Mylan Australia Holding Pty Ltd v Commissioner of Taxation - [2024] FCA 253 - FCA 2024 case summary — Zoe

CTHFCA

Mylan Australia Holding Pty Ltd v Commissioner of Taxation

[2024] FCA 253

Federal Court of Australia|2024-03-20|Before: Button J

View original source

At a glance

Source facts

Court

Federal Court of Australia

Decision date

2024-03-20

Before

Button J

Catchwords

[2015] FCAFC 57 - points not necessary to be decided

Source

Original judgment source is linked above.

Catchwords

[2015] FCAFC 57 - points not necessary to be decided

Judgment (89 paragraphs)

[1]

INTRODUCTION 1 The Applicant, Mylan Australia Holding Pty Ltd (MAHPL), brought two proceedings against the Respondent, the Commissioner of Taxation (the Commissioner) under Pt IVC of the Taxation Administration Act 1953 (Cth) (the TAA). By determinations issued under s 177F of the Income Tax Assessment Act 1936 (Cth) (the ITAA36), the Commissioner disallowed MAHPL's deductions for interest expenses under an intercompany promissory note referred to as PN A2, and consequential carry forward losses. PN A2 had flexible terms, permitting interest to be capitalised and allowing the prepayment of principal without penalty. The deductions for interest claimed by MAHPL also reflected the fixed interest rate that was determined (exactly when was in dispute) to apply to PN A2. 2 The Commissioner issued amended assessments for the income years ending 31 December 2009 to 31 December 2020. 3 MAHPL is the head company of a tax consolidated group, which includes Mylan Australia Pty Ltd (MAPL). MAHPL is the immediate parent company of MAPL. The ultimate holding company of MAPL and MAHPL is Mylan Inc (formerly known as Mylan Laboratories Inc) (Mylan). Mylan is the head of the Mylan group of companies. 4 MAPL acquired all of the shares in Alphapharm Pty Ltd (Alphapharm) in October 2007. Alphapharm was one of the operating subsidiaries of Merck KgaA (Merck) which, together with its related entities, carried on a global generics pharmaceutical business (Merck Generics). Merck Generics was acquired by members of the Mylan group in October 2007. The acquisition was a USD 7 billion transaction. The proceedings concern the application of Pt IVA of the ITAA36 to the funding arrangements associated with MAPL's acquisition of the shares in Alphapharm. In short, MAPL was funded with a mix of interest-bearing debt and equity at a 3:1 ratio. The debt component was constituted by PN A2, issued by MAPL to Mylan Luxembourg 1 S.a.r.l. (Lux 1). As its name suggests, Lux 1 was a Luxembourg company. 5 In applying Pt IVA, the Commissioner identified a wider scheme (also referred to as the primary scheme), and a narrower scheme (also referred to as the secondary and tertiary schemes - the secondary and tertiary schemes being identical). 6 The Commissioner considered that the entry into the wider scheme (which included the incorporation of the local Australian holding company structure (MAPL and MAHPL)) generated a tax benefit, being all the interest deductions on PN A2 (and a subsequent note, entered into in 2014, referred to as PN A4). This was on the basis of the Commissioner's view that, had the wider scheme not been pursued, the shares in Alphapharm would not have been separately acquired through a local Australian holding company structure. Rather, Alphapharm would have remained a subsidiary of the Netherlands company, Merck Generics Group B.V. (MGGBV) and would have become part of the Mylan group with the acquisition of MGGBV. In this scenario (described as the primary counterfactual), MAPL would not have acquired the shares in Alphapharm and would not have incurred interest expenses under PN A2. 7 At the objection stage, the Commissioner identified the narrower scheme, and developed two alternate counterfactuals (being the secondary counterfactual and the tertiary counterfactual). The narrower scheme does not include the establishment of the Australian holding company structure with MAHPL as the head entity. According to the counterfactuals the Commissioner developed in respect of the narrower scheme, MAPL and MAHPL would still have been incorporated and MAPL would still have acquired the shares in Alphapharm but, on the secondary counterfactual, MAPL would have borrowed a lesser sum (ie it would have had a lower gearing ratio) and would have borrowed under the same facility that Mylan and another group company in fact borrowed to fund the Merck Generics acquisition (ie MAPL would have taken on external debt at a floating rate). The tertiary counterfactual was the same, save that it posited the lender being Mylan or another US subsidiary of Mylan. 8 Although entry into PN A4 formed part of the wider and narrower schemes, the Commissioner did not take issue with the terms on which PN A4 was issued. Rather, his case was that, on the primary counterfactual MAPL would not have incurred any debt, therefore there would have been nothing to refinance and PN A4 would not have come into existence. How PN A4 featured in the secondary and tertiary counterfactuals was not clear. I can only assume that, by parity of reasoning, the deductions obtained as a result of the interest incurred under PN A4 was only contended to form part of the tax benefit to the extent that, on either of those counterfactuals, a different (and lesser) amount would have been refinanced by PN A4, with lower associated deductions for interest. 9 Initially, the Commissioner also defended his amended assessments on transfer pricing grounds, but dropped that part of his case before trial. Accordingly, the matter went to trial only on the Pt IVA issue (and some minor issues whose outcome rested on the determination of the Pt IVA case). 10 The conclusions I have reached on the principal issues are as follows: (a) MAHPL did not obtain a tax benefit in connection with the primary scheme that may be calculated by reference to the primary counterfactual; (b) had none of the schemes been entered into or carried out, the most reliable - and a sufficiently reliable - prediction of what would have occurred is what I have termed the "preferred counterfactual"; (c) the principal integers of the preferred counterfactual are as follows: (i) MAPL would have borrowed the equivalent of AUD 785,329,802.60 on 7 year terms under the SCA (specifically the term applying to Tranche B), at a floating rate consistent with the rates specified in the SCA; (ii) MAPL would otherwise have been equity funded to the extent necessary to fund the initial purchase of Alphapharm and to stay within the thin capitalisation safe harbour ratio from time to time; (iii) Mylan would have guaranteed MAPL's borrowing under the SCA; (iv) Mylan would not have charged MAPL a guarantee fee; (v) interest on the borrowing would not have been capitalised; (vi) MAPL would have been required to pay down the principal on a schedule consistent with that specified in the SCA and would have made voluntary repayments to reduce its debt as necessary to stay within the thin capitalisation safe harbour, from time to time; (vii) MAPL would not have taken out hedges to fix some or all of its interest rate expense; (viii) MAPL would have taken out cross-currency swaps into AUD at an annual cost of 3.81% per annum over AUD 3 month BBSW; and (ix) if MAPL's cashflow was insufficient to meet its interest or principal repayment obligations, Mylan would have had another group company loan MAPL the funds necessary to avoid it defaulting on its obligations, resulting in MAPL owing those funds to that related company lender by way of an intercompany loan, accruing interest at an arm's length rate; (d) MAHPL did (subject to matters of calculation) obtain a tax benefit in connection with the schemes, being the difference between the deductions for interest obtained in fact, and the deductions for interest that would be expected to be allowed on the preferred counterfactual; and (e) MAHPL has discharged its onus in relation to the dominant purpose enquiry specified by s 177D of the ITAA36 and so has established that the assessments issued to it were excessive.

[2]

THE EVIDENCE 11 The parties tendered a substantial number of documents at trial. Following the conclusion of the trial, the parties marked up versions of the indexes to the Court Book and the Supplementary Court Book, striking out documents that were not referred to at trial (and were therefore not treated as having been tendered). The parties also prepared a Second Supplementary Court Book, all of which was referred to at trial (such that there was no need to prepare a marked up index of that court book). 12 There was limited lay evidence. MAHPL read the following affidavits of lay witnesses: (a) an affidavit of Paul Campbell dated 4 August 2022; (b) an affidavit of Thomas Salus dated 23 August 2023; and (c) a further affidavit of Thomas Salus dated 20 October 2023. 13 Mr Campbell is the Senior Vice President, Controller and Chief Accounting Officer at Viatris Inc, the present ultimate parent of MAHPL. In 2007, Mr Campbell was Mylan's Vice President Corporate Accounting and Reporting, Business Development, Strategic Development. His responsibilities were mainly focused on accounting, finance business development activities, purchase accounting for acquisitions and preparation of consolidated financial statements. Mr Campbell gave some very limited evidence concerning the acquisition of the Merck Generics business. He deposed to Mylan having completed the acquisition in accordance with version 17 of a "step plan" prepared by Mylan's advisors. 14 Mr Salus is the Assistant Secretary of Viatris Inc, and the Deputy Global General Counsel of Mylan. Mr Salus's August 2023 affidavit set out the names of individuals employed in various roles and when they were last employed by any member of the Mylan group. This evidence was to the effect that, the CFO, two treasurers, the Vice President Tax, the Director - International Tax and various others, had all left the employ of members of the Mylan group years prior to the present litigation. 15 Mr Salus's October 2023 affidavit is addressed in more detail below. It produced documents that relate to when the interest rate on PN A2 was fixed and when intercompany accruals reflecting the fixed rate, were effected. 16 Mr Campbell and Mr Salus were not cross examined. MAHPL also relied on an affidavit of its solicitor Daniel James Slater dated 2 May 2023 (sworn in connection with a pre-trial discovery application) in relation to some points concerning what documentation had been produced to the Commissioner, and when such documentation was produced. 17 The parties also relied on reports of experts. 18 The Commissioner and MAHPL each called an expert on US tax law. The Commissioner relied on an expert report of Harry David Rosenbloom dated 28 November 2022, an attorney engaged in private practice and a visiting professor of law at New York University School of Law. MAHPL relied on two reports of Kevin Glenn dated 2 August 2022 and 10 April 2023. Mr Glenn is a practising attorney at law and a partner of DLA Piper LLP (US). Prof Rosenbloom and Mr Glenn prepared a joint expert report, which was also tendered at trial, dated 16 June 2023. 19 Mr Glenn and Prof Rosenbloom both gave evidence and were cross-examined. 20 MAHPL also called evidence from two further experts: Terence Stack, and Mozammel Ali. Mr Ali is a financial markets expert, and is the Managing Director of Theorem Consulting PtyLtd, a firm specialising in advising on mergers and acquisitions, acquisition financing, capital raisings and capital structuring. Mr Ali prepared two reports: dated 10 August 2022, and 9 April 2023. Mr Stack is an expert in corporate treasury functions, including in relation to capital structuring, capital allocation, debt and equity market transactions, financial market risk management, liquidity management and related matters. Mr Stack prepared two reports: dated 5 August 2022 and 8 April 2023. 21 The Commissioner called expert evidence from Gregory Johnson. Mr Johnson is a capital markets expert and is the Managing Director of Global Capital Advisors LLC. He prepared a report dated 12 February 2023. 22 Mr Stack, Mr Ali and Mr Johnson prepared a joint expert report, which was tendered at trial, dated 18 July 2023. 23 Mr Stack and Mr Ali also participated in the preparation of a further joint expert report, dated 18 July 2023, along with David Bernard. Mr Bernard had been retained by the Commissioner, and had prepared a report. However, the Commissioner determined not to rely on his evidence at trial. Nevertheless, and as the joint expert report involving Mr Bernard contained material on which MAHPL wished to rely even though Mr Bernard was not being called, a redacted version of that report was tendered at trial. 24 Mr Stack, Mr Ali and Mr Johnson all gave evidence at trial and were cross-examined. 25 All of the experts were amply qualified to give opinion evidence on the topics covered by their reports.

[3]

THE LEGISLATIVE SCHEME 26 The parties jointly put forward, as the version of Pt IVA of the ITAA36 according to which the issues arising are to be determined, the version operative from 1 October 2007 to 31 December 2007. Relevant extracts were provided by the joint book of authorities. 27 This proceeding falls to be determined under the "old" Pt IVA regime, being the provisions in place prior to the amendments introduced by the Tax Laws Amendment (Countering Tax Avoidance and Multinational Profit Shifting) Act 2013 (Cth). 28 Section 177F provided for the making of determinations to disallow tax benefits. Section 177F(1)(b) provided as follows in the applicable version of Pt IVA: (1) Where a tax benefit has been obtained, or would but for this section be obtained, by a taxpayer in connection with a scheme to which this Part applies, the Commissioner may: ... (b) in the case of a tax benefit that is referable to a deduction or a part of a deduction being allowable to the taxpayer in relation to a year of income - determine that the whole or a part of the deduction or of the part of the deduction, as the case may be, shall not be allowable to the taxpayer in relation to that year of income; 29 As may be seen, the key concepts are the existence of a "tax benefit" obtained in connection with a "scheme" to which Pt IVA applies. 30 The term "scheme" was broadly defined by s 177A(1) and (3), as follows: scheme means: (a) any agreement, arrangement, understanding, promise or undertaking, whether express or implied and whether or not enforceable, or intended to be enforceable, by legal proceedings; and (b) any scheme, plan, proposal, action, course of action or course of conduct. ... (3) The reference in the definition of scheme in subsection (1) to a scheme, plan, proposal, action, course of action or course of conduct shall be read as including a reference to a unilateral scheme, plan, proposal, action, course of action or course of conduct, as the case may be. 31 Section 177C governs determination of whether a taxpayer has obtained a "tax benefit in connection with a scheme", and the amount of the tax benefit. Sections 177C(1)(b) and (d) provided as follows in relation to deductions: (1) Subject to this section, a reference in this Part to the obtaining by a taxpayer of a tax benefit in connection with a scheme shall be read as a reference to: ... (b) a deduction being allowable to the taxpayer in relation to a year of income where the whole or a part of that deduction would not have been allowable, or might reasonably be expected not to have been allowable, to the taxpayer in relation to that year of income if the scheme had not been entered into or carried out; ... and, for the purposes of this Part, the amount of the tax benefit shall be taken to be: ... (d) in a case to which paragraph (b) applies - the amount of the whole of the deduction or of the part of the deduction, as the case may be, referred to in that paragraph; 32 Section 177D provided for Pt IVA to apply only to some schemes. Relevantly, it provided that Pt IVA applies only to schemes entered into with the requisite purpose (objectively determined): This Part applies to any scheme that has been or is entered into after 27 May 1981, and to any scheme that has been or is carried out or commenced to be carried out after that date (other than a scheme that was entered into on or before that date), whether the scheme has been or is entered into or carried out in Australia or outside Australia or partly in Australia and partly outside Australia, where: (a) a taxpayer (in this section referred to as the relevant taxpayer) has obtained, or would but for section 177F obtain, a tax benefit in connection with the scheme; and (b) having regard to: (i) the manner in which the scheme was entered into or carried out; (ii) the form and substance of the scheme; (iii) the time at which the scheme was entered into and the length of the period during which the scheme was carried out; (iv) the result in relation to the operation of this Act that, but for this Part, would be achieved by the scheme; (v) any change in the financial position of the relevant taxpayer that has resulted, will result, or may reasonably be expected to result, from the scheme; (vi) any change in the financial position of any person who has, or has had, any connection (whether of a business, family or other nature) with the relevant taxpayer, being a change that has resulted, will result or may reasonably be expected to result, from the scheme; (vii) any other consequence for the relevant taxpayer, or for any person referred to in subparagraph (vi), of the scheme having been entered into or carried out; and (viii) the nature of any connection (whether of a business, family or other nature) between the relevant taxpayer and any person referred to in subparagraph (vi); it would be concluded that the person, or one of the persons, who entered into or carried out the scheme or any part of the scheme did so for the purpose of enabling the relevant taxpayer to obtain a tax benefit in connection with the scheme or of enabling the relevant taxpayer and another taxpayer or other taxpayers each to obtain a tax benefit in connection with the scheme (whether or not that person who entered into or carried out the scheme or any part of the scheme is the relevant taxpayer or is the other taxpayer or one of the other taxpayers). (Emphasis added.) 33 The concept of the "purpose" behind entry into the scheme was elaborated upon by s 177A(5), which provided that: A reference in this Part to a scheme or a part of a scheme being entered into or carried out by a person for a particular purpose shall be read as including a reference to the scheme or the part of the scheme being entered into or carried out by the person for 2 or more purposes of which that particular purpose is the dominant purpose.

[4]

FACTS 34 While some very limited lay evidence was given by affidavit, the factual dimensions of the case were entirely documentary. In its opening submissions, MAHPL set out an account of the facts by reference to the documents. Its account was, very substantially, couched in neutral terms; the account of the facts was not used as an opportunity for advocacy. In his opening submissions, the Commissioner addressed the factual background in one paragraph: The Commissioner apprehends MAHPL's statement of material facts at AS1 [12]-[58], [65]-[107] to be broadly accurate. The weight to be given to any particular fact, and the inferences that may be drawn from such facts, are likely to turn on the evidence given at trial and will therefore be addressed by the Commissioner in closing. 35 In view of the Commissioner's acceptance that MAHPL's account of the facts was "broadly accurate", in the course of opening submissions, I requested that the Commissioner detail any respect in which he contended the narrative was inaccurate. I also requested, noting the paragraphs carved out of the Commissioner's concession, that the Commissioner "be a bit more granular about what it is about those paragraphs that [he did not] accept". The Commissioner then sent a letter to MAHPL dated 18 October 2023, a copy of which was provided to the Court. The letter stated, in respect of wide ranges of paragraphs in MAHPL's written opening submissions, that the Commissioner "put [MAHPL] to proof". This entirely unhelpful response was staunchly defended by counsel for the Commissioner on the basis that the taxpayer has the onus of proof and had chosen not to call lay evidence. Ultimately, on 19 October 2023, a somewhat more helpful document was provided by the Commissioner, which set out the basis for the Commissioner's "non-admission" (as it was characterised in the Commissioner's letter) of certain facts. 36 I will not dwell on this episode further, save to observe that the taxpayer's onus in Pt IVC appeals does not absolve the Commissioner of his obligations under s 37N of the Federal Court of Australia Act 1976 (Cth) to conduct the proceeding in a way that is consistent with the overarching purpose. The overarching purpose includes the resolution of disputes as quickly, inexpensively and efficiently as possible, and the efficient use of the judicial and administrative resources of the Court (s 37M). I would not have thought that asking a litigant to identify with some specificity the respects in which it does not accept, or wishes to supplement, the document-based factual summary of his opponent is straining at the edges of the obligation of a litigant to conduct the proceeding in a way that is consistent with the overarching purpose. 37 Somewhat surprisingly, in view of the contents of its written opening and the interactions referred to above, the Commissioner's closing submissions contained an annexure - which ran to more than 30 pages and whose delivery had not been foreshadowed - setting out the facts. 38 Nevertheless, the Commissioner's factual narrative did not differ significantly from the factual narrative presented by MAHPL in opening. The main factual matters on which the parties differed concerned: whether or not Mylan had settled on an acquisition structure when it initially signed the SPA; whether or not Mylan intended that 100% of free cash flow would be repatriated; and when the interest rate on PN A2 was in fact fixed and retroactively applied. 39 Given the very limited compass of the divergence in views of the facts and the documentary source of the factual narratives, what follows is an account of the facts that, in part, reproduces and expands upon the parties' accounts of the facts. MAHPL's account in opening constitutes the base, but significant additional facts that were included by the Commissioner in his account in closing, but omitted from MAHPL's account, have been incorporated. Most of the additional facts addressed in the Commissioner's factual summary, but not in MAHPL's summary in opening, concern what the Commissioner characterised as the "debt pushdown" (viz, the creation of intercompany debt at the MAPL level). In their respective narratives, the parties highlighted different aspects of some documents. In the setting out the factual summary below, I have had regard to those differences of emphasis or construction. I have also had regard to, and included, factual matters canvassed in oral submissions, as well as additional factual matters that I considered ought to be addressed. I have also had regard to the Commissioner's comments (in his letter of 19 October 2023) regarding MAHPL's account of the facts in opening. 40 I have addressed the more significant factual controversies in the course of my reasons on "tax benefit" and "dominant purpose", and have addressed more minor factual controversies in the context of the narrative that follows. 41 The factual narrative below is arranged chronologically, within topics.

[8]

Bid to acquire Merck Generics 52 On or about 6 March 2007, Mylan received a letter from Bear, Stearns & Co Inc on behalf of Merck. The letter stated that Merck was exploring the potential sale of its generics pharmaceuticals business (being Merck Generics). It also enclosed a Confidential Information Memorandum, and invited Mylan to provide a preliminary, non-binding indication of interest in acquiring Merck Generics (the Acquisition). The Acquisition was code-named "Project Genius", with Merck given the code name "Mastermind". 53 From about March 2007, for the purposes of preparing its indicative offer (and subsequently its bid) for the Acquisition, Mylan engaged, among others: (a) Merrill Lynch as its primary financial advisor; (b) Cravath Swaine & Moore LLP (Cravath) as its external counsel; and (c) Deloitte & Touche LLP (Deloitte) as its accounting and tax advisor. 54 On 12 March 2007, the Mylan Board of Directors (Mylan Board) met and resolved to submit a preliminary non-binding indication of interest for the Acquisition with a proposed purchase price of EUR 4.2 to 4.7 billion. Mylan submitted an indicative offer for the Acquisition on the same day. Mylan was subsequently invited to participate in a "second round" process for the Acquisition, with a final bid to be submitted by 30 April 2007. 55 In March, April and May 2007, Merrill Lynch provided Mylan with modelling, analysis and structuring alternatives for the Acquisition. The modelling included scenarios involving both wholly debt and partial debt/partial equity funding for the Acquisition. Mylan also received due diligence materials and analysis. 56 On 30 April 2007, the Mylan Board met and resolved (inter alia) to approve the submission of an updated non-binding proposal for the Acquisition, attaching revised drafts of a "Share Purchase Agreement", "Transitional Services Agreement" and "Brand License Agreement", as well as a copy of a "commitment letter" for the financing of the Acquisition received by Mylan from a syndicate of lenders (Merrill Lynch, Citibank and Goldman Sachs) (referred to in more detail below). Mylan submitted its updated proposal on the same day, proposing a base purchase price in the range of EUR 4.4 to 4.75 billion. 57 The marked-up draft Share Purchase Agreement and Transitional Services Agreement attached to Mylan's updated proposal included the following notation: Note to Sellers: Purchaser's acquisition structure to be further discussed with Sellers. We understand based upon our discussions during due diligence process that Mastermind [Merck] is willing to discuss and accommodate an optimal acquisition structure for Purchaser. 58 The marked-up draft Brand Licence Agreement attached to the proposal similarly included the notation: "Note to Sellers: "Purchaser's acquisition structure to be further discussed with Sellers".

[9]

Share Purchase Agreement 59 On or about 3 May 2007, Mylan was invited to participate in the final stage of the Merck Generics sale process. 60 On 6 May 2007, Christian Brause from Cravath sent an email to representatives from Mylan with the subject "Genius/Structuring". It stated, inter alia: As you know, the negotiations on Monday and Tuesday will move very fast. Thus, we will have no time to come up with a fully agreed upon acquisition structure. We therefore intend to built [sic] into the SPA some flexibility to rearrange the acquisition structure between signing and closing. We intend to do that by incorporating a new section that would essentially look like the one set forth at the end of this email 61 The end of the email set out a new proposed clause which ultimately became the basis for clause 3.1.5 of the Share Purchase Agreement. It stated: Structure of Transaction. At the election of Purchaser's Guarantor, (i) any one or more Affiliates of Purchaser's Guarantor may be substituted for Purchaser in the transaction and (ii) Purchaser or any such substituted purchaser may directly acquire the Interests in any Subsidiary … In any such event, the parties will cooperate in good faith to effectuate any such substitution any/or change in the acquisition structure 62 On 8 May 2007, the Mylan Board met and resolved that it approved the submission of the updated, non-binding proposal for the acquisition of Merck Generics for EUR 4.9 billion. 63 On 12 May 2007, the Mylan Board again met and resolved (inter alia) that it approved the execution and delivery of the Share Purchase Agreement and all of the transactions contemplated thereby. The Mylan Board also approved and authorised the financing described in the lenders' Commitment Letter accompanying the Share Purchase Agreement (referred to in more detail below). 64 Also on 12 May 2007, Mylan, Merck Generics Holding GmbH, Merck S.A., Merck Internationale Beteiligung GmbH and Merck KgaA executed the Share Purchase Agreement (SPA), which provided for the Acquisition by Mylan of all of the shares in Merck dura GmbH, MGGBV, EMD Inc, Merck Generics Belgium B.V.B.A and Merck Genericos S.L. for a cash purchase price of EUR 4.9 billion (subject to certain adjustments) (being approximately USD 6.7 billion). 65 Mylan issued a press release on 12 May 2007 which stated (inter alia, emphasis added): Mylan Laboratories Inc. (NYSE: MYL) and Merck KGaA today announced the signing of a definitive agreement under which Mylan will acquire Merck's generics business ('Merck Generics') for EUR 4.9 billion ($6.7 billion) in an all-cash transaction. The combination of Mylan and Merck Generics will create a vertically and horizontally integrated generics and specialty pharmaceuticals leader with a diversified revenue base and a global footprint. On a pro forma basis, for calendar 2006, the combined company would have had revenues of approximately $4.2 billion, EBITDA of approximately $1.0 billion and approximately 10,000 employees, immediately making it among the top tier of global generic companies, with a significant presence in all of the top five global generics markets. … Under terms of the transaction, which have been unanimously approved by Mylan's Board of Directors, Mylan will acquire 100% of the shares of the various businesses comprising Merck Generics for a cash consideration of EUR 4.9 billion ($6.7 billion). Mylan has secured fully committed debt financing from Merrill Lynch, Citigroup and Goldman Sachs. The transaction is anticipated to be dilutive to full-year cash EPS in year one, breakeven in year two, and significantly accretive thereafter based on management's internal projections. The company is committed to reducing its leverage in the near term through the issuance of $1.5 billion to $2.0 billion of equity and equity-linked securities. The combined company will generate substantial free cash flow that will further enable it to rapidly reduce acquisition-related debt. Reflecting its more leveraged capital structure and focus on growth, Mylan is suspending the dividend on its common stock. 66 While it was bound to proceed with the Acquisition pursuant to the SPA as signed, at the time that the SPA was executed, Mylan had not settled upon its preferred structure for the acquisition of the Merck Generics business. The terms of the SPA included provisions which allowed Mylan to put forward a finalised transaction structure, with provision also being made for indemnification of the Merck Generics side for any increase in its tax costs arising from such changes. 67 Clause 3.1.5 of the SPA provided as follows: Structure of Transaction. At the election of Purchaser [Mylan], (i) any one or more Affiliates of Purchasers may be substituted for Purchaser in the transaction and (ii) Purchaser or any such substituted purchaser or purchasers may directly acquire Interests in any Subsidiary … The parties will cooperate in good faith to effectuate any such substitution and/or change in the acquisition structure, including executing any necessary or advisable amendments to this Agreement in order to reflect the foregoing. Purchaser will agree to an appropriate full indemnification arrangement with Sellers and Sellers' Affiliates to the extent such change in acquisition structure increases the tax costs to Sellers and Sellers' Affiliates above the amount of costs that would have been incurred in connection with the sales and transfers set forth in this Section 3.1 as of the Signing Date. 68 Clause 21.5 (headed "Assignment and Designation of Transferors") relevantly provided that Mylan: may designate any of its direct or indirect, wholly owned subsidiaries as a transferee of the Shares and Purchaser may assign its rights under this Agreement by way of security in connection with the Financing. 69 Notwithstanding the terms of the SPA that provided for changes to the stated acquisition structure, the Commissioner sought to emphasise that the original SPA was an agreement which Mylan was obliged to perform. The Commissioner contended that the fact Mylan entered into the original SPA prevented MAHPL from arguing that it would not have acquired Alphapharm in a way analogous to that which was provided for under the agreement. While I perceive the Commissioner's point to be one concerning the implications of this matter for the primary counterfactual (as distinct from a factual contest), to the extent he disputed the fact, I find that Mylan had not settled on a final acquisition structure when it signed the SPA. This matter is further addressed in relation to the primary counterfactual below.

[10]

