Commissioner of Inland Revenue v BNZ Investments Ltd

IN THE COURT OF APPEAL OF NEW ZEALAND

CA 147/00

BETWEEN	THE COMMISSIONER OF INLAND REVENUE
Appellant

AND	BNZ INVESTMENTS LIMITED
Respondent

Hearing:	13, 14, 15 and 16 November 2000
Coram:	Richardson P Thomas J Keith J Blanchard J Tipping J
Appearances:	J G Fogarty QC, J H Coleman and V L Heine for Appellant A R Galbraith QC, G J Harley and G S Jansen for Respondent
Judgment:	22 May 2001

JUDGMENTS OF THE COURT

Judgments
Para No
Richardson P, Keith and Tipping JJ  [1] -    [58]
Thomas J  [59] - [167]
Blanchard J  [168] - [183]

RICHARDSON P, KEITH AND TIPPING JJ
(DELIVERED BY RICHARDSON P)

Table of Contents
Para No
Introduction  [1]
The statutory provisions  [5]
The facts  [7]
The one arrangement issue: the rival arguments
in the High Court  [17]
McGechan J's conclusions  [21]
Discussion: factual findings  [30]
Arrangement: analysis of s99  [38]
Arrangement: conclusions  [49]
Result  [58]

Introduction

The question for decision is whether 9 redeemable preference share (RPS) investments made by BNZ Investments Limited (BNZI) in the 1989 and 1990 tax years were affected by the anti‑avoidance provisions of s99 of the Income Tax Act 1976.
In each instance the Bank of New Zealand (BNZ) borrowed and funded BNZI, its wholly owned subsidiary, to subscribe for RPS in companies provided for that purpose by Capital Markets Limited (CML), a member of the Fay Richwhite Group. Those investments by BNZI provided a return in the form of dividends from the date of issue of the particular RPS until the date of maturity between 9 months and 3 years later depending on the particular investment. CML utilised the proceeds in complex off‑shore transactions which placed the funds in off‑shore interest bearing deposits. Dividends on the RPS were at rates negotiated between BNZ for BNZI and CML on the basis that no allowance was required for New Zealand tax on downstream income and that the dividends paid on the RPS would be exempt income as inter‑company dividends in the hands of BNZI.
Following a 16 day hearing in the High Court, McGechan J, in a judgment delivered on 10 July 2000 and reported at (2000) 19 NZTC 15,732, held in favour of BNZI and against the Commissioner of Inland Revenue. He held first that on the proper application of s99 the downstream transactions were not part of the arrangement entered into affecting BNZI; second, had there been a single composite arrangement including downstream transactions, those cases where mandatory convertible notes had been employed downstream by CML (the MCN transactions) did not have the requisite purpose or effect of tax avoidance within s99(2) but other cases described as Alasdair/Fenstanton transactions which employed s195 debentures at the equivalent point would be caught by s99(2); and third, had there been a tax avoidance arrangement within s99(2), the adjustments made by the Commissioner relying on s99(3) would have been excessive and should have been limited to assessing the increment in RPS dividends received by BNZI attributable to tax avoidance by entities downstream.
The Commissioner appeals and BNZI cross‑appeals against the Judge's purpose or effect finding in respect of the Alasdair/Fenstanton transactions.

The statutory provisions

At the material times s63(2) relevantly provided that "dividends derived by any company that is resident in New Zealand ... from companies ... shall be exempt from income tax in New Zealand". That exemption from income tax was substantially affected as from 30 July 1991 (and so not affecting the dividends the subject of the Commissioner's assessments) by s63(2C) which excluded the application of s63(2) where the dividend was payable in respect of a fixed rate share as defined in very broad terms in s63(2D). That definition of "fixed rate" included a rate determined by "a fixed relationship to economic, commodity, industrial or financial indices or to banking rates or general commercial rates of interest", or adjusted by a fixed relationship to a rate of income tax or where, having regard to various specified factors, the Commissioner considered "the payment of dividends in respect of the share is equivalent to the payment of interest in respect of money lent".
The immediately relevant provisions of s99 read:
(1)For the purposes of this section--
"Arrangement" means any contract, agreement, plan, or understanding (whether enforceable or unenforceable) including all steps and transactions by which it is carried into effect:
"Liability" includes a potential or prospective liability in respect of future income:
"Tax avoidance" includes--
(a)   Directly or indirectly altering the incidence of any income tax:
(b)   Directly or indirectly relieving any person from liability to pay income tax:
(c)   Directly or indirectly avoiding, reducing, or postponing any liability to income tax.
(2)Every arrangement made or entered into, whether before or after the commencement of this Act, shall be absolutely void as against the Commissioner for income tax purposes if and to the extent that, directly or indirectly,--
(a)   Its purpose or effect is tax avoidance; or
(b)   Where it has 2 or more purposes or effects, one of its purposes or effects (not being a merely incidental purpose or effect) is tax avoidance, whether or not any other or others of its purposes or effects relate to, or are referable to, ordinary business or family dealings,--
whether or not any person affected by that arrangement is a party thereto.
(3)Where an arrangement is void in accordance with subsection (2) of this section, the assessable income . . . of any person affected by that arrangement shall be adjusted in such manner as the Commissioner considers appropriate so as to counteract any tax advantage obtained by that person from or under that arrangement, and, without limiting the generality of the foregoing provisions of this subsection, the Commissioner may have regard to such income as, in his opinion, either--
(a)   That person would have, or might be expected to have, or would in all likelihood have, derived if that arrangement had not been made or entered into; or
(b)   That person would have derived if he had been entitled to the benefit of all income, or of such part thereof as the Commissioner considers proper, derived by any other person or persons as a result of that arrangement.

