Federal Court agrees that deductions should be denied for interest on financing

A win for the Commissioner of Taxation at first instance

On 17 December 2021, the Federal Court of Australia published the decision of Justice Moshinsky in Singapore Telecom Australia Investments Pty Ltd v Commissioner of Taxation [2021] FCA 1597. Taxpayers with significant international financing in Australia, or transfer pricing investigations underway by the ATO, may wish to reflect upon the lessons from this latest instalment in the judicial consideration of the Australian transfer pricing provisions.^[1]

In summary, Moshinsky J held that Singapore Telecom Australia Investments Pty Ltd (STAI^[2]) was not entitled to deductions of approximately $894 million in relation to the years ending 31 March 2010-2013 inclusive, as it had not established that the interest paid under the Loan Note Issuance Agreement (LNIA, as amended)^[3] to SingTel Australia Investment Limited^[4] (SAI) was an arm's length price. There were two aspects of the transfer pricing legislation that the Court focused on, being whether:^[5]

• there were conditions operating between the two enterprises in their commercial or financial relations which differed from those that might be expected to operate between independent enterprises dealing wholly independently with one another (Conditions Issue); and
• but for those conditions, an amount of profits might have been expected to accrue to the entity; and by reason of those conditions, the amount of profits did not so accrue (Profits Issue).

Even though this was not a case in relation to Subdivision 815-B, in relation to whether it was permissible to reconstruct the conditions of a related party transaction (other than the price), the case put by the Commissioner was on a consistent basis to his approach in Glencore and Chevron. This continues to be a key concern for taxpayers given the Commissioner’s willingness to substitute key terms of a transaction (in this case, amendments to the LNIA).

To support its position, STAI had adduced evidence (lay and expert) to show that its actual cost of borrowing under the LNIA was not greater than the costs that a party in STAI's position might be expected to have paid to a party acting wholly independently in an arm's length debt capital markets (DCM) transaction. STAI’s expert, Mr Chigas, applied the actual borrowing and payments over ten years, including with deferred and capitalised interest.  

The Commissioner’s expert, Mr Johnson, also concluded that an independent party would have obtained debt in the DCM, but instead found that independent parties would have obtained:

1. a short 9 month bridging $5.2 billion facility (the Base Debt) that would have been refinanced on 31 March 2003 (the date of the Secondment Amendment); and
2. a $1.5 billion medium-term revolving bank facility, which would have enabled STAI to make regular interest payments on the Base Debt.

Mr Johnson concluded that undertaking these facilities would have saved STAI $1.23 billion in interest over the life of the loan. The Commissioner argued that the commercially viable alternative third party financings identified by Mr Johnson conformed closer to what transpired because the actual transaction was a sale of shares and not a loan of money, bond raising or other borrowing in financial markets.

Moshinsky J found that the principal difficulty with STAI's approach was that it departed too far from the actual transaction and the characteristics of the parties to the transaction. He found that:

• conditions were operating which differed from those which might be expected to operate between independent enterprises dealing wholly independently with one another. The actual transaction did not involve a DCM bond issue. It involved a vendor and purchaser of shares and the issue of loan notes totalling approximately $5.2 billion by way of partial consideration of the shares; and
• in relation to the Profits Issue, the calculation of the rate advanced by STAI adopted hindsight and was not by reference to what the parties, dealing with each other independently, might have done in June 2002, and at the time of the relevant amendments to the LNIA. The Court held that the original interest rate under the LNIA (1 year BBSW + 1% x 10/9) might have been expected to have continued throughout the life of the loan (and thereby ignoring the subsequent amendments made to the LNIA in relation to the interest rate).

The parties have orders to appear before Moshinsky J in February 2022 to provide submissions to give effect to the Court’s reasons for judgment. If an appeal is filed, it must be filed within 28 days from the date final orders are made.

