Frucor Suntory New Zealand Limited v Commissioner of Inland Revenue

IN THE SUPREME COURT OF NEW ZEALAND

I TE KŌTI MANA NUI O AOTEAROA

SC 81/2020
[2022] NZSC 113

BETWEEN	FRUCOR SUNTORY NEW ZEALAND LIMITED Appellant
AND	COMMISSIONER OF INLAND REVENUE Respondent

SC 92/2020
BETWEEN	COMMISSIONER OF INLAND REVENUE Appellant
AND	FRUCOR SUNTORY NEW ZEALAND LIMITED Respondent

Hearing:	8–10 June 2021
Court:	Winkelmann CJ, William Young, Glazebrook, O’Regan and Ellen France JJ
Counsel:	L McKay, M McKay and H C Roberts for Frucor Suntory New Zealand Ltd J B M Smith KC, E J Norris and L K Worthing for Commissioner of Inland Revenue
Judgment:	30 September 2022

JUDGMENT OF THE COURT

AThe appeal is dismissed.
BThe cross-appeal is allowed with the result that the appellant’s challenge to shortfall penalties is dismissed.

CThe appellant must pay the respondent costs of $45,000 plus usual disbursements.

____________________________________________________________________

REASONS

Winkelmann CJ, William Young, O’Regan and Ellen France JJ	[1]
Glazebrook J	[144]

WINKELMANN CJ, WILLIAM YOUNG, O’REGAN AND ELLEN FRANCE JJ

(Given by William Young J)

Table of Contents

Para No.
What this case is about	[1]
The funding arrangement	[13]
The background to the funding arrangement	[13]
The planning documents—prior to DHNZ acquiring FBGL	[16]
The planning documents—after DHNZ acquired FBGL	[22]
Tax treatment in Singapore	[34]
The components of the funding arrangement	[36]
Explanatory documents written after execution	[39]
Accounting treatment	[42]
Steps at maturity and subsequent return of capital	[47]
Tax avoidance—the legal framework	[49]
The High Court and Court of Appeal judgments as to the application of s BG 1	[59]
The High Court judgment	[59]
The Court of Appeal judgment	[65]
The application of s BG 1(1) to the funding arrangement: our assessment	[68]
Reconstruction	[88]
Shortfall penalties	[94]
The legislative scheme	[94]
The approaches of the High Court and Court of Appeal	[108]
Frucor Suntory’s argument on the application of the “about as likely as not to be correct” standard	[113]
The relevant authorities	[115]
Were the tax positions of DHNZ “about as likely as not to be correct” when they were adopted?	[128]
The significance of the alternative assessments	[135]
Conclusion as to unacceptable position	[136]
Was the dominant purpose of the arrangement to avoid tax?	[137]
A brief response to the dissent of Glazebrook J	[139]
Disposition	[143]

What this case is about

On 17 March 2003 Deutsche Bank advanced $204 million[1] to Danone Holdings NZ Ltd (DHNZ) in exchange for a convertible note (the note) redeemable at maturity in five years’ time at Deutsche Bank’s election by the issue of 1,025 non‑voting shares in DHNZ. Interest on the advance (often referred to in the documents as “coupons”) was payable semi‑annually in arrears at 6.5 per cent per annum. Over the five-year duration of the note, this amounted to $66 million and was paid by DHNZ to Deutsche Bank.
[1]Except where otherwise stated in this judgment, amounts have been rounded for ease of expression.
DHNZ claimed deductions in respect of the interest payments, a treatment that, as will be apparent, follows the form of the transaction as between it and Deutsche Bank. That transaction, however, was a component of a broader funding arrangement under which:
(a)Deutsche Bank would, at maturity, elect to take shares in DHNZ and in this way extinguish the liability of $204 million; and
(b)Danone Asia Pte Ltd (DAP) (which owned DHNZ) paid Deutsche Bank $149 million on inception of the funding arrangement to acquire these shares at maturity (the forward purchase agreement).
The position of the Commissioner of Inland Revenue is that the net economic effect of that funding arrangement was that:
(a)Deutsche Bank advanced only $55 million to DHNZ (being the difference between the $204 million advance and the $149 million paid by DAP under the forward purchase agreement); and
(b)the $66 million paid by DHNZ to Deutsche Bank amounted to repayment of that $55 million and interest on an amortising basis.
Section BG 1(1) of the (now repealed) Income Tax Act 2004 (the Act)[2] provides that a tax avoidance arrangement is void as against the Commissioner for income tax purposes. Tax avoidance is defined by s OB 1 to include “directly or indirectly altering the incidence of any income tax”. The Commissioner maintains that the funding arrangement purported to alter the incidence of DHNZ’s liability to income tax by facilitating the claim of deductions for payments which were, in substance, repayment of debt. She says that DHNZ was only entitled to deductions in respect of payments in excess of $55 million.
[2]Except where otherwise stated in this judgment, all references to sections in the Income Tax Act are references to those provisions in the Income Tax Act 2004. Despite now being repealed, they are referred to in the present tense because they are operative here.
Frucor Suntory New Zealand Ltd (Frucor Suntory) is a successor to DHNZ and is responsible for its liabilities. It is the appellant. Frucor Suntory’s position throughout has been that s BG 1(1) is not engaged and that the interest deductions DHNZ claimed were legitimate.
This appeal concerns only the deductions claimed by DHNZ for the 2006 ($10,827,606) and 2007 ($11,665,323) income years,[3] along with shortfall penalty assessments premised on the contention that DHNZ adopted unacceptable and abusive tax positions.[4] In issue are:
(a)whether s BG 1(1) was engaged;
(b)the Commissioner’s reconstruction under s GB 1(1), under which the taxable income of DHNZ was adjusted by disallowing the deductions said to have been claimed illegitimately; and
(c)as to shortfall penalties, whether the tax positions adopted by DHNZ were “unacceptable” as not meeting the “about as likely as not to be correct” standard stipulated in s 141B(1) of the Tax Administration Act 1994 and, if so, whether they were “abusive” on the basis that DHNZ had acted with the “dominant purpose” of obtaining tax advantages (s 141D).
[3]The Commissioner of Inland Revenue accepts that she cannot disallow deductions claimed for prior years in respect of this arrangement because of the four-year time bar in s 108 of the Tax Administration Act 1994. The dispute between the Commissioner and Frucor Suntory New Zealand Ltd in relation to the 2008 and 2009 income years remains in abeyance pending resolution of the present dispute.
[4]The shortfall penalties imposed were $1,786,555 for 2006 and $1,924,779 for 2007.

In the High Court, Muir J upheld Frucor Suntory’s challenge to the assessments.[5] The Commissioner’s appeal to the Court of Appeal was successful as to the disallowance of the deductions but unsuccessful as to shortfall penalties.[6]
[5]Frucor Suntory New Zealand Ltd v Commissioner of Inland Revenue [2018] NZHC 2860, (2018) 28 NZTC ¶23-078 [HC judgment].
[6]Commissioner of Inland Revenue v Frucor Suntory New Zealand Ltd [2020] NZCA 383, (2020) 29 NZTC ¶24-075 (Kós P, Gilbert and Courtney JJ) [CA judgment].
Both sides now appeal to this Court.[7]
[7]Leave to appeal and cross-appeal was granted in Frucor Suntory New Zealand Ltd v Commissioner of Inland Revenue [2020] NZSC 150, (2020) 29 NZTC ¶24-086.
We dismiss the appeal and allow the cross-appeal. This is on the grounds explained at length in these reasons and which we now summarise.
As to tax avoidance our reasons are:
(a)Section BG 1(1) applies to tax arrangements (such as those associated with the note) which, but for its invocation, would have been effective in producing the desired tax advantage.
(b)Such application is justified if the tax advantage results from the use of a tax provision outside the parliamentary contemplation of that provision’s purpose.
(c)Use of a tax provision intended to provide relief in relation to a particular economic burden (such as a cost, a loss or a reduction in income), where such a burden has not, in economic substance, been suffered, will usually lie outside of the relevant parliamentary contemplation. This is particularly so where such use is contrived and artificial.
(d)In this instance the tax provisions relied on by DHNZ provide relief in relation to “interest incurred”. In economic substance, however, the payments in respect of which DHNZ sought the disallowed deductions were repayments of principal. The arrangements on which DHNZ relied to categorise these principal repayments as interest were contrived and artificial. Deductibility for such repayments is not within the purpose of allowing deductibility for “interest incurred”. Accordingly, DHNZ’s use of the deductibility provisions lay outside of the relevant parliamentary contemplation. This means that s BG 1(1) applies to void the arrangements.
Since the purpose and effect of the tax avoidance arrangements were to provide deductibility for what in economic substance were repayments of principal, the Commissioner correctly applied s GB 1(1) to adjust the taxable income of DHNZ to disallow the deductions illegitimately claimed.
As to shortfall penalties:
(a)In this case at least, application of the “about as likely as not to be correct” standard must be against the background of the facts as the Court finds them to be.
(b)On the basis of the facts as we find them to be, the tax positions adopted by DHNZ did not meet that standard and were thus unacceptable.
(c)DHNZ acted with the dominant purpose of obtaining tax advantages with the result that the tax positions were abusive.

The funding arrangement

The background to the funding arrangement

DAP is a Singapore-based wholly-owned subsidiary of Groupe Danone SA (Groupe Danone). DAP formed DHNZ to acquire the then New Zealand-owned Frucor Beverages Group Ltd (FBGL).
The purchase by DHNZ of FBGL was completed in January 2002. The purchase price was $298 million. This was funded as to:
(a)$150 million by DAP subscribing for 1,000 ordinary shares in DHNZ at $150,000 per share; and
(b)$148 million by a loan from Danone Finance SA (Danone Finance), another Groupe Danone subsidiary, to DHNZ.
As we are about to explain, prior to the January 2002 acquisition of FBGL, Groupe Danone and Deutsche Bank had considered putting in place a funding arrangement along the same lines as that eventually entered into in March 2003. It is a fair inference that the advance by Danone Finance to DHNZ was an interim measure intended to hold the position until the funding arrangement could be finalised. It is also a fair inference that the intended end result of the funding arrangement as implemented in March 2003 was to repay the advance by Danone Finance to DHNZ (of approximately $148 million), with DHNZ borrowing from Deutsche Bank some of the necessary funds and DAP contributing the balance by way of a top-up of its capital investment in DHNZ.[8]

The planning documents—prior to DHNZ acquiring FBGL

[8]It is possible that there was a shortfall in the amount required to repay Danone Finance SA and that this was met by DHNZ. The documents are not entirely consistent as to this. Whether this is so is not material and for ease of discussion we will generally proceed on the basis that the repayment of Danone Finance was entirely effected under the funding arrangement.