The facts

The primary focus in the judgment and in the argument of the appeal was on the first question, whether the downstream transactions formed part of a single arrangement affecting BNZI. Much of the detail of the agreed statement of facts, the briefs of witnesses and oral evidence and the exhibits and, as well, the narrative in the judgment, is directed at the downstream structures and flows reflected in comprehensive and complex wiring diagrams. The upstream transactions are relatively simple and straightforward. The downstream are anything but. There is a clear line in terms of overt participation on the part of BNZI between upstream and downstream and in considering that first question it is unnecessary to go into all the detail of the downstream transactions. As well, the report in the New Zealand Tax Cases of McGechan J's judgment amply provides any further detail that may facilitate understanding of that question.
It is sufficient then to refer at this point to the broader background of RPS investments, to the Judge's factual narrative, and then to set out his essential findings of fact and law on this branch of the case.
Over a long period company law and the successive statutes have recognised RPS as a conventional equity means of funding companies. The extent of their practical utilisation has necessarily been affected by the commercial climate of the times. McGechan J explains:
From deregulation of financial markets in 1984 RPS financing became commonplace. It carried taxation advantages. An investor could borrow funds to invest in RPS and deduct interest on the borrowings. Dividends on the RPS, when received, would be tax exempt under the then inter-company dividend exception. The dividends were not deductible to the issuer, but if the issuer had tax losses or received only exempt income, that was not significant. There was room for splitting the tax margin, otherwise payable by the investor, between the investor and the issuer to mutual gain. There was one cloud on the horizon in the late 1980's. It was the Commissioner's Public Information Bulletin PIB163 which recorded doubt whether interest on borrowings by the investor for the purpose of such schemes was deductible. The Commissioner's negative viewpoint was widely regarded as wrong, but probably is the genesis of the habitual extension of issuer indemnities to encompass tax aspects.
[12] BNZ, along with many others, entered into RPS on a widespread basis. The receipt of exempt dividends seems to have carried particular attraction in allowing the reduction of the bank's ratio of tax to receipts. By November 1987 the bank held some $500m so invested.
By the late 1980s BNZ and Fay Richwhite, a merchant bank, had a strong working relationship and BNZ was Fay Richwhite's principal banker. However, they were competitors. McGechan J noted (para [13]):
FR in particular kept its affairs so far as possible to itself, appreciating that the bank could move on opportunities of which the bank became aware. The two organisations did not involve themselves in joint ventures. BNZ remained a financier. There was an exception, to some degree, in the form of the Euronational Group. Euronational was a company incorporated in the Cook Islands by FR and Brierley Investments Limited which the BNZ was invited to join as an equal shareholder. It did so, taking up 28% of the capital, and having a Board representative. This reflected a then viewpoint of the BNZ that it should have an offshore banking capacity to match that of competitor banks.
BNZ built up a large RPS portfolio and at the time of these transactions BNZI had entered into some 57 RPS transactions. None of the other transactions is in question. Following an indication to a CML executive by a BNZ executive of the bank's potential interest in undertaking RPS transactions, and following a meeting or meetings, BNZ indicated by letter of 14 July 1988 its willingness in principle to proceed. McGechan J summed up the position in this way (para [17]):
The parameters were up to $100m for up to 3 years at dividend rates to be agreed, with so-called normal indemnities, including within their scope any tax liability in respect of exempt dividends, disallowance of deductions, and taxation of redemption payments. BNZI was to have the benefit of a secured Put Option to cover repayment. Negotiations, particularly as to the bank's margin, dividend rate, securities, and indemnities, and draft documentation followed. The principle was attractive to BNZ. CML had significant bargaining power accordingly. The dividend rate was agreed at the after tax bank bill rate plus 50% of tax savings generated. Securities, to be made available from a pre-approved stable, were blue-chip. Indemnities, under CML pressure, were kept to a point where BNZI would only make a small margin; and arrangements were reached over conduct of objections. It followed the bank could end up carrying some tax risk. I accept the BNZ did not think there was any significant risk indemnities would be called on. All was normal enough RPS business.
The Judge had earlier noted that CML was developing the MCN transaction template and that the proposed structures would not utilise tax losses or exempt income in the orthodox way. Instead, although not known to BNZ, the transactions would involve conversion of interest received abroad into dividends of an exempt character by movement through tax haven jurisdictions. The process, complicated enough already by the accruals rules brought in in 1987, was about to become further complicated by intended legislation implementing international tax reforms which had been announced in March 1988 to take effect from 1 April 1988 but which had not yet actually been enacted. The structure eventually finalised obviously involved very considerable expertise and effort.
Importantly, the Judge found that Fay Richwhite had good commercial reasons for keeping this structure secret as far as and for as long as possible. It would not have wished to see that effort picked up and used by competitors, including BNZ.
And referring to the first RPS investment offered by CML, the Judge expressly accepted that CML did not disclose any aspect downstream of the issue vehicle CMIL and continued (para [15]):
I accept BNZ had no reason to think the proposal was anything other than commonplace RPS financing utilising tax losses, whether on the part of CMIL or an associated entity. Tax loss companies abounded following the October 1987 crash. BNZI, subject to one exception, did not ask CMIL what CML planned to do with the money. CML was a known securities trader, and BNZ assumed funds raised would be used for that purpose. As noted, CML was a competitor, and secretive. BNZ had a high regard for FR expertise. BNZ was adequately covered by indemnities and security of a high quality. I accept that in principle, and certainly in the context of smaller businesses or suspect security, a bank does generally inquire as to funds utilisation intentions. However this transaction was of the special character noted, and I see nothing suspicious in absence of inquiry by the BNZ.
[16] As I have said, there was one exception. Interestingly, a relatively junior bank executive responsible for negotiations did ask Mr Tompkins of CML what CML would do with the funds raised. I consider this inquiry arose from personal curiosity, and was not an official inquiry on behalf of the bank. Mr Tompkins said something close to the effect it was none of the bank's business. Matters were left at that.
Documentation and settlement followed the agreement in principle and, to return to McGechan J:
[18] ... BNZ, from banking instructions received and actioned, became aware that BNZI payments to CMIL were transferred by CMIL to a European Pacific New Zealand account with the BNZ, then onward by EuroPacific to a Royal Bank of Canada ANZ account. BNZ became aware, to that extent, that the RPS funds were paid to a bank based in the Cook Islands tax haven. The bank could not know one way or the other whether there were intermediate transactions involved, with funds simply held on trust meantime in the EPBC BNZ account. Given the apparent immediate onward payment to the Royal Bank of Canada there would be no particular reason to speculate along those lines. Fortuitously, BNZ did come to learn that along with the transfer by EPBC to the Royal Bank of Canada the Fay Richwhite Group had taken up Royal Bank of Canada notes, as an announcement to that effect was picked up from the Reuters trading screens the next day. ...
[19] BNZI knew nothing of the circuitous character of the route used to fund dividends due from CMIL. It simply received dividends as and when due.
What became the second MCN transaction followed. Next, CML worked up the Alasdair/Fenstanton scheme, deploying s195 debentures. McGechan J found that CML had never informed the BNZ of the character of its downstream structures, and unsurprisingly saw no reason to disclose this change. He continued (para [28]):
In the course of negotiating the dividend rate, CML found it necessary to disclose the fixed security rate it would receive from the Sanwa bank. The BNZ thus knew the destination of the RPS funds (apart from the intended onward placement by Sanwa with HCNZ) and knew CML's gross return. BNZ had no other details of downstream transactions. There were changes to indemnities and indeed a (new) loan agreement document reflecting alteration to fixed rate securities. Kensington Swan shelf companies were used again. The bank again knew that funds received went to an EPBC BNZ account, and then onward to Sanwa at a local Sanwa Westpac account. BNZI knew nothing of the route along which dividends were funded.

The one arrangement issue: the rival arguments in the High Court

The question was whether as affecting BNZI the upstream (issuer and upwards) transactions and the downstream (below issuer) transactions were to be regarded for the purposes of s99 as one arrangement.
The argument for BNZI was that an arrangement must have an element of mutuality, that being drawn from the definition of arrangement in s99(1), from the application of the operative provisions of subs (2) to an "arrangement made or entered into", from the reference in subs (3) to "party" and from dicta in various cases.
The argument for the Commissioner was that the court had at least four policy options:
22.1     First, the Court could consider it sufficient that the RPS deals are interdependent tax-driven transactions built around simple cashflows, and including shared tax advantages. There is no requirement for notice of any risk of avoidance on the part of the investor.
22.2     Secondly, the Court could require knowledge that there would be a plan or structure, and consider that ignorance of its content is no excuse for not including a person within that plan or structure if the content of the plan was left to another party.
22.3     Thirdly, it could require, or rely on as present in any event, notice by the investor of the risk of avoidance.
22.4     Fourthly it could require actual notice of the content which makes the transaction void, before including in the arrangement the steps to which the Objector was a formal participant.
The Commissioner supported 1, 2 and 3 and rejected 4. In discussing the submissions the Judge noted that the Commissioner fairly conceded that there was considerable appeal in an interpretation of s99 which would not apply the section to “innocent” taxpayers, who are in fact affected by avoidance behaviour, but not under notice that they might be deriving advantage from such behaviour, and that there may be some need for notice of risk of avoidance. Next, as to option 2, the Commissioner's submission was that, given that the RPS market was built primarily on exploitation of tax losses, a reasonable person would be put on inquiry as to how dividends would be funded. In the absence of such inquiry such a taxpayer was to be regarded as indifferent to an avoidance mechanism or as relying on the promoter. Then, as to option 3, the Commissioner submitted that, while BNZ may not have known the details of downstream transactions, it knew there was a risk of tax avoidance.

