Assistant Treasurer that there are potentially 380 targets of the
MAAL – the number has been growing.38
One suspects that many of those targets will seek to
negotiate an approach with the ATO for the future that involves either setting up an actual PE much
as Amazon has indicated it is doing in Europe and paying tax in the normal way for a PE, or
alternatively perhaps seeking to negotiate on the transfer pricing for the Australian subsidiary which is
creating the concern so that the ATO is happy to leave the matter at that. While the ATO and
government may be keen to test the MAAL in court, it may be that taxpayers can imagine nothing
worse.
38 Assistant Treasurer, ‘The importance of free enterprise, fairness and planning for the future’ – Address to the National Press
Club Canberra, 3 February 2016, “the tax office has already identified 80 taxpayers as having arrangements in the general
scope of the law and a further 300 taxpayers are being profiled.”