Appeal in PepsiCo Royalty Case

by Graeme Cooper, Paynter Hugh, Ryan Leslie, Geraldine Chan

A majority of the Full Federal Court has upheld an appeal by PepsiCo against the first instance decision which had held it liable to royalty withholding tax and (in the alternative) found it had triggered the diverted profits tax provisions. Given the split in the Full Federal Court, and the disagreement with the judge at first instance, an application for special leave to appeal by the ATO to the High Court seems highly likely, especially on the diverted profits tax issues. Whether the High Court will hear an appeal is another question, but a 2-2 split between Federal Court judges, involving untested law, increases the chances.

Is there an unpriced royalty?

The royalty withholding tax aspect of the dispute revolved around two features of the agreements between PepsiCo group entities and Schweppes Australia. The first was that the PepsiCo group sold concentrate to Schweppes and granted it licences to use various trademarks, but the contracts allocated the price entirely to the goods, and none to the licence.

At first instance, Moshinsky J held, ‘it is apparent from the terms of the [agreements] (considered in their business and commercial context) that the payments were, to some extent, consideration for the use of, or the right to use, the relevant trademarks and other intellectual property.’ He came to this conclusion because, ‘without a licence of the trademarks and other intellectual property, [Schweppes] would not be able to package and sell the beverages under the Pepsi, Mountain Dew and Gatorade brands … [Schweppes] would not have agreed to make the payments without the licence of the trademarks and other intellectual property.’

On appeal, Colvin J, who was in the minority on this issue, took a similar approach: ‘regard to the whole of the terms of the [agreements] reveals that the prices to be paid … are not simply consideration for the concentrate. Rather they are also … for the right to use the valuable brands ... It follows in my view that the amounts provided for by the [agreements] as the prices for units of concentrate were partly amounts in consideration for the use of the trademarks which [Schweppes] was licensed to use.’

Yet the majority took a different view. The majority judges, Perram and Jackman JJ, based their opinion on a completely different analysis of what the bottling agreements involved: ‘[the licence rights] were intertwined with [Schweppes’] obligations to distribute the beverages in Australia under the Pepsi, Mountain Dew and Gatorade trade marks using PepsiCo … can and bottle designs and to do so under the strict quality control of PepsiCo … and in accordance with its detailed stipulations about how the trade marks were to be used.’ In other words, what might look like a right conferred on Schweppes was actually a burden imposed on it, or at least a mixture of rights and obligations. And that obligation was imposed so that, ‘PepsiCo … obtained the benefit of having their goodwill sustained and enhanced …’ ‘To see the trade mark licence as simply a contractual grant of a right solely for the benefit of [Schweppes] is significantly to misunderstand the nature of the rights involved.’

If the licence was better understood as part of an obligation imposed on Schweppes for the benefit of PepsiCo, it is unsurprising that Schweppes would not pay for it. But the analysis turns entirely on this conclusion: that what looks like a right is actually a burden.

For the majority, the result may well turn on the way the parties approached the case. As Perram and Jackman JJ noted, both the taxpayer and the ATO ran the case on the basis that ‘the question of whether payments are consideration for the right to use intellectual property rights, and therefore a ‘royalty’ … is determined by the construction of the relevant agreement.’ That was consistent with earlier case law. The majority judges this focussed on the words appearing in the text of the agreements, rather than some overarching business and commercial context. In that regard, there was no evidence of whether the bundle of trademark rights and obligations assumed by Schweppes were in fact burdensome or not: ‘whether these benefits were ultimately in favour of PepsiCo… or [Schweppes] is unclear and likely in any event to be a close function of market dynamics.’ As there was no contextual evidence led, the majority were prepared to follow the text of the agreements and conclude that the payments were exclusively for the concentrate.

Who derived the income, and to whom was it paid?

The second feature of the agreements revolved around the cash flows. Royalty withholding tax is triggered where a non-resident derives income in the form of a royalty, and a resident pays the royalty to a non-resident (although the royalty will be viewed as paid to the non-resident if the non-resident has directed that the royalty be paid elsewhere).

In this case, the parties to the bottling agreements were Pepsi group entities in the US and Schweppes in Australia. With regard to the supply of concentrate, Pepsi could either supply the concentrate itself, or ‘cause [the concentrate] to be sold by one of its subsidiaries ...’ In the event, Pepsi incorporated an Australian subsidiary and nominated it as the seller of the concentrate to Schweppes. Consequently, Schweppes was paying an Australian-resident subsidiary in the PepsiCo group, not a non-resident entity.

