Court limits ATO’s ability to reconstruct related party transactions
17 September 2019
The Australian Taxation Office (ATO) has been arguing that Australia’s transfer pricing rules include a broad power to reconstruct transactions with related parties. However in the recent Federal Court decision in Glencore Investments Davies J rejected the ATO’s arguments and held that the transfer pricing analysis ‘should be based on the form of the actual transaction entered into between the associated enterprises.’
The issue in Glencore revolved around the pricing of copper concentrate, mined by a Glencore company in Queensland, and sold in an unrefined state to the parent company in Switzerland, which then sold the concentrate to smelters in Japan, Korea, India, China and the Philippines, which turned the concentrate into copper. In the ATO's view, the Australian entities were underpaid almost $241m over three years; the ATO claimed tax, interest and penalties exceeding $92m. After a 10-day trial, the evidence of 5 witnesses, and almost 130 pages of judgment, the Court concluded that the price received was within the acceptable range for tax purposes.
In Chevron, the case largely turned on the question of a hypothetical financing agreement and the extent to which this hypothetical agreement can differ from the actual agreement between the related parties. In Glencore, the issue was the nature of a hypothetical copper concentrate offtake agreement, and the extent that this hypothetical offtake agreement could differ from the actual offtake agreement. Consequently, the contract terms were examined in some detail in the case but the terms to which the ATO objected most strenuously were –
- the pricing formula in the contract lasted for several years without revision,
- the Swiss company had choices about some elements in the pricing formulae which set the prices it would pay, with some of those choices giving the buyer the benefit of hindsight, and
- the Australian company accepted a discount from the sale price, said to reflect some of the cost for the (upstream) smelter of turning the concentrate into copper. In fact, this clause exposed the seller (ie, the mining company) to a portion of the gains (and losses) from movements in the market price of refined copper.
The ATO's argument was that an independent miner would not have agreed to sell under contracts in their current form; it would have sold under contracts with different variants of these clauses:
- the pricing formula would have been renegotiated every year,
- the formula for setting the price would have had fewer options, and would not have allowed choices based on numbers from the past,
- there could be a discount from the price for expected refining costs etc, but the discount should have been calculated using either an annual benchmark for expected refining costs or a spot price or even a combination, but not, as in this contract, a discount which moved depending on the price of refined copper in world markets.
Unless the actual versions of the clauses were excluded, the resulting price would be, 'unrealistic, artificial and contrary to the purpose of the transfer pricing provisions and would result in a commercially unrealistic outcome ... controlled by the taxpayer.'
The taxpayer countered by the arguing the ATO did not have the power to re-write so much of the actual contract – it could not assess Glencore on the basis of, 'a wholly different agreement ... to that which the parties in fact agreed.' And in case that argument did not succeed, Glencore also tendered evidence of contracts, between independent market participants, on foot in the relevant years, containing terms which resembled the actual contract terms.
The judge seems to have accepted both of the taxpayer's arguments and held that:
The Commissioner’s approach impermissibly restructures the actual contact entered into by the parties into a contract of a different character.
Substituting [the ATO's preferred term] for [the actual contract term] would produce a hypothetical transaction based upon different commercial considerations and a different commercial structure to that which was actually entered into. It involves and requires fundamentally rewriting the actual arrangement and engaging in an extensive exercise in commercial judgment which does not accord with, or give effect to, the objective of the arm’s length principle.
Unlike the judges in Chevron, Davies J took seriously the admonition in the OECD Transfer Pricing Guidelines (as they stood at that time) that only 'in exceptional circumstances' could the actual transaction be disregarded and another transaction substituted for it (and the price of that transaction be used for tax purposes).
And with regard to the contracts tendered by the taxpayer as evidence of current market practice, the Court rejected the ATO's argument, 'that those contracts cannot provide a probative basis for the Court to conclude that such kinds of terms might be expected to have been agreed ...' The ATO tried to discount the reliability of these agreements because they lacked other features of the Glencore deal – eg, they didn't involve a buyer taking 100% of a mine's output, they didn't have a 3-year term, etc. But the judge said the ATO was setting the test of comparability too high: no transaction is ever likely to be absolutely identical, and insisting that it be almost identical would mean a taxpayer 'could never succeed' in a transfer pricing case.
Some observations on the decision
There is much one might say about this decision but we will confine our comments to a few key points.
Re-pricing v re-construction. As in Chevron, the ATO’s case was not that the price was wrong, but that the deal was wrong: unrelated sellers would not have agreed to this deal. The ATO has gone to some effort in the last few years to assert this idea, that the ATO can impose tax based on a deal which the taxpayer did not do, but the ATO considers it should have done.