McGechan J's conclusions

Focussing on s99(1), McGechan J concluded that all four elements, contract, agreement, plan or understanding, had one essential common factor inherent in the term "arrangement" itself, namely a conscious involvement by the taxpayer. The concept of conscious involvement and multiple parties inherent in the terms used is reinforced by reference in s99(2) to "arrangement" and to "party thereto". "Arrangements", it seems, have "parties". The reference to "parties" also points to specifically identifiable persons rather than to vague or inchoate groupings. The overall concept is one of conscious involvement by other parties to an end. The Judge preferred to avoid the BNZI term "mutuality" as carrying certain contractual baggage but the concept, he said, was similar.
The Judge saw the authorities cited for BNZI as providing only slender support but they could be viewed as consistent with the conscious involvement test. Turning to policy considerations, McGechan J rejected the proposition advanced by the Commissioner that, s99 being intended to suppress tax avoidance, it should be interpreted in whatever way it does so. Parliament's objective approach envisaged limits, the Judge said, the first being the requirement for an arrangement. That deliberately limited focus is to be respected and advanced, not subverted by expansionist approaches passing beyond the normal meaning of terms used which predicate conscious involvement.

There were, too, in the Judge's view, some real difficulties in a concept which drags a taxpayer within a multi‑step arrangement on a simple basis of taxpayer knowledge. The Judge saw the terminology of s99(1) as stemming from the observations of Lord Denning in Newton v Federal Commissioner of Taxation [1958] AC 450, 465:
... the word 'arrangement' is apt to describe something less than a binding contract or agreement, something in the nature of an understanding between two or more persons — a plan arranged between them which may not be enforceable at law. But it must in this section comprehend, not only the initial plan, but also all the transactions by which it is carried into effect — all the transactions, that is, which have the effect of avoiding taxation, be they conveyances, transfers or anything else.
Consistently with Newton, Parliament in enacting the definition of arrangement confined itself to consensual activity towards an end and that intention should be respected. And, on the threshold question of arrangement, he did not find the fiscal nullity approach in the Ramsay line of cases (W T Ramsay Ltd v Inland Revenue Commissioner [1982] AC 300) helpful. The Ramsay doctrine was not intended to determine the scope of an arrangement within an exhaustive definition as in s99. Also, if invoked, it would not add much of present assistance. Unlike the BNZI case, those cases pointed towards taxpayers who were fully informed, whether in person or through imputed advisor knowledge, in relation to the full chain of transactions, and who actively promoted that entirety.
The Judge concluded that the Commissioner's "composite interdependent preordained transaction" encompassing both upstream and downstream transactions was seriously flawed because of agreed inapplicability to "innocent" transactions. The BNZI theory, based on a requirement for conscious involvement ("mutuality"), provided a narrower and stronger definition which did not require an illogical exception, and had greater attraction accordingly.
McGechan J turned to consider the Commissioner's second and third options (knowledge a structure would exist but ignorance of content; and notice of risk of avoidance). To the extent that those options might rest solely on such knowledge or notice, they were rejected. Arrangement, he said, required conscious involvement. There was a conceptual divide between grounds for suspicion an activity may occur and arranging such activity. And, he said (para [64]):
While to suspect, or to have grounds for suspicion, and even to know, do not in themselves predicate involvement in a “contract, agreement, plan or understanding” , it does not take very much more to move a situation onward to a point where tacit involvement may be found. ... If, for example, there are factual matters which point to an interconnected downstream scheme at risk of avoidance under s 99, and the downstream counterparty is correspondingly justified in assuming the taxpayer is aware of those matters and is comfortable with any such risk, there may be room in some cases for a factual finding the transaction proceeded on the basis of a tacit “understanding” those downstream matters would occur. The situation in that way could move past mere suspicion, or even knowledge, to one of “mutuality” , albeit tacit. The same will follow, of course, in Nelsonian cases of wilful blindness. A taxpayer who deliberately refuses to see the obvious, but proceeds with a transaction in which the obvious occurs downstream, readily enough could be held to be part of at least an “understanding” to that effect. A taxpayer who actually knows all the details, and proceeds nevertheless, is of course at equal or greater risk.
...
In such latter factual situations, there would then be one composite transaction, comprising both upstream and downstream elements. The question will always be highly fact specific.
Turning to the facts the Judge found that BNZI started with two factual features strongly in its favour: (i) the upstream transactions were routine RPS investments, of a type widely considered to fall outside s 99 avoidance; (ii) there were good grounds to believe CML would utilise tax losses in conventional fashion for RPS investments. BNZI started from the position of strength that its own upstream transactions were not avoidance, and that the probable course of whatever downstream transactions CML intended would not involve avoidance. From that perspective, there would not be notice of possible downstream avoidance, and there would be little room for a finding of tacit mutuality.
He then considered whether there were any weaknesses in that initial position. He saw the concern over the risk of avoidance in relation to deduction of interest as between BNZ and BNZI as slight and as related to upstream rather than downstream matters. Second, and in answer to the point suggested that Fay Richwhite was an innovative bank quite capable of putting together downstream avoidance arrangements, his finding on the evidence was that the almost overwhelming likelihood, and indeed the BNZ assumption, was that tax losses would be the shelter utilised. Third, responding to the view expressed by a banker called by the Commissioner that in general banks should make inquiries as to the transaction to be undertaken by the recipient with the money provided by the bank and be able to fully evaluate risk accordingly, he did not regard the absence of inquiry as suspicious, and concluded that inquiries were neither needed, and nor would they have been productive. This was not a case of neglect, or a turning of a blind eye. BNZ could accept it did not need to know downstream proposals beyond the standard assumptions which it made. Fourth, he observed in relation to the evidence that RPS funds had to BNZI's knowledge gone to EPBC, a company operating in the Cook Islands, that BNZI's actual knowledge went no further than immediate payment by EPBC into an apparent ultimate investment in a prime overseas bank. It had no knowledge or reason to suspect the existence of complicated intermediate downstream transactions. Transactions of that character through an offshore bank part owned by Fay Richwhite could have been for reasons other than utilisation of tax haven status and did not by any means necessarily displace the apparent likelihood of primary reliance on tax losses in the normal and approved way.
Finally, the taking of tax indemnities had developed to become an invariable market practice in all RPS transactions and the Judge's finding on the evidence was that the requirement for indemnities in these transactions could not be taken as betraying any particular banking knowledge or concerns in relation to s 99.
On the facts, BNZI was entitled to view the RPS investments as normal in character, very probably turning on tax losses recognised as outside s 99. The features noted such as lack of exact knowledge, an inability for commercial reasons to enquire as to accuracy of understandings, the passage of RPS proceeds through the Cook Islands tax haven company, and tax and general indemnities negotiated, were not such in my view as would or should have created a suspicion of downstream tax avoidance, let alone to a degree and in circumstances capable of amounting to an understanding such would occur. The pointers toward suspicion or knowledge simply do not go far enough to overcome the clear evidence BNZ regarded this as a standard transaction which would fall outside s 99.