Acquisition structuring 70 During May and June 2007, Mylan's advisors presented it with various potential structures for the acquisition of Merck Generics. They contemplated (inter alia) that: (a) various US and non-US entities would be created, including indirect Australian, French, and Canadian subsidiaries of Mylan - MAPL, Mylan France SAS (Mylan France) and Mylan Canada NSULC (Mylan Canada), respectively; and (b) MAPL, Mylan France and Mylan Canada would acquire the shares in local MGGBV subsidiaries, financed by a mixture of equity (common stock) and debt. 71 One of the slide decks prepared by Mylan's advisors was a Deloitte slide deck dated 27 April 2007 titled "Project Genius - Tax Overview". That report assumed a total acquisition price of USD 7 billion, of which USD 1 billion was to be allocated to the Australian component of the Merck Generics business. In relation to the entities to be acquired in Australia, Canada and Japan, Deloitte's report contemplated that: (a) the Australian entity would borrow from a "US/UK third party financial institution" the AUD equivalent of USD 750 million and otherwise be capitalised by USD 250 million equity, reflecting a debt capitalisation proportion of 75%; (b) the Canadian entity would borrow from a "Canadian branch of a US/UK third party financial institution" the CAD equivalent of USD 437.5 million and otherwise be capitalised by USD 62.5 million equity, reflecting a debt capitalisation proportion of 87.5%; and (c) the Japanese entity would borrow from a "Japanese branch of a US/UK third party financial institution" the JPY equivalent of USD 437.5 million and otherwise be capitalised by USD 62.5 million equity, also reflecting a debt capitalisation proportion of 87.5%. 72 On the same day (27 April 2007) there was also an email from Mr Todd Izzo (Deloitte) to Mr Jeffrey Mensch (Deloitte). In that email, Mr Izzo stated that he "would like to limit the foreign loans to Japan, Australia and Canada if possible", to which Mr Mensch responded, "Australia is 3:1 safe harbour", and quoted the following extract from "IBFD" (which is apparently a service provider in relation to cross-border tax affairs): From 1 July 2001, new thin capitalization measures (the "safe harbour" test) contained in Div. 820 of the ITAA97 apply a debt-to-equity ratio of 3:1 to all debt of an entity and not just related foreign party debt. For financial entities the debt-to-equity ratio is 20:1. Nevertheless, if the safe harbour test is failed, debt deductions will not be denied if the entity is able to demonstrate that the debt amount is at arm's length (i.e. an independent party would be able to raise the same amount of debt under the same terms and conditions). Further, Australian entities with overseas investments may avoid the application of the thin capitalization provisions by satisfying the worldwide gearing test, which requires that the average value of their Australian assets be at least 90% of their worldwide assets. 73 Mr Izzo responded "fine" and also said as follows: Also, let's not do a debt push down to Ge [apparently Germany]. Too much hassle. So, in sum, one bank, loaning to Japan and Canada through branches and directly to Austr[a]lia and Lux. 74 As further addressed below, MAHPL did not dispute that the thin capitalisation rules in each relevant jurisdiction influenced the amounts of debt being contemplated for acquisitions in a number of jurisdictions, but disputed that adopting structures that stayed within thin capitalisation limits could be characterised as tending to suggest a desire to maximise tax deductions. 75 On 1 May 2007, Deloitte issued a slide deck to Mylan entitled "Project Genius - Tax Overview", which was labelled as "Draft: For Discussion Purposes Only (Subject to Review by Non-U.S. Tax Professionals)". The slide deck contemplated (inter alia) that: (a) Mylan and a newly established Luxembourg S.a.r.l. would borrow funds from third party lenders; (b) various US and non-US entities would be created, including MAPL, Mylan France and Mylan Canada; (c) MAPL would acquire the shares in Merck's Australian subsidiary, Alphapharm, from MGGBV, financed as follows: (i) a US subsidiary of Mylan "contributing the Australian dollar equivalent of US$250,000,000 in exchange for common stock" in MAPL; and (ii) MAPL borrowing "the Australian dollar equivalent of US$750,000,000 from a US/UK third party financial institution … secured by a guarantee from Mylan US and all Mylan's non-US assets"; and (d) each of Mylan France and Mylan Canada would also acquire the shares in a local Merck subsidiary, financed with a mixture of equity (common stock) and external debt. 76 MAHPL relied on the fact that the subsequent slide decks prepared by Deloitte dated 11 May 2007, 30 May 2007 and 4 June 2007 also contemplated the formation of MAPL, Mylan France and Mylan Canada, each of which was to acquire the shares in a local Merck subsidiary using a mixture of equity and debt. MAHPL relied on the debt component of MAPL's financing having been expressed in each slide deck as: the Australian dollar equivalent of US$750,000,000 … from a US/UK third party financial institution … secured by a guarantee from Mylan US and all Mylan's non-US assets. 77 The Commissioner sought to focus the Court's attention to other aspects of these slide decks, which he contended demonstrate Mylan's tax structuring objectives. In respect of the 11 May 2007 presentation, the Commissioner observed that the presentation continued to contemplate that the Australian acquisition entity would obtain external debt funding, while the source of funding for the Canadian and Japanese acquisition entities had been varied to include internal borrowing. 78 The Commissioner pointed to the following statements in the 30 May 2007 revised slide deck as illustrating the influence of thin capitalisation rules in Australia, France and Canada on the planned capitalisation of each of the Australian, French and Canadian entities: The debt:equity ratio of Mylan France is 1.5:1, which is in line with the new French thin capitalization rules. … Related party debt push-down may have adverse Australian tax consequences; therefore, Genius Pty. Ltd. should be acquired before Lux Holdco's acquisition of Genius BV. … Interest Deductions - Australia's thin capitalization rules are based on accounting book values rather than issued capital (total debt cannot exceed 75% of the Australian asset values). … Since Mylan Canada is funded with intercompany debt, the Canadian thin capitalization rules come into play, which limits the debt:equity ratio to 2:1. Therefore, assuming a $500m total price for Genius Canada, the debt:equity ratio should be $333m of debt at $167m of equity. 79 In relation to the 4 June 2007 Deloitte slide deck, the Commissioner drew attention to the "Alternative A" structure, which assumed third party lenders loaning funds into each of Mylan Australia, Mylan Japan and Mylan Canada, and the relationship between the gearing ratios derived in relation to the posited quantum borrowed, and the thin capitalisation limits in those jurisdictions. 80 On 12 June 2007, Deloitte circulated a further slide deck dated 11 June 2007 entitled "Project Genius - 'Simple' Alternative" (marked as a draft for discussion purposes). MAHPL contended that this slide deck contemplated a structure under which (inter alia) MAPL would acquire the shares in Alphapharm using USD 250 million in equity and USD 750 million in debt in the form of a note from "Lux Holdco" (Mylan Luxembourg Sarl), rather than external financing from a third party lender. The Commissioner disputed this description and drew attention to step 18. The Commissioner stated that that step showed that, under the structure being contemplated, only the Australian Merck entity was to be acquired separately (and for cash consideration) whereas the Canadian and Japanese Merck (and other) entities were to be acquired indirectly. Step 18 of the slide deck stated: Mylan Australia acquires 100% of the outstanding shares of Genius Pty. Ltd. from Genius Genericos Group BV ("Genius BV") in exchange for the Australian dollar equivalent of US $1,000,000,000 in cash. 81 On 10 July 2007, Mr Joseph Vitullo (PwC US) sent an email to Mr Tony Carroll (PwC Australia) regarding PwC US having been engaged by Mylan. Mr Vitullo sought Mr Carroll's assistance in relation to a "debt pushdown into Australia related to the acquisition of the Merck generics business". That email included the following: Steve White (ITS partner) and I (ITS director) have been engaged by Mylan to assist them in the structuring of a debt pushdown into Australia related to the acquisition of the Merck generics business. We would like to have a conversation with you this week to discuss alternative means by which we can accomplish this goal. 82 On 13 July 2007, Mr David Kennedy (Vice President of Corporate Taxation, Mylan) signed a "Statement of Work" (SOW) between Mylan and its subsidiaries and PwC. The purpose of the SOW was stated as follows: This SOW covers services in connection with the acquisition and integration of Merck's Generics Business ("MGB"). Examples of the types of services covered by this SOW include evaluation of the tax implications and possible tax planning strategies at the federal, state and international level, associated with the acquisition and integration of MGB. 83 The SOW also described the nature of the services to be provided by PwC, which were grouped in three categories - namely, "Analyze", "Develop" and "Implement". The description of the services to be provided did not refer to the repatriation of foreign income or Mylan's stated de-leveraging plans (the absence of which was a matter to which the Commissioner called attention). 84 Following Mylan's engagement of PwC to provide tax advice in relation to the Acquisition, PwC's personnel undertook a range of activities, recorded in email correspondence and other documents, in furtherance of its retainer. 85 On 14 July 2007, Mr Vitullo sent an email to various PwC personnel attaching "a projection by country of Merck's operating profits from 2007 through 2010". The email stated that "[t]his should help in assessing each country's interest capacity". The attachment to that email included hand-written notations that highlighted Alphapharm, among other entities. As is discussed below, MAHPL and the Commissioner put diametrically opposed constructions on this document. The Commissioner said it showed an analysis directed at working out how much interest had to be charged to eliminate taxable income, whereas MAHPL contended it showed that there was (contrary to the Commissioner's submission) analysis of MAPL's capacity to service debt. 86 On or around 18 July 2007, Mylan prepared a document titled "Weekly Update - Finance: 6 - Tax Plan & Compliance - Week of 071607". In a section with the heading "Issues/Risks/Key Decisions", that document included the following: • Conclude which alternative acquisition structure is optimal from a tax perspective • Assess taxable income capacity, on a country-by-country basis, to absorb acquisition finance interest expense • Assess optimal levels of local country debt giving consideration to income capacity, debt:equity restrictions, income tax rate arbitrage, and fair values 87 On 19 July 2007, Mr Carroll sent an email to Mr Vitullo attaching a slide deck prepared by PwC entitled "Mylan Laboratories Structure Alternatives" (marked as "Draft Report"). In the email, Mr Carroll stated that: I have suggested limiting the borrowing level to the same proportion of the total borrowing to the total purchase price. We could always stretch this further within the safe harbour rules in Australia but we need to be comfortable from an anti avoidance perspective that we can justify a greater amount form a commercial perspective. 88 The slide deck set out five alternative structures for the acquisition of Alphapharm in Australia. Structure 1 contemplated external borrowing by an Australian subsidiary of Mylan to fund the Acquisition. Structure 1 included a note that: It is recommended that the level of borrowing be limited to the worldwide debt funding proportion for this acquisition. Where there is any increase above this level, the Australian anti-avoidance provisions would need to be considered. 89 Structures 2, 4 and 5 contemplated "internal" debt funding into Australia. Structure 3 contemplated an external borrowing by a partnership that would be treated as part of the Australian tax consolidated group. 90 Also on 19 July 2007, Mr Vitullo responded to Mr Carroll's email stating that: we have agreed with Mylan that on 8/3 we will deliver a comprehensive holding company structure along with proposed debt pushdown structure for Australia, France, Canada, and Japan. We will incorporate each separate country's debt pushdown strategy into this presentation. 91 Also on 19 July 2007, Mr Carroll sent an email to Mr Vitullo stating: As you will see from my note I did not want to push the debt to the limit unless we have strong commercial reasons for doing so. 92 Also on 19 July 2007, Mr Steve White (PwC US) sent an email to Mr Carroll stating that: Mylan has asked if we would give them the names of law firms that we have worked with on similar debt pushdowns/financings in that they would hope that this would help expedite the implementation of any strategy that we develop. 93 On 1 August 2007, Mr Vitullo sent an email to his PwC colleagues, attaching a PowerPoint file titled "Merck Acq Structures". In his email, Mr Vitullo commented on the Canada, Australia and France acquisitions (among other things) (emphasis in original): October 1 Structure - This represents the proposed minimum structure required to be in place at the date of the closing of the Merck transaction. The following are specific country questions with respect to this structure. Canada Would it be possible to simply put in place a loan/equity from Bermuda 1 to a newly formed ULC which would be used to acquire the Canadian target entity? We would then after Oct 1 drop the note down into the structure and form the Canadian holding partnership structure which ultimately generates the tax savings element of the structure. Australia Would it be possible to simply put in place a required loan/equity from the 80/20 company to Aus Holdco to acquire the target entity? We would then after Oct 1 drop the note down into the structure and form the Australian high/low tax structure which ultimately generates the tax savings element of the structure. … 1) France - we contemplate establishing internal debt levels of 1.5:1. Are we correct that as long as Mylan maintains this relationship, there should not be a thin-cap exposure? 2) All Countries - Please indicate if there is any principal repayment requirements for the internal debt that we are putting into place. In other words, is a demand loan that Mylan keeps in place for a significant period of time acceptable, is there a requirement that principal payments are made over the life of the loan or at a point in time. 94 On 2 August 2007, Mr Garth Drinkwater (PwC Australia) sent an email ("on behalf of Tony Carroll") to Mr Vitullo which, among other things, stated that: Interest payments by Aust Hold should be deductible for Australian income tax purposes (subject to thin capitalisation provisions - broadly 75% of Australian assets less non-debt liabilities). … There are no requirements for principle [sic] payments to be made over the life of the loan (i.e. principle [sic] can be repaid at the end of the loan term). However, interest would need to capitalised (if not paid). Interest withholding tax would continue to be payable as the interest accrues. … We note that if interest is capitalised to the loan balance, rather than being paid, it may put pressure on the Australian group's thin capitalisation position where there is no corresponding increase in the book value of the assets (e.g. via increases in retained profits or asset revaluations). 95 On 3 August 2007, PwC prepared a slide deck titled "Merck Tax Integration August 2007". Under the heading "Tax Integration Goals/Objectives". MAHPL accepted on the transcript that this document was received by Mylan even though the covering email was not in evidence. The slides included the following statements: (1) Allow for redeployment of foreign excess cash via tax efficient Treasury Centre (2) Foreign tax reduction A. Use of debt-pushdown to effect immediate ETR reduction B. Consider utilizing a tax-efficient Principal in developing the new centralized supply chain management structure. … (2) Foreign tax reduction A. Use of debt-pushdown to effect immediate ETR reduction - Tax efficient internal debt utilized in Australia, France, Canada and Japan. Approximately $40M-$50M of annual tax savings over the first five years (resulting in immediate ETR benefits) may be realized by Mylan 96 On 10 August 2007, Mr Drinkwater sent an email to Mr Vitullo with the subject "Mylan acquisition - Stamp duty comments". That email provided comments in relation to a "Direct Acquisition", "Indirect Acquisition" and an "Alternative" structure. 97 Regarding the "Direct Acquisition", Mr Drinkwater stated that a liability would arise for "New South Wales share transfer duty at 0.6% on the greater of market value or consideration paid". 98 Regarding the "Indirect Acquisition", being an "indirect acquisition of Alphapharm by acquiring a foreign holding company further up the chain", Mr Drinkwater stated that no liability for share transfer duty would arise "provided the foreign company does not have a share register in Australia or a registered office in South Australia". Mr Drinkwater also stated that "[i]f an indirect acquisition occurs, it would not be possible to push debt into Australia until Alphapharm is later moved under the Australian holding company". 99 Mr Drinkwater commented on the "Alternative" structure as follows: Merck Generics Group BV could incorporate a new Australian holding company in Victoria ("Newco") and transfer Alphapharm under the Newco prior to Mylan's acquisition of the three global Merck companies. This transfer should be eligibile [sic] for the New South Wales corporate reconstruction exemption (subject to land rich issues etc as noted above). In addition, the debt pushdown would be effective on acquisition by Mylan (and therefore we would not need to wait one year before the debt pushdown could be effected). Newco and Alphapharm could be transferred and incorporated into your preferred structure one year later. Of course, this requires the co-operation of Merck and for Merck Generics Group BV to apply for the corporate reconstruction exemption. That said, this is not an unusual transaction here in Australia. 100 On 23 August 2007, Mr Drinkwater sent an email to Mr Vitullo. Among other things, that email confirmed that "we should be able to get a step up in the accounting values of the Australian Group" and then commented that "this was important for Australia's thin capitalisation rules". 101 On 1 September 2007, Mr Vitullo sent an email to Messrs Carroll and Drinkwater with the subject "Time of Transactions". That email included the following (underlining in original): We have been reviewing all of the Oct 1 steps and researching some of the US tax issues associated with the transactions which achieve the purchase of Merck targets in France and Australia prior to the acquisition of Merck BV. Some of the US tax, legal, and govenmental [sic] approval issues are particularly troublesome and we are now wondering if we may want to reconsider the timing of the debt pushdowns into France and Australia. Accordingly, we would like you to provide us, in a return email, confirmation of our understanding that it would be possible to push debt into your respective countries after Mylan acquires Merck BV. We anticipate that the internal debt pushdowns would occur within days or weeks of Oct. 1. 102 On 4 September 2007, Mr Carroll sent an email to Mr Vitullo with the subject "Acquisition structures". That email included the following: Further to our discussions this morning, I confirm that if the purchaser of Alphapharm is a wholly owned subsidiary of New Australian Hold Co, there are no adverse consequences from an Australian perspective. Your need for this from a US perspective also assists me in any arguments I might have regarding why we set up a two tiered structure and formed a tax consolidated group, prior to acquiring the Alphapharm company, from an Australian thin capitalisation perspective, so I welcome that addition. In relation to the alternative acquisitions [sic] structures, I would confirm that my preferred option, would be to establish New Australian Hold Co and its wholly owned subsidiary, underneath the proposed Bermuda structure and debt fund either, Australian Hold Co or Australian Interposed Co, to fund the acquisition. Under this arrangement an agreement would be entered into with BV prior to your acquisition of BV but conditional on Mylan's acquisition of BV. In my view from an income tax perspective, I believe there are considerably stronger arguments in relation to the debt push down, under this alternative than the one set out below. The alternate structure would involve the establishment of Australian Hold Co and Australian Interposed Co by BV and then an acquisition from BV after Mylan has acquired BV. In my view this proposal substantially increases the risk that interest deductions may be denied under the debt push down arrangements. The stamp duty corporate reconstruction exemptions is on the basis of an internal reorganisation and whilst the technicalities of the relief are available it is not really intended that there would be a change in ownership of BV and bearing in mind that the NSW government, to whom this duty would be payable, is as I understand it, one of Alphapharm's largest customers, I am not sure you necessarily want to push the letter of the law to this extent, bearing in mind the commercial relationships between Alphapharm and the NSW government from whom you are obtaining the concession. Further from an income tax perspective, the debt pushdown is on the basis this is a third party acquisition. There is a clear conflict between the reasons for obtaining the stamp duty relief and the reasons for undertaking the debt push down transaction. The Australian Revenue are very wary of internal reorganisations that achieve a debt push down. As originally discussed in one of our earlier conference calls, I believe the avoidance of the AU$6 million in stamp duty whilst potentially available, does increase the risks both from a tax perspective, in respect of the debt pushdown and secondly has a potential commercial outcome which could be adverse. I would strongly advise adopting the original proposal and pay the stamp duty. As mentioned from a thin capitalisation perspective, the establishment of a two company structure in Australia is our preferred route in any event. 103 On 10 September 2007, Mr Drinkwater responded to an email from Mr Vitullo concerning the timing of the transfer of legal title of Alphapharm. Mr Drinkwater's email stated as follows: Following on from your email below, I understand that the legal transfer of Alphapharm will be effected minutes before the legal transfer of Merck Generics Group BV (despite issues around the timing of cash transfers). This should not give rise to an Australian income tax problem and the debt pushdown should still be effective in Australia. 104 Also on 10 September 2007, Mr Drinkwater sent an email to Mr Vitullo which included comments regarding "timing of steps", "foreign exchange gains/losses" and "incorporation of companies". Among other things, that email noted that it would be necessary to undertake "a thin capitalisation calculation to ensure that the transfer of Note A1 to the Australia 1 would provide sufficient equity value from a thin capitalisation perspective". This document was objected to when the Commissioner sought to tender it. The Commissioner did not press the tender at that time. However, the Commissioner's annexure detailing facts annexed to his closing submissions did refer to this document and the document was not struck through in the marked up index to the court book, prepared by the parties. Accordingly, I have treated it as in evidence, notwithstanding the initial objection and withdrawal of the initial tender. 105 On 11 September 2007, Mr Drinkwater responded to an email from Mr Vitullo, by which Mr Vitullo sought comments on "copies of the draft intercompany notes to effectuate the transfers". Mr Drinkwater's email included comments regarding "transfer pricing", "thin capitalisation" and "legal review". Among other things, that email included the following: I have spoken to my transfer pricing colleagues regarding the terms of the Al and A2 PNotes. An interest rate 400 basis points above the 1 month AUD LIBOR rate may be on the high side of what is acceptable to the Australian Taxation Office. … we would recommend that a benchmarking exercise be carried out for the A2 and Lux8 PNotes to determine an appropriate interest rate. A benchmarking exercise would take approx 4 weeks to determine an appropriate rate and another approx 4 weeks to pull the documentation together as supporting evidence for the interest rate. As part of this exercise we could incorporate terms that would justify a higher interest rate (e.g. duration of the loan, fixed rate, early repayment at discretion of the borrower and subordinating the loan to any external borrowing). … I understand that the thin capitalisation position of the Australian Group (including Alphapharm) will be determined post-acquisition. On this basis, we will need flexibility as to the amount of the A2 or Lux8 PNotes coming into Australia. As such, we recommend that the A2 and Lux8 PNotes contain a clause which enables them to be partly paid down if required. 106 On 12 September 2007, there was a meeting of the Mylan Board. The minutes of that Board meeting record that: "the primary purpose of the meeting was to update the Board with regard to the upcoming closing of the Merck Generics acquisition and related matters"; there was discussion of "the status of the financing"; and certain Merrill Lynch personnel "gave an overview of the debt capital markets including the impact of supply and demand imbalances with respect to newly issued debt". The Commissioner observed that the minutes indicate that no representative of PwC was present at the meeting and emphasised the absence of any record of consideration of the proposed debt push-down structure and its relationship with the debt-servicing capacity of Mylan's subsidiaries, or the Merck Generics entities that were to be acquired. 107 On 13 September 2007, PwC Australia sent a note to PwC US titled "Mylan - Australian acquisition of Alphapharm - List of tax issues considered". That note addressed a number of issues under the headings "US considerations", "Australian income tax" and "Australian stamp duty". Among other things, under the bullet point which reads "Deductibility of interest", PwC Australia referred to "[t]hin capitalisation", "[t]iming of recapitalisation of the Australian group" and "[t]iming of acquisition of Alphapharm compared to acquisition of Merck Generics Group BV". 108 On 20 September 2007, Mr Drinkwater sent an email to Mr Vitullo which included the following comments with respect to the topic of the "timing of legal transfer": It would be preferable from an Australian perspective for a sale and purchase agreement to be drafted in relation to the transfer of Alphapharm and signed before the sale and purchase agreement to transfer Merck Generics Group BV is signed. The actual transfer of Alphapharm would be conditional on the transfer of Merck Generics Group BV. If the agreement is structured this way the debt pushdown would be effective in Australia. If it is not possible for the agreements to be drafted in this way, we expect that the debt pushdown would still be effective (given that it would occur contemporaneously with the acquisition of Merck Generics Group BV) but at a marginally higher risk of being challenged by the Australian Taxation Office. 109 Also on 20 September 2007, Mr Drinkwater sent an email to Mr Vitullo stating that he had received confirmation that the transfer agreement for Alphapharm would be signed before the transfer agreement for MGGBV. Mr Drinkwater then stated: Further, the transfer will be effected before the transfer of Merck Generics Group BV will be effected. As such, the debt pushdown would be effective in Australia. 110 Also on 20 September 2007, Mr Drinkwater responded to an email from Mr Vitullo by which Mr Vitullo requested comments on "revised drafts of the notes". Mr Drinkwater stated as follows: The terms of the Promissory Notes seem fine, although l would add one clause to PNote A2 to allow it to be partially repaid if needed (any partial repayment would need to be funded by an equity injection). This is to provide flexibility from a thin capitalisation perspective. Whilst broadly the thin capitalisation rules work on a ratio of 75% debt to Australian assets, there are adjustments which could impact this. We would not be in a position to accurately forecast the actual allowable debt level until the valuation is complete. 111 On 26 September 2007, there was a further meeting of the Mylan Board. The minutes of the Board meeting indicate that Merrill Lynch representatives and Cravath representatives were present, and PwC representatives were not present. The Commissioner noted that, while management addressed the board on a number of matters concerning the Acquisition, there is no record that the Board was addressed on the debt push-down structure. I do not regard that as a matter of real significance; there is no reason why the Board would, or ought, not have left the detail such as internal financial structuring to management, without requiring a presentation on the topic. 112 On 28 September 2007, Mr Vitullo sent an email to (among others) Messrs Carroll and Drinkwater (both of PwC) with the subject "Final Version of Oct 1 & and latest ver of Post-Acq Slide Decks". In that email (which was an internal, PwC communication), Mr Vitullo said that "[t]he client has asked that we keep the momentum going with regard to the implementation of the post-closing steps as it is critical for Mylan to attaining the intended tax benefits". The email does not make reference to Mylan. Indicating that implementation of the post-closing steps was critical to attaining any non-tax benefits. 113 The Commissioner relied on the above emails in support of his contention that the debt push-down structure was developed by PwC independent of any non-tax (eg, corporate finance or debt capital markets) discipline.

[12]

Acquisition financing - debt 117 Beginning in about April 2007, Mylan's advisors exchanged with counsel for a syndicate of external lenders drafts of a "Commitment Letter" (including Term Sheets) under which the lenders agreed to provide finance for Mylan's acquisition of Merck Generics, and refinancing of its existing indebtedness, through a series of Senior Credit Facilities and an Interim Loan. 118 A draft Term Sheet for the Senior Credit Facilities labelled "CS&M Draft - 4/24/07" contemplated "[s]enior secured credit facilities… in an aggregate principal amount of up to $4,250.0 million" and contained the following definition of "Borrower" (emphasis in original): With respect to the US First Lien Term Loan Facility and the First Lien Revolving Facility, Mylan Laboratories Inc. ("US Borrower"). With respect to the Euro First Lien Term Loan, [ ] (the "Euro Borrower" and, together with the US Borrower, the "Borrowers"). [To be discussed: additional foreign borrowers] 119 A subsequent draft of the Term Sheet for the Senior Credit Facilities dated 26 April 2007 retained the notation in bold, above. 120 In a draft Term Sheet for the Senior Credit Facilities labelled "CS&M 4/27/07", the definition of "Borrower" was as follows (emphasis in original): With respect to the US Term Loan Facility and the Revolving Facility, Mylan Laboratories Inc. ("US Borrower"). With respect to the Euro Term Loan, [ ] a European subsidiary of the US Borrower to be mutually agreed (the "Euro Borrower" and, together with the US Borrower, the "Borrowers"). [To be discussed: If requested by the US Borrower, one or more additional foreign borrowers may be added on terms and conditions to be agreed between the US Borrower and the Lead Arrangers.] 121 On 30 April 2007, the lenders issued a final Credit Facilities Commitment Letter to Mylan. The Term Sheet for the Senior Credit Facilities contemplated "[s]enior secured credit facilities … in an aggregate principal amount of up to $4,850.0 million" and defined "Borrower" as follows (emphasis in original): With respect to the US Term Loan Facility and the Revolving Facility, Mylan Laboratories Inc. ("US Borrower"). With respect to the Euro Term Loan, a European subsidiary of the US Borrower to be mutually agreed (the "Euro Borrower" and, together with the US Borrower, the "Borrowers"). If requested by the US Borrower, one or more additional borrowers (including non-U.S. borrowers) may be added on terms and conditions to be mutually agreed between the US Borrower and the Lead Arrangers. 122 On 11 May 2007 and 18 June 2007, the lenders issued a further Credit Facilities Commitment Letter and an Amended and Restated Credit Facilities Commitment Letter, respectively, to Mylan. In each case, the definition of "Borrower" in the Term Sheet for the Senior Credit Facilities remained the same. Under the Amended and Restated Credit Facilities Commitment Letter, the lenders committed to provide the following Senior Credit Facilities: (a) Euro Term Loan: EUR equivalent of USD 1.6 billion, maturing seven years after the closing date; (b) US Tranche A Term Loan: USD 500 million, maturing six years after the closing date; (c) US Tranche B Term Loan: USD 2 billion, maturing seven years after the closing date; and (d) Revolving Facility: USD 750 million, maturing six years after the closing date. 123 On or about 20 June 2007, representatives from Mylan gave a presentation to the lenders. MAHPL contended that that the presentation included financial projections for Mylan (based on modelling undertaken by Merrill Lynch) that were consistent with the use of 100% of free cash flow across the Group (adjusted for certain specified changes in Mylan's cash balance, and other than free cash flow referable to Matrix) to repay debt. 124 The sources and uses of funds was presented in the following form: 125 The presentation also included a slide depicting "pro forma capitalization" after the Acquisition (but before the anticipated capital raising and repayment of the USD 2.85 billion interim loan). That chart was as follows and showed a debt to equity ratio of 83% debt and 17% equity: 126 The presentation referred to Mylan intending to reduce leverage in the near term through the issuance of a mix of USD 1.5 to 2.0 billion of common stock and mandatory convertible notes. It also referred to dividends being suspended. 127 One of the "modelling assumptions" identified in the presentation was that there would be a "100% cash flow sweep with the exception of Matrix cash flow which is assumed to remain at Matrix subsidiaries". 128 The Commissioner disputed MAHPL's characterisation of this presentation to the lenders. The Commissioner did not accept that the projections for Mylan should be interpreted as evidence of an intention to use 100% free cash flow to repay debt, nor that reference to a "100% free cash flow sweep" should be interpreted as a warranty or representation that all free cash flow across the group would be used to repay debt. Rather, the Commissioner contended that the assumption was intended to be "point in time" such that any 100% cash flow sweep was to be confined to 2007. 129 This dispute about what was conveyed to the lenders is addressed below. As set out there, the Commissioner's construction of the presentations is incorrect. 130 In or about July 2007, Deloitte prepared a memorandum headed "Summary of Third Party Borrowing Considerations". The memorandum stated (inter alia) that: (a) "direct borrowings" by "newly established Mylan entities in Australia, Canada and Japan … in their local currency from third party lenders or local branches of third party lenders" may offer "multiple tax benefits for Mylan relative to Mylan financing these entities through related party loans"; and (b) in Australia, if Mylan used intercompany/related party loans to finance the acquisition of Alphapharm, Merck's Australian subsidiary (rather than "direct borrowing" from a third party lender), interest payments would be subject to a 10% Australian withholding tax resulting in approximately USD 4,500,000 of withholding tax (which Mylan "may or may not" be able to credit for US foreign tax credit purposes). 131 The Commissioner directed attention to the three tax considerations identified and addressed by Deloitte in respect of the Australian direct borrowing option, namely, Australia's corporate tax rate, Australia's thin capitalisation limits, and Australian withholding tax on interest. The Commissioner highlighted Deloitte's focus on the general interest withholding tax rate, being 10%, and certain exemptions from interest withholding tax arising under the AUS-US DTA (that is, the Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with respect to Taxes on Income, US-Australia, signed 6 August 1982 (entered into force 31 October 1983)) and s 128F of the ITAA36. 132 Between 27 September 2007 and 2 October 2007, Mylan executed: (a) a Senior Credit Agreement (SCA) with a syndicate of lenders comprising Lasalle Bank, National Association, The Bank of Tokyo-Mitsubishi UFJ, Ltd., New York Branch, Merrill Lynch, Pierce, Fenner & Smith Incorporated, Citibank, N.A. and JPMorgan Chase Bank, National Association. Under the SCA, the lenders provided the following loans to Mylan and Mylan Luxembourg 5 S.a.r.l. (Lux 5) (which was a disregarded entity for US purposes) to finance the Acquisition and refinance Mylan's existing indebtedness: (i) US Trance A Term Loan to Mylan: USD 500 million, maturing on 2 October 2013; (ii) US Tranche B Term Loan to Mylan: USD 2 billion, maturing on 2 October 2014; (iii) Euro Term Loan to Lux 5: EUR 1,130,703,095.33, maturing on 2 October 2014; and (iv) Revolving Facility to Mylan or Lux 5: USD 750 million, maturing on 2 October 2013; (b) "Term Borrowing Requests" under the SCA between Mylan, Lux 5 and the lenders, requesting the following from the lenders on 2 October 2007: (i) the Euro equivalent of USD 1.6 billion under the Euro Term Loan; (ii) USD 500 million under the US Tranche A Term Loan; and (iii) USD 2 billion under the US Tranche B Term Loan; (c) an "Irrevocable Funding Indemnity Agreement" in relation to the Eurocurrency loans; (d) an "Interim Loan Borrowing Request" under a (then) draft Interim Loan Agreement between Mylan and the lenders, requesting a loan of USD 2.85 billion; and (e) an "Irrevocable Funding Indemnity Agreement" in relation to the (then) draft Interim Loan Agreement. 133 On 2 October 2007, Mylan entered into an Interim Loan Agreement with Merrill Lynch and other lenders for a principal amount of USD 2.85 billion. 134 As addressed further below, funds borrowed under the SCA and Interim Loan Agreement were on-lent to other Mylan group entities to fund the acquisition of MGGBV and other Merck companies. 135 On or about 20 December 2007, the SCA was amended and restated with effect from 28 December 2007. The amended and restated SCA added a number of financial institutions to the syndicate of lenders, split the Euro Term Loan into two tranches and converted a portion of US Tranche A Term Loans to US Tranche B Term Loans. The outstanding amounts and maturity dates under each of the loans became as follows: (a) US Tranche A Term Loan to Mylan: USD 312.5 million, maturing on 20 October 2013; (b) US Tranche B Term Loan to Mylan: USD 2.556 billion, maturing on 2 October 2014; (c) Euro Tranche A Term Loan: EUR 350,414,947.37, maturing on 20 October 2013; (d) Euro Tranche B Term Loan: EUR 525 million, maturing on 2 October 2014; and (e) Revolving Facility: USD 300 million, maturing on 2 October 2014. 136 The US Tranche A Term Loan and the US Tranche B Term Loan bore interest at LIBOR plus 3.25% or at a base rate (defined to be equal to the greater of (a) prime rate and (b) the Federal Funds Effective rate plus one half of one percent) plus 2.25%. The Euro Term Loans bore an interest rate of the Euro Interbank Offered rate (EURIBO) plus 3.25%. Borrowings under the Revolving Facility bore interest at LIBOR (or EURIBO) plus 2.75%. The interest rates could vary based on a calculation of the borrowers' consolidated leverage ratio.

[13]

Credit ratings and engagement with ratings agencies 137 As at February 2007, Mylan had a credit rating of BBB- (Standard & Poor's Ratings Services (S&P)) and Ba1 (Moody's Investors Service (Moody's)). 138 On or about 14 May 2007, following the announcement of the Acquisition, S&P downgraded Mylan's corporate credit rating to BB+. 139 Also on 14 May 2007, Mylan gave a presentation to Moody's Rating Agency titled "Mylan to Acquire Merck Generics - Creates a World Class Global Generics Leader" which resulted in Moody's announcing that it had placed Mylan under review for possible downgrade. 140 The Commissioner relied on the following statements made in the presentation to Moody's regarding Mylan's intentions as to target gearing levels: • Mylan intends to reduce leverage through the issuance of a mix of at least $1.5bn - $2.0bn of common stock and mandatory convertible securities shortly after close • Strong free cash flow expected to further de-lever balance sheet • Near-term leverage target of less than 6x Net Debt/EBITDA • Long-term leverage target of less than 4x Net Deb/EBITDA 141 The presentation to Moody's included a slide titled "Sources and Uses and Pro Forma Capitalization". Like the similar presentation made to the lenders (referred to above), this slide set out the anticipated pro forma position immediately after the Acquisition as follows: 142 On 27 and 28 September 2007, Mylan delivered presentations to Moody's and S&P. 143 The presentation included reference to the USD 2.85 billion unsecured interim loan, and also stated "Mylan intends to reduce leverage in the near term through the issuance of a mix of $1.5 - $2.0 billion of common stock and mandatory convertible preferred equity securities". The presentation included the following slides showing the sources and uses of funds, the pro forma capitalisation following the Acquisition, as well as the pro forma capitalisation after the anticipated equity raising and the anticipated issue of USD 850 million in senior notes (referred to as "PF Permanent") as follows: 144 Each of the presentations delivered to Moody's and S&P on 27-28 September 2007: (a) incorporated modelling provided by Merrill Lynch; and (b) disclosed that Mylan anticipated that its capital structure would: (i) comprise approximately 98.62% debt immediately following the 2 October 2007 acquisition of Merck Generics (including the interim loan); and (ii) reduce to 73.84% debt by 31 December 2007 following an anticipated USD 2 billion capital raising post acquisition. 145 The significant step up in the percentage debt immediately following completion (as compared with percentages presented earlier) was not explored by the parties but appears to be driven by the figure presented for "Shareholders' Equity" following closing being significantly lower than in previous iterations of the equivalent slide in other presentations. This change was not explained, but I note that MAHPL's submissions generally relied on the lower debt percentage presented earlier (81.5% to 83% as against the 98.62% figure in the 27-28 September 2007 presentations). 146 As with other presentations: (a) MAHPL contended that these 27-28 September 2007 presentations to the ratings agencies contained financial projections and ratios which were based on, and assumed, that free cash flow from Mylan's international operations (with specified adjustments, and excluding free cash flow referable to Matrix) would be used to service and repay debt; but (b) the Commissioner did not accept this to be correct and referred to those matters which he said demonstrate that Mylan did not, prior to the Acquisition, hold an expectation of repatriation of 100% of free cash flow. 147 Following Mylan's issue of common and preferred stock on 13 November 2007, S&P downgraded Mylan's corporate credit rating from BB+ to BB- and senior unsecured debt rating from BB+ to B. Moody's downgraded Mylan's corporate family rating from Ba1 to B1 and assigned B1 ratings to Mylan's new senior secured credit facilities.

[16]

Expected repatriation of funds and the OFL position 150 Because of the borrowings necessary to fund an approximately USD 7 billion acquisition, the Mylan group's interest expenses rose from USD 31.3 million in 2006, the year before the Acquisition, to USD 357 million in 2008, the first full year after the Acquisition. The Acquisition was, on any view, very highly leveraged. As noted above, the gearing ratio for Mylan as a whole was expected to be over 80% debt immediately following the Acquisition (and prior to capital raising and repayment of the interim loan). This represented a significant increase from the pre-Acquisition gearing (just over 50% debt). 151 MAHPL submitted that Mylan recognised that its US-based free cash flow would be insufficient to service the debt for the acquisition of Merck Generics (and cover its other post-Acquisition cash outflows), and that free cash flow from the acquired Merck Generics subsidiaries would need to be used for this purpose. It relied on the projections contained in the Merrill Lynch modelling and the fact of that modelling having been disclosed by Mylan to parties including the lenders and ratings agencies. Those presentations referred to the "modelling assumption" that there would be a "100% cash flow sweep" with the exclusion of Matrix cash flow. 152 The Commissioner did not accept that Mylan held this view and took issue with the characterisation of the presentations to the ratings agencies and lenders as evidence of a representation or warranty from Mylan that 100% of free cash flow would be repatriated to service external debt. The Commissioner relied on a number of matters said to provide "context" to the modelling assumption (regarding 100% cash flow sweep) contained in the presentations. 153 First, the Citigroup presentation to Mylan dated 25 June 2007, titled "Sizing Hedge Notional". The Commissioner pointed to: (a) the statement, under the heading "Assumptions", that "50% of FCF available is used to pay down debt (as opposed to no debt paydown from FCF, other than mandatory amortizations, for the first 3 years)"; (b) the statement under the heading "Swap Notional", that "50% of FCF used to pay down debt", following which Citigroup included modelling for 2008 to 2014 which applied a 50% pay down to identified amounts representing "FCF Available for Debt Paydown"; and (c) there being no reference in the Citigroup presentation to an intention to repatriate 100% of the free cash flow from Mylan's subsidiaries to the US. 154 The Commissioner's reliance on this document fails to acknowledge that it was a presentation concerning potential floating to fixed hedging strategies, and also that the assumptions referred to regarding free cash flow were expressed in terms of paying down debt. I do not take a reference to paying down debt to refer to merely servicing the interest due and amortisation of the principal in accordance with the terms of the SCA, but to refer to reducing the debt over and above the minimum repayment terms of the SCA. An assumption regarding using half the free cash flow to pay down the principal does not invite an inference that the other half was not also expected to be required to be repatriated for debt servicing obligations. The calculations in the Citigroup presentation support this reading as they show the principal being reduced on a different schedule from the figures after the application of the "pay down" based on the use of 50% of the free cash flow for that purpose. 155 The second matter of context referred to by the Commissioner was the PwC slide deck dated 3 August 2007 which: (a) stated, under the heading "Tax Integration Goals/Objectives", "[a]llow for redeployment of foreign excess cash via tax efficient Treasury Centre" and detailed how Lux 1 may be used to achieve that end; and (b) did not refer to an intention to repatriate 100% of the free cash flow from Mylan's subsidiaries to the US. 156 The third matter of context referred to by the Commissioner was Mylan's Form 1118 for the year ending 31 December 2008 which disclosed receipt of: (a) total "deemed dividends" from Luxembourg of USD 373,312,934 (plus a gross-up of USD 268,000); and (b) total "other dividends" from Bermuda of USD 1,000,000. 157 Specific findings of fact on Mylan's repatriation expectations and what was conveyed by these presentations are set out below (at paragraphs 267 to 288). 158 Mylan also had (and expected that it would continue to have) an "Overall Foreign Loss" (OFL) in the US. An OFL is a US tax law concept that limits the availability of foreign tax credits (FTCs) to be applied against taxable US income. At the time of the Acquisition, Mylan was in an unfavourable OFL position which was expected to limit its ability to claim FTCs in future taxable years. The effect of this was that dividends or other payments from the earnings of the acquired subsidiaries for the benefit of the US parent were expected to be subject to US tax at the rate of 35% with no, or no full, credit for foreign taxes paid. 159 The expected OFL was described by Paul Martin (from Mylan) as a "mega OFL" in an email he sent Jeffrey Mensch (from Deloitte) on 19 July 2007. In that email, Mr Martin stated: While we may not go down the route of local-country borrowing (i.e., may borrow in the U.S., do as much I/C financing as possible, and manage the mega-OFL with basis reduction, and live with the W/H tax cost), we need to give the Lenders a sense of where things might go, so they can say whether things do or don't work. 160 As at 31 December 2008, Mylan had an OFL of over USD 189 million.

[17]

Implementation of the Acquisition - Australia 161 Mylan implemented the acquisition of the global Merck Generics business generally in accordance with a "Step Plan" prepared by its advisors titled "Merck Tax Integration - October 2007, Version 17 11.2.07". The structure immediately following the Acquisition is depicted in the following diagram: 162 A simplified version of the structure, so far as it concerned Australia, was presented by the Commissioner in the following diagram: 163 In relation to the acquisition of Alphapharm in Australia, the following steps took place. 164 On 10 August 2007, MAPL was incorporated with one ordinary share which was held by Mylan International Holdings Inc (MIHI). 165 On 13 September 2007, MAHPL was incorporated with one ordinary share held by MIHI. 166 On 17 September 2007, MIHI transferred its share in MAPL to MAHPL in exchange for one ordinary share in MAHPL. On the same day, MAHPL and MAPL formed a consolidated group for Australian income tax purposes. 167 On 2 October 2007 and before the acquisition of the shares in MGGBV: (a) Lux 1 issued Promissory Note Lux 1 (PN Lux 1) to MAPL in exchange for Promissory Note A1 (PN A1) and PN A2; (b) MGGBV transferred its ordinary and redeemable preference shares in Alphapharm to MAPL in exchange for PN Lux 1; and (c) as a result, Alphapharm joined the MAHPL tax consolidated group. 168 The principal amount of PN Lux 1 was: EUR 670,000,000, less those amounts owed by Alphapharm Pty. Ltd. to related parties as of the date of this promissory note, estimated as of such date to be the equivalent of EUR 17,548,000 (together, "the principal"). The principal shall be automatically adjusted with retroactive effect and considered finally determined to equal the amount of the valuation as of October 1, 2007 of Alphapharm Pty. Ltd. by PricewaterhouseCoopers LLP using generally accepted valuation principles. 169 The interest rate for PN Lux 1 was: 400 basis points above the 1-month EUR LIBOR rate in effect at any time, provided, however, that such interest rate, as well as other material terms of this promissory note, shall be finally determined as agreed upon between the parties within 90 days of the execution of this promissory note. If no such agreement is achieved, the interest rate shall be finally determined based upon an independent transfer-pricing study of arm's length terms. (Emphasis in original.) 170 PN A1 and PN A2 represented 25% and 75% of the face value of PN Lux 1, respectively. The principal amount of PN A1 was: the AUD equivalent of EUR 167,500,000 … to be automatically adjusted with retroactive effect and considered finally determined to equal 25% of the amount of the valuation as of October 1, 2007 of Alphapharm Pty. Ltd. by PricewaterhouseCoopers LLP using generally accepted valuation principles. The AUD equivalent of the EUR value shall be based upon the imputed AUD-EUR rate for the date this promissory note is executed, based on the AUD-USD and USD-EUR rates published in the Wall Street Journal with respect to that date. 171 The principal amount of PN A2 was: the AUD equivalent of EUR 502,500,000, less those amounts owed by Alphapharm Pty. Ltd. to related parties as of the date of this promissory note, estimated as of such date to be the equivalent of EUR 17,548,000 … The principal shall be automatically adjusted with retroactive effect and considered finally determined to equal 75% of the amount of the valuation as of October 1, 2007 of Alphapharm Pty. Ltd. by PricewaterhouseCoopers LLP using generally accepted valuation principles. The AUD equivalent of the EUR value shall be based upon the imputed AUD-EUR rate for the date this promissory note is executed, based on the AUS-USD and USD-EUR rates published in the Wall Street Journal with respect to that date. 172 The interest rates for both PN A1 and PN A2 were as follows: 6.25 basis points above the rate paid by Mylan Luxembourg 1 Sarl to Mylan Luxembourg 2 Sarl on that instrument known by such parties as "Note Lux1" as may be in effect at any time or at such rate as may be ultimately determined by the Luxembourg Taxing Authority, provided, however, that such interest rate, as well as other material terms of this promissory note, shall be finally determined as agreed upon between the parties within 90 days of the execution of this promissory note. If no such agreement is achieved, the interest rate shall be finally determined based upon an independent transfer-pricing study of arm's length terms. (Emphasis in original.) 173 Each of PN A1, PN A2 and PN Lux 1: (a) was "prepayable at any time, in whole or in part, without permission or penalty"; (b) had a maturity date of 30 September 2014 which could "be extended with agreement from both parties"; (c) provided for interest to be calculated and accrued on either a calendar quarter basis (PN A1 and PN A2) or monthly basis (PN Lux 1); (d) provided for any unpaid interest amounts to be added to the principal balance and to bear interest; and (e) provided for interest to be payable upon demand to the holder of the note. 174 On or about 19 December 2007, 24 January 2008, 29 May 2008 and 30 May 2008, PwC produced draft valuations of Merck Generics (including Alphapharm) as at 2 October 2007. A final valuation was prepared on 2 February 2009. 175 Later, on 2 October 2007: (a) Lux 2 acquired all the shares in MGGBV; and (b) MGGBV distributed PN Lux 1 to Lux 2. The PwC step plan said that the purpose of this together with the distribution of other notes was "to consolidate intercompany financing in Mylan Luxembourg 2 S.a.r.l. which will serve as Mylan's foreign treasury centre".

