FEDERAL COURT OF AUSTRALIA

Singapore Telecom Australia Investments Pty Ltd v Commissioner of Taxation [2024] FCAFC 29

ORDERS

THE COURT ORDERS THAT:

1.    The appeal is dismissed.

2.    The appellant pay the respondent's costs to be assessed if not agreed.

Note:    Entry of orders is dealt with in Rule 39.32 of the Federal Court Rules 2011.

REASONS FOR JUDGMENT

THE COURT:

1    On 28 June 2002, Singapore Telecom Australia Investments Pty Ltd (STAI) acquired all of the shares in a company that came to be named SingTel Optus Pty Ltd (SOPL). The vendor of the shares was Singtel Australia Investments Limited (SAI). Both SAI and STAI were wholly owned subsidiaries of Singapore Telecommunications Limited (SingTel). SOPL operated the Optus telecommunications business in Australia.

2    Funds to purchase the shares in SOPL were provided by SAI. The instrument recording the terms of the vendor finance was termed a Loan Note Issuance Agreement (LNIA). An amount of $5.2 billion was advanced under the terms of the LNIA.

3    The LNIA had some unusual terms. In particular, it contained a mechanism by which the payment of interest was not required until a 'variation notice' had been issued by SAI. The effect of that mechanism was that SAI could determine precisely when, over the 10 year term of the LNIA, STAI was required to pay interest and in what amount. It was intended to operate in a manner that would ensure that the liability to pay interest corresponded with the periods when STAI was earning sufficient profits to be able to meet the interest expense. The ability of SAI to determine the timing of the interest obligations of STAI had particular significance in circumstances where, in the first few years of the loan term, STAI was not expected to earn profits because SOPL was committed to undertaking substantial capital investment and it was not until after it had done so that profits were expected to flow to its shareholder, STAI. However, under the terms of the LNIA as originally agreed, the liability to pay interest still accrued. It was just the timing of the obligation to make payment that could be deferred by use of the 'variation notice' procedure.

4    Although it was STAI that needed to be able to defer repayment of interest until it had sufficient cash flow to make repayments, it was SAI that had the power to effect deferral through the issue of a variation notice. However, as has been explained, the arrangements were put in place between two wholly owned subsidiaries of SingTel in respect of the acquisition of shares in SOPL.

5    The interest rate payable under the LNIA was the one year bank bill swap rate from time to time plus 1%. That is to say, the rate payable was to be adjusted each year by reference to the bank bill swap rate at that time plus 1%. It was then to remain fixed for the next year. It was then to be reset again by reference to the then current bank bill swap rate plus 1%. The formula for calculation of interest included a further factor that was designed to require STAI to pay the 10% withholding tax applicable by reason that interest was to be remitted to SAI, a company incorporated in the British Virgin Islands. However, for present purposes that aspect of the interest formula assumes little significance.

6    STAI could repay the loan notes issued under the LNIA at any time. The LNIA also provided that SAI could at any time require STAI to redeem the loan notes issued under the terms of the LNIA. In effect, the whole of the borrowing could be repaid voluntarily by STAI or could be required to be repaid by STAI at any time during the 10 year term. Therefore, it provided no security of ongoing funding for STAI nor any security of return at the agreed interest rate for SAI.

7    The LNIA was amended on three occasions. Of present significance are the second and third amendments. They were made on 31 March 2003 and 30 March 2009 respectively.

8    By the time of the second amendment, STAI had not paid interest nor was any interest due to be paid because SAI had not issued any variation notices under the LNIA. Even so, under the accounting standards applicable to the preparation of accounts for SingTel on a consolidated group basis, SingTel would have been required to account for accrued liabilities arising from the LNIA. Those accrued liabilities would then fall due for payment when variation notices were issued by SAI. In the meantime, it appeared that withholding tax would be payable in respect of the accrued liabilities. On the evidence, it was those prospects that prompted the second amendment to the LNIA.

9    Broadly speaking, the second amendment did three things. First, it forgave the accrued obligation to pay interest for the period from the commencement of the LNIA until the amendment. In that respect, it was common ground that the second amendment relieved STAI of an accrued obligation to pay an amount of approximately $286 million. Second, it introduced a profitability benchmark (with retroactive effect) with the consequence that there could be no liability for interest (and hence no withholding tax liability) until the benchmark was met. Third, it added a further factor of 4.552% of the principal debt to the formula for the interest calculation. This factor was said to have been calculated on the basis of an expectation as to when the benchmark would be met. The additional factor was designed to equate the overall interest to be paid over the term of the LNIA to the equivalent economic value of the interest that would have been payable if the amendment had not been made (namely interest over the 10 year term at the agreed rate of the one-year bank bill swap rate plus 1%). However, the equivalence was dependent upon the benchmark being met at the date that was used for the purposes of making the calculation. At various points in the submissions the additional 4.552% was referred to as an interest premium. The use of the term 'premium' is somewhat inconsistent with the notion of equivalence that was used to determine the amount. Nevertheless, these reasons will also use that terminology.

10    The second amendment had considerable significance for the timing of the accrual of any obligation on the part of STAI to pay interest under the LNIA as well as the timing of the obligation to pay that interest. Whereas under the original LNIA, an agreed rate of interest (which varied according to the bank bill swap rate) applied from the outset of the term of the LNIA with the timing of the obligation to pay that interest dependent upon the issue of a variation notice, the LNIA as amended by the second amendment deferred the timing of any obligation to pay interest until the point in time when the benchmark was met. At that time the bank bill swap rate would still determine the base rate of interest but both the margin of 1% and the premium of 4.552% would be added to reflect the fact that interest was only being paid once the benchmark was met. It represented the interest that would have been paid before the benchmark was met if an interest liability had accrued at that time (as had been the case under the original terms of the LNIA).

11    The use of the benchmark mechanism gave rise to another timing issue. It arose because the premium of 4.552% would only result in the total interest under the LNIA being equivalent to the application of the rate of interest that had originally been agreed if two assumptions came to be satisfied in the events which subsequently occurred. The first assumption was that the benchmark was met on the expected date which had been used to calculate the premium. The second assumption was that the interest continued to be calculated according to the original rate plus the premium for the balance of the term of the LNIA. If the benchmark was met at an earlier date (or there was some subsequent agreed change to the way the interest rate was determined) then there would be no economic equivalence with the amount as originally agreed. Indeed, if the benchmark was not met during the term of the LNIA (a possibility which was accepted by STAI's main expert to exist as at the time of the second amendment) then there would be no interest payable at all.

12    In the result, the benchmark was met earlier and also the LNIA was subsequently further amended by the third amendment so that the base interest rate was fixed. Each of those events resulted in considerably more interest being paid by STAI in respect of the funds advanced under the LNIA than would have been the case if the original rate had applied over the whole of the term. Therefore, in two respects, the premise for economic equivalence was not met in the circumstances as they unfolded.

13    As to the third amendment, it introduced the fixed rate. It replaced the original variable rate (to which the premium of 4.552% was added) with a fixed rate for the balance of the loan term. It produced an overall fixed rate of interest of 13.2575% for the balance of the term of the LNIA. As has been observed, in the events which occurred, the fixed rate resulted in more interest being payable than would have been the case if the variable rate had continued to apply as the base rate.

14    The financial years of the SingTel entities including STAI commenced on 1 April each year. Therefore, the taxation liabilities of STAI were determined according to the position in each of those years.

15    The Commissioner made determinations both under the cross-border transfer pricing provisions in Subdivision 815-A of the Income Tax Assessment Act 1997 (Cth) (ITAA97) and under the arm's length consideration provisions of Division 13 of the Income Tax Assessment Act 1936 (Cth) (ITAA36) and then issued notices of amended assessment for STAI for the financial years ending March 2011, 2012 and 2013 (being the final years in the 10 year term of the LNIA). The effect of the amendments was to disallow substantial deductions for interest payments under the LNIA in those years.

16    STAI objected to the assessments. The Commissioner disallowed the objections. STAI brought an appeal pursuant to Part IVC of the Taxation Administration Act 1953 (Cth) against the Commissioner's objection decisions. The appeal was unsuccessful: Singapore Telecom Australia Investments Pty Ltd v Commissioner of Taxation [2021] FCA 1597 (PJ). STAI now brings an appeal against the decision of the primary judge.

17    STAI articulated its appeal on the basis of some 49 separate appeal grounds. However, the written and oral submissions in support of the appeal were developed in terms that were grouped under seven alleged errors (each said to relate to a group of appeal grounds). The case for STAI on the appeal was put by reference to the seven alleged errors. It was not advanced by reference to the individual appeal grounds in any meaningful sense. The Commissioner responded on the same basis and also relied upon a notice of contention.

18    Before considering the particular matters raised by the grounds and contentions, it is necessary to consider four matters by way of introduction, namely:

(1)    the statutory context;

(2)    the history concerning the relevant taxation assessments;

(3)    the nature of the expert evidence before the primary judge; and

(4)    the reasoning pathway of the primary judge.

The statutory context

19    As has been indicated, the Commissioner pursued alternate statutory pathways as the foundation for disallowing the objections, being Subdivision 815-A of ITAA97 and Division 13 of ITAA36. It will be necessary to deal with both of them noting that the primary judge approached the issues in the case principally through the framework of Subdivision 815-A.

20    If Subdivision 815-A applies to support the Commissioner's position (with the consequence that STAI fails in its attempt to demonstrate that the assessments are excessive) then the only significance of considering whether Division 13 applies is for the purposes of liability to penalties. If Subdivision 815-A does not apply to support the Commissioner's position (with the consequence that STAI succeeds in demonstrating that the assessments are excessive to the extent that they rely on those provisions) then it is necessary to consider whether Division 13 applies in determining whether the assessments were excessive.

21    The Court is concerned with the statutory provisions as they applied at the relevant time. In that regard, up until 1 July 2004, only Division 13 applied. Then for tax years starting between 1 July 2004 and 28 June 2013, both Division 13 and Subdivision 815-A applied. Then those provisions ceased to operate and Divisions 815-B to 815-D of the ITAA97 were the provisions that applied to transfer pricing. In the following reasoning, reference is made to the provisions as they applied during the term of the LNIA.

The relevant provisions in Subdivision 815-A

22    Subdivision 815-A operates by empowering the Commissioner to make a determination 'for the purpose of negating a transfer pricing benefit an entity gets': s 815-10(1). As to such a benefit, s 815-15(1) provides:

An entity gets a transfer pricing benefit if:

(a)    the entity is an Australian resident; and

(b)    the requirements in the *associated enterprises article for the application of that article to the entity are met; and

(c)    an amount of profits which, but for the conditions mentioned in the article, might have been expected to accrue to the entity, has, by reason of those conditions, not so accrued; and

(d)    had that amount of profits so accrued to the entity:

(i)    the amount of the taxable income of the entity for an income year would be greater than its actual amount; or

(ii)    the amount of a tax loss of the entity for an income year would be less than its actual amount; or

(iii)    the amount of a net capital loss of the entity for an income year would be less than its actual amount.

The amount of the transfer pricing benefit is the difference between the amounts mentioned in subparagraph (d)(i), (ii) or (iii) (as the case requires).

23    In terms of overall structure, the provision identifies four circumstances that must exist for an entity to get a transfer pricing benefit and the amount of the transfer pricing benefit is an amount for an income tax year (being an amount of taxable income, an amount of a tax loss or an amount of net capital loss). It is that income year amount which may be negated by the Commissioner making a determination to that effect.

24    Turning then to the provisions of s 815-15(1) and the way in which each of the four circumstances may apply to SAI and STAI in the present case.

25    As to the first circumstance specified in (a), it was common ground that STAI was an Australian resident.

26    As to the next circumstance, it can be seen that (b) incorporates by reference 'the requirements in the associated enterprises article for the application of that article'.

27    It is common ground that the relevant associated enterprises article for present purposes is Article 6 of the Agreement between Australia and Singapore for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, and the protocols to that agreement. It is expressed in the following terms:

(1)    Where-

(a)    an enterprise of one of the Contracting States participates directly or indirectly in the management, control or capital of an enterprise of the other Contracting State; or

(b)    the same person participates directly or indirectly in the management, control or capital of an enterprise of one of the Contracting States and an enterprise of the other Contracting State,

and in either case conditions operate between the two enterprises in their commercial or financial relations which differ from those which might be expected to operate between independent enterprises dealing wholly independently with one another, then any profits which, but for those conditions, might have been expected to accrue to one of the enterprises, but, by reason of those conditions, have not so accrued, may be included in the profits of that enterprise and taxed accordingly.

28    There are two 'requirements' stated in Article 6 for its application. First, there must be the requisite degree of management, control or capital as specified in (a) or (b). In the present case it was common ground that this first requirement was met as between SAI and STAI. Second, there must be conditions that operate between the two enterprises in their commercial or financial relations which differ from those which might be expected to operate between independent enterprises dealing wholly independently with one another. In these reasons, these conditions are referred to as the Non-Independence Conditions.

29    It is important to bear in mind that the Non-Independence Conditions are not confined to any agreed terms which govern the dealings between the two entities, whether by formal agreement or by understanding. Rather, they encompass all of those conditions which operate in the commercial and financial relations between SAI and STAI. It includes their relevant commercial and financial characteristics as separate entities and the way in which those characteristics affect the relations between them. In this respect, s 815-15(1) applies in a different way to Division 13 of ITAA36. The latter applies where (a) there was a supply of property; (b) the parties were not dealing at arm's length in relation to the supply; and (c) the consideration received or given by the taxpayer was less than arm's length consideration (see below). Therefore, Division 13 focusses much more closely upon whether there was non-arm's length consideration received or given by the taxpaying in a particular dealing.

30    For present purposes, the second circumstance as expressed in s 815-15(1)(b) is met if the conditions that operate between SAI and STAI are Non-Independence Conditions.