Discussion: factual findings

In the light of the factual findings, Mr Fogarty for the Commissioner did not pursue on appeal the third argument advanced in the High Court, namely that BNZI had notice of the risk of avoidance. The Commissioner submitted that, as interdependent transactions including sharing of tax advantages, the RPS deals were part of the arrangement within s99. Alternatively, if knowledge of how the tax advantage was to be obtained was required as part of the plan or understanding, BNZI had left that part of their arrangement to CML and must be taken to have authorised what was done. (See paras [19] and [20] above.)
Before turning to analyse the elements of an arrangement under s99 it is as well to summarise the Judge's material findings of fact relative to any ultimate arguments in respect of the first and second propositions as to BNZI's involvement or complicity in, knowledge of, or suspicion about downstream transactions, and to provide the factual setting for the legal analysis in this case.
The Judge accepted that BNZ had no reason to think CML's proposal which led to the first MCN transaction was anything other than commonplace RPS financing utilising tax losses whether on the part of CMIL or an associated entity. There was nothing suspicious in the absence of inquiry by BNZ into CML's intended utilisation of funds to be invested by BNZI in the RPS. BNZ became aware from banking instructions that the BNZI payments to CMIL were paid to a bank based in the Cook Islands and then onward to a Royal Bank of Canada account, but there was no reason to speculate as to whether there were intermediate transactions involved and BNZI knew nothing of the circuitous character of the route used to fund dividends due from CMIL.
Next, as to the Alasdair/Fenstanton scheme worked up by CML deploying s195 debentures, McGechan J found that CML had never previously informed BNZ of the character of its downstream structures and unsurprisingly saw no reason to disclose that change. He went on to find that BNZ knew that the funds were to be paid to the Sanwa Bank and the fixed security rate it would pay but BNZI knew nothing of the route along which dividends were funded.
After he had discussed the elements of arrangement under s99 and the submissions of counsel in that regard, the Judge returned to factual matters. Importantly, he found that BNZI started with two factual features strongly in its favour. The first was that the upstream transactions were routine RPS investments of a type widely considered to fall outside s99 avoidance. The second was that there were good grounds for BNZI to believe CML would utilise tax losses in conventional fashion for RPS investments. Its own upstream transactions were not avoidance and the probable course of whatever downstream transactions CML intended would not involve avoidance. From that perspective, there was no notice of possible downstream avoidance and there would be little room for a finding of tacit mutuality.
In his further findings of fact, responding to suggested contra‑indications, the Judge concluded that the almost overwhelming likelihood, and indeed the BNZ assumption, was that tax losses would be the shelter utilised; that the absence of inquiry of CML as to the intended utilisation of the funds was not suspicious, and was not a case of neglect or turning a blind eye; that the requirement for indemnities could not be taken as betraying any particular banking knowledge or concern; that the features noted were not such as would have or should have created a suspicion of downstream tax avoidance; and that on the facts BNZI was entitled to view the RPS investments as normal in character and very probably turning on tax losses.
The evidence in the High Court was extensive and detailed. BNZI called their two bank executives particularly concerned with the transactions, a solicitor who was involved in the preparation of the documentation, three expert witnesses who testified as to RPS transactions, including two who had extensive experience in constructing such transactions and, as well, the CML executive who was particularly involved in negotiation with BNZ and BNZI. For the Commissioner, the senior international audit officer of the department and an experienced banker were called as witnesses.
On the argument of the appeal Mr Fogarty for the Commissioner sought to challenge certain of those factual findings and we were taken through parts of the briefs and oral evidence‑in‑chief and cross‑examination of particular witnesses. McGechan J had the advantage as trial Judge of assessing the witnesses and their evidence. Having carefully considered the evidence and counsel's submissions, we are satisfied that there was ample evidence to support the Judge's material findings of fact and we cannot possibly say that he was wrong in the factual conclusions he reached on that evidence.

Arrangement: analysis of s99

In terms of s99(1), "'Arrangement' means any contract, agreement, plan, or understanding (whether enforceable or unenforceable) including all steps and transactions by which it is carried into effect". By s99(2) "Every arrangement made or entered into ... shall be absolutely void ... " if it has a more than a merely incidental tax avoidance purpose or effect "whether or not any person affected by that arrangement is a party thereto". And the adjustment provisions of s99(3) apply to the assessable income of "any person affected by that arrangement".
For the reasons discussed in the cases (e.g. Challenge Corporation Limited v Commissioner of Inland Revenue [1986] 2 NZLR 513, 545), s99 as the expanded successor of the old s108 of the Land and Income Tax Act 1954 is perceived legislatively as an essential pillar of the tax system designed to protect the tax base and the general body of taxpayers from what are considered to be unacceptable tax avoidance devices. By contrast with specific anti‑avoidance provisions which are directed to particular defined situations, the legislature through s99 has raised a general anti‑avoidance yardstick by which the line between legitimate tax planning and improper tax avoidance is to be drawn.
Line drawing and the setting of limits recognise the reality that commerce is legitimately carried out through a range of entities and in a variety of ways; that tax is an important and proper factor in business decision making and family property planning; that something more than the existence of a tax benefit in one hypothetical situation compared with another is required to justify attributing a greater tax liability; that what should reasonably be struck at are artifices and other arrangements which have tax induced features outside the range of acceptable practice - as Lord Templeman put it in Challenge at p562, most tax avoidance involves a pretence; and that certainty and predictability are important but not absolute values.
The function of s99 is to protect the liability for income tax established under other provisions of the legislation. The fundamental difficulty lies in the balancing of different and conflicting objectives. Clearly the legislature could not have intended that s99 should over‑ride all other provisions of the Act so as to deprive the taxpaying community of structural choices, economic incentives, exemptions and allowances provided by the Act itself. Equally the general anti‑avoidance provision cannot be subordinated to all the specific provisions of the tax legislation. It, too, is specific in the sense of being specifically directed against tax avoidance; and it is inherent in the section that, but for its provisions, the impugned arrangements would meet all the specific requirements of the income tax legislation. The general anti‑avoidance section thus represents an uneasy compromise in the income tax legislation.
Line drawing represents the legislature's balancing of the relevant public interest considerations. In terms of s99, that line drawing is directed to three elements, each of which contains its own limits. There must be an arrangement coming within the section. The arrangement must have a more than merely incidental purpose or effect of tax avoidance. And where those two ingredients are present, the assessable income of any person affected by the arrangement is adjusted so as to counteract any tax advantage obtained by that person from or under that arrangement.
The definition of arrangement closely follows the meaning given to the composite expression "contract, agreement or arrangement" in Newton and other decisions under the former s108 and its Australian counterpart, s260 of the Income Tax Assessment Act 1936. In Davis v Federal Commissioner of Taxation (1989) 86 ALR 195, 227 Hill J saw the bilaterality requirement as founded in the very nature of the words of s260, contract, agreement or arrangement. And an arrangement cannot exist in a vacuum. As did the former s108, s99 bites on an "arrangement made or entered into". It presupposes there are two or more participants who enter into a contract or agreement or plan or understanding. They arrive at an understanding. They reach a consensus.
The crucial issue in this case is the extent of the understanding: how much knowledge is required and how and where the line is to be drawn when it is contended that A has left downstream matters to the decision of B. The inquiry is also relevant under s99(3) which provides that, if the arrangement is void against the Commissioner, then any person affected by that arrangement can have his or her income adjusted accordingly.
The words contract, agreement, plan and understanding appear to be in descending order of formality. A contract is more formal than an agreement, and in ordinary usage is usually written while an agreement is generally more formal than a plan, and a plan more formal or more structured than an understanding. And it is accepted in the definition of arrangement that the contract, agreement, plan or understanding need not be enforceable. Section 99 thus contemplates arrangements which are binding only in honour.
In Jacques v Federal Commissioner of Taxation (1924) 34 CLR 348, 359 Isaacs J said that arrangement in s260 meant an arrangement which was in the nature of a bargain but which might not legally or formally amount to a contract or an agreement. And in Bell v Federal Commissioner of Taxation (1953) 87 CLR 548, 573, the High Court of Australia described arrangement as extending beyond contracts and agreements "so as to embrace all kinds of concerted action by which persons may arrange their affairs for a particular purpose or so as to produce a particular effect". Newton v Federal Commissioner of Taxation (para [23] above) is to similar effect. Their Lordships considered arrangement apt to describe something less than a binding contract or agreement, something in the nature of an understanding between two or more persons - a plan arranged between them which may not be enforceable at law. Lord Denning went on in the same paragraph to say that the whole set of words in the section denoted concerted action to an end; (at 455) that the "the whole complicated series of transactions must have been the result of a concerted plan"; (at 467) that looking at the whole of the arrangement, "the whole of the transactions show that there was concerted action to an end"; (and at 468) that the exposition of the law given by the High Court of Australia in Bell was a valuable guide to the true understanding of the section. Similarly, in Rowdell Pty Ltd v Federal Commissioner of Taxation [1963] 9 AITR 177, 194, Kitto J said: "The operation of s260 extends, of course, beyond the arrangement (in the limited sense of the consensus between the parties) to everything done as part of the concerted means adopted for the avoidance of a liability to tax."
The more difficult question under s99 is how and where the line is to be drawn in those cases where it is argued that the understanding extended to giving one party a free hand as to how it would produce revenue of a particular character and whether by legitimate or illegitimate means.
In that regard we do not read the additional words in the definition of arrangement, "including all steps and transactions by which it is carried into effect", on which Mr Fogarty relied, as assisting his argument. Clearly they are concerned with the implementation of the established contract, agreement, plan or understanding. The word "it" in "by which it is carried into effect" refers back to the applicable "arrangement" and does not extend it.