[19]

Acquisition financing - equity raising 178 Between about July and November 2007, Merrill Lynch advised Mylan on strategies for raising equity to partly fund or refinance the Acquisition. The initial Merrill Lynch presentation was dated 20 July 2007. The Merrill Lynch presentation modelled an "all debt" scenario, noting that was the base case during the board approval process for the Acquisition. It also modelled raising USD 1 billion in equity (only with mandatory convertible notes), and raising USD 1.5 billion (USD 1 billion in mandatory convertible notes and USD 500 million in common stock). Merrill Lynch recommended the latter option, noting that there was the ability to raise USD 1.5 to 2 billion "with greenshoe/upsize". 179 In August 2007, Mylan internally circulated a model with no equity raising (ie all debt). 180 On or about 25 September 2007, representatives from Mylan made a presentation to the company's external lenders. The presentation noted (amongst other matters) that Mylan remained committed to reducing debt in the near term by USD 1.5 billion to USD 2 billion through the issuance of equity and equity-linked proceeds. 181 On or about 28 September 2007, Merrill Lynch provided Mylan with the following: (a) a financial model scenario entitled "Pro Forma Model Case 2", which was a scenario under which there would be no divestitures and USD 2.0 billion in equity raised (including USD 1.5 billion of mandatory convertible notes); (b) a financial model scenario entitled "Pro Forma Model Case 10", which was an all-debt scenario under which there were no divestitures and no new equity raised; (c) a financial model scenario entitled "Pro Forma Model Case 13", which was a scenario under which there were no divestitures and USD 1.5 billion in equity raised (including USD 1.2 billion of mandatory convertible notes); (d) a summary of key metrics for the three scenarios referred to in subparagraphs (a) to (c) above; and (e) a copy of the Excel model from which the above documents were generated, with the file name "v146". 182 There was another Merrill Lynch presentation dated 28 October 2007. This presentation, titled "Equity Financing Considerations", modelled a USD 1.5 to 2.5 billion equity raising (with different blends of mandatory convertible notes and common stock), in addition to an anticipated issue of high yield bonds in amounts ranging from USD 350 million to USD 1.35 billion. 183 As the Commissioner noted, Merrill Lynch prepared a large number of models. However, what is most relevant is the modelling provided to the Mylan Board via its Finance Committee. That modelling was provided in a Merrill Lynch presentation dated 29 October 2007, titled "Confidential Discussion Material Prepared for: Finance Committee of the Board of Directors of Mylan. Regarding: Equity Financing". The pack modelled various different equity raising scenarios, which went as high as USD 3.5 billion in equity. However, the presentation did not recommend attempting to raise equity at that level, but recommended (by dotted line selection) the USD 2 billion equity raising with modelling assumptions drawing out the likely adverse impact of higher equity raising attempts on the assumed credit rating and rate on the mandatory convertible note component. 184 On 30 October 2007, the Mylan Board met. The Board: (a) was informed that it was the underwriters' recommendation that USD 2 billion in equity be raised, comprising USD 1.4 billion of mandatory convertible preferred stock and USD 600 million of common stock; and (b) resolved (broadly) to proceed with an offering of common stock, mandatory convertible preferred stock and senior notes, with a net of up to USD 3.0 billion. 185 On 13 November 2007, both the Finance Committee of the Mylan Board and the full Mylan Board met. Minutes of the meeting record the following: (a) "of the 40 million shares of common stock and [USD] 1.4 billion of mandatory convertible preferred stock expected to be sold, the demand had been 84 million shares and [USD] 10 billion, respectively"; (b) Merrill Lynch was recommending: (i) "an increase in the number of shares of common stock to 53.5 million at a [USD] $14 price per share"; and (ii) an offering of USD 1.86 billion of preferred stock; (c) "the proceeds were nearly a 50% increase from where we began" and "management was not recommending going to the high-yield market as had been initially anticipated" (i.e. raising debt though an issue of notes); and (d) it was resolved that the recommended equity offerings (of common stock and preferred stock) at both the amounts and the prices mentioned in subparagraph (b), above, be approved by the Finance Committee in all respects. 186 On 13 November 2007, Mylan issued approximately 53.5 million shares of common stock and 1.86 million shares of 6.50% mandatory convertible preferred stock at USD 1,000 per share. The total raised was approximately USD 2.8 billion, which enabled Mylan to repay the USD 2.85 billion interim loan in full on 19 November 2007. 187 Mylan did not proceed with an offering of notes into the high-yield market.

[20]

Interest rate swaps 188 Between May and November 2007, Mylan received advice from advisors including Merrill Lynch, JP Morgan, Goldman Sachs and Calyon Credit Agricole on interest rate risk management. The advisors: (a) acknowledged that, following the acquisition of Merck Generics, Mylan would have significant floating rate exposure (under the SCA); and (b) recommended that Mylan decrease its interest rate risk by hedging (swapping a substantial portion of its floating rate liabilities to fixed liabilities). 189 The advice received by Mylan was to fix a very large portion of the floating rate exposure. In a presentation titled "Interest Rate Risk Management Discussion" dated 24 May 2007, Merrill Lynch advised that, at 70% floating rate exposure, Mylan was "significantly exposed to interest rate risk" and recommended "[l]owering floating rate exposure to 7%". Also in May 2007, JP Morgan recommended a "70%/30% fixed/floating rate" to increase certainty around Mylan's future interest expenses. Similar recommendations were made by Goldman Sachs, who recommended in November 2007 that Mylan "target about 80% to 90% fixed rate debt". Earlier (in July 2007) Calyon Credit Agricole had recommended Mylan consider hedging "at least 50% of its floating rate debt". 190 The Commissioner contested the assertion by MAHPL that Mylan "sought" the interest rate risk advice. The Commissioner said he did not accept that the advice was "sought" by Mylan and put MAHPL to proof on this point. Quite why the point was not accepted was not explained. Should it matter, my factual finding is that the advice was sought by Mylan. I infer that to be the case on the basis that it is wholly improbable that the authors of that advice sent it to Mylan unsolicited. The Commissioner also made the point, which I accept, that the advice went to the interest rate management at the group level, cf being advice about the floating and fixed rate exposure of any particular subsidiary. 191 In December 2007, Mylan entered into five interest rate swaps to fix the interest rate on USD 1 billion of its USD denominated debt under the SCA. Those swaps were as follows: (a) on 17 December 2007 - four swaps totalling USD 500,000,000 arranged by Merrill Lynch, syndicated in equal proportions of USD 125,000,000, with Merrill Lynch Capital Services, Inc., Goldman Sachs Capital Markets, JP Morgan Chase Bank NA and Citibank NA respectively. Each swap had an effective date of 28 December 2007, a maturity date of 30 December 2010 and a fixed interest rate of 4.234% per annum, payable quarterly; and (b) on 20 December 2007 - a USD $500,000,000 swap with Citibank NA, with an effective date of 28 December 2007, a maturity date of 30 December 2010 and a fixed interest rate of 4.00% per annum. 192 According to MAHPL's calculations, these two hedges resulted in Mylan having fixed an average interest rate of 7.37% until December 2010 (including the 3.25% spread under the SCA). 193 In the 2008 calendar year, Mylan entered into further interest rate swap arrangements to fix the rate on more of its USD denominated debt under the SCA. As at 31 December 2008, Mylan had swapped a total of USD 2 billion of floating rate debt to fixed rate debt. 194 In February 2009, Mylan executed an additional EUR 200 million of notional interest rate swaps in order to fix the interest rate on a portion of the Euro denominated debt under the SCA.

[21]

Post-acquisition steps - Australia 195 On 26 December 2007, the Mylan group implemented a series of steps which had the effect of capitalising MAPL with further equity. Those steps included the following: (a) Mylan Bermuda Limited (Mylan Bermuda) transferred two ordinary shares in MAHPL (being all of its shares in MAHPL) to Mylan Gibraltar 4 Limited (Gibraltar 4) by way of capital contribution; (b) Gibraltar 4 transferred two ordinary shares in MAHPL (being all of its shares in MAHPL) to Lux 1 in exchange for EUR denominated promissory note Lux 7; (c) Lux 1 subscribed for 20,000 ordinary shares in MAHPL in exchange for AUD denominated PN A1; (d) MAHPL subscribed for 20,000 ordinary shares in MAPL in exchange for PN A1; and (e) PN A1 was cancelled as a capital contribution to MAPL. 196 Following these steps, MAPL's only debt instrument was PN A2. Steps were then taken in late December 2008, which resulted in AUD 105,087,273 being paid down on the principal of PN A2, and AUD 122,815,968 of the capitalised interest on PN A2 being paid. There was a letter issued by Lux 1 to MAPL (dated 31 December 2008) that demanded "full repayment of an amount of the accrued and outstanding interest on Note A2 of AUD 122,815,968" (which amount implies interest being calculated at a fixed rate of 10.15%). The letter from Lux 1 demanded payment by 31 December 2008. 197 On 30 December 2008, the directors of MAHPL made resolutions authorising: (a) the transfer by Lux 1 of 20,002 shares in MAHPL to Lux 2; (b) a capital contribution by Lux 2 to MAHPL in the form of a promissory note A3 (PN A3); and (c) the assignment by MAHPL of PN A3 to MAPL by way of capital contribution by MAHPL to MAPL. 198 Also on 30 December 2008, the directors of MAPL made certain resolutions by circulating resolution regarding: (a) the capital contribution of PN A3 by MAHPL; and (b) the assignment of PN A3 to Lux 1 following a demand issued by Lux 1 to MAPL requiring payment of outstanding interest and a repayment of principal on PN A2.

[23]

Fixing of the interest charged under, and subsequent formal amendment of, PN A1 and PN A2 200 PN A1 and PN A2 contemplated that their interest rates would be determined within 90 days of 2 October 2007. On MAHPL's account of the facts, during November and mid-December 2007, the interest rates applicable on PN A1 and PN A2 were determined to be a fixed rate of 10.15%. 201 The Commissioner contended that the weight of the evidence supports a finding that no formal decision had been made to fix the interest rate on PN A2 at 10.15% until October 2008 when a decision was made to formally vary the interest rate on PN A2 with retrospective effect. 202 The Commissioner pointed to an email exchange between Mylan personnel dated 6 November 2007 regarding the topic of the conversion of a variable interest rate on intercompany debt into a fixed rate. In the course of that correspondence, Mr Gregory Weixel (of Mylan) wrote: I already have the loans in finavigate at the fixed rate plus 350. In process of putting together loan agreements based on the Matrix templates. 203 The Commissioner also relied on two emails dated 28 November 2007. The first being an email sent by Mr Fraser to Miharu Maeda-san (of the Merck Generics Japanese entity acquired by Mylan) enquiring as to the intercompany loan documentation. The second email was between Mr Drinkwater (PwC) and Mr Martin (Mylan) arranging a telephone conference to discuss the interest rate on PN A2. Relevantly, Mr Martin wrote: Did we ever discuss the Australian transfer-pricing documentation requirements associated with converting the current, variable rate on Note A2 to a fixed-rate equivalent? I don't believe we did, and if not would like to have such call, as we are obliged in the Notes to arrive at an arm's length rate by end of this year. 204 Mr Drinkwater responded: We didn't discuss the conversion from variable to fixed and the Australian documentation required. That said, we did briefly touch on the need for the transfer pricing to be right given the large amounts involved. I thought a global transfer pricing piece was going to be run from the US in this regard…? 205 Between 1 and 13 October 2008, various PwC personnel exchanged email correspondence about undertaking analysis to "verify whether a 7 year AUD loan at 10.15% completed last October is arm's length". 206 The Commissioner relied on a 2 October 2008 email from Mr Vitullo to Mr Tim Hogan-Doran (PwC Australia) and Mr Carroll which included the following: Steve and I took some time today to review the structure to see if there was a US reason for the establishment of two entities in Australia. As we actually did the transaction it turns out that there was no specific US tax reason for Mylan Australia Holding Pty Ltd (the DRE which owns Mylan Australia Pty Ltd). However, we both remembered that when we were designing the acquisition structure there was a great deal of discussion and planning about whether it would be more advantageous from a US tax perspective for the debt between Lux and Aus to be regarded or disregarded debt. It is very likely that the reason for the two Aussie entities was to allow Mylan the flexibility to either regard or disregard this I/C loan between Lux and Aus. For example, as we did the transaction, Mylan Aus Pty issued notes A1 adn [sic] A2, resulting in the debt with Lux being a regarded loan for US tax purposes. On the other hand, it would have been just as easy to have A1 and A2 notes issued by the DRE in which case the Lux - Aus I/C debt would have been disregarded. By forming two Australian entities, and treating one as a DRE and the other as a corporation, we maintained total flexibility regarding the status of the debt. We utilized a financial model which we modified continuously, almost all the way up to the date of the acquisition, to evaluate the US tax effectiveness of either regarded or disregarded debt. Also, keep in mind that we always knew that post acquisition Mylan was going to check the box to elect to treat Alphapharm as a DRE which means for US tax purposes it would be liquidated. This was necessary to insure [sic] that we had a fallback position that the acquisition would be a D reorganization for US tax purposes and not a Section 304 transaction if the IRS were to argue that the acquisition was not a qualified purchase. Because Alphapharm was going to be treated as liquidated for US tax purposes with the check the box election, we had to have one of the Aussie entities be a regarded corp to insure that Alphapharm's E&P did not flood into the top Bermuda holding company. 207 It was not clear what the Commissioner sought to draw from this email in relation to the issues concerning fixing the interest under PN A2. 208 The Commissioner also referred to an amount of AUD 8,450,455 having been recorded in a general ledger account reflecting "September interest on I/C Notes" with a backdated posting date of 30 September 2008. 209 The demand issued by Lux 1 to MAPL on 31 December 2008, requiring payment of some principal and accrued interest (referred to above) demanded an amount of interest, the calculation of which shows the application of the 10.15% fixed interest rate. 210 Amendments were formally made to PN A1 and PN A2 by instruments dated 8 January 2010. Amendments were made, with contractually retroactive effect from 2 October 2007, to: (a) PN A1 to have a principal of AUD 316,337,043 and a fixed interest rate of 10.15%, with retroactive effect to 2 October 2007; (b) PN A2 to have a principal of AUD 923,205,336 and a fixed interest rate of 10.15%, with retroactive effect to 2 October 2007; and (c) PN Lux 1 to have a principal of EUR 774,445,358 and a fixed interest rate of 7.81%, with retroactive effect to 2 October 2007. 211 MAHPL characterised these amendment instruments as paperwork tidying up the position that had been substantively agreed and implemented in December 2007. While the Commissioner accepted there was a meeting of minds and the rate of 10.15% was agreed in principal by October 2008, he continued to point to formal amendments to PN A2 only having been executed in January 2010. 212 I address and determine the factual controversy concerning when the interest rate on PN A2 was set and applied below (see paragraphs 476 to 506).

[33]

The narrower scheme and secondary and tertiary counterfactuals 224 In the alternative to the wider scheme, the Commissioner relied on the narrower scheme, which comprises steps (c), (d), (f), (g) and (h) of the wider scheme. That is, the wider scheme allows for the incorporation of MAHPL and MAPL, the amendment of the original SPA to include Alphapharm as the target entity and MAPL as the purchaser, and the subsequent choice to form a tax consolidated group. 225 In respect of the narrower scheme, the Commissioner relied (in the alternative) on two counterfactuals, referred to as the secondary and tertiary counterfactuals. The secondary counterfactual posits that, had the narrower scheme not been carried out, it might reasonably be expected that the following would have occurred: (a) Mylan might reasonably be expected to have had MAPL borrow under the SCA on the following terms: (i) a set principal amount being the AUD equivalent of approximately EUR 356.239 million, that is AUD 571.72 million, reflecting the debt level for the Mylan group's 2007 acquisitions, making the loan an AUD obligation at a floating interest rate above an AUD base rate such as AUD LIBOR or BBSW; (ii) alternatively to paragraph 225(a)(i), a set principal amount of approximately EUR 356.239 million, that is AUD 571.72 million, reflecting the debt level for the Mylan group's 2007 acquisitions (or USD equivalent thereof) which is then swapped into an AUD equivalent amount at a floating interest rate above an AUD base rate such as AUD LIBOR or BBSW; (iii) with a requirement that the debt be fully serviced with interest amounts being paid, that is there would be no provision for interest to be capitalised, and principal repayments would be made (consistent with the SCA which required the amortisation of agreed levels of principal over time and at a rate not less than the actual principal repayments made under the SCA and the actual principal repayments made under PN A2); (iv) with a credit margin based on Mylan's credit rating, as applicable from time to time; (v) with no retroactive adjustments to the principal amount or interest rates; (vi) which was secured by certain assets of the Mylan group; and (vii) which was guaranteed by certain entities of the Mylan group; and (b) MAPL would then acquire the shares in Alphapharm (using the funds raised by the borrowing by MAPL referred to in paragraph 225(a) above together with other funds provided to MAPL, directly or indirectly, by Mylan in the form of equity or non-interest bearing loans). 226 The tertiary counterfactual is materially the same as the secondary counterfactual, save that it posits the lender being Mylan or a US subsidiary of Mylan. In closing, the Commissioner abandoned the non-interest bearing loans aspect of the secondary and tertiary counterfactuals.

[34]

MAHPL's counterfactuals 227 For its part, MAHPL relied, in the alternative, on two counterfactuals to the wider and narrower schemes: (a) Counterfactual A - that MAPL might be expected to have funded the acquisition of Alphapharm using 25% equity injected by its parent and 75% debt borrowed from Mylan (or a US subsidiary, rather than Lux 1) on the same, or similar, terms as those set out in PN A2. (b) Counterfactual B - that MAPL might be expected to have funded the acquisition of Alphapharm using 25% equity injected by its parent and 75% debt borrowed from an external lender or lenders. 228 In elaborating on its counterfactuals in closing, MAHPL added further detail. 229 Counterfactual A was said to consist of the following steps: (a) in or about August or September 2007, MAPL and MAHPL would or might reasonably be expected to have been incorporated and formed a tax consolidated group (with MAHPL as head company); (b) at some time on or before 1 October 2007, the SPA would or might reasonably be expected to have been amended to (inter alia) include Alphapharm as an "Additional Target Company" and MAPL as an "Additional Purchaser"; (c) a subsidiary of Mylan would or might reasonably be expected to have capitalised MAPL with equity in an amount equal to 25% of the estimated value of Alphapharm as of 1 October 2007, subsequently to be adjusted to 25% of the valuation of Alphapharm as of 1 October 2007 by PwC using generally accepted valuation principles, or such other amount that satisfied the thin capitalisation rules for Australian income tax purposes; (d) the balance of the funding required to purchase the shares in Alphapharm would or might reasonably be expected to have been provided to MAPL by Mylan (or a US subsidiary) by way of debt on the terms of PN A2 or on substantially similar terms; (e) on or about 2 October 2007, MAPL would or might reasonably be expected to have acquired 100% of the shares in Alphapharm from MGGBV under the SPA, and Alphapharm would or might reasonably be expected to have joined the MAHPL tax consolidated group; (f) in or about September 2014, the debt from Mylan or its US subsidiary would or might reasonably be expected to have been extended or refinanced on terms broadly equivalent to those in PN A4; and (g) in the relevant years, MAPL would or might reasonably be expected to have paid interest to Mylan or its subsidiary (and/or capitalised interest), and made repayments, in line with what actually occurred under PN A2 and PN A4. 230 Counterfactual B was said to involve the following steps: (a) in or about August or September 2007, MAPL and MAHPL would or might reasonably be expected to have been incorporated and formed a tax consolidated group (with MAHPL as head company); (b) at some time on or before 1 October 2007, the SPA would or might reasonably be expected to have been amended to (inter alia) include Alphapharm as an "Additional Target Company" and MAPL as an "Additional Purchaser"; (c) a subsidiary of Mylan would or might reasonably be expected to have capitalised MAPL with equity in an amount equal to EUR 213.75 million (approximately AUD 342 million); (d) MAPL would or might reasonably be expected to have been a borrower under the SCA, borrowing EUR 641.25 million or more likely the USD equivalent under the "Term Loan B" facility (and correspondingly, Mylan, or Lux 5, would have borrowed a lesser amount under that facility); (e) MAPL would or might reasonably be expected to have entered into an internal cross-currency and interest rate swap(s) (ie within the Mylan group) to convert its borrowings under the SCA to AUD (approximately AUD 1.029 billion) and to fix the interest rate on the debt at a rate of at least 10.15%; (f) MAPL would or might reasonably be expected to have paid Mylan (and/or US subsidiaries of Mylan) an arm's length guarantee fee in respect of the guarantee provided to it for its borrowings under the SCA; (g) on or about 2 October 2007, MAPL would or might reasonably be expected to have acquired 100% of the shares in Alphapharm from MGGBV under the Amended SPA, and Alphapharm would or might reasonably be expected to have joined the MAHPL tax consolidated group; (h) in or about September 2014, MAPL's debt under the SCA would or might reasonably be expected to have been extended or refinanced on arm's length terms; and (i) in the relevant years, MAPL would or might reasonably be expected to have incurred a cost of funds equivalent to, or greater than, the claimed deductions.

[36]

The proper approach on the authorities 232 Part IVA operates by permitting the Commissioner, in specified circumstances, to cancel a "tax benefit". Accordingly, the existence of a tax benefit is an essential component in the application of Pt IVA to a particular scheme: Federal Commissioner of Taxation v Hart (2004) 217 CLR 216 (Hart) at [33] (Gummow and Hayne JJ); Federal Commissioner of Taxation v Trail Bros Steel & Plastics Pty Ltd (2010) 186 FCR 410; [2010] FCAFC 94 (Trail Bros) at [23] (Dowsett and Gordon JJ); CPH Property Pty Ltd v Federal Commissioner of Taxation (1998) 88 FCR 21 at 32 (Hill J). 233 The taxpayer bears the onus of establishing that it did not obtain a tax benefit in connection with an alleged scheme: Trail Bros at [35] (Dowsett and Gordon JJ). How the taxpayer establishes the non-existence of a relevant tax benefit is a matter for it: Trail Bros at [36] (Dowsett and Gordon JJ); Federal Commissioner of Taxation v Ashwick (Qld) No 127 Pty Ltd (2011) 192 FCR 325; [2011] FCAFC 49 (Ashwick) at [153] (Edmonds J, Bennett and Middleton JJ agreeing); RCI Pty Ltd v Federal Commissioner of Taxation (2011) 84 ATR 785; [2011] FCAFC 104 (RCI) at [133]-[136] (Edmonds, Gilmour and Logan JJ). 234 While a taxpayer may lead evidence regarding what it would have done in lieu of the scheme, it may not do so for any number of reasons, and any direct evidence will be useful insofar as it reveals "facts or matters that bear upon the objective determination of the alternative postulate": Trail Bros at [36] (Dowsett and Gordon JJ); see also Ashwick at [153(8) and (10)] (Edmonds J, Bennett and Middleton JJ agreeing). Expert evidence may also be relied upon to establish an alternative postulate (or counterfactual): eg Futuris Corporation Ltd v Federal Commissioner of Taxation (2010) ATC ¶20-206, referred to in Ashwick at [153(8)] and upheld on appeal: Federal Commissioner of Taxation v Futuris Corporation Ltd (2012) 205 FCR 274; [2012] FCAFC 32 at [68], [79]-[81] (Kenny, Stone and Logan JJ). 235 Whether or not a taxpayer obtains a relevant tax benefit is an "objective fact": Commissioner of Taxation v Peabody (1994) 181 CLR 359 (Peabody) at 382; Trail Bros at [23] (Dowsett and Gordon JJ); Ashwick at [153] (Edmonds J, Bennett and Middleton JJ agreeing). In the case of a deduction, its existence (or non-existence) is arrived at following "an objective inquiry as to what would have been allowed or might reasonably be expected to have been allowed as a deduction had the scheme not been entered into or carried out": Ashwick at [153(2)] (Edmonds J, Bennett and Middleton JJ agreeing) citing Epov v Federal Commissioner of Taxation (2007) 65 ATR 399; [2007] FCA 34 (Epov) at [62] (Edmonds J); Peabody at 385-6; Trail Bros at [24] (Dowsett and Gordon JJ). 236 The legislation requires a comparison between the relevant scheme and an alternative postulate, or counterfactual: Federal Commissioner of Taxation v AXA Asia Pacific Holdings Ltd (2010) 189 FCR 204; [2010] FCAFC 134 (AXA) at [128] (Edmonds and Gordon JJ) citing Hart at [66]; see also Trail Bros at [25] (Dowsett and Gordon JJ). 237 As Edmonds and Gordon JJ explained in AXA at [129] (emphasis in original): The alternative postulate requires a "prediction as to events which would have taken place if the relevant scheme had not been entered into or carried out and that prediction must be sufficiently reliable for it to be regarded as reasonable" (emphasis added). "A reasonable expectation requires more than a possibility": Lenzo [Federal Commissioner of Taxation v Lenzo (2008) 167 FCR 255] at [122] citing Peabody at 385. The question posed by s 177C(l) is answered on the assumption that the scheme had not been entered into or carried out: Lenzo at [121]. 238 In a deduction case, if it is determined that the alternative postulate, or counterfactual, would give rise to tax deductions, then the tax benefit is the differential between the amount claimed and the deductions arising from the counterfactual: Ashwick at [153(16)] (Edmonds J, Bennett and Middleton JJ agreeing) citing Trail Bros at [54] and [67] (Dowsett and Gordon JJ); see also Trail Bros at [30] and [47]-[49] (Dowsett and Gordon JJ). 239 While the entirety of a scheme must be assumed not to have been entered into when predicting the events that would or might reasonably be expected to have taken place, the alternative postulate or counterfactual may include elements of the scheme, provided it is not essentially the same set of steps as comprise the scheme: Ashwick at [153(3)-(4), (6)] (Edmonds J, Bennett and Middleton JJ agreeing), referring to AXA at [131]-[133] (Edmonds and Gordon JJ) and Trail Bros at [28]-[29] (Dowsett and Gordon JJ). 240 It can be seen that the essence of the tax benefit analysis mandated by s 177C is twofold. First, the court must identify a "sufficiently reliable" prediction of the events which would have, or might reasonably be expected to have, taken place in the absence of the scheme (the alternative postulate). Secondly, the tax consequences of the alternative postulate are to be compared with the tax consequences of the scheme in fact. In the case of deductions, the tax benefit will be the differential between the amount claimed and the deductions arising from the alternative postulate.

[37]

MAHPL's submission on the need to be able to anticipate a tax benefit as at the time of entry into the scheme 241 MAHPL submitted that there will not be a "tax benefit" for the purposes of s 177C(1) unless, at the time the scheme was entered into, its proponents could foresee that the course they were embarking on would be more advantageous from a tax perspective than an alternate course. MAHPL contended that the reference, in s 177C(1)(b) to whether a deduction (or part of a deduction) "might reasonably be expected" not to have been allowable means that the "expectation" must be assessed at the time of entry into the scheme. 242 This was not a submission advanced by MAHPL in opening, or in writing at any point. On the contrary, MAHPL's written opening summarised the enquiry required in order to determine the existence of a tax benefit as follows: It is an objective fact whether or not a taxpayer obtained a "tax benefit" in connection with a scheme to which Pt IVA applies [citing AXA at [126] (Edmonds and Gordon JJ); Trail Bros at [23] (Dowsett and Gordon JJ); Ashwick at [153] (Edmonds J)]. Paragraph 177C(1)(b) posits an objective inquiry as to what would have been allowed, or might reasonably be expected to have been allowed, as a deduction had the scheme not been entered into or carried out [citing Ashwick at [153] (Edmonds J), in turn citing Epov at [62], Peabody at 385-6 and Trail Bros at [24]]. It involves "the objective inquiry of predicting the particular activity or the events that would or might reasonably be expected to have taken place in the absence of the scheme" [citing Ashwick at [153] (Edmonds J); AXA at [131] (Edmonds and Gordon JJ) and RCI at [127] (Edmonds, Gilmour and Logan JJ)]. The subjective intentions of parties to the scheme play no part in determining this issue. 243 Applying MAHPL's new proposition to the present case in its oral closing, MAHPL submitted as follows (emphasis added): So we say, in order to get [a] tax benefit, you have to identify what's going to happen. And just assuming a slice of the term loans under the SCA for a moment, even without any swaps, interest rate swaps, you have to be able to say - have a reasonable expectation - and as the High Court said, that's not just a guess; it's got to be sufficiently reliable to form a reasonable expectation that the interest payable under the external loan would have been less than the interest payable under the internal loan, and you simply couldn't do that on 2 October [2007]. 244 In response to a request to identify authority supporting its construction, MAHPL relied on Federal Commissioner of Taxation v Lenzo (2008) 167 FCR 255; [2008] FCAFC 50 (Lenzo) and in particular [128], being a passage in the reasons of Sackville J. In that passage, Sackville J said as follows: By parity of reasoning, in determining whether the particular deduction claimed by the taxpayer would or might reasonably have been allowable, the Court must consider, in the absence of the scheme, what activity the taxpayer would have undertaken. The taxpayer can satisfy the onus of showing that he or she has not obtained a tax benefit in connection with a scheme if: • he or she would have undertaken or might reasonably be expected to have undertaken a particular activity in lieu of the scheme; and • the activity would or might reasonably be expected to have resulted in an allowable deduction of the same kind as the deduction claimed by the taxpayer in consequence of the scheme. 245 While, in focusing on the second bullet point, MAHPL acknowledged that it did not go expressly to the timing point, it nevertheless maintained that "you've got to have the expectation at some point in time, and the expectation, we say, has got to be when the scheme is entered into". 246 I do not accept MAHPL's submission on this point. In my view, the authorities are clear that the exercise required by s 177C(1)(b) does not require that, in order for there to be a tax benefit, the specific advantage gained through entry into the scheme - which is objectively determined at a later point in time - be anticipated and expected at the time of entry into the scheme. As the authorities referred to above make clear, in referring to reasonable expectation, s 177C(1) directs attention to the qualitative likelihood of the prediction put forward as a counterfactual. 247 That is not to say that an inability to predict, at the time of entry into a scheme, the financial consequences of an alternate course of action is irrelevant to the "tax benefit" enquiry. On the contrary, the fact that the consequences of an alternate course of action could not be assessed at the time may (depending on the circumstances) be relevant to whether or not that alternate course of action is one that the taxpayer was likely to have entered into. 248 In some circumstances, it may be the case that an inability to anticipate the consequences of a course of action put forward as a counterfactual is such that the counterfactual would not be a sufficiently reliable prediction of the events that would, or might reasonably be expected to, have taken place in the absence of the scheme. But that is not this case. MAHPL's submission on this point focused on no-one having a crystal ball on interest rates so as to anticipate the consequences of adopting fixed or floating rate borrowings. Given the expert evidence adduced by MAHPL on fixed-floating rate equivalence - which posits that, at any point in time, fixed and floating interest rates are at points that reflect the same future economic consequences so that there is no real arbitrage opportunity - it may be accepted that Mylan could not anticipate whether MAPL's final interest bill would be higher or lower on a fixed or floating rate basis. But this also suggests that Mylan would not have rejected entry into a floating rate borrowing (entry into floating rate borrowings was a feature of the Commissioner's counterfactuals).

[39]

The primary counterfactual 251 Under the primary counterfactual, Mylan would not have incurred any debt at the Australian level related to its indirect acquisition of Alphapharm. Alphapharm's earnings would have been taxed in Australia, with funds then remitted upstream to Mylan as dividends, which would have to have been declared by each intermediary entity up the chain, or possibly by upstream intercompany loans. 252 The Commissioner's primary counterfactual does not constitute a prediction of the events which might have taken place (had the primary scheme not been entered into), which is sufficiently reliable, such that it may be regarded as reasonable: Peabody at 385; Ashwick at [150]-[152] (Edmonds J, Bennett and Middleton JJ agreeing). That is so for two principal reasons. First, the primary counterfactual would inflexibly have tied up funds equivalent to the purchase price of Alphapharm as equity, when debt is significantly more flexible than equity and a mix of debt and equity is generally the preferred means of funding subsidiaries. Secondly, Mylan's OFL position in the US was such that it would have been unable to claim any foreign tax credits for income taxes paid in Australia, exposing it to an effective worldwide tax rate of 65% on Australian-generated earnings. I expand on these two matters below, before turning to the other counterfactuals. 253 Having regard to these matters, I do not consider that the primary counterfactual is reasonable as I do not consider that, had it not entered into the primary scheme, Mylan would, or might reasonably be expected to, have made the acquisition of Alphapharm with 100% equity funding at the Australian level. Rather, I consider it much more likely that the acquisition of Alphapharm would have been funded with a mix of debt and equity. Rejection of the primary counterfactual in favour of a mixed debt and equity scenario does not require any conclusion to be reached about what the debt to equity ratio would be, whether the loan would be from an external or internal lender, whether it would be fixed or floating, and related questions. Those matters are all considered in relation to the other counterfactuals.