31    The third circumstance as specified in (c) is that there is 'an amount of profits' which but for the Non-Independence Conditions 'might have been expected to accrue' to STAI but, by reason of the Non-Independence Conditions, has not so accrued. So, there is a 'but for' causative requirement as between the Non-Independence Conditions and the failure to accrue an amount of profits. What must be identified is an amount of profits which might have been expected to accrue to STAI if the Non-Independence Conditions were not present which has not accrued by reason of the Non-Independence Conditions. It invokes a hypothetical concerning the profits that might have been expected to accrue to STAI.

32    Many of the issues in the appeal concern the nature of the third circumstance and its application in the conditions that pertained as between SAI and STAI.

33    The fourth circumstance concerns the taxation consequences of the third circumstance. It proceeds on the premise that the third circumstance has been met (namely the amount of profits that might have been expected to accrue but for the Non-Independence Conditions did actually accrue to STAI) and requires one of three taxable outcomes to have occurred in consequence - greater taxable income for an income year, a reduced tax loss for an income year or a reduced net capital loss for an income year. In the present case, the assessments were issued by the Commissioner on the basis that the taxable income of STAI would have been greater in each of the financial years ending 31 March 2011, 2012 and 2013 because the interest expense charged by SAI under the LNIA would have been significantly less but for the Non-Independence Conditions.

34    In summary, s 815-15(1) operates by reference to (a) an amount of profits that might have been expected to accrue; (b) that did not accrue; and (c) which had it accrued would have had a particular tax consequence in a particular income year. It describes an entity as getting a 'transfer pricing benefit' if each of those matters pertain. It identifies the amount of the transfer pricing benefit as an amount of taxable income, tax loss or capital loss of the entity that would have received the profits 'for an income year' (see concluding words to s 815-15(1)). That is to say, the amount of the transfer benefit relates to an income year. It is that amount which the Commissioner can negate by making a determination under s 815-30 (see s 815-10). Section 815-30 relevantly provides:

(1)    The determinations the Commissioner may make are as follows:

(a)    a determination of an amount by which the taxable income of the entity for an income year is increased;

(b)    a determination of an amount by which the tax loss of the entity for an income year is decreased;

(c)    a determination of an amount by which the net capital loss of the entity for an income year is decreased;

(d)    a decrease of a particular amount in particular capital losses of the entity for an income year.

(2)    If the Commissioner makes a determination under subsection (1), the determination is taken to be attributable, to the relevant extent, to such of the following as the Commissioner may determine:

(a)    an increase of a particular amount in assessable income of the entity for an income year under a particular provision of this Act;

(b)    a decrease of a particular amount in particular deductions of the entity for an income year;

(c)    an increase of a particular amount in particular capital gains of the entity for an income year;

(3)    If the Commissioner makes a determination under subsection (1), the Commissioner must make a determination under subsection (2), unless it is not possible or practicable for the Commissioner to do so.

35    Therefore, Subdivision 815-A results in a determination that adjusts the taxable income, tax loss or net capital loss, as the case may be, 'for an income year'. It requires the Commissioner to identify the amount of the transfer pricing benefit that pertains to an income year and determine an amount that negates that benefit. The determined amount 'is taken to be attributable' to that income year: s 815-30(2).

36    The express object of Subdivision 815-A is to ensure that certain amounts 'are appropriately brought to tax in Australia': s 815-5. It is fundamental to the tax laws that assessment for taxation is imposed in respect of an income year. Generally speaking, amounts are brought to taxation in Australia in accordance with the year of income to which they relate.

37    Therefore, regard to the scheme of Subdivision 815-A as a whole requires a determination by the Commissioner not only of the amount of profits which, but for the non-arm's length conditions, might have been expected to accrue to the relevant entity, but also the income year in which those profits might have been expected to accrue.

38    In addition, Subdivision 815-A includes detailed consequential adjustment provisions which confer authority upon the Commissioner to make further determinations that make consequential adjustments in any income year. Those provisions empower the Commissioner to ensure that the negation of a transfer pricing benefit in one income year does not operate unfairly or unreasonably in respect of any other income year. They are referred to below in dealing with equivalent aspects of Division 13.

39    The possibility that Non-Independence Conditions might have profit consequences across a period of years does not mean that the amount of profits with which s 815-15 is concerned is an aggregate amount of profits across those income years. Subdivision 815-A applies with respect to the amounts of profits that correspond with a particular income year. That is, the statute operates in a way that affords significance both to the amount of the profits and the income year in which they accrued.

40    Take a simple example. A taxpayer enters into a non-arm's length dealing which provides that there will be no interest payable unless the loan amount is still outstanding after three years. Absent Non-Independence Conditions, the interest would have accrued and been payable by equal instalments in each month of the three year term. The interest is a deductible expense. Each year of the loan term corresponds to an income tax year. If interest is incurred because the loan amount is still outstanding, the effect of the agreement is that an amount of profits which but for relevant Non-Independence Conditions might have been expected to accrue in income tax year three is not accrued in that year because the interest expense in that year is three times higher than might have been the case without the Non-Independence Conditions. The Commissioner reaches the view that the greater amount of taxable income that would otherwise have been earned in income tax year three is a transfer pricing benefit and makes a determination negating the transfer pricing benefit.

41    In such a case, the taxpayer could not contend that there is no basis to make a determination for the purpose of negating a transfer pricing benefit because the interest paid in year three is the same as that which would have been payable over three years if there had been an arm's length dealing. The timing difference has significance for an income tax year and produces a transfer pricing benefit for that year.

42    Of course, it may be said that it would be unfair and unreasonable for the Commissioner to make a determination that would increase the taxable income in the third year without making an adjustment to years one and two to allow for the deductible interest expense that would have been available in those years if the hypothesis used to make the determination also applied to those years. However, as has been mentioned, Subdivision 815-A has detailed provisions which confer power upon the Commissioner to make a further determination in respect of those other years if the Commissioner considers it fair and reasonable to do so (and to make consequential adjustments to taxable income in other tax years). The taxpayer may request the Commissioner to make such a determination and has a right to object if dissatisfied. Indeed, the existence of these provisions reinforces the evident structure of Subdivision 815-A which is concerned with an amount of profits that might have accrued in an income year.

43    As will emerge, much of the case for STAI sought to ignore the timing effects of the way the LNIA operated (both in its original and amended forms). It sought to demonstrate that, viewed with the benefit of hindsight, over the entire 10 year term of the LNIA the total interest that was in fact paid by STAI was less than the interest that might have been expected to be paid if an arm's length rate had been agreed at the outset. On that basis, it contended that there was no amount of profits which, but for the relevant Non-Independence Conditions as between SAI and STAI, might have been expected to accrue to STAI. It treated the 'amount of profits which … might have been expected to accrue' as being concerned with the profits that accrued from the dealing as a whole (in the present case, the performance of the LNIA over its 10 year term). In consequence, STAI maintained that it was not until the end of the 10 year term of the commercial dealing that the Commissioner could make a determination under Subdivision 815-A.

44    For reasons that have been given, this was not the correct approach because it ignored the timing effects of when profits were earned by STAI. It assumed that Subdivision 815-A was concerned only with whether the total interest cost charged over the whole term of the borrowing was less than the total interest cost that might have been charged if the Non-Independence Conditions did not apply. However, as has been explained, Subdivision 815-A was also concerned with the effect upon taxable income (tax loss and net capital loss) in a particular income year. That is to say, it was concerned with the years in which there was an effect upon taxable income (tax loss and net capital loss) by reason of the Non-Independence Conditions.

45    This flaw in the case advanced by STAI permeates many of its contentions in the appeal insofar as it concerns Subdivision 815-A (and, for reasons explained below, is also a flaw that permeates the case as to Division 13). As has been explained, the statutory concept of a 'transfer pricing benefit' includes an assessment as to the effect upon the amount of taxable income for an income year. It requires an assessment as to whether non-arm's length dealings have affected the profits in a particular tax year and consequently the amount of taxable income or loss (or net capital loss) for that year. Consequently, there is no foothold in the legislation for concluding that the making of any determination under Subdivision 815-A had to be deferred until the course of performance of a long-term transaction has concluded. If circumstances exist at any point in the performance of a long-term contract which have produced a transfer pricing benefit for an income year then the Commissioner may make a determination for the purpose of negating that transfer pricing benefit.

46    STAI sought to rely upon statements by Allsop CJ in Chevron Australia Holdings Pty Ltd v Commissioner of Taxation [2017] FCAFC 62; (2017) 251 FCR 40 at [17] to support its claim that the evaluation as to whether an amount of profits might have accrued but for the non-arm's length conditions must be made retrospectively once it is known how much interest was paid over the whole term of the LNIA. Those statements were made in the context of a consideration of the application of Division 13 which operates in a different manner to Subdivision 815-A (see below). Further, they concerned an instance where no issue arose as to the significance of the timing of interest payments. The issue in Chevron was whether the rate of interest exceeded the arm's length consideration. In determining whether that was the case, the Court had regard to the fact that the borrowing was for a five year term. However, it did not have to consider the accrual or deferment aspects that were fundamental to the way the interest liability was determined under the LNIA. Therefore, the reasoning in Chevron was not directed to answering that question.

47    STAI also contended that the primary judge had approached the matter on the basis that the relevant profits were those that had been earned over the term of the STAI rather than in the income years. The primary judge did say (PJ[300]):

Further, in circumstances where the LNIA was a 10 year transaction, it is necessary to consider the issues in relation to the whole life of the LNIA, not just the LNIA as it stood during the years ending 31 March 2010, 2011, 2012 and 2013. There does not appear to be any issue between the parties about this. Both parties, in their evidence and submissions, addressed the whole life of the LNIA. Unless one goes back to the beginning of the transaction, that is, when the LNIA was entered into, one cannot sensibly apply the provisions of Subdiv 815-A to the years ending 31 March 2010, 2011, 2012 and 2013.

48    However, two matters must be noted about that part of his Honour's reasoning. First, it is apparent from the final sentence that his Honour approached the matter on the basis that Subdivision 815-A required application of its provisions to each of the relevant income years. Second, in the reasoning that follows on from PJ[300], the primary judge dealt with the different circumstances that applied in each of three periods that corresponded with when the initial terms of the LNIA applied, when the second amendment applied and when the third amendment applied: see, particularly PJ[302]. His Honour reached particular conclusions about whether the terms that applied in each period differed from those that might be expected on an arm's length basis and the profits that might be expected to accrue but for those non-arm's length conditions in each of those periods: PJ[303]-[344]. His Honour then reached conclusions as to the position in each of the relevant tax years: PJ[345]-[348]. His Honour did not approach the statutory task on the basis contended for by STAI in its submissions on the appeal.

49    STAI's approach would also give rise to serious practical difficulties for the administration of Subdivision 815-A. First, on STAI's approach, in the case of any long term dealing, the required comparison with the reliable hypothesis could not be undertaken until the end of the term of the dealing. In the present case that period was 10 years, but it may be many more. In consequence, the Commissioner could only make the required determinations for the operations of Subdivision 815-A and Division 13 many years after the income years in question. Issues would arise as to how those provisions would operate to affect taxation liabilities that arose many years previously. Second, as the facts in the present case demonstrate, long term arrangements are susceptible to amendment. It is impractical to determine the application of the provisions with hindsight to the whole term of an agreement when its terms have changed part way through that term. Third, the transitional provisions for the introduction of Subdivision 815-A provide for it to apply 'to income years starting on or after 1 July 2004'. Any long term transaction which straddled that date would be subject to its terms for only part of its term. It is difficult to see how STAI's aggregating, hindsight-based approach could be applied in such a case. Where, as here, there have been changes to the transaction over time, it would be unclear as to the point in time when the assessments to be made as to whether the Non-Independence Conditions are met but for which an amount of profits might have been expected to accrue. Indeed, the use of an income year for the transitional provision supports the construction that has been outlined.

50    These practical difficulties are further reasons not to accept STAI's construction as to the operation of Subdivision 815-A.

The relevant provisions in Division 13

51    The relevant operative provision in Division 13 of ITAA36 was s 136AD(3) which provided:

Where:

(a)    a taxpayer has acquired property under an international agreement;

(b)    the Commissioner, having regard to any connection between any 2 or more of the parties to the agreement or to any other relevant circumstances, is satisfied that the parties to the agreement, or any 2 or more of those parties, were not dealing at arm's length with each other in relation to the acquisition;

(c)    the taxpayer gave or agreed to give consideration in respect of the acquisition and the amount of that consideration exceeded the arm's length consideration in respect of the acquisition; and

(d)    the Commissioner determines that this subsection should apply in relation to the taxpayer in relation to the acquisition;

then, for all purposes of the application of this Act in relation to the taxpayer, consideration equal to the arm's length consideration in respect of the acquisition shall be deemed to be the consideration given or agreed to be given by the taxpayer in respect of the acquisition.

52    It was common ground that SAI had supplied property to STAI under an international agreement. Property included services which in turn was defined to include any rights or benefits to be provided or conferred under an agreement for or in relation to the lending of monies. Therefore, the interest terms agreed under the LNIA were 'property' supplied to STAI. It was also common ground that SAI and STAI were not dealing at arm's length with each other.

53    As to s 136AD(3)(c), the term 'arm's length consideration' was to be construed having regard to the terms of s 136AA(3)(d) which provided:

a reference to the arm's length consideration in respect of the acquisition of property is a reference to the consideration that might reasonably be expected to have been given or agreed to be given in respect of the acquisition if the property had been acquired under an agreement between independent parties dealing at arm's length with each other in relation to the acquisition;

54    Therefore, the terms of s 136AD(3)(c) were met if the amount of the interest given or agreed under the LNIA exceeded that which might reasonably be expected to have been given or agreed had the interest been paid under an agreement between independent parties dealing at arm's length with each other.

55    It is common ground that the Commissioner determined that s 136AD should apply in relation to STAI for each of the income years ending 31 March 2011, 2012 and 2013. Separate determinations were made for each of those years. Regard to the determinations makes plain that each one determined an arm's length amount of consideration that was to be substituted for the interest actually paid in that year by STAI for the acquisition of property under an international agreement (namely the procurement of financial accommodation under the terms of the STAI by the issue of notes).