Overview of key facts

The origins of this case emerged from the takeover in 2001 by Singapore Telecommunications Limited (SingTel) of the majority of the shares in Cable & Wireless Optus Ltd (Optus) from Cable & Wireless plc (C&W). In June 2002, SAI sold 100% of the issued capital of Singtel Optus Pty Ltd (SOPL) to STAI for consideration of $14.2 billion. This comprised $9 billion of ordinary shares issued  by STAI to SAI and approximately $5.2 billion of debt issued by STAI to SAI pursuant to the LNIA.

The Commissioner often examines debt pushdowns from a number of perspectives under the tax acts, including the anti-avoidance and transfer pricing provisions. In this case, he issued determinations under the transfer pricing provisions in October 2016 and also issued amended assessments.

STAI presented evidence from one lay witness who had held various senior leadership roles in STAI from 2001,^[6] as well as two expert witnesses (one in relation to credit rating, the other in relation to DCM).^[7] The Commissioner also called two expert witnesses in relation to credit rating and DCM.^[8] Joint reports were prepared by the experts in relation to credit rating and DCM.

Hypothetical characteristics

The challenge for all taxpayers when trying to apply Subdivision 815-A is to hypothesise what conditions “might be expected” to have been agreed between independent parties. The parties agreed (at [321]) that the hypothesis should be “as close as possible” to the actual transaction.  However, Moshinsky J selected certain elements of the actual transaction to include in the hypothesis but then rejected other elements because, in His Honour’s view, third parties would not be expected to enter into such terms. This makes it difficult for taxpayers self-assessing transactions because the ATO or a court may find characteristics are not relevant that, in the taxpayer’s view, are relevant. This means it is important for taxpayers to substantiate why each term and condition of their related party borrowings are commercial. Practically for organisations who are mindful of commercially sensitive data, this may present challenges in discharging the onus. Ideally this evidence would be collected contemporaneously to the arrangements being put in place.

As noted above, STAI and the Commissioner approached the hypothetical differently and Moshinsky J held that STAI’s approach departed too far from the actual transaction and the characteristics of the parties. The hypothetical conditions applied by Moshinsky J included that independent parties:

• in the position of STAI would be a member of a multinational corporate group like the SingTel group and a holding company of an operating subsidiary like SOPL;
• might have agreed to an interest rate of the 1 year BBSW plus 1%;
• would likely have a withholding tax gross up clause in the agreement;
• might have agreed that interest could be deferred and capitalised; and
• would likely have been given a guarantee from a parent company because it wouldn’t want a subsidiary to pay a much greater amount of interest (and therefore impact the parent’s financial position).

However, His Honour found that independent parties would not have agreed to, among other things:

• the Second and Third Amendments to the LNIA;
• to change the 1 year BBSW component of the interest rate to a fixed rate of 6.835% during the global financial crisis; or
• that the subsidiary would pay the parent a guarantee fee in return for the provision of the guarantee.

In reaching these conclusions, His Honour stated (at [320]) that “[t]he overarching consideration is that the enterprises in the hypothetical should generally have the characteristics and attributes of the actual enterprises. Giving effect to this consideration requires the party in the position of STAI to be treated as a member of a group like the SingTel group. It is not necessary for the purposes of the hypothetical to address the precise ownership structure whereby this is the case.” (emphasis added)

The approach to the hypothetical based on Moshinsky J’s analysis is to look for independent hypothetical lenders and borrowers with the same characteristics as the actual lender and borrower. If this approach was sustained on any appeal it would present practical difficulties for taxpayers in discharging the onus of proof. In the present case, based on Moshinsky J’s approach, His Honour was unable to accept the DCM expert evidence. We think that the result of this approach would require the presentation of evidence of preparedness to lend to subsidiaries of other competing multinationals in the same industry or conversely, of subsidiaries borrowing from other competing multinational parent entities in the same industry. This is certainly a point to watch with interest. We observe that in Chevron, each of Allsop CJ and Pagone J seemed not to focus on the characteristics and expected behaviour of the hypothetical lender; instead focusing on the characteristics and expected behaviour of the hypothetical borrower.^[9]