[40]

Debt and equity 254 Mr Stack described making an acquisition with a mix of debt and equity as a common and efficient practice, particularly where the acquisition is large relative to the size of the acquirer (as was this transaction). Mr Stack explained the ways in which equity ties up capital to a much greater extent than intercompany loans, contrary to the interests of the parent's shareholders, and contrary to the shareholders' general preference for the parent company to maintain access to liquidity to invest in growth. 255 As Mr Stack explained, there are a number of steps involved in, and regulatory restrictions that may apply to, a subsidiary remitting funds to its ultimate parent by way of dividends. Those constraints include the payment of dividends from profit (which may not correlate with free cash flow), dividends often only being allowed to be paid once a year (and often only after audited statutory accounts have been prepared), and the need to repeat that process in each intermediary holding company, by reference to the profit of that company. Mr Johnson agreed with Mr Stack regarding the difficulties associated with relying on dividends to remit excess cash to a parent entity and also noted that often board involvement is required in order for dividends to be declared. Mr Johnson's view, like that of Mr Stack, was that debt is more flexible than equity. 256 The Commissioner identified five matters in submitting that the primary counterfactual was reasonable. First, he pointed to the structure of the original SPA and the fact that some other operating subsidiaries within the Merck Generics group were acquired with equity funding, and some planning documents having other substantial operating subsidiaries (in Canada and France) equity-funded. Secondly, he observed that, in a 100% equity funding scenario, dividends could still have been "parked" in Mylan Bermuda until required to be remitted upstream to Mylan in the US, allowing US income tax on those funds to be deferred until remitted. Thirdly, the Commissioner noted that the 100% equity funding scenario would not have split up the external borrowing between the US and Australia (which is a point of distinction between the primary counterfactual and the counterfactuals that posited external borrowing). Fourthly, the Commissioner submitted that the 100% equity scenario is straightforward and simple. Fifthly, the Commissioner stated that the 100% equity scenario would also have been attractive on the basis that there would be no risk that it would be set aside under Pt IVA. 257 I will address the last two points briefly, before turning to the points of greater substance. I accept that the 100% equity counterfactual is straightforward and simple. However, I do not accept that it would have commended itself to Mylan for that reason. Mylan was making a USD 7 billion acquisition of an international generics group; some complexity was to be expected. In my view, it is inconceivable that Mylan would have been willing to accept the significant downsides of the 100% equity scenario for the acquisition of Alphapharm for the sake of simplicity. As to avoiding any risk of Pt IVA applying, of course not claiming any acquisition-related tax deductions in a particular jurisdiction is likely to result in a position where the acquisition funding escapes adverse scrutiny by the local taxing authorities. However, I do not accept that Mylan was so timid that it would adopt an otherwise unsuitable funding structure for the Acquisition, so far as it concerned Alphapharm, simply to have the comfort of knowing the ATO would not scrutinise it. 258 The first point raised by the Commissioner requires some explanation. The original SPA provided for Mylan, as purchaser, to acquire five target entities, which included MGGBV. MGGBV held most (if not all) of the non-US based operating entities. One of its subsidiaries was Alphapharm. While Mylan was, as the Commissioner submitted, bound to proceed with the original SPA, the evidence is overwhelming that there was no intention for the final acquisition structure to be simply constituted by Mylan acquiring the five target entities. Clause 3.1.5 of the original SPA expressly provided for Mylan to elect to have any one or more of its affiliates purchase interests in any "Subsidiary", either in lieu of, or in addition to, the acquisition of the shares in the five target companies. The legal entities held by the five target companies were the "Subsidiaries" (cl 2.2). Clause 3.1.5 of the original SPA also provided for a full indemnification of the Sellers and their affiliates to be agreed, should changes to the Acquisition structure increase tax costs on the vendor side. 259 The structure provided for by the original SPA is readily explained by the fact that Mylan's advisors recognised that there would be "no time to come up with a fully agreed upon acquisition structure" by the time the original SPA was signed on 12 May 2007. It was a fast-paced transaction; the initial notification that Merck was exploring the potential sale of its generics business having been issued only two months earlier, on 6 March 2007. The original SPA was signed less than two weeks after Mylan submitted its updated non-binding proposal, on 30 April 2007. 260 In my view, it was always on the cards that the Acquisition structure would be settled after the original SPA was signed. Indeed, three iterations of the Deloitte presentation, being those dated 27 April 2007, 1 May 2007 and 11 May 2007, prepared prior to the original SPA's execution, canvassed the acquisition of Merck subsidiaries, including Alphapharm, directly by local holding companies (cf the indirect acquisition posited by the primary counterfactual). Accordingly, the fact that the original SPA provided, absent any changes, for Mylan to acquire MGGBV (which would have come into the Mylan group holding Alphapharm) does not provide any material support to the reliability of the primary counterfactual. 261 The Commissioner also pointed to the fact that: (a) Deloitte's June 2007 "Simple Alternative" structure slide deck presented a scenario under which all of the operating subsidiaries, other than Alphapharm, would be indirectly acquired without debt, such that subsidiaries that included Merck Canada and Merck France (both of which had roughly the same expected operating profit as Alphapharm) would have been indirectly acquired (ie acquired without debt at the local subsidiary level); (b) ultimately, on 2 October 2007, subsidiaries other than Alphapharm, Merck Canada and Merck France, were indirectly acquired (ie without a local holding company financing the acquisition of the subsidiary partly with debt); and (c) Deloitte's "Project Genius - Tax Overview" slide deck dated 1 May 2007 (ie before the original SPA was signed on 12 May 2007) proposed that Alphapharm, Merck Canada and Merck Japan would be acquired by local holding companies taking on external debt, but, for reasons that are not explained, Merck Japan was not ultimately funded with any debt, but was among the group of indirectly acquired subsidiaries. 262 The thrust of MAHPL's response on these points was that there must have been commercial reasons why Mylan ultimately elected to acquire the Australian, Canadian and French operating subsidiaries by partly debt-funded local holding companies, and to acquire the other subsidiaries indirectly (without debt funded local holding companies). MAHPL stressed the economic contribution that the Australian, Canadian and French subsidiaries were expected to make to the high debt servicing burden taken on by Mylan (and Lux 5), and the anticipated need for a 100% "cash sweep" to be undertaken to fund the obligations incurred under the external loans. In closing, the Commissioner hotly contested the 100% cash sweep point. I deal with the cash sweep point separately below (see paragraphs 267-288 below), as it is also relevant to the OFL and foreign tax credit issue. 263 Most of the structuring slide decks do not detail why certain steps were being proposed. Nor, to generalise, did the documentary record reveal the reasons why what Mylan's advisers were proposing changed from one slide deck to the next. At some points, emails explain some decisions that were taken, but such explanations were generally the exception rather than the rule. In short, the evidence does not disclose why Mylan did not proceed with the Deloitte June 2007 "Simple Alternative", or why it was that Merck Japan was moved from one group of subsidiaries (those to be partially debt-funded) to another (those being acquired without local level debt, by indirect acquisition). Nevertheless, the evidence does not support an inference that, because, in June 2007 - which was still relatively early in the piece - Deloitte modelled a scenario by which other sizeable subsidiaries (Merck Canada and Merck France) would have been acquired without debt at the local level, Mylan would likewise have been prepared to proceed with the primary counterfactual (cf the Commissioner's submission). 264 The Commissioner's observations cannot simply be swept away on the basis that "there must have been commercial reasons"; there is no evidence of the reasons, and there was no exploration of the local tax regimes in the multiple jurisdictions involved, which may otherwise allow inferences to be drawn about the structures proposed and ultimately adopted. Nevertheless, in my view, the significant lack of flexibility entailed in tying up such extensive funds as equity in Australia, together with the OFL issue (addressed below) mean that the: (a) fact that other structures were contemplated which would have seen Merck Canada and Merck France acquired with equity; (b) decision not to have Merck Japan acquired with debt at a local level; and (c) ultimate acquisition of various other subsidiaries indirectly and without local holding company debt, do not make the primary counterfactual a reasonable prediction of what is likely to have occurred in the absence of the scheme. 265 As to the Commissioner's observation that the primary counterfactual would still allow dividends to be "parked" outside the US, the proposition is correct as a matter of the operation of US tax law (according to the expert evidence). However, it is only a point that goes anywhere if there was no expectation, as at October 2007, that funds from Alphapharm would not need to be repatriated to the US. In other words, if it was expected that funds would need to be repatriated to the US to meet the external debt service obligations (and targeted deleveraging), the ability to "park" funds outside the US would not be a significant factor motivating Mylan to proceed according to the primary counterfactual. For the reasons set out below, I accept that, as at October 2007, it was expected that funds would need to be repatriated to service and reduce the very substantial external debt assumed under the SCA. 266 As to the Commissioner's third point - that the primary counterfactual avoids splitting the overall USD 7 billion borrowing into two (with USD 1 billion being borrowed by an Australian holding company) - I accept that a reduction in the portion of the overall loan in respect of which Mylan could claim tax deductions for interest in the US would reduce its tax deductions and would do so to an extent that exceeds the percentage value of the deduction in Australia (as the overall US tax rate is 35-40% depending on the applicable state and local taxes, as against 30% in Australia). However, this also is but one factor and not one that persuades me that, in the absence of the primary scheme, Mylan would have preferred a 100% equity funding of the acquisition of Alphapharm, as distinct from a funding that combined debt and equity. Again, the combined disincentives of tying up that much equity, and the OFL issue, suggest strongly that Mylan would have preferred a combination of debt and equity.

[41]

Repatriation of free cash flow 267 I turn, then, to the expectations that were held in October 2007 concerning repatriation of funds to meet Mylan's external debt obligations and reduce its leverage. MAHPL stressed, at numerous points in the argument, that Mylan intended to repatriate 100% of free cash flow to service external debt. That contention was advanced relying heavily on documents recording statements made by Mylan to its lenders, the ratings agencies and the market. As set out further below, Mylan made repeated reference to using anticipated strong free cash flow to reduce its leverage. 268 The Commissioner did not accept that Mylan had the stated intention to utilise free cash flow from its overseas subsidiaries. The Commissioner relied heavily on the fact that, in the year to 31 December 2008, Mylan's "Form 1118" shows that only USD 1 million was repatriated from Bermuda, and USD 373 million was included as a deemed dividend from Mylan Luxembourg 3 S.a.r.l. (a Luxembourg entity) (Lux 3). In the absence of an explanation of why such limited funds were remitted from Bermuda in 2008 (despite the debt obligations owed by MAPL, Mylan Canada and Mylan France all being owed to subsidiaries of Mylan Bermuda), the Commissioner submitted that "[t]he most reasonable explanation is that Mylan Inc. did not have an intention to actually repatriate 100% of its free cash flow from its foreign subsidiaries to the US to repay Mylan Inc.'s external debt". 269 The Commissioner urged that what Mylan in fact did (in the years between 2008 and 2012) should be used to construe what it intended to communicate by earlier documents (including documents dating back to May 2007). In my view, documents need to be construed by reference to their context, constituted by matters in the period leading up to when they were written. When an entity's subsequent actions are inconsistent with earlier documents, those subsequent events may logically be used to question why there was the divergence observed, but it makes no sense to try to use later events to construe documents created years earlier. 270 Nevertheless, the Commissioner sought to use later events to suggest that earlier documents should be construed as not revealing an intention to repatriate free cash flow to the US. The Commissioner's contention that there was no intention to repatriate free cash flow is not consistent with the basis upon which the experts, including Mr Johnson, proceeded. In their joint report, Messrs Johnson, Stack and Ali recorded that Mr Johnson was of the view that "Mylan subsidiaries were expected to distribute available cash to Mylan and that such cash distributions were essential to Mylan meeting its debt service obligations". Mr Stack was of the view that "Mylan was clear that it had to sweep cash from subsidiaries … to service debt". Mr Ali's first report set out his opinion that "Mylan would have expected that it might need to repatriate all (or substantially all) of the earnings from its acquired offshore subsidiaries, including Alphapharm, in order to meet its anticipated cash flow requirements consistent with its representations to the market". 271 In considering the significance of statements made in contemporaneous documents, and what was in fact done between 2008 and 2012, it should be recalled that, under the terms of the SCA, Lux 5 was also a borrower. As such, while some cash could be swept from group entities for use by Lux 5 in debt servicing, without that cash coming in to the US as actual or deemed dividends (a point that Mr Stack explained in one of his responses in the Group B joint experts' report), the repayment schedule for the US Tranche A Term Loan under the SCA was far more aggressive than the repayment schedule for Lux 5's Euro Term Loans component of the borrowing. It follows from this that the terms of the SCA support the need to repatriate funds to the US (cf having them sent to Lux 5). In addition, it should be noted that the structure of the Mylan group was such that all (or virtually all) of the operating subsidiaries outside the US sat beneath Mylan Bermuda, whereas Lux 5 was situated in another branch of the corporate structure. Accordingly, in order for dividends from Alphapharm to be used to meet interest obligations under the SCA and pay down external debt, those dividends would need to have gone up the corporate chain into the US (cf being used by Lux 5, which did not sit in the dividend chain between Alphapharm and Mylan). 272 In addition, the terms of the SCA provide support to the veracity of Mylan's statements about its intentions to reduce leverage. Under the SCA (pursuant to a clause requiring mandatory additional repayments over and above the scheduled payments based on free cash flow), the percentage of free cash flow that was required to be used to pay down debt operated so that the percentage of free cash flow that had to be devoted to these mandatory additional debt repayments reduced as the consolidated leverage ratio (consolidated total debt to EBITDA) reduced. The applicable interest rates on at least some components of the debt were also set to reduce as the consolidated leverage ratio reduced. 273 In its presentation to Moody's in mid-May 2007, Mylan presented on the Merck acquisition. Its presentation made reference to an intention to "reduce leverage" through the issuance of common stock and mandatory convertible securities shortly after the close of the transaction. It also referred to an anticipation that "[s]trong free cash flow expected to further de-lever balance sheet". The presentation then identified a near-term leverage target of "less than 6x Net Debt/EBITDA" and a "[l]ong-term leverage target of less than 4x Net Debt/EBITDA". 274 Mylan's June 2007 presentation to SMA Lenders sang the praises of the Acquisition and the profile of the combined entity. The presentation recorded that Mylan intended to reduce leverage through the issue of at least USD 1.5 to 2 billion of common stock and convertible securities in the near term, expected that strong free cash flow would "further de-lever" the balance sheet and that the near-term leverage target was less than 6.0x Net Debt/EBITDA. The presentation included slides recording "Modelling Assumptions", one of which was (under the heading "Pro Forma"): "Model assumes 100% cash flow sweep with the exception of Matrix cash flow which is assumed to remain at Matrix subsidiaries". The modelling assumptions followed slides that presented, inter alia, "Pro Forma Capitalization and Credit Statistics". The figures presented a pro forma view of the combined entity. While this view presented "PF Permanent", it is clear that the figures there presented merely recorded the anticipated position following issue of the equity Mylan anticipated issuing after closing (and so without the bridging finance). It was not, therefore, presenting an anticipated "permanent" debt profile. Other slides went on to present pro forma financials (in bar graphs) from 2008-2014, including a projection of bank and total leverage reducing to a total debt/EBITDA ratio of 0.6x in 2014. 275 I do not accept the Commissioner's contention that the statement in "Modelling Assumptions" regarding the 100% cash flow sweep was confined to 2007. First, many other assumptions set out in the reported modelling assumptions were not, on their face, confined to 2007. Secondly, the pro forma financials, to which I have referred, only presented the post-equity raising position of Mylan; they did not purport to set out figures regarding debt paydown by the cash flow sweep in 2007. Thirdly, other parts of the presentation referred to longer term intentions to reduce leverage by using anticipated strong cash flows. Fourthly, it defies logic that, in making a presentation to its intended lenders in relation to a seven year commitment, Mylan would only be presenting on its use of cash flow for a few months in 2007. For these reasons, in my view, the statement regarding the "100% cash flow sweep" did constitute a general statement of intention on Mylan's part. I reject the Commissioner's contentions to the contrary. 276 Merrill Lynch prepared a document "Pro Forma Model Case 2" dated 26 September 2007. This model presented a debt schedule which estimated the reduction of total debt between 2007 (pro forma) and 2008 to 2014 (estimates), with an estimate of USD 850 million in debt at the close of 2014. That slide also modelled the free cash flow available to pay down debt over the same period. The model also differentiated between the funds to be deployed to mandatory debt paydown, and funds available for optional payments. The model detailed that (other than for the 10 year senior notes) "yes" was recorded against the note "Sweep?" for all components of the optional debt paydown. 277 Presentations were also made to Moody's and S&P on 27 and 28 September 2007 respectively. These presentations repeated the "100% cash flow sweep" statement. For the same reasons as set out in respect of the June 2007 presentation to SMA Lenders, that statement was not confined to 2007. 278 While there was limited evidence of internal communications within Mylan, an email of 20 August 2007 from Paul Martin of Mylan to personnel at Deloitte advised that: "As regards 956 [deemed dividends] exposure up to that point, we currently are modelling full repatriation in the first couple [of] years anyway as a means of servicing the debt". 279 In his first report, Mr Ali also set out a number of other presentations made by Mylan (including its presentation to investors in October 2007) in which the "100% cash flow sweep" statement was made. Mr Ali also opined that the financial projections disclosed in the presentations were all based on the assumption of all, or substantially all, available free cash flow being used to repay debt. Mr Ali was not cross-examined on this evidence. 280 As noted above, in 2008, Mylan's Form 1118 showed that only USD 1 million was repatriated from Bermuda, and USD 373 million was included as a deemed dividend from Lux 3. The Commissioner contended that the fact that substantial amounts of free cash flow were not remitted (via Bermuda) in 2008 shows that Mylan did not, in 2007, intend to engage in a 100% cash flow sweep. I do not accept that submission. First, as set out above, the statements made by Mylan to its lenders, the ratings agencies, and the investors all clearly conveyed that intention. It would be a remarkable thing for Mylan to have broadcast that intention far and wide, and to such critical audiences, had it harboured intentions to the contrary. 281 Secondly, while the Commissioner focused on Form 1118 of Mylan's 2008 tax return - its other tax returns were not in evidence - what occurred in relation to dividends in that particular year must be approached with some caution as there was evidence of Mylan having received a USD 370 million cash payment in early 2008, arising from the sale of rights to a hypertension product. Given the obligations in the SCA to use the proceeds of sale to pay down debt, this sale may have resulted in a reduced need for repatriated cash flow to meet debt servicing obligations. 282 Further, and as highlighted by MAHPL in its closing submissions, the US business performed substantially in excess of expectations. The picture concerning the source of the USD 373 million in deemed dividends in the 2008 tax return is also muddied by reference to the source of the deemed dividend being explained (in MAHPL's submission) by reference to PwC's "80/20" memo of 30 September 2007. According to MAHPL's submission, the USD 373 million deemed dividend recorded by Mylan in its 2008 Form 1118 was not unexpected, nor did it reflect funds sourced economically from "outside the Mylan system". Rather, on MAHPL's submission, those funds were "a wash", as there was a planned "circle" of funds involving interest payments from MIHI going to Lux 3, and payments going the other way (as set out in PwC's "80/20" memo of 30 September 2007), albeit by reference to a different amount. MAHPL submitted that, according to that memo, it was intended that Lux 3 would wind up as the holder of two promissory notes to be issued by MIHI and that these aspects of the 80/20 memo were reflected in the PwC step plan (in particular, steps 17A to 22 of version 17 of the step plan). According to the PwC step plan, the interest income distributed up would be classified as a deemed dividend in the US, but would not be taxable at the US state level. 283 PwC's 80/20 memo referred to MIHI funding its interest payment obligations by means of dividends received from Bermuda, and loans extended by another Mylan group company. MAHPL tendered tax documents of MIHI, which showed an interest expense roughly corresponding with the USD 373 million. What remains unclear, however, is the source of funds used by MIHI in 2008. Its 2008 tax return refers to only USD 1 million in "[g]ross foreign dividends not previously taxed". 284 While the evidence did not fully explain the exact way in which cash would be repatriated to fund the interest obligations under the SCA, and the desired deleveraging profile, it is tolerably clear that the receipt of deemed dividends, relating to interest on intra-group loans, was part of the plan and not obviously inconsistent with the repatriation representations made by Mylan. In addition, given the matters referred to above concerning events in 2008, I do not conclude that Mylan did not, in structuring the transaction, intend to repatriate free cash flow simply because Mylan's 2008 tax return (which was its only tax return in evidence) did not show significant funds repatriated from Bermuda. 285 Thirdly, I also do not accept the Commissioner's submission that Mylan's actual debt profile between 2008 to 2012 reveals its "real" intentions in October 2007 as to debt repayment. That submission entails an implicit suggestion that the presentations made in 2007 falsely presented Mylan's intentions. While it is true that, between 2008 and 2012, Mylan did not reduce its debt in accordance with its presentations in 2007 and that regard may be had to what actually occurred in considering what Mylan in fact intended in October 2007, I do not accept that Mylan's debt profile (from 31 December 2008 to 2012) suggests that the statements of intention in 2007 were false. Subsequent events also do not support the retrospective construction that the Commissioner sought to impose on the May 2007 statement regarding the intended debt/EBITDA ratio being "less than" 4x, by which he contended Mylan in fact was conveying that its intended long-term ratio was a plateau of 4x. The various arguments mounted by the Commissioner do not, as he contended, undermine Mr Ali's analysis (even putting to one side that these matters were not put to him), and sit uncomfortably with the "Pro Forma Projected Financial Metrics" contained within the same presentation, which, as noted above, projected a decline in total debt/EBITDA from 2008 to 2014. 286 The terms of the SCA, and the terms sheets that preceded execution of the formal agreement, support a conclusion that Mylan did in fact intend to remit substantial free cash flow to service its debt and reduce its leverage. Those terms required that at least 50% of annual "excess cash flow" (provided Mylan's consolidated leverage ratio was greater than 4.5), and 100% of net cash proceeds of any asset sales, be applied to pre-payments. 287 A further point supporting Mylan's anticipated need for cash is that it announced that it was suspending the payment of dividends on its common stock. 288 In my view, as at October 2007, Mylan did intend that group-wide free cash flow would be deployed to service external debt and to reduce group-wide leverage.

[42]

Mylan's OFL position and foreign tax credits 289 The Commissioner submitted that, had Alphapharm been acquired entirely with equity (and no debt at the Australian level), funds might reasonably be expected to have been remitted up the corporate chain from Alphapharm by way of dividends (or possibly by loans from Alphapharm or Lux 2). It was not suggested that Alphapharm making loans to upstream group companies would have addressed the OFL issue that was explored by reference to dividends being issued by Alphapharm. 290 It was common ground that Mylan's post-acquisition OFL position was such that it would have been unable to claim any foreign tax credits on income tax paid in Australia. The effect of an inability to claim foreign tax credits under US tax law was that the effective worldwide tax rate of Mylan (on Australian-generated income) would have increased markedly, from 35% (being the US federal corporate tax rate) to 65% (with the inclusion of the 30% Australian corporate tax rate). In my view, the extent of the increase - or, in other words, the effect of the OFL meaning that Mylan would have been unable to claim any foreign tax credits under US law in respect of taxes paid in Australia - was so extreme as to be intolerable. I find that, had the scheme not taken place, Mylan would not, or might not reasonably be expected to, have proceeded in the way that the counterfactual posits. 291 In his first report, Mr Glenn set out the following table illustrating the impact of the OFL position on Mylan's worldwide effective tax rate: Alternative 1 Alternative 2 Alternative 3 (100% equity funded) (Partially equity and partially debt from U.S. CFC - Int Exp is 50% of EBIT (Partially equity and partially debt from U.S. CFC - Int Exp is 75% of EBIT Alphapharm / MAPL Earnings of Alphapharm/MAPL before interest and taxes 100 100 100 Interest expense (if applicable) - (50) (75) Earnings of Alphapharm/MAPL before taxes 100 50 25 Australian income taxes (30) (15) (8) After tax earnings of Alphapharm/MAPL 70 35 18 U.S. CFC/Mylan Bermuda Dividend income 70 35 18 Interest income - 50 75 Income taxes/Australian Withholding Tax - (5) (8) After tax earnings of U.S. CFC 70 80 85 Mylan Dividend income 70 80 85 Section 78 Gross Up 30 20 15 Interest Income - - - Australian Withholding Tax on Interest Income - - - Foreign Taxes Deemed Paid under Section 901/902 30 20 15 Total U.S. Taxable Income 100 100 100 U.S. Tax Liability before tax credits (35) (35) (35) With No OFL Permitted Foreign Tax Credits 30 20 15 Residual U.S. Tax Benefit (Liability) (5) (15) (20) Total U.S. and Foreign Taxes Paid (35) (35) (35) After Tax Earnings of Mylan 65 65 65 With OFL Permitted Foreign Tax Credits - - - Residual U.S. Tax Liability (35) (35) (35) Total U.S. and Foreign Taxes Paid (65) (55) (50) After Tax Earnings of Mylan 35 45 50

[43]

292 As that table illustrates, under a 100% equity funded scenario with an OFL, assuming a notional $100 of income earned by Alphapharm, resulting in a dividend of $70 up the corporate chain, Mylan would be liable to pay US tax at 35% on the grossed-up amount of the dividend ($100) but Mylan would have been unable to claim foreign tax credits for the $30 of income tax paid in Australia, resulting in an overall effective tax rate of 65% on that $100 earned. 293 Mr Glenn's table also illustrates that the impact (on Mylan's after tax earnings) of the inability to claim foreign tax credits in the US decreases the more debt (relative to equity) is assumed. With Alphapharm paying interest on debt, its taxable income would naturally reduce as a consequence, meaning that less tax was payable on Alphapharm's earnings that may be remitted up the corporate chain. 294 The Australian and foreign tax considerations of a proposed course are not irrelevant considerations in the analysis of whether a taxpayer might reasonably be expected to adopt that particular course in the absence of a scheme. As Gleeson CJ and McHugh J observed in Hart at [15]: "[t]axation is part of the cost of doing business, and business transactions are normally influenced by cost considerations". To like effect are the observations of Harlan J of the United States Supreme Court in Commissioner of Internal Revenue v Brown (1965) 380 US 536 at 579-80, quoted by the majority judgment in Spotless Services at 416, that: the tax laws exist as an economic reality in the businessman's world, much like the existence of a competitor. Businessmen plan their affairs around both, and a tax dollar is just as real as one derived from any other source. 295 So it was, for example, that in RCI, the Full Court (Edmonds, Gilmour and Logan JJ) observed at [136] that "as Peabody itself establishes, the absence of any tax benefit obtained in connection with the scheme might be established by demonstrating the illogicality of the taxation consequences upon which the Commissioner's counterfactual is predicated". 296 Here, the US tax experts were agreed that what was actually done was a far superior outcome for Mylan when compared with all of the three counterfactuals proposed by the Commissioner; the US tax experts were not asked to consider the two counterfactuals raised by MAHPL. Nevertheless, in my view, given Mylan's OFL position, the effects of 100% equity funding would have been unacceptable to Mylan, and the primary counterfactual is not a reasonably reliable prediction of what would, or might reasonably be expected to, have occurred, had the primary scheme not been entered into.

[44]

Conclusion on the primary counterfactual 297 I am mindful that it is open to the court to consider counterfactuals that depart from the precise bounds of the counterfactuals put up by the parties (subject to procedural fairness being afforded to the parties to address any further counterfactual). In Peabody, the High Court explained (at 382) that the Commissioner's discretion to cancel a tax benefit under s 177F(1) depends on objective facts (cf the formation of an opinion or state of satisfaction by the Commissioner). The question in every case, as their Honours put it, is "whether a tax benefit which the Commissioner has purported to cancel is in fact a tax benefit obtained in connexion with a Pt IVA scheme and so susceptible to cancellation at the discretion of the Commissioner": Peabody at 382. Consequently, in Peabody, the Court found that the Commissioner was entitled to proceed at trial (and on appeal) by reference also to a scheme that was narrower than the 10 steps he previously identified. Such a departure from the originally described scheme was permitted "provided it causes no undue embarrassment or surprise to the other side": Peabody at 382. 298 Similarly, in discussing the nature of the taxpayer's onus under s 14ZZO(b)(i) of the TAA, insofar as it concerns the "tax benefit" enquiry, the Full Court in RCI explained relevantly (at [130]-[131]): Even if a taxpayer establishes that the Commissioner's counterfactual is unreasonable, it will not necessarily follow that he has established that the assessment is excessive. That is because the issue is not whether the Commissioner puts forward a reasonable counterfactual or not; it is a question of the court determining objectively, and on all of the evidence, including inferences open on the evidence, as well as the apparent logic of events, what would have or might reasonably be expected to have occurred if the scheme had not been entered into. Thus, even if a taxpayer establishes that the Commissioner's counterfactual is unreasonable, that will not discharge the onus the taxpayer carries if the court determines that the taxpayer would have or might reasonably be expected to have done something which gave rise to the same tax benefit. That such an articulation of the onus is at worst erroneous and at best unhelpful, can also be illustrated from the other side of the coin, because it implies that if the Commissioner's counterfactual is reasonable that is the end of the matter; even if the court were to conclude, on all the evidence, inferences and logic referred to, that if the scheme had not been entered into the taxpayer would have or might reasonably be expected to have done something which did not give rise to a tax benefit, or which gave rise to a tax benefit less than that thrown up by the Commissioner's counterfactual. In our view, that cannot be correct. (Emphasis added.) 299 In other words, consistent with the passages of Peabody cited above, the court is not bound by the counterfactuals put forward by the parties. The Commissioner accepted as much in oral submissions and MAHPL did not contend otherwise. 300 While the primary counterfactual has a number of features, its primary distinguishing characteristic is the proposition that the acquisition of Alphapharm would be 100% equity funded. As I do not consider that a 100% equity funded acquisition is a sufficiently reliable prediction of what would, or might reasonably be expected to, have occurred, there would be no utility in considering whether a variation of the primary counterfactual might be devised. Rather, that entire counterfactual is driven from the 100% equity funding assumption. It is that assumption that stands behind the proposition that no debt funding would have occurred in Australia at all, and hence there would have been no interest deductions in Australia. 301 MAHPL has discharged its onus of showing that it did not receive a tax benefit calculated by reference to the primary counterfactual.

[45]

Other counterfactuals 302 I turn, then, to what would most likely have occurred had none of the schemes been entered into. Clearly, Alphapharm would still have been acquired by the Mylan group; not acquiring it is not a realistic scenario. 303 Given my rejection of a counterfactual based on 100% equity, the acquisition of Alphapharm would have involved at least some interest-bearing debt, which would have been incurred by MAPL (cf the primary counterfactual in which no Australian holding companies would have been formed). The key issues that arise for determination are: (a) whether the lender would have been an external lender, or Mylan (or another US subsidiary); (b) how much debt would have been incurred, relative to equity (or non-interest bearing debt, as allowed for by the Commissioner's secondary counterfactual and tertiary counterfactual); (c) whether the interest on that debt would have been at a fixed or floating rate, or a mixture of the two, and what interest rates would have applied; (d) whether the borrowing would have allowed for interest to be capitalised; and (e) whether the borrowing would have required the borrower to make payments amortising the principal consistently with the repayments of principal actually made under PN A2 (cf allowing the borrower not to repay any of the principal over the term of the loan, while permitting it to make repayments at its election). 304 Those points are to be addressed, to determine the most reliable counterfactual (assuming there is a sufficiently reliable counterfactual). Those matters fall to be addressed in the context of some points that are not controversial. 305 First, no acquisition structure would have been adopted that would have seen greater debt assumed in Australia than the then-applicable thin capitalisation rules would have allowed. It should be noted that none of the four remaining counterfactuals posited debt in excess of those thin capitalisation limits. In 2007, Australia's thin capitalisation rules operated to limit the deductibility of interest where, broadly speaking, the taxpayer had a debt to equity ratio of more than 3:1. 306 Secondly, the Australian entity would have borrowed in AUD, or would have taken out cross-currency swaps so that its indebtedness was effectively in AUD. An ultimate economic borrowing in AUD was a feature common to all four remaining counterfactuals and I accept that such a borrowing would reasonably be expected so that an operating subsidiary did not take on foreign exchange risk. In this regard, the three relevant experts (Mr Stack, Mr Johnson and Mr Ali) agreed that an international company such as Mylan would commonly manage its currency risk in a centralised manner at the group or treasury level (and not at the level of operating subsidiaries). In Mr Stack's view, centralisation would better support the management of counterparty risk, and facilitate proper accounting of interest rate and currency volatility and hedge instruments. Mr Ali gave evidence that he would particularly expect foreign exchange risk to be centralised where subsidiaries did not have the treasury functionality, or the capability or wherewithal to manage currency risks themselves, as the group would be able to manage all of its currency risk more efficiently and minimise any duplication of treasury functionality. In Mr Stack's view, it was reasonable to expect that Mylan would not have treasury staff in subsidiaries such as the consolidated MAHPL group (and MAHPL did not in fact have its own treasury function). 307 Thirdly, if MAPL were to borrow externally, its borrowing would be supported by a guarantee from Mylan, such that MAPL could borrow at an interest rate reflecting Mylan's credit rating. This point was assumed by the relevant experts in giving their evidence as to the quantum of the debt. 308 I turn, now, to features of the potential counterfactuals that were contested by the parties.

[46]

Debt to equity ratio 309 The debt to equity ratio posited by the Commissioner's secondary and tertiary counterfactuals was 54.6% debt and 45.4% equity. According to the Commissioner's appeal statement, this was the proportion of debt and equity advised by MAHPL during the audit as the funding mix of the Mylan group in making the 2007 acquisition. MAHPL accepted that the figure came from it, but said it was wrong as the figures provided reflected the Mylan group's overall gearing as at 31 December 2007 (cf the gearing of the Acquisition, or the gearing immediately following the Acquisition). MAHPL also contended that the counterfactual borrowing of MAPL should not be set having regard, with the benefit of hindsight, to the gearing of Mylan after the 2007 capital raising as that capital raising turned out to be more successful than anticipated (meaning that Mylan was able to pay out its entire USD 2.85 billion bridging loan without having to issue USD 850 million in senior notes (ie more debt), as had originally been anticipated). 310 It is convenient to start by setting out some key debt to equity ratios, established on the evidence (noting further ratios are presented in the table at paragraph 148 above): (a) in its June 2007 presentations to the SMA Lenders and the Bridge Lenders, Mylan projected that its pro forma full year 2007 capitalisation (prior to equity raising) would comprise 83% debt; (b) Merrill Lynch's Pro Forma Model for Case 2 (dated 26 September 2007) assumed that USD 2 billion in equity would be raised and projected a total debt to equity ratio of 73.1% (pro forma 2007); (c) the capitalisation figures presented to Moody's on 27 September 2007 (assuming USD 2 billion in equity would be raised after closing) equated to approximately 74% debt; and (d) as at the end of December 2007, Mylan's overall capitalisation was 54.6% debt and 45.4% equity but, as calculated by Mr Stack, its acquisition funding mix remained at approximately 3:1 debt to equity (ie 75% debt). 311 Mr Ali also calculated that the midpoint of the range of Mylan's anticipated aggregate acquisition funding mix for the Acquisition as a whole was 74.8% debt, which was consistent with the funding mix for Alphapharm (75% debt). Mr Ali's opinion was that the shareholder risk appetite for a subsidiary like MAPL would be determined at the group level. He explained that, while the funding mix chosen by the parent for different subsidiaries in different jurisdictions would take into account local factors (as well as the projected cash flows of the subsidiary), because Australia is a relatively stable jurisdiction and a mature market, he would expect a high tolerance for debt in the acquisition of Alphapharm, consistent with or higher than the gearing level for the overall acquisition. In cross-examination, Mr Ali accepted that, although he had not considered these jurisdictions with the same rigour as he had with Australia, Canada and France (as well as Japan and the UK) were also jurisdictions which shared those characteristics with Australia, and so an acquirer may, all else being equal, have a risk tolerance in those jurisdictions that supported a gearing level that was consistent with the gearing level for the overall transaction. 312 In my view, if the schemes had not been entered into, MAPL would still have been capitalised with 75% debt. It is clear from the evidence that, when Mylan considered having local acquisition entities take on debt to acquire relevant Merck subsidiaries, the debt levels it projected tracked the applicable thin capitalisation limits in the various jurisdictions. As noted above, at the relevant time, the applicable debt to equity ratio to stay within the thin capitalisation safe harbour in Australia was 3:1 (with no difference depending on whether the lender was an external lender or a group company). According to a memorandum by Deloitte dated 20 July 2007, there was no thin capitalisation limit on third party debt in Canada or Japan, but a limit of 2:1 and 3:1 for internal debt in Canada and Japan respectively. Although the Commissioner said that there was a thin capitalisation limit of 7:1 applicable to external debt in those countries, nothing turns on the difference between the parties' positions on whether there was a thin capitalisation limit on external debt in Canada and Japan. The relevant point is that, when Mylan ultimately proceeded with Canada taking on intercompany (not external) debt, the amount borrowed dropped to correspond with the safe harbour limit in Canada that applied to internal debt. 313 Although the Commissioner made much of Mylan tracking the safe harbour limits in relation to the dominant purpose enquiry, the relevant point for present purposes is that, where it decided to capitalise a local acquisition vehicle with a mix of debt and equity, Mylan was careful not to exceed the safe harbour; that much is to be expected. Moreover, given the inability to claim foreign tax credits given its OFL position, there is no reason to think that, had it not proceeded with either of the secondary or tertiary schemes, Mylan would have had MAPL take on less debt than in fact it did take on. As Mr Glenn's analysis demonstrated, the more equity relative to debt, the greater the impact of the OFL on the worldwide tax burden on earnings of Alphapharm. 314 While the counterfactual by which the existence of a tax benefit is assessed cannot be the same as the scheme, it can share some features in common with the scheme: Ashwick at [153(4)] (Edmonds J, Bennett and Middleton JJ agreeing) citing AXA at [131]-[133] (Edmonds and Gordon JJ) and Trail Bros at [28]-[29] (Dowsett and Gordon JJ). In my view, the amount of debt is such a common feature. The scenario most likely if MAPL had not been capitalised with loan funds from Lux 1 via PN A2, is that it would have borrowed 75% of the purchase price of Alphapharm, either from an external lender, or from Mylan or another US subsidiary. 315 It was common ground that the principal of the loan posited under the counterfactuals needed to be set as at 2 October 2007. The debt to equity ratio that underpins the Commissioner's counterfactuals (54.6% debt to 45.4% equity) was the group-wide ratio that was derived only at year end in December 2007, and after the equity raising in November 2007 turned out to be more successful than had been anticipated, to the tune of USD 889 million (raising USD 2.89 billion, against an anticipated target of USD 2 billion). I can think of no rational reason why Mylan would have decided, at the start of October 2007, to gear MAPL at the 54.6:45.4 ratio, which was as yet unknown. 316 Of course, the precise debt level of MAPL proposed by the Commissioner (or figures lower than the 75% ratio proposed by MAHPL) should not be rejected only because the ultimate group gearing after the equity raising could not have been anticipated in early October 2007. However, given the points I have already noted, in my view the most likely scenario in the absence of the schemes would still have seen MAPL take on the same level of debt (75% of the price of Alphapharm) as it in fact assumed under the schemes (cf taking on some lower level of debt). 317 By the time the parties closed their cases, it was common ground that the balance of MAPL's funding would have comprised equity; in closing his case, the Commissioner abandoned the suggestion (raised by his articulation of the secondary counterfactual and the tertiary counterfactual) that MAPL may have been partly capitalised by non-interest bearing loans.