56    As to those determinations, s 136AD(4) provides:

For the purposes of this section, where, for any reason (including an insufficiency of information available to the Commissioner), it is not possible or not practicable for the Commissioner to ascertain the arm's length consideration in respect of the supply or acquisition of property, the arm's length consideration in respect of the supply or acquisition shall be deemed to be such amount as the Commissioner determines.

57    An issue arises as to the effect of the provision that the determined consideration shall be deemed to be the consideration 'for all purposes of the application of this Act': see s 136AD(3) (quoted above). It is addressed below. However, what is clear is that the determinations in issue were not in fact based upon some view being taken as to the total consideration that was paid by STAI over the 10 year term of a kind that could only be made once that total amount as actually paid was known. Further, the terms of Division 13 did not require any such determination to be made. They did not require the Commissioner to defer making any such determination until all consideration payable under the terms of LNIA were known. If and when consideration has been given or agreed (see the terms of s 136AD(3)(c) as quoted above) then the determination may be made to substitute arm's length consideration for that which has been given or agreed.

58    It may be observed that Division 13 operates in a different way to Subdivision 815-A. Division 13 is concerned with ascertaining the arm's length consideration in respect of the relevant dealing (in the present case the acquisition of finance services) and substituting the arm's length consideration for the actual consideration, thereby affecting the computation of any taxation payable by reason of the amount of that consideration. Its focus is upon adjusting the consideration actually given or agreed. It can make that adjustment in respect of the agreed consideration (to the extent that the fact of agreement is relevant for taxation purposes, such as when it affects accruals) or consideration that has been given, that is performed (to the extent that it is relevant for taxation purposes). In the latter case, the determination takes the time at which the consideration was given and replaces the amount with arm's length consideration for that time. Conceptually, it may require regard to all of the terms of the relevant international agreement in determining the arm's length consideration to be ascertained (or determined by the Commissioner and deemed to be the consideration in a case where s 136AD(4) is applied) and substituted at that time. However, it does not require anything more than the fact that consideration has been agreed or given for property that has been acquired under an international agreement. The determination adjusts that consideration to arm's length consideration.

59    The same is the case where the arm's length consideration 'received or receivable' is less than the arm's length consideration (s 136AD(1)) or no consideration is 'received or receivable' (s 136AD(2)) thereby resulting in less income to the Australian taxpayer. Those provisions also apply to deem the consideration 'received or receivable' to be the arm's length consideration (or an amount which the Commissioner determines if it is not possible or practicable to ascertain the arm's length consideration). In consequence, for tax purposes, the consideration that is received or receivable at a particular time is replaced by the arm's length consideration. Otherwise, tax is payable based upon the facts as they have occurred, including when the consideration is received or receivable. Accordingly, the determination will have consequences for the income year to which the consideration relates.

60    No aspect of Division 13 involves some form of adjustment or allowance as between income years as part of the replacement of the arm's length consideration. Nor does it involve some form of evaluation of the consideration agreed or given (received or receivable) over the whole term of an agreement (in the present case the LNIA).

61    The above analysis is consistent with the reasoning in WR Carpenter Holdings Pty Ltd v Commissioner of Taxation [2008] HCA 33; (2008) 237 CLR 198, noting that the High Court was there concerned with the provisions in Division 13 that apply to instances where Australian taxpayers are assessed on the basis that they have received insufficient consideration (rather than, as in the present case, where they are said to have paid excessive consideration): see reasoning at [23]-[27], [34]. In that case, the High Court approached the provisions of Division 13 on the basis that they require a decision to be made as to a year of income and may result in inclusion of income or disallowance of deductions where, in the Commissioner's opinion 'it is fair and reasonable' that the amount not be included or the deduction be allowed in that year of income (being a reference to the power of the Commissioner by a further determination to adjust the consequences of a determination substituting arm's length consideration). However, as the High Court emphasised, the power conferred upon the Commissioner to deal with those timing consequences is by a further determination. It is by the exercise of that separate statutory power that unfair or unreasonable consequences of a determination as to the arm's length consideration may be ameliorated. It is a power that a taxpayer can request the Commissioner to exercise and in respect of which the taxpayer can appeal. That is to say, the focus is upon substituting arm's length consideration for the actual consideration at the time it was agreed or given with any consequential determination as to what is fair and reasonable to be dealt with by way of further adjustment to be made separately.

62    In that regard, s 136AF(1), relevantly provided as follows:

Where, by reason of the application of section 136AD in relation to the supply or acquisition of property by a taxpayer, an amount is included in the assessable income of the taxpayer of a year of income or a deduction is not allowable or is not, in part, allowable, to the taxpayer in respect of a year of income, the Commissioner may, in relation to any taxpayer (in this subsection referred to as the relevant taxpayer):

…

(b)    if, in the opinion of the Commissioner:

(i)    an amount would have been allowed or would be allowable to the relevant taxpayer as a deduction in relation to a year of income if the property had been supplied or acquired, as the case may be, under an agreement between independent parties dealing at arm's length with each other in relation to the supply or acquisition, being an amount that was not allowed or would not, but for this subsection, be allowable, as the case may be, as a deduction to the relevant taxpayer in relation to that year of income; and

(ii)    it is fair and reasonable that that amount or a part of that amount should be allowable as a deduction to the relevant taxpayer in relation to that year of income;

determine that that amount or that part of that amount, as the case may be, should have been allowed or shall be allowable, as the case may be, as a deduction to the relevant taxpayer in relation to that year of income;

and the Commissioner shall take such action as the Commissioner considers necessary to give effect to any such determination.

63    As has been mentioned, Subdivision 815-A has a similar structure whereby a determination may be made which negates a pricing benefit, but the Commissioner may, by further determination, adjust the consequences of that determination if the Commissioner considers that 'it is fair and reasonable' that the relevant amount in the income year be so adjusted: s 815-35. However, as has been explained, Subdivision 815-A operates in a different manner. It is concerned with the amount of profits that have accrued to an entity in an income year by reason of Non-Independence Conditions and determining the taxable income, tax loss or net capital loss in that income year. Its focus is upon adjusting the taxation consequences to negate the taxation benefit in that income year.

64    As observed by Allsop CJ in Chevron at [15] in considering the terms of Division 13, 'the concept of the consideration that might reasonably be expected to have been given or agreed to be given had the agreement been reached between independent parties dealing at arm's length involves a suppressed premise or assumption that, in the real world, commercial parties could (and so would) enter into such a transaction such that an arm's length consideration can be hypothesised'. That is to say, it is not concerned with considering whether taxation should be imposed on the basis of a different transaction with different timing consequences and, therefore, effects in different income years. Similarly, as has been explained, Subdivision 815-A is concerned with how an arm's length dealing may have affected the profits (and consequently the taxable income, tax loss or net capital loss) in an income year.

65    STAI sought to rely upon statements concerning the application of the provisions in Division 13 that were made by Middleton J in SNF (Australia) Pty Ltd v Commissioner of Taxation [2010] FCA 635 at [53]-[54] (said to have been approved on appeal). At that point, his Honour was dealing with a submission by the Commissioner to the effect that the taxpayer could not succeed in demonstrating that the prices paid were arm's length prices on the basis that the prices paid by the taxpayer did not exceed prices paid by unrelated parties in similar circumstances 'if the taxpayer paid prices that generated losses year after year'. As to that submission, Middleton J reasoned as follows:

… If losses occurred over many years, this may give rise to a close examination as to whether the prices for the acquisition of the goods were inflated. However, if after analysis and evaluation looking at the evidence and submissions made by the parties, the Court comes to the view that the prices were equivalent to or lower than the arm's length consideration, sustained loses even over a substantial period are not otherwise determinative against the taxpayer in applying Div 13. As I have indicated, genuine losses may be sustained for many reasons, and not necessarily because the transfer prices of the goods were artificially inflated. In any event, the sole enquiry is to consider the arm's length consideration, assuming (as in this proceeding) the other objective criteria in s 136AD are satisfied.

66    His Honour was there concerned with the evidentiary significance of the fact that prices were set at a level that resulted in sustained losses over many years. His Honour was not concerned with whether the determination as to whether particular consideration that was relevant to the assessment of taxation in a particular income year was arm's length consideration had to be undertaken over some form of extended period. The reasoning is not directed to the timing issue that arises in the present case.

Relevance of differences between Subdivision 815-A and Division 13

67    Both STAI and the Commissioner approached the appeal on the basis that there was no identified reason why Division 13 did not apply if it was determined that Subdivision 815-A applied, and vice versa. Reference was made to the theoretical possibility of a different outcome in a different factual circumstance by reason of the difference in the statutory concepts. However, neither party identified how such a difference might arise in the present case.

68    For the most part, the submissions for STAI were presented through the lens of Subdivision 815-A. However, those submissions were presented on the basis that even though Subdivision 815-A and Division 13 impose different tests 'the object of each is the same and [STAI] will in either case succeed if the amount of interest actually paid over the 10-year period was equal to or less than that which might be expected to have been paid between independent parties in similar circumstances'. Accordingly, STAI advanced its 10 year analysis as the basis for why the assessments were excessive irrespective of whether it was the determinations under Subdivision 815-A or Division 13 that were relied upon. As has been explained, that approach was fundamentally flawed.

The history of the relevant tax assessments

69    For present purposes, the relevant history of the tax assessments the subject of the appeals that were dismissed by the primary judge may be shortly stated.

70    The course of events commenced with the making of determinations by the Commissioner in October 2016 relating to STAI in respect of the 2010, 2011, 2012 and 2013 tax years. The primary judge set out examples of the determinations made by the Commissioner under Subdivision 815-A and Division 13 at PJ[102]-[103]. It is sufficient to note at this point that each of the determinations under Subdivision 815-A specified an increase in the taxable income for a particular income tax year and each of the determinations under Division 13 determined the arm's length consideration in respect of acquisition of property (that is the interest payable in respect of the loan provided under the LNIA) for a particular income tax year (on the express basis that it was not practicable for the Commissioner to ascertain the arm's length consideration).

71    On 28 October 2016, the Commissioner issued notices of amended assessments for each of the 2011, 2012 and 2013 tax years. There was no amended assessment issued for the 2010 tax year because the determinations for that year brought to account carry forward losses and did not result in STAI having taxable income because it remained in a loss position for that year.

72    In December 2016, STAI lodged objections against the amended assessments.

73    In September 2019, the Commissioner disallowed STAI's objections. STAI appealed to this Court. On the appeal, STAI carried the burden of demonstrating that the amended assessments were excessive or otherwise incorrect. The Commissioner did not have to justify the assessments nor the process by which they were issued. The Court was required to address a particular question and does so on the basis of material received in the statutory appeal. The Commissioner was required to respond to the case as advanced by the taxpayer. As to these matters: see Bosanac v Commissioner of Taxation [2019] FCAFC 116; (2019) 267 FCR 169 at [47]-[48].

74    Before the primary judge, the Commissioner advanced an affirmative answer to STAI's appeal which involved three alternatives. STAI maintained that each of the alternatives departed from the reasons that had been given for disallowing the objections. The Commissioner did not dispute that proposition but maintained that the question on appeal was whether the amended assessments were excessive, not whether there had been deliberative error in the basis for the issue of the amended assessments or the reasons for disallowing STAI's objections.

75    STAI maintains that before the primary judge the Commissioner could not depart from the reasoning that provided the foundation for each of the determinations. It will be necessary in due course to consider this aspect of the grounds advanced by STAI in the present appeal.

The nature of the expert evidence relied upon by STAI before the primary judge

76    As has been mentioned, before the primary judge STAI contended that its actual cost of borrowing over the whole 10 year term was significantly less than that which a party in STAI's position might be expected to have paid to an independent party acting wholly independently so as to achieve the same cash flow advantages as STAI actually achieved. It relied upon expert evidence to support a contention that the 'effective credit spread' determined by reference to the amount of interest that STAI actually paid over the term of the LNIA was lower than the credit spread that might reasonably be expected to have been agreed in an arm's length debt capital markets transaction between independent parties at the time of entry into the LNIA. The primary judge summarised STAI's case in those terms at PJ[16] being terms that were not criticised on appeal.

77    The Commissioner's case before the primary judge involved three alternative scenarios or models. For each of those alternatives there were said to be three periods of relevance, namely (a) when the initial terms of the LNIA applied; (b) when the LNIA as amended by the second amendment applied; and (c) when the LNIA as amended by the third amendment applied. (No party contended that the first amendment materially altered the position as between SAI and STAI).

78    The Commissioner's position was that in order to apply each of Subdivision 815-A and Division 13 it was necessary to look at the circumstances at the relevant time, being when the amount of profits did not accrue (by reason of the commercial and financial relations between the parties), in the case of Subdivision 815-A, or when the consideration was agreed or given, in the case of Division 13. On the Commissioner's case, there was no transfer pricing benefit (for the purposes of Subdivision 815-A) and the consideration was nil (for the purposes of Division 13) in the first of the three periods because no interest was paid at that time. As to the time when interest accrued and was paid in each of the second and third periods, the relevant amount was the difference between the interest that was actually paid in the income year and the amount that would have been incurred for that period if the interest for that income year had been determined by adjusting for the effect of the Non-Independence Conditions (in the case of Subdivision 815-A) or on an arm's length basis (in the case of Division 13). For the purposes of the Commissioner's approach, it appeared that the amount may have been incurred in an income year on an accruals basis or on the basis of the amount paid in that year, whichever was applicable.

79    As to the way in which that arm's length basis should be determined, the Commissioner had three alternatives. They were described as (a) the without bridge model (which had both a 'with' and 'without' guarantee calculation); (b) a no amendment model (which assumed that the original LNIA terms remained in place over the life of the financing); and (c) a no third amendment model (which assumed that after 31 March 2003 the LNIA as amended by the terms of the second amendment continued to apply and there was no third amendment to the LNIA). The alternatives were described by the primary judge at PJ[292]. On the no third amendment model, the assessments would be excessive only in part.

