Addendum (8 March 2024): The Full Federal Court has handed down its judgment on the appeal in this matter and found for the taxpayer.
30 September 2022
In Brief
On 16 September 2022, a single judge of the Federal Court (O'Callaghan J) handed down his judgement in Minerva Financial Group Pty Ltd v Commissioner of Taxation [2022] FCA 1092. The case related to the ATO's challenge under the general anti-avoidance provisions in Part IVA of the Income Tax Assessment Act 1936 to the Liberty group's structure.
Whilst the ATO was largely successful with the judge finding that Part IVA applied to two of three identified schemes, it was the unusual features of the arrangement (including a discretion to distribute income to the corporate side of the structure), rather than the use of a stapled structure in general, that attracted the application of Part IVA. It is as yet unknown if the taxpayer or the ATO will seek to appeal the decision.
In Detail
Although the judgment runs to 161 pages in the Federal Court’s version, the facts on which the case was ultimately decided can be summarised as follows. The Liberty group had a series of securitisation trusts that derived income from loan receivables. In 2007-2008, the group was restructured into a corporate/trust structure, involving a corporate side and a trust side - with the ultimate owners of the group directly owning shares in the corporate side and units in the trust side. Two entities on the corporate side also held 'special units' in the trust side of the group. A number of the securitisation trusts were held under the trust side of the group – the intention being that new securitisation trusts would be on the trust side while some of the existing securitisation trusts remained on the corporate side. In other words, over time the interest flowing from the securitisation trusts reduced on the corporate side and increased on the trust side. The ATO contended that Part IVA of the Income Tax Assessment Act 1936 applied to the structure as the existence of the trusts (and various other facts) allowed the underlying net 'interest' income to flow through to the non-resident unitholders at a 10% tax rate, whereas a 30% rate would have applied if the securitisation trusts had continued to be held within the corporate group.
Prior to the ultimate IPO, the shares in the corporate side and the units in the trust were not stapled (the group ultimately stapled prior to the IPO in 2020). Although not stapled, the position of the ATO and much of the analysis is based upon there being a ‘stapled’ structure. For ease of reference, we have referred to a stapled structure in the remainder of our analysis.
The ATO put forward three schemes under Part IVA - as the second and third schemes are similar, we have simply referred to them as the second scheme. The first scheme related to the formation and existence of the stapled structure. The second scheme related to the existence of the 'special units' and a series of loan arrangements that resulted in the corporate side gaining access to much of the cash for which present entitlements had been created for the non-resident unitholders and largely satisfied by various intragroup transactions.
The taxpayer won on the first scheme with the judge holding that there were commercial factors driving the existence of the stapled structure including, in particular, the ability for investors to derive higher gross cash returns (rather than a franked dividend) and the advice that the group had consistently received that a stapled structure would be a more effective and preferred structure to take to market as part of an IPO of the group. While the stapled structure was not listed until 2020 (on the third attempt), the consistent advice from the various investment banks assisting was that it was preferable to go to market with a stapled structure.
On the second scheme, the judge held in favour of the ATO. In this regard, although the judge found that the loan arrangements and the requirement for the corporate side of the group to access the cash was legitimate and commercial, he nevertheless held that Part IVA applied. His decision was based upon the taxpayer being unable to demonstrate commercial reasons (i.e., non-tax reasons) why the trustee decided to distribute virtually all of the income derived by the trust to the non-resident unitholders (rather than to all the unitholders, including the corporate side pursuant to the 'special units'). Under the stapled structure the income flowed by default to the non-resident owners but with a discretion to make distributions on the special units.
Both the taxpayer and the ATO are unlikely to be happy with the judgement. The taxpayer will be unhappy because it lost the case based on an application of Part IVA to the second scheme. The ATO is also likely to have misgivings about the judgment as it was almost certainly hoping for a Part IVA 'win' on stapled structures more generally - indeed, although the ATO was successful in applying Part IVA to the second scheme such cross staple units (i.e., the special units) are very unusual so the case is unlikely to assist the ATO in attacking stapled structures more generally. As such, it is likely that one or both parties will look to appeal the decision.
Some observations
The problems of complex Part IVA cases
The judge correctly considered the three schemes put forward separately but the evidence and the parties’ written and oral submissions tended to treat everything in globo making it necessary for the judge to untangle for each scheme the relevant findings on the facts and expert evidence and the relevant arguments to be considered. He also took the cautious course, often found in Part IVA judgments, of considering each factor in section 177D(2) separately for each scheme. As a result, the judge’s views on the facts, expert evidence and sorting out arguments relevant to each scheme take up a large proportion of the judgment and the reasoning on dominant purpose only 12 pages at the end under 24 headings (the ATO’s 3 schemes with 8 factors each) replete with cross references to the earlier material. Most space is given to the first scheme. On each scheme the judge finds most factors neutral and only one or two pointing in either direction.
The impression is that the critical factor in relation to the first scheme was the consistent intention to list with a stapled structure on commercial grounds of obtaining the highest value for the business both on listing and subsequently. On the second (and third) scheme the impression is that the change wrought by the stapled structure and the decisions in the exercise of discretion by the relevant trustee on the trust side to distribute only small amounts if any on the ‘special units’ to the corporate side was driven by the lower interest withholding tax rate compared to the corporate tax rate. We comment on each further below.
The first scheme - the stapled structure
For the first scheme (i.e., the stapled structure) the judge placed much greater emphasis on the evidence provided by the taxpayer (in particular, its expert witness, a former Deutsche Bank investment banker) than that provided by the ATO. The ATO's expert witness got off to a bad start by having to admit that he had never been involved in or worked on an IPO – this admission leading the judge to quickly discount his evidence related to the benefits (or not) of having a stapled structure in order to implement a successful IPO. The judge in particular accepted that some classes of investors would favour a higher cash return than a lower cash return plus franking credits. The taxpayer’s expert dealt comprehensively with the often contested issue of the value of franking credits and the importance of cash flows for a variety of reasons to many of the institutional investors which make up the great bulk of investment in the ASX.