[47]

Capitalisation of interest 318 The Commissioner's secondary and tertiary counterfactuals posited that interest would not be capable of being capitalised. By contrast, counterfactual A (internal lending), put forward by MAHPL, posited a borrowing on the same terms as PN A2 (such that interest could be capitalised). 319 There was some debate at trial regarding whether MAPL's earnings would have been sufficient to support a borrowing equivalent to 75% of the purchase price of Alphapharm. While this debate primarily arose in relation to the debt to equity ratio, it has consequences for the question of whether interest would have been capitalised. 320 Analysis undertaken by Mr Stack, albeit based on fixed interest rates, did show that, having regard to projected earnings, the borrowing (assuming either the initial principal amount of PN A2, the amended principal amount, and the loan balance as at the end of 31 December 2008) was affordable without capitalising interest on the basis that there was free cashflow in excess of interest expenses. 321 In addressing Mr Stack's evidence, the Commissioner was critical of his reliance on earnings projections provided by the vendor side. 322 I consider that a company in Mylan's position would likely not rely wholly on vendor projections of earnings, but would, for example, look at the stability of historic earnings (of which there was only limited evidence in this proceeding - principally PwC's draft valuation, which set out historical earnings figures from 2004-2006). 323 Nevertheless, in the context of this acquisition, in my view, Mylan would have been prepared to have MAPL enter into a facility that did not allow for interest to be capitalised on the basis that interest rate volatility risk could be internally managed (as set out above), and the risk associated with Alphapharm's earnings falling short of expectations could be managed by recourse to intercompany loans, if necessary. In short, I do not consider that a counterfactual which did not allow for interest to be capitalised would have been rejected by Mylan as unsuitable for MAPL for that reason. Accordingly, the Commissioner's counterfactuals are not unreasonable merely because they provide for facilities that did not allow for interest to be capitalised.

[48]

Internal or external lender 324 Both the Commissioner and MAHPL put forward counterfactuals positing MAPL borrowing from an external lender (the secondary counterfactual and counterfactual B), and borrowing internally from Mylan or another US Mylan subsidiary (the tertiary counterfactual and counterfactual A). In closing, the Commissioner contended an external borrowing was more likely than an internal borrowing (while of course contending his primary counterfactual was the most likely counterfactual - a proposition I have already rejected). MAHPL, however, contended that it was "marginally more likely" that Mylan would have opted for an internal loan over external debt. Counsel for MAHPL characterised the choice between an internal and external borrowing as a "close call". 325 By the close of the trial, it was also common ground that, if MAPL were to have borrowed externally, it would have been introduced as an additional borrower under the SCA, and would have borrowed in USD or EUR and swapped the borrowing to its functional currency, AUD. 326 I accept that, as MAHPL submitted, an internal borrowing would have been simpler than an external borrowing from Mylan's perspective. In addition, an intra-group borrowing would inherently be more flexible than an external borrowing. 327 In his evidence, Mr Stack explained that MAPL borrowing funds externally would add unnecessary complexity. As Mr Stack explained, US-based multinationals typically have a corporate treasury function with two offices: one in the US, and one in Europe. He explained that, typically, multinationals would borrow externally at this level, and then send funds where needed by internal financing arrangements (such as intercompany loans), and it would make no sense to add in a third, Australian-based borrower, when that was not necessary. Mr Stack also explained that, from the lenders' perspective, they also prefer a single point of contact (rather than distinct borrowings by subsidiaries). That said, Mr Stack considered that it would be less inefficient to add a company like MAPL in as an additional borrower under the SCA, than have it enter into a standalone facility. 328 Of course, and as Mr Stack recognised, his evidence was given on the basis of his experience as a corporate treasurer. He acknowledged that ultimately financing decisions would take into account considerations other than those advanced by the Treasurer of a multinational, including tax advice coming to the CFO from others in the company. 329 Based on treasury concerns alone, it would appear more likely that MAPL would have borrowed internally than externally. But the contemporaneous documents make it clear that Mylan seriously considered an external borrowing by an Australian holding company subsidiary as an alternative in its transaction planning. While the choice was ultimately made that MAPL should borrow via the intercompany notes, the documents do not record there being any consequence of MAPL borrowing externally that would have made it an intolerable choice. On the contrary, the fact that an external borrowing was repeatedly included as an option in planning documents over a period of three months (from late April 2007 to late July 2007) in what was otherwise a fairly tight timeline for such a significant acquisition speaks to external borrowing by the Australian subsidiary being an option Mylan would (and did in fact) countenance. 330 Between 27 April 2007 and 30 May 2007, Deloitte produced four versions of its "Project Genius - Tax Overview" document, which included MAPL borrowing USD 750 million (being 75% of the assumed USD 1 billion for Alphapharm) from an external lender. Deloitte's "Financing Overview" dated 4 June 2007 also provided for MAPL to borrow USD 750 million from an external lender. This was proposed for both alternative funding structures, covered by that document. When PwC was developing structure alternatives, as at 19 July 2007, it described external funding of MAPL as the "base case", but went on to consider internal funding structures (although I note that it is not clear on the evidence whether Mylan received this document from PwC contemporaneously). The serious, and ongoing, consideration being given to external funding of MAPL (and some other acquiring subsidiaries) is reflected in the detailed Deloitte memorandum "Summary of Third Party Borrowing Considerations" dated 20 July 2007. That memorandum noted benefits of direct borrowing from third parties (albeit in local currencies) including relief from thin capitalisation limits in Canada and Japan that applied where the lender was related, reducing or avoiding withholding tax on interest payments, and avoiding intercompany interest income that would be subject to tax in the lender's country. The more detailed section of the memorandum addressing Australia set out the impact of withholding tax, and noted that Mylan may or may not be able to credit that tax for US foreign tax credit purposes. 331 Further, provision was made in the draft terms sheet, and then in the final Commitment Letter issued by the lenders, allowing for additional borrowers (including non-US borrowers) to be added on terms and conditions to be agreed, if requested by Mylan. Provision for such additional borrowers to be included in the SCA remained when a Further Commitment Letter was issued on 11 May 2007, and an Amended and Restated Commitment Letter was issued on 18 June 2007. Clearly, the idea of additional Mylan subsidiaries borrowing externally was still in serious contemplation and had not been ruled out when these letters were obtained from the lenders. In addition, in July 2007, Deloitte was still considering third party borrowing, and assessing the relative merits of external borrowings as compared with internal loans (noting, inter alia, the withholding tax that would apply to intercompany loans). 332 By contrast, no documents contemplated an internal loan being made by Mylan or another US subsidiary. In closing, MAHPL sought to refute the point that an internal loan by Mylan had never been contemplated, and pointed to a document that did refer to "Mylan US" making an internal loan to "Aus Hold Co" (which was "Structure 2" of five structures set out in the document). That document was a PwC document (in draft form) titled "Mylan Laboratories Structure Alternatives" and was dated July 2007. The Commissioner accepted that that document was in Mylan's possession by August 2022, but did not accept that the document was provided to Mylan contemporaneously, and so may have been internal work at PwC that did not reflect thinking in Mylan at the time. 333 This document was the only document MAHPL pointed to as suggesting there had ever been any contemplation of Mylan (or any US subsidiary) lending to the Australian holding company. I am not satisfied, based on that document, that there was any serious contemplation of an internal loan from the US. Not only is that the only document concerning that possibility, but there is no evidence it was ever presented to, or seriously considered by, Mylan. Further, the "Outcomes" and "Considerations" recorded in relation to "Structure 2" are at odds with the basis upon which it was accepted (at trial) that any loan made by Mylan or a US subsidiary would be structured and taxed. In particular, it was accepted at trial that the interest income on any internal loan to MAPL (once that income was remitted in some form to the US) would have been taxed in the US. However, "Structure 2" proceeded on the basis that interest payments would be disregarded. It was also accepted at trial that Mylan's OFL position would have prevented foreign tax credits being claimed for income and withholding taxes paid in Australia, whereas this presentation assumes that foreign tax credits could be claimed for both such taxes paid in Australia. 334 I should also mention three further matters that affect the choice between the internal and external loan scenarios. First, if MAPL borrowed from a related party, 10% withholding tax would have been payable in Australia. Given Mylan's OFL position, no foreign tax credits could be claimed for this. The Commissioner raised this point in support of an external borrowing counterfactual. However, Mylan was prepared to bear essentially the same cost in entering into the schemes, and there was expert evidence of Mr Stack that the 10% withholding tax may well be regarded as a "cost of doing business". Accordingly, I regard the fact that an external borrowing would avoid this 10% withholding tax cost as only providing modest support for the external lending counterfactual. 335 Secondly, if Mylan or another US subsidiary were the lender, interest would be taxed as income in the US immediately, and there would be no ability to take advantage of the "look through" rule. However, given the then-anticipated need to repatriate free cash flow to manage the group's external debt, it is doubtful that the inability to defer receipt of interest income would have constituted a significant disadvantage to an internal borrowing. 336 Thirdly, while the consequence of the desire to repatriate cash might, at first blush, be thought to favour an internal borrowing, that does not follow. With an external borrowing, it was common ground that Mylan and Lux 5 would have borrowed a commensurately lower amount. Accordingly, as earnings from Alphapharm would be deployed to reduce its share of the external debt under the SCA, Mylan and Lux 5 would not need or expect to be able to access free cash flow from Alphapharm's operations to meet their obligations and repayment objectives in respect of the balance of the external debt. Accordingly, I do not consider that an external borrowing would have been unattractive due to Mylan's expectations of being able to sweep cash from the subsidiaries to service the external debt. This is borne out by the documents, referred to above, which demonstrate that MAPL taking on external debt was seriously considered by Mylan. 337 While MAHPL was correct to characterise the choice between internal and external debt as a "close call", in my view Mylan was more likely to have had MAPL participate in the SCA than to borrow from Mylan (or another US subsidiary) directly. I have reached that view largely because Mylan did seriously consider, and take steps to lay the groundwork for, MAPL to borrow externally, whereas an internal borrowing from the US was never seriously explored.

[49]

The quantum of the counterfactual loan 338 The initial principal amount of PN A2 was EUR 502,500,000 (AUD equivalent 785 million) less those amounts owed by Alphapharm to related parties (estimated at EUR 17,548,000). That amount was subject to adjustment to the sum that was equal to 75% of the valuation of Alphapharm, as at 1 October 2007, as determined by PwC. That valuation was to be undertaken after closing. It ultimately was not finalised until February 2009. The initial combined total of PN A1 (which was replaced with equity) and PN A2 was EUR 670 million (equivalent to AUD 1.04 billion) (subject to the EUR 17.548 million adjustment just mentioned). The value placed on Alphapharm of EUR 670 million was the bottom end of the range of PwC's draft valuation as it stood at the time the Amended SPA was finalised. The valuation range at that stage was EUR 670 million to EUR 1.04 billion. 339 In December 2007, when those in Mylan were establishing the best estimate of the valuation of Alphapharm for use as a figure in calculating the provisional revised value of PN A2, they used the mid-point of the then-current draft of the PwC valuation. The mid-point valuation was EUR 760 million. 340 When PwC completed its valuation in February 2009, it valued Alphapharm at approximately AUD 1.23 billion, which resulted in the principal of PN A2 being retroactively increased to AUD 923 million (up from AUD 785 million). 341 It was ultimately common ground that, if MAPL had borrowed externally, the principal would be fixed once and for all as at 2 October 2007, and would not be varied retroactively, as in fact occurred under the internal borrowing. 342 The Commissioner's secondary counterfactual applied his preferred debt to equity ratio to the original combined value of Alphapharm adopted when the SPA, PN A1 and PN A2 were concluded, which resulted in his posited principal being EUR 356.239 million (54.6% of EUR 670 million less the adjustment for related party liabilities), which was the equivalent of AUD 571.72 million. In closing, the Commissioner revised the projected borrowing figure down to AUD 471.8 million, giving a debt to equity ratio of approximately 45:55. 343 In closing, MAHPL submitted that, if the principal were to be fixed once and for all as at 2 October 2007 (ie with no opportunity to revise it when PwC finalised its valuation), the mid-point of PwC's then-current draft valuation (EUR 855 million) would have been chosen as the figure of which 75% would constitute the principal of the loan (cf the selection of 75% of the low end of PwC draft valuation, used for PN A2, which could be retroactively adjusted once the valuation was finalised). Seventy-five per cent of the mid-point, converted into AUD is AUD 1,029,206,250. That is obviously significantly more than the initial principal of PN A2, which was the equivalent of AUD 785 million, and is also more than AUD 100 million more than the revised figure that was in fact adopted (which was the equivalent of AUD 923 million). 344 The AUD 1.03 billion figure was then used by MAHPL in an appendix to its submissions (Appendix C), which applied principal repayment obligations as they stood under the SCA to that principal, and then calculated interest payments due based on a 10.15% fixed interest rate. Applying that methodology to the high starting notional principal, MAHPL submitted that, in fact, MAPL would have paid substantially more interest under an external borrowing scenario to 2014 than it did under the facts as they occurred. Of course, that conclusion was driven by application of the same fixed interest rate, and the higher starting principal, as well as an assumption that MAPL would only have made the minimum repayments of principal (cf the lump sum payments of principal it in fact made under PN A2). 345 In order to evaluate MAHPL's submission as to the likely quantum of MAPL's borrowing on the counterfactual, it is necessary to address how the price for Alphapharm fitted in to the Amended SPA. Under the Amended SPA, Alphapharm was added as an "Additional Target Company", and MAPL was added as an "Additional Purchaser" and "Transferee". Pursuant to cl 4.1.1(b), MGGBV was to sell and transfer the shares in Alphapharm to MAPL in exchange for a promissory note in the form attached to the Amended SPA as Exhibit 4.1.1(b)(ii). That exhibit was PN Lux 1, which was in the amount of EUR 670 million (less the estimated related party liabilities of EUR 17.548 million). Under Exhibit 4.1.4, 65% of the total purchase price (after some deductions for specific assets) was allocated to MGGBV. 346 Had MAPL borrowed under the SCA, the structuring documents in existence when external borrowing was being contemplated provided for MAPL to pay cash to MGGBV for the shares in Alphapharm (cf the shares being transferred in exchange for PN Lux 1). Given that the changes to the original SPA were driven by Mylan, and it appears from the original SPA and the recitals to the Amended SPA, that Merck was prepared to accommodate Mylan's preferred structure (provided it obtained the whole purchase price and suitable indemnities were given for any adverse effects of those changes), I see no reason to suppose that there would have been any real impediment to Mylan choosing a different point in PwC's valuation range. 347 However, one significant point that MAHPL's submissions overlooked is that, if the quantum of an external borrowing were determined by adopting the mid-point of the valuation range, MAPL would have been exposed to a thin capitalisation ratio risk if the final valuation was lower than the mid-point of the draft valuation. It appears from documents tendered by MAHPL (in answer to the Commissioner's submission that MAPL's borrowings were determined to max out the thin capitalisation limit) that in fact MAPL maintained substantial "headroom" (approximately AUD 74 million in the year ending 31 December 2007, and AUD 98 million in the year ending 31 December 2008). This suggests a desire on the part of those in control not to run the risk of having MAPL exceed the thin capitalisation threshold. Accordingly, I am not persuaded that, had MAPL borrowed externally, Mylan would have determined that MAPL should run the risk of finding itself having exceeded the thin capitalisation ratio as a result of the final valuation coming in lower than the mid-point of the valuation as it stood at 2 October 2007. 348 Consideration must also be given to questions of debt servicing capacity. Could MAPL have afforded to borrow the greater sum posited by MAHPL in its closing submissions, and would its debt servicing capacity have served as a real constraint on the borrowing? At the stage of fixing the quantum of the counterfactual borrowing, I am concerned with what likely would have occurred, not with whether the paucity of debt analysis in the facts as they occurred, supports the Commissioner's dominant purpose contentions. 349 While Mylan was considering external debt being assumed by MAPL, Deloitte's "Project Genius - Tax Overview" dated 27 April 2007 assumed a ballpark value of USD 1 billion for Alphapharm, with MAPL borrowing USD 750 million from a third party lender. This figure appears to have been based on a ballpark allocation of a part of the overall assumed purchase price for the Merck group to Alphapharm, and applying the 75% thin capitalisation ratio; the debt figure does not appear to have been based on any analysis assessing MAPL's capacity to service debt at that level. 350 This figure for external borrowings by MAPL, and the concern not to stray over the thin capitalisation limits, continues through the documents referred to above, which set out transaction structures involving external borrowing by the Australian holding entity. 351 It is not surprising not to see debt servicing analysis in a tax structing advice by Deloitte, but nonetheless, there is also no evidence of specific attention being paid to the debt servicing capacity of the Australian business when external debt was being considered, in other documents tendered by the parties. At one level, this indicates that, at least when assessing matters at a broad brush level and in assessing its structuring options, Mylan was not concerned to establish debt levels by reference to the servicing capacity of individual subsidiaries. 352 While Mylan would have taken (and did, it seems, take) a broad brush approach to the quantum of external borrowings by MAPL when looking at how it would structure the Acquisition, that does not mean that it would have been indifferent to MAPL's debt servicing capacity when (on the counterfactual) having MAPL borrow under the SCA. That said, in my assessment, Mylan would also not have been overly cautious as MAPL's borrowing under the SCA would be guaranteed by it, and intra-group financing could be readily arranged if necessary for Alphapharm to meet its obligations to the SCA lenders, particularly in the early years of the loan when the quantum of the servicing obligations would have been higher. Further, to the extent that the Australian holding entity made losses due to an interest to earnings mismatch, carry forward losses would have been available (subject to limitations, as PwC noted in its 19 July 2007 "Mylan Laboratories Structure Alternatives" document). 353 While, as noted, there was very little evidence of consideration by Mylan of debt serviceability in Australia, Mr Stack conducted an analysis of projections for Alphapharm's business, by which he concluded that Alphapharm was expected to have sufficient free cash flow to service the after-tax interest expense on PN A2. Mr Stack calculated that, applying an interest rate of 10.79% to the initial principal (EUR 502,500,000, which he converted to AUD 804 million), the annual interest cost would have been AUD 87 million, which, after being assessed on an after-tax basis, would leave free cash flow in excess of the interest expense of AUD 30 million in 2008, AUD 50 million in 2009 and AUD 82 million in 2010. Mr Stack also calculated the remaining free cash flow based on the increased principal of note PN A2 (AUD 923,205,336 million) at the fixed interest rate of 10.15% and his figures continued to show an excess of free cash flow after interest expenses. Mr Stack also repeated the exercise based on the paid-down principal of PN A2 at the end of December 2008, which showed improved post-interest free cash flow in 2009 and 2010 (similar to the figures on his first scenario). 354 The Commissioner advanced two criticisms of this analysis. First, that Mr Stack's analysis was not in fact available to Mylan in 2007. Secondly, that Mr Stack used projections from Alphapharm that came from KPMG, which was acting for the vendor. The Commissioner pointed to PwC (acting for Mylan) having adopted more conservative figures. 355 I do not consider the first criticism to be relevant when the earnings expectations regarding Alphapharm are being considered from the perspective of whether a proposed borrowing would have been unaffordable for MAPL. Of course, the absence of consideration of such figures may still feature in the analysis of what the likely quantum of debt would be on the counterfactual. 356 As to the second criticism: Mr Stack justified his use of sell-side projections on the basis that those figures were the most complete information that would have been available by October 2007. Mr Stack appears to have been of the understanding that PwC's figures were not available before October 2007, although in fact PwC did provide figures in a draft valuation in August 2007. There was some debate at the closing stages of the trial regarding whether Mr Stack was provided with those August 2007 PwC figures (cf a later iteration of PwC's draft). However, as Mr Stack clearly stated why he used the KPMG figures, and what his understanding was about the availability of PwC figures before 2 October 2007, it does not matter whether or not he was provided with the pre-closing draft valuation of PwC. 357 A comparison may be made between two sets of EBITDA projections (in EUR), one set of figures being those derived by Mr Stack from KPMG's vendor-side due diligence report, and the other set being those presented in PwC's August 2007 estimates, which were prepared for Mylan. It should be noted that the KPMG EBITDA figures excluded R&D (research and development), whereas R&D was not excluded from PwC's figures. Including a line to normalise for R&D, those figures were as follows: 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 PwC August 2007 EBITDA figures (EUR millions, inclusive of R&D costs) 35 50 49 61 103 120 126 135 141 149 157 PwC projected R&D costs (15) (15) (13) (13) (12) (12) (12) (13) (13) (13) (14) PwC August 2007 EBITDA figures exclusive of R&D costs 50 65 62 74 115 132 138 148 154 162 171 KPMG EBITDA (EUR millions, exclusive of R&D costs) - used by Mr Stack 48 63 72 80 94

[50]

358 As may be seen, once the adjustment for R&D is made, PwC's figures were somewhat higher than KPMG's figures in 2007, more conservative in 2008 and 2009, but then substantially higher in 2010. Further, PwC was projecting substantial growth beyond the period referred to by Mr Stack (see PwC's projections for 2011 onwards), including in the period through to 2014, which represents the period of the initial financing. 359 As such, Mr Stack's use of the KPMG figures does not really undermine his analysis, particularly once it is taken into account (as I think it must be) that Mylan would approach any analysis of MAPL's capacity to service external debt with a longer term view than just the first couple of full years (2008-2009). Both Mr Stack and Mr Ali gave evidence to the effect that they would expect Mylan to take a relatively long term view and be prepared to assist with any short-term shortfalls in the capacity of a subsidiary to service its debts. They were not challenged on this evidence. 360 I have reservations about the utility of approaching questions of serviceability by simply comparing the gross anticipated EBITDA with the anticipated interest expense (as MAHPL sought to do in closing, including to justify the affordability of the larger external loan it proposed (in Appendix C to its closing submissions)). A serviceability analysis requires (as the Commissioner pointed out) consideration of some more granular factors. Nevertheless, I do consider that MAHPL's comparison between the gross EBITDA anticipated by PwC and the actual interest expense incurred by MAPL under PN A2 (AUD 1.2 billion in EBITDA versus AUD 551 million in interest, being the total interest calculated at the fixed rate of 10.15% on the AUD 923 million principal, as calculated by Mr Johnston) to illustrate at a broad level, alongside Mr Stack's analysis (which was not challenged in cross-examination), that Mylan would not have regarded MAPL taking on an external borrowing of EUR 502,500,000 (or EUR 484,952,000 if adjusted to account for the EUR 17.548 million adjustment) as a proposal it would reject on the basis that it was excessive and unserviceable by MAPL (albeit that some group support may have been anticipated to be required in the early years). 361 As I have mentioned, the Commissioner, in closing, proposed that the quantum of any external debt assumed by MAPL would be lower than was stated in his articulation of the secondary counterfactual. In his closing submissions, the Commissioner developed a methodology which resulted in a proposed external debt of AUD 471.8 million. That figure was derived by applying a Net Debt/EBITDA ratio of 6x to Alphapharm's projected EBITDA in 2008 (based on PwC's August 2007 draft figures) of EUR 49 million (converted to AUD 78.6 million). The Net Debt/EBITDA ratio was said to be the "near-term" gearing target. Mylan presented a ratio of 8.3x in its presentations to the ratings agencies at the end of September 2007 as the immediate, post-acquisition pro forma ratio and a pro forma "permanent" ratio of 6.1x. However, as even PwC's more conservative EBITDA projections for Alphapharm specifically situated 2008 as a low point, I do not consider it realistic to suggest that Mylan would have limited MAPL's borrowings based on Alphapharm's projected 2008 EBITDA, essentially disregarding the very substantial up-swing in EBITDA that was being projected. In addition, as Mr Ali explained, the amount of debt taken on by subsidiaries may not track the group's gearing exactly given local factors that may suggest a higher, or lower, gearing ratio for the subsidiary. 362 In summary, having regard to debt serviceability considerations and the thin capitalisation risks associated with a higher borrowing in Australia, I do not consider it likely that Mylan would have taken a more aggressive position (had MAPL borrowed externally) and chosen the mid-point of the PwC draft valuation. As I have set out, I am also not of the view that Mylan would have preferred a lower level of debt to be taken on by MAPL on the external borrowing counterfactual. Accordingly, in my view, the borrowing would have, or might reasonably be expected to have, been EUR 489,339,000 (being 75% of (EUR 670 million less EUR 17.548 million)), converted into AUD 785,329,802.60. 363 For completeness, I note that the parties' submissions on serviceability also referred to an email dated 14 July 2007 circulated within PwC. That email attached a country by country analysis to assess "each country's interest capacity". While MAHPL positioned this as a document by which debt service capacity was assessed, the Commissioner contended it was a document directed to ascertaining how much interest each country would need to bear to eliminate taxable profits. I will return to this document in the context of dominant purpose, but note that, for the counterfactual analysis, I have not found the document to be of assistance as it did not tie the figures set out for Alphapharm to any provisional interest expenses or loans.

[51]

Interest: Fixed or floating borrowing 364 As mentioned above, it was ultimately common ground that, if the preferred counterfactual involved MAPL borrowing externally, it would have done so under the SCA. Consequently, MAPL's borrowing would have been at the floating interest rates provided for in the SCA (but not subject to the special terms that applied to some components of the debt such as the US Tranche A Term Loan, on which the applicable interest rate varied with the consolidated leverage ratio, and not just with the credit rating). 365 Given the common ground just noted as to the counterfactual external borrowing being under the SCA, it is not strictly necessary to address the expert evidence on this point. However, I note that the evidence was consistent with it being more likely that MAPL would borrow on the bank market at floating rates, than on the bond market, at fixed rates. The expert evidence was to the effect that, while not without exceptions, in general the bank market loaned at floating rates, and the bond market loaned at fixed rates. 366 It was not disputed - and was the evidence of Mr Johnson, the debt capital markets expert called by the Commissioner - that floating rate obligations can be paid out at any time without significant penalty, and therefore are very flexible. Mr Johnson's evidence was also that adding hedges to fixed interest would not undermine that flexibility as floating to fixed hedges can be exited at any time, with little or no penalty. By contrast, the penalties associated with early termination of fixed rate borrowings are significant, and set at a level to act as a real disincentive to the early repayment of the fixed rate borrowing. 367 To the extent that MAPL (contrary to my preferred counterfactual) was more likely to have borrowed internally, then I consider it would have borrowed at a fixed rate of interest so as to minimise the exposure of a subsidiary to interest rate risk. I refer to the expert evidence concerning fixed and floating interest rates below, but in the context of an intercompany borrowing, there would be no reason not to fix the interest rate. As such a borrowing would, on this hypothesis, have been entered into on 2 October 2007, and as there was no evidence that that was not a market interest rate for a fixed interest borrowing with a seven year tenor, the borrowing would have been at 10.15% and otherwise on the same flexible terms as MAPL in fact borrowed from Lux 1 pursuant to PN A2.

[52]

Fixing floating rate exposure of MAPL 368 The next issue is whether, had MAPL borrowed externally under the SCA at a floating rate, it would have entered into hedges to fix some or all of its floating rate exposure. 369 There are some features of the expert evidence of Mr Stack and Mr Ali, which I accept, that puts this debate in context. Their evidence was to the effect that: (a) the predictability of fixed interest rates means that fixed interest debt presents less risk (it removes interest rate risk); (b) where an entity's leverage is relatively high, and also where an operating entity is newly acquired, predictability of debt service obligations will typically be regarded as having added importance; (c) in multinational groups, the tolerance that a subsidiary will have for both foreign exchange risk and interest rate risk is a matter that will typically be determined at the group level by the group's treasury function; and (d) in multinational groups, the general preference is for both foreign exchange risk and interest rate risk to be centralised, managed and borne at the group level - cf being borne by subsidiaries whose task is to conduct operations and generate profit, and not to attempt to trade on the financial markets and seek advantage through interest rate speculation - while group level treasury management could set the group position having regard to natural hedges from intra-group exposures. 370 Mr Johnson did not disagree with these points, per se. Mr Johnson also considered that multinationals may have centralised borrowing and hedging operations, accepted that there would be a centralised group view of overall group exposure, but considered that foreign exchange and interest rate risk may, in some groups, be borne by subsidiaries (albeit guided by, and limited by, any group policies and positions). The greater focus of Mr Johnson's disagreement was with the other experts' views as to the benefits of fixed rate borrowings. Mr Johnson stressed that fixed rate borrowings are typically inflexible (given break costs) and considered that most companies would consider a 50:50 fixed and floating rate exposure as a "neutral" position, with deviations from that position being considered, responsive to the entity's circumstances. Of course, a fixed rate borrowing should not be confused with a floating rate borrowing that is then hedged by swaps. The evidence of Mr Johnson was that interest rate swaps could be readily arranged, and exited at little to no cost. Given my determination that MAPL would have borrowed under the SCA, it follows that the borrowing would have been a floating rate borrowing, so the question is how much of its debt MAPL would have hedged, whether the hedges would have been internal or with external parties, and the cost of the hedges. 371 Mr Johnson modelled the extent to which MAPL would have incurred less interest than it in fact incurred, had it fixed different proportions of its debt. He modelled a fixed percentage of debt at 22%, 41% and 30%. Mr Johnson's view was that fixing approximately 30% of MAPL's debt would meet what he took to be its objectives. What is relevant at this point is the percentage that would be fixed, and not the savings in interest that may have been achieved had MAPL retained some floating rate exposure. It should be noted that the expert called by the Commissioner was supportive of at least some of MAPL's interest rate exposure being fixed, whereas (as set out in the Commissioner's appeal statement) the secondary counterfactual posited an entirely floating rate exposure. In his closing submissions, the Commissioner submitted that, if it were contended that MAPL would have entered into interest rate swaps, "[i]t could logically be assumed that MAPL would have been party to the same swaps that Mylan Inc. in fact entered into in relation to the financing provided under the SCA". That point leads, then, to the question of what proportion of Mylan's group debt was fixed. 372 On Mr Johnson's analysis, approximately 22% of Mylan's acquisition debt of USD 4.5 billion was hedged through USD 1 billion of 3 year interest rate swaps. Mr Stack disagreed with Mr Johnson's analysis, and pointed to the following matters: (a) Mr Johnson had only based his calculation on debt taken on for the Merck acquisition, and not Mylan's overall exposure to fixed versus floating rates; (b) Mr Johnson excluded USD 600 million in senior convertible notes which, on generally accepted accounting principles, was treated by Mylan as debt; (c) if those notes, and the further USD 850 million in senior notes that Mylan intended to issue, were included, then Mylan's percentage of fixed rate debt would rise to 25% and, if account were taken of the USD 1 billion in swaps that Mr Johnson referenced, the fixed rate percentage would climb to 42%; (d) given the success of the equity raising, Mylan's fixed percentage was in fact 30%; (e) the proportion of Mylan's fixed rate debt would climb as the principal due under the SCA amortised; and (f) in the period following the end of 2007, steps taken by Mylan (which included issuing new cash convertible notes fixed rate debt in September 2008) resulted in Mylan's fixed rate percentage being 60% at the end of December 2008, and 65% when additional liabilities under the SCA were fixed in February 2009. 373 The matters highlighted by Mr Stack show a number of things. First, it is not always straightforward to calculate how much of an entity's debt is fixed or floating. Secondly, steps may be taken at different points in time to manage the fixed to floating rate exposure, as a company sees fit. Thirdly, while there may be room for argument about the exact number, at the group level, Mylan had tolerance for substantial floating rate exposure initially, but quite rapidly reduced that exposure. 374 It is not known what plans, if any, Mylan had in October 2007 in relation to the extent to which it would fix its interest rate exposure. What is known is that Mylan obtained advice from a number of sources, all of which recommended fixing significant portions of its debt. In a document dated 24 May 2007, Merrill Lynch referred to Mylan's pro forma liability portfolio (being the post-acquisition pro forma) having 70% floating rate debt, cautioned about the risk posed by interest rate volatility, and recommended that Mylan target a floating rate mix of only 7% floating. On 25 June 2007, Citi recommended Mylan fix all of its interest rate exposure between 2008 and 2010, then introduce 10% to 20% floating rate exposure between 2011 and 2014. In a document headed "Interest Rate Risk Management Discussion" dated 27 June 2007, Merrill Lynch observed that: "Leveraged companies typically maintain 80-100% of their liabilities in fixed interest rates to immunize against adverse effects on interest expense and cash flow generation." Merrill Lynch repeated its 7% floating rate exposure recommendation in the same document. While Mylan did fix an increasing portion of its debt over time, it clearly had a higher tolerance for interest rate risk than its advisers were recommending, despite being highly leveraged. 375 How, then, does all that translate to the counterfactual borrowing of MAPL as a borrower under the SCA? On this counterfactual, the principal to be borrowed by MAPL would have been a substantial portion of the overall SCA quantum, but it would not have approached the total quantum of floating rate debt that Mylan was willing to retain. In my view, it is likely that Mylan would have sought to manage its group-wide exposure to floating interest rates at a group level, even if MAPL were a borrower under the SCA. The group-wide proportion of fixed interest rates that Mylan desired to achieve did not require that any portion of MAPL's debt be fixed. In my view, and given the manifest willingness of Mylan to have subsidiaries enter into intra-group financing arrangements, Mylan would have left the MAPL exposure as a floating rate exposure, initially at least, on the basis that intra-group financing via promissory notes could be arranged if necessary to manage MAPL's obligations to the lenders under the SCA. There is evidence of liquidity being managed within the Mylan group by intercompany promissory notes, which were (at least in some respects) treated somewhat flexibly (such flexible treatment being evidenced by, eg, Lux 1 calling on MAPL, on 31 December 2008, to pay a sizeable amount of accrued and outstanding interest under PN A2 when it had no right to demand that payment under the terms of the note). While Mylan would, in my view, have kept fixing some of MAPL's debt open as an option, it would not have fixed the entirety of that debt on day one (2 October 2007), as MAHPL contended (and as the figures proffered by MAHPL in Appendix C to its closing submissions supposed). 376 I accept MAHPL's submission that, as at 2 October 2007, Mylan and MAPL could not have foreseen that floating interest rates would drop as they did, and I accept Mr Ali's evidence about "fixed-floating equivalence", by which the deep and liquid swap market ensures that, on a particular day, there is no arbitrage opportunity in the market by picking fixed over floating, or vice versa. But the fact that interest rates did drop (as Mr Johnson's evidence demonstrates), and in light of Mylan having "locked in" some benefit from falling interest rates through the steps it took to fix other components of its debt (as set out by Mr Stack in his second report), suggests that there would have been no imperative for Mylan to arrange separate hedges of MAPL's exposure. 377 MAHPL suggested that Mylan would have fixed the whole of MAPL's counterfactual borrowings under the SCA internally. I am not persuaded that that is likely to have occurred as there is no evidence that entities in the Mylan group offered internal hedges. Such internal financing as was evidenced within the Mylan group was constituted by relatively straightforward promissory notes. By contrast, as Mr Ali explained, where the principal of a liability is paid down over time, it is necessary to enter into a series of hedges, to match the varying quantum. If MAPL were a borrower under the SCA, it would have quarterly principal repayment obligations. Even if a rather rougher (or less exact) hedge profile were assumed, a great many hedge instruments would still have been required. I doubt that Mylan would have embraced such "fiddly" internal financing arrangements, given its evident preference for keeping internal financing arrangements simple. In so concluding, I am mindful that, on the facts as they ultimately occurred, there was no occasion for Mylan to consider internal hedges, but there was no consideration on the evidence of hedging MAPL's exposure, even at a time when external financing of MAPL was being actively considered by Mylan. 378 Another reason why I consider it unlikely that Mylan would have arranged for 100% of MAPL's borrowing to be fixed from the outset (as MAHPL contended for) is that, under the terms of the SCA, the credit spread of the borrowings reduced as Mylan's credit rating improved. As at October 2007, Mylan intended to reduce the group's leverage, and it is apparent from the ratings agencies' reports in evidence that Mylan's leverage was a significant factor impacting its rating. It was not in dispute that, had MAPL borrowed under the SCA, its obligations would have been guaranteed by Mylan, and the pricing of its debt would have been the same as the pricing available to Mylan. Accordingly, MAPL stood to benefit from lower interest rate expenses as Mylan's credit rating improved (as it was expected to with reducing group-wide leverage). Locking in a fixed rate swap for the entirety of the MAPL borrowing on day one would have deprived MAPL of the ability to benefit from positive changes to Mylan's credit rating.