No guarantee fee assumed in the hypothesis

Taxpayers should ensure they put on sufficient evidence to support alternative positions, to the extent possible. Although determining that independent parties would provide a parent guarantee in the hypothetical (at [324]), it is surprising that His Honour also concluded (at [328]) the parties could not expect a guarantee fee would be charged by the parent to the subsidiary in the hypothesis. Moshinsky J’s reasons for concluding this were that the proposition was based on the expert evidence of Mr Chigas, whose evidence on this point His Honour found to be expressed in general terms. His Honour also noted there was no evidence that SingTel charged a fee for guaranteeing the obligations of Optus Finance Pty Ltd (a subsidiary of SOPL) under a separate $2 billion facility in May 2002.

If Moshinsky Js’ view is sustained, it would lead to a number of practical difficulties for taxpayers. Providing such a significant guarantee would have implications for the parent’s credit rating and ability to raise capital in the future. For that reason it would likely expect to be remunerated in the form of a guarantee fee. Also, even if a guarantee fee was not paid in 2002 in respect of the separate facility, it does not mean that a fee would not have been paid on the hypothetical if a guarantee is assumed to be given.  In this case, as a guarantee was not provided, it would have been difficult for the taxpayer to put on evidence about something that did not occur. This highlights the inherent difficulty when comparing the actual transaction to a hypothetical. 

For completeness we note that in Chevron, Pagone J accepted the proposition that, if it is assumed that the hypothetical borrower would have received the benefit of the parental guarantee, then the arm’s length consideration which is provided for the acquisition of the identified property (i.e. the loan) can include a guarantee fee paid by the hypothetical borrower to its parent (or another third party).^[10] However, His Honour’s judgement also makes it clear that, whether or not it can be concluded that a guarantee fee “might reasonably have been expected to have been paid”, turns on the evidence.^[11]

Getting the brief to experts right

As with all taxation disputes, the onus remains on the taxpayer to demonstrate the excessiveness of the amended assessments. Predicting how independent parties dealing in an arm’s length transaction would price a wholly controlled transaction is difficult and complex. The lessons for taxpayers in utilising expert evidence to support their positions or to defend against adverse transfer pricing assessments are:

• The expert’s expertise, qualifications and experience are closely examined by the Court in order to determine the weight to be attributed to the evidence. For example, Moshinsky J held that in relation to credit rating, he preferred the evidence of Dr Chambers, expert for STAI, more so than Mr Weiss. A range of reasons were provided for this preference, including that Mr Chambers held a more senior position at Standard & Poors and was with the credit rating agency for a longer period of time compared to Mr Weiss.
• The formulation of the question continues to be a complex task. In relation to the DCM issue, Moshinsky J critiqued the evidence of both Mr Johnson and Mr Chigas on the basis that their reports were directed to responding to questions that did not reflect the statutory test. That is, the reports were responsive to a question regarding only the conditions in the LNIA, as distinct from conditions operating between STAI and SAI in their commercial or financial relations, which is the question raised by Article 6 of the Singapore DTA.
• Asking the expert to present views on alternative assumptions may be prudent case management. Moshinsky J found it difficult to accept the evidence of Mr Johnson (expert for the Commissioner) in so far as it relied upon the credit rating analysis of Mr Weiss which the Court did not prefer.
• In order to provide assistance to the Court, the evidence presented should be particular and consistent. For example, in evaluating the reliability of the evidence, Moshinsky J had particular regard to whether the views expressed in the respective expert reports were sustained or conceded in cross-examination. Moshinsky J took into account the extent of any inconsistencies between the individual expert reports and the joint report. Mr Weiss was found to have lost some credibility given that the joint expert report in relation to credit rating agreed that a “top down” approach is required under the Standard & Poors criteria, and Mr Weiss’ analysis had been founded on a “bottom up” approach.^[12]

Lay witness evidence and its assistance to the Court

As with all taxation disputes, there is often a lengthy period of time between the relevant events and when the matter is before a Court. This can present issues in case management and evidentiary risk for taxpayers.