Davies J rejected the ATO’s arguments and held that the hypothetical transaction:
should be based on the form of the actual transaction entered into between the associated enterprises but on the assumption that the parties are independent and dealing at arm’s length.
Although, as discussed below, Davies J referred to the exceptions to this principal in the 1995 OECD Transfer Pricing Guidelines.
Davies J’s construction of Division 13 and Subdivision 815-B suggests that where related party transactions are on terms observed between independent parties it will be difficult for the ATO to apply Division 13 or Subdivision 815-A. However it should also be noted that the offtake agreements in the Glencore case were for a single mine. It may be more challenging for taxpayers to refute the ATO’s arguments where transactions are not observed between related parties.
It should also be noted that Subdivision 815-B (which applies for tax years commencing after 29 June 2013) specifically affirms that the hypothetical transaction must be based on the actual conditions. Transactions can only be ‘reconstructed’ if it can be established that exceptions detailed in section 815-130 apply.
The ATO has asserted in Tax Ruling TR 2014/6 that the section 815-130 reconstruction provisions are similar to the OECD Guidelines exceptions, however the different wording of section 815-130 means this is debatable.
Use (and abuse) of the OECD Transfer Pricing Guidelines. Another complicating factor in this case is the role and impact of the OECD Transfer Pricing Guidelines in the interpretation of Division 13 and Subdivision 815-A. The Guidelines are both an interpretation of article 9 of a treaty and a guide for tax authorities and multinational enterprises in order to reduce transfer pricing disputes. Division 13 does not refer to the Guidelines, but Subdivision 815-A requires that determining whether an entity receives a ‘transfer pricing benefit’ should be done consistently with the 2010 Guidelines to the extent they are relevant.
Davies J considered the 1995 Guidelines were relevant to applying both Division 13 and Subdivision 815-A, and in particular focused on the approach in the Guidelines that:
in other than exceptional circumstances, the actual transaction should not be disregarded or other transactions substituted.
Davies J then referred to the 1995 Guidelines which limit the ability to reconstruct a related party transaction to two narrow exceptions, which in her view did not apply to the facts in this case.
The judges in Chevron did not focus on the Guidelines in the same way as Davies J. However it should be noted that the judges in Chevron were dealing with different facts, and in particular with arguably less persuasive evidence about the commerciality of the related party arrangements.
Use of hindsight. The Glencore subsidiary entering into the offtake agreement chose a particular pricing option from a range of available options. At that time, the commercial outcomes from the different pricing options was unknown. With the benefit of hindsight, the ATO didn’t like the fact that the chosen pricing option resulted in lower Australian profit than if a different pricing option had been chosen.
Davies J held, following the Canadian Cameco case, that the use of hindsight was not appropriate, and it was irrelevant to the transfer pricing enquiry whether different pricing options resulted in different financial outcomes.
Transfer pricing ‘economists.’ One final point to note is the way in which the case was run, and in particular the evidence led by both sides.
Prior transfer pricing cases in Australia such as Roche, SNF and Chevron have been heavily reliant on the evidence of transfer pricing experts. Their evidence has been unpersuasive to judges where it has not addressed the specific questions raised by the statutory text: in Chevron one judge said, ‘I give no weight to the opinions of transfer pricing economists where those opinions appear not to be founded in the statutory language which the Court must apply.’
None of these kinds of transfer pricing experts were called by either side in Glencore. Instead, both sides led evidence from finance experts. They gave evidence about the profitability of Glencore’s mining operations, its future cash flow needs, expected production costs, and so on. This evidence was also dismissed as ‘totally irrelevant to the issues for determination.’
Instead, the judge paid most attention to the evidence of practices in the copper industry and market. Glencore led evidence from the editor of the leading copper industry journal and from a senior Glencore executive experienced in the company’s operations and the copper market. The ATO also led evidence about the industry practice. Unfortunately for the ATO, the evidence of one of its experts was largely dismissed because, ‘he had no experience at all with respect to price sharing contracts and had no specialist knowledge qualifying him to express any opinion about price sharing contracts [and] he had no experience of what a trader would do …’ Clearly that evidence about industry practice was what the judge was looking for.
Comparables. The ATO argued thatthe third party offtake agreements put forward by Glencore were not sufficiently comparable to provide evidence of arm’s length pricing. Davies J rejected this argument and held, following the SNF case, that the offtake agreements did provide probative evidence of the range of prices which might ‘be expected to operate between hypothetical arm’s length mine producers and traders.’ This affirms the SNF decision that comparables don’t have to be perfect to be relevant to transfer pricing analysis.
The ATO has already invested a lot of time and effort into this case and it may be unwilling to throw in the towel at the end of Round 1. The ATO's media release says simply it, 'will consider ... whether an appeal is appropriate.'