NEW ZEALAND FIRE SERVICE COMMISSION V INSURANCE BROKERS ASSOCIATION OF NEW ZEALAND INCORPORATED CA56/2013 [2014] NZCA 179

NEW ZEALAND FIRE SERVICE COMMISSION V INSURANCE BROKERS ASSOCIATION OF NEW ZEALAND INCORPORATED CA56/2013 [2014] NZCA 179 [13 May 2014]IN THE COURT OF APPEAL OF NEW ZEALANDCA56/2013[2014] NZCA 179BETWEEN NEW ZEALAND FIRE SERVICE COMMISSIONAppellantAND INSURANCE BROKERS ASSOCIATION OF NEW ZEALAND INCORPORATEDFirst RespondentAND VERO INSURANCE LIMITEDSecond RespondentHearing: 7 November 2013 (further submissions received 28 February 2014)Court: Ellen France, Wild and White JJCounsel: A R Galbraith QC and C M Stevens for AppellantR G Simpson and S P Elliott for RespondentsJudgment: 13 May 2014 at 11.30 amJUDGMENT OF THE COURTA The appeal is dismissed.B The declarations made by the High Court, as set out in [45], [49] and [76]of this judgment, are affirmed.____________________________________________________________________REASONS OF THE COURT(Given by Wild J)ContentsPara NoIntroduction [1]A legislative solution is needed [7]Appropriateness of declarations about the split tier policies [11]Split tier policies [16]Indemnity policy [17]Excess of indemnity policy [18]The New Zealand Ports Collective policy [19]Placing slip [20]Policy document [21]Invoices for the premiums [22]Section 48 of the Act [23]Application of s 48 to the split tier policies [26]Analysis of the New Zealand Ports Collective policy [46]Costs [78]Result [82]Introduction[1] In issue on this appeal is the correct method of computing levies under s 48 of the Fire Service Act 1975 (the Act). As most of the funding for New Zealand's Fire Service comes from levies struck under s 48, the issue is important.[2] The appeal is by the New Zealand Fire Service Commission against a judgment given by Heath J in the High Court at Auckland on 17 December 2012. In that judgment, upon the application of the Insurance Brokers Association of New Zealand Inc and Vero Insurance New Zealand Ltd, the Judge made declarations as to the correct method of computing levies under s 48.1[3] As we mentioned, the Commission and, through it, the New Zealand Fire Service, is funded largely by the levies computed and payable under s 48.We will refer to the parties, respectively, as the Commission, IBANZ and Vero.21 Insurance Brokers Association of New Zealand Inc v New Zealand Fire Service Commission[2012] NZHC 3437, (2012) 17 ANZ Ins Cas ¶61-969 [High Court judgment].Under s 48(2) the Executive prescribes the rate of the levies which are payable to the2 Counsel for the Commission advised the Court that approximately 97 per cent of the Commission's funding currently comes from the levies struck under s 48 of the Fire Service Act 1975 (the Act), the balance from Government.Commission by insurance companies on each contract of fire insurance made in New Zealand.3[4] The proceeding brought by IBANZ and Vero, and this appeal by the Commission, turn on the correct interpretation of the computation provisions in s 48, and their application to the two types of fire insurance policy which Heath J was asked to rule on. These are split tier insurance policies and multi-insured composite policies.[5] IBANZ and Vero indicated in a notice on 8 February 2013 that they would be seeking to support Heath J's judgment on an additional ground. Their concern was that the meaning of "indemnity value" was implicit rather than explicit in Heath J's judgment. The notice asserted:The term "indemnity value" means the depreciated replacement cost of insured property, or its current market value, depending on the nature of the property and the purpose for which it is held by the insured.We understood Mr Galbraith QC for the Commission to accept this meaning, though stressing the s 48(6)(c)(i) requirement for objectivity – "a fair and reasonable indemnity value in relation to the replacement value of the property". We also agree with the meaning set out in the notice.[6] As the rate of the levy is fixed by the Executive the Court inquired, before the hearing, whether the Crown ought not to have been a party to the proceeding. Mr Galbraith advised us that the Department of Internal Affairs had been notified of the proceeding and had elected not to join. This is an appropriate point to record concern expressed by Mr Simpson about the impact of the levy penalty provisions, should the appeal be allowed and the declarations set aside.43 The levies are prescribed by the Governor-General by order in council (s 48(2)) and the rate of the levy is reviewed annually by the Minister of Internal Affairs (s 48(3)).In response, Mr Galbraith informed the Court that the Commission