At first instance, Moshinsky J concluded these requirements for triggering royalty withholding tax were satisfied. ‘PepsiCo [was] entitled to receive the payments made by Schweppes … from the fact that PepsiCo [was the party to the agreement] and [Schweppes’] payment obligations … were owed to [Pepsi]. This remained the case even though PepsiCo … nominated [the Australian subsidiary] as the seller of the concentrate. [It] was not, and did not become, a party to the agreements. The nomination of the Australian subsidiary] constituted a direction to Schweppes to pay [the Australian subsidiary] rather than PepsiCo …’

As the majority had concluded there was no royalty, the operation of these requirements did not arise for them. Nevertheless, all three judges in the Full Federal Court disagreed with the analysis at first instance: ‘where a related entity was nominated as the seller the [agreements] were clear that it was that related entity and not PepsiCo … that would be selling the concentrate… And, if there was no sale of concentrate by PepsiCo …, it cannot be the case that [Schweppes] was ever obliged to pay [Pepsi] for something [it was not] not selling… Since there was no antecedent monetary obligation, it is not possible that by paying the [Australian subsidiary, Schweppes] was paying PepsiCo ...’

But it is worth noting, the agreements did not contemplate a similar nomination process with respect to the licence agreements – Pepsi could not nominate someone else presumably because PepsiCo owned the intellectual property and only it could licence Schweppes to use its brand names. The majority pondered this difference and said, ‘it is possible, although we draw no firm conclusions, that the royalty component of the payments might have been seen as belonging to PepsiCo ...’ but given that the majority thought there was no ‘royalty component’, the passage is mere musing.

But the issue is definitely relevant for Colvin J as he concluded there was a supply of trade mark rights. Elsewhere in his judgment he observed, ‘on the conclusion I have reached, it is the mechanism by which the Related Entity could be nominated as the Seller … that means that there is no income in the form of a royalty that may be the subject of a withholding tax liability on the part of PepsiCo …’ His logic seems that PepsiCo simply didn’t derive any income, and consequently could not derive any part of the payment which included the royalty. The Commissioner had also accepted Schweppes was not the agent of PepsiCo and did not hold the money it collected on trust for PepsiCo.

The result is all three judges agreed Pepsi did not owe royalty withholding tax but could not agree why.

Diverted profits tax

The Full Federal Court’s analysis of the possible application of the diverted profits tax focussed on three issues:

1. what was the ‘tax benefit’;

2. was there a plausible counterfactual; and

3. was the ‘purpose’ test satisfied?

Again, the judges disagreed among themselves and with the analysis of Moshinsky J.

Tax benefit

The first issue involves a discrepancy in the drafting of Part IVA. Most parts of the definition of ‘tax benefit’ identify various items which are relevant to the calculation of tax – missing income, over-claimed deductions, net capital losses’ and so on. In the case of royalties, however, the tax benefit is escaping tax: ‘the taxpayer not being liable to pay withholding tax on an amount.’ The conceptual problem was, if PepsiCo did not earn a royalty, it could not obtain a tax benefit in the form of not having to pay royalty withholding tax. The majority were happy to accept the approach of the parties that the tax benefit was ‘the amount of the royalty which did not come into existence … not the amount of royalty withholding tax that would have been due ...’ – in other words, the item which is relevant to the calculation of royalty withholding tax.

Counterfactual

The second issue involved identifying the counterfactual. The events occurred in the 2018 and 2019 income years, and so it was necessary for the Court to identify the counterfactual in light of the 2012 amendments. The case was run on the basis that the counterfactual would be determined by applying the second limb test: the tax outcome that would arise from ‘a reasonable alternative [having] particular regard to the substance of the scheme and any result or consequence for the taxpayer that is or would be achieved by the scheme …’

The Commissioner presented two counterfactuals: either the agreements would have said expressly that the price being paid was for both concentrate and the licences, not just the concentrate, or else the agreements would have identified a separate price for each component.

The majority began by observing,

          … the taxpayer … bears the burden of proving that the assessments are excessive... Proving that the Commissioner’s postulates are unreasonable does not in itself discharge that burden. It remains the burden of the taxpayer to show on all of the evidence that the tax benefit would not reasonably be expected to have been obtained if the schemes had not been entered into … PepsiCo must show that there is no reasonable postulate ... Naturally this will include demonstrating that the Commissioner’s postulates are not reasonable but PepsiCo must also demonstrate on the evidence that there is no other reasonable postulate.

After reading this, the ATO might have been forgiven for thinking they had won the day.