What follows in this and the succeeding section of these reasons is largely derived from the judgment of the Court of Appeal.
The funding arrangement put in place in March 2003 was based on a generic tax‑driven convertible note funding structure developed by Deutsche Bank. Deutsche Bank had executed the structure in various different jurisdictions, including New Zealand, and claimed to have received “positive rulings” from tax advisors. Groupe Danone was familiar with the structure, having considered using it in conjunction with Deutsche Bank for an earlier proposed transaction in Argentina.
In late 2001, Deutsche Bank discussed with Groupe Danone the possibility of using the convertible note structure as a means of financing the then proposed acquisition of FBGL. In January 2002, it presented two “Efficient Financing Alternatives for [Groupe] Danone in New Zealand”. One of these was a convertible note structure.
The proposal was that:
(a)Deutsche Bank would advance money ($x) to the Danone acquisition vehicle against a convertible note carrying interest over a five‑year term.
(b)This advance ($x) would be repaid by the issue of shares by the Danone acquisition vehicle to Deutsche Bank.
(c)Another Danone subsidiary would enter into a forward purchase agreement with Deutsche Bank under which it would pay ($y) to acquire the shares at termination of the funding arrangement.
(d)The effect (albeit not spelt out in the document) would be that:
(i)Deutsche Bank’s net injection of funds was to be the difference ($z) between the advance to the Danone acquisition vehicle ($x) and the money paid by the other Danone subsidiary under the forward purchase agreement ($y).
(ii)The Danone acquisition vehicle would pay interest on the amount of the advance ($x), with the amount paid as interest being sufficient to reimburse/repay Deutsche Bank for its net injection of funds ($z) on an amortising principal and interest basis.

To correlate this to what eventually happened, DHNZ became the Danone acquisition vehicle, DAP the other Danone subsidiary, $x became $204 million, $y became $149 million and $z became $55 million which, as noted, the Commissioner regards as the amount actually (in economic substance) advanced by Deutsche Bank to DHNZ.

Deutsche Bank explained the proposed tax consequence: the interest payable by the Danone acquisition vehicle on the convertible note would be fully deductible; there should be no capital gains tax on the acquisition of the shares purchased by the other Danone subsidiary and the funding costs in relation to setting up the funding arrangement should also be fully deductible.
The advantages of the proposal were said to be that the structure was familiar to Groupe Danone and there would be no VAT or GST issues. The single constraint mentioned was that a “75% thin capitalisation rule applies in New Zealand and should be taken into account in order to determine the size of the transaction”. Assuming a $300 million acquisition, the note could not exceed $225 million.

The planning documents—after DHNZ acquired FBGL

Groupe Danone advised Deutsche Bank in early February 2002—the month after DHNZ acquired FBGL—that it wished to go ahead with the convertible note structure. This was confirmed in an internal Deutsche Bank email on 4 February 2002:
Actually Yes!
They’ve now confirmed they want to go ahead with the convertible structure. Next steps they’ve asked for are (i) New Zealand memorandum/opinion confirming deductibility of coupons; (ii) UK memorandum/opinion relating to forward purchase; and (iii) termsheet.
The UK side of this I had prepared before when we looked at the Argentinian deal. Can you get something from an NZ lawyer for them? On the termsheet I’ll start a draft and send it over to you.
Concerning fees they have suggested upfront arrangement fee of [USD1 million] plus credit spread and costs (the idea would be that the credit spread is set by Corporate Bank in Paris who provide risk weighted assets and take the credit risk in return for earning the credit spread. Accordingly [Deutsche Bank Structured Capital Markets] just keeps the upfront fee but has no credit risk etc). Danone’s justification for this level of fee is:
1.        Fees for these transactions in Europe are generally 1% of the principal. Here the principal on the notes is only about $80mio;
2.        We had agreed to execute the Argentinian transaction for this pricing (although this is because it would have been a ground-breaking transaction for Emerging Markets in Argentina. Also we expected to earn more by selling the notes to a tax sparing investor);
3.        They have (apparently) been inundated by other banks willing to execute this structure with them in New Zealand (they have a moral commitment to us arising out of Argentina).
Accordingly we should probably accept this but let me know what you think (there is also a lot of glory in this with [Deutsche Capital Management] who have been trying to develop the relationship with Danone).
…

By way of explanation, we note that “the principal on the notes” of “about $80mio” is a reference to what was then thought to be the approximate amount to be advanced by Deutsche Bank to the Danone group, an amount that in the end was $55 million.

The email proceeds on the basis that full deductibility of the interest payable by DHNZ was critical. It was also critical that the other Danone subsidiary (not yet identified as DAP) should not incur a tax liability in relation to the forward purchase agreement. The apparent gain under that agreement (the difference between the amount paid under the forward purchase agreement and the face value of the note) equated to the net injection of funds to be made by Deutsche Bank. If that apparent gain was taxable it would counteract the tax advantage in New Zealand of deductibility of repayment of the net injection of funds. This is presumably what was to be addressed in the “UK memorandum/opinion related to forward purchase” referred to in the email.
As at 24 May 2002, it was proposed that the face value of the note be $225 million and the forward purchase payment $154 million. The $225 million equated to 75 per cent of $300 million, thus complying with the thin capitalisation rules. The calculation of the $154 million was explained in a document distributed on that date headed “Project Falcon”, the project name ascribed to the transaction by Deutsche Bank. This explained that the purchase price payable by DAP:
… on day one will be calculated as the face value of the convertible note less the present value of convertible note coupon payments discounted at the applicable zero coupon swap rate plus credit margin (0.35%).

By way of illustration, if the face value of the note was $225 million, then the purchase price payable under the forward purchase agreement would be $154 million, being $225 million less $71 million (the present value of semi‑annual coupons of $8.7 million at the then applicable interest rate of 7.736 per cent per annum). Deutsche Bank would fund the “net investment” (approximately $71 million) from its normal market sources for New Zealand dollars “swapped to an amortising flow that matches the profile of the net investment”.

The purpose of the arrangement was set out under a heading “Summary” in the “Project Falcon” document:
The structure provides term funding to DHNZ at an after tax cost that is significantly below the Group’s normal cost of funds (ie. pre‑tax equivalent of approximately minus 1.50%).

This document was used as a template and updated from time to time as the transaction progressed towards completion.

By 18 September 2002, the figures had changed. The face value of the note was now expected to be $215 million and the forward purchase price $151 million, being $215 million less $64 million (the present value of semi-annual coupons of $7.69 million at an interest rate of 7.15 per cent per annum). As well, there was some additional protection for the Danone group as it was agreed that the shares to be issued by DHNZ would be non-voting.
The net funding figure of $64 million had dropped to $61 million by 29 November 2002. An internal Deutsche Bank email sent on that date explained the net effect of the transaction—“Danone raises approx NZD61m for [five] years on amortising basis”.

An internal Deutsche Bank approval document prepared on 2 December 2002 provides further confirmation of the purpose of the five-year structured transaction—it “is designed to provide cheaper, tax efficient funding to [DHNZ]”.[9] The difference between the face value of the note and the forward purchase payment “will be amortised from the convertible coupons”. The payments to be made by DHNZ represented repayment of principal and interest over the five-year term on an amortising basis. DHNZ would claim interest deductions for what were described as “nominal payments” of interest on the face value of the note but were, in effect, payments of principal and interest on the net funding amount.
[9]The document refers to “a subsidiary of Groupe Danone SA in New Zealand”. For ease of discussion we will refer to this subsidiary as DHNZ.
An internal Deutsche Bank email of 3 March 2003 marked a change of approach in calculating the values to be attributed to the components of the financing arrangement. Up until this time, the starting figure was the face value of the note, the upper limit on which was the maximum amount on which interest could be claimed under the thin capitalisation rules. The other components, being the forward purchase price and the net advance to DHNZ, were calculated by reference to this figure. In contradistinction, the 3 March 2003 email used as a starting point for other calculations the forward purchase amount fixed at $149 million, with the face value of the note becoming a function of that amount.
The reasons for the shift to the forward purchase amount and why it was fixed at $149 million are not made explicit in the contemporaneous documents. A possible explanation for this is provided in the Court of Appeal judgment but, as we see it, nothing turns on why this figure was chosen.
One of the few available Danone group documents is an internal memorandum sent on 11 March 2003 by Pierre-André Terisse, the person responsible for agreeing to the “net funding amount” and the final amount of the note. Mr Terisse explained that his memorandum was “intended to give a brief description of the transaction for signatories”. His memorandum included the following summary:

The structure, established by Deutsche Bank, works as follows:
-issuance by DHNZ of 215 m NZD convertible bonds, subscribed by Deutsche Bank
-Deutsche Bank keeps the principal amount, which gets reimbursed over 5 years,
-But sells the conversion rights to [DAP] for an amount of 149 m NZD
-At the end of the 5 years, shares issued in repayment of the bonds are transferred to [DAP], or, as a fallback, to Compagnie Gervais Danone

Benefits obtained
-Financing cost: extremely attractive for NZD financing.
-NZD financing: putting a debt in the same currency as cash-flows of the company acquired provides us with a natural hedging; furthermore, interest [is] located in the same country as operating income.

Issues
-Legal / tax issues have been checked by France Hasselman, tax opinions have been obtained

There are three points about this memorandum which warrant mention or explanation. The first is that the $149 million to be paid under the forward purchase agreement is referred to in the document but the other figures are not correlated to it. Second, the reference to Compagnie Gervais Danone relates to the contingency arrangement which we discuss briefly below at [38]. The third and most significant is that the attractiveness of the financing cost was a function of tax efficiency, the benefits of which, when calculated in dollar terms, were always assessed by reference to the tax position in New Zealand of DHNZ. This is illustrated by an email from Deutsche Bank to Mr Terisse soon afterwards which indicated that the “pre-tax equivalent benefit from the transaction” should be approximately:
[Net Funding Amount]*[Tax Rate]/[1 - Tax Rate] - US$ 1mm, which is roughly NZ$ 24mm
The rate setting and final calculation of the face value of the note did not occur until 14 March 2003. An internal Deutsche Bank email sent that day following execution of the documents confirms:
… For your info, the rates/amounts agreed with Danone today are as follows:
Convertible Note Principal                 NZD204,421,565
Interest rate  6.50% pa payable 18 Sept/18 Mar
Issue Date  18 Mar 2003
Maturity Date  18 Mar 2008
Forward Purchase Price  NZD149,000,000
Therefore net funding amount is NZD55,421,565

Tax treatment in Singapore

Under the funding arrangement, DAP prepaid $149 million for the purchase from Deutsche Bank of the shares to be issued by DHNZ when the note matured. These were to have a face value of $204 million. It was critical to the overall tax efficiency of the funding arrangement that DAP not be taxed in Singapore on the “profit”, representing the difference between the value of shares acquired on maturity (assumed to be $204 million) and the price paid five years earlier ($149 million). If this $55 million “profit” was taxable in Singapore, it would substantially or completely negate the tax advantage DHNZ (and thus the Danone group) derived from being able to treat as deductible in New Zealand what in substance were principal repayments of $55 million.
An opinion was obtained from PricewaterhouseCoopers in Singapore which was to the effect that they did not expect that DAP would be subject to tax in Singapore by reason of the steps taken on the maturity of the funding arrangement.