would apply any different method of computing levies prescribed by this Court only to future fire insurance contracts.4 The penalty provisions are in ss 53 and 53A of the Act. They provide for interest at 1.5 per cent per month and penalty surcharges on unpaid levies.A legislative solution is needed[7] Early in the argument before us it became apparent that s 48 was drafted before the two types of policy in question became common, and perhaps before the advent of at least the multi-insured composite policy. Counsel differed over the extent to which this is true. But, undoubtedly, s 48 has not kept pace with developments in New Zealand in the structuring of fire insurance policies covering commercial properties. Notwithstanding its apparent purpose, it is fair to say that s 48 is now at breaking point.[8] We make that comment notwithstanding s 6 of the Interpretation Act 1999 and the "ambulatory" or "dynamic" approach to interpretation s 6 supports. The section provides:6 Enactments apply to circumstances as they ariseAn enactment applies to circumstances as they arise.The ambulatory interpretative approach aims to apply the purpose of a statutory provision(s) to the changing circumstances of today. The nature and scope of that approach are well described in Burrows and Carter's Statute Law in New Zealand.5[9] The solution is legislative not judicial. A comprehensive look at how levies to fund the Commission are most effectively and fairly fixed is needed. It has, in fact, taken place. On 5 September 2013 the Minister of Internal Affairs announced reforms to the Act, including to s 48. That announcement followed the publication in December 2012 of a comprehensive report by the Fire Review Panel.That approach does not assist here, because the structure of the two types of policies in issue differs fundamentally from earlier "conventional" policies.6[10] In the meantime, this Court must consider whether it was appropriate for the High Court to make the declarations it did under the Declaratory Judgments Act 1908.5 J F Burrows and R I Carter Statute Law in New Zealand (4th ed, LexisNexis, Wellington, 2009) at 397–405.6 Report of the Fire Review Panel (Department of Internal Affairs, Wellington, 11 December 2012).Appropriateness of declarations about the split tier policies[11] The declarations under appeal in respect of the split tier policies were based on sample (or template) forms of those policies, and not on actual policies. Heath J put to one side the Commission's submission that this was not an appropriate case for a declaratory judgment. He regarded that as a "narrow view" of the circumstances in which declaratory relief may be appropriate, one which was rejected by the Supreme Court in Mandic v The Cornwall Park Trust Board (Inc).7Heath J's view was that the questions of statutory interpretation arising here are suitable for declaratory relief. That accords with the view this Court took inElectoral Commission v Tate.8[12] We describe in [33]–[36] below what occurred on this appeal in respect of the split tier policies. Essentially, although this appeal was argued on two sample split tier policies, the Court post-hearing was asked to look at different sample policies. These events underline the soundness of the principle that a court should not make a declaration if the underlying facts are in dispute, including where the question is one of mixed (but disputed) facts and law. The authorities establishing that principle were collected by White J in Taylor v The District Court at North Shore.9[13] In her judgment in Mandic, the Chief Justice observed:[9] The jurisdiction under the Declaratory Judgments Act enables anyone whose conduct or rights depend on the effect or meaning of an instrument, including an agreement, to obtain an authoritative ruling. Implicit in that observation is that there be no doubt as to which instrument or agreement the court is being asked to rule on. Especially bearing in mind the binding precedential effect of a judgment of this Court under the Declaratory Judgments Act,107 At [5] citing Mandic v The Cornwall Park Trust Board (Inc) [2011] NZSC 135, [2012] 2 NZLR 194 at [5]–[9] and [82].it is not satisfactory to be asked to rule on sample policies – policies that have not actually been issued, and are not complete as to their relevant8 Electoral Commission v Tate [1999] 3 NZLR 174 (CA) at [30]–[37].9 Taylor v The District Court at North Shore HC Auckland CIV-2009-404-2350, 24 March 2010 at [22]. The authorities collected there include New Zealand Insurance Co Ltd v Prudential Assurance Co Ltd [1976] 1 NZLR 84 (CA) and Electoral Commission v Tate, above n 8.10 Declaratory Judgments Act 1908, s 12.contents. Having heard argument based on two such policies, it is even less satisfactory now to be invited to consider other policies. This judgment is accordingly based on the two sample split tier policies that were before us when we heard this appeal on 7 November last.