80    The terminology 'without bridge' was adopted because the expert evidence relied upon by the Commissioner had included an analysis based upon STAI initially obtaining short term or bridging finance before replacing that borrowing with long term debt. Criticism of that approach was accepted and further calculations were undertaken on a 'without bridge' approach. It was those 'without bridge' calculations that were relied upon before the primary judge as the basis for the Commissioner's principal case.

81    Before the primary judge, STAI and the Commissioner each relied upon expert evidence. Broadly speaking, the expert evidence addressed two matters. First, the appropriate credit rating to be applied to STAI as the borrower. Second, the interest that STAI would have been required to pay if it had raised the $5.2 billion amount at the time of entry into the LNIA. The credit rating was one of the inputs used by the experts in undertaking an analysis of a hypothetical debt capital market borrowing by STAI.

82    As to the credit rating, evidence was given by Dr Chambers (for STAI) and Mr Weiss (for the Commissioner). The primary judge preferred the analysis of Dr Chambers: PJ[223]-[228]. No issue arises in the appeal as to that part of the reasoning by the primary judge.

83    As to the debt capital market analysis, STAI relied upon the evidence of Mr Chigas. The Commissioner relied upon evidence given by Mr Johnson. They expressed different views concerning the credit spreads that would be likely to have applied if STAI had borrowed the $5.2 billion in the debt capital market by a US bond issue.

84    Mr Chigas prepared a report in which he presented his opinion as to the 'effective credit spread' represented by the actual borrowings and payments made by STAI over the whole term of the LNIA. Relevantly for present purposes, the effective credit spread is a measure of the difference in yield (interest paid to the lender) between a benchmark debt instrument on the US debt capital market and another transaction (in the present case, the LNIA).

85    Mr Chigas calculated the effective credit spread on the actual borrowings and actual payments over the whole 10 year term of the borrowing based on when interest payments were actually made by STAI as being 'approximately 144 bps [basis points]'. One basis point is equivalent to 0.01%.

86    The manner in which he undertook that calculation was described by the primary judge in terms that were not disputed in the appeal: PJ[247]-[257]. The calculated 'effective credit spread' was not a market based figure. Rather, it was a figure that was derived from the actual payments that had been made by STAI over the life of the STAI. Conceptually it involved the derivation of a credit spread figure that was equivalent to the margin above a base rate represented by the interest payments actually made by STAI under the terms of the LNIA.

87    Mr Chigas then expressed the opinion that a borrowing by STAI (allowing for its circumstances and the nature of its business) in a debt capital market transaction with deferred and capitalised interest over 10 years would have been expected to be concluded on the basis of a credit spread of 400 bps.

88    A significant part of this further analysis by Mr Chigas concerned the degree of uplift in the credit quality of STAI, if any, that would have been allowed by lenders in the debt capital market for the implicit support available to STAI by reason that both STAI and SAI were ultimately 100% owned by SingTel, a Singapore-resident publicly listed company. An issue also arose as to whether a company in the position of SAI acting independently in financing its acquisition of shares in SOPL would have sought a guarantee from its parent in order to enhance its credit rating so as to take advantage of the lower interest rates at which a parent like SingTel could access funds - and, if so, whether a company like SingTel, acting rationally, would have provided such a guarantee (and whether it would have charged a fee for doing so).

89    Mr Chigas gave an alternative spread of 205 to 215 if it was assumed that STAI had gone to the debt capital market to raise funds with the support of a parent guarantee from SingTel with no guarantee fee, and 300 to 360 bps with the support of a parent guarantee with a guarantee fee payable. The difference between the two figures represented his assessment of the likely guarantee fee.

90    The primary judge described these aspects of Mr Chigas' analysis in terms that were not disputed in the appeal: PJ[258]-[267].

91    One of the inputs into the analysis undertaken by Mr Chigas to determine what he described as the effective credit spread for the total interest actually paid by STAI over the term of the LNIA was an interest rate of 5.6143% which was used as the base rate for the period from 1 April 2009 until the end of the term of the loan. For the period prior to that Mr Chigas used the bank bill swap rate (reset annually as contemplated by the initial terms of the LNIA) as the base rate. As to the use of the fixed rate of 5.6143% (rather than the one year bank bill swap rate), the primary judge explained the position concerning its use in the following terms (at PJ[254]):

Senior counsel for STAI said that the figure of 5.6143% came from an Australian Taxation Office schedule of calculations, and that it was a base rate that STAI had agreed to use. Senior counsel for the Commissioner then said that it is a figure which the Commissioner had accepted as a fixed rate from 1 April 2009. I therefore proceed on the basis that there is no issue about Mr Chigas's use of this figure in his calculations.

92    The circumstances relating to the rate of 5.6143% and its significance, if any, for the appeal were the subject of competing submissions. Those matters are addressed below when dealing with those submissions.

93    Finally, in considering the hypothetical interest rate that would have been agreed, Mr Chigas proceeded on the basis that the hypothetical dealing would have had the characteristics of the LNIA as amended by the second amendment: PJ[236]. It was on that basis that Mr Chigas concluded that the overall structure of the LNIA is consistent with a bond issued in the debt capital market: PJ[237]. That is to say, in broad terms, he equated the LNIA (as amended by the second amendment) with a bond issue.

94    Mr Johnson undertook an analysis which was based upon his opinion as to the financing arrangements that were likely to have been put in place if STAI had raised its finance on an arm's length basis, but still in circumstances where it was ultimately a subsidiary of SingTel. In his view, STAI could have borrowed initially on a short-term or bridging basis on attractive terms from Australian financiers, deferring any long-term borrowing. In his opinion, such a course would have been followed where short term funding could have been obtained at lower cost from Australian financiers at the time and STAI would have expected to obtain more favourable terms for long term debt funding by reason of its then expected improvement in its financial position.

95    As to that long term borrowing, Mr Johnson expressed the opinion that, based upon the support of SingTel and the opinion of Mr Weiss as to the strength of SingTel's credit rating, it could have been arranged in the debt capital market on more favourable interest terms than the interest rate originally agreed under the terms of the LNIA. As to the credit rating analysis, as has been noted, the primary judge preferred the evidence of Mr Chigas over that of Mr Weiss.

96    Further, Mr Johnson's opinion was that STAI would have taken a 50/50 fixed and floating position at the time of the second amendment and, therefore, would not have been interested in considering whether to fix the interest rate at the time of the third amendment during the Global Financial Crisis.

97    Mr Johnson, like Mr Chigas, expressed a view as to the extent of the additional profits that should have accrued to STAI over the 10 year term of the borrowing had it borrowed on the terms outlined in his opinion rather than the actual terms of the STAI as amended by the second and third amendments.

98    However, Mr Johnson also produced further calculations that were deployed by the Commissioner in submissions. They corresponded to the three alternative cases advanced by the Commissioner before the primary judge (as described above), namely (a) the without bridge model (with and without SingTel guarantee); (b) the no amendment model; and (c) the no third amendment model. Though the term model was used to describe each alternative, Mr Johnson simply produced the calculations. The three alternatives were then advanced by the Commissioner by way of submissions based upon the evidence more generally.

The reasoning pathway of the primary judge

99    The primary judge described the primary case advanced by STAI in the following terms:

(1)    'The financial position of STAI in June 2002 was that the $5.2 billion borrowing had to take into account the fact that, due to the capital expenditure programme SOPL had embarked upon, its cash flow was not expected to be sufficient to meet interest payments and other expenditure for the short to medium term': PJ[157(a)].

(2)    The LNIA (as amended) provided a mechanism by which to defer interest payments by STAI until funds were expected to be available and did so in a manner suited to the parent and subsidiary relationship between SAI and STAI as the parties to the LNIA: PJ[157(b)].

(3)    The correct and reasonable comparator for the terms of the LNIA was a commercial loan taken out in June 2002 'under which interest could be deferred and capitalised to suit the borrower's expected cash flows': PJ[157(c)].

(4)    'No amount of profits might have been expected to accrue to STAI' in the relevant years 'because the interest rate which might be expected to be payable for a loan such as this might be expected to be no less that the effective interest rate under the LNIA': PJ[157(d)].

(5)    '…the most economical, and likely only, source of debt in the amount of $5.2 billion for a term of 10 years for a party in the position of STAI was the [United States debt capital market] in the form of a new issue transaction with a USD:AUD Cross Currency Interest Rate Swap (CCIRS)': PJ[158(a)].

(6)    'The terms and economic effect of the LNIA…are substantially similar and of similar commercial effect to the terms that would be expected in a [debt capital market] transaction between independent parties where the borrower sought to defer interest payments to suit its expected cash flow': PJ[158(b)].

(7)    'STAI's primary case is that the actual cost of the borrowing to STAI under the LNIA was not greater (in fact, it was significantly less) than the cost that a party in STAI's position might be expected to have paid to an independent party acting wholly independently so as to achieve the same case flow advantages which STAI actually achieved': PJ[158(c)].

100    The primary judge found that the principal difficulty with STAI's case was the difference between the actual transaction and the debt capital market transaction by way of bond issue that was used by STAI's expert to reach the conclusion that STAI's borrowing cost was significantly less than might be expected to have been incurred through an arm's length dealing: PJ[17]. His Honour found that the terms of the hypothetical bond issue 'departs too far from the actual transaction and the characteristics of the parties to that transaction': PJ[17], [343].

101    His Honour also found that there was a further difficulty with STAI's approach because it was undertaken with hindsight rather than by considering what parties in the position of SAI and STAI, dealing independently with each other, might be expected to have agreed at the outset and at the time of each relevant amendment to the LNIA: PJ[17], [343].

102    In the result, the primary judge found that the commercial and financial operations that pertained as between STAI and SAI (that is the Non-Independence Conditions) differed from those which may be expected to operate between independent enterprises dealing wholly independently with each other: PJ[18(a)]. His Honour also found that it was a reliable hypothesis that independent parties in their position would have agreed to a variable interest rate for the borrowing at the same rate as was actually agreed by the terms of the LNIA. However, they would not have agreed to the variations to the interest rate that were agreed at the time of the second and third amendments to the terms of financing: PJ[18(b)]. Based on those ultimate findings, his Honour concluded that for each of the taxation years in issue, but for the conditions which differed from those expected to operate between independent enterprises dealing independently, an amount of profits that would have accrued to STAI did not so accrue.

103    That is to say, ultimately, the primary judge accepted the second alternative advanced by the Commissioner being the 'no amendment' model: PJ[355].

104    Based upon those conclusions, his Honour found that 'STAI has not demonstrated that the amended assessments are excessive': PJ[19], see also PJ[355].

105    After recounting the history and directing attention to the key legislative provisions, the primary judge considered the applicable principles as they had emerged from two judgments of this Court, namely Chevron and Commissioner of Taxation v Glencore Investment Pty Ltd [2020] FCAFC 187; (2020) 281 FCR 219. His Honour emphasised passages in those authorities concerning the approach to be taken in formulating a hypothetical to be used for undertaking the required comparison between the actual circumstances and those which have the requisite arm's length characteristics. His Honour also referred to passages concerned with the formulation of the amount of the substituted arm's length consideration for the purposes of Division 13 and the amount of profits that might have been expected to accrue for the purposes of Subdivision 815-A.

106    After a consideration of those two decisions his Honour summarised some key propositions that emerge from them relating to Subdivision 815-A. Those propositions, which were not meaningfully disputed in the appeal, were to the following effect (at PJ[156], noting that the following excludes the source of each of the propositions as recorded by the primary judge):

(1)    in relation to s 815-15(1)(b) and whether (in the language of the relevant Treaty) the 'commercial and financial relations' between the two enterprises 'differ from those that might be expected to operate between independent enterprises dealing wholly independently with one another', the identification of those conditions permits a broad and wide-ranging inquiry into relations existing between the enterprises concerned;

(2)    in relation to the causal test in s 815-15(1)(c) and whether, but for the differing conditions, an amount of profits might have been expected to accrue to the taxpaying entity, the test 'is a flexible comparative analysis that gives weight, but not irredeemable flexibility, to the form of transaction actually entered into between the associated enterprises';

(3)    the form of the transaction 'may, to a degree, be altered if it is necessary to do so to permit the transaction to be analysed through the lens of mutually independent parties';

(4)    the required comparison 'will generally require that the parties in the hypothetical will generally have the characteristics and attributes of the actual enterprises in question' but where 'the conditions operating between them were between independent enterprises dealing wholly independently with each other'; and

(5)    the non-price terms of a transaction may be the subject of substitution in the comparative analysis required by s 815-15(1)(c).

107    In respect of the expert evidence of Mr Chigas, the primary judge made the following findings of relevance to the appeal:

(1)    Mr Chigas was asked to identify any conditions which operated between SAI and STAI 'in respect of the LNIA' which differed from those which might be expected to operate between independent enterprises dealing at arm's length in substantially similar circumstances, which was an approach which did not conform to the statutory test under Subdivision 815-A. Instead, the statutory test required the identification of conditions that operated between the two enterprises 'in their commercial or financial relations' which was a broader enquiry: PJ[233]-[234].

(2)    Mr Chigas undertook a comparison between the terms of LNIA and the terms of a debt capital market bond issue but did so by reference to 'the LNIA as amended by the Second Amendment, as distinct from the LNIA as originally entered into' (original emphasis): PJ[235]-[236].

(3)    Mr Chigas accepted that under the original LNIA interest could be deferred and capitalised whereas under the LNIA as amended by the second amendment the obligation to pay interest was contingent on reaching the benchmark. As to that difference, Mr Chigas accepted in relation to a bond issue to raise $5.2 billion 'he would not expect to see the lender being prepared to accept interest on a contingency or related to the financial performance of the issuer': PJ[243]-[244].

(4)    Mr Chigas accepted that he had not addressed the question whether an arm's length borrower in the position of STAI would be expected to agree, at the time of the second amendment, to pay a premium amount of 4.552% for the deferral of interest: PJ[245].

(5)    As to the third amendment which substituted a fixed rate, Mr Chigas accepted that at the time the fixed rate was agreed 'the 1 year [bank bill swap] rate [being the rate used as the benchmark to calculate the credit spread] was dropping quite significantly': PJ[246].