The components of the funding arrangement

At inception of the funding arrangement:
(a)Deutsche Bank paid $204 million to DHNZ.
(b)Upon receipt of this $204 million, DHNZ returned $60 million of capital to DAP in a share buyback and the balance of $144 million was paid in full or substantial satisfaction of the amount then owing to Danone Finance.
(c)DAP borrowed $89 million from BNP Paribas. This, along with the net $55 million provided by Deutsche Bank (which it borrowed from its internal treasury), either substantially or completely funded repayment of the amount owed to Danone Finance.[10]
(d)DAP paid $149 million to Deutsche Bank under the forward purchase agreement. This was funded as to $89 million by the advance from BNP Paribas and, as to the balance of $60 million, by the return of capital from DHNZ (which in turn had been funded by the $204 million advance from Deutsche Bank).
(e)DHNZ paid a fee of $1.8 million to Deutsche Bank.
[10]See above at n 8.
Aspects of the documentation of the funding agreement are reasonably complex.
As will be appreciated, the forward purchase agreement was to ensure that DAP retained complete ownership of DHNZ on termination of the funding arrangement. This required Deutsche Bank to transfer to DAP the shares it was expected to elect to receive. The complexity of the documentation was a function of the need to cover various risks that might prevent that transfer occurring. Since such risks were always seen as remote and did not crystallise, there is no need to review in these reasons the associated contractual arrangements. There is, however, one point we should mention. This is that although in form the note was for optional conversion of the $204 million advance into shares in DHNZ, the surrounding contractual arrangements made that conversion effectively mandatory.

Explanatory documents written after execution

A Project Falcon summary document prepared by Deutsche Bank post‑execution confirmed how the note issue price and the net funding amount were derived and how the net funding amount was to be serviced:
The net funding requirement is therefore the difference between the convertible note issue price and the share forward purchase price. This funding will be serviced by the convertible note interest payments. … The funding for the net amount (i.e. note subscription less prepaid forward purchase price) was provided by [Deutsche Bank Treasury] to [Deutsche Bank Structured Capital Markets] by way of a 5 yr NZD amortising loan, to be fully serviced by the note interest payments.
This understanding was shared by Groupe Danone. A document prepared on 15 October 2003 concerning the DHNZ funding arrangement, under a heading “Purpose and ‘débouclage’ of the operation”, states:
During the 5 years, DHNZ pays coupons to Deutsche Bank. Those coupons are analyzed differently according to tax/statutory and consolidated accounts:
·For statutory, coupons are considered as interest expenses deductible for tax purposes. They amount to a total of some [$66 million] … The necessary cash is provided by dividends received from Frucor.
·For consolidation, coupons paid are analyzed in two separate elements 1. Reimbursement of Deutsche Bank loan for [$54 million] and 2. interest expense on this loan for the difference, i.e. [$12 million]. As this is a permanent difference, no deferred tax shall be recorded in consolidated accounts.
At maturity date:
·DHNZ reimburses the remaining [$150 million] loan from Deutsche Bank by delivering its 40% own shares previously bought back from [DAP]
·Deutsche Bank delivers those shares to [DAP], without receiving any cash, as [$150 million] were paid in advance in 2003
·[DAP] holds 100% of DHNZ shares, as it was prior to the refinancing scheme.
There is one other (undated) Danone group document which is material. It includes the following section:
What was the point of the scheme?
The scheme allowed DHNZ to finance the purchase of [FBGL] in a way that would entitle it to tax credits for the life of the scheme.
Under the arrangement DHNZ made two coupon payments to [Deutsche Bank] each year. The coupon payments were approximately $7m per payment and were funded by payment of a fully imputed dividend up from FBL [Frucor Beverages Ltd, which was amalgamated with DHNZ in 2009] to FBGL and finally to DHNZ. These coupon payments were treated differently for Management and Statutory purposes.
For Stat (and Tax) purposes, the whole payment was treated as an interest expense. The interest payment was 100% deductible. Total payments over the life of the scheme added up to $66m, which equated to $21.8m of tax credits (approx $4.4m for each year of the scheme’s life).
For Management purposes, part of the payment was treated as an interest expense, and part was treated as repayment of the principal of the convertible note loan.
…
An estimated … NZD 21.6m was saved in taxes over the 5 years as interest expense of NZD 65.5m arising from the convertible bond from Deutsche Bank of NZD 204m can be claimed as a deduction against the income of the Frucor Group as NZ practices consolidated tax filing.

The reference to the tax that was “saved” presupposes a counterfactual in which $55 million was borrowed and deductibility was confined to interest payments on that sum.

Accounting treatment

During the currency of the funding arrangement, DHNZ recorded the convertible note as “borrowings” of $204 million and the coupon payments as “interest expense”. On the basis of the material we have seen and the logic of the structure of the transaction, we infer that DHNZ accounted in its statutory accounts for the disbursement of the $204 million by:
(a)debiting the inter-company loan and expensing the fee of $1.8 million and crediting cash, $144 million; and
(b)debiting share capital and crediting cash, $60 million.
Although we do not have the management accounts of DHNZ and DAP, internal company documents recording the way in which the funding arrangement was to be treated were in evidence (including the document referred to at [41]).
Internal DHNZ documents suggest that its management accounts differed from the statutory accounts in two significant respects. In the management accounts:
(a)the transaction was treated as involving an advance of $55 million and the interest payments under the note as payments of principal and interest on that advance; and
(b)the repurchase of the shares was reversed, by debiting a receivable from DAP and crediting capital of $60 million.

The corollary of these entries (and their logic) is that DHNZ treated DAP as having contributed a further $89 million in capital. According to the documents we have, this was to be recorded in DHNZ’s management accounts with an entry recognising an $89 million “share premium”.

In DAP’s management accounts, the $60 million repayment of capital was to be reversed (matching the corresponding reversal in DHNZ’s management accounts) and the $89 million borrowed from BNP Paribas to be treated as capital contributed by DAP to DHNZ; this matching what we see as the likely corresponding entry in the DHNZ management accounts.
An internal DAP accounting note recorded that at the maturity of the scheme, the $89 million was to be capitalised; this because:
a)This is pseudo capital and should be treated as part of the investment
b)It can be seen as the price that we pay to hold Frucor, indirectly though through DHNZ.
…

Steps at maturity and subsequent return of capital

On 20 February 2008, Deutsche Bank, in accordance with the convertible note deed, gave notice to DHNZ requiring it to satisfy its obligations to repay the principal amount outstanding by issuing shares on the maturity date. At maturity on 18 March 2008, DHNZ issued 1,025 new non‑voting shares to Deutsche Bank which immediately transferred those shares to DAP pursuant to the forward purchase agreement.
On 22 December 2008, DHNZ repurchased from DAP 747 non‑voting shares and 307 ordinary shares thereby returning $204 million of surplus capital to DAP. This was followed in February 2009 by DAP selling all shares in DHNZ to Suntory (NZ) Ltd, a subsidiary of a Japanese beverage manufacturer and distributor.

Tax avoidance—the legal framework

Section BG 1(1) of the Act is in these terms:
BG 1 Tax avoidance
Avoidance arrangement void
(1)A tax avoidance arrangement is void as against the Commissioner for income tax purposes.
Section OB 1 defines arrangement, tax avoidance and tax avoidance arrangement as follows:
arrangement means an agreement, contract, plan, or understanding (whether enforceable or unenforceable), including all steps and transactions by which it is carried into effect
…
tax avoidance includes—
(a) directly or indirectly altering the incidence of any income tax:
(b)directly or indirectly relieving a person from liability to pay income tax or from a potential or prospective liability to future income tax:
(c)directly or indirectly avoiding, postponing, or reducing any liability to income tax or any potential or prospective liability to future income tax
tax avoidance arrangement means an arrangement, whether entered into by the person affected by the arrangement or by another person, that directly or indirectly—
(a) has tax avoidance as its purpose or effect; or
(b)has tax avoidance as 1 of its purposes or effects, whether or not any other purpose or effect is referable to ordinary business or family dealings, if the purpose or effect is not merely incidental
Provisions such as s BG 1(1) are often referred to as general anti-avoidance rules (GAARs). Historically, the main problem with the application of a GAAR (such as s BG 1(1), its precursors in New Zealand and its equivalents in other jurisdictions) has been the apparent awkwardness of applying a generally expressed rule in the otherwise highly prescriptive legislative context of income tax legislation. When tax avoidance is in issue, the question of whether the GAAR, or the provision(s) relied on by the taxpayer, should prevail is not susceptible to resolution on the basis of usual interpretative techniques addressed to the meaning of a single provision.
The first of the modern cases, Challenge Corporation Ltd v Commissioner of Inland Revenue, represents an attempt to resolve this awkwardness.[11] There, the Privy Council concluded that the effect of the GAAR (at that time, s 99 of the Income Tax Act 1976) was that the entitlement to group losses under s 191 of the 1976 Act was dependent upon the companies concerned having been associated at the time the losses were incurred, a condition which was not explicit and could also hardly be said to be implicit in s 191, at least if looked at without reference to the GAAR. Instead, this conclusion was arrived at on the basis that the taxpayer had not suffered an economic burden of the kind envisaged by Parliament as warranting the tax advantage claimed.
[11]Challenge Corporation Ltd v Commissioner of Inland Revenue [1986] 2 NZLR 513 (PC) [Challenge PC judgment].
As Challenge illustrates, and the approach expressly and authoritatively adopted by this Court in Ben Nevis Forestry Ventures Ltd v Commissioner of Inland Revenue now recognises,[12] the GAAR applies to void tax arrangements which, but for the invocation of the GAAR, would be effective, with the reconciliation of the GAAR and the provision relied on by the taxpayer achieved through application of a parliamentary contemplation test. This was explained in the reasons of Tipping, McGrath and Gault JJ in Ben Nevis:
[107] When, as here, a case involves reliance by the taxpayer on specific provisions, the first inquiry concerns the application of those provisions. The taxpayer must satisfy the court that the use made of the specific provision is within its intended scope. If that is shown, a further question arises based on the taxpayer’s use of the specific provision viewed in the light of the arrangement as a whole. If, when viewed in that light, it is apparent that the taxpayer has used the specific provision, and thereby altered the incidence of income tax, in a way which cannot have been within the contemplation and purpose of Parliament when it enacted the provision, the arrangement will be a tax avoidance arrangement. …
…
[109] … The ultimate question is whether the impugned arrangement, viewed in a commercially and economically realistic way, makes use of the specific provision in a manner that is consistent with Parliament’s purpose. If that is so, the arrangement will not, by reason of that use, be a tax avoidance arrangement. If the use of the specific provision is beyond parliamentary contemplation, its use in that way will result in the arrangement being a tax avoidance arrangement.
[12]Ben Nevis Forestry Ventures Ltd v Commissioner of Inland Revenue [2008] NZSC 115, [2009] 2 NZLR 289 at [107] per Tipping, McGrath and Gault JJ.
In these passages, the expression “specific provision” is used in contradistinction to the GAAR, rather than as indicating a tax provision which is expressed with great particularity. Indeed, it can cover general features of the tax system, such as income tax being a tax on income, as in Penny v Commissioner of Inland Revenue (Penny and Hooper),[13] or the general right to deduct expenses incurred in generating income, as in many other cases.
[13]Penny v Commissioner of Inland Revenue [2011] NZSC 95, [2012] 1 NZLR 433.
In applying the parliamentary contemplation test, the courts have rejected claims to tax advantages where the taxpayers have not suffered the economic burden that is the usual corollary of, and purpose for, the conferral of those advantages. This is exemplified by Penny and Hooper. In that case the taxpayers had transferred their businesses to companies which then paid them salaries which were distinctly lower than what they had been earning. The companies were owned by trusts which they had established. The transfers to the companies occurred at a time when the marginal tax rate on income was increasing. The surplus income (that is, after payment of the salaries) available ultimately to the trusts continued to be at the disposal of the taxpayers. Blanchard J observed that:[14]
… Parliament has deliberately preserved, and in fact enlarged, the New Zealand general anti-avoidance provision … It continues to have work to do whenever a taxpayer uses specific provisions of the Act and otherwise legitimate structures in a manner which cannot have been within the contemplation of Parliament. … Woodhouse P said in Challenge Corporation Ltd v Commissioner of Inland Revenue^[[15]] that there must be a weapon able to thwart technically correct but contrived transactions set up as a means of exploiting the Act for tax advantages. That is what the artificially low salary settings did in this case. They reduced each taxpayer’s earnings but at the same time enabled the company’s earnings (derived only because of the setting of the salary levels) to be made available to him through the family trusts. In reality, the taxpayers suffered no actual loss of income but obtained a reduction in liability to tax as if they had, to adapt Lord Templeman’s dictum in Challenge.
[14]At [47] (footnotes omitted).
[15]Challenge Corporation Ltd v Commissioner of Inland Revenue [1986] 2 NZLR 513 (CA).