[14] Had we been sitting as a court of first instance, we would probably have declined to make declarations, at least in relation to the split tier policies. That is simply because there was no actual policy or policies to rule on.[15] For the Commission, Mr Galbraith reiterated the Commission's reservations about the appropriateness of declarations. As the appeal was fully argued on the two sample split tier policies we give judgment, though reiterating our concerns about the appropriateness of doing so.Split tier policies[16] We detail the two sample policies in turn.Indemnity policy[17] This policy is headed "Fire (Indemnity) Insurance Policy". Its important features are:(a) Indemnity cover: It provides cover up to a fixed cap which is lower than the indemnity value of the property. Take, as an example, a portfolio of 10 commercial buildings, geographically dispersed through New Zealand. The buildings have an indemnity value of $600 million and a replacement value of $1 billion. These two levels of value are diagrammatically illustrated in the Appendix to this judgment,1111 This was appended by counsel for the respondents to their submissions to illustrate their case.in the two vertical columns on the left hand side. In this example, the indemnity cover is the portion coloured yellow (or lightly shaded). It is capped at $300 million – 50 per cent of the indemnity value of $600 million.(b) Sum insured: The policy contains a sum insured. This sum was not completed in the sample policy in evidence but Mr Simpson advised us that it is always a sum less than the indemnity value of the properties, because otherwise there would be no point in stipulating a sum insured. We comment further on this in [33]–[36] below.(c) Uninsured tier: The policy does not provide cover for that portion of any loss which is above the cap on the indemnity cover but less than the indemnity value of the property. Thus, in the example in the Appendix, there is no cover for a loss or losses above $300 million but less than the $600 million indemnity value of the properties. This uninsured "tier" is white (unshaded) in the Appendix.Excess of indemnity policy[18] This is headed Fire (Excess of Indemnity) Insurance Policy. The important features of this policy are:(a) Cover: Its operative clause states that it provides cover for the excess over the indemnity value of the property insured. The clause further defines the cover as being for that part of the value of the property insured representing its reinstatement value as new, less its indemnity value. The operative clause also states:2. This Policy will not provide any element of cover for the indemnity value of the Property Insured, and in particular will not provide cover for any shortfall in the cover provided by the underlying Fire (Indemnity Value) Insurance Policy due to the existence of a limit on the sum insured in that policy.In the example in the Appendix, this excess of indemnity cover is coloured pink (or heavily shaded).(b) Sum insured: This policy also provides for a sum insured, which was also not completed in the sample policy in evidence. Again, we draw attention to our comments in [33]–[36] below.(c) Tandem policies: This policy contains the following clause:UNDERLYING POLICYThe Underlying Policy is defined as being the Fire (Indemnity) Insurance Policy in the name of the Insured.This Policy shall be subject to the same terms, clauses, conditions and exceptions as such Underlying Policy, other than as varied herein.The Underlying Policy shall be maintained in full effect during the currency of this Policy.Mr Simpson accepted that either both policies are in effect or neither is. We elaborate in [26] below on the consequences of this.The New Zealand Ports Collective policy[19] The important documents evidencing this policy are the placing slip, the policy document itself and the invoices for the premium.Placing slip[20] The placing slip was issued by the broker to the insured. The slip described the insured as "The Port Collective" being a purchasing group comprising the eight port companies "for their respective rights and interests". The eight companies are listed. The slip also stipulated the policy limits which include:• Fire Damage (Indemnity) $250 million aggregate limit• Fire Damage (Excess of Indemnity) $250 millionThe slip sets