The meaning given to an expression such as "arrangement" or "understanding" in other statutory contexts is no sure guide to its meaning in s99. But it is, we think, helpful to note the approach taken under the Commerce Act 1986. Section 27 of the 1986 Act is directed to contracts, arrangements or understandings substantially lessening competition; and "arrive at" in relation to an understanding is in turn defined as including "reach, and enter into" (s2(1)). The governing words are "enter into a contract or arrangement or arrive at an understanding". In New Zealand Apple and Pear Marketing Board v Apple Fields Ltd [1991] 1 NZLR 257, 261 Lord Bridge said: " 'Arrangement' is a perfectly ordinary English word and in the context of s27 involves no more than a meeting of minds between two or more persons, not amounting to a formal contract, but leading to an agreed course of action". Again, in Commerce Commission v Caltex NZ Ltd (1999) 9 TCLR 305, 314 Salmon J, after citing from Lord Bridge and an earlier dictum of Tipping J in NZ Magic Millions Ltd v Wrightson Bloodstock Ltd [1990] 1 NZLR 731 as to the need for a meeting of minds between the parties to the alleged contract, arrangement or understanding, held that mutuality or meeting of the minds is an essential ingredient of an arrangement or understanding under s27. He also endorsed the observation of Jeffries J in Commerce Commission v Wellington Branch NZ Institute of Driving Instructors (1994) TCLR 19, 24 that "Arrangements and understandings result from an apprehension shared by two and more persons that there will be accord among them as to future acts in a specified area". To similar effect, the Laws of Australia 30.2.52 state: "The essential element of an arrangement and an understanding is a meeting of minds between one or more persons. ... There must be a consensus as to what is to be done."
In our view that reasoning is also applicable under s99. In short, an arrangement involves a consensus, a meeting of minds between parties involving an expectation on the part of each that the other will act in a particular way. The descending order of the terms "contract, agreement, plan or understanding" suggests that there are descending degrees of enforceability, so that a contract is ordinarily but not necessarily legally enforceable, as is perhaps an agreement, while a plan or understanding may often not be legally enforceable. The essential thread is mutuality as to content. The meeting of minds embodies an expectation as to future conduct. There is consensus as to what is to be done.
The justification for construing the concept of arrangement in that way is that it would be inequitable for a taxpayer who enters into an apparently unobjectionable transaction to be deprived of its rights thereunder merely because, unknown to the taxpayer, the other party intended to meet its obligations under that transaction, or in fact did so, in a legally objectionable way. In that regard the effect at common law of illegality in the performance of a contract by one party, where the other party is not implicated, is of some assistance as an indirect analogy. As stated in Chitty on Contracts, 28th ed 17‑011:
But when the contract does not necessarily involve the commission of a legally objectionable act and the legally objectionable intention or purpose of one party is unknown to the other, the latter is not precluded from enforcing the contract. ... The justification for this result is that it would be inequitable for a person who enters into an apparently unobjectionable contract to be deprived of his rights thereunder merely because the other party had an unlawful object in mind in entering into the contract.
In order to avail the Commissioner, the consensus - the meeting of minds - necessary to constitute an arrangement under s99 must encompass explicitly or implicitly the dimension which actually amounts to tax avoidance; albeit the taxpayer does not have to know that such dimension amounts to tax avoidance. Whether there has been a meeting of minds as to what is subsequently done in a particular respect by one party to an arrangement, and whether in answering that question the concept of wilful blindness (discussed by McGechan J - see para [26] above) may provide guidance, will depend on the particular facts.
In assessing the extent to which the relevant minds have met the following considerations may be helpful. One is the assumption which each party may be entitled to make, other things being equal, that the other will act consistently with the justified expectations of the first, in relation to the way their common purpose is to be achieved. An unexpected departure from those expectations should not, without more, be regarded as part of the meeting of minds and hence as part of the arrangement.
On the other hand, a commercially realistic approach should be adopted when assessing the extent of the meeting of minds, particularly in cases where a significant feature of the arrangement is the obtaining, and sometimes the sharing, of tax benefits. Where that feature is present, a court is unlikely to find persuasive the stance of a taxpayer who professes to have had no knowledge or expectation of the mechanism by which the benefit was to be delivered. In such a situation the taxpayer may well appropriately be regarded as having authorised or accepted whatever mechanism was actually used. In such circumstances a consensus could properly be found in respect of the use of that mechanism.
By contrast, if the taxpayer believes on reasonable grounds that the particular and legitimate tax saving mechanism is to be used by the other party, whereas in fact the other party uses a mechanism amounting to tax avoidance, it would be difficult to conclude that the taxpayer had entered into an arrangement extending that far. In such circumstances there would ordinarily be no consensus in respect of the dimension which constituted the tax avoidance. But as we have emphasised the extent of the arrangement entered into by the taxpayer will always depend on the facts of the particular case. That inquiry, of course, precedes consideration of its purpose or effect under s99(2).

Arrangement: conclusions

On the present facts as found by McGechan J, there was no meeting of minds as to what steps or activities CMIL would undertake downstream. CML did not become BNZI's agent entrusted to act on its behalf. CML acted throughout in the course and furtherance of its own business. BNZI did not know CMIL's plans. CMIL would not disclose its intended template. The course actually undertaken by CMIL could not be legally challenged by BNZI. Certainly there was no expectation on BNZI's part or any other basis for knowing that CMIL would or might act illegally or in breach of tax avoidance provisions. There was a natural divide between the upstream and downstream transactions. The upstream arrangement was a standard commercial RPS investment which in terms of the Income Tax Act entitled BNZ to a deduction for interest on the sums borrowed for investment in the RPS and provided that the dividends on the RPS would be exempt income as inter‑company dividends. The RPS investments are a far cry from the self cancelling and circular schemes that have come before the New Zealand and Australian courts under the general anti‑avoidance provisions. On the facts as found by McGechan J, BNZI's arrangement to invest in the RPS and the downstream transactions organised by CML were separate and different arrangements for the purposes of s99.
In the result it is unnecessary to consider other issues raised on the appeal and cross‑appeal, all dependent on a conclusion, which we have rejected, that there was a single arrangement encompassing upstream and downstream transactions.