In his reasons for the majority in Challenge, to which Blanchard J made reference, Lord Templeman observed:[16]
In an arrangement of tax avoidance the financial position of the taxpayer is unaffected (save for the costs of devising and implementing the arrangement) and by the arrangement the taxpayer seeks to obtain a tax advantage without suffering that reduction in income, loss or expenditure which other taxpayers suffer and which Parliament intended to be suffered by any taxpayer qualifying for a reduction in his liability to tax.

As is at least implicit in what Lord Templeman said, and as was made explicit by Blanchard J, the “reduction in income, loss or expenditure” in the passage above is to be assessed as one of substance rather than on the form of the arrangement under challenge. In Penny and Hooper, while the taxpayers had undoubtedly suffered a reduction in income in a formal sense, as Blanchard J recognised, the reason why they lost the case was because under the substance of the arrangement the “lost income” remained available to them.

[16]Challenge PC judgment, above n 11, at 562.

There is one aspect of the Ben Nevis two-stage approach which warrants brief comment. The “first inquiry” referred to was said to be whether the taxpayer’s use “of the specific provision is within its intended scope”.[17] If so, the court should move to a second inquiry as to whether the taxpayer “has used the specific provision … in a way which cannot have been within the contemplation and purpose of Parliament”.[18] On a literal approach to this formulation, the result arrived at in Ben Nevis might seem paradoxical—the taxpayers’ use of the specific provisions relied on was:
(a)within their “intended scope”; but at the same time
(b)“cannot have been within the contemplation and purpose of Parliament”.
[17]Ben Nevis, above n 12, at [107] per Tipping, McGrath and Gault JJ.
[18]At [107].
Ben Nevis cannot sensibly be applied on the basis that a conclusion that the use of the provisions relied on was within their “intended scope” for the purposes of the first inquiry precludes resort to s BG 1(1) at the second stage of the exercise. Instead, making sense of the Ben Nevis approach means that in the context of tax avoidance, scheme and purpose considerations of a kind that in other contexts would be applied to the construction of a statutory provision, largely come into play at the second stage of the inquiry. On this basis, the first inquiry can be dealt with on a largely textual basis. If, on the wording of the provisions relied on, an arrangement appears to warrant the tax position adopted, the court can move to the second inquiry. When tax avoidance is in issue, it is the second inquiry that is of paramount importance.

The High Court and Court of Appeal judgments as to the application of s BG 1

The High Court judgment

The High Court Judge found in favour of Frucor Suntory. We outline what we see as the key elements of his reasoning.
He saw the purposes of the funding arrangement as being:
(a)To put in place what was seen within the Danone group as a more appropriate debt/equity ratio for DHNZ from approximately 50:50 to approximately 63:37.[19]
(b)To do so in a way which avoided the adverse tax consequences for other Danone companies associated with the implementation of what he described as “alternative structures”.[20]
[19]HC judgment, above n 5, at [36(g)], [165]–[166] and [198].
[20]At [141(j)]–[141(l)], [165]–[166], [198] and [203].
He summarised the alternative structures at [141(j)] as including:
(i)[DHNZ] borrowing an additional $60 million under the Cash Management Agreement [with Danone Finance] to achieve its desired debt/equity rebalancing.
(ii)DAP (or Danone Finance) lending $204 million to [DHNZ] for five years at 6.5 per cent either by way of interest-bearing debt or convertible note.
(iii)[DHNZ] issuing a convertible note to [Deutsche Bank] on the same terms but without the Forward Purchase between [Deutsche Bank] and DAP and [Deutsche Bank] funding the $149 million Forward Purchase amount internally or from another bank or third party or by a loan from DAP or Danone Finance or by a sub-participation by a third party or Danone entity.
He then went on to say:[21]
… I accept that in respect of each of alternatives (i) and (ii) (and likewise (iii) with funding or sub-participation by a Danone entity) these alternative arrangements would have given rise to assessable interest in the hands of the relevant offshore Danone entity. By contrast, the distinguishing feature of the transaction entered into was that, although the same level of deduction was available in New Zealand, the Forward Purchase provisions negated foreign‑assessable income.

On the basis that the avoidance of foreign tax is not tax avoidance for the purposes of s BG 1(1), the corollary of these findings was that Frucor Suntory’s challenge to the assessments succeeded.

[21]At [141(l)].

As will be apparent, it follows that the High Court judgment rests in part on the factual finding that the purpose of the scheme was to adjust the debt/equity ratio of DHNZ and do so in a way which was efficient in respect of liability to foreign tax. This was in relation both to DAP not being liable to tax in Singapore for the $55 million apparent gain under the forward purchase agreement and what he saw as an advantage to Danone Finance in not having to pay tax on interest receipts from DHNZ.[22]
[22]At [165]–[166].
In the course of his judgment, the High Court Judge made other findings or observations, some of which we should mention:
(a)He concluded that it was not appropriate to examine the funding arrangement in terms of “overall impact at a group or consolidated level looking at the net external position of entities under common control”.[23] Such an approach would be to subvert what he described as the “separate entity principle”,[24] which required separate recognition of the roles of different companies within a single group.
(b)He considered that the funding arrangement involved real money flows[25] and, to the extent that there was circularity, it was not material, as real liability was affected and there was real change to DHNZ’s funding structure.[26]
(c)He accepted that “the very particular sum of the [n]ote” (which as noted in [33] was $204,421,565) was unusual and was a “strong indicator that the company’s medium term corporate financing requirements did not drive its face value”.[27] But he did not regard this or the fact that the note lacked the usual characteristics of a note typically issued when “new ‘debt’” is raised as “a significant indicator of avoidance”.[28] He described as “compelling” the argument that the financial arrangement rules and the Commissioner’s determinations applicable to notes did not draw distinctions between the issue of shares to a company’s parent (directly or, as in this case, indirectly) or third parties.[29]

The Court of Appeal judgment

[23]At [133].

[24]At [122].

[25]At [141(a)].

[26]At [163].

[27]At [141(c)].

[28]At [141(f)].

[29]At [183].