(6)    Mr Chigas' analysis could be summarised as follows (see PJ[263]):

(a)    the effective rate under the actual LNIA was a credit spread of 144 bps;

(b)    a [debt capital market] transaction with a parent guarantee from SingTel, with no guarantee fee payable by STAI, would have had a credit spread of 205 to 215 bps;

(c)    a [debt capital market] transaction with a parent guarantee from SingTel, with a guarantee fee payable by STAI, would have had a total credit spread of 300 to 360 bps; and

(d)    a [debt capital market] transaction with no parent guarantee would have had a credit spread of 400 bps.

(original emphasis)

(7)    Mr Chigas concluded that the interest rate condition (whereby the obligation to pay interest was conditional upon reaching a benchmark) would not have produced a different spread if there was an operating guarantee provided by SOPL and that a 400 bps spread (without a parent guarantee) 'is reasonable for the market conditions at the time and assuming a BBB-/Baa3 credit rating': PJ[265];

(8)    The analysis by Mr Chigas produced a result whereby the difference between the interest incurred under the LNIA and the interest under what he identified as the 'market scenario' (a credit spread of 400 bps) was approximately $2.8 billion: PJ[267].

108    The primary judge recorded that by a joint report prepared by Mr Chigas and Mr Johnson, a number of matters were agreed. They included the following (see PJ[290]):

(1)    that the provisions 'that gave [STAI] the ability to require payment of interest earned at any time of its choosing [until the second amendment, thereafter once a benchmark event has occurred] has not been observed by [either expert] in a debt capital markets transaction';

(2)    that 'STAI's most economic and likely only source for the quantum of $5.2 billion of third-party, long term funds would have been a US [debt capital market] new issue transaction coupled with a CCIRS; and

(3)    'Optus or STAI could have raised the equivalent of $5.2 billion in the US [debt capital market] without a SingTel Guarantee'.

109    The primary judge characterised the approach of Mr Chigas as one which priced the LNIA as if it were a debt capital market bond issue by STAI in June 2002 and then compared this 'arm's length price' with the actual effective price of the LNIA. Having done so, Mr Chigas concluded that the interest actually paid under the LNIA was less than the 'arm's length price': PJ[295].

110    The primary judge made three important findings concerning Mr Chigas' evidence. First, his comparison of conditions only addressed conditions in the LNIA as distinct from conditions operating between STAI and SAI 'in their commercial or financial relations': PJ[296(a)]. This was significant for his Honour's conclusion that the analysis failed to account for the fact that the LNIA recorded the terms of a very large vendor financing provided to a subsidiary of SingTel to acquire the shareholding in SOPL.

111    Second, his Honour had difficulty in accepting the view expressed by Mr Chigas that the overall structure of the LNIA was consistent with a bond issued in the debt capital market: PJ[296(b)]. In particular, his Honour had 'difficulty in accepting that the Benchmark concept and the associated clauses are consistent with a [debt capital market] bond issue. Further, the terms of the LNIA regarding the giving of Variation Notices do not appear to be consistent with a [debt capital market] bond issue'. As to these matters, the primary judge found that they 'created significant uncertainty for the parties, which is unlikely to have been agreed in a [debt capital market] bond issue'. This too was significant for any comparison that may be made between the overall interest paid under the terms of the LNIA and the terms of a bond issue on the basis that the timing of those payments was disregarded. A significant feature of the LNIA, as originally agreed, was that it allowed considerable flexibility as to when interest payments may be made of a kind that the primary judge did not accept allowed for comparison with a debt capital market bond issue.

112    Third, his Honour found that 'by addressing the LNIA, as amended, Mr Chigas has not directly addressed whether the changes effected by the Second Amendment might be expected to have been made if the parties were independent' (original emphasis): PJ[296(c)]. In that regard, his Honour gave the example that Mr Chigas 'has not directly considered whether an independent SAI might be expected to have agreed to have relieve an independent STAI of the obligation to pay approximately $286 million in accrued interest'.

113    As to the application of Subdivision 815-A, the primary judge reasoned in the following way.

114    There were conditions that operated between SAI and STAI at the time of entry into the LNIA and throughout the term of the LNIA which differed from those that might be expected to operate between independent enterprises dealing wholly independently with each other: PJ[301]-[316]. As part of that reasoning, his Honour observed that the Commissioner did not suggest that the interest rate that was applicable under the original terms of the LNIA was other than arm's length consideration: PJ[308].

115    In determining, but for the conditions that differed from those that may be expected to operate, an amount of profits that might be expected to have accrued to STAI that did not accrue to STAI, it was necessary to form 'a reliable hypothesis based on probative material as to what independent parties in the positions of SAI and STAI might have been expected to have done': PJ[318]. In forming that hypothesis SAI and STAI were to be given their actual characteristics and attributes, particularly that STAI is a member of a multinational corporate group like the SingTel group and the holding company of an operating subsidiary like SOPL: PJ[319]-[321].

116    As to that reliable hypothesis, the primary judge found that SAI and STAI would have been expected to act in the following way (PJ[322], reasoning in following paragraphs):

(1)    they would have agreed in June 2002 to the interest rate that was actually agreed in the original LNIA;

(2)    they would have agreed in June 2002 that interest could be deferred and capitalised;

(3)    they would not have agreed to make the changes contained in the second amendment, particularly the introduction of the benchmark terms and the agreed premium of 4.552%; and

(4)    at the time of the third amendment, they would not have agreed to change the interest rate from the variable rate set annually by reference to the bank bill swap rate as originally agreed to a fixed amount of 6.835% plus 1% (and retaining the uplift to cover withholding tax), being an applicable rate of 13.2575% (see PJ[95] as to this rate).

117    As to Division 13, his Honour found that STAI had acquired property under an international agreement and that 'the interest actually paid by STAI to SAI under the LNIA during each year of the life of the LNIA' (emphasis added) was consideration actually given or agreed to be given for the property: PJ[351(b)]. For the reasons already given in relation to Subdivision 815-A, his Honour concluded that the arm's length rate 'for each year of the life of the LNIA' (emphasis added) was to be determined on the basis that it was the interest rate originally agreed under the LNIA: PJ[351(a)]. Those findings were made by reference to the calculations that had been undertaken by Mr Johnson which were produced as Annexure D to the reasons. Those calculations were made for each tax year of STAI. They identified the arm's length interest income that would have been earned for each tax year.

118    On that basis, the primary judge concluded that 'consideration equal to the arm's length consideration in respect of the acquisition is deemed to be the consideration given or agreed to be given by STAI in respect of the acquisition': PJ[352]. Although the final expression of the conclusion does not refer to any particular time period, in the context of the preceding references to each year of the life of the LNIA and the contents of Annexure D, his Honour found that for each of the relevant tax years the subject of the assessments, there was an identified amount which was the arm's length amount for the purposes of Division 13 and, on that basis, the amended assessments were not shown to be excessive to the extent that they were based upon Division 13: PJ[355].

119    The primary judge also rejected a submission by STAI to the effect that the Commissioner could not defend the amended assessments in the Court on a basis that was inconsistent with his determinations: PJ[353].

STAI's grounds of appeal

120    The case for STAI on appeal (as before the primary judge) was constructed on the primary proposition that if the amount of interest actually paid over the 10 year period was equal to or less than that which might be expected to have been paid between independent parties in similar circumstances over the same period then neither Subdivision 815-A nor Division 13 applied so as to impose additional taxation liability upon STAI. It was a case that rested upon the analysis of Mr Chigas as to the 'effective credit spread' (which he had calculated based upon what was actually paid by STAI) when compared with the amount of interest that, in the opinion of Mr Chigas, may have been paid for finance over the same 10 year period as the LNIA if the required funds had been raised by way of bond issue by a party dealing at arm's length in the United States' debt capital market with an associated agreement to cover USD to AUD exchange rate liabilities.

121    STAI's case depended upon the debt capital market bond issue analysis of Mr Chigas being accepted as demonstrating the relevant arm's length consideration for the purposes of Division 13 and Subdivision 815-A. It also depended upon a comparison as between that amount for the 10 year term and the effective credit spread analysis undertaken by Mr Chigas for that same period. However, as has been explained, the primary judge found that the interest rate payable under the terms of LNIA as originally agreed was to be used in determining the relevant arm's length consideration in each of the relevant income years.

122    The focus by STAI upon the consideration payable over the whole of the 10 year term meant that STAI ignored timing issues that arose from when interest was paid. It treated the case as one in which the relevant inquiry concerned the total amount of interest over the whole term. As has been explained that approach was fundamentally flawed. It failed to engage with an important distinction when it came to the amount of interest, being the distinction between:

(1)    the extent to which the total interest paid over the 10 year term of the LNIA exceeded that which would have been payable on an arm's length basis; and

(2)    the extent to which the timing of the accrual and payment of that interest (with no interest accruing or being paid in the first five years of the LNIA) resulted in higher interest expense amounts being paid in the relevant years of assessment than would have been the case on an arm's length basis.

123    The Commissioner maintained that STAI had failed to demonstrate that the assessments in the relevant income years were excessive. The Commissioner also disputed other aspects of STAI's case as to the arm's length interest that STAI contended would have been payable over the 10 year term of the LNIA.

124    Significantly, Mr Chigas did not seek to demonstrate that the amount of interest that accrued and was paid by STAI in each of the years of assessment equated to an amount of interest that would have accrued in respect of that year for a borrowing of $5.2 billion in an arm's length dealing in the circumstances that pertained to the actual borrowing.

125    STAI prepared its tax returns on the basis that interest both accrued and was paid in later years. As has been explained, on a proper construction of the statutory provisions, in order to support that position, it needed to demonstrate that an arm's length dealing between SAI and STAI would have resulted in interest obligations accruing when they did (as a result of the benchmark arrangement that was introduced by the second amendment and the fixing of the interest rate by the third amendment) and in the amounts that they did in the relevant income years.

126    As to that aspect, the primary judge found (at [244]-[245]):

Mr Chigas accepted during cross-examination that all the examples that he had given in his reports and the [joint report of the experts concerning debt capital markets] of bonds issued in the [debt capital market] where interest was deferred were bonds where interest was payable 'come what may, albeit it might be deferred in terms of when the investor receives its money'. Mr Chigas was asked whether he would expect to see in the [debt capital market] a bond where the payment of interest is contingent upon the issuer meeting certain financial benchmarks. He responded that they are not common, but they exist. In relation to a bond to raise $5.2 billion, he accepted that he would not expect to see the lender being prepared to accept interest on a contingency or related to the financial performance of the issuer.

In cross-examination, Mr Chigas accepted that he had not addressed in any of his reports whether an arm's length borrower in the position of STAI would be expected to agree, at the time of the Second Amendment, to pay a premium of the amount of 4.552% for the deferral of interest under the Second Amendment.

127    By its written submissions, STAI consolidated its appeal grounds into seven alleged key errors by the primary judge, namely:

(1)    rejecting Mr Chigas' evidence as to the arm's length interest rate (based on debt capital market issues to similarly rated telecoms within 14 days of entry into the LNIA) and the adoption of the interest rate as agreed in the initial LNIA as the appropriate rate;

(2)    concluding that independent parties might not reasonably have chosen to change from a floating base rate to a fixed base rate for the last three years of the term of the LNIA (notwithstanding that the agreed base rate was the market rate at the time for swapping from a 1 year bank bill swap rate to a three year fixed base rate);

(3)    concluding that the comparative hypothetical should be based on a non-arm's length guarantee provided by a hypothetical parent of STAI;

(4)    finding that interest should be capitalised on an annual basis rather than a quarterly or half yearly basis;

(5)    concluding that the second amendment was irrational;

(6)    reaching an incorrect conclusion as to the legal effect of the determinations by the Commissioner; and

(7)    failing to have regard to losses arising in the years before the relevant years of assessment.

128    The Commissioner responded to the seven alleged key errors as formulated.

129    In those circumstances, these reasons treat the seven alleged key errors (and the submissions advanced to support them) as capturing all relevant matters raised by the extensive grounds of appeal.

130    The Commissioner also advanced three matters by way of notice of contention. They were to the following effect:

(1)    the primary judge ought to have found, further or in the alternative, that each of the assessments for the 2011, 2012 and 2013 years was not excessive on the basis of Mr Johnson's guaranteed 'without bridge' model (that is to say, the primary judge ought to have upheld the principal case advanced by the Commissioner);

(2)    the primary judge ought to have found, in the further alternative, that STAI had failed to discharge its burden of proving that each of the assessments was excessive; and

(3)    the primary judge ought to have found, in the further alternative, that the Commissioner's no third amendment model should have been upheld and, on that basis, determined that each of the assessments was excessive only in part.

131    The discursive manner in which the appeal was presented for STAI makes the isolation of particular respects in which the primary judge was said to have erred somewhat difficult, especially as to respects in which it was said that the primary judge was said to have overlooked or ignored aspects of the evidence or should have made different factual findings. In the reasons that follow the main points that were developed are sought to be identified and addressed in dealing with each of the alleged errors.

132    Before dealing with the specific allegations of error, it is necessary to address the extent to which there was a divergence as between STAI and the Commissioner concerning the way in which the comparison with the hypothetical was to be undertaken for the purposes of Subdivision 815-A and Division 13.

Conceptual differences between the parties on appeal concerning the way the comparison with the hypothetical was to be undertaken

133    As has been observed, the primary judge summarised the state of the authorities as to the formulation of the hypothetical that is required for the purposes of applying Subdivision 815-A and Division 13. There was no substantive dispute as to that summary.