[54]

Amortisation of principal 382 Under the terms of the SCA, the required repayments were set quarterly, but the borrowers were permitted to pre-pay at any time, without penalty (provided notification requirements were met). The SCA provided a repayment schedule for each component of the debt, specifying the repayments of principal required each quarter. Other than in respect of the US Tranche A Term Loan, the repayment schedules for the other actual components of the debt provided for the same payment to be made each quarter and were relatively modest when compared to the principal. For example, the repayment schedule for the US Tranche B Term Loan only required a total repayment of USD 130 million of the initial USD 2 billion principal. By contrast, the repayment schedule for the US Tranche A Term Loan increased the quarterly payments due each year (from an initial USD 25 million in 2008 to USD 125 million in 2012) and, overall, were set to reduce the principal from USD 500 million to USD 62.5 million by the maturity date. 383 In addition, the SCA required certain additional mandatory prepayments be made based on free cash flow (the percentage to be applied as mandatory prepayments depended on the consolidated leverage ratio) and the net sales of assets (subject to conditions and calculations). 384 Given the differential treatment of the US Tranche Term Loans, it seems more likely that, had MAPL been included as a borrower under the SCA, its repayment obligations would have followed the pattern adopted for the other term loan components of the SCA (0.25% of the principal due each quarter). However, MAPL would, or might reasonably be expected to, also have been able to make voluntary prepayments of principal without penalty. 385 In my view, and as the Commissioner submitted, MAPL would also have made such additional payments as were necessary to ensure it remained within the thin capitalisation safe harbour limit (including as it was varied - to 1.5:1 (for non-ADI entities) - by Pt 1, Sch 1 to the Tax and Superannuation Laws Amendment (2014 Measures No. 4) Act 2014 (Cth)), which would have been funded (as they were on the facts as they occurred) by additional equity. Without detailed calculations, it is not possible for me to say at this stage whether such additional equity would have been required, and, if so, how much would have been required. To avoid doubt, I am not of the view that voluntary prepayments under the external borrowing counterfactual would necessarily have tracked the payments of principal made under PN A2 as it is not clear whether, and to what extent (if at all), the payment of principal made in 2008 of AUD 105 million was necessary to avoid breaching the thin capitalisation safe harbour ratio given that interest under PN A2 had, to that point, been capitalised (and, after payment down of some principal, continued to be capitalised thereafter).

[55]

Effective interest rates 386 As set out above, had it borrowed under the SCA, MAPL would have borrowed subject to the floating interest rates prescribed by that agreement, as they varied from time to time. However, as noted, it was common ground that the borrowing would have been in USD or EUR, and would have been swapped into AUD, being MAPL's functional currency. Swapping the borrowing to AUD would mean that Australian operations would not bear foreign exchange risk. That is consistent with the evidence in relation to the centralisation of foreign exchange risk, as described above. Mr Ali also gave substantial evidence on why he would expect MAPL to have hedged its foreign exchange exposure, in Section 12 of his first report. He was not cross-examined on that evidence, and I accept it. 387 Accordingly, some costs would have been incurred in swapping the exposure to AUD, which must be added to the floating interest rate. In his first report, Mr Ali calculated the prices of AUD/USD cross currency basis swaps, and AUD/EUR cross currency basis swaps as set out below: 78. Based on an underlying USD floating rate loan with a margin of 325bps over USD LIBOR, I estimate the aggregate costs associated with an AUD/USD cross currency basis swap for MAPL to have been 51 - 61 bps per annum, resulting in an AUD equivalent margin range of 376 - 386 bps over AUD 3 month BBSW, as shown in Table 3.1 below. 79. Based on an underlying EUR floating rate loan with a margin of 325bps over EURIBOR, I estimate the aggregate costs associated with an AUD/EUR cross currency basis swap for MAPL to have been 52 - 59 bps per annum, resulting in an AUD equivalent margin range of 377 - 384 bps over AUD 3 month BBSW, as shown in Table 3.2 below. 388 As may be seen, there is very little difference depending on whether the swap is from USD or EUR. As Mr Ali set out a high and a low range, the mid-point should be selected. While I note that Mr Ali's analysis is based on a greater principal for the loan than I have settled on, I am unable to conclude from his analysis that lowering the principal would affect the per annum basis point cost of the swap. Accordingly, the mid-point cost that I consider MAPL would have incurred to swap the exposure into AUD is 3.81% per annum over AUD 3 month BBSW.

[56]

Guarantees 389 It was common ground that, if MAPL were to have borrowed externally, the borrowing would have been guaranteed by Mylan, meaning that the borrowing could have been secured with pricing reflecting Mylan's credit rating (rather than MAPL's credit rating, which would have been inferior). There was, however, debate as to whether Mylan would have charged a guarantee fee. If a guarantee fee were to have been charged, this would have increased the cost of the borrowing in Australia, and thereby the tax deductions in Australia, but would also have resulted in a corresponding uplift in the gross income of Mylan in the US. 390 The actual transactions entered into involved a number of guarantees having been given. Mylan guaranteed Lux 5's borrowing under the SCA, and Mylan guaranteed Lux 1's obligations under PN Lux 1. There is no evidence that any guarantee fee was charged by Mylan, and I infer that no guarantee fee was in fact charged for the giving of these guarantees. I therefore conclude that, to the extent that Mylan was even mindful that a transfer pricing adjustment may be made in the absence of a guarantee fee being charged, it in fact ran that risk in 2007. Had MAPL borrowed externally in 2007 to fund the Alphapharm acquisition, I conclude that Mylan would likewise not have charged any guarantee fee. As I consider it more likely that MAPL would have borrowed externally (than internally) in 2007 (in the absence of the schemes), this is the relevant conclusion. 391 For completeness, I will note that, had MAPL borrowed from Mylan or a US subsidiary in 2007, my conclusion is that no guarantee fee would have been charged. It would be economically pointless for Mylan to guarantee its own loan to MAPL and, had another US subsidiary been the lender, if Mylan considered the risk of default by MAPL ought to have been borne by the parent company (and not the lending US subsidiary), no guarantee fee would have been charged, as no such fee was charged in respect of PN Lux 1.

[57]

Conclusion on the counterfactual to be adopted 392 I am mindful that it is open to the court in a Pt IVA case to conclude that the most realistic counterfactual is a set of circumstances and events that includes features of the scheme: Ashwick [153(4)] (Edmonds J, Bennett and Middleton JJ agreeing); AXA at [131]-[133] (Edmonds and Gordon JJ). Between the parties, three counterfactuals were posited in respect of the primary scheme: the Commissioner's primary counterfactual, MAHPL's counterfactual A and MAHPL's counterfactual B. As I have concluded that the Commissioner's primary counterfactual does not constitute a sufficiently reliable (and thereby reasonable) prediction of the events which might have taken place (had the primary scheme not been entered into), I do not have regard to it in assessing whether MAHPL obtained a tax benefit in connection with the primary scheme, within the meaning of s 177C(1)(b). 393 The facts that I consider would have, or might reasonably be expected to have, occurred, in the absence of any of the schemes (including the primary scheme), are closest to MAHPL's counterfactual B and the Commissioner's tertiary counterfactual. 394 The features of what I will refer to as the preferred counterfactual are as follows: (a) MAPL would have borrowed the equivalent of AUD 785,329,802.60 on 7 year terms under the SCA (specifically on the same terms as applied to the US Tranche B Term Loan), at a floating rate consistent with the rates specified in the SCA; (b) MAPL would otherwise have been equity funded to the extent necessary to fund the initial purchase of Alphapharm and to stay within the thin capitalisation safe harbour ratio from time to time; (c) Mylan would have guaranteed MAPL's borrowing under the SCA; (d) Mylan would not have charged MAPL a guarantee fee; (e) interest on the borrowing would not have been capitalised; (f) MAPL would have been required to pay down the principal on a schedule consistent with that specified in the SCA and would have made voluntary repayments to reduce its debt if necessary to stay within the thin capitalisation safe harbour, from time to time; (g) MAPL would not have taken out hedges to fix some or all of its interest rate expense; (h) MAPL would have taken out cross-currency swaps into AUD at an annual cost of 3.81% per annum over AUD 3 month BBSW; and (i) if MAPL's cashflow was insufficient to meet its interest or principal repayment obligations, Mylan would have had another group company loan MAPL the funds necessary to avoid it defaulting on its obligations, resulting in MAPL owing those funds to that related company lender by way of an intercompany loan, accruing interest at an arm's length rate. 395 Upon the maturity date of the hypothetical borrowing under the SCA on a seven year term, in my view it is likely that MAPL would have refinanced the outstanding balance internally by a promissory note on the same terms as PN A4. I note that little turns on this particular point as the Commissioner did not take issue, per se, with the deductions claimed by MAHPL for interest on that note save to contend that, under the primary counterfactual, MAPL would not have any debt (and so there would have been no occasion for PN A4 to have been entered into in 2014). The parties did not otherwise engage with any distinct counterfactual for the tax years between 2014 and 2017, when interest under PN A4 was claimed as a deduction by MAHPL. 396 Had any of the schemes not been entered into, I consider that the most likely course of events is that set out above (ie the preferred counterfactual). The preferred counterfactual constitutes a prediction of the events which might have taken place, had any of the schemes not been entered into, which is sufficiently reliable, such that it may be regarded as reasonable.

[58]

Conclusion on "tax benefit" 397 By entering into the schemes, it appears that MAHPL obtained a tax benefit. I say "appears", as the tax benefit is the difference between the deductions it in fact claimed for interest expenses incurred by MAPL and the deductions that MAHPL would have, or might reasonably be expected to have, claimed had it proceeded according to the preferred counterfactual, but no calculations have yet been made to quantify that tax benefit. Although no calculations have been made, I am satisfied that there is likely to have been a tax benefit - and performing the necessary calculations will not yield a nil or negative number - because: (a) It was common ground that, in the years after 2007 (and in particular as the GFC took hold) interest rates dropped. MAHPL emphasised this point repeatedly in its submissions concerning no one having a "crystal ball" so as to be able to foresee that the fixed interest rates in fact set for MAPL's borrowings would turn out to be higher than a floating rate borrowing. As the preferred counterfactual involves a floating rate borrowing, it is clear that, other things being equal, the interest cost would be less than MAPL in fact incurred. (b) The preferred counterfactual does not involve any increase to the initial principal of the loan (contrary to MAHPL's submission that the counterfactual should involve an increased quantum). (c) The preferred counterfactual does not allow interest to be capitalised, thereby increasing the quantum of the borrowing. (d) Under the preferred counterfactual, the principal would be amortised through regular payments of principal. (e) The preferred counterfactual does not involve costs being incurred for floating to fixed hedges. (f) While the preferred counterfactual does involve cross-currency swaps, the cost of those swaps ought not absorb the full benefit of the significant interest rate advantage associated with the preferred counterfactual (noting also that MAPL stood to benefit from further reductions in interest rates as group leverage reduced). 398 I note that MAHPL contended that, because the amount of the tax benefit on the Commissioner's secondary and tertiary counterfactuals had not been specified in dollar terms, or by way of a formula, the Commissioner had not discharged an evidentiary onus that he bore. MAHPL's submissions on the consequences of this were not clear, but the point seemed to be that its appeals ought, therefore, to succeed. In view of my conclusion on the dominant purpose analysis, the outcome of these appeals does not depend on this contention. However, if it were necessary to decide the point, I would conclude that the fact that the precise amount of a tax benefit has to be calculated once the Court has determined the relevant counterfactual to be used, does not mean that the taxpayer, for that reason alone, has succeeded in showing that the assessments are excessive and its appeals against the objection decisions should be allowed in full. The Full Court's decision in Trail Bros (upholding the primary judge's decision below) supports this conclusion. 399 In Trail Bros, the Commissioner contended that the relevant tax benefit (in both of the tax years in dispute) comprised a $210,000 deduction paid by the taxpayer to a "Welfare Fund" established for the benefit of two persons employed by the taxpayer. Those deductions were claimed in circumstances where: (a) the employees' contracts of employment previously provided for superannuation contributions to be made on their behalf to a superannuation fund; and (b) in response to certain legislative changes limiting deductibility of superannuation contributions under those employment contracts to annual age-based deduction limits for each employee, an amount of $210,000 was (instead of being contributed to a superannuation fund) paid to a "Welfare Fund" established on behalf of the employees, in both tax years in dispute. 400 The trial judge held that the taxpayer had obtained a tax benefit in each year, being the difference between the $210,000 and the superannuation age-based deduction limits for that year (cf the whole of the $210,000 in each year, contended for by the Commissioner): Federal Commissioner of Taxation v Trail Bros Steel & Plastics Pty Ltd (2009) 75 ATR 916; [2009] FCA 1210 at [61] (Greenwood J). That finding was upheld by the Full Court: Trail Bros at [53]-[54] (Dowsett and Gordon JJ), [62] and [65] (Edmonds J). No attempt was made by either the trial judge or the Full Court to quantify that tax benefit as found.

[61]

The proper approach on the authorities 404 Close attention must be paid to the exercise required by s 177D. It does not require the Court to determine the actual or subjective motives of any of the persons who "entered into or carried out the scheme or any part of it": Hart at [65] (Gummow and Hayne JJ). Rather, it requires the Court to assess whether, having regard to the eight matters listed in s 177D(b) "it would be concluded that" any of those persons so acted. The inquiry is objective: Hart at [37] (Gummow and Hayne JJ). What is determined is the "purposes to be attributed to [the] relevant persons who entered into or carried out the scheme or any part of the scheme": Hart at [63] (Gummow and Hayne JJ). 405 The distinction between the objective determination of why a taxpayer (or other relevant person) actually acted as it did, and the question posed by s 177D may be subtle, but it is real. As Gummow and Hayne JJ explained in Hart at [65], s 177D is not directed to determining why such persons acted as they did; rather, it is directed at the drawing of a conclusion about purpose from the eight identified matters. This distinction is helpfully exposed by the reasons of the plurality in Federal Commissioner of Taxation v Spotless Services Ltd (1996) 186 CLR 404 (Spotless Services). Their Honours, having noted that the eight matters are "posited as objective facts" (citing Peabody at 382), said (at 422) that the phrase "it would be concluded that": also indicates that the conclusion reached, having regard to the matters in par (b), as to the dominant purpose of a person or one of the persons who entered into or carried out the scheme or any part thereof, is the conclusion of a reasonable person. In the present case, the question is whether, having regard, as objective facts, to the matters answering the description in par (b), a reasonable person would conclude that the taxpayers entered into or carried out the scheme for the dominant purpose of enabling the taxpayers to obtain a tax benefit in connection with the scheme. (Emphasis added.) 406 To similar effect, Hespe J (Perry and Derrington JJ agreeing) observed (in relation to the current s 177D) in Federal Commissioner of Taxation v Guardian AIT Pty Ltd ATF Australian Investment Trust (2023) 115 ATR 316; [2023] FCAFC 3 (Guardian) at [180]: The s 177D inquiry is not concerned with the actual purpose of a party to a scheme. It requires a conclusion to be drawn about an objectively ascertained intention: News Australia Holdings [2010] FCAFC 78 [30] (Stone, Jessup and Jagot JJ). 407 Given the nature of the enquiry required by s 177D, contemporaneous documents are typically the best guide to the events that occurred which, often in Pt IVA cases, took place many years before the matter comes to trial: see, eg, Macquarie Bank Ltd v Federal Commissioner of Taxation (2011) 85 ATR 409; [2011] FCA 1076 at [41] (Edmonds J). 408 The objective purpose is generally to be determined when the scheme was entered into: Ashwick at [141] (Edmonds J, Bennett and Middleton JJ agreeing), citing CPH Property Pty Ltd v Federal Commissioner of Taxation (1998) 88 FCR 21 at 42 (Hill J). However, it has been recognised that some of the s 177D factors lend themselves to assessment at other points in time, after the initial entry into the scheme. In Guardian, Hespe J (Perry and Derrington JJ agreeing) observed, of the time at which dominant purpose was to be tested (at [182]): The conclusion as to purpose is required to be drawn from the application of each of the eight factors referred to in s 177D. Each of those factors is to be applied according to their respective terms. Some of those factors refer to when the scheme was entered into or carried out and some refer to the consequences of the scheme. 409 See also Federal Commissioner of Taxation v Sleight (2004) 136 FCR 211; [2004] FCAFC 94 (Sleight) at [224], where Carr J said of the "change in financial position" s 177D factor: "In those circumstances, I consider that, on a proper construction of s 177D(b) the assessment should be made, in respect of this factor but not necessarily in respect of every factor, as at the time of entry into the scheme." 410 It is accepted on the authorities that tax is a cost and it is rational for a taxpayer to take into account total costs (including taxation costs) in deciding how to proceed: Hart at [3] (Gleeson CJ and McHugh J). It follows that, where a particular commercial transaction is chosen from a number of alternative courses of action because of the tax benefit associated with its adoption, that "does not of itself mean that there must be an affirmative answer to the question posed by s 177D": Hart at [15] (Gleeson CJ and McHugh J); Ashwick at [189] (Edmonds J, Bennett and Middleton JJ agreeing). The fact that a particular form of transaction carries a tax benefit does not mean that obtaining the tax benefit is the dominant purpose of the taxpayer in entering into the transaction: Hart at [15]. 411 Merely establishing that the taxpayer pays less tax by adopting one form of a transaction over another does not, of itself, yield a positive answer to the s 177D enquiry: Hart at [53] (Gummow and Hayne JJ). The same point was made by McHugh J, writing separately in Spotless Services (at 425), where his Honour observed that the requisite dominant purpose conclusion "will seldom, if ever, be drawn if no more appears than that a change of business or investment has produced a tax benefit for the taxpayer". 412 In Metal Manufactures Ltd v Federal Commissioner of Taxation (1999) 43 ATR 375; [1999] FCA 1712 at [261] (decision affirmed on appeal), Emmett J observed that the conclusion that the relevant scheme was entered into for the requisite sole or dominant purpose should not be drawn "if no more appears than that a taxpayer adopted one of two or more alternative courses of action, being the alternative that produces a tax benefit". While always fact-dependent, examples of such choices that have been given in the case law include the decision to rent, rather than buy, business premises (Hart at [15]), or to borrow money rather than raise capital (Macquarie Finance Ltd v Federal Commissioner of Taxation (2005) 146 FCR 77; [2005] FCAFC 205 at [213] (Hely J, French J agreeing)). 413 As the Full Court (Besanko, Colvin and Hespe JJ) recently explained in Minerva Financial Group Pty Ltd v Commissioner of Taxation [2024] FCAFC 28 (Minerva) (at [62]): The requisite dominant purpose is not to be drawn merely because, as a matter of objective fact, it is to be concluded that "but for" the tax benefit, another course of action would have been adopted. Part IVA does not require that a taxpayer choose a form of transaction which results in the most tax or more tax being payable. 414 The Full Court continued (at [65], emphasis added): Purpose directs attention to object or aim. It is concerned with the reason why something has occurred or been allowed to occur. The objective dominant purpose of a party to a scheme (such as an action or course of action) that has enabled a person to obtain a tax benefit is determined by regard to what has happened and evaluating why it has happened. Obtaining the tax benefit is not enough. Desiring the tax benefit is not enough. The obtaining of the tax benefit must have been the main object or aim of what is said to be the scheme when viewed objectively in its surrounding context. 415 Nevertheless, the fact that a transaction is entered into in pursuit of a wider commercial objective does not mean that it will necessarily fall outside s 177D: Spotless Services at 416; Hart at [16] (Gleeson CJ and McHugh J) and [64] (Gummow and Hayne JJ). It has been observed that any perceived dichotomy between a "rational commercial decision" and "the obtaining of a tax benefit as the dominant purpose" is a false dichotomy: Spotless Services at 415; Hart at [64]. 416 In Spotless Services, the plurality explained (at 416) that a particular course of conduct may be both "tax driven" and a rational commercial decision. In such circumstances the plurality explained (at 416, emphasis added) that: Much turns upon the identification, among various purposes, of that which is "dominant". In its ordinary meaning, dominant indicates that purpose which was the ruling, prevailing, or most influential purpose. 417 In some cases - the "wealth optimiser" structure in Hart was one such example - the transaction structure depends entirely on the tax benefits generated by its adoption, and the structure has no explanation other than the fiscal consequences: Hart at [18] (Gleeson CJ and McHugh J), [68] (Gummow and Hayne JJ). Similarly, in Spotless Services, the investment of excess funds with a related company in the Cook Islands at below market deposit rates only made sense due to the ability of the taxpayers to achieve a tax benefit in the form of an exemption under s 23(q) of the ITAA36 (on the basis that the interest income was exempt from income tax as it had been derived in the Cook Island, where withholding tax had been paid). As the plurality put it, without that benefit the proposal to invest in the Cook Islands as the taxpayer did would have "made no sense": Spotless Services at 422. 418 The exercise mandated by s 177D, as it has been explained in the leading cases, is consistent with the observations made in the Explanatory Memorandum (EM) introducing the Income Tax Laws Amendment Bill (No 2) 1981 (Cth), which introduced Pt IVA. The EM (which has been referred to in many cases, including Hart (at [86])) stated (at 2) that the new regime was directed at tax avoidance arrangements that "are blatant, artificial or contrived". That statement was immediately followed by reference to the objective view prescribed by the dominant purpose test. 419 The exercise mandated by s 177D also involves comparing the scheme and the alternative postulate; a conclusion about purpose from the eight matters listed in s 177D(b) requires consideration of what other possibilities existed: Hart at [66] (Gummow and Hayne JJ). The facts in Hart well illustrate the point. There, the "wealth optimiser" loan structure enabled the taxpayers to split a single loan, devoting all their repayments to the home loan portion, and accumulating more tax deductible interest on the portion of the loan directed to their investment property. There were obvious ways in which money might have been borrowed, other than the split loan structure. The availability of other means exposed that the only explanation for the "wealth optimiser" split loan structure was the tax consequences: Hart at [67]-[68] (Gummow and Hayne JJ). 420 All eight matters referred to in s 177D(b) must be considered, but in many cases, one or more will be of greater importance than others, depending on the facts of the case; some factors may be irrelevant or neutral: Hart at [58], [70] (Gummow and Hayne JJ), [92] (Callinan J). Further, it is not necessary to refer to each of the eight matters individually, where all of the matters are taken into account in forming a "global assessment of purpose": Federal Commissioner of Taxation v Consolidated Press Holdings Limited (2001) 207 CLR 235 (Consolidated Press) at [94] (Gleeson CJ, Gaudron, Gummow, Hayne and Callinan JJ).

[62]

DOMINANT PURPOSE: CONSIDERATION 421 As MAHPL observed in its closing submissions, the case was not run on the basis that the dominant purpose enquiry differed depending on whether the focus was on the primary, or the secondary/tertiary, scheme (although it continued to rely on its submissions in opening, which articulated some supplementary points in respect of the narrower (secondary/tertiary) scheme). 422 The case was also not run on the basis that there was any need to examine the conclusions that would be drawn as to the purpose of MAPL, MAHPL or Lux 1 (being the lending entity under PN A2) as distinct from the purpose of Mylan. It was not disputed that the financing and structuring arrangements were decided at the parent company level. Accordingly, I will refer to the purposes of Mylan in this part of my reasons. 423 I will address some further, general matters, before addressing the specific topics on which the parties engaged in relation to the dominant purpose enquiry. 424 The dominant purpose enquiry is an objective enquiry. It is not concerned with the subjective or actual purpose of any person or persons within Mylan or its subsidiaries. The events in question also date back to 2007, some 16 years ago, and most of the individuals involved have moved on. In these circumstances (and contrary to submissions made by the Commissioner highlighting the absence of lay witnesses of fact to explain the transaction) I do not consider that the absence of a lay witness speaking to the purposes for which aspects of each of the schemes were entered into is of significance in undertaking the analysis required by s 177D(b). As the Full Court stated in Minerva (at [68]): "The question posited by s 177D is not addressed … by testimony of a person as to their reasons for taking a particular action or step". 425 Further, MAHPL relied on the affidavit of Paul Campbell, then Mylan's Vice President Corporate Accounting and Reporting, Business Development, Strategic Development. Mr Campbell's evidence included that the Acquisition was carried out in accordance with PwC's step plan version 17. While Mr Campbell did not seek to delve into Mylan's reasons for proceeding as it did (or its reasons for not proceeding by any other available series of steps), the Commissioner could have cross-examined Mr Campbell if he regarded direct lay evidence of the processes or reasoning at the time to be important. The absence of lay evidence (or, more accurately, its limits) is not particularly telling in this case. 426 Another matter concerns the use to which stray comments in emails may be put. While the statements and documents of advisers may be probative of the purpose of the adviser, to be attributed to the taxpayer - see, eg, Consolidated Press at [95] (Gleeson CJ, Gaudron, Gummow, Hayne and Callinan JJ) - they nevertheless need to be approached with some care. The scope of the retainer of each set of advisers involved was not clear. Mylan was a highly sophisticated taxpayer with significant internal expertise. It obtained advice on a number of topics from a number of advisers, but ultimately had to make its own decisions regarding the transaction. This is not a case where the taxpayer can be assumed to have had no purpose distinct from its advisers (cf Consolidated Press at [95]), so comments in discrete internal emails of advisers cannot be elevated to impute an overarching purpose to Mylan in relation to broader structuring decisions. Finally, it appears that advisers also addressed some proposals that did not go ahead, which means that not all adviser documentation can simply be assumed to relate to the steps that were taken and which constitute the schemes. MAHPL referred, as an example of such a document, to a PwC email chain which referred to post-acquisition restructuring that did not occur, but which was relied on by the Commissioner as it contained the words "intended tax benefits". 427 A related point arises from the paucity of internal Mylan documentation - emails, memoranda and the like - in evidence. While various board and committee minutes were in evidence, they did not tend to descend into the reasons for which Mylan did, or did not do, various things. Through his audit, the Commissioner collected a great deal of documentation, including internal Mylan documents. To the extent that either the Commissioner, or MAHPL, regarded such material as probative on dominant purpose, it was in evidence. In that context, the relative lack of such documentation suggests that it did not exist, and I am not able to simply speculate about why that is the case. I do not consider that the relative lack of contemporaneous documentation expressing reasons for doing (or not doing) certain things assists the Court in determining whether either of the schemes (or part thereof) was entered into or carried out for the requisite dominant purpose. MAHPL bears the onus of establishing that neither scheme (or part thereof) was entered into for the requisite dominant purpose. The relative lack of internal documentation of the kind to which I have referred does not determine that question, one way or the other. 428 The parties both made a number of submissions on dominant purpose that, while they were not articulated by reference to any of the eight specific matters, are relevant to most of the discrete matters mentioned in s 177D(b). Many of those factors overlap. Most of those matters have been addressed in relation to "manner" (s 177D(b)(i)) below, but I have addressed the "double deduction" contention of the Commissioner, and the intersection between the Australian tax outcomes and the consequences of Mylan's OFL position in addressing the group of consequence-based matters referred to in ss 177D(b)(v)-(vii). While I have addressed specific topics on which the parties made submissions under these headings (whereas the parties advanced many of their submissions generally and not under any specific paragraph of s 177D(b)), I have done so mindful of the overlap and that many particular topics addressed are relevant to more than one of the matters referred to in ss 177D(b)(i)-(viii).

[65]

Commercial factors 433 MAHPL's submissions on dominant purpose stressed commercial considerations as the driving force behind MAPL's acquisition of Alphapharm, and the funding mix employed for that acquisition. I accept that, as MAHPL submitted, the acquisition of the Merck Generics group was an "enormous and highly geared global acquisition for the Mylan Group". 434 While not quite a case of "the minnow swallowing the whale", the Mylan group pre-acquisition was dwarfed by the scale of the Mylan group post-acquisition. As Mr Stack observed, Merck Generics' estimated revenue for 2006 (EUR 1.802 billion) was larger than Mylan's total revenues for the financial year ended 31 March 2007 (USD 1.611 billion). Alphapharm alone was larger than Matrix, which was the Mylan group's single largest acquisition prior to Merck Generics. Mylan's press release announcing the Acquisition stated that, on a pro forma basis, "for the calendar year 2006, the combined company would have had revenues of approximately [US] $4.2 billion, EBITDA of approximately [US] $1.0 billion". 435 The scale of the Acquisition, relative to the pre-acquisition Mylan group, was also reflected in the gearing of the Mylan group sharply increasing following the Acquisition. As illustrated by the figures recorded in the table at paragraph 148 above, the gearing of the Mylan group was expected to increase from approximately 50% debt pre-acquisition, to at least 81% immediately after completion. 436 MAHPL submitted that the manner in which the Acquisition was structured was appropriate and commercially expedient as it provided a flexible and straightforward means of repatriating cash from Australia. MAHPL relied on the fact that similar structures were put in place for the acquisitions of the Canadian and French subsidiaries, which were also expected to be major cash-generating units for the post-acquisition business. While these submissions were principally directed at the dominant purpose issue in relation to the primary scheme, the commercial imperatives to which MAHPL referred are also relevant as part of the background to the entry into, and carrying into effect of, the narrower scheme. The gearing level and decision to borrow internally via PN A2 are also matters that arise for consideration in relation to the wider and narrower schemes (and are addressed separately below). 437 The other commercial matters to which MAHPL referred were as follows: (a) the structure adopted (whereby an intra-group promissory note was used as the source of debt funding) provided for flexibility which is common and often preferred in multinational groups, as substantiated by the expert evidence of Mr Stack and Mr Ali, with Mr Johnson joining in agreeing with them that intercompany loans present a "flexible mechanism for multinational corporations to manage liquidity, interest rate risk and currency risk"; (b) the commercial utility of the flexibility attendant on intercompany debt was sufficiently valued by Mylan that it was prepared to wear the 10% withholding tax that arrangement attracted in Australia as a cost of doing business; and (c) the structure adopted allowed Mylan to take advantage of the US "look through" rules such that interest paid to Lux 1 would not be taxed in the US unless and until repatriated to the US or used for the benefit of Mylan. 438 As I have addressed above, in relation to the tax benefit analysis, there was substantial evidence to support the contention that repatriation of cash was a commercial objective when the scheme was entered into in October 2007. That is so notwithstanding that, as matters transpired, in 2008 there was only very limited repatriation of funds from Bermuda. The repatriation objective was also one that was facilitated by intra-group financing. Repatriation of cash by means of dividends alone, by contrast, would have been more complex (requiring resolutions at each stage up the corporate chain, and possibly also constrained by limits on when and how often dividends could be paid, and/or restrictions relating to the payment of dividends from profits). While the utility of being able to take advantage of the "look through" rule may not have been regarded as a substantial benefit given the anticipated need to repatriate cash (cf parking it in Bermuda), there is an obvious commercial attraction in establishing a structure that allows for the "look through" rule to be utilised if and when possible. Mr Stack's evidence was that it was common practice for US multinationals to establish similar structures that allowed for advantage to be taken of the look through rule at that time. The adoption of such a structure does not of itself indicate the requisite dominant purpose. 439 The flexibility attendant upon intra-group financing - about which all three financing experts agreed - is also a powerful commercial factor which means that the switch from external to internal borrowing is not a matter that itself suggests the existence of the dominant purpose to which s 177D refers. 440 While the Commissioner drew attention to the absence of contemporaneous Mylan documents speaking to the commercial benefits of the selected structure, as Mr Stack, Mr Ali and Mr Johnson agreed in their joint report, matters such as the flexibility of intra-group debt are well understood. In discussion with the Court, Mr Stack also confirmed that the utility of flexibility in financing arrangements is well known and accepted (cf being something that necessarily needs to be spelled out). In that regard, neither the absence of reference to the commercial objectives raised by MAHPL on this appeal in contemporaneous internal Mylan documentation, nor the fact that funding structures appear to be generated by Mylan's advisers (cf internally, for the commercial reasons highlighted by MAHPL in this appeal), suggest that Mylan should be understood to have proceeded as it did to obtain tax benefits for MAHPL.

[66]

Structuring of the transaction and changes to the SPA 441 The wider scheme included, as an element, "[t]he amendments to the Original SPA to include Alphapharm as a target entity and MAPL as a purchaser". This step was not an element of the narrower scheme. Nevertheless, the same substantive matter falls to be considered as context and part of the overall "manner" in which the narrower scheme was entered into or carried out. 442 The timeline by which Mylan and Merck negotiated and entered into the SPA was relatively tight, particularly for an acquisition of such magnitude. Provision was made within the terms of the SPA for the acquisition structure to be settled prior to closing. The SPA also anticipated that amendments may be made to the SPA to accommodate the buyer's final acquisition structure. In this context, I do not consider the bare fact that the SPA was amended to introduce MAPL as an additional purchaser, and Alphapharm as a specific target company, tells in favour of the requisite dominant purpose in relation to the schemes. Nor, for the record, do I consider it at all suggestive of the requisite dominant purpose for the head of a global group making a significant acquisition to establish a local holding company structure to acquire one or more local subsidiaries of the seller. 443 However, there are aspects of the changes to the acquisition structure that do have a more direct bearing on the issues falling for consideration under s 177D, and which the Commissioner emphasised. The steps that the Commissioner raised as telling were as follows: (a) Under the original SPA, Mylan was the only named purchaser. It was to acquire the shares in five subsidiaries of the Merck parent company. One of those subsidiaries was MGGBV, an entity which held most of the operating entities, including Alphapharm. The total base price (subject to adjustments) was EUR 4.9 billion. Payments were due to be made by wire transfer (cl 4.4.1). (b) In the period before closing, the SPA was amended. Under the Amended SPA, Alphapharm was added as a named target company and MAPL was included as an additional purchaser and transferee company (cll 2.1.2 and 2.3.2). The Amended SPA included additional actions to be taken "on or before the Closing Date or as otherwise indicated". Those actions included the sellers causing MGGBV to transfer all the shares in Alphapharm to MAPL in exchange for a promissory note in the attached form, which was PN Lux 1 in the amount of EUR 670 million (and adjusted to account for any amounts owed by Alphapharm to related parties at the date of the note). Accordingly, MAPL would hold the shares in Alphapharm, and MGGBV would hold PN Lux 1 (which was in the sum of EUR 670 million less adjustments, equivalent in AUD 1.04 billion) before the shares in MGGBV were transferred. After closing, MGGBV (holding as an asset, PN Lux 1) would be a subsidiary of the Mylan group of companies, so the economic value of the note PN Lux 1 would stay with the Mylan group. The transfer of the shares in Alphapharm was one of a number of steps to be taken simultaneously at closing, but after payment to the sellers of the purchase price (cl 7.3.1). The purchase price was unchanged. (c) According to step 24 of the final PwC step plan (version 17), MAPL was to exchange PN Lux 1 for the shares in Alphapharm before the acquisition of MGGBV. Mylan Bermuda being capitalised with sufficient cash to purchase the shares in MGGBV (and other targets) was step 30B. Lux 2 acquiring the shares in MGGBV was step 35A. (d) PN Lux 1 was issued by Lux 1 to MAPL in exchange for PN A1 and PN A2 (respectively 25% and 75% of the value of PN Lux 1). This step constitutes the first step in the narrower scheme. (e) PN Lux 1 was then assigned by MAPL to MGGBV in exchange for the shares in Alphapharm, in accordance with the Amended SPA. This constitutes the second step in the narrower scheme. 444 The point that the Commissioner sought to make was that the change in the structure of the Acquisition facilitated the creation of intercompany financing arrangements that (on his view) duplicated a portion of the external debt with internal debt. In short, the point was that Mylan was still paying the same price to Merck and it (and Lux 5) were still borrowing the same amount from the external lenders under the SCA, but, instead of a portion of the external debt being borne by MAPL (and so reducing the amount of external debt borne by Mylan and Lux 5), that external debt remained unchanged and an additional intercompany debt was created. The Commissioner emphasised that this arrangement did not result in any additional cash resources being made available. 445 MAHPL submitted that no adverse inference was to be drawn by reason of MAPL's acquisition of Alphapharm, and its funding for that acquisition, by way of promissory notes. It relied on the evidence of Mr Stack that, in a transaction as complex as the Merck Generics acquisition, trying to execute many steps with physical cash transfers (instead of intercompany notes) would have posed an elevated risk to closing. MAHPL also submitted that it made sense for MAPL to borrow from Lux 1 so as to take advantage of the recently introduced "look through" rules in Subpart F of the US tax code, as this enabled US taxation to be deferred to the extent that profits were not repatriated to the US. It relied on Mr Stack's evidence that US companies "commonly developed structures to enable the movement of liquidity between non-US units" without US tax consequences being taken on until funds were repatriated to the US. 446 I do not consider that features drawn out by the Commissioner and referred to above suggest a dominant purpose within the terms of s 177D. It would have been administratively simpler for the external debt to be assumed by Mylan and Lux 5, without having each acquiring subsidiary assume a portion of the debt. Of course, the preferred counterfactual involves MAPL taking on external debt, but that is not to gainsay that leaving the overall external debt in place at the Mylan and Lux 5 level was simpler. The inefficiency and complexity that would be associated with introducing an Australian subsidiary as a separate borrower under the SCA was explained in evidence by Mr Stack. With the external debt at the group parent and treasury level, the next step involves some of that debt being pushed down the corporate chain. Subject to matters of method and quantum, I do not consider that there is anything that bespeaks the requisite dominant purpose in a corporate group parent such as Mylan distributing debt to the local holding company level. The commercial rationales for not equity funding all subsidiary holding companies have been addressed above. 447 Nor, again subject to questions of quantum and method (by which I include the terms of the borrowing), is there anything in debt being taken on by local holding companies (relevantly here MAPL) by way of intercompany loans, established by promissory notes, that necessarily bespeaks the dominant purpose to which s 177D refers. The commercial rationales for financing subsidiaries by intercompany loans have also been addressed above. The mechanism by which PN Lux 1 was issued in Euro, in return for PN A1 and PN A2, denominated in AUD, was explained in the PwC step plan as serving two purposes: eliminating foreign exchange risk on the Euro denominated note that would be issued to MGGBV, and facilitating the acquisition by MAPL of Alphapharm. There is a comprehensible commercial rationale for the exchange of notes, when considered within the selected transaction structure, whereby the shares in Alphapharm would be transferred out of MGGBV before MGGBV was transferred. 448 I also note that having MAPL take on debt by way of intercompany loan was, in the circumstances of the Mylan group, a more expensive option as 10% withholding tax had to be paid in circumstances where foreign tax credits were not expected to be able to be recouped due to Mylan's OFL position. As such (subject to questions of method), the switch from the plan for MAPL to borrow externally to having it borrow internally (by promissory note) is not suggestive of the dominant purpose to which s 177D refers. 449 In the opening stages of the trial, the Commissioner made submissions to the effect that the structure of the Amended SPA enabled Mylan to nominate the price it wished to pay for Alphapharm. However, as MAHPL established that the price was set within the SPA - on the basis that PN Lux 1 was referred to in the SPA - the Commissioner's earlier submission fell away. 450 These matters do not, however, fully explain what occurred here. As set out above, the headline "price" for the Merck acquisition was left unchanged, notwithstanding that, under the Amended SPA, the shares in Alphapharm would be taken out of the Merck group just prior to the transfer of the shares in MGGBV. PN Lux 1 would be transferred to MGGBV, only to remain with that entity following closing. 451 Once a decision was taken to have Alphapharm (and some other local subsidiaries including Merck Canada and Merck France) held by local country Mylan holding companies, and where a great many local operating subsidiaries were held by MGGBV, adopting a transaction structure that separated out, and transferred, various of MGGBV's operating subsidiaries before the "rump" (MGGBV with all remaining subsidiaries) was transferred has an appreciable commercial rationale. It is not a step that can be explained only by reference to achieving tax benefits. 452 For these reasons, I do not accept the Commissioner's arguments that there was a duplication of debt that tends to suggest the requisite dominant purpose. Similarly, I do not regard the amendments to the original SPA to introduce Alphapharm as a target and MAPL as a purchaser, or the pricing features referred to above, suggest the requisite dominant purpose in relation to either scheme. 453 In this regard, it is important not to lose sight of the tax benefit in question. The tax benefit is the difference between the tax deductions claimed under the schemes and the tax deductions that would have been (or would reasonably be expected to have been) allowable on the preferred counterfactual. It is the assumption of debt by MAPL that results in some tax deductions being allowed for interest on borrowings. 454 There were more and less complex ways in which MAPL could have ended up holding the shares in Alphapharm and having debt associated with that acquisition. The quantum of tax deductions was driven by the amount of debt taken on by MAPL, and the terms on which it was taken on. 455 It is appropriate, then, to now focus on the quantum and terms of the debt assumed by MAPL.