The Court of Appeal saw the case very differently:[30]
[82]     It seems to us to be reasonably plain that the funding arrangement had New Zealand tax avoidance as one of its purposes or effects and this was not merely incidental to some other purpose. The primary purpose of the funding arrangement was the provision of tax efficient funding to [DHNZ]. That was its stated goal. The tax advantage was gained in New Zealand through the interest deductions [DHNZ] claimed. DAP (in effect) paid $149 million to [DHNZ] for the shares on day one but with the payment being structured to enable [DHNZ] to claim interest deductions on it over a five‑year term. This tax beneficial outcome was achieved by calculating the amount that needed to be added to the $149 million to enable interest payments on that grossed-up sum over five years to match the amount required to repay over the same period the amount of the gross-up plus interest (the funds introduced by Deutsche Bank). [DHNZ] was thereby able to repay the $55 million advanced by Deutsche Bank plus interest over the five‑year term of $11 million but claim the entire $66 million as an interest deduction.
[83]     DAP’s subscription for equity was effectively repackaged as a loan from Deutsche Bank to achieve the intended tax benefits for [DHNZ]. DAP’s equity subscription was bundled with an amortising loan from Deutsche Bank in an artificial and contrived manner to enable [DHNZ] to claim interest deductions on the loan which were, in substance, repayments of principal and interest payable to Deutsche Bank in respect of the funds it introduced to facilitate the arrangement.
[84]     ... We agree that avoiding offshore tax was an important consideration. The funding arrangement gave rise to the prospect of generating an unwelcome offshore taxable gain because of the notional increase in the value of the non‑voting shares over the five‑year period from $149 million to $204,421,565 (or whatever other figure happened to be generated by the formula at the date of closing). This potential problem needed to be managed for the funding arrangement to succeed. However, we view the need to avoid capital gains tax on the notional share value growth as a condition precedent to the intended funding arrangement rather than being its object. The condition was always understood to be readily satisfied simply by choosing a Danone entity resident for tax purposes in a jurisdiction that would not impose capital gains tax on the inferred uplift in the value of the shares represented by the difference between the forward purchase amount and the face value of the convertible note. Thus, this condition had to be ticked off before the funding arrangement could proceed but that was never going to be a problem. This consideration does not detract from our assessment that a more than incidental purpose and effect of the funding arrangement was to engineer tax deductions for interest expenses claimable by [DHNZ] in New Zealand of sufficient magnitude to repay Deut[s]che Bank. That purpose was not to come at the expense of creating tax problems elsewhere.
[85]     We consider the funding arrangement fits within the Supreme Court’s formulation in Ben Nevis as one enabling the taxpayer to gain the benefit of the specific provision in an artificial and contrived way. In our view, this transaction was in many respects artificial and it was clearly contrived for the very purpose of enabling [DHNZ] to gain the benefit of the specific provision allowing interest deductions. The artificial and contrived features of the funding arrangement are not seriously in dispute and most were accepted by the Judge. Taken together, they reveal that the purpose of the arrangement was to dress up a subscription for equity as an interest only loan to achieve a tax advantage. It is hard to discern any rational commercial explanation for the artificial and contrived features of the arrangement, other than tax avoidance.
…
[90]     As a matter of commercial and economic reality, the payment of $149 million made by DAP did not carry any liability for [DHNZ] (or Deutsche Bank) to pay interest. The only amount that did attract interest was the $55,421,565 independently advanced by Deutsche Bank.
[30]CA judgment, above n 6.
The Court rejected the challenge by Frucor Suntory to the Commissioner’s reconstruction:[31]
[103] Section GB 1 does not require the Commissioner to consider other arrangements the taxpayer might have entered into had it not chosen to proceed with the tax avoidance arrangement under review. Further, the tax advantage with which the section is concerned is the New Zealand tax advantage achieved by the New Zealand taxpayer — [DHNZ]. While DAP is a party to the funding arrangement, its funding costs and tax position are irrelevant to the analysis that must be conducted under s GB 1.
[104] We have already concluded that the principal driver of the funding arrangement was the availability of tax relief to [DHNZ] in New Zealand through deductions it would claim on the coupon payments. The benefit it obtained under the arrangement was the ability to claim payments totalling $66 million as a fully deductible expense when, as a matter of commercial and economic reality, only $11 million of this sum comprised interest and the balance of $55 million represented the repayment of principal. The tax advantage gained under the arrangement was therefore not the whole of the interest deductions, only those that were effectively principal repayments. We consider the Commissioner was entitled to reconstruct by allowing the base level deductions totalling $11 million but disallowing the balance. The tax benefit [DHNZ] obtained “from or under” the arrangement comprised the deductions claimed for interest on the balance of $149 million which, as a matter of commercial reality, represented the repayment of principal of $55 million.
[31]Footnote omitted.
The Court also rejected the Commissioner’s cross‑appeal on shortfall penalties. It concluded that DHNZ had not taken an unacceptable tax position.[32] The Court was “not persuaded that [DHNZ’s] arguments could be dismissed as lacking in substantial merit”.[33] This was primarily on the basis that these arguments were accepted as correct by the High Court Judge.

The application of s BG 1(1) to the funding arrangement: our assessment

[32]At [107].

[33]At [107].

For the purposes of stage one of the Ben Nevis framework, we accept that the funding arrangement and its various components were not shams and that, subject to the effect of s BG 1(1), DHNZ was entitled to deduct the $66 million it paid to Deutsche Bank as interest. In issue is whether such deductibility is consistent with s BG 1(1).
Under s DB 7(1), a deduction is allowed for “interest incurred”. “Interest” is relevantly defined in s OB 1 in this way:
(a)       for a person’s income,—
(i)means a payment made to the person by another person for money lent to any person, whether or not the payment is periodical and however it is described or computed; and
(ii)      does not include a redemption payment; and
(iii)     does not include a repayment of money lent:
…
(d)       in [section] … DB 7 …—
(i)includes expenditure incurred under the financial arrangements rules or the old financial arrangements rules …

The word “incurred” in s DB 7(1), along with the exclusion of “repayment of money lent” in s OB 1(a)(iii) and the repetition of the word “incurred” in s OB 1(d)(i), give a good indication of the kinds of interest cost for which deductibility is provided.

Subpart EW of the Act (in which the financial arrangement rules are found) contemplates both repayment of debt by the issue of shares and advance subscription for shares (payment now for shares to be issued in the future) and, as well, provides for tax treatment where the payment is less than the assumed or agreed future value of the shares. The application to the note of these rules, in conjunction with Determination G22,[34] involves no substantive complexity; this because their effect, on the arrangement as implemented, was that the amount advanced under the note was treated as debt until conversion and the coupon payments were treated as interest.
[34]Te Tari Taake | Inland Revenue “Determination G22: Optional Conversion Convertible Notes Denominated in New Zealand Dollars Convertible at the Option of the Holder” (24 October 1990) < Determination G22 has been replaced by Determination G22A: Te Tari Taake | Inland Revenue “Determination G22A: Optional Convertible Notes Denominated in New Zealand Dollars” (26 September 2006) <>
In Accent Management Ltd v Commissioner of Inland Revenue, the decision of the Court of Appeal upheld in Ben Nevis, the Court commented:[35]
Given the generality of cases to which specific tax rules necessarily apply, it would be unrealistic to confine the application of general anti-avoidance provisions to transactions which lie outside of a discernible specific legislative purpose. When construing such specific rules and looking for their scheme and purpose, it is necessary to keep general anti-avoidance provisions steadily in mind. On this basis, it will usually be safe to infer that specific tax rules as to deductibility are premised on the assumption that they should only be invoked in relation to the incurring of real economic consequences of the type contemplated by the legislature when the rules were enacted.
[35]Accent Management Ltd v Commissioner of Inland Revenue [2007] NZCA 230, (2007) 23 NZTC 21,323 (CA) [Accent Management CA judgment] at [126].
There is nothing in the relevant deductibility provisions (just referred to at [69]) to suggest a parliamentary contemplation that deductibility should extend to what in substance are repayments of principal and which, in economic effect, are not “interest incurred” or “expenditure incurred”. Nor is there any indication of legislative indifference to tax avoidance. Indeed, it would be strange to attribute to Parliament a purpose of allowing deductibility for payments that in substance are repayments of principal and able to categorised as “interest” only through the mechanism of artificial and contrived arrangements. No post-Challenge case was cited to us in which a court has upheld the deductibility of a cost incurred in legal form but, in the opinion of the court, not in economic substance, where the arrangement giving rise to the cost was artificial and contrived.
As we have explained, the High Court judgment in the present case was premised on two findings of fact: first, that the primary purpose of the funding arrangement was to put in place what was seen within the Danone group as a more appropriate debt/equity ratio for DHNZ from approximately 50:50 to approximately 63:37; and, secondly, that there was a related purpose to do so in a way which did not attract unwelcome foreign tax in Singapore and relieved Danone Finance of a liability to tax on interest payments from DHNZ. The Judge’s conclusion that s BG 1(1) was not engaged flowed from those findings of fact.
In company with the Court of Appeal, we see the facts very differently.
Despite the apparent complexity of the funding arrangement as a whole, its economic substance was straightforward. Netting off the transactions which occurred when the funding arrangement came into effect, in the manner in which we infer, they were separately treated by DHNZ and DAP in their management accounts. There were two primary effects:
(a)DHNZ’s relevant debt was reduced from the $144 million owed to Danone Finance to the $55 million owed to Deutsche Bank; and
(b)DAP’s “investment” in DHNZ increased by $89 million (corresponding to the money it borrowed from BNP Paribas).
In economic substance, Deutsche Bank advanced $55 million to DHNZ which was fully repaid on an amortising principal and interest basis over the term of the note. There was never any alteration of substance in relation to the ownership by DAP of DHNZ. DAP started off owning 100 per cent. And as was always intended, it wound up owning 100 per cent. The net effect of everything that happened was that the advance of $55 million was repaid on a basis that ostensibly permitted DHNZ to deduct all payments made, including repayments of principal. This was undoubtedly the effect of the arrangement in terms of s BG 1(1). And because the arrangement had been structured so as to produce this effect, and there was no plausible commercial reason other than tax avoidance for so structuring the arrangement, this effect was distinctly more than incidental.