Result

In accordance with the views of the majority, the appeal is dismissed. Costs in the sum of $20,000 to BNZI together with all reasonable disbursements as fixed, if necessary, by the Registrar.

THOMAS J

Table of Contents

Para No

Introduction 59

The English approach 63

1)   The role of the Privy Council

2)   The cost to New Zealand of the English approach

3)   A world apart – a different perception

Section 99 – objective

64

70
73
80

1)   A public interest enactment

2)   The Courts’ traditional reaction

3)   Back to Parliament’s intent

4)   Parliament is disappointed

5)   The unrealistic perception of certainty

6)   An exaggerated fear of deterring commercial activity

Form over substance

1)   …and a general tax avoidance provision

2)   Macniven

(i)  Lord Tomlin’s dicta reinterpreted
(ii) And now “business substance”

     3) Substance over form

81
84
85
89
91

97
100

101
103
105
109
113

Section 99 – interpretation

   1)   A purposive approach

   2)   An analysis of s 99

   3)   It is the “effect” that counts

   4)   Some specific points refuted
117
117
119
126
129

Section 99 – application 138

   1)   The accepted facts – common ground

   2)   BNZI’s knowledge

   3)   BNZI makes no inquiries

   4)   The upstream/downstream analysis

   5)   The real scope of the arrangement

   6)   The expected use of tax losses

   7)   The ignorance of a taxpayer in
         perspective

   8)   Tax avoidance facilitated

   9)   No inhibition of commercial activity
         in this case

10)   Those who fly blind …

138
143
145
146
150
153

155
157

160
162

Conclusion 166

Introduction

The Commissioner of Inland Revenue’s appeal puts the scope of this country’s statutory general anti-avoidance provision directly in issue. On the facts of this case as found by McGechan J at first instance, this appeal would seem to me to give rise to two key questions:

Does the transaction between BNZ Investments Ltd (BNZI) and Capital Markets Ltd (CML) fall within an “arrangement” for the purposes of s 99, notwithstanding that BNZI was not consciously involved in or aware of the exact nature or details of the tax avoidance transactions to be undertaken by CML; and/or
Is the transaction between BNZI and CML part of the arrangement for the purposes of s 99 having regard to the fact that BNZI thereby obtained a tax advantage which it knew required CML to use a tax shelter to obtain, and that it failed to make any or sufficient inquiries to ascertain the nature and details of the tax shelter to be used?

An affirmative answer to the first question would make any requirement of “conscious involvement” or “mutuality” in the specific transactions avoiding the tax irrelevant for the purpose of determining the scope of the arrangement for the purposes of s 99. The consensus or meeting of minds of the participants would not need to embrace the steps or transactions used to carry the arrangement into effect or an awareness that any such steps or transactions would constitute tax avoidance. The scope of the arrangement, as well as its effect, would be determined objectively. A positive answer to the second question would mean that a transaction incorporating a tax advantage for a taxpayer based on the use of a tax shelter would fall within the scope of the arrangement if the taxpayer failed to ascertain the nature of that tax shelter. No such concept as “negative knowledge” would pertain. In failing to make the necessary inquiries, the taxpayer would be said to have accepted the risk that the tax shelter which it left the other party to determine and implement will involve tax avoidance contrary to s 99.
These questions are essentially questions of statutory interpretation. The sole objective must be to give effect to Parliament’s intent. For this purpose, the Court is required to examine the text of the section in the light of its purpose, the scheme of the statute and, as far as it can be gleaned, the legislative policy. The resulting answers can be said to represent Parliament’s intent.
In pressing these issues, I will respectfully differ from the majority. I do not intend to address the submission pressed by BNZI’s counsel that the steps or transactions which were undertaken by CML do not constitute tax avoidance. As this is a dissenting judgment, it would be futile to do so. I do not, however, feel uncomfortable about proceeding on the assumption that tax avoidance took place, and restricting my opinion to the question of the scope of the arrangement.

The English approach

But the straight-forward approach for which I contend would seem to be precluded by the glosses, concepts, distinctions and doctrines which have been judicially developed in relation to tax statutes. There is, in respect of almost any issue relating to tax, a baggage of judicial rules that impedes an orderly and logical decision-making process. In no other branch of the law is formalism more prevalent. Legalistic requirements and suppositions overlay the plain task of statutory interpretation.

1) The role of the Privy Council

Because it is the highest Court in this country’s judicial hierarchy, the Privy Council bears ultimate responsibility for this approach. The Board’s decisions handed down from London over the years have set the tone and direction of judicial debate in the area of tax law. The Privy Council (or the House of Lords, the permanent members of which also sit on the Board) has dictated the judicial methodology.
This different approach has been acknowledged. In Commissioner of Inland Revenue v Wattie [1999] 1 NZLR 529, the Privy Council upheld a decision of the majority of this Court to the effect that a lump sum incentive payment paid by a lessor to an intending lessee to take up a lease at an inflated rental represented a “negative premium” and was therefore exempt from income tax. The Board quoted at length (at 531-534) the commercial background to the payment as described by Fisher J at first instance, in the course of which the learned Judge identified the premium payment as a payment in lieu of rent. It accepted (at 538) that the premium was commercially, financially and mathematically linked to the rental payments due from the lessee under the lease. Consequently, while the legal form of the payment was a premium, the payment was in substance a rental subsidy. As no consideration moved from the lessee, the notion of a “negative premium” has subsequently been widely perceived as a contradiction in terms and essentially untenable.
Prior to the Privy Council’s decision, the Supreme Court of Canada had unanimously arrived at the opposite conclusion in Ikea Ltd v Canada [1998] 1 SCR 196. In reaching this conclusion the Supreme Court upheld a decision of the Federal Court of Appeal of Canada [1996] 3 CTC 307; 96 DTC 6526. That Court had in turn dismissed an appeal from a judgment of Bowman J [1994] 1 CTC 2140; 94 DTC 1112, holding that a lessee had to include a tenant inducement payment in its income in the year of receipt.
The Supreme Court of Canada is a prominent and respected Court in the common law world. This Court has benefited from its jurisprudence in many areas of the law. How, then, did the Privy Council deal with the Canadian decision? It did so with astonishing abruptness. The Board simply stated (at 539):
Their Lordships would wish to make no comment upon the decision of the Supreme Court of Canada in the Ikea case save to observe that the Canadian Courts appear to have adopted a different approach from that of the Courts of New Zealand and the United Kingdom, and of Their Lordships’ Board. (Emphasis added).
Subsequent to their Lordships’ advice in Wattie, the Privy Council’s decision was expressly rejected by a majority of the High Court of Australia in FCT v Montgomery (1999) 198 CLR 639. The majority were unable to accept the reasoning of the Privy Council and fully explained why that was so (at 670-671).
In Auckland Harbour Board v Commissioner of Inland Revenue (1999) 19 NZTC 15,433, at para [95], I expressed a clear preference for the approach of the Supreme Court of Canada and the High Court of Australia. I could see no reason to adopt any different or particular approach simply because the subject matter is tax. Other instances which are consistent with this reaction may be mentioned. In Peters v Davison (No. 3) (1998) 18 NZTC 14,027, at 14,063, I referred to what has happened in practice with the over-zealous application of the doctrine of form over substance by various corporate taxpayers and their tax advisers. The doctrine has spawned a culture in certain sections of the community and the specialist tax advice industry dedicated to extreme legalism in the application of the doctrine. In Wattie v Commissioner of Inland Revenue (1997) 18 NZTC 13,297, at 13,311, I was critical of the way the so-called doctrine of economic equivalence is applied, and in Colonial Mutual Life Assurance Society Ltd v Commissioner of Inland Revenue (2000) 19 NZTC 15,614, at 15,641 para [125], I confirmed that the doctrine is an extremely flexible and portable concept all too often invoked to exclude recognition of the substance of a transaction or even to avoid a rigorous analysis of the legal arrangement actually entered into. In Wattie, supra, at 13,310-13,311, I also suggested that the “sham or genuine, no halfway house” rule could not withstand scrutiny.