[67]

Quantum of debt and the thin capitalisation limits 456 Three principal matters were raised in relation to the quantum of the debt assumed under PN A2. First, the absolute quantum selected and the reasons for the selection of the 75% figure (which resulted in a debt to equity ratio of 3:1). Secondly, the selection of a debt percentage of the value of Alphapharm that tracked the thin capitalisation limits. Thirdly, the lack of consideration of MAPL's capacity to service debt at that level. 457 I will address these matters in that order. 458 The Commissioner emphasised that the 75% debt figure could not be explained on the basis that Mylan established structures that would maximise the repatriation of funds for purposes of repaying external debt. This was said to be clear from the fact that Mylan still planned for MAPL to borrow 75% of the value of Alphapharm when it was exploring structures whereby MAPL would borrow from third party funders. Obviously enough, if MAPL borrowed externally, its interest payments would not be available to help service group level debt. That may be so, but overlooks the fact that, had MAPL borrowed externally, group level debt would have been commensurately lower. 459 For its part, MAHPL highlighted that the 75% debt level mirrored, at a local level, the target post-acquisition debt profile of the group. It also relied on the expert evidence of Mr Ali that, given Mylan's equity raising was more successful than expected, it was reasonable to retain that excess capital at the group level, and so leave the original funding mix for the Alphapharm acquisition in place, rather than inject additional equity into MAPL and take its debt portion down from 75%. MAHPL also relied on Mr Stack's evidence that, once the gearing level of a subsidiary has been set, it is typically not reviewed very often. 460 There was ample evidence to support MAHPL's contention that Mylan's target level of debt related to the Acquisition was approximately 75%. The slide decks used in Mylan's presentation to Moody's and S&P on 27 and 28 September 2007 respectively, set out the expected "PF Permanent 2007E" total debt of USD 5,777.6 million and equity of USD 2,046.5 million, reflecting a ratio of 26% of equity to total capitalisation or a ratio of 3:1. Mylan's presentation to the lenders in September 2007 reflected the same expected permanent funding mix. The slide deck prepared for Mylan's presentation to the lenders in June 2007 also calculated the expected permanent funding mix based on an expected USD 2.0 billion of equity raised. It should be noted that the term "permanent" in the context of these presentations indicated the anticipated position following the raising of equity (as distinct from the position immediately following the Acquisition). 461 Mr Ali calculated that the mid-point of the planned capital raising would have put group gearing at 74.8%. I also accept Mr Ali's evidence that, where the capital raising was more successful than anticipated, it is reasonable in a large corporate group to retain the excess at the head entity level, to use flexibly, as opposed to devoting it to reducing the leverage of subsidiaries by injecting additional equity. Mr Stack's evidence also suggests that there is nothing warranting an adverse inference that arises from the failure to revisit the chosen gearing level for MAPL after Mylan's capital raising proved more successful than anticipated. 462 The 3:1 gearing ratio that Mylan implemented for MAPL was also supported by the expert evidence of Mr Stack and Mr Ali that the funding structure, and the level of debt, were not excessive from a group treasury perspective, and constituted a reasonable funding mix that was broadly consistent with Mylan's anticipated funding mix for the Acquisition as a whole, reflecting Mylan's overall risk appetite. 463 The Commissioner also referred to the fact that there was not a matching of assets and debts at the 75% ratio all the way down the corporate chain as the funds borrowed externally under the SCA were mostly deployed to capitalising Mylan Bermuda. In submissions, the figure of 4 billion (I assume USD) was referred to as the capital of Mylan Bermuda. The capitalisation of Mylan Bermuda enabled the establishment of a downstream funding structure that took advantage of the US "look through" rules as it allowed funds to be retained in Bermuda and not taxed in the US unless and until they were repatriated to the US (cf the position if Mylan Bermuda had obligations to pay interest upstream to the US had it been debt funded). In any event, and irrespective of this point, I do not regard the fact that the 75% debt to asset ratio was not reflected at every layer of the corporate chain down to MAPL is a point that undermines the economic matching of assets and debt at the MAPL level. In other words, it is not a point that supports a conclusion that Mylan had the requisite dominant purpose. 464 I turn, next, to the thin capitalisation point. 465 It was common ground that the quantum of debt that Mylan planned for MAPL to assume in connection with the acquisition of Alphapharm corresponded with the thin capitalisation limits then in place in Australia, although there was evidence that in fact there was some "head room" in the actual post-acquisition figures. While Mylan did not plan for all Merck subsidiaries to be acquired through a holding company structure that involved local holding entities taking on debt, its planning for those countries where there was to be local country debt likewise tracked the applicable thin capitalisation limits. 466 In his submissions, the Commissioner cast this tracking of the thin capitalisation limits as supportive of his position that the debt assumed by MAPL was not driven by business considerations but was driven by a desire to load MAPL up with the maximum possible debt in order to obtain deductions in Australia. For its part MAHPL cast tracking the thin capitalisation limits as a virtue, submitting that limiting borrowings to stay within the thin capitalisation safe harbour does not found an adverse conclusion under s 177D of the ITAA36. 467 The significance of setting the debt level at, or close to, the thin capitalisation limits must also be considered in light of the significant body of expert evidence concerning the benefits of debt (particularly intra-group debt) over equity, coupled with the impact of Mylan's OFL position. Both of these factors - independently and together - would, other things being equal, support a commercial decision to have MAPL take on more, rather than less, debt. As I set out below, I do not accept that seeking to avoid suffering the consequences of Mylan's substantial OFL is properly to be characterised as a strategy to reduce Australian tax. That is, as I explain below, an oversimplification. 468 In these circumstances, it is correct to characterise Mylan as having limited MAPL's debt to conform with the thin capitalisation rules then in place in Australia. I do not consider that transaction planning that saw internal local country debt track the thin capitalisation limits in place in Australia (and elsewhere) is indicative of Mylan having the dominant purpose of enabling MAHPL (being the relevant taxpayer) to obtain a tax benefit in connection with the scheme. 469 That takes me to the point about serviceability. 470 The Commissioner highlighted the lack of evidence of any contemporaneous assessment of MAPL's capacity to support debt at the level it took on. MAHPL had two responses to this line of attack. First, it said that, contrary to the Commissioner's submissions, there was evidence that Mylan's advisers did, in fact, perform an assessment of each country's interest capacity based on projected operating profits. Secondly, MAHPL submitted that the Commissioner's emphasis on the asserted lack of analysis of debt servicing capacity was misdirected where the financing arranged was intra-group (cf third party) lending. 471 In my view, the second response has substance, but the first response fails on the evidence. 472 MAHPL relied, in support of the first contention, on an internal PwC email dated 14 July 2007. That email was addressed "Team Mylan" and stated as follows: "Attached below is a projection by country of Merck's operating profits from 2007 through 2010. This should help in assessing each country's interest capacity." The email attached a document with a footer date reference of "20.02.2007" and a title date reference of "Apr 03, 2007". The document stated that it was "according to Genius forecast figures", which in context I take to be indicating that Merck's projections were being used. 473 The document set out, for the years in question, figures including profit before tax (anticipated and adjusted), any carry forward losses (and amounts used), as well as provisional income tax (figures and percentage rates). The document itself does not reveal any analysis of debt carrying capacity. While calculating the anticipated raw figure of profit before tax may well be useful to someone wishing to carry out an analysis of debt servicing capacity, the mere statement of that figure in respect of Alphapharm does not, in my view, constitute an assessment of the level of debt MAPL could bear. Besides anything else, arriving at a figure for the quantum of debt an entity can sustain would require assumptions to be made about the terms of the borrowing (interest rate, whether interest could be capitalised, the repayment schedule etc) but the document in question did not address these matters. Nor am I willing to assume that PwC carried out an actual debt servicing analysis in documents that were not in evidence just because this email indicated that the attachment would be helpful in assessing each country's interest carrying capacity. Nor do I accept, however, that this email shows an approach by which debt levels were set to ensure that profits were, to the maximum extent possible, absorbed by deductions, as the Commissioner suggested. 474 Returning to the second submission advanced by MAHPL, it is, in my view, important to keep firmly in mind that the scheme involved an intra-group promissory note on terms that were very flexible. Under PN A2, interest could be capitalised and the principal need not be repaid until maturity (but could be repaid in any amount at any time, at MAPL's election). It is not surprising that Mylan, as the parent company of a group with global treasury functions, would not be concerned to closely analyse the debt carrying capacity of a holding company subsidiary such as MAPL. This is supported by the evidence of Mr Stack to the effect that Mylan would be expected to take a long-term view and would not be expected to be overly concerned by any short-term deficiency in MAPL's expected cashflows relative to its interest expenses. 475 This is not a transfer pricing case. The dominant purpose enquiry need not, and should not, treat departures from the terms that may be seen between parties dealing at arm's length, or the absence of conduct of a kind that one would expect to see in connection with third party transactions (such as specific debt servicing capacity), as ipso facto demonstrating the requisite dominant purpose.

[68]

Fixing the interest rate 476 Two related matters were raised in relation to the fixing of the interest rate of PN A2. First, the election to fix the interest rate (cf leave the rate at a floating interest rate). Secondly, the timing of when the interest rate of PN A2 was fixed was in issue. 477 The terms of PN A2 provided for the principal to be retroactively adjusted following a post-acquisition valuation to be conducted by PwC. Clause 2 of PN A2 also provided for the interest rate to be adjusted and was expressed, relevantly, in the following terms (emphasis in bold added): This promissory note shall pay interest on the principal (as adjusted with retroactive effect) at a rate of 6.25 basis points above the rate paid by Mylan Luxembourg l S.a.r.l. to Mylan Luxembourg 2 S.a.r.l. on that instrument known by such parties as "Note Lux l" as may be in effect at any time, or at such rate as may be ultimately determined by the Luxembourg Taxing Authority, provided, however, that such interest rate, as well as other material terms of this promissory note, shall be finally determined as agreed upon between the parties within 90 days of the execution of this promissory note. If no such agreement is achieved, the interest rate shall be finally determined based upon an independent transfer-pricing study of arm's length terms. 478 As may be seen, PN A2 provided for the interest rate (and other terms) to be varied within 90 days, or to be determined by a transfer pricing study. 479 I do not consider that the decision to fix the interest rate (cf retain a floating rate) suggests any purpose of maximising tax deductions for interest expenses. The expert evidence of Mr Stack and Mr Ali supported fixing the interest rate exposure of subsidiaries and handling floating rate exposure at the group level. As Mr Stack pointed out, fixed interest rates better facilitate forward planning and mitigate the risk of profit and loss volatility, and adverse outcomes from risks associated with the movement of interest rates. Mr Ali and Mr Stack also explained the ways in which fixing the interest rates of subsidiaries' financing arrangements facilitates management of interest rate risk centrally, at a group treasury level. Their evidence on these points was consistent and compelling. Mr Johnson also agreed with the thrust of much of their evidence, joining with Mr Stack and Mr Ali in the joint report, observing that "international companies commonly manage their liquidity, interest rate and currency risk in a centralised manner at a group level" and that "fixed rate instruments or borrowings eliminate the negative or positive effects arising from changes in interest rates for a borrower in respect of the term of the fixed rate". 480 I do not accept that the decision to fix the rate is a matter that, of itself, suggests the existence of the requisite dominant purpose (as distinct from the level of the interest rate at which the borrowing was fixed). 481 It is clear that the Mylan group did operate on a group treasury model, and its actions were consistent with managing the balance between fixed and floating rate exposure at that level. This also means that the various presentations MAHPL relied on which showed advisers suggesting Mylan fix higher levels of its floating rate exposure than it in fact did, are not of assistance in assessing the fixing of MAPL's intra-group interest exposure. Nor, in that context, is the decision to fix the interest rate on 100% of MAPL's debt (cf fixing lower proportions of group-level external debt) telling. 482 Having regard to these matters, I do not consider that the decision to fix the interest rate on PN A2 (as well as a number of other intra-group notes) points to any inconsistency in Mylan's risk appetite (as the Commissioner submitted), or tends to suggest the existence of a dominant purpose to which s 177D refers. While a decision to fix interest rates at a level that exceeds market rates may speak to that dominant purpose, as I go on to set out, that is not what occurred. 483 MAHPL's position was that the decision to fix the interest rate on PN A2 and to fix it at 10.15% was taken in, or by, December 2007, which was within the 90 day period following 2 October 2007. The Commissioner's position was that the interest rate was not so fixed in December 2007. He contended that the decision to vary (and fix) the interest rate of PN A2 was only revisited "at the cusp of the GFC in September 2008" when, very soon thereafter, interest rates dropped dramatically following the collapse of Lehman Brothers which triggered what we now call the "GFC". The Commissioner further contended, in his written closing submissions, that the fixed rate of 10.15% was not "agreed-in-principle until October 2008". 484 There was some confusion about the significance of a document titled "Amendment to Note A2". That document was signed by MAPL and Lux 1 on 8 January 2010. Omitting recitals, the document stated that the parties agreed to amend PN A2 as follows: The interest rate provided for in Section 2, with retroactive effect to the Effective Date [2 October 2007], shall be a fixed rate of 10.15% for the entire term of the Note; and The principal amount of the Note, as determined in accordance with 1, and with retroactive effect to the Effective Date, shall be AUD 923,205,336. The remaining terms of the Note shall remain in effect; nothing in this Amendment shall preclude the parties from subsequent modifications of the Note in accordance with its terms. 485 That document was characterised by MAHPL as merely "housekeeping" to formalise and clarify actions that had already been taken. In oral opening, the Commissioner characterised that document as having given effect to the valuation undertaken by PwC in late 2008 to 2009, supporting his contention that the interest rate was in fact set by reference to PwC's work, and had not earlier been set pursuant to cl 2 of PN A2. I proceed on the basis that the Commissioner's submission in closing - by which he accepted that the rate was agreed upon in October 2008 - has overtaken this proposition, which was advanced in opening, given the further light that was shed on this issue during the course of the trial. 486 In support of its contention that the interest rate was fixed in December 2007, MAHPL relied heavily on a spreadsheet that (according to metadata) was last modified on 14 December 2007. The spreadsheet included cells for "Principal Amount in FC [functional currency]" and "Interest Rate". In respect of PN A2, the interest rate was specified as 10.15%. Pages of the spreadsheet titled "Note Balances & Int Accruals" also included "accruals" for October to December 2007. It is not clear why the December 2007 accruals show 20 days for December for PN A2 when other notes had 31 days. 487 MAHPL submitted that this spreadsheet also illustrated that the 10.15% rate was a product of the seven year fixed swap rate (6.90%) and a "spread" of 3.25%, mirroring the spread applicable to Mylan under the SCA. However, the cells setting out the seven year fixed rates were all highlighted in yellow, indicating "final input still required". 488 This spreadsheet was attached to an email from Paul Martin of Mylan to Canadian legal counsel, copied to a number of people at PwC (all "ca.pwc" which I take to refer to PwC Canada, as well as Joe Vitullo at US PwC). Mr Martin's email attached two documents (in addition to a set of documents that it appears had been attached by the Canadian counsel in his email to Mr Martin). Those two documents were "IC Notes, Offshore Interest, Q4 2007.xls" and "Promissory Note-GenpharmCAD.doc". In his email, Mr Martin said (emphasis added): Dan, Attached is the proposed Genpharm Note, as yet unexecuted, as well as a spreadsheet to determine the values of Notes Lux10, C4, C4.1, and C5. Dan Fontaine, Joe, and I will run through the computations of these Notes' principal amounts hopefully today, as the math is something of a moving target. Regards, Paul 489 As the email shows, the spreadsheet relied on by MAHPL was provided by Mr Martin to the Canadian lawyer on the basis that it would be used to determine the value of a different series of notes. I note that the principal of note Lux 10 appears to have been set to include figures drawn from the "accruals" part of the spreadsheet. Mr Martin's email said nothing about the status of the matters recorded in relation to MAPL. It is not clear from Mr Martin's email whether the spreadsheet reflected a concluded position vis-à-vis the fixing of interest on PN A2, or merely constituted a planning or working document. The fact that the seven year fixed rate interest figures were still recorded as "final input still required" does not support a conclusion that a finalised position had been reached to fix the interest rate of PN A2 at 10.15% in December 2007. The document may well have been only a working document. References to "accruals" in this document are not conclusive evidence to the contrary. 490 That is the position notwithstanding that, as MAHPL noted, there were earlier emails that indicated an internal awareness within Mylan of the 90 day timeline for setting a different interest rate would expire at the end of December 2007. An email from Mr Martin to Brian Byala (Mylan's treasurer) and Gregory Weixel on 6 November 2007 referred to an earlier meeting where they decided they would prefer to have the intercompany debt converted to "fixed-rate equivalent[s]". Mr Martin then asked that he be provided with the fixed rate equivalents, "and associated documentation as supports such fixed rate[s] fairly soon". Mr Martin noted the 31 December 2007 deadline and the need to get local country input in the next fortnight and "then with real effort, to have proper TP documentation of the rates being arm's length in place". In another email the same day (part of the same thread), Mr Martin said: It's a matter of timing, we have agreed per the notes to true them up by December 31, 2007, and that means as a practical matter presenting the methodology for converting to fixed, and any associated documentation, to local tax counsel no later than two weeks from now, in order to have any chance of getting their timely input (and time to adjust our methodology, as need be) prior to December 21, 2007, when I suspect everyone will disappear until January 2008. 491 In the final email in the thread of 6 November 2007, Mr Weixel said he was in the "process of putting together loan agreements based on the Matrix template" but the email does not elaborate on whether these agreements referred to agreements fixing interest rates, or something else. Emails exchanged between PwC and Mylan indicate that Mylan had not yet raised transfer pricing analysis in relation to the fixed interest rate with PwC in late November 2007. Contrary to the Commissioner's submission, that does not mean that Mylan had not discussed the fixing of interest rates internally as indicated by the internal Mylan email correspondence of 6 November 2007 (referred to above). 492 I take three things from these emails. First, there was an internal decision to move the intercompany notes to fixed interest. Given the roles of those involved and the fact that they were dealing with intercompany finance, I consider it is a decision they were likely able to take without board signoff (cf the Commissioner's reliance on the absence of board minutes concerning the fixing of interest rates). As such, the decision to move to fixed rates appears to have been firm at this time, although the reasons for that were not explained. 493 Secondly, those involved recognised that it was one thing to decide to fix the rates, but that actually giving effect to that decision involved completing various steps. While there was a desire to take the first option outlined in cl 2 of PN A2, which was to set the interest rate and terms of PN A2 within 90 days (rather than have such matters fixed later by PwC's transfer pricing study), those involved also appeared to recognise that it was not at all certain that they would have the necessary paperwork and transfer pricing signoff in place in time to do that. 494 Thirdly, but importantly, the emails do not indicate any attempt to maximise the interest rates in fixing them. On the contrary, the emails referred to determining the fixed rate equivalents. An email between Mr Weixel and Mr Martin on 19 October 2007 also referred to essentially using a notional conversion of the interest rate on the group's external debt to a fixed rate and reflecting that in the intercompany notes. 495 In the period between the evidence of the last witness concluding and the parties returning for closing submissions, MAHPL filed a further affidavit of Mr Salus, the Assistant Secretary of Viatris Inc and Deputy Global General Counsel of Mylan. That affidavit exhibited a copy of the Consolidating Income Statement for the year ending 31 December 2007 of "Mylan Laboratories Inc. & Subsidiaries", which extract was time stamped 23 February 2008. This document records an interest expense of MAPL in the amount of AUD 29,953,626 and a net earnings figure of AUD 23,214,060. 496 The net earnings figure in this income statement reconciles to another document - described as a "loan schedule recording MAPL payments made under [PN A2] to [Lux 1] for the period October 2008 to October 2014". This entry, along with some others, are recorded as "AUG 2310 LOAD". Mr Salus explained that 2310 was the accounting code for MAPL, and that Mylan transitioned records from the financial system "Hyperion" to "SAP" in around August 2008. There was no record as to when that document was produced. 497 Nevertheless, it is clear that by at least 23 February 2008, the internal accounting system recorded interest expenses being incurred by MAPL that were consistent with the interest rate being fixed at 10.15% on the whole of PN A2 by that date. The fact that interest was already being accrued at the fixed rate of 10.15% by that date in February 2008 is confirmed by the terms in which PwC described the task it was undertaking in its transfer pricing analysis. Bill Yohana of PwC Australia emailed Joe Vitullo (PwC USA) on 17 October 2008 as follows (emphasis added): Thank you for taking the time to speak with me earlier today to verify my understanding of the related-party loan transactions which PricewaterhouseCoopers Australia will evaluate from an Australian transfer pricing perspective for Mylan Inc. In brief, PwC Australia will benchmark two transactions. In the first transaction, we will verify that the rate applied to a loan (note A2) from a Luxembourg entity to an Australian borrower, with a seven-year tenor and a start date of 2 October 2007, is arm's length. I understand that Mylan applied a rate of 10.15% to this note (which is denominated in AUD and is fixed rate), which is based on the prevailing seven-year AUD swap rate and a margin of approximately 3.50%, which reflects the rate at which Mylan Inc raises funds from third parties at the parental level. We have already conducted an initial analysis of this transaction and believe that we can support its arm's-length nature using this 'parental cost of funds' approach. 498 As this email makes clear, PwC understood it was conducting a transfer pricing analysis of a rate that had already been set and implemented. An earlier PwC email of 1 October 2008 also referred to their understanding that "the Australian loan is currently set at 10.15 percent" (emphasis added), although I acknowledge that an email later in that chain referred to arrangements that are proposed. 499 Mr Salus also exhibited a further document to his 20 October 2023 affidavit. That document was described as a report generated from the Hyperion system at his request, to show journal entries recorded "in December 2007" in respect of MAPL in that system. However, each of the entries showing interest expenses has a "date created" entry of 9 June 2008, as do a number of other initial establishment entries recording the initial capitalisation of MAPL by PN A1 and PN A2 (it appears this document uses American date formats). MAHPL prepared an aide memoire showing how the interest expenses shown in this document reflect interest on the principal of PN A2 at 10.15%. While I accept the mathematics of those calculations, the fact that the "date created" is in mid-2008 means that I do not accept that this document establishes that the interest rate was fixed and was being accrued in the group's internal accounts at 10.15% by and during December 2007. 500 In my view, the only date that has been firmly established for internal accrual in the accounting ledgers of the Mylan group is the Consolidating Income Statement for the year ending 31 December 2007 of "Mylan Laboratories Inc. & Subsidiaries", time stamped 23 February 2008. I consider this date to be reliable notwithstanding the ledger the Commissioner pointed to which showed records posting interest on 30 September 2008. 501 What are the consequences of this conclusion? 502 First, MAHPL has established that, whether or not arrangements were in place within 90 days of 2 October 2007 and whether or not internal arrangements would have been legally binding if tested (which is hard to see occurring in a group setting), Mylan, MAPL and Lux 1 were operating on the basis that interest on PN A2 had been fixed and was accruing at 10.15% by no later than 23 February 2008. To the extent that it was maintained, I reject the Commissioner's contention that in fact the interest rate was only fixed following, and consequent upon, PwC's transfer pricing study, or was only set in October 2008. I also do not accept that it is material whether or not the interest rate was fixed by "exercising", in a legal sense, the faculty provided by cl 2 of PN A2 by 31 December 2007. What matters for the dominant purpose analysis is when the rate was set and applied internally, even if neither of the two options set out in cl 2 of PN A2 was followed. I also note that a recognition that formal documentation will be needed at some stage does not gainsay the practical implementation of a decision that has been taken and implemented internally. 503 Secondly, this brings the time at which the interest rate was fixed on the whole of PN A2 to a time before the point (from about September 2008 onwards) where the Commissioner contended that interest rates were falling in the wake of the collapse of Lehman Brothers and the GFC starting to manifest itself. 504 Thirdly, given this point of timing, and the evidence establishing that the fixed interest rate was set to achieve equivalence with the external funding - and there being no evidence that the fixed rate was set to maximise borrowing costs (and therefore deductions in Australia) - I do not consider that the fixed interest rate of 10.15% was "over the odds", or that its selection tends to support a conclusion that this part of the scheme was entered into or carried out for the dominant purpose of obtaining a tax benefit. 505 I am fortified in this conclusion by the expert evidence of Mr Ali that the AUD equivalent fixed rate cost of an external borrowing would have been in the range of 10.72-10.86% (that is, that 10.15% was not excessive). It appears that those within Mylan who came up with the 10.15% rate did so by applying a margin above the LIBOR for the currency in question (ie AUD for PN A2). This is evidenced by the seven year fixed rate figures referred to in the spreadsheet attached to the 13 December 2007 email (to which the margin was added) varying with the currency and the explanation given by Mr Martin in his email of 6 November 2007, referred to above. While the internal email correspondence at PwC reveals a concern on the part of PwC Australia that the rate of 10.15% was aggressive and that Mylan was justifying it on the basis of what it would cost to borrow in AUD (cf the lower costs of USD or EUR), the derivation of an equivalent interest rate on the basis of applying a margin to LIBOR on an AUD borrowing was not a topic that the Commissioner explored as part of his case on dominant purpose so I will say no more about it. 506 For completeness, I note that I do not accept MAHPL's submission that, because the Commissioner abandoned his transfer pricing case, the Commissioner was precluded from making any submissions about the interest rate being excessive as part of his Pt IVA case. Entering into an intercompany loan at an excessive interest rate may be a factor that falls within several matters referred to in s 177D(b). That is so whether or not the Commissioner advances a transfer pricing case seeking to uphold amended assessments on that basis. In other words, abandoning a transfer pricing case does not, in my view, involve an abandonment of any contention that the interest rate was excessive in a Pt IVA case. However, the issue is not determinative in this case given I have concluded that the fixed interest rate of 10.15% was not excessive at the time it was entered into (by no later than 23 February 2008).

[69]

Other terms of the borrowing 507 The terms on which MAPL borrowed under PN A2 were very flexible. Interest could be capitalised, and principal could be repaid at will, without penalty, while no repayments were required prior to maturity. The evidence of Mr Stack and Mr Ali was to the effect that such flexible terms are common in intra-group lending. While Mr Johnson pointed out that such terms are not seen in arm's length borrowings, MAHPL submitted that his evidence was beside the point as the lending in question was internal lending, not arm's length lending. There was also some evidence, given by Mr Ali, about bonds available on the market, known as "PIK toggle" bonds, which allow the borrower to capitalise interest. 508 The Commissioner also relied on evidence suggesting that the original impetus for the inclusion of a term allowing prepayment of principal without penalty was to ensure sufficient flexibility to stay within the thin capitalisation limits when the principal was to be retrospectively fixed following the post-acquisition valuation. That may be so, but, for reasons already canvassed, I do not consider a desire to stay within the thin capitalisation limits (and, by extension, the adoption of terms to ensure that could be done) is indicative of the dominant purpose to which s 177D refers. 509 Nor do I consider that the inclusion of terms permitting the interest rate and principal to be retroactively set indicate such a dominant purpose on the facts of this case. 510 The mechanism to adjust the principal simply reflects a decision to ensure the debt financing of the Alphapharm acquisition remained at 75% of its value, when the final valuation of the component parts of the overall Merck Generics acquisition was to be undertaken post-acquisition. There is no separate point here, over and above the initial selection of the 75% debt level, which has been addressed above. 511 As to the provisions to subsequently re-fix the interest rate, while I accept that such terms would not be seen in arm's length borrowings, that is not to the point. Here, the Acquisition was undertaken in a relatively compressed timeframe, and Mylan - in an intra-group financing setting - deferred the making of final decisions on a number of matters. The evidence referred to above in connection with fixing the interest rate also shows a mindfulness that the selected interest rate needed to have support from a transfer pricing point of view. This is also consistent with the terms of PN A2 providing for the interest rate to be set by a transfer pricing study if not already determined within 90 days. 512 For these reasons, I do not consider that the flexibility of the terms on which MAPL borrowed tends to suggest the existence of a dominant purpose of the kind referred to in s 177D.

[70]

Failure to refinance 513 Under the terms of PN A2, MAPL could prepay the principal at any time without penalty. It retained that capacity even after the principal was retrospectively fixed (following the post-acquisition valuation by PwC) and after the interest rate was fixed at 10.15%. Accordingly, and as the Commissioner emphasised, MAPL had the ability to refinance as interest rates reduced as the GFC took hold. The Commissioner submitted that Mylan's failure to cause MAPL to refinance, or renegotiate its interest obligation on PN A2, when economic circumstances changed indicates that the scheme was carried out in a manner so as to maximise MAPL's interest outgoings. 514 MAPL did not refinance PN A2 until September 2014, when it refinanced the outstanding balance (AUD 436,504,514) through Lux 2 subscribing for additional shares in MAHPL and MAHPL subscribing for additional shares in MAPL, with the balance after the share subscriptions (AUD 274,194,206) accounted for by PN A4, which was issued by MAPL to Lux 2 at 5.073%. MAPL was, with those transactions, then able to repay PN A2 to Lux 1. The refinancing of PN A2 was the final step in both the wider and narrower schemes. However, the Commissioner did not advance any particular criticisms of, or submissions concerning, that refinancing. 515 It was common ground, and Mr Johnson opined, that interest rates declined significantly following the collapse of Lehman Brothers in September 2008 which marked the beginning of the GFC. MAHPL itself emphasised that Mylan did not have a "crystal ball" so as to foresee this decline. 516 Nevertheless, MAHPL's submissions did not adequately grapple with the point raised by the Commissioner as to MAPL's failure to refinance (prior to September 2014) to take advantage of falling interest rates. MAHPL advanced three points on this: first, that a failure to renegotiate the interest rate is a "matter of commercial judgment that is not the concern of Pt IVA"; secondly, that MAPL had no unilateral power of amendment under PN A2; and thirdly, that as a failure to refinance is an event that never occurred, the failure to refinance can have no relevance to "manner" under s 177D(b)(i). None of those arguments holds water. A matter is not insulated from, and put beyond the reach of consideration under, Pt IVA simply because it is a "commercial judgment". Further, while MAPL did not have a unilateral right of amendment under PN A2, it was able to pay out the note early if it obtained finance from elsewhere, and could have at least sought to renegotiate the rate on the note. I also do not accept that the examination of "manner" in relation to a scheme cannot include considering steps that could have been, but were not, taken. It is clear from Hart at [66] (Gummow and Hayne JJ), that the dominant purpose analysis includes reference to alternative courses of action that were available. 517 In my view, MAPL's failure to refinance PN A2 to take advantage of lower interest rates, or to use declining interest rates as an occasion on which to renegotiate the interest rate attached to PN A2 does suggest an indifference on the part of MAPL, and Mylan, to the level of interest expenses being incurred by MAPL. That suggests that the level of interest expenses being incurred by MAPL was not regarded as something that should be minimised wherever possible. To an extent, this does point to the tax deductions for interest expenses being welcome. 518 However, it must be recalled that, as the Full Court said in Minerva at [65], obtaining the tax benefit is not enough, and nor is desiring the tax benefit. In addition, to the extent that the failure to refinance indicates that the deductions for interest expenses were welcome, this matter cannot be considered in isolation from other factors. With intra-group finance, Mylan operated a group treasury. Mylan's willingness to leave in place a situation whereby one subsidiary continued to pay interest to another subsidiary at a fixed rate set prior to market declines in interest rates would likely have been informed by that group perspective. However, as interest was mostly capitalised, whatever group perspectives were at play, the continuation of PN A2 without any attempt at refinancing or reviewing the interest rate cannot be explained on the basis that the group perspective prioritised obtaining additional liquidity by leaving PN A2 in place. 519 In my view, this matter does tend to support the existence of the requisite dominant purpose in relation to how the scheme was carried out (cf the purpose to be attributed to entry into the scheme).

