What just happened? The fight about the 2013 Part IVA amendments

The Pepsi litigation, currently before the High Court, involves several specific issues about Australia’s royalty withholding tax and diverted profits tax legislation. But alongside those specific disputes is a question with much broader impact: just what did the 2013 amendments to Part IVA achieve?  The case presents the High Court with its first opportunity to resolve significant uncertainty about how these amendments operate – an issue with implications far beyond this individual case.

The prediction problem

The 2013 amendments came as the direct result of ATO dissatisfaction with the Federal Court judgments in 4 key cases on Part IVA: Futuris (2010), AXA (2010), Noza Holdings (2011) and RCI (2011). All the cases exposed a fundamental challenge for the ATO when applying Part IVA: the tax result which the taxpayer’s actual transaction produced was struck down, but what happens now? 

The test in s. 177C ITAA 1936 set out the substitute to be taxed as the amount ‘that … would have been …, or might reasonably be expected to have been [involved] if the scheme had not been entered into or carried out …’ This formula required an assessment of what the particular taxpayer would likely have done absent the scheme.

As the High Court explained in Peabody, the test required, ‘a prediction as to events which would have taken place if the relevant scheme had not been entered into or carried out and the prediction must be sufficiently reliable for it to be regarded as reasonable.’

This formulation posed a fundamental practical problem for the ATO.  If the ATO had to establish a plausible alternative scenario, the substitute had to thread a difficult needle – to involve sufficiently more tax than the actual transaction to justify issuing the amended assessment, but not be so expensive that it was implausible to claim the taxpayer would have undertaken that transaction. 

Futuris is a good example of the difficulty. Besanko J dismissed the ATO’s proposed counterfactual because it generated so much tax that the taxpayer would never have done the deal that way:

I am satisfied that, had the scheme not been entered into or carried out, the [ATO’s] Presumed Counterfactual would not have happened as a matter of reasonable expectation [because] the effect of the Presumed Counterfactual … would be a doubling up of capital gains in respect of essentially the same assets.

The ATO’s prediction also had to identify the right taxpayer. In Futuris the substitute which the Court did consider likely, would have put the gain in the hands of a different taxpayer from the one the ATO had assessed.

The 2013 amendments: Replacing prediction or supplementing it?

Against this background, Parliament sought to address these challenges by amending Part IVA in 2013.  The fight in Pepsi is about what those amendments achieved. Crucially, the 2013 amendments did not touch the text of s. 177C. Instead, the amendments added a new s. 177CB into the legislation which established two alternative bases for identifying a tax benefit:

• The ‘annihilation’ test: taxing the taxpayer on a modified version of the actual transaction by removing from the events or circumstances that actually happened or existed those steps that form part of the scheme; and
• the ‘reasonable alternative’ test: taxing the taxpayer on a transaction which produces similar commercial outcomes to what the taxpayer did (‘… a reasonable alternative to entering into or carrying out the scheme … [having] regard to the substance of the scheme’ and the non-tax results it produced). 

This legislative structure created uncertainty: did these new tests replace the original ‘prediction’ test, or supplement it? If they supplement it, how do they interact?

Given the mischief which the amendments were meant to address, and the Explanatory Memorandum to the 2013 Bill, it seems tolerably clear the two new tests were meant to displace the old prediction test:

[The old test] is viewed as an open-ended inquiry into what, if anything, the taxpayer might reasonably have done if it had not participated in the scheme…  From a policy perspective, the operation of Part IVA as a general anti-avoidance provision would be better served if the inquiry focused on whether or not there were other ways (for example, more convenient, or commercial, or frugal ways) in which the taxpayer might reasonably have achieved the substance and effect (tax implications aside) that it achieved from, or in connection with, the scheme.

However, by grafting the amendments onto the old statutory wording and not explicitly repealing the old test, and saying instead that the test ‘must be based on a postulate …’ with certain characteristics, the amendments created considerable uncertainty: as between the old test and the new tests, to what extent was the old test now defunct?  Or was the old test a gateway which must be met as a step to the two new tests?  Or do the 3 tests accumulate so that all have to be met?  Or are they substitutes: it is sufficient to meet any one?  And the same questions could be asked about how to assemble just the two new tests: are they substitutes, do they accumulate, does the taxpayer elect, and so on? 

The arguments in Pepsi on the interaction between the tests

While Pepsi is not the first time these questions have arisen, in the way that the case developed, it is the case that puts the questions most directly in issue.  The ATO had argued Pepsi should be taxed as if the agreements had expressly provided for the payment of a royalty, rather than assigning all the consideration as payment for the supply of concentrate.  But do the tests allow the ATO to impose tax on this substitute, instead of the transaction actually undertaken?

At first instance, Moshinsky J accepted the ATO’s alternative postulate finding both that, ‘… had the scheme not been entered into or carried out, it might reasonably be expected that the relevant EBA would have provided for the payments to be made by SAPL to be for all of the property provided …’ [and this was] ‘a reasonable alternative, as required by s. 177CB (3).’  His Honour’s reasoning mentions both the old prediction test and the new reasonable alternative test.