2) The cost to New Zealand of the English approach

I do not doubt that the Privy Council’s approach to tax cases has cost this country inestimable millions of dollars in tax revenue. See Auckland Harbour Board v Commissioner of Inland Revenue, supra, para [95]. The formalism of that approach has provided a fecund breeding ground for dubious schemes to avoid tax. The glosses, concepts, distinctions and doctrines are exploited and have created a commercial environment in New Zealand in which tax avoidance has been a significant feature. The evidence before the Commission of Inquiry into Certain Matters Relating to Taxation, Report of the Wine Box Inquiry, August 1997 (The Wine-Box Inquiry) provided ample evidence of the extent of tax-driven schemes in this country. Inevitably, the tax avoidance industry has thrived on such concepts as form over substance, economic equivalence, the sham or nothing classification, “legal substance” (as distinct from the actual substance), the choice principle, “tax mitigation”, and the like. In the result, the tax base of this country has been hugely eroded and the equitable incidence of tax on its taxpaying citizens correspondingly distorted.
It would appear that no empirical study of the economic cost of tax avoidance in New Zealand has been carried out. (See Robert McLeod, “Tax Avoidance Revisited”, (2000) 6(2), NZJTLP 103, at 104). Related measures, however, confirm that the cost, including the dead-weight loss, would run into billions of dollars. See Report to the Treasurer and Minister of Revenue by a Committee of Experts on Tax Compliance, December 1998 (Tax Compliance Report 1998), at 151-152, 217, and 224-225. One particular avoidance scheme alone had the cumulative effect on the tax revenue of New Zealand costing the state tens of millions of dollars. But as it is incontrovertible that the cost of tax avoidance, as distinct from evasion, has amounted to billions of dollars and represents a sizeable percentage of this country’s gross national product, the calculation of a precise figure is neither here nor there.
I am not, of course, suggesting that the entire cost to the revenue of tax avoidance in this country is attributable to the particular approach of the Privy Council. Some degree of tax avoidance is inevitable, whatever the system or approach adopted. Nor am I unaware that there have been favourable landmarks in the House of Lords’ and Privy Council’s succession of tax cases. WT Ramsay Ltd v Inland Revenue Commissioners [1982] AC 300 and Inland Revenue Commissioners v McGuckian [1997] 1 WLR 991 are notable departures from the legalistic approach which has otherwise been preferred. Overall progress has certainly been made towards a more substantive approach devoid of undue literalism, technicalities or complex concepts. But there is a way to go yet, and the promise of better to come does not alleviate the past and present cost of the English approach to the revenue base of this country.