134    To those propositions, the following may be added:

(1)    the function of the reliable hypothesis is to identify a sufficiently reliable substitute for that which actually occurred when the Non-Independence Conditions were operating (Subdivision 815-A) or the parties were not dealing at arm's length (Division 13) to determine the reasonably expected arm's length consideration for the type of dealing that occurred;

(2)    formation of the reliable hypothesis requires a common sense approach by which the hypothesis is made to work in a reliable way to determine the arm's length consideration for the type of dealing that occurred with due regard to the position of the taxpayer and the commercial context for the actual dealing;

(3)    formation of the reliable hypothesis requires a focus upon the characteristics of the dealing itself (not the subjective or special concerns of the taxpayer and other parties involved in the dealing);

(4)    where relevant and appropriate to the task, the reliable hypothesis may include aspects of the corporate structure to which the taxpayer belongs and also the characteristics of the taxpayer and other parties where they would be relevant if the parties were dealing at arm's length (for example, where the reliable hypothesis would involve a dealing in market circumstances where the price would be affected by such characteristics);

(5)    for any dealing, there may be a range of arm's length outcomes that might reasonably be expected and if there are then each would be a reliable hypothesis;

(6)    where the appetite for risk of the parties may affect the arm's length consideration then there is no need for the hypothesis to be approached on the basis that profit will be preferred over prudence;

(7)    Subdivision 815-A and Division 13 each call for an inquiry into what might have been expected and not what would have been expected; and

(8)    the evaluative determination by the Court of what might reasonably be expected by way of arm's length consideration must be based upon the evidence that has been led bearing in mind that it is the taxpayer that bears the onus.

As to these additional propositions, see Commissioner of Taxation v SNF (Australia) Pty Ltd [2011] FCAFC 74; (2011) 193 FCR 149 at [98]-[99], [125] (Ryan, Jessup and Perram JJ); Chevron at [42]-[45] (Allsop CJ), [126]-[128], [156] (Pagone and Perram JJ agreeing); and Glencore at [167]-[188] (Middleton and Steward JJ), [271], [295]-[299] (Thawley J).

135    For STAI, emphasis was placed upon the need for the inquiry to be directed at ascertaining what might reasonably have been expected to be agreed or paid rather than upon any view as to what independent parties would have agreed. That is to say, it was enough for STAI to demonstrate that the amount of the profits that actually accrued (Subdivision 815-A) and the consideration actually given or agreed (Division 13) was within a range that may be observed if the same dealings had occurred on an arm's length basis.

136    The Commissioner accepted that the inquiry concerned what might have been agreed not what would have been agreed. In doing so, the Commissioner emphasised that a possibility was not enough. The Commissioner submitted that what is required is the demonstration by probative material of a sufficiently reliable prediction based upon evidence. That is to say, the use of the term 'might' is not to encompass matters advanced on the basis of conjecture. That submission should be accepted: Glencore at [184], approving reasoning of Pagone J in Chevron at [127].

137    As to the extent to which the hypothesis should disregard the actual circumstances between the parties (described by the parties as 'depersonalisation'), both parties accepted the statement by Allsop CJ in Chevron at [45] (approved in Glencore at [176]) as correctly stating the required approach, namely that the extent of depersonalisation depends upon what is appropriate to the task. It is an approach which requires regard to the essential commercial characteristics of the transaction, the extent to which there is a market in which transactions with those characteristics take place and the factors that influence price in that market, not 'the utter disembodiment of both parties from the circumstances of reality if one is seeking to understand … the consideration that would be given for acquiring a characterised loan from an independent lender': see the analysis by Allsop CJ in Chevron at [42]-[44]. The task concerns what may reasonably be expected as an arm's length outcome for a hypothetical transaction that is sufficiently close to the actual transaction to perform the required statutory comparison.

138    As to whether the terms of Subdivision 815-A and Division 13 required the formulation of a hypothetical only at the time of entry into the LNIA or also at the time of the relevant amendments, the logic of the case advanced by STAI was that there was only one hypothetical (namely, what would have been agreed at the outset). On STAI's case, it was a comparison between Mr Chigas' opinion as to the interest rate that would have been agreed for the whole borrowing and the total amount of interest that was ultimately paid that is required. It ignored the timing of the payments. It did not seek to support the actual dealings on the basis that they would be arm's length. In effect, it treated the actual dealings as if they were an agreement made at the outset to pay the amount of interest which ultimately came to be paid without regard to when. It did not analyse whether the terms of the LNIA (and the second and third amendments) were dealings that might have been reached if the Non-Independence Conditions did not apply (Subdivision 815-A) or the consideration paid by STAI exceeded arm's length consideration for the relevant acquisition of property (Division 13). Nor did it analyse what arm's length dealings might have been agreed at the time of the second and third amendments to address the commercial circumstances at that time.

139    However, the fact that there was a series of transactions, relevantly the initial LNIA, the second amendment to the LNIA and the third amendment to the LNIA, could not be ignored. As has been explained, what was required was a focus upon (a) the interest actually claimed by STAI by way of deduction on the basis that the amount related to the particular income year the subject of assessment when compared with (b) the interest that may be claimed by way of deduction in that same income year on the basis of a hypothetical by which the terms of borrowing were agreed with the Non-Independence Conditions not applying (Subdivision 815-A) or on an arm's length basis (Division 13). Therefore, the required comparison must have regard to the actual conditions that gave rise to the interest for which STAI claimed a deduction. Those conditions were not those that existed at the time that SAI and STAI entered into the LNIA. They were the conditions that existed at the time of the actual events that were said to justify the interest in the particular income year.

140    The 'no amendment' case for the Commissioner was founded on the contention that the relevant hypothetical was one in which neither the second nor third amendments would have been agreed and interest payments were made by STAI when they were actually made in the events which had occurred. The alternative 'without bridge' case for the Commissioner advanced a different hypothetical to that relied upon by STAI. As has been explained it involved a 50/50 split between fixed and floating interest over the term of the borrowing with the ability to defer and capitalise interest.

141    Finally, it was common ground that the relevant provisions did not depend for their application upon the existence of some tax avoidance or profit shifting motive or purpose. However, as was submitted for the Commissioner, if it was shown that there was no apparent commercial reason for the actual dealing (or particular terms of the actual dealing) then that would be a reason why that aspect would not be included in the reliable hypothesis.

Alleged Error (1): Did the primary judge err in rejecting Mr Chigas' evidence based upon debt capital market issues by similarly rated telecoms and finding that the interest rate as initially agreed in the LNIA was the appropriate rate for the hypothetical?

142    STAI advanced a number of submissions as to why the primary judge was allegedly in error in adopting for the hypothetical the rate as initially agreed between the parties in the LNIA, being the bank bill swap rate plus 1% (the Initial Rate) and in doing so on the basis of a parental guarantee at no cost to STAI rather than accepting certain parts of the evidence of Mr Chigas as establishing the appropriate rate for the hypothetical. Those submissions were based upon the following principal contentions:

(1)    the evidence relied upon by the primary judge as the basis for adopting the Initial Rate was of little, if any, probative value;

(2)    the Initial Rate was outside (that is to say, below) the outer limits of what was in issue as far as the experts were concerned;

(3)    the adoption of the debt capital market analysis to determine the arm's length rate was a 'common approach' as between the experts;

(4)    each of the experts acknowledged that the actual transaction involved vendor finance;

(5)    if SingTel had provided a guarantee then its rating would have been adversely affected which would have limited the extent to which the parent guarantee could have resulted in a more favourable interest rate for STAI; and

(6)    the evidence of the actual issue by similarly rated telecoms at about the time of entry into the LNIA as relied upon by Mr Chigas should not have been rejected.

143    STAI's case was that the primary judge should have found a rate within the limits of what was in issue between the experts and that if his Honour had done so then STAI would have demonstrated that the assessments are excessive. Like much of STAI's case, this submission assumed that the comparison to be made was over the life of the LNIA, not as between (a) the interest that was paid in each relevant income year; and (b) the interest that would have been paid in that income year based upon a reliable hypothesis as to what might have been expected if the parties were dealing at arm's length. Noting that significant point (which was also made by the primary judge at PJ[343]), the principal contentions advanced by STAI are addressed below.

The findings by the primary judge as to the interest rate for the reliable hypothesis

144    The relevant findings by the primary judge were expressed in the following terms (at PJ[322], [328]):

In my view, having regard to the facts and circumstances as described earlier in these reasons, a reliable hypothesis is that independent parties in the positions of SAI and STAI (and SingTel) might have been expected to have agreed in June 2002 that: the interest rate applicable to the loan notes would be the 1 year BBSW plus 1%, with the resulting amount grossed-up by 10/9 (that is, the same rate as was actually agreed in the original LNIA); interest under the loan notes could be deferred and capitalised; and, there would be a parent guarantee from a company like SingTel of the obligations of the company in the position of STAI. Further, having agreed to a transaction with these components in June 2002, a reliable hypothesis is that independent parties in the positions of SAI and STAI would not have agreed to make the changes contained in the Second Amendment. In particular, they would not have agreed to introduce the benchmark terms and add the Premium of 4.552%. Further, having agreed to a transaction with the components described above in June 2002, a reliable hypothesis is that independent parties in the positions of SAI and STAI would not have agreed to make the changes in the Third Amendment. That is, they would not have agreed to change the component of the interest rate that was the 1 year BBSW to a fixed amount of 6.835%. It follows that, in my view, in the hypothesis, the interest rate of the 1 year BBSW plus 1%, with the resulting amount grossed-up by 10/9, would have (or might be expected to have) continued through the whole life of the LNIA.

…

Insofar as STAI contends that, if the hypothesis includes the provision of a parent guarantee, it might be expected that a guarantee fee would be charged by the parent to the subsidiary, I do not accept that contention. This proposition was largely based on Mr Chigas's evidence. However, Mr Chigas's evidence on this point (see [261]-[262] above) is expressed in very general terms. Moreover, his evidence regarding the amount of any fee appears to be speculative. He states in his first report that a fee of 95.5 bps to 146 bps 'could have been' paid to SingTel. There does not appear to be any probative evidence to hypothesise that, assuming the provision of a parent guarantee, a guarantee fee might be expected to be charged. I note also that there is no evidence that SingTel charged a guarantee fee for guaranteeing the obligations of Optus Finance Pty Ltd under its $2 billion bank facility.

(original emphasis)

145    The earlier reasoning concerning the evidence of Mr Chigas to which his Honour referred was as follows (PJ[261]-[262]):

Mr Chigas considers the possibility that a guarantee fee might be charged for the provision of a guarantee. He expresses the following views at paragraph 125 of his first report:

A guarantee imposes a real liability and incremental costs on the guarantor. Therefore, it is reasonable that a guarantee fee is paid to the guarantor assuming an arm's length transaction. Fees for guarantees and credit support are commercially reasonable and commonly observed in the financial markets. In my experience guarantee fees are paid to a guarantor in many different types of transactions.

Mr Chigas calculates the potential fee STAI may have paid SingTel assuming an arm's length transaction, applying a 50% to 75% fee sharing methodology. He expresses the view at paragraph 129 that, had the LNIA been guaranteed by SingTel, a fee of 95.5 bps to 146 bps 'could have been' paid to SingTel. If this fee is incorporated, the resulting total credit spread to STAI is, in Mr Chigas's view, in the range of 300 bps to 360 bps.

146    As to the interest rate in particular, the primary judge reasoned by the following steps:

(1)    there is evidence that the Initial Rate was considered to reflect an arm's length consideration: PJ[329];

(2)    in a message summarising the historical basis and key commercial drivers for the acquisition of shares in SOPL by STAI and the LNIA, Ms Low, SingTel's Group Financial Controller said that introducing the debt via SAI rather than directly from SingTel allowed 'SingTel to more efficiently manage its interest cashflow in the future, without repatriating interest income to Singapore': PJ[303(c)], [305];

(3)    in the course of describing the 'key attributes' of the LNIA, Ms Low's message referred to '[a]n arm's length interest rate (grossed up to take into account interest withholding tax …) was included in the [LNIA] as provided by Optus Treasury': PJ[329];

(4)    an email between senior financial officers of SOPL sent three days before the LNIA was entered into, set out a proposed definition of the interest rate which matches the definition in the LNIA and stated (PJ[330]):

The 1% margin is based on several factors:

•    Our latest three year issue was in the range 0.35% to 0.4% p.a. (but is for a considerably shorter term). The 1 Year Swap Rate assumes interest is paid / compounded quarterly, therefore a 'zero coupon' rate would be higher than the reference rate, which I am taking into account in the 1% p.a. margin.

•    SingTel's 10 year Euro Bonds are currently trading at 0.99% p.a. above LIBOR (Source: report on Credit Spread Movement by ABN AMRO's International Bond Research Desk dated 25 June 2002).

(5)    it is 'notable that, in identifying what was considered to be an arm's length rate, regard was had to the margin applicable to SingTel's bonds' and 'the actual parties approached the matter of identifying an arm's length interest rate by reference to SingTel's credit rating' (original emphasis): PJ[331]; and

(6)    the expert evidence was not a basis to depart from the Initial Rate as the interest rate that might be expected between independent parties: PJ[332]-[334].

147    His Honour also found that 'independent parties in the positions of SAI and STAI might be expected to have agreed that interest could be deferred and capitalised. The hypothetical includes that the party in the position of STAI is the holding company of an operating subsidiary like SOPL. This brings into the hypothesis the cashflow position and the projected cashflow position of SOPL. In light of this, the independent parties might be expected to agree that the company in the position of STAI could defer and capitalise interest' (emphasis added): PJ[337].