Substance of the scheme. But the majority then went on to hold, ‘there is no postulate which is a reasonable alternative to the scheme.’ They did so because the ATO’s alternatives weren’t the actual arrangement, which is why it is called, “a counterfactual” and ATO had not led any evidence, which the judges have just said, isn’t their responsibility. The majority said:

          the Commissioner contended … the commercial and economic substance of the scheme was that the payments for concentrate included a royalty for the use of the trade marks and other intellectual property … [but] there is nothing in the terms of the scheme from which this may be inferred and there is no evidence which could support the Commissioner’s submission.

and later –

          … neither the scheme advanced by the Commissioner nor any of the evidence provides material from which it may be inferred that the commercial and economic substance of the scheme was that the concentrate price included a royalty for the licence of the intellectual property. In that circumstance, there is no basis upon which we could conclude that the price for concentrate in the scheme included a royalty component …

When it came to identifying ‘the substance of the scheme …’ the majority concluded, ‘the commercial and economic substance of the scheme was that the price agreed for concentrate was for concentrate.’ If that is the ‘commercial and economic substance’ then any arrangement which involves buying a licence would be too dissimilar, almost by definition.

Colvin J, on the other hand, was quite happy to conclude, without evidence, that, once he found there was an unpriced royalty –

          … the [agreements] resulted in a tax benefit because, if the [agreements] had not been entered into, then a reasonable postulate was that the [agreements] would have provided for the royalty to be paid to PepsiCo … as the holder of the rights to the trade mark.

And he continued, ‘as the joint reasons explain, it is the burden of PepsiCo … to prove that there was no reasonable alternative to entering into the scheme and for the reasons I have given they have not done so.’

Result or consequence. The direction to identify a reasonable alternative having regard to ‘any result or consequence for the taxpayer that is or would be achieved by the scheme …’ went in what might be regarded to be an unexpected direction.

The majority, quoting from the EM to the 2012 amendments, said, ‘a postulate will be a reasonable alternative to the scheme if it may reasonably be expected that … the taxpayer would have entered into or carried out the postulate. This is an exercise in prediction.’ On this reading, the drafting effort that went into eradicating the need for the ATO to second-guess what the taxpayer might have done (instead of what it did), was wasted.

And the majority tied their prediction test back to the language of the original s. 177C: ‘an amount [which] might reasonably be expected to …’ The majority said, ‘the approach which ought to be taken to [the new test] is similar in kind to the prediction referred to in Peabody.’ At the time of the 2012 amendments, it was generally accepted that the new test was meant to supplant (or at least supplement) the prediction test, even though that test was not being repealed. It was noted at the time that it was dangerous to assume language was being replaced while the new test was built upon it. Shirking the drafting effort needed to repeal s. 177C may prove to have been a bad call if a different test was in fact intended.

The majority concluded, ‘neither postulate is a reasonable alternative to the scheme …’. Thus there was no tax benefit and on that basis the taxpayer prevailed on DPT.

Purpose

Having found no tax benefit, the majority nevertheless undertook the task of trying to identify whether Pepsi had entered into or carried out the scheme with the requisite purpose, assuming ‘our conclusion that there are no postulates that are reasonable alternatives to the scheme is incorrect.’ The majority largely agreed with the analysis of Moshinsky J, concluding ‘we would have concluded that the requisite purpose … had been established.’ Not surprisingly, Colvin J agreed with that assessment as well.

The result is, the majority held the diverted profits tax was not triggered; Colvin J held it was.

Conclusions

Given the split within the Full Court, and the disagreement with the judge at first instance decision, an application by the ATO for leave to appeal to the High Court seems highly likely.

No doubt such an application would include the royalty withholding tax issue. This is because the DPT issues need to be considered in light of the royalty withholding tax issues – that is, the DPT is an alternative to those issues. Therefore the High Court would be unlikely to hear an appeal on the DPT issues and not the royalty withholding tax issues.

The royalty withholding tax issues does raise some particular issues, namely the extent to which business and commercial context is relevant to the characterisation. It is also worth noting that the case was determined on domestic law rather than on the law as affected by international tax treaties.

The diverted profits tax outcome is probably the greatest irony in the case. All four judges concluded the taxpayer had acted with a purpose of avoiding tax, and the 2012 amendments meant to ensure this conclusion would be the end of the Part IVA journey. The amendments were meant to ensure that a taxpayer, found to have acted with a tax avoidance purpose, could never again escape Part IVA because of disputes about the counterfactual. Yet this is exactly where the majority landed.

Moreover, the majority’s approach to the counterfactual would seem to be unconventional. The 2012 test is meant to find a transaction with produces similar commercial outcomes to the one undertaken. The question which the Act poses is, is collecting (a) a smaller price but with a royalty, a ‘reasonable alternative’ to (b) collecting a larger price, without a royalty? Does option (a) produce the same ‘substance’ as option (b); are the ‘results and consequences’ the same, once you ignore the royalty withholding tax (as the Act directs)? There is plenty of scope to argue about the answer, but one must start with the right question, and that question is not a question that involves any element of prediction. The test is commercial resemblance, not likelihood. And it is for the taxpayer to discharge the onus in this regard.

Revisiting how to find a counterfactual after the 2012 amendments is definitely worth the High Court’s attention. But to do so, it would seem necessary to also look at the royalty withholding tax issues.