It follows that, contrary to the finding of fact in the High Court, the purpose of the arrangement was not to alter, by increasing, the debt/equity ratio of DHNZ. In substance its effect and purpose was quite the reverse. DHNZ’s relevant indebtedness decreased from $144 million to $55 million and the capital investment of DAP in DHNZ increased by $89 million (representing the money it borrowed from BNP Paribas). As well:
(a)While it is correct, as the Court of Appeal pointed out, that it was essential to the scheme that DAP not be liable to tax in Singapore on the notional gain in the new shares which it took when the funding arrangement was unwound, it is unrealistic to treat the avoidance of tax on that gain as a purpose of the arrangement. The “gain” was itself entirely contrived; being in effect the other side of the coin to the $55 million in principal that DHNZ expensed against its New Zealand income.
(b)Income (in the form of interest payments from DHNZ) foregone by Danone Finance once the advance was repaid (and any consequential tax consequences) would presumably have been matched either by a reduction in Danone Finance’s cost of funding or by income derived from the redeployment of the money represented by the repayment of the $144 million it was owed. There is no indication in any of the planning documents that the tax position of Danone Finance was a material consideration.
None of the counterfactuals accepted in the High Court or advanced to us are true comparators for the actual arrangement entered into, under which DAP injected into DHNZ (through Deutsche Bank) sufficient funds (being the money it borrowed from BNP Paribas) to enable DHNZ to retire $89 million of debt owing to Danone Finance, and DHNZ borrowed $55 million to discharge the balance of the liability to Danone Finance. On any counterfactual which replicated these essential features of the funding arrangement, DHNZ would not be able to offset its repayments of the $55 million against its revenue.
Contrary to the view of the High Court Judge and the arguments of Frucor Suntory, this approach does not involve inappropriately conflating DHNZ and DAP:
(a)We accept that New Zealand companies in overseas ownership are generally taxed on a standalone basis (that is, in accordance with what the High Court Judge called the “separate entity principle”). This principle is most evident in the context of the application of what can be described as the “black letter” provisions of the Act. But when it comes to avoidance and the stage‑two Ben Nevis exercise, there can be no sensible objection to looking at a transaction, which is a component of a wider arrangement, in the round, that is, in the context of that wider arrangement.[36] Indeed, that is exactly what happened in Ben Nevis, in which offshore components of the overall transaction were taken into account.
(b)In any event, the conclusion we have reached reflects DHNZ’s assessment of its standalone position. The management accounts of DHNZ and DAP, as described in the documents to which we have referred, show that those responsible for the management accounts of both companies recognised that the substance of what happened in respect of each of those companies, considered separately, is exactly as we have described it.
[36]This is consistent with the approach taken in Alesco New Zealand Ltd v Commissioner of Inland Revenue [2013] NZCA 40, [2013] 2 NZLR 175 at [31]. See also Craig Elliffe “Discerning Commercial and Economic Reality: Applying the GAAR to Frucor” (2021) 27 NZJTLP 223 at 235–236.
The picture which emerges from the planning documents which we have reviewed is clear. The whole purpose of the arrangement was to secure tax benefits in New Zealand. References to tax efficiency in those planning documents are entirely focused on the advantage to DHNZ of being able to offset repayments of principal against its revenue. The anticipated financial benefits of this are calculated solely by reference to New Zealand tax rates. The only relevance of the absence of a capital gains liability in Singapore was that this tax efficiency would not be cancelled out by capital gains on the contrived “gain” of DAP under the forward purchase agreement.
There were many elements of artificiality about the funding arrangement. Of these, the most significant is in relation to the note itself.
Orthodox convertible notes offer the investor the opportunity to receive both interest and the benefit of any increases in the value of the shares over the term of the note. For this reason, the issuer of a convertible note can expect to receive finance at a rate lower than would be the case for an orthodox loan.
The purpose of the convertible note issued by DHNZ was not to enable it to receive finance from an outside investor willing to lend at a lower rate because of the opportunity to take advantage of an increase in the value of the shares. The shares were to wind up with DAP which already had complete ownership of DHNZ. As well, Deutsche Bank had no interest in acquiring shares in DHNZ. Instead, it had structured a transaction that generated tax benefits for DHNZ in return for a fee. Leaving aside the purpose of obtaining tax advantages in New Zealand, the convertible note structure that was adopted had no point.
The other elements of artificiality discussed in the High Court and Court of Appeal judgments are functions of this fundamental artificiality. The most significant of these elements is the way in which the value was attributed to various components of the funding arrangement. As we have explained at [29]–[30], as initially proposed, the starting point for attribution of value was to be the face value of the note calculated by reference to the maximum amount that DHNZ could borrow and still meet the thin capitalisation rules, with the values attributed to the other components (the amount paid under the forward purchase agreement and net contribution from Deutsche Bank) to be worked out backwards from this figure. As it happened, the eventual starting point was the amount paid under the forward purchase agreement. Again, the other figures were then worked out by reference to this different starting point and the applicable interest rate.
As is often the case with tax avoidance, there were substantial elements of contrivance, circularity and cancellation. The artificial features of the funding arrangement to which we have just referred are all indications of contrivance. As between DAP, DHNZ and Deutsche Bank, there was complete circularity as to $60 million of the $149 million paid under the forward purchase agreement.[37] As well, if DHNZ and DAP are treated as a group, there was, in economic substance, complete circularity in relation to the $149 million. As to cancellation, the effect of the forward purchase was to cancel the note save as to the liability for $55 million which was to be discharged by DHNZ in the guise of interest payments on an advance of $204 million. There was also further cancellation in relation to $60 million paid by DHNZ back to DAP as a return of capital; this because it was immediately reversed in the management accounts of both companies, leaving DAP “owing” DHNZ $60 million as unpaid capital.
[37]This treats the payment of $149 million by DAP to Deutsche Bank as funded by the $60 million it received from DHNZ and the $89 million it borrowed from BNP Paribas, a treatment which follows the flow of money.
Against this background, the effect of the arrangement was that DHNZ sought to obtain deductions in relation to $55 million in principal repayments. These are provided for in the Act to meet financing expenses and not repayments of principal. DHNZ was thus claiming deductions for expenses which, in economic substance, it had not incurred. This use of the relevant deduction provisions of the Act lay outside of parliamentary contemplation as to the use of those provisions. It was therefore tax avoidance and the Commissioner was entitled to void the arrangement under s BG 1(1).
There is one final consideration we should mention. In these proceedings the onus of proof was on Frucor Suntory. In other words, it was for Frucor Suntory to show that the funding arrangement did not fall foul of s BG 1. But the only witness called by Frucor Suntory was a Danone group employee who provided linking evidence in relation to documents which he produced but who had not participated in the design and implementation of the funding arrangement. And, as the Court of Appeal noted,[38] very few documents which originated within the Danone group were produced. There was nothing from those involved in the funding arrangement on the Danone group side and nothing apparent on the face of the contemporaneous documents which offered any explanation for the funding arrangement other than tax avoidance. This is perhaps not surprising given the tenor of the planning and post‑March 2003 documents—a tenor that is not susceptible to being easily explained away. That said, an assertion that the funding arrangement was not predicated on tax avoidance could only be credible if it was based on contemporaneous Danone group documentation or evidence from someone involved in the transaction as to a commercial context or effect other than tax avoidance. In the absence of such evidence, the view taken by the Court of Appeal might be thought to have been inevitable.

Reconstruction

[38]CA judgment, above n 6, at [74].

Section BG 1(2) provides:
Reconstruction
(2)Under Part G (Avoidance and non-market transactions), the Commissioner may counteract a tax advantage that a person has obtained from or under a tax avoidance arrangement.
Section GB 1(1) provides:
GB 1 Agreements purporting to alter incidence of tax to be void
(1)Where an arrangement is void in accordance with section BG 1, the amounts of assessable income, deductions, and available net losses included in calculating the taxable income of any person affected by that arrangement may be adjusted by the Commissioner in the manner the Commissioner thinks appropriate, so as to counteract any tax advantage obtained by that person from or under that arrangement, and, without limiting the generality of this subsection, the Commissioner may have regard to—
(a)such amounts of assessable income, deductions, and available net losses as, in the Commissioner’s opinion, that person would have, or might be expected to have, or would in all likelihood have, had if that arrangement had not been made or entered into; …
…
These sections confer broad powers of reconstruction. Where tax avoidance has been established a taxpayer challenging the Commissioner’s reconstruction faces some hurdles, as explained in Ben Nevis:[39]
[171] Furthermore, when taxpayers challenge an assessment based on a reconstruction adopted by the Commissioner, the onus is on them to demonstrate, not only that the reconstruction was wrong, but also by how much it was wrong. Unless the taxpayer can demonstrate with reasonable clarity what the correct reconstruction ought to be, the Commissioner’s assessment based on his reconstruction must stand. This is settled law. …
[39]Footnote omitted.
Frucor Suntory’s challenge to the reconstruction is largely based on counterfactuals. The underlying argument was that DHNZ could have funded the purchase of FBGL and later the refinancing in ways that would have resulted in similar or perhaps greater deductible interest.
As will be apparent, s GB 1(1) provides for, although it does not mandate, counterfactual analysis along the lines of what the tax position would have been had the arrangement not been entered into. Importantly, the generality of the power to adjust is expressly not limited by s GB 1(1)(a). This is made explicit by the phrase “without limiting the generality of this subsection” which precedes s GB 1(1)(a).
We accept that in some circumstances counterfactual analysis may provide the most appropriate basis for challenging reconstruction. But, in this case:
(a)For present purposes, what is critical is that, in economic substance, DHNZ borrowed $55 million from Deutsche Bank which it repaid along with interest. The tax advantage was the deductibility of its principal repayments. The Commissioner was entitled to cancel that tax advantage, which is what she has done.
(b)In any event, none of the counterfactuals proffered by Frucor Suntory replicate the substance of what we have found to be the transaction: repayment of what was owed to Danone Finance by (a) a capital injection by DAP into DHNZ of $89 million and (b) an advance from Deutsche Bank to DHNZ of $55 million.

Shortfall penalties

The legislative scheme

Section 139 of the Tax Administration Act provides:
139 Purposes of this Part
The purposes of this Part are—
(a)to encourage taxpayers to comply voluntarily with their tax obligations and to co-operate with the department; and
(b)to ensure that penalties for breaches of tax obligations are imposed impartially and consistently; and
(c)to sanction non-compliance with tax obligations effectively and at a level that is proportionate to the seriousness of the breach.
Part 9 of the Tax Administration Act provides for penalties for a wide-range of conduct such as late filing of returns and late payment of tax. It addresses in particular:
(a)failure to take reasonable care in adopting a tax position (for which a penalty of 20 per cent of the tax shortfall may be imposed under s 141A);
(b)taking an unacceptable tax position (for which a penalty of 20 per cent of the tax shortfall may be imposed under s 141B);
(c)gross carelessness in taking a tax position (for which a penalty of 40 per cent of the tax shortfall may be imposed under s 141C);
(d)taking an abusive tax position (for which a penalty of 100 per cent of the tax shortfall may be imposed under s 141D); and
(e)evasion or a similar act (for which a penalty of 150 per cent of the tax shortfall may be imposed under s 141E).
In issue in this case is the application of ss 141B (unacceptable tax position) and 141D (abusive tax position).
Section 141B(1) of the Tax Administration Act provides that a “taxpayer takes an unacceptable tax position if, viewed objectively, the tax position fails to meet the standard of being about as likely as not to be correct”. Whether a position meets the “about as likely as not to be correct” standard is determined at the time it was taken.[40] The matters which must be considered include:[41]
(a)the actual or potential application to the tax position of all the tax laws that are relevant (including specific or general anti-avoidance provisions); and
(b)decisions of a court or a Taxation Review Authority on the interpretation of tax laws that are relevant (unless the decision was issued up to 1 month before the taxpayer takes the taxpayer’s tax position).
[40]Section 141B(5).
[41]Section 141B(7).
Section 141D relevantly provides:
(1)The purpose of this section is to penalise those taxpayers who, having taken an unacceptable tax position, have entered into or acted in respect of arrangements or interpreted or applied tax laws with a dominant purpose of taking, or of supporting the taking of, tax positions that reduce or remove tax liabilities or give tax benefits.
…
(4)This section applies to a taxpayer if the taxpayer has taken an unacceptable tax position.
…
(6)A taxpayer’s tax position may be an abusive tax position if the tax position is an incorrect tax position under, or as a result of, either or both of—
(a)       a general tax law; or
(b)      a specific or general anti-avoidance tax law.
(7)For the purposes of this Part … an abusive tax position means a tax position that,—
(a)is an unacceptable tax position at the time at which the tax position is taken; and
(b)      viewed objectively, the taxpayer takes—
(i)in respect, or as a consequence, of an arrangement that is entered into with a dominant purpose of avoiding tax, whether directly or indirectly; or
(ii)where the tax position does not relate to an arrangement described in subparagraph (i), with a dominant purpose of avoiding tax, whether directly or indirectly.
There are two aspects of these provisions which warrant brief comment.
In Ben Nevis, a majority in the Supreme Court commented on the s 141B test:[42]
[184] On its terms this standard does not require that the appellants’ tax position had a 50 per cent prospect of success but, subject to that qualification, the merits of the arguments supporting the taxpayer’s interpretation must be substantial. The stipulation of an objective test means that the taxpayer’s belief that the position taken was correct, or not unacceptable, is irrelevant.

This passage has sometimes been construed as substituting for the statutory language a test of substantiality.[43]

[42]Ben Nevis, above n 12, per Tipping, McGrath and Gault JJ.

[43]See, for example, ASB Bank Ltd v Commissioner of Inland Revenue [2014] NZHC 2184, (2014) 26 NZTC ¶21-098 at [13].