148    The reasoning of the primary judge to the effect that the reliable hypothesis did not include the matters agreed by the second and third amendments was as follows:

(1)    there did 'not appear to be any commercial rationale' for the second amendment as it was not suggested that it was needed to address SOPL's cashflow issues and the original LNIA permitted interest to be deferred and capitalised. 'It seems more likely that, as submitted by the Commissioner, the aspect of the Second Amendment that stopped the accrual of interest (at least for a time) was directed at withholding tax issues': PJ[338]-[339];

(2)    the evidence of Mr O'Sullivan, a director of STAI, as to why the third amendment was agreed was to the effect that at the time of the third amendment to the LNIA external credit markets were 'experiencing extreme volatility as a result of the Global Financial Crisis' and 2008 was a difficult year for SOPL and SingTel: PJ[184]. However, 'Mr O'Sullivan's main focus appears to have been upon cash and liquidity, rather than on whether the interest rate under the existing facility (with more than three years left to run) was floating or fixed': PJ[189];

(3)    it was not accepted that the need to obtain certainty as to the interest rate payable under the LNIA or certainty as to ongoing funding was the rationale for the change in interest rate to a fixed rate: PJ[340]-[341];

(4)    further, and in any event, there was no expert evidence that a fixed base rate of 6.835% at the time of the third amendment was an arm's length substitute for the variable rate agreed at the outset of the LNIA in circumstances where Mr Chigas made clear that he had not addressed the figure: PJ[342];

(5)    the analysis of Mr Chigas departed too far from the actual transaction and the characteristics of the parties to that transaction and there were significant differences between the LNIA and the terms of a typical debt capital market bond issue: PJ[343];

(6)    there was further difficulty with the approach of Mr Chigas because 'it involves the calculation (in hindsight) of the effective credit spread of the LNIA, and a comparison of this credit spread with that of an STAI-issue [debt capital market] bond issue, rather than an approach which focusses on what independent parties in the positions of SAI and STAI, dealing independently with each other, might be expected to have agreed in June 2002, and at the time of each relevant amendment to the LNIA' (emphasis added): PJ[343]; and

(7)    the reasoning in Glencore did not support the conclusion that the possibility of a parent guarantee from SingTel should be excluded from the hypothesis (or the imputation of a guarantee fee if it was to be included): PJ[344].

149    It will be necessary to return to aspects of the reasoning of the primary judge concerning the matters the subject of the second and third amendments when dealing with Alleged Errors (2) and (5). These further Alleged Errors advance separate contentions as to why the Initial Rate was not appropriate for the whole of the term of the LNIA (even if the finding by the primary judge as to the Initial Rate is upheld as to the starting interest rate for the purposes of the reliable hypothesis).

(1)    the contention that the evidence relied upon by the primary judge as the basis for adopting the Initial Rate was of little, if any, probative value

150    When it comes to the interest rate for the reliable hypothesis, the question of what is probative must be evaluated having regard to the nature of the reliable hypothesis. The primary judge found that a capital raising on the debt capital market departed too far from the actual transaction and the characteristics of the parties to constitute the required reliable hypothesis. STAI does not challenge that finding in a direct way. Instead, it seeks, in effect, to rely upon the evidence as to interest rates in the debt capital market as being relevant in some way to the determination of the interest rate for the reliable hypothesis as formulated by the primary judge.

151    In that regard, broadly speaking, STAI says that the evidence that was relied upon by the primary judge to determine the interest rate was of little or no probative value and then advances reasons why the evidence of Mr Chigas should have been used to determine that rate.

152    It may be accepted that the evidence relied upon by the primary judge was not expert opinion evidence as to the appropriate interest rate for the vendor financing in the sense that it was not evidence based upon a demonstrated field of expertise. However, it was evidence of senior personnel with financial experience who were involved in the actual dealings. It was also evidence that the Initial Rate was set on the basis that it was an arm's length rate for the kind of financing that was being provided by SAI as vendor. The relevant business records contained statements by market actors who were involved in making finance decisions that were made in the ordinary course of their business dealings. If it had been shown that in forming views as to what was an arm's length rate they were influenced by Non-Independence Conditions or non-arm's length circumstances, their views as to what was an appropriate independent interest rate might be criticised as being of little probative value. However, there is no suggestion that their efforts in seeking to set an appropriate arm's length rate were not sincere or that they were formed by individuals whose daily responsibilities did not include forming views as to such matters. In those circumstances, it was open to the primary judge to treat the documents as recording the views of experienced market participants acting genuinely to set an arm's length rate based upon their experience.

153    It is the fact of the rate being identified at the relevant time as an appropriate arm's length rate by experienced market actors actually involved in the transaction and by reference to the character of the transaction that makes it probative.

154    STAI sought to criticise the Initial Rate as relied upon by the primary judge on the basis that the logic set out in the email between senior officers of SOPL did not support the rate that they identified. In particular, STAI submitted that the reliance by those officers upon the two data points identified in the email was flawed. They were also said to be views that were inadmissible as expert opinion evidence. However, as has been explained, the views of the officers concerned were not received as having been reasoned by the application of expertise derived from undertaking a recognised field of study that enabled conclusions to be reached from an available data set. The email was received as a record of the contemporaneous view of two people with finance responsibilities within the SingTel group who were seeking to determine an appropriate arm's length rate for the vendor financing for the share purchase by SOPL. In circumstances where those views were not said to have been criticised before the primary judge, it was open to his Honour to give the record of those views weight.

155    As to the criticisms of the logic of using the two data points, it was submitted by STAI that the first data point was 'of little value' because it concerned short term borrowings. The submission to that effect is not supported by any evidence. It overlooks the fact that the email makes an upward adjustment to the rate (which was for a three year issue) to allow for the longer term of the LNIA.

156    The second data point was questioned on the basis that historical trading data was not in existence prior to 1 July 2002. However, the email is not based upon data of that kind. It is based upon the rate at which SingTel's long term debt then on issue was being traded. There is no basis to suggest that this was not contemporaneous information available to those who were involved in the email exchange at the time. Further 'difficulties' were raised by STAI in submissions concerning the comparison with SingTel 10 year bonds then on issue. The bonds on issue were fixed interest bonds when the LNIA was variable. The bonds were denominated in Euro and had not been swapped into AUD. At the time, part of the term had expired. Reliance was also placed on evidence of Mr Johnson and Bloomberg Swap Manager calculations.

157    These submissions are based upon the premise that the two data points were the full extent of the foundation for the view that the Initial Rate was a market rate. However, the email did not indicate that the data points were used in such a direct way to make the contemporaneous determination that the bank bill swap rate plus 1% was an appropriate rate. The email records that the rate was 'based on several factors' and then lists the two data points. Therefore, the document does not indicate that it is a direct application of the data points.

158    Further, these submissions fail to engage with the significance of the fact that the evidence relied upon by the primary judge was a contemporaneous assessment by those with responsibility to undertake such assessments within SingTel that is 'based on' the two data points in an unspecified manner. No doubt there is room for a divergence in views as to the rate at which a vendor might have provided finance to STAI for a share sale transaction of the kind that was financed by the LNIA. However, evidence from within the SingTel group of what, at the time, was considered to be an arm's length rate, which is not said to record views that lacked bona fides, does not lack probity because it might be said that an expert who had to rely upon data rather than day to day experience might have approached the task in a different way.

159    In addition to the communications identified by the primary judge, there was advice from external tax advisors to STAI on 16 December 2002, at the time of the first amendment, that the LNIA's 1% spread was 'genuinely an arm's length interest rate'. There were also other documents at the time of the second amendment which confirmed that the benchmark arrangement was formulated on the basis that the initial rate for the LNIA was an arm's length rate.

160    As has been noted, the second amendment resulted in the deferral of the accrual of interest payment obligations (and the potential for higher interest payments overall if the benchmark was met earlier than expected, as it was). The third amendment then fixed the interest rate at higher levels. Instead of seeking to justify these further dealings as producing interest payments in the each of the relevant income years that were equivalent to arm's length payments, STAI sought to rely upon the analysis of Mr Chigas which was to the effect that the appropriate arm's length rate exceeded the 'effective credit spread' of 144 bps determined on the basis of the US bond market (where the bank bill swap rate was used as the risk free base rate at the time of entry into the LNIA) over the whole 10 year term. That is to say, despite having acted on the basis that the bank bill swap rate plus 1% was an arm's length rate which was an appropriate variable rate at the time of entry into the LNIA, STAI sought to justify a much higher rate based upon the expert evidence of Mr Chigas as the arm's length rate. It did so without explaining why the contemporaneous assessments that had been made as to an appropriate arm's length rate for the vendor financing provided by the LNIA were considered to be too low.

161    Further, the principal analysis by Mr Chigas was to the effect that a 'market' rate in the debt capital market would have resulted in an additional $2.8 billion in interest over and above the $4.9 billion in interest actually paid by STAI over the 10 year term. That is to say, his evidence was that despite the contemporaneous efforts to set the rate at an arm's length rate, the actual market rate was at a level that would have resulted in over 57% more interest being paid than had actually been paid by STAI. These figures indicated a very substantial disconnect between the analysis of Mr Chigas and the real world contemporaneous assessments of those involved at the time. The differential was a matter that was noted by the primary judge: PJ[267], [343]. As his Honour observed as to the differential of $2.8 billion (PJ[343]):

Mr Chigas does not ask whether an independent party in the position of STAI in June 2002 might be expected to have agreed to the interest rate applicable under his 'market' scenario; it is not the way that he has approached the issues. However, for the reasons already indicated, I consider that the matter needs to be approached in that way.

162    As is explained further below in dealing with the remaining contentions advanced by STAI to support its Alleged Error (1), there was no expert evidence led by STAI of the rate that might be agreed by a vendor providing finance for the sale of shares in a company like SOPL.

163    For reasons that have been given, no error has been demonstrated in the primary judge giving significance to the contemporaneous evidence in reaching a conclusion as to the interest rate for the reliable hypothesis.

(2)    the contention that the Initial Rate was outside (that is to say, below) the outer limits of what was in issue as far as the experts were concerned

164    For STAI, it was submitted that as between the experts there was never a suggestion that the arm's length margin for a 10 year loan raised on the debt capital market by bond issue as at 28 June 2002 was less than 1.4%. In particular, it was submitted that Mr Johnson did not give evidence that the margin of 1% as agreed in the initial LNIA was the relevant rate. These propositions may be accepted, but they are confined to the debt capital market.

165    It was submitted for STAI that reliance was placed upon the following agreed statements included in a joint report provided by Mr Chigas and Mr Johnson:

STAl's most economic and likely only source for the quantum of $5.2 billion of third-party, long-term funds would have been a US [debt capital market] new issue transaction coupled with a CCIRS.

…

[Debt capital market] desks use both recent new issue transactions and current secondary market trades for pricing new issues.

An adequate quantity of verifiable new issue pricing and secondary trading data of comparable pricing benchmarks is ideally sourced as close to the pricing date as possible.

166    The approach of the experts was also said to be logical because it would not be expected that a purchaser would pay a materially higher rate of interest to a vendor than it would pay if it raised the funds under a commercial debt financing. However, that was a matter advanced as a submission, not by reference to any aspect of the evidence before the primary judge. Vendor finance denominated in AUD enabled the saving of the cost of arranging a CCIRS. The primary judge ultimately found that the Initial Rate was comparable to vendor financing of $5.2 billion for the acquisition of shares in a company like SOPL by a company like STAI undertaken on the basis that a parent guarantee would be provided whereas a bond issue in the debt capital market may not be supported by such a guarantee, a matter that may affect the relative interest rates. The significance of these matters is addressed separately. For present purposes what is significant is that it is not possible to equate the reliable hypothesis as formulated by the primary judge with a borrowing by bond issue in the US debt capital market.

167    STAI pointed to no expert evidence to support the proposition that there may be some form of equivalence between interest rates for vendor financing for a purchase of Australian shares and prevailing interest rates in US debt capital markets. Rather, the contention for STAI was to the effect that Mr Chigas and Mr Johnson had agreed that a bond issue in the US debt capital market meant that the interest rates they had considered should be used for the reliable hypothesis. However, that was to submit that the appropriate reliable hypothesis was a matter for agreement between the experts. It was not. It required an evaluative judgment by the primary judge to reach a conclusion on a question of mixed law and fact applying the authorities concerning the way in which to formulate the hypothesis. As has been mentioned, the primary judge undertook that task and reached a conclusion that a capital raising in the debt capital market departed too far from the actual transaction and the characteristics of the parties. That conclusion is not directly challenged in the appeal.

168    To the extent that the contention involves some form of claim that the proceedings before the primary judge were conducted on the basis that the relevant interest rate would be somewhere within the disputed range as between Mr Chigas and Mr Johnson, there was no foundation put forward for a procedural submission of that kind. It could not be sustained in circumstances where the Commissioner relied upon the 'no amendment' calculations which used the Initial Rate.

(3)    the contention that the adoption of the debt capital market analysis to determine the arm's length rate was a 'common approach' as between the experts

169    For reasons that have been given, the fact that Mr Chigas and Mr Johnson both undertook analysis by reference to the debt capital market did not mean that there was error by the primary judge in not applying their interest rate analysis to a different hypothesis which his Honour found to be the reliable hypothesis that was appropriate in the circumstances of the case.

170    Further, the experts did not adopt a common approach. As the primary judge correctly explained (at PJ[295]):

Mr Chigas and Mr Johnson have undertaken quite different exercises. Mr Chigas has, in simple terms, 'priced' the LNIA as if it were a [debt capital market] bond issue by STAI in June 2002. He has then compared this 'arm's length price' with the actual effective price of the LNIA, and concluded that the interest actually paid under the LNIA was less than the 'arm's length price'. On the other hand, Mr Johnson has considered what STAI and SAI might reasonably be expected to have done if they had been independent parties. Further, he has considered this as at the date of the original LNIA, the date of the Second Amendment and the date of the Third Amendment. He has described the transactions that he considers STAI and SAI, assuming they were independent parties, might reasonably be expected to have entered into.

171    Further, at PJ[343], the primary judge observed (in a passage to which reference as to part has already been made):

… A second difficulty with Mr Chigas's, and thus STAI's, approach is that it involves the calculation (in hindsight) of the effective credit spread of the LNIA, and a comparison of this credit spread with that of an STAI-issued [debt capital market] bond issue, rather than an approach which focuses on what independent parties in the positions of SAI and STAI, dealing independently with each other, might be expected to have agreed in June 2002, and at the time of each relevant amendment to the LNIA … Mr Chigas does not ask whether an independent party in the position of STAI in June 2002 might be expected to have agreed to the interest rate applicable under his 'market' scenario; it is not the way that he has approached the issues. However, for the reasons already indicated, I consider that the matter needs to be approached in that way.

172    It was the failure by Mr Chigas to express any opinion as to what parties may have done if the reliable hypothesis as found by the primary judge applied that led to the rejection of his evidence: see, in particular, PJ[296] to which reference has already been made.

173    There was no relevant sense in which the evidence of Mr Chigas and Mr Johnson could be said to have proceeded upon some common basis.