What is disappointing (but not surprising) is that there is no discussion about tax policy settings either in the hearings or judgment. The ATO is clear that policy has nothing to do with the application of Part IVA, only complexity and contrivance – this is most clearly evident in the Consolidation Reference Manual where the ATO does consider policy issues while making clear they are not relevant to the Part IVA analysis. The absence of policy is very conspicuous here as Australia has recently had a lengthy policy debate over the extent to which stapled structures facilitated tax avoidance and 65 pages of legislation in 2019 to deal with the problems. As a result, the tax system has been set up to allow certain types of passive income (e.g., rent and interest income) to flow through trusts in a way that allows non-residents to derive that income at a lower rate of tax than if such income was held via an Australian corporate group. This is part of the bedrock of stapled groups generally. While the Liberty case concerned income years pre-dating the legislation, the ATO objection decision which generated the appeal postdated it. While, as noted above, the facts of the Liberty staple were unusual, the ATO’s first scheme was a direct attack on the stapled structure, rather than its unusual features. It will be a significant concern if the ATO in future seeks to apply Part IVA to stapled structures as a matter of general principle.
The second scheme - including the 'special units'
As noted above, the conclusions of the judge on the second scheme are unlikely to impact stapled groups more generally. What is interesting about the second scheme is that the 'scheme' that the ATO put forward was not simply the decision by the trustee to distribute the income to the non-resident unitholders (rather than distributing all or a significant part of the income to the corporate group) but rather also the structural arrangements that allowed the trustee to distribute income to the non-resident unitholders and the arrangements that were in place between the group members and the non-resident unitholders (through various loan arrangements) such that the cash that was available to be distributed to the unitholders was moved across to the corporate group – as the corporate group needed this cash in order to continue to run its activities. The precise mechanics of these loan / cash arrangements are not entirely clear from the judgement - however, what does appear to be apparent is that (i) distributions of the income were validly made to the unitholders (and 10% withholding tax paid), (ii) the cash was moved to the corporate group (it appears that no cash was received by the unitholders) and (iii) there were various loan arrangements in place (flowing in various directions).
The judge did not find these other factors compelling – he had no difficulty in finding, having regard to the facts, that the loan arrangements served a commercial purpose. Ultimately, he decided the second scheme having regard to the decision to distribute most of the income of the trusts (approximately 98% of the income was distributed to the non-residents). On this he stated (at [565]):
I agree with the Commissioner's submission that, viewed objectively, the exercise of the choice in each of the relevant years (the manner in which the second scheme was carried out) was driven by the tax benefit of directing that income away from [the corporate side] ... I agree with the Commissioner's submission that, objectively, the manner in which the second scheme was entered into is indicative of a dominant purpose of obtaining that tax benefit.
This conclusion may have wide ranging implications as it is not clear how the analysis and reasoning should apply to other trusts, including discretionary trusts. On one view of the world, there is a potential Part IVA issue with any decision that a trustee makes to distribute income to A, rather than B, based on tax attributes. The position is clearly much more nuanced than this as a distribution to A, rather than B, has commercial consequences as well - i.e., the money or property actually flows to A rather than B. Accordingly, the decision may be interpreted narrowly as applying to a situation where the funds are effectively kept in-house (because the ultimate owners of both parts of the structure are the same entities). Furthermore, the discretion was narrowly confined here in the sense that there were only two possibilities (a 30% taxpayer or a 10% taxpayer) – most discretionary trust situations do not present this stark choice. Hence, it may be argued that a normal discretionary trust is not exposed. Although such an interpretation / argument can be put forward (and probably will be put forward by the ATO) the position is certainly far from clear. Indeed, is the position here really that different from a family trust where there are two beneficiaries who are spouses and the decision is made to distribute income to one spouse rather than another because of their tax attributes?
At a more general level, the fact that the result flowed simply because there were non-residents with a different tax profile, was rejected in a recent UK tribunal decision in Burlington Loan [2022] UKFTT 00290 (TC) in the context of the main purpose test in the UK-Ireland tax treaty. This test was the predecessor of the principal purpose test (PPT) in the Multilateral Instrument and now in the Australian Multinational Anti-Avoidance Law and Diverted Profits Tax. In that case, the fact that a mid-east tax exempt person without treaty reduction for interest from the UK 20% default rate sold a debt for 92% of its face value to an Irish company which benefited from a zero tax rate in the UK-Ireland tax treaty was not enough to deny treaty benefits as there are a range of tax favoured investors in the market and the reality is that the tax status of the buyer affects the amount they are willing to pay. Otherwise, the tribunal found there would be considerable disruption of international capital markets. Given that the PPT is a lower threshold than the dominant purpose test applicable to the general operation of Part IVA under section 177D, one would expect the UK tribunal to reach a similar result for widely held stapled structures for a dominant purpose test.
The expert evidence for the taxpayer in the Liberty case accepted by the judge made a similar point. The only difference is that in the years in question the Liberty stapled group was closely held, though its owners were consistently seeking to turn it into a widely held group through an IPO. It is not clear why this should make a difference.
The Takeaway
This case raises a number of interesting questions regarding the use of stapled structures and discretionary trusts, and is likely to go on appeal. Following a lengthy debate over the use of stapled structures and significant tax law reforms in 2019, it is clear that the ATO is not yet through with stapled structures.
Related content
Contact us
Hayden Scott
Partner, Tax Controversy & Dispute Resolution Leader, PwC Australia
Tel: +61 488 221 199