We agree that a mathematical assessment is not appropriate but do not regard substituting “substantiality” for the statutory language as particularly helpful. “[A]bout as likely as not” is an ordinary, perhaps slightly colloquial, English expression with its own nuances of meaning—nuances not necessarily fully captured by the language of substantiality. Indeed, we doubt whether such a substitution was intended in the passage just cited. As we will indicate, when applying s 141B in Ben Nevis, this Court reverted to the statutory language.[44]
[44]Ben Nevis, above n 12, at [203] per Tipping, McGrath and Gault JJ.
Turning now to s 141D, this Court in Ben Nevis rejected an argument for the taxpayers “that the statute required an objective assessment of each [taxpayer’s] dominant purpose in entering the arrangement”.[45] In doing so the Court explained that:
[206] … The proposition that the concept relates to the taxpayer’s mind may on first impression appear arguable, but we are satisfied that there are allied provisions within the section which make it untenable. Our conclusion is that s 141D(7)(b)(i) refers to a tax position that is taken by a taxpayer by means of an arrangement which has a dominant purpose of avoiding tax.
[207] The qualification in s 141D(7)(b), “[v]iewed objectively”, substantially answers the [taxpayers’] argument. It directs attention to features of the arrangement rather than intentions of a taxpayer in taking a tax position linked to the arrangement. Subparagraph (ii) of s 141D(7)(b) reinforces this interpretation in its reference to “an arrangement described in subparagraph (i), with a dominant purpose” (emphasis added). This makes it clear that it is the purpose of the arrangement itself, not the purpose in the mind of the taxpayer, that is referred to in s 141D(7)(b)(i). This aspect of the definition of an “abusive tax position” is concerned with the means employed rather than intentions of taxpayers in taking a tax position. The section requires that the arrangement itself be examined to ascertain its dominant purpose from its terms, irrespective of what may be known or inferred concerning the motives of individual investors.

This approach was criticised by Shelley Griffiths in a 2013 article, “An ‘abusive tax position’”.[46] She argued that in the context of s 141D, “purpose” should be taken to refer to an objective assessment of the taxpayer’s state of mind.

[45]At [206].

[46]Shelley Griffiths “An ‘abusive tax position’” [2013] NZLJ 392.

In most situations, taxpayer purpose, objectively assessed, is likely to coincide with the purpose of the arrangement. Indeed, this is the situation in this case. However, it is possible to conceive of situations where this may not be so. For instance, it sometimes happens that the taxpayer was not privy to all the relevant features of the tax avoidance arrangement. Peterson v Commissioner of Inland Revenue[47] provides one example of this and Commissioner of Inland Revenue v BNZ Finance Ltd provides another.[48]
[47]Peterson v Commissioner of Inland Revenue [2005] UKPC 5, [2006] 3 NZLR 433.
[48]Commissioner of Inland Revenue v BNZ Investments Ltd [2002] 1 NZLR 450 (CA).

The Privy Council, by majority, held that Mr Peterson was entitled to the full deduction.[192] The majority held that the fact that the investment was funded by a non‑recourse loan, which was unlikely ever to be repaid, did not alter the fact that the investors had suffered the economic burden of paying the full amount.[193] Even taking the wider context into account, including the inflated production costs and the recycling of the loan to the lender, the majority considered that the focus is on the party who acquires the assets and not what happens to the funds in the hands of the vendor.[194]
[192]At [54] per Lord Millett, Baroness Hale and Lord Brown.
[193]At [44] per Lord Millett, Baroness Hale and Lord Brown. The majority set out the test under Challenge (PC) (see above at [189]) at [37].
[194]At [40]–[42] Lord Millett, Baroness Hale and Lord Brown.
The minority disagreed. They considered that the non-recourse loan was in fact a device to produce a higher sum to be depreciated; these sums were never received or recorded by the respective production companies as premiums that had to be paid as part of the investors’ acquisition costs.[195] In the minority’s view, it was not possible to ignore what happened to the money once it had left the investors’ hands.[196] In addition, the findings in the Taxation Review Authority meant that in fact the loans had never been made (although the investors were not aware of this and were also not aware of the inflated production costs).[197] The Privy Council minority said that:[198]
The source of the $y the investors allegedly paid was the non-recourse loans. But the loans were never made. … a clearer case can hardly be imagined of an arrangement that has not in fact involved the taxpayer “in the loss or expenditure which entitles him to that reduction” (per Lord Templeman in the Challenge case cited at [37] of Lord Millett’s opinion).

Discussion of Peterson

[195]At [91] per Lord Bingham and Lord Scott.

[196]At [101] Lord Bingham and Lord Scott.

[197]At [95]–[98] per Lord Bingham and Lord Scott.

[198]At [96] per Lord Bingham and Lord Scott. The minority made this comment in relation to the Utu loans, although it also concluded that the loans had not been made in the case of the Lie of the Land either: at [98].

Although the majority of the Privy Council in Peterson purported to consider whether Mr Peterson had suffered the economic burden of the whole payment, they in fact applied an (arguably extreme) legal form over substance approach. Because of the circumstances of the non-recourse loan, the Commissioner’s view of the transaction was well available on the then-current approach to tax avoidance, as is shown by the minority judgment.
The Privy Council majority’s view of the particular transaction can arguably be seen as somewhat of an outlier, even on the then-current approach to tax avoidance.[199] Nevertheless, when Frucor took its tax position and filed its 2006 and 2007 returns, no judicial retreat could have been discernible from the legal substance approach. If anything, it would appear from the Privy Council Peterson majority decision that the pendulum had swung further in the direction of an even more formalistic approach than had been applied up to that point.

Trinity scheme

[199]It may be the case that the Privy Council majority was influenced by the fact that the investors had not been aware of the way in which the scheme as a whole operated. The High Court and the Court of Appeal majority in BNZ Investments were clearly influenced by BNZ Investments’ apparent lack of knowledge as to the so-called downstream transactions: see above at [212] and [214]. However, both the Peterson, above n 136, Privy Council majority (at [34]) and the minority (at [93]) said, contrary to the view of the majority in BNZ Investments (CA), above n 149, that an “arrangement” did not require a meeting of the minds, that the taxpayer need not be party to the arrangement and need not know the details of the arrangement.

The majority in the present case also rely on the High Court and Court of Appeal judgments in Accent Management which relate to the Trinity scheme.[200]
[200]Accent Management (CA), above n 145, was decided before Frucor filed its 2007 tax return but after it had filed its 2006 tax return. Accent Management (HC), above n 145, was decided before Frucor filed its 2006 and 2007 tax returns: see Appendix Three.
The Trinity scheme involved land owned by the subsidiary of a charitable foundation being licensed to a syndicate of investors. The licensees were obliged to plant, maintain and harvest a forestry plantation through a forestry management company. The investors paid $1,350 per plantable hectare for the establishment of the forest and $1,946 per plantable hectare for options over the land, namely, an option to buy the land in 50 years for half of its then market value. This and sundry other payments, including a $50 annual licence fee, sufficed to cover the costs both of acquiring the land and planting and maintaining the forest. The land had been bought for around $580 per plantable hectare.
Despite the upfront payments already covering the cost of the land and planting and maintenance of the forest, the syndicate investors agreed to pay a licence premium of some $2 million per plantable hectare payable in 50 years’ time after the trees were harvested and sold. The syndicate purported to discharge its liability for the licence premium immediately by the issuing of a promissory note redeemable in 50 years’ time. It was unlikely that ultimately the venture would be profitable given the amounts paid already by the investors.[201]
[201]Ben Nevis, above n 92, at [122].
There was also a purported insurance arrangement to cover the risk that the harvest would not yield sufficient income to pay the licence premium. There was an upfront and a deferred insurance premium payable, at $1,307 per plantable hectare and $32,791 per plantable hectare respectively. The insurance provider was controlled by one of the promoters of the scheme and, because of an arrangement with the landowner, did not in fact carry any risk.[202] Some 90 per cent of the upfront premiums paid found their way back to the promoters’ family trusts.[203]
[202]At [148].
[203]At [144].
The investors claimed an immediate deduction for the insurance premium (both the current and deferred portions) and claimed a depreciation deduction of the $2 million licence premium amortised over the 50 years of the arrangement.
All three Courts in the judicial hierarchy held that the Trinity scheme was tax avoidance.[204] All three Courts considered that the taxpayers had taken an abusive tax position in terms of s 141D and imposed 100 per cent shortfall penalties.[205]

High Court decision in Accent Management

[204]Accent Management (HC), above n 145, at [401]; Accent Management (CA), above n 145, at [146]; and Ben Nevis, above n 92, at [1] per Elias CJ and Anderson J and [156] per Tipping, McGrath and Gault JJ.

[205]Accent Management (HC), above n 145, at [402]; Accent Management (CA), above n 145, at [170]; and Ben Nevis, above n 92, at [1] per Elias CJ and Anderson J and [209] per Tipping, McGrath and Gault JJ.

The High Court, noting the uncertainty of the profitability of the forest venture contrasted with the certainty and extent of the deductions as well as the degree of relationship and circularity between payments, concluded that the Trinity scheme was established and structured to achieve the mismatch of the deductibility of the license and insurance premiums.[206] Applying Challenge (PC) and the majority approach in BNZ Investments,[207] it held the arrangement to be tax avoidance.[208]

Court of Appeal decision in Accent Management

[206]Accent Management (HC), above n 145, at [308], [312], and [322].

[207]At [282]–[286] and [305].

[208]At [146], [304] and [306].