In the Full Federal Court, the majority adopted a rather different approach treating the tests as complementary. The majority stated:

The question posed by s 177CB(3) requires the Court to assess whether entry into either of the postulates is a reasonable alternative to entering into and giving effect to the scheme… [The old test] calls for an assessment of what might reasonably be expected to have occurred in the absence of a scheme and uses the same language of ‘might reasonably be expected’ as s 177CB(3)…  This is the approach which ought to be taken to s 177CB(3)… In assessing the reasonableness of the postulate one must therefore ask whether it is reasonable to think that, but for the scheme, the taxpayer would have entered upon and carried into effect the postulate.

This approach is surprising: the judges are saying, what makes something commercially similar is that it is likely.  But likelihood and commercial resemblance are very different ideas. 

On the appeal to the High Court, competing interpretations about how to assemble the tests were put to the High Court by counsel for the taxpayer and the ATO.  Counsel for the ATO put forward the view that the Part IVA world changed in 2013:

Following the introduction of 177CB in 2013 … the inquiry has changed from one of predicting what would or what is most likely to have taken place if the scheme had not been entered into or carried out. The question of whether a particular postulate is sufficiently reliable now arises in the sense of whether it is ‘a reasonable alternative’ to the scheme as described in 177CB(3)…  So, this provides the context to the introduction of 177CB in this respect, and it marks a change from any of the cases on the previous law, which had indicated that the task was to find the one true alternative postulate.

Counsel for Pepsi put a similar position: ‘Our position, your Honours, is that section 177CB(3) overrides section 177C(1)(bc).’  

But a short time later Counsel for Pepsi makes a slightly different argument – that the new test of commercial similarity is actually what the old prediction test always was:

So, your Honour, I think the starting point is 177C(1)(bc), being the provision that has always been in the legislation. So, section 177CB was introduced to modify the way that that would operate. We would say that even prior to the introduction of section 177CB, Peabody tells you that there must be a reasonable alternative postulate, and 177CB(3) is really only confirming that position that has always been the law.

This second argument is not the obvious reading of Peabody. Peabody says the alternative must be reasonably likely as a prediction of events, not that the alternative must be reasonable. 

The substance debate: What is a ‘reasonable alternative’?

Beyond the interaction between the tests, Pepsi also highlights important questions about determining the ‘substance’ of a scheme under s. 177CB(4)(a)(i). The Commissioner argued that the substance of the EBAs was that SAPL received both concentrate and valuable intellectual property rights from the Pepsi in exchange for the payments, but the terms of the agreements allocated all consideration to the concentrate. Pepsi countered that the substance of the arrangement was precisely what the terms reflected – payments for concentrate, with the intellectual property rights exchanged for non-monetary consideration (SAPL's promises to build the brand in Australia, promote the products, and make specific investments).

This debate was central to determining whether the Commissioner's alternative postulates could satisfy the ‘reasonable alternative’ standard in s. 177CB(3). As Pepsi’s counsel put it:

‘… if the substance of the arrangement is that the price agreed for concentrate was for concentrate... that is the substance, and it cannot be otherwise.’

Pepsi also raised the onus question in relation to the operation of Part IVA: can a taxpayer prevail by showing no reasonable alternative exists? A significant portion of the High Court hearing addressed a critical procedural question: how does a taxpayer discharge its onus of proving an assessment excessive when the Commissioner relies on Part IVA?  The Commissioner submitted that a taxpayer must not only demonstrate that the Commissioner's alternative postulate is unreasonable, but also establish a positive alternative and that this alternative would not give rise to a tax benefit.  Pepsi argued it is sufficient to show no reasonable alternative exists. Counsel for Pepsi cited the Full Court's decision in Guardian that the provision has been ‘historically described as a gateway to Part IVA,’ and ‘it secures a sensible result that a taxpayer will not be taxed on a hypothetical state of affairs which is unreasonable.’  This question has significant practical implications. If Pepsi prevails, taxpayers could potentially defeat Part IVA assessments by demonstrating there was no commercially reasonable way to achieve their objectives other than the path they chose.

What next?

There is always the possibility that the High Court will dodge this issue by deciding Pepsi on the withholding tax aspect of the case (and if Pepsi is to lose, given the different amounts at stake on the withholding tax as opposed to the Diverted Profits Tax, one suspects it would prefer to lose on this basis).

But hopefully, the High Court will address the Part IVA issues, even if in obiter, including –

• Tax benefit test interaction: Does the ‘reasonable alternative’ test in s. 177CB(3) replace, supplement, or qualify the original prediction test in s. 177C?
• Onus of proof: Can a taxpayer discharge its onus by showing there is no reasonable alternative to the scheme it implemented?
• Substance determination: How should courts determine the ‘substance’ of a scheme under s. 177CB(4)(a)(i)?
• Commercial similarity: What makes an alternative ‘reasonable’ under s. 177CB(3)? Does reasonableness incorporate both commercial similarity and likelihood?
• Purpose assessment: How does the existence of an established business model affect the purpose analysis under s. 177D?

Taxpayers with existing structures involving cross-border IP arrangements will obviously be watching the outcome closely. But the High Court's approach to the Part IVA issues may have much broader implications for how anti-avoidance provisions are applied to complex commercial arrangements generally, especially in light in of the 2013 amendments.