out the deductibles and general and specific endorsements. Some of these differ between the insured companies. The slip states "premium – as quoted". Finally, the placing slip is dated and is signed by all three insurers.Policy document[21] The policy document is described as "The New Zealand Ports Collective Material Damage/Business Interruption Policy". It starts with a table of contents and is a lengthy document running to about 60 pages. Notable features are:(a) The insured: The policy states:THE INSUREDSouth Port New Zealand LtdEastland Port LtdLyttelton Port Company LtdPort of Napier LtdPort Nelson LtdPort Marlborough New Zealand LtdCentrePort LtdPorts of Auckland LtdThe policy then stipulates:This policy is to be interpreted as if the Policy were issued separately to each of the Insured for their respective rights and interests.(b) Sums insured: The policy states that the sums insured apply to losses incurred "by all Insureds collectively", and reduce by any amount paid. Thus, if a $50 million claim is paid under the Fire Damage (Indemnity) section, the remaining cover reduces to $200 million.(c) Operative clause (payment of the premium): The first part of this states:In consideration of the Insured having paid or promised to pay the required premium and subject to the terms of this Policy, the Company agrees to indemnify the Insured as set out in the Sections of this Policy.(d) Cancellation: The cancellation clause, where the cancellation is by the insured, provides:The Insured may cancel this Policy at any time with immediate effect by giving written notice to the Company. The Company will then retain a pro-rata proportion of thepremium (subject to any adjustment required by the terms of this Policy) for the time during which the Policy has been in force, and will refund the unearned balance to the Insured.(e) Fraud and misrepresentation: The fraud clause in the policy provides:If the Insured or anyone acting on behalf of the Insured makes any claim that is in any respect fraudulent, or makes any false declaration in support of any claim, or uses any other fraudulent means or devices to obtain benefit under this Policy, all benefit in respect of that claim will be forfeited. For the purpose of this condition, each of the Insured (if more than one) will be treated as having been issued with a separate policy.The misrepresentation clause in the policy contains an identical last sentence.Invoices for the premium[22] The total premium, some $2.5 million, was invoiced initially by the broker, Marsh Ltd, to "The Port Collective" at Marsh Ltd. There followed a series of invoices by Marsh, one to each of the eight port companies for that company's share of the total premium. All of these invoices were dated on the same day.Section 48 of the Act[23] It is only s 48(1) to (7) that are relevant to this appeal. Those subsections provide:48 Levy(1) Subject to this Act, every insurance company with which any property is insured against fire under any contract of fire insurance made in New Zealand in respect of any period commencing on or after 1 July 1986 shall pay a levy to the Commission.(2) The Governor-General may from time to time, by Order in Council, prescribe—(a) the rate of the levy that shall be computed at a uniform rate per annum on every motor vehicle which is insured in terms of any contract of fire insurance, whether or not the contract specifies the sum insured; and(b) the rate of the levy that shall be computed on all other property on—(i) the amount for which the property is insured for the period of the contract of fire insurance; and(ii) the period of the contract of fire insurance:provided that where the period of the contract is in respect of any period other than a complete year, the levy shall be calculated as a pro rata proportion of the levy for a complete year.(2A) An Order in Council made under subsection (2) is a legislative instrument and a disallowable instrument for the purposes of the Legislation Act 2012 and must be presented to the House of Representatives under section 41 of that Act.(3) The rate of the levy shall be reviewed annually by the Minister.(4) In reviewing the rate of the levy the Minister shall have regard to—(a) the total amount for which all properties in respect of which the levy is payable are insured at the latest available date, and the likelihood of any increase or decrease in that total amount:(b) the necessity of ensuring that the amounts received by the Commission in respect of the levy in that financial year are sufficient to meet—(i) the requirements of the Rural Fire Fighting Fund; and(ii) the actual net expenditure that, in the case of the Commission, is required to be met by way of the proceeds of levy in terms of section 47:(c) the desirability of ensuring, as far as is reasonably practicable, that any increases or decreases in the rate of the levy are designed to maintain the overall level of stability of the levy in the long term.