The Court of Appeal conducted a review of the caselaw to date on the GAAR. It identified the “scheme and purpose” approach of Richardson J in the Challenge Court of Appeal decision as “current” and of “particular significance”.[209] The Court identified several indicators of tax avoidance such as where transactions are “technically correct but contrived” or where there are “elements of pretence”.[210] The Court said that, although there is no taxation on the basis of “economic equivalence”, issues relating to economic reality are nevertheless important,[211] referencing the approach in Challenge (PC) and quoting the comment of Lord Nolan that:[212]
The hallmark of tax avoidance is that the taxpayer reduces [their] liability to tax without incurring the economic consequences that Parliament intended to be suffered by any taxpayer qualifying for such reduction in [their] tax liability.
[209]Accent Management (CA), above n 145, at [112]. I also discuss this above at n 145.
[210]At [118](a)–(b) citing Challenge (CA), above n 145, at 532 per Woodhouse P; and Challenge PC, above n 124, at 562 per Lord Templeman.
[211]At [118](c).
[212]Inland Revenue Commissioners v Willoughby [1997] 1 WLR 1071 (HL) at 1079. The Court of Appeal in Accent Management commented (at [118]) that the remarks in Willoughby had been cited by the majority in Peterson, above n 136, at [38], as applicable in the New Zealand context.
The Court of Appeal referred with approval to the comments on the proper approach to tax avoidance by Richardson P in BNZ Investments and, in particular, the comments on line drawing.[213] It then examined in some depth the Privy Council decision in Peterson,[214] concluding both the minority and the majority used a similar test and commenting that:
[123] … the difference between the two approaches seems to have come down to a difference of opinion as to whether the investors had, in truth, suffered the pretax economic consequences which were intended by the legislature to be the prerequisite of deductibility.
[213]Accent Management (CA), above n 145, at [119]. The relevant quotes from BNZ Investments (CA), above n 149, are set out above at [215].
[214]At [120]–[123]. Peterson, above n 136, was decided after the High Court decision in Accent Management and before the Court of Appeal decision: see Appendix Three below.
The Court of Appeal thus accepted the test of whether the requisite economic consequences had been suffered was the correct one. It said that in some cases the legislature must have intended to encourage certain types of behaviour.[215] The anti‑avoidance provisions could not be taken to apply in such cases. Likewise, it may be “obvious that the specific tax rules relied on were not intended to confer the tax benefit in issue”.[216] Such cases would probably fail on a proper interpretation of the specific provisions without need to have recourse to the GAAR. The Court went on to say:
[126] Cases which lie in between the two extremes just identified still raise a question of statutory interpretation but one which, in our view, cannot be addressed solely by reference to the specific tax rules relied on by the taxpayer. The relevant general anti-avoidance provisions are also relevant. Given the generality of cases to which specific tax rules necessarily apply, it would be unrealistic to confine the application of general anti-avoidance provisions to transactions which lie outside of a discernible specific legislative purpose. When construing such specific rules and looking for their scheme and purpose, it is necessary to keep general anti-avoidance provisions steadily in mind. On this basis, it will usually be safe to infer that specific tax rules as to deductibility are premised on the assumption that they should only be invoked in relation to the incurring of real economic consequences of the type contemplated by the legislature when the rules were enacted. Further, it also seems reasonable to assume that deductibility rules are premised on a legislative assumption that they will only be invoked by those who engage in business activities for the purpose of making a profit. Further, schemes which come within the letter of specific tax deductibility rules by means of contrivance or pretence are candidates for avoidance. The result in any given case comes down to a question of evaluation, or as Richardson P put it in BNZ Investments, an exercise in “line drawing”.
[215]At [125].
[216]At [125].
Applying the analysis required by the past caselaw, the Court of Appeal held that, because the license fee payable by the taxpayers was “essentially voluntary”, the taxpayers had “not suffered the pre-tax economic burden (as opposed to a technical legal liability) which Parliament intended as the pre-condition of deductibility”.[217] Although there was a business in the “real and tangible sense”:[218]
… the real purpose of the arrangement is not the conduct of a forestry business for profit, but rather generation of spectacular tax benefits. The end result (ie the profitability or otherwise of the venture) was never seen as being material. The corollary of this statement is that there never was a “real” purpose of making a profit from the harvesting of trees.
[217]At [144].
[218]At [140]–[141].
The Court of Appeal concluded that there was thus no economic burden borne and no business purpose. The arrangement was “technically correct but contrived” and “an artifice”.[219] In the light of those conclusions, the Court of Appeal was of the view that the scheme was “well and truly across the ‘line’ referred to by Richardson P in BNZ Investments.”[220]

Discussion of Accent Management

[219]At [144](d).

[220]At [146].

Both the High Court and the Court of Appeal in Accent Management applied the then-current legal substance test to the Trinity scheme, as they were of course obliged to do in terms of the then-current Privy Council authority. There is nothing in the Accent Management judgments therefore that suggests a different approach from the previous authorities, including Auckland Harbour Board.[221] Frucor’s liability for penalties must therefore be judged on the basis of the legal substance approach. Under this approach, it is true taxpayers must incur the economic burden envisaged by the relevant provision in terms of Challenge PC, but this is judged on the basis of legal form and not economic substance.[222]
[221]There were arguably indications in the Court of Appeal judgment that, absent current authority, the Court may have preferred a different approach (see for example at [114]–[115]). But that cannot change the fact the Court was applying the then-current caselaw, including Challenge PC, above n 124, in its analysis and therefore there could have been nothing to alert to any change in the test to be applied.
[222]The majority’s comments at [139]–[142] misunderstand the previous caselaw in this regard and therefore, as noted earlier, the majority seriously understates the fundamental changes brought about by Ben Nevis, above n 92.
The “sea change” of moving to economic substance and contemplation of Parliament[223] occurred at the time of this Court’s decision in Ben Nevis which post‑dated the tax positions taken by Frucor.
[223]Elliffe and Cameron, above n 109.
While the majority acknowledges that the Trinity scheme was “worse” than the arrangement in the case before us, they still suggest that the Court of Appeal decision on the Trinity scheme is important in assessing whether Frucor was taking an unacceptable or abusive tax position.[224]
[224]See the majority above at [131]–[132].
I disagree. The Frucor convertible note involved deductions for actual payments made that, on the majority’s view of economic substance, were partly capital in the guise of interest. But, note that legal and not economic substance was the test at the time Frucor took its tax positions. Auckland Harbour Board, the most analogous case, involved a gift of capital.[225]
[225]See above at [205]–[209].
Neither Auckland Harbour Board nor the present case bear any relationship to the elaborate and artificial Trinity scheme which involved large deferred sums that would never in fact be paid and that had absolutely no business justification. The licence premium purportedly related to forests and land which had in fact already been paid for by the investor,[226] while the deferred insurance premium related to what was in effect an illusory insurance scheme.[227]
[226]See above at [227]–[228].
[227]See above at [229]–[230].
In Frucor’s case, the refinancing was of a debt incurred for the acquisition of a very successful investment by way of the issuing of a convertible note on which interest was actually paid and thus the economic burden suffered.[228] This provides a very stark contrast to the Trinity scheme where, as the Court of Appeal held, there was no economic burden borne (because the deferred payments would never be paid), no commerciality of any kind and effectively no business.[229]
[228]There may have been some contrivance and artificiality involved in the offshore elements of the scheme but under the legal substance test it would have been the economic burden of the interest under the convertible note that was at issue.
[229]See above at [236]–[237].
As Professor Elliffe and Jess Cameron say:[230]
The Ben Nevis scheme in particular was an especially egregious arrangement in many respects. There was a large disparity between the magnitude of the tax advantages claimed, as compared to the economic burden borne. It was also clear from both the statements made by the architect of the scheme, Dr Muir, and the insertion of various artificial steps into the arrangement that this scheme had been very carefully engineered to maximise tax benefits, with little regard for the purported business purpose of the transactions.
[230]Elliffe and Cameron, above n 109, at 443. See also Susan Glazebrook “Statutory Interpretation, Tax Avoidance and the Supreme Court: Reconciling the Specific and the General” (2014) 20 NZJTLP 9 at 24–25.
For all of the above reasons, the majority’s reliance on the fact that penalties were imposed in Ben Nevis is misplaced.[231]

Frucor under the legal substance test

[231]See the majority above at [134].

Applying the legal substance approach, at issue would be the correct treatment of the interest under the convertible note pursuant to the interest deductibility provisions, interpreted purposively. The legal form of the convertible note was a debt instrument until converted.[232] The interest paid on the debt was in fact paid. Under Challenge (PC), the taxpayer (Frucor) therefore had the economic burden of paying the interest and thus suffered the expenditure other taxpayers suffer and which Parliament intended to be suffered by any taxpayer qualifying for a reduction of tax liability.[233] There are few restrictions on interest deductibility and Frucor met all of them.[234] Further, the convertible note was issued to refinance a very successful business acquisition.[235] In these circumstances and, judged by the state of the law at the time, the tax position Frucor took was (at least) about as likely as not to be correct.[236] It follows that Frucor did not take an unacceptable tax position.

Dominant purpose

[232]The need to consider the legal form of arrangements is discussed above at [192].

[233]Challenge (PC), above n 124, at 562 per Lord Keith, Lord Brightman, Lord Templeman and Lord Goff. The majority say above at [130] that it is clear that Frucor (DHNZ in their judgment) did not suffer the economic burden of the amounts paid in interest on the convertible note. On the legal substance approach that is just not correct. The interest was actually paid to Deutsche. Indeed, this is true even on the majority’s economic substance approach. The interest was paid and the economic burden of payment suffered by Frucor. That economic burden actually suffered has just been recharacterised by the majority (and the Commissioner) as partly a payment of capital.

[234]See above at [173]. Frucor’s debt/equity ratio (63 per cent) was well within the allowable levels under the thin capitalisation regime – it could have further increased the convertible note amount up to thin capitalisation limits (75 per cent) to increase its interest deductions. But it did not do so.

[235]See above at [150] and [206].

[236]Tax Administration Act, s 141B(1). Given the state of the law as at the time Frucor filed its returns and, in particular, the legal substance as against the economic substance approach, the result in Auckland Harbour Board and the majority approach in Peterson, I cannot understand the general response of the majority above at [142]. I reiterate that the matter is not viewed through a post‑Ben Nevis lens but in terms of the position at the time the tax returns were filed.

The majority say that the dominant purpose of the arrangement in this case was to reduce the tax liabilities of Frucor.[237] This despite the fact that the whole reason for the restructuring was to ensure that Danone Asia did not incur tax liabilities in Singapore, unlike the position before the refinancing where direct debt funding was provided by Danone Finance.[238] Given that, before the refinancing, Frucor was deducting interest payments roughly equivalent to the amounts it claimed deductions for under the current arrangement, it is difficult to see how its purpose could have been to achieve a result it was already receiving (deductibility of interest) and thus difficult to see its dominant purpose as being to reduce its tax liabilities or to achieve an illegitimate tax advantage in New Zealand.

Credible view of the law

[237]See the majority above at [138].

[238]The majority accept that it was an essential part of the scheme that Danone Asia (called DAP in the majority judgment) not incur tax liabilities in Singapore: see above at [137].

As discussed above, I take a different view of the economic substance of the arrangement from the majority.[239] Even on the majority’s view of economic substance, however, most of the points I make still apply.[240] There is therefore not only a credible view of the facts that would lead to a finding there had not been tax avoidance but also a credible view of the law, even based on the majority’s view of economic substance, that the arrangement was not tax avoidance. In other words, there was a credible view that Frucor was not using the interest deductibility provisions in a manner that was outside the contemplation of Parliament.

Conclusion

[239]See above at [171], in particular, n 110.

[240]See above at n 111 and n 233.

For the above reasons, I do not consider penalties should have been imposed. I would have dismissed the Commissioner’s cross appeal, even if I had taken the same view as the majority on the question of tax avoidance.

Solicitors:
Bell Gully, Auckland for Frucor Suntory New Zealand Ltd

Crown Law Office, Wellington for Commissioner of Inland Revenue

Appendix One: Pre-refinancing diagram

Appendix Two: Post-refinancing diagram

Appendix Three: Timeline of Frucor’s tax returns and relevant judgments

[2022] NZSC 113

30 September 2022