The Commissioner advanced two arguments against this view. First, he said, as the Bank was admittedly seeking a tax advantage in taking up the redeemable preference shares, it took the risk that the issuer company’s ability to fund the covenanted dividends on those shares might be dependant on the downstream transactions and that those transactions might involve tax avoidance mechanisms. (For present purposes I treat those transactions as in themselves constituting a tax avoidance scheme). The Commissioner’s argument was that a taxpayer who enters into interdependent transactions which may involve tax avoidance is caught by s99 if it transpires that avoidance is in fact being practised, whether or not the taxpayer understood that this might occur. On this argument of the Commissioner, the taxpayer would be liable even if it was the taxpayer’s understanding that there would be no tax avoidance and even if that view were reasonably held by the taxpayer.
The Commissioner’s alternative argument was that where a taxpayer does not know how a tax advantage will be produced but expressly chooses, or must be taken to have chosen, to authorise someone acting on behalf of the taxpayer to procure such an advantage, being indifferent to whether or not what is to occur will involve tax avoidance, that other person is the taxpayer’s agent and the agency will encompass the avoidance mechanisms. The taxpayer is thus a party to the avoidance arrangements and is caught by s99.
I consider that the first of these arguments fails as a matter of law. I accept the Commissioner’s alternative formulation in principle but his appeal must fail on the particular facts of this case as they have been properly found by McGechan J.
It is a fundamental pre-requisite to the use of s99 against a taxpayer that there be a contract, agreement, plan or understanding (the words the legislature chose to use in s99(1) in defining “arrangement”) in which the taxpayer is a participant. This state of affairs cannot exist for the taxpayer unless there has been formally or informally – even if unenforceably – a consensus between the taxpayer and another or others as to what, in general terms, will occur pursuant to the arrangement. The taxpayer does not have to know all the detail or be able to discern exactly how the arrangement will avoid tax by producing the illegitimate tax advantage, by which I mean an advantage which the legislature cannot have contemplated as flowing from the legislation. But the taxpayer must at least have a broad appreciation of the character of what is occurring.
Hence the taxpayers in Hancock v Federal Commissioner of Taxation (1961) 108 CLR 258 were liable because they were found to have agreed to a plan which they knew had an effect which could otherwise have been produced only by their retaining their shares, receiving the dividends free of tax and applying most of the money in purchasing the shares of their partners after the dividends had been paid. As Kitto J said (at p291), the arrangement was, therefore, a means for avoiding the income tax which the taxpayers would have been liable to pay if they had achieved the same results without an arrangement.
And in Federal Commissioner of Taxation vGregrhon Investments Pty Ltd (1987) 87 ATC 4,988, a case much relied upon by counsel for the Commissioner, though the taxpayers themselves may have professed ignorance of the scheme, the Court plainly considered that their accountant and bank manager were aware that the structuring of the deal was done for tax reasons. As in Hancock, there was no way, other than a tax avoidance arrangement, whereby the advantage sought could be achieved. The remarks of Fisher J (at p4996), highlighted by counsel for the Commissioner, concerning the significance, viewed objectively, of the purpose and effect of an arrangement and the undesirability of allowing a taxpayer to point to and rely upon naivety, ignorance and calculated abstention from knowledge, do not in my view take the matter any further. Indeed Fisher J immediately referred to circumstances which should have alerted the taxpayers and their accountant, a tax specialist, to a scheme of tax avoidance.
As the principal judgment records at para [42], there are three successive inquiries. The first is as to the extent of the arrangement; the second is as to whether it has the purpose or effect of tax avoidance and the third, which arises only where the second is answered affirmatively, is as to the adjustment to be made to counteract the tax advantage. The adjustment can be made against both a party to the arrangement and a person affected, who is not necessarily a party. But it can be made only where a tax advantage has been obtained “under that arrangement”. The Commissioner therefore cannot make an adjustment as against someone who is not a party merely because that person has received a payment subsequent to the operation of an arrangement but outside the arrangement. BNZI received payments funded by means of the tax advantage obtained by the Cook Island entities under the downstream avoidance arrangement, but the dividends paid to it were not obtained under that arrangement.
It is said for the Commissioner that unless his first argument is accepted the promoters of tax structures will be able to insulate their customers from allegations of participation in tax avoidance by ensuring that they remain in ignorance of the mechanism whereby a tax advantage is delivered. That concern, if it were realistic, is not I think a reason for departing from the language of s99, which defines an arrangement in terms of a meeting of minds. But in any event I consider that the concern is exaggerated. Taxpayers are not often likely to be found to have been willing to part with large sums of money and to incur the risk of an assessment and substantial penalties on the basis of a prospective tax advantage without their advisers first having gained a sufficient understanding of what is to occur so as to feel comfortable with it. Professions of ignorance about the existence of tax avoidance are likely to be treated with the same scepticism as in the Australian cases just mentioned, especially where what occurs has no business, professional or family purpose of the taxpayer or the final position could not have been achieved by any legitimate means. The taxpayer carries the burden of proof. It will be for the taxpayer to prove that it genuinely and reasonably had an understanding that the person or persons with whom it was dealing did not intend to practice tax avoidance from which the taxpayer would directly or indirectly benefit.
Furthermore, there is an additional protection for the Revenue. The party which undertakes the formation of the scheme may itself attract a liability for taxation, as it appears that certain of the Cook Island entities did in the present case, where the Commissioner missed his opportunity of assessing them.
The Judge has found that the Bank satisfied him on the evidence that it had no knowledge of what was occurring in the downstream transactions. My survey of the relevant evidence does not persuade me that he was wrong. It was open to him to conclude that the Bank believed that Capital Markets Ltd would fund its special purpose company to the extent necessary to pay the dividends on the redeemable preference shares by reducing its own taxation (on a group basis) by legitimate utilisation of tax losses. Because of the provision of prime securities and the Bank’s general confidence in the integrity of the Fay Richwhite group, which is to be judged in the light of the Bank’s knowledge at the time and not on the basis of information which has subsequently become known, the Bank was content to proceed on the basis of limited information. It made no stipulation concerning what use was to be made of its share subscriptions because it held a put option from Capital Markets Ltd, which was the economic equivalent of a guarantee, and in the event of default it had an entirely satisfactory means of resort to the securities. It came to know, after the first transaction, where the money was finally placed, but the Judge was well entitled to hold that such knowledge and an awareness of some form of involvement of an associated bank in the Cook Islands (European Pacific) was not sufficient to alert the Bank to the presence of an illegitimate tax scheme when it was considering subsequent transactions.
The tax advantage which the Bank sought from the redeemable preference shares appeared to be achievable legitimately. Redeemable preference share transactions, structured as this one was in the upstream transaction, were relatively common-place. So far as the Bank was concerned, the structure consisted of and went no further than the subscription for the preference shares, the put option and the securities. (The securities were part of the upstream transaction, being provided at the same time as the share subscription moneys were paid.) It is not suggested for the Commissioner that such an arrangement, of which there were many in the market place at the relevant time, in itself could constitute a tax avoidance arrangement within s99. The transaction merely sought to take legitimate advantage of the provision by virtue of which tax was not assessable in the hands of a company on dividends received from another company. There was nothing unusual in the use of a special purpose company and the provision of a put option. The former was necessary to insulate the supplier of funding from the business risks of Capital Markets Ltd; the latter was required to avoid falling foul of s62 of the Companies Act 1955. The giving of a tax indemnity for each transaction was a sensible and normal precaution in view of the uncertainty for the time being created by the Commissioner’s stated view, which he later changed, that redeemable preference share dividends were liable for tax in the hands of a recipient company. The Bank was aware that Capital Markets would take the position that what it might choose to do with the money was not to be dictated by the Bank and was a commercial secret. When Mr Turley made his informal inquiry of Mr Tompkins he got exactly that kind of response. Reasonably enough, I consider, secrecy in this context was not taken as indicative of the existence of an illegitimate scheme. The Bank has discharged the burden of showing that it did not believe that there would be any tax avoidance and that it had good reason to so believe.
By the same token, there is no proper basis for the suggestion by the Commissioner, in his alternative argument, that the Bank engaged the Fay Richwhite group as its agent to construct a scheme consisting of the whole of the upstream and downstream transactions, with the Bank being indifferent to whether tax avoidance was involved and being prepared to accept the advantages if it was. The Bank gave no authority to Capital Markets Ltd to formulate any downstream transactions on its behalf. Certainly it made no inquiry about what steps might come between its own subscription for shares in each special purpose company and the placement of the money at its final destination but, as I have mentioned, it had an acceptable reason for taking that attitude.
In the recent decision of the Privy Council in Commissioner of Inland Revenue v Auckland Harbour Board (24 January 2001, No 30 of 2000), delivered by Lord Hoffman, the Board remarked that some of the work of a provision like s99 has now been taken over by “the more realistic approach to the construction of taxing acts exemplified by WT Ramsay Ltd v Inland Revenue Commissioner [1982] AC 300 although their Lordships should not be taken as casting any doubt upon the usefulness of such tax avoidance provisions as a long stop for the revenue”. [Emphasis added].
In view of this dictum, which, it should be noted, appears to be in conflict with views expressed by the High Court of Australia (John v Federal Commissioner of Taxation (1989) 166 CLR 417) and the Supreme Court of Canada (Stubart Investments Ltd v The Queen (1984) 10 DLR (4^th) 1), I have considered the possible application to the present case of the fiscal nullity doctrine, although the Commissioner has very understandably not sought to rely upon it. As the Board says, it is a method of construction. It is quickly apparent, however, that the approach taken in the fiscal nullity cases would not have been of assistance to the Commissioner on the present facts. The doctrine has been applied in the United Kingdom where a series of transactions is preordained or composite so that they constitute a single and indivisible whole or a “unitary arrangement” as it is called in O’Neil v Commissioner of Inland Revenue (Privy Council, 10 April 2001, No 40 of 2000). It proceeds on the basis that the legislature intended tax to be imposed by reference to the business substance of the composite of the transactions. Accordingly, artificial steps inserted for a tax purpose and having no commercial purpose are disregarded in applying particular taxing provisions (Macniven (Inspector of Taxes) v Westmoreland Investments Ltd [2001] 2 WLR 377). But, as Lord Keith of Kinkel said in Inland Revenue Commissioners v Fitzwilliam [1993] 1 WLR 1189,1204, it must from a construction point of view be possible realistically and intellectually to treat a series of transactions as one composite whole. That cannot be done in the present case, for the reasons I have given in relation to s99. Even if it could, there might be a difficulty in applying the doctrine so as to enable the dividend income received by the Bank to be taxed in its hands.
I agree that the appeal should be dismissed.

Solicitors
Crown Law Office, Wellington, for appellant
Rudd Watts & Stone, Wellington, for respondent

Table of Contents
Para No
Introduction	59
The English approach	63
1) The role of the Privy Council 2) The cost to New Zealand of the English approach 3) A world apart – a different perception Section 99 – objective	64 70 73 80
1) A public interest enactment 2) The Courts’ traditional reaction 3) Back to Parliament’s intent 4) Parliament is disappointed 5) The unrealistic perception of certainty 6) An exaggerated fear of deterring commercial activity Form over substance 1) …and a general tax avoidance provision 2) Macniven (i) Lord Tomlin’s dicta reinterpreted (ii) And now “business substance” 3) Substance over form	81 84 85 89 91 97 100 101 103 105 109 113
Section 99 – interpretation 1) A purposive approach 2) An analysis of s 99 3) It is the “effect” that counts 4) Some specific points refuted	117 117 119 126 129
Section 99 – application	138
1) The accepted facts – common ground 2) BNZI’s knowledge 3) BNZI makes no inquiries 4) The upstream/downstream analysis 5) The real scope of the arrangement 6) The expected use of tax losses 7) The ignorance of a taxpayer in perspective 8) Tax avoidance facilitated 9) No inhibition of commercial activity in this case 10) Those who fly blind …	138 143 145 146 150 153 155 157 160 162
Conclusion	166

[2001] NZCA 184

22 May 2001