(5) For the purposes of subsection (4)(b)(ii) of this section, the Minister shall take into consideration any shortfall in the levy as determined in accordance with section 47(3) of this Act.(6) For the purposes of subsection (2)(b) of this section, the amount for which the property is insured for the contract of fire insurance shall be—(a) in the case of residential building as defined in section 2(1) of the Earthquake Commission Act 1993, the amount for which that property is insured pursuant to section 18 of that Act:(b) in the case of personal property as defined in section 2(1) of the Earthquake Commission Act 1993, the amount for which that property is insured pursuant to section 20 of that Act:(c) in the case of other property, where the contract of fire insurance provides for the settlement of any claim for damage to or destruction of the property upon any basis more favourable to the insured person than its indemnity value or where there is no sum insured in the contract, be computed on the basis of the indemnity value of the property as stated by either of the following:(i) a declaration signed by the owner to the effect that the indemnity value declared by the owner for the purposes of the levy is a fair and reasonable indemnity value in relation to the replacement value of the property:(ii) a valuation certificate—(A) given by a registered architect, a valuer registered under the Valuers Act 1948, an engineer with qualifications suitable for the purposes of this Act, or a quantity surveyor possessing qualifications and experience suitable for the purposes of this Act, or a plant and machinery valuer possessing qualifications and experience suitable for the purposes of this Act, being in any case a person who is competent to give such a valuation; and(B) establishing clearly the indemnity value of the property for the purposes of the levy:(d) in any case where the indemnity value cannot be established under paragraph (c) of this subsection, be computed—(i) where the contract specifies the sum insured, on that sum:(ii) where the contract does not specify the sum insured, in the manner determined by the Fire Service Commission.(6A) For the purpose of subsection (6) of this section, the insurance company shall hold for the purpose of audit under section 51 of this Act all declarations and certificates furnished to it under that subsection.(6B) Where the Commission considers that the indemnity value declared in respect of any property by the owner under subsection (6) of this section is not a fair and reasonable indemnity value in relation to the replacement value of the property, the following provisions shall apply:(a) the Commission shall, before the expiry of the contract, determine a fair and reasonable indemnity value and, subject to paragraph (d) of this subsection, the levy shall be computed on the basis of that determination:(b) the Commission shall notify the owner in writing of—(i) its determination; and(ii) the owner's right of objection under paragraph (c):(c) within 28 days after receiving the notice under paragraph (b), the owner may object in writing to the Commission's determination; and every such objection shall be supported by a valuation certificate that complies with subsection (6)(c)(ii):(d) if the owner furnishes a valuation certificate under paragraph (c), the Commission shall be liable to pay 50% of the costs incurred in obtaining the valuation, and the levy shall be computed on the basis of that valuation.(7) This section shall not apply to any contract of fire insurance that is limited to an excess over the indemnity value of the property or to any portion thereof which is in excess of the indemnity value.[24] Section 48 was amended in 1986, presumably following this Court's decision in AMP Fire and General Insurance Co (NZ) Ltd v The Earthquake and War Damage Commission,12 [25]to which we refer in the final bullet point in below. New subsections 6 and 6B – both pertinent to this appeal – were substituted with effect from 1 January 1994. There were inconsequential amendments to the new subsection 6 in 2002 and 2004.[25] As it applies to the two types of fire insurance policy in issue, we interpret s 48 as operating in the following way:• It applies to both types of policy because both are contracts of fire insurance in terms of s 48(1). We need not set out the definition in s 2 of the Act of "Contract of Fire Insurance", since there is no dispute about this first point.