STAI called one lay witness, Mr O’Sullivan. Aspects of Mr O’Sullivan’s evidence regarding implicit support were accepted by Moshinsky J. For example, Mr O’Sullivan gave evidence that had there been a requirement to dispose of SOPL, the integrations would not have affected the disposition, they could have been maintained under commercial arrangement, or otherwise separated.  Moshinsky J found this evidence relevant to the issue of implicit support in the consideration of the credit rating of STAI.

It should also be noted that Moshinsky J found that Mr O’Sullivan’s evidence in relation to whether SingTel would have provided a parental guarantee was not of any assistance including because Mr O’Sullivan was not involved in the preparation of the LNIA nor in communications about its terms. The evidence was also expressed in general terms. Moshinsky J held that Mr O’Sullivan was not in a position to give any probative evidence about whether SingTel would have (or might be expected to have) provided a guarantee of STAI’s obligations under the LNIA (or an arm’s length transaction in place of the LNIA). Given the Court’s comments regarding parental guarantees, the question remains as to whether evidence from the perspective of SAI and/or SingTel would have been of greater assistance in resolving this factual finding.

^[1] Litigation of the transfer pricing provisions in Australia is rare, with only a handful of disputes proceeding to Court: Commissioner of Taxation v Glencore Investment Pty Ltd [2020] FCAFC 187 (Glencore); Chevron Australia Holdings Pty Ltd v Commissioner of Taxation [2017] FCAFC 62 (Chevron); Federal Commissioner of Taxation v SNF (Australia) Pty Ltd (2011) 193 FCR 149; Re Roche Products Pty Ltd v Federal Commissioner of Taxation (2008) 70 ATR 703

^[2] A company incorporated and resident in Australia

^[3] Under the original terms of the LNIA, the interest rate under the LNIA was the 1 year Bank Bill Swap Rate (BBSW) plus 1% per annum. The applicable rate was the interest rate multiplied by 10/9. Subsequent to June 2002, the LNIA was amended three times:

• On 31 December 2002, in relation to maturity of a Loan Note (not later than the tenth anniversary of the issue date less one day) (the First Amendment);
• On 31 March 2003, regarding the terms for accrual and payment of interest, including adding a premium of 4.552% (the Second Amendment);
• On 30 March 2009, by replacing the 1 year BBSW component of the interest rate to a fixed rate of 6.835%. The applicable rate became (a) the interest rate (6.835% plus 1%) multiplied by 10/9 plus (b) the premium (of 4.552%) resulting in an applicable rate of 13.2575% (the Third Amendment).

^[4] A company incorporated in the British Virgin Islands and resident in Singapore

^[5] Although the case involved both Division 13 of the Income Tax Assessment Act 1936 (Cth) (ITAA 1936) and Subdivision 815-A of the Income Tax Assessment Act 1997 (Cth) (ITAA 1997), the focus of the argument by the parties was on Subdivision 815-A.

^[6] Mr Paul O’Sullivan (Mr O’Sullivan), the Chairman and a director of SOPL.  Mr O’Sullivan was the Chief Operating Officer of SOPL from September 2001 to August 2004.  He was the Chief Executive Officer of SOPL from September 2004 to March 2012.  He was a director of STAI from 27 September 2004 to 8 October 2014.  

^[7] Dr William J Chambers, credit rating; and Mr Charles W Chigas, DCM

^[8] Mr Robert A Weiss, credit rating; and Mr Gregory Johnson, DCM

^[9] For example, refer to Chevron at [44] per Allsop CJ and at [128] per Pagone J.

^[10] Chevron at [133] per Pagone J.

^[11] Chevron at [133] per Pagone J. Also refer to Chevron at [60] per Allsop J.

^[12] At [226]