(4)    the contention that each of the experts acknowledged that the actual transaction involved vendor finance

174    It was submitted for STAI that in their joint report Mr Chigas and Mr Johnson acknowledged that the actual transaction involved vendor finance. It appeared to be submitted that this meant that their analysis was applicable to the reliable hypothesis as found by the primary judge. What was stated in their joint report was: 'The $5.2 billion financing under the LNIA was partial consideration for the purchase of Optus shares by STAI'. A statement in those terms provides no foundation for the view that the experts were expressing some opinion concerning the interest rate that may be applicable to vendor financing of a share sale. The statement did no more than refer to the background circumstance in which their joint opinions were expressed. Those joint opinions concerned the terms upon which financing might be obtained on an arm's length basis in the debt capital markets, not from a vendor.

175    A separate submission to the effect that the experts used the debt capital market rate 'as a proxy for the interest rate that a hypothetical purchaser might be expected to pay a hypothetical vendor were they independent parties' should not be accepted. The experts did not express their opinions on that basis. Had they done so then issues may well have arisen as to whether that was a valid approach having regard to wider aspects of the transaction. Instead, the experts proceeded on the view that the appropriate alternative to vendor financing for the purposes of the statutory analysis was a capital raising by a buyer in the debt capital market. The primary judge did not accept that approach and his reasoning as to that aspect is not challenged.

(5)    the contention that if SingTel had provided a guarantee then its rating would have been adversely affected which would have limited the extent to which the parent guarantee could have resulted in a more favourable interest rate for STAI

176    The primary judge's reasoning that the Initial Rate was one which would have applied depended in part upon his Honour's conclusion that the reliable hypothesis to be used for the application of the statutory provisions should include the provision of a parental guarantee by SingTel at no cost to STAI. By Alleged Error (5), STAI claims that his Honour was in error in including the provision of a parental guarantee (alternatively a guarantee at no cost to STAI) in the reliable hypothesis. That aspect of STAI's case is dealt with below in addressing Alleged Error (5).

177    At this point in its case on appeal, the contention by STAI is that in determining the interest rate that would apply in the conditions of the reliable hypothesis, the primary judge should not have concluded, in effect, that a guarantee from SingTel would have allowed STAI to enjoy the benefit of SingTel's better credit rating at the time. Rather, so it is submitted, the primary judge should have concluded (based upon the evidence of Mr Chigas) that the rate of interest that might reasonably have been expected from an arm's length dealing was higher than the Initial Rate because a guarantee from SingTel would not have caused the market to conclude that the credit risk of STAI was reduced.

178    The contention appears to involve the following steps:

(1)    the primary judge found that the Initial Rate was an arm's length rate at which STAI might have been expected to be provided with vendor finance of $5.2 billion as at the time of entry into the LNIA;

(2)    part of his Honour's reasoning was that the Initial Rate was one which would have applied because a guarantee at no cost would have been provided by a parent like SingTel which would have improved STAI's credit risk;

(3)    the evidence of Mr Chigas was that a guarantee from a parent like SingTel would not have improved STAI's credit risk or would not have improved it to the extent assumed by the primary judge; and

(4)    therefore, the Initial Rate was too low.

179    Precisely how the Initial Rate might be adjusted if the contention was accepted was not articulated. The argument seemed to proceed on the basis that it was a matter for subsequent calculation if the submission was upheld. However, STAI bore the onus of putting forward some evidence upon which a finding might be made as to the interest rate. STAI did not point to evidence which might form the basis for some adjustment of the Initial Rate to reflect the extent to which it factored in support by way of a guarantee from a parent like SingTel. In that regard, a fundamental difficulty for STAI in advancing the contention is that the primary judge did not reason that the Initial Rate included any form of identifiable reduction for a guarantee from a parent. Rather, the primary judge reasoned that because the Initial Rate had been determined on the basis that it was a rate at which SingTel could have borrowed then it was an appropriate rate to use for a reliable hypothesis that included the provision of a parent guarantee. That is to say, the Initial Rate was not built up from some other rate that was found to be the arm's length rate that might reasonably be expected if the reliable hypothesis did not include a parent guarantee. Consequently, there was no identifiable component that might be adjusted if STAI's submission was accepted.

180    In any event, for reasons which follow, STAI's contention should not be accepted.

181    The primary judge's conclusion that the Initial Rate was appropriate did rest, in part, upon reasoning to the effect that the evidence of Mr Chigas that a credit spread of 400 bps was a 'market based pricing' was not of assistance and did not provide a basis for departing from the Initial Rate as the interest rate that might be expected between independent parties (that is, in the circumstances of the reliable hypothesis as found by the primary judge): PJ[332]-[333].

182    To support that finding, the primary judge made two points. The first point was that Mr Chigas' analysis was conducted on the basis of the LNIA as amended by the second amendment. There were two important aspects of the second amendment that were found by the primary judge to be outcomes of Non-Independence Conditions that operated at the time. First, the second amendment had the effect that interest of approximately $286 million that had accrued was treated as not having accrued. Second, the benchmark was introduced which meant that the amount of interest that was to be paid depended upon when and if SOPL became profitable.

183    It was suggested that the primary judge was confused as to the nature of the task that was undertaken by Mr Chigas. It was said that Mr Chigas determined the actual interest rate based on the cash payments (that is the payments made over the whole 10 year term of the LNIA). The primary judge was not confused as to that aspect. He had carefully described that part of the way in which the analysis by Mr Chigas was undertaken: PJ[253]-[257]. In any event, it was not that part of the analysis by Mr Chigas to which the primary judge was referring. His Honour was concerned with the interest rate that should be applied on the hypothesis that he had described (namely vendor finance to STAI with a parent guarantee and with deferred and capitalised interest payments). In that context, his Honour pointed out that Mr Chigas had conducted his analysis based upon the LNIA as amended by the second amendment. That was correct: see the earlier unchallenged finding of the primary judge at PJ[236]. In consequence, there was uncertainty as to the extent to which the aspects of the second amendment described above (which did not form part of the reliable hypothesis as found by his Honour) were aspects that affected the credit spread estimated by Mr Chigas as being the spread that would have applied to a capital raising on the terms of the LNIA as amended by the second amendment. In particular, the benchmark term introduced considerable risk for the lender, a matter that may be expected to be reflected in a higher interest rate. The point being made by the primary judge was that, by reason of the way Mr Chigas had approached the task, the effect of the conditions that were included in the LNIA as amended by the second amendment (that were not to be included in the reliable hypothesis) was 'unclear' and '[raised] a question as to whether his analysis [can be] relied on' for the purposes of determining the interest rate on the hypothesis to be applied: PJ[332]. That aspect of the reasoning of the primary judge has not been demonstrated to be in error.

184    The second point made by the primary judge concerned a view expressed by Mr Chigas that a debt market transaction with a parent guarantee from SingTel (with no guarantee fee) would have a credit spread of 205 to 215 bps. His Honour found that evidence not to be of assistance because there did not appear to be any firm foundation for Mr Chigas' assumption that SingTel would likely have been placed on negative watch or downgraded prior to the financing. It was submitted in the appeal that 'under either of these two scenarios [i.e. 205 bps or 215 bps] the taxpayer would win whatever the precise amount of the spread'. The submission advanced by STAI in those terms, like many of its submissions, assumes the acceptance of the legally flawed analysis by which there is regard to the total interest over the 10 year term rather than a focus upon the interest paid in the relevant income years.

185    As to the whether there was a firm foundation for Mr Chigas' opinion about the basis for his opinion about the credit spread based upon a parent guarantee without fee from SingTel, as the primary judge correctly observed, it depended upon a view that SingTel's credit rating would likely have been placed on negative watch or downgraded. It was that aspect that his Honour described as an 'assumption' without 'any firm foundation': PJ[333].

186    There is no analysis in the report of Mr Chigas to support the view that a negative watch or downgrade for SingTel's credit rating was 'likely' at the time of entry into the LNIA. Instead, there is a list of matters that, in Mr Chigas' opinion experienced investors would have incorporated 'for a SingTel guaranteed transaction'. One of those is 'S&P rating adjustment for SingTel to A+, resulting in 3 bps better than [an identified comparator] with a negative outlook for Moody's'. Expressed in those terms, it assumes that such an adjustment would occur and expresses an opinion as to the effect that the adjustment would have (with other matters) on the spread that would be required by experienced investors in the debt capital market.

187    Therefore, the primary judge was correct to refer to the negative watch or downgrade as an assumption without a firm foundation.

188    A submission was made to the effect that Mr Chigas was not challenged on his opinion that SingTel's credit rating would likely have been placed on negative watch or downgraded. But, as the primary judge correctly found, it was not a matter on which Mr Chigas expressed a reasoned opinion.

189    It was then said that there was ample evidence from which the primary judge could and should have adjusted the margin suggested by Mr Chigas. It was suggested that, on appeal, the Court should undertake some assessment as to an adjustment to the Initial Rate that should be made to reflect the credit rating effect for SingTel if it had guaranteed the debt of $5.2 billion. There are a number of problems with the submissions advanced by STAI on that basis. First, and fundamentally, as has been explained the figures referred to by Mr Chigas were not supported. They were based upon an assumption not analysis. Consequently, it was not possible to make a meaningful adjustment to that figure as assumed by Mr Chigas. Second, it was not suggested that the primary judge was invited to undertake an adjustment of the kind contended for by STAI in the appeal. In consequence there is no reasoning by the primary judge concerning the evidence that might be brought to bear on such a topic nor any indication of the findings that may have been made by the primary judge based upon hearing all of the evidence. Third, the whole of the analysis by Mr Chigas was predicated on the view that there was broad consistency between the terms of the LNIA as amended by the second amendment. The interest rate that was relevant on the approach of the primary judge was for a different reliable hypothesis. Fourth, the finding about the effect of credit rating is a matter within the fields of expertise of Mr Weiss and Dr Chambers. The primary judge found that the level of expertise of Mr Chigas as to such matters was not as great as that of Mr Weiss and Dr Chambers. Yet the submission is not advanced by reference to any aspect of their evidence.

190    Finally, even if (contrary to the above conclusions) it was appropriate to have some regard to the evidence of Mr Chigas to adjust the Initial Rate, the parties made competing submissions as to the significance of various aspects of the evidence for any finding concerning the credit rating effect of a guarantee by SingTel. It is not for this Court on appeal, in the absence of any expert evidence as to the manner in which such an adjustment might be made in the context of the reliable hypothesis being applied by the primary judge, to attempt some form of unguided adjustment to the Initial Rate.

(6)    the contention that the evidence of the actual issue by similarly rated telecoms at about the time of entry into the LNIA as relied upon by Mr Chigas should not have been rejected

191    Mr Johnson gave evidence to the effect that there was a disconnect between the performance of companies in the telecommunications markets in the United States and Australia at the relevant time. In his opinion, by reason of the disconnect, interest rates at which US telecommunications companies borrowed funds in the debt capital market at the time did not indicate rates at which Australian telecommunications companies could borrow funds. Mr Johnson sought to demonstrate that difference by an analysis of the credit spread for a SingTel issued bond. The approach of Mr Johnson was the subject of successful challenge before the primary judge: PJ[334].

192    STAI now submits that having rejected the analysis of Mr Johnson, the primary judge should not have 'overlooked' the analysis of Mr Chigas based upon the US telecommunications companies. However, there was no need for the primary judge to address that analysis which concerned the rates that prevailed in the debt capital market which the primary judge found not to be the appropriate reliable hypothesis. His Honour found that there were significant differences between the terms of the reliable hypothesis (in which a vendor would provide finance with deferred and capitalised interest: PJ[337]) and a debt capital market transaction. Mr Chigas and Mr Johnson agreed that the variation notice procedure in the LNIA had not been observed by either of them in a debt capital market transaction: PJ[290]. The unchallenged finding of the primary judge is that there were significant differences between the terms of the LNIA and the terms of a typical debt capital market bond issue: PJ[244], [343].

193    There was no error by the primary judge in not using the analysis by Mr Chigas in reaching a conclusion as to the arm's length interest rate based upon a different reliable hypothesis to that which was implicit in the analysis used by Mr Chigas (and which was not informed by an inquiry as to what might have been expected to have been agreed as part of an arm's length vendor financing of the purchase of the shares in SOPL).

Conclusion as to Alleged Error (1) and more fundamental difficulty for STAI's case

194    For the reasons that have been given, Alleged Error (1) has not been made out by STAI.

195    STAI sought to deploy the interest rate analysis of Mr Chigas within the context of the reliable hypothesis as found by the primary judge. As has been explained, that approach involved a comparison between (a) the 'effective credit spread' (as computed by him) based upon interest actually paid by STAI over the 10 year term; and (b) the interest margin over the base rate that might have been expected as the arm's length rate for the reliable hypothesis. STAI says the primary judge should have found that the interest margin for the reliable hypothesis was considerably more than 1% over the base rate as found by the primary judge. As has been explained, that approach disregards the timing of the accrual of the obligation to pay interest. It fails to identify and compute the extent to which interest that actually accrued and was paid in each of those years was no more than would have accrued and been paid if the Non-Independence Conditions did not apply (as to Subdivision 815-A) or arm's length consideration was agreed or given (Division 13).

196    Therefore, even if it had been shown that a different interest rate should have been used by the primary judge, it does not follow that the primary judge was in error in finding that STAI has not shown the amended assessments to be excessive. That is because, STAI's case as to how the relevant provisions of Subdivision 815-A and Divisions 13 should be applied was fundamentally flawed for reasons that have been given. It failed to address the amount of interest that would have been deductible in the relevant income years based upon evidence of a reliable hypothesis that demonstrated that:

(1)    there was no transfer pricing benefit in each of the relevant income years (or that the benefit was less than the amount assessed) (in the case of Subdivision 815-A); or

(2)    the interest consideration claimed by STAI to be deductible on the basis that it accrued and was paid in the relevant income years (by reason of when it was given or agreed) did not exceed the arm's length consideration in respect of that income year (in the case of Division 13).