• The levy is computed on "the amount for which the property is insured" and the period of the contract of fire insurance: s 48(2)(b).12 AMP Fire and General Insurance Co (NZ) Ltd v The Earthquake and War Damage Commission(1983) 2 ANZ Ins Cas ¶60-529 (CA).• "The amount for which the property is insured" for the contract of fire insurance is to be ascertained pursuant to s 48(6). It is s 48(6)(c) and (d) that are relevant. Section 48(6)(c) provides that the levy is computed on the basis of the indemnity value of the property in two situations:o where the contract provides for settlement of a claim on a basis more favourable to the insured than the indemnity value of the property;13 oro where there is no sum insured.In either of those two situations indemnity value must be fixed by the owner's declaration or the valuation certificate stipulated. Both means aim to ensure the indemnity value is accurate – "is a fair and reasonable indemnity value in relation to the replacement value of the property", to use the words of s 48(6)(c)(i).• Where the indemnity value cannot be established under s 48(6)(c), but there is a sum insured, then the levy is computed on that sum insured: s 48(6)(d)(i). This applies only in the first of the two situations where s 48(6)(c) applies, namely where "the contract of fire insurance provides for the settlement of any claim upon any basis more favourable to the insured person than its indemnity value ".• Where the indemnity value cannot be established and there is no sum insured, then the amount for which the property is insured is determined by the Commission: s 48(6)(d)(ii).• Where the Commission does not accept, as fair and reasonable, the indemnity value declared pursuant to s 48(6)(c)(i), then the procedure stipulated in s 48(6B) applies.13 Counsel were agreed that s 48(6)(c) was directed to the straightforward type of replacement (or reinstatement) policy current when the present s 48(6) was inserted in 1994. Those policies were "more favourable" than indemnity policies and generally did not have a sum insured.• Section 48(7) exempts from levy any contract of fire insurance or any "portion" of such a contract which is limited to cover over the indemnity value of the property. The subsection is ambiguously, and therefore poorly, drafted. Our interpretation accords with Mr Simpson's submission that the "thereof" refers to the contract of fire insurance and not to the "excess over the indemnity value". Thus, the subsection is intended to read:This section shall not apply to any contract or to any part of a contract of fire insurance that is limited to an excess over the indemnity value of the property.There can be little doubt about this. The words "or to any portion thereof which is in excess of the indemnity value" were added in the 1986 amendment, we surmise in consequence of this Court's judgment in AMP.14That case concerned s 14(2B) of the Earthquake and War Damage Act 1944, which was in essentially the same terms as s 48(7). The Court said:15So we conclude without difficulty that in sec. 14(2B) the words "any contract of insurance" include any part of any such contract. In other words sec. 14 does not apply with respect to any contract of insurance (including any part of any such contract) that is limited to an excess over the indemnity value of the property.Application of s 48 to the split tier policies[26] Because the policies operate in tandem or not at all, we consider they are part of the same contract of fire insurance. We think Mr Simpson accepted that. He certainly submitted that it mattered not whether there was one or two or more policies providing for the different tiers of cover.[27] It is obvious from Heath J's judgment that the argument Mr Galbraith put to us differed from that advanced by the Commission in the High Court.16[28] The essential part of Heath J's judgment on the split tier policies is this: 1714 Although there is no confirmation of this in the contemporaneous Parliamentary materials.15 AMP, above n 12, at 78,021.16 Mr Galbraith was not counsel in the High Court.17 High Court judgment, above n 1.[50] In my view, the position for which IBANZ and Vero contend is correct. Section 48(7) is designed to exempt from the scope of the value on which the levy is calculated any amount that is in excess of the "indemnity value" to which s 48(6)(c) refers. In short, where an insured has more cover than the indemnity value the additional part is "excess", for the purposes of s 48(7).