[71]

Other alternatives 520 In Federal Commissioner of Taxation v Macquarie Bank Ltd (2013) 210 FCR 164; [2013] FCAFC 13 (Macquarie Bank), Middleton and Robertson JJ explained the place of analysis of alternatives as follows at [210]-[211]: Despite the fact that s 177D does not expressly refer to possibilities other than the scheme that was implemented, the High Court in both Spotless Services (1996) 186 CLR 404 and Hart considered it necessary to invoke this concept for the purpose of the s 177D analysis. In Spotless Services, the Court acknowledged the "other alternatives which had been under consideration by Spotless Services" (at 418) and the "[v]arious courses of action" that were considered before the relevant scheme was carried out (at 422) in the course of setting out the events and circumstances "to which regard may be had for the purposes of pars (i) and (ii) of s 177D(b))" (at 420). It was in part by reference to these other possibilities that the Court was ultimately able to conclude that, under s 177D, the dominant purpose of the taxpayers was to obtain the tax benefit in question, as "[w]ithout that benefit, the proposal would have 'made no sense'" (at 422). Similarly, in Hart, Gummow and Hayne JJ confirmed in relation to their analysis of s 177D(b)(i) that "[t]he conclusions as just described, as being indicated by the manner in which the scheme was entered into or carried out, are indicated by a consideration of how else the loan might have been arranged" (at 244). In light of this authority, it is clear that, where appropriate, regard may be had to the other possibilities that existed for the purpose of conducting the s 177D analysis. This does not mean that the s 177D inquiry merely becomes a "but for" test (as was the subject of express warning in Citigroup Pty Ltd v Federal Commissioner of Taxation 81 ATR 412 at [24]; see also British American Tobacco Australia Services Ltd v Federal Commissioner of Taxation at 162). That is clearly not required - or permitted - on the face of the statute. But from a practical perspective, if the s 177D(b) analysis were to be carried out without any consideration of the other possibilities that may have been open to the relevant taxpayer/s at the relevant time, the analysis would risk being artificial and sterile. Accordingly, we consider that reference to such other possibilities as may have existed at the relevant time is a necessary constituent of a number of the factors set out in 177D(b) (a conclusion that we consider is harmonious with the warning administered by Edmonds J in Citigroup Pty Ltd v Federal Commissioner of Taxation 81 ATR 412 set out above). 521 In his submissions concerning s 177D(b)(vi), the Commissioner addressed alternative means by which funds could have been remitted up the corporate chain by MAPL. I will address that matter here, although consideration of alternatives overlaps with s 177D(b)(vi). 522 The alternatives raised by the Commissioner were: repatriating cash by way of dividends (cf interest on debt), or by loans from Alphapharm or Lux 2 (to the extent that that entity had a central treasury role). The Commissioner submitted that the availability of other methods of repatriating cash to Mylan supports a conclusion that MAPL's borrowings were not necessary to facilitate the repatriation of cash to Mylan. The Commissioner also submitted that the cash repatriation rationale was undermined by the fact that Mylan Canada and Mylan Japan were not capitalised at a 3:1 gearing ratio (but instead were capitalised to stay within the thin capitalisation limits). 523 The alternatives raised by the Commissioner (repatriation of cash by dividends and repatriation by intra-group loans) are alternatives which assume either that MAPL did not exist, or that (if it existed) MAPL did not have debt funding, but was capitalised wholly with equity. The Commissioner's alternatives raise how else, on those assumptions, cash might have been repatriated to Mylan. This reflects a point of contention between the parties concerning the inflexibility of equity funding and its impact on cash repatriation intentions. That debate loomed large in the case, particularly in relation to identifying the counterfactual in the tax benefit analysis. 524 As I have explained in rejecting the primary counterfactual, funding the acquisition of a foreign operating subsidiary wholly with equity and relying on dividends to send profits up the corporate chain involves complications and potential restrictions. In addition, and as already addressed, there are disadvantages (arising from inflexibility) in wholly equity-funding subsidiaries, whereas there are clear advantages (principally in terms of flexibility) in using at least some intercompany debt to capitalise a subsidiary. Consistently with that analysis, I do not regard the availability of alternative means of repatriating cash to support an adverse conclusion on dominant purpose. 525 Looking at alternatives to the narrower scheme - and so considering alternatives which necessarily involve MAPL being capitalised with a mixture of debt and equity - the alternatives are presented by the counterfactuals identified by the parties, as well as the preferred counterfactual identified above (other than to the extent the counterfactuals adjusted the principal amount of the borrowing). In addressing the counterfactuals, I did not adjust the principal of MAPL's borrowing up (as MAHPL sought) or down (as the Commissioner sought). In considering alternative debt to equity ratios as part of the dominant purpose, while it is true that borrowing less was an "alternative" to the amount in fact borrowed by MAPL, the failure to pursue a "lesser debt" alternative does not suggest the existence of the requisite dominant purpose. 526 The selection of the form of transaction represented by the scheme, and not any of those alternatives, is not a matter that is explicable only by reference to enabling MAHPL to obtain a tax benefit in connection with the scheme: cf the "wealth optimiser" structure in Hart (at [18] (Gleeson CJ and McHugh J) and [68] (Gummow and Hayne JJ)) or the below market rate investment of excess funds with a related party in the Cook Islands in Spotless Services (at 422). The choice of related party (cf third party) funding provided obvious commercial benefits in flexibility as to capitalisation of interest and the terms concerning repayment of principal. The choice of a scheme which involved fixing the interest rate (rather than leaving a floating rate in place) does not, at the time that choice was made (as to which the evidence has been addressed above), suggest a dominant purpose of the kind referred to in s 177D as it could not then be foretold that a floating rate would, over the course of the borrowing, be more advantageous than the fixed rate. The failure to refinance PN A2 or, under Mylan's direction, re-set the interest rate is a point that I have addressed separately above.

[72]

Other points 527 It remains to address a few further, discrete points, raised by the parties. 528 MAHPL referred to the structure adopted in respect of the acquisition of other Merck operating subsidiaries in other jurisdictions in submitting that Mylan's behaviour across jurisdictions suggests that local tax considerations were not the focus. In particular, MAHPL drew attention to the fact that Mylan did not debt-fund the acquisitions in Japan and the UK - despite having identified that there might be foreign tax advantages in doing so - and also that the Canadian local holding company took on less debt in order to come within the Canadian thin capitalisation limits applicable to intra-group (cf external) debt when that limit would not have applied (or would have been higher) had the debt been external. MAHPL also referred to an observation made in an email between Deloitte personnel suggesting there be no debt push down for "Ge" (likely Germany) as it would be "[t]oo much hassle". 529 In my view, Mylan's approach to the structure of acquisitions in other jurisdictions is of limited relevance. In order to draw any conclusions of real utility, it would be necessary to have a fuller understanding of local country conditions and tax laws. Quite appropriately, the evidence in this case did not go into such matters. Nevertheless, to the extent that Mylan did adopt structures that did not pursue identified local country tax advantages elsewhere, and also proceeded with a structure that resulted in additional tax being paid in Australia (withholding tax), that provides some support for the proposition that Mylan was not generally striving to minimise overseas tax. 530 MAHPL also referred to its transparency in describing the transaction structure in providing notice under s 26 of the Foreign Acquisitions and Takeovers Act 1975 (Cth). I consider Mylan's honesty and accuracy in describing the transaction structure to be a neutral factor in relation to dominant purpose.

[74]

(ii) The form and substance of the scheme 532 In Hart, Callinan J said as follows at [88], in a passage quoted in Mills v Commissioner of Taxation (2012) 250 CLR 171 at [71] (Gageler J, with French CJ, Hayne, Kiefel and Bell JJ agreeing): The reference in s 177D(b)(ii) to the "substance of the scheme" invites attention to what in fact the taxpayer may achieve by carrying it out, that is to matters whether forming part of, or not to be found within the four corners of an agreement or an arrangement. They also require that substance rather than form be the focus. 533 Regard must be had to whether the substance of the scheme diverges from its form. In Macquarie Bank, Middleton and Robertson JJ said as follows in relation to the nature of the enquiry required by s 177D(b)(ii) (at [263]): We understand this criterion to relate to whether there are material differences between the form and substance of a scheme - one example might be where a comparison of the form and substance of a scheme reveals that despite its form, in reality, it is effectively a sham (see the comments of Emmett J in Metal Manufactures Ltd 43 ATR 375 at [289]-[290]). We consider that this criterion requires a direct evaluation of the extent to which the form of the scheme adopted matches the outcome achieved. 534 The Commissioner submitted that the schemes' form did diverge from their substance in two principal respects. First, he submitted that, in substance, the schemes involved only one economic borrowing (the borrowing by Mylan and Lux 5 under the SCA), whereas the form of the schemes involved amendments being made to the SPA to introduce the separate transfer of the shares in Alphapharm, create duplicate debt in Australia, and to do those things in circumstances where the headline price paid remained unchanged. This argument has already been considered above. In my view this point does not expose a divergence between substance and form, as both the substance and form involved debt being distributed internally while maintaining a streamlined external debt profile. 535 The second matter raised by the Commissioner was the lack of economic risk borne by MAPL. In this regard, the Commissioner pointed to the indulgence afforded by Lux 1 by reason of MAPL being able to capitalise a significant amount of interest, and the payment in fact of part of the principal and interest owing on PN A2 by further equity raising: (a) by the issuance by Lux 2 of PN A3 in favour of MAHPL on 31 December 2008 (in the value of EUR 113,267,297), which promissory note was transferred down to MAPL by MAHPL, and then used by MAPL to pay off AUD 122.8 million in interest, and AUD 105 million in principal; and (b) in September 2014, when MAPL issued further equity to raise AUD 162,310,308, to make part payment of PN A2 and then refinance the unpaid balance by issuing PN A4. 536 Having asserted that MAPL bore no real economic risk, and having referred to the facts, as they transpired, in relation to the capitalising of interest and the issue of equity, the Commissioner submitted that "[t]his indicates a dominant purpose of Mylan causing MAPL to issue PN A2 was not to create an economic debt obligation (in form) but to create an intercompany instrument which would (in substance) generate interest deductions in Australia". 537 I do not accept that submission. The stated premises simply do not sustain the sweeping conclusion. 538 A promissory note is a real economic obligation. As Gordon J observed in Noza Holdings Pty Ltd v Federal Commissioner of Taxation (2011) 82 ATR 338; [2011] FCA 46 at [163], "a promissory note is a written promise to repay a loan or debt under specific terms - usually at a stated time, through a specified series of payments, or upon demand". The form of the scheme here involved a promissory note with very flexible terms. There was no divergence between form and substance. To the extent that the Commissioner's underlying contention was as to the flexibility of the terms of PN A2, that matter has been addressed above. Further, to the extent that the Commissioner pointed to the fact that interest was in fact capitalised, I do not consider that matter tells against MAHPL in relation to the dominant purpose enquiry. As MAHPL noted in its submissions, the performance of Alphapharm post-acquisition fell far short of expectations. 539 In relation to the subsequent reduction of debt by the issuance of further equity, there was no explanation on the evidence for why Lux 1 issued a demand in late 2008 when it had no right to do so under the terms of PN A2. Nevertheless, putting aside whether Lux 1 had the right to make the demand that it issued, the likely explanation for the issue of equity to pay down interest and debt lies in remaining under the thin capitalisation limits. The Commissioner raised, in explaining this equity injection, that it put MAPL back within the thin capitalisation limits. In opening, MAHPL acknowledged that the 2014 equity injection was also likely related to a reduction in Australia's thin capitalisation limits. 540 As the injection of equity reduced MAPL's debt burden (and therefore its interest deductions), the dominant purpose in question cannot be supported by the bare fact of equity having been injected. Rather, the Commissioner seemed to be contending that what was telling was that MAPL "bore no real economic risk". While the level of risk taken on by a debtor pursuant to an intra-group loan will necessarily reflect their corporate relationship, I do not accept that the prospect of lenience by Lux 1 when MAPL's obligations would fall due supports any conclusion of dominant purpose of the kind to which s 177D refers.

[76]

(iv) The result in relation to the operation of the Act that, but for this part, would be achieved by the Scheme 544 In order for this factor to apply and support a finding as to dominant purpose that is adverse to the taxpayer, it is not enough merely to point to the fact that less tax has been paid under the form of transaction that was selected and executed: see the discussion in Hart at [53] (Gummow and Hayne JJ), Ashwick at [189]-[190] (Edmonds J, Bennett and Middleton JJ agreeing) and Macquarie Bank at [273]-[274] (Middleton and Robertson JJ); Minerva at [99] (Besanko, Colvin and Hespe JJ). The cases consistently point out that the bare fact that a taxpayer pays less tax than it otherwise would have, or obtains a particular tax consequence by virtue of the operation of Australia's taxation laws, does not, of itself warrant an inference of the requisite dominant purpose. 545 As the Commissioner submitted, the scheme resulted in MAHPL being allowed deductions in respect of interest expenses of approximately AUD 589,540,023. The Commissioner submitted that the tax benefit was significant in its quantum, but his submission focused on the fact that the large tax benefit was obtained without any entity in the Mylan group experiencing any "substantial change in [its] cash position". This position arose due to the intra-group nature of the financing arrangements and the fact that arrangements were effected by promissory notes. In particular, the Commissioner highlighted that no additional external borrowings were incurred over and above the external debt taken on by Mylan and Lux 5 under the SCA. 546 I do not accept the Commissioner's submission that these features lead to the conclusion that Mylan entered into or carried out the scheme for the dominant purpose in question. Any time a local holding company acquires an operating subsidiary and takes on debt, there will be deductions for interest. The quantum of those deductions will reflect the amount borrowed and the terms on which it is borrowed. Those matters have already been addressed above. Nor, for reasons already canvassed, do I consider that the fact that Mylan structured group financing arrangements on the basis that all external debt (SCA debt) was held at the group parent and group treasury (Lux 5) level, and distributed down to local holding companies by intercompany, and intra-group, debt suggests the dominant purpose for which the Commissioner contended. The obtaining of tax deductions for interest expenses was an ordinary incident of the financing structure adopted. The withholding tax paid by MAHPL was another ordinary taxation consequence of the internal financing.

[81]

Reduction of Australian tax as a means to address the consequences of Mylan's OFL 555 In his closing submissions, the Commissioner accepted that Mylan would have expected to have an OFL as a consequence of the Acquisition. In addition to addressing various of the s 177D(b) factors, the Commissioner's argument on dominant purpose highlighted that Mylan's OFL position meant that it could not reasonably expect to utilise any FTCs to reduce its US tax liability arising from repatriation or accrual of foreign source income, including credits for tax paid in Australia. 556 There was some discussion regarding whether the payment of US tax and foreign tax when an OFL situation precluded recourse to FTCs can be aptly described as "double taxation", as Mr Glenn had done in his report. In his oral evidence, Mr Glenn described the tax in such circumstances as "certainly incremental taxation" that would not be expected were it not for US tax rules precluding access to foreign tax credits. Prof Rosenbloom referred to the foreign tax incurred (where there is no ability to reduce US tax by foreign tax credits) as going "directly to the bottom line", and as a "direct out-of-pocket cost" which could not be credited. 557 As the Commissioner recognised, with Mylan in an OFL position, income generated in Australia would be taxed at 30% in Australia and, upon being repatriated to the US, would bear further tax at a rate between 35% to 40% (US corporate tax and state tax rates combined). The focus of the Commissioner's argument was, then, that in all the scenarios illustrated by Mr Glenn's table (which has been reproduced at paragraph 291 above), the US corporate tax on income remitted from Alphapharm to the US remains constant at 35% (USD 35 on each USD 100 of remitted income). The Commissioner sought to highlight that what changes in the various scenarios, and what drives the total tax burden and Mylan's earnings after tax, is the amount of Australian tax paid. As MAPL's interest expense increases: (a) the amount of Australian taxable income and, hence, Australian tax paid, reduces; (b) the overall combined US and foreign tax liability reduces in percentage terms; and (c) the after tax earnings of Mylan increase. 558 Although the withholding tax expense goes up the more interest MAPL pays, the combined income and withholding tax burden goes down. This is illustrated by the fact that, in Mr Glenn's alternative two, the total Australian income tax and withholding tax is USD 20 whereas, when the amount of assumed debt (and therefore interest expense) is increased in alternative three, the total Australian income tax and withholding tax is USD 15. 559 Against that background, the Commissioner accepted that it was economically rational for Mylan to seek to reduce its overall tax burden (given the inability to claim FTCs) but submitted that "[t]he reduction of Australian tax [by incorporating MAPL and having it borrow funds] was a response to Mylan being in an OFL position, which created the motivation to reduce non-US tax on foreign source income repatriated to Mylan". In effect, the Commissioner's submission was that the purpose of entry into the schemes was to address the consequences of Mylan's OFL position by reducing the amount of tax paid in Australia (the rationale being, the less tax paid, the less the impact of an inability to claim FTCs). 560 In my view, it is artificial to seek to characterise avoiding the impact of the OFL as avoiding Australian tax. As Mr Glenn's evidence showed, (other things being equal) a taxpayer in Mylan's position would be indifferent to the quantum of debt to equity of a subsidiary such as Alphapharm, provided it could obtain full FTCs for foreign tax paid. 561 The net taxes paid by Mylan after foreign tax credits have been applied is a function of whether or not Mylan could claim foreign tax credits. That is a function of the US revenue laws, not Australian tax laws. The Commissioner's analysis on this issue seeks - artificially in my view - to focus only on the line of Mr Glenn's table that sets out the US tax liability of Mylan before the application of foreign tax credits, and to treat differences in the "After Tax Earnings of Mylan" lines as driven only by the non-payment of Australian taxes. In my view, this ignores the fact that the net US tax position is a function of whether or not, as a matter of US revenue law, foreign tax credits can be claimed for the Australian taxes, which vary with each of Mr Glenn's alternatives.

[83]

(viii) The nature of any connection (whether of a business, family or other nature) between the relevant taxpayer and any person referred to in subparagraph (vi) 564 MAHPL is a wholly owned subsidiary of the Mylan group, as is MAPL, Lux 1 and Lux 2. The only submissions made by the Commissioner on this factor: (a) referred to his submission, made in connection with s 177D(b)(ii), that the scheme involved an internal reorganisation of no real economic substance; and (b) referred to and reinforced his submission that the absence of contemporaneous analysis of debt servicing capacity is significant in the context of the related party relationships. 565 I have addressed elsewhere the matter of the lack of contemporaneous analysis of MAPL's capacity to service debt. 566 The Commissioner's submissions on the second point included the contention that (emphasis added): the relationship between Mylan Inc. and MAPL enabled the parties to collude to produce an outcome whereby interest deductions could be claimed in Australia and the US, with no corresponding amount of interest income being included in Mylan Inc.'s US tax base. 567 This submission asserts a conclusion, not a reasoned basis for a conclusion. 568 Although MAHPL submitted that this factor was "at best, neutral", the related party relationships between MAHPL and the other Mylan group companies involved in the scheme explains much. MAHPL has not shied away from that point, emphasising throughout that aspects of the Commissioner's analysis and expert evidence of Mr Johnson proceeded from the perspective that arm's length dealings were still to be expected, and departures from outcomes that might be expected between arm's length parties were, ipso facto, colourable. 569 The related party relationships explain why the schemes involved MAPL being funded by intercompany promissory notes, as well as the flexible terms of PN A2 (which were utilised by capitalisation of interest and apparently ad hoc repayments of principal that were connected with the issue of further equity). The related party feature also explains why the focus was on group financing arrangements, without the kind of close attention to debt servicing capacity that might be expected where a standalone company enters into a loan with a third party lender. It also explains the final refinancing of the balance of PN A2 with another intercompany note, PN A4. 570 While the related party status of the protagonists looms large in explaining the shape of the schemes, it does not do so in a way that points to the requisite dominant purpose. As I have addressed above, what generated the deductions was MAPL having interest-bearing debt. The related party status of MAPL, Lux 1 and Mylan explains the way in which that debt was structured, but the tax deductions were driven by the quantum of, and interest rate on, MAPL's debt. I am satisfied, based on the evidence referred to above, that the initial fixing of the interest rate at 10.15% occurred by no later than 23 February 2008 and that the rate was not driven by a desire to maximise deductions. 571 However, and as also set out above in addressing s 177D(b)(i), the failure to consider refinancing PN A2, or resetting the interest rate on PN A2, following marked declines in rates is explicable by reference to the related party status of the companies involved. That inaction does indicate an indifference to the cost to MAPL of interest on its finance which tends to support the existence of the requisite dominant purpose in relation to an aspect of how the scheme was carried out (cf the purpose to be attributed to entry into the scheme).

[87]

REMITTER AND FURTHER DETERMINATIONS 579 On 21 April 2021, the Commissioner, by his delegate, made determinations under s 177F(1)(b) of the ITAA36 for the income years ended 31 December 2007 to 31 December 2017 cancelling MAHPL's deductions for interest expenses on the debt under PN A2. Each of these determinations specified that a stated sum "being a tax benefit that is referable to a deduction being allowable to Mylan Australia Holding Pty Ltd" for the stated year of income "shall not be allowable to the taxpayer in relation to that year of income". 580 Nothing was said in any determination regarding the conceptual basis upon which the Commissioner had determined a tax benefit capable of being disallowed by the exercise of the Commissioner's powers under s 177F(1)(b). Nothing was required to be said about such matters (in particular, the "scheme" identified by the Commissioner, and the counterfactual that was applied in calculating the tax benefit for the determinations). Nevertheless, as it was common ground that the Commissioner's secondary and tertiary counterfactuals were only developed after MAHPL put forward alternatives to the Commissioner's primary counterfactual when the Commissioner was determining MAHPL's objections to the amended assessments, it is clear enough that the Commissioner issued those determinations having devised the primary scheme and having calculated the tax benefit by reference to the primary counterfactual. 581 In his appeal statements, the Commissioner stated that, if the Court were to find that the secondary or tertiary counterfactual was the preferable counterfactual, the Court should remit the matter to the Commissioner to make further determinations under s 177F(1)(b) to disallow so much of the deductions for interest claimed by MAHPL as exceeds the amount of interest it would have or might reasonably have incurred under those counterfactuals. 582 In response to the Commissioner having raised the prospect of further determinations being issued, MAHPL contended in its opening submissions that the Court could not remit the matter to the Commissioner to make further determinations. In advancing this argument, MAHPL drew attention to the fact that, whereas the primary counterfactual proceeded on the basis that Alphapharm would have joined the Mylan group as a result of the acquisition of its former immediate parent, MGGBV (and MAPL would never had been incorporated), the secondary and tertiary counterfactuals and associated tax benefits rest on the premise that MAPL would have been incorporated and would have acquired Alphapharm. MAHPL observed that, during the objection process, the Commissioner could have, but did not, make new or alternative determinations, addressing the narrower scheme and calculating the tax benefits by reference to the secondary and tertiary counterfactuals, in reliance on s 169A(3). 583 Returning to the secondary and tertiary counterfactuals, MAHPL submitted as follows (footnotes omitted): It is also important to note that the Commissioner's proposal that the Court remit the matter to him to make further determinations under s 177F(1) in the event that either the secondary or tertiary counterfactual is 'preferred' is contrary to the authority of this Court in Federal Commissioner of Taxation v Jackson (1990) 27 FCR 1. It follows from the decision in Jackson that prior to the introduction of s 169A(3) of the 1936 Act, determinations purporting to cancel a tax benefit had to be made prior to assessment and could not effectively be made by the Court. Section 169A(3) partially ameliorated the effect of Jackson by permitting the Commissioner to make determinations at the time he determined an objection, but that was not the course the Commissioner adopted in this case. 584 In oral submissions, MAHPL contended that: (a) Federal Commissioner of Taxation v Jackson (1990) 27 FCR 1 (Jackson) is authority for the proposition that the Commissioner can only make a Pt IVA determination at or prior to the issue of an assessment and that the assessment must give effect to the determination; (b) as a result of Jackson, parliament amended s 169A and introduced s 169A(3), which gave the Commissioner the power to make fresh determinations at the objection stage, thus overcoming the problem in Jackson; and (c) the Commissioner could have, but did not, make fresh determinations at the objection stage, and is out of time to do so now. 585 The Commissioner's position was that Jackson and s 169A(3) were not relevant, and that: Any further amended assessment issued to MAHPL to reflect the tax benefit worked out by reference to the Secondary Counterfactual or Tertiary Counterfactual (or some variant thereof) would either give effect to the pre-existing s 177F determinations or, if necessary, fresh s 177F determinations made at the conclusion of the appeal. 586 MAHPL's contentions based on Jackson and the legislative history of s 169A were misconceived. Section 169A(3) provides as follows: In determining whether an assessment is correct, any determination, opinion or judgment of the Commissioner made, held or formed in connection with the consideration of an objection against the assessment shall be deemed to have been made, held or formed when the assessment was made. 587 Section 169A of the ITAA36 was introduced by s 19 of the Taxation Laws Amendment Act 1986 (Cth), and took effect on 24 June 1986, upon receiving royal assent. Sub-section (3) was part of s 169A when it was originally inserted into the ITAA36; it was not added by amendment of s 169A in response to Jackson, which was only decided at first instance on 21 April 1989, and on appeal on 31 October 1990. Section 169A was amended on two occasions after it was initially introduced into the ITAA36, but neither amendment concerned s 169A(3). 588 Further, Jackson involved a circumstance where a Deputy Commissioner of Taxation purported to make s 177F determinations while an appeal against an objection decision disallowing objections to amended assessments was already on foot. At first instance, Gummow J noted that Pt IVA had, before the making of the determinations, played no part on the processes of assessment for the two years in question and that to allow Pt IVA to intrude into the matters of which the Court was seized, when litigation was already on foot, would be to "change the nature of those matters from challenges to the decisions of the respondent some years ago to disallow the taxpayer's objections to particular amended assessments": Jackson v Federal Commissioner of Taxation (1989) 87 ALR 461 at 475. 589 Both Gummow J at first instance and Hill J (with whom Burchett and Von Doussa JJ agreed) on appeal (at 17), confirmed that determinations are to be carried into effect by the making of amended assessments, not vice versa. In explaining his Honour's view, Hill J likewise stressed (at 19-20) that the alternative view would mean that, if at the time the taxpayer elected to proceed in the Federal Court (and not the AAT), the Commissioner had not, by a determination, put Pt IVA in issue, the taxpayer's election as to forum (which is a matter of real consequence in view of the administrative functions of the AAT) would be compromised by the later introduction of Pt IVA issues; the taxpayer would irrevocably lose its right to merits review on the Pt IVA issues. 590 In my view, Jackson is not authority for propositions of the breadth stated by MAHPL. As set out, Jackson involved the making of determinations to bring in Pt IVA while litigation was on foot, when no Pt IVA issues had hitherto arisen. That is very different from this case, which has been conducted on the basis of Pt IVA issues all along, and the taxpayer and the Commissioner have fully engaged on the secondary and tertiary counterfactuals. 591 While the terms of s 169A(3) do not avail the Commissioner if it be the case that fresh determinations are required, neither side addressed any submissions to the ambit of the Court's power under s 14ZZP of the TAA, or the Commissioner's capacity to amend his existing determinations under s 14ZZQ of the TAA, although the Commissioner referred to them briefly in opening. Those provisions provide as follows: 14ZZP Order of court on objection decision Where a court hears an appeal against an objection decision under section 14ZZ, the court may make such order in relation to the decision as it thinks fit, including an order confirming or varying the decision. 14ZZQ Implementation of court order in respect of objection decision (1) When the order of the court in relation to the decision becomes final, the Commissioner must, within 60 days, take such action, including amending any assessment or determination concerned, as is necessary to give effect to the decision. (2) For the purposes of subsection (1): (a) if the order is made by the court constituted by a single Judge and no appeal is lodged against the order within the period for lodging an appeal - the order becomes final at the end of the period; and (b) if the order is made by the court constituted other than as mentioned in paragraph (a) and no application for special leave to appeal to the High Court against the order is made within the period of 30 days after the order is made - the order becomes final at the end of the period. 592 Of course, the question of the issue of further or amended determinations only arises if the Commissioner were to have prevailed in the ultimate outcome. It is not immediately apparent why s 14ZZQ would not allow the Commissioner to amend his determinations (if amendment be necessary at all) in order to give effect to this Court's decision. However, as I have concluded the dominant purpose issue in MAHPL's favour, the question of the amendment of determinations does not strictly arise and it is not necessary to express a concluded view. 593 My observations regarding s 14ZZQ ought also not be understood to suggest a concluded view that the existing determinations were not effective to support any amended assessments that may subsequently have been issued, had I concluded the dominant purpose enquiry adversely to MAHPL. As I set out below, the relationship between the determinations and the secondary and tertiary counterfactuals (and necessarily by extension, the preferred counterfactual) and the impact of Channel Pastoral Holdings Pty Ltd v Federal Commissioner of Taxation (2015) 232 FCR 162; [2015] FCAFC 57 (Channel Pastoral) on the capacity of the existing determinations to support counterfactuals other than the primary counterfactual is a matter that is of some complexity, but which does not arise for determination on this case, and was not fully argued.

[88]

CHANNEL PASTORAL ARGUMENT 594 MAHPL also argued that, contrary to the stipulation in Channel Pastoral that an assessment issued to give effect to a s 177F determination must be "consistent, in all material respects, with the postulate upon which that determination is predicated" (Channel Pastoral at [81] (Edmonds and Gordon JJ)), the assessments issued to MAHPL were not consistent with the primary counterfactual because, on that counterfactual, MAHPL and MAPL would not have been incorporated at all. This contention was elaborated upon by MAHPL in its written and oral submissions. 595 MAHPL's Channel Pastoral contention has apparent merit in relation to the primary counterfactual. However, as I have rejected the primary counterfactual, it is not necessary to say any more about it. 596 MAHPL also raised the Channel Pastoral point in relation to the secondary and tertiary counterfactuals, submitting as follows (footnote omitted): If the Secondary or Tertiary Counterfactual is adopted, it would still be the case that the determinations and assessments were not "consistent, in all material respects, with the postulate upon which that determination is predicated". The postulate upon which the determination was predicated was the Primary Counterfactual, being that MAHPL is not incorporated and would not have formed a TCG [tax consolidated group]. That postulate is inconsistent with the Secondary and Tertiary Counterfactuals which provide that MAHPL would have been incorporated and would have formed a TCG … This is not merely a matter concerning the "quantum" or "amount" of a tax benefit to which the comments made by Dowsett and Gordon JJ in Federal Commissioner of Taxation v Trail Bros Steel and Plastics Pty Ltd (2010) 186 FCR 410 in the passage cited at RS[320] were directed. Rather, the problem is more fundamental; the postulate underpinning the current s 177F determinations is materially different from that underpinning the Secondary and Tertiary Counterfactuals. 597 MAHPL did not otherwise elaborate on the Channel Pastoral argument insofar as it concerned the secondary and tertiary counterfactuals save to submit in oral closing submissions that those counterfactuals, which posited a lesser borrowing (but still a borrowing by MAPL), were not consistent with the postulate behind the determinations, which was that MAPL would never have been incorporated and there would not have been a consolidated group. 598 In view of the conclusion I have reached as to the dominant purpose enquiry and in light of the lack of detailed argument on the point, it is not necessary to reach a concluded view on the Channel Pastoral point raised by MAHPL in relation to the secondary and tertiary counterfactuals. 599 Nevertheless, I will make two observations. The first is that the basis upon which MAHPL contended that the secondary and tertiary counterfactuals were inconsistent with the determinations has a logical flaw. As MAHPL submitted, in issuing amended assessments which assumed the continued existence of MAHPL as a taxpayer and parent of MAPL, when the primary counterfactual proceeded on the basis that both MAHPL and MAPL would never have been incorporated, the determinations were ostensibly inconsistent with the primary counterfactual. But those features make that conceptual premise of the determinations consistent with the secondary and tertiary counterfactuals, not inconsistent. 600 Secondly, to the extent that the determinations were issued on the basis of calculations of the tax benefit which assumed no debt financing of the acquisition of Alphapharm, whereas the secondary and tertiary counterfactuals assumed significant debt financing, it is not obvious that that is an issue of the kind referred to in Channel Pastoral, cf being a matter of detail or calculation within the ambit of Trail Bros. 601 As the Commissioner emphasised, the issue of a determination is an administrative matter: Deputy Commissioner of Taxation (Cth) v Richard Walter Pty Ltd (1995) 183 CLR 168 at 178 (Mason CJ); Macquarie Bank at [137] (Middleton and Robertson JJ); WR Carpenter Holdings Pty Ltd v Federal Commissioner of Taxation (2007) 161 FCR 1; [2007] FCAFC 103 (WR Carpenter) at [43] and [48] (Heerey, Stone and Edmonds JJ). The authorities emphasise that the exercise of the Commissioner's discretion under s 177F(1) is not made to depend on the Commissioner's opinion or satisfaction (that is, his state of mind) as to whether there is a tax benefit and, if so, whether it was obtained in connection with a scheme; they are matters of objective fact: Peabody at 382; Trail Bros at [57]-[58] (Dowsett and Gordon JJ) and WR Carpenter at [48]. Relatedly, in a Pt IVC appeal against an objection decision, the taxpayer's burden is to show that the assessment is excessive or otherwise incorrect and what the correct amount should be: s 14ZZO of the TAA. 602 Channel Pastoral does not have the effect of binding the Commissioner to the approach taken in calculating the relevant tax benefit the subject of the determination. As the Full Court (Wigney, Banks-Smith and Colvin JJ) recently stated in Singapore Telecom Australia Investments Pty Ltd v Commissioner of Taxation [2024] FCAFC 29 (at [292]): [Channel Pastoral] did not state a broad proposition that once a determination is made under s 177F then the Commissioner is bound to adhere to the reasoning upon which the determination was based for the purposes of any statutory appeal to this Court in which liability to tax is said to depend upon the determination. A proposition of that kind would be counter to well established authority that, on an appeal to this Court, the Commissioner does not need to demonstrate the correctness of the analysis by which the disputed assessment was made. 603 The determinations themselves disallow deductions in the stated amounts. The determinations do not effectively incorporate by reference the detail of the analytical path by which those amounts have been identified by the Commissioner as tax benefits liable to be disallowed in accordance with Pt IVA. While there are cases - Channel Pastoral is one such case - where the assessment has no coherent relationship with the anterior determination, I am not persuaded that this is such a case, insofar as the secondary and tertiary counterfactuals (and any variations of them) are concerned. However, for the reasons noted, it is not necessary to reach a concluded view on this point.

Parties

Applicant/Plaintiff:

Mylan Australia Holding Pty Ltd

Respondent/Defendant:

Commissioner of Taxation

Legislation Cited (8)

Measures No. 4) Act 2014(Cth)

Tax Laws Amendment (Countering Tax Avoidance and Multinational Profit Shifting) Act 2013(Cth)

Taxation Laws Amendment Act 1986(Cth)s 19

Income Tax Assessment Act 1936(Cth)ss 23, 128F, 169A, 177A, 177C, 177D, 177F

Federal Court of Australia Act 1976(Cth)ss 37M, 37N

Income Tax Assessment Act 1997(Cth)

Taxation Administration Act 1953(Cth)ss 14ZZO, 14ZZP, 14ZZQ

Foreign Acquisitions and Takeovers Act 1975(Cth)

Cases Cited (39)

205 FCR 274; [2012] FCAFC 32 Federal Commissioner of Taxation v Guardian AIT Pty Ltd ATF Australian Investment Trust (2023) 115 ATR 316; [2023] FCAFC 3 Federal Commissioner of Taxation v Hart (2004) 217 CLR 216 Federal Commissioner of Taxation v Jackson (1990) 27 FCR 1 Federal Commissioner of Taxation v Lenzo (2008) 167 FCR 255; [2008] FCAFC 50 Federal Commissioner of Taxation v Macquarie Bank Ltd (2013) 210 FCR 164; [2013] FCAFC 13 Federal Commissioner of Taxation v Sleight (2004) 136 FCR 211; [2004] FCAFC 94 Federal Commissioner of Taxation v Spotless Services Ltd (1996) 186 CLR 404 Federal Commissioner of Taxation v Trail Bros Steel & Plastics Pty Ltd (2010) 186 FCR 410; [2010] FCAFC 94 Federal Commissioner of Taxation v Trail Bros Steel & Plastics Pty Ltd (2009) 75 ATR 916; [2009] FCA 1210 Futuris Corporation Ltd v Federal Commissioner of Taxation (2010) ATC ¶20-206 Jackson v Federal Commissioner of Taxation (1989) 87 ALR 461 Macquarie Bank Ltd v Federal Commissioner of Taxation (2011) 85 ATR 409; [2011] FCA 1076 Macquarie Finance Ltd v Federal Commissioner of Taxation (2005) 146 FCR 77; [2005] FCAFC 205 Metal Manufactures Ltd v Federal Commissioner of Taxation (1999) 43 ATR 375; [1999] FCA 1712 Mills v Commissioner of Taxation (2012) 250 CLR 171 Minerva Financial Group Pty Ltd v Commissioner of Taxation [2024] FCAFC 28 Noza Holdings Pty Ltd v Federal Commissioner of Taxation (2011) 82 ATR 338; [2011] FCA 46 RCI Pty Ltd v Federal Commissioner of Taxation (2011) 84 ATR 785; [2011] FCAFC 104 Singapore Telecom Australia Investments Pty Ltd v Commissioner of Taxation [2024] FCAFC 29 WR Carpenter Holdings Pty Ltd v Federal Commissioner of Taxation (2007) 161 FCR 1; [2007] FCAFC 103 Commissioner of Internal Revenue v Brown (1965) 380 US 536

(2015) 232 FCR 162

(1994) 181 CLR 359