[51] As Cooke J suggested in AMP, the kind of contract which Parliament intended to take out of the scope of the levy calculation section was "much more likely to have been simply a contract purporting to give cover in excess of the amount of indemnity insurance". Any type of cover, at whatever level, that is more than the stated insured sum gives "cover in excess of the amount of the indemnity insurance". In my judgment, the words contained in s 48(7) are wide enough to encompass any type of insurance cover for damage by fire that is in excess of the stated "indemnity value", as defined in AMP.(Footnotes omitted.)[29] In summary, Mr Galbraith submitted Heath J had erred because:• The contract of insurance comprising the two split tier policies is within the "more favourable settlement" aspect of s 48(6)(c), and therefore levies are to be computed on the indemnity value of the policy.• Section 48(7) does not apply for two reasons. First, because the contract is not "limited to an excess over the indemnity value of the property". Second, because the exemption only applies where the property is insured up to its replacement value and there is insurance for any loss in excess of that value.[30] Each of these points was rejected by Mr Simpson for the respondents. First, he advised that in practice there was no declaration and no valuation certificate, with the consequence that the indemnity value of the property could not be established. Section 48(6)(d) therefore applied. As we have already noted, Mr Simpson advised us that the split tier policies always specified a sum insured which was less than the indemnity value. The levies were to be computed on the sum insured: s 48(6)(d)(i).[31] As mentioned in [17](b) and [18](b) above, each of the two sample tandem split tier policies contained a sum insured clause. This is the clause in the indemnity policy:THE SUM INSUREDThe sum insured for any one loss and all losses in the aggregate during the Period of Insurance under both this policy and the Insured's Fire (Excess of Indemnity) Insurance Policy is $[32] And this is the one in the excess of indemnity policy:SUM INSUREDThe sum insured for any one loss and all losses in the aggregate during the Period of Insurance under both this policy and the Insured's Fire (Indemnity) Insurance Policy is $[insert amount][33] Having considered those clauses as they apply to the example in the Appendix, the Court on 5 February 2014 issued a minute seeking clarification from counsel for the respondents, given their advice that the sum insured is always less than the indemnity value of the properties.18[34] The Court's minute stated:[2] ... On our reading, when the sample [split tier] policies are applied to the Appendix A example, the sum insured is higher than the indemnity value. We assume the sum insured is $700 million, the total of the indemnity cover (capped at $300 million) plus the excess of indemnity cover (which is $400 million). That is more than the indemnity value given in the Appendix A example, which is $600 million.[35] In their 13 February memorandum in response, counsel for the respondents:• Reiterated their advice that the sum insured "will always be less than the indemnity value of all the property insured", including "where split tier policies contain a combined sum insured for both fire indemnity insurance and separate fire excess of indemnity insurance", as with the two sample split tier policies in the Case on Appeal. If that were not the case, they pointed out the owner would "simply purchase a single replacement policy and provide valuation certificates or declarations".• Submitted the Court's query arises from the incorrect assumption that the relevant "sum insured" in the sample policies comprises the indemnity sum18 This is the advice we record in [17](b) and [30] above.insured (capped at $300 million in the Appendix) plus the $400 million excess of indemnity value cover, giving a total of $700 million.• Suggested the Court's query may also have arisen because the contracts of insurance depicted in the Appendix differ from the two sample policies in the Case on Appeal. Counsel stated:Unusually, the sample policies provide a combined sum insured across both the indemnity and excess of indemnity policies, as follows:The sum insured for any one loss and all losses in the aggregate during the Period of Insurance under both this and the Insured's Fire [(Indemnity)/Excess of Indemnity) Insurance Policy is $[insert amount](Counsel's emphasis.)• Advised that selection of the two sample policies in the Case on Appeal: was in hindsight unfortunate as they contain the less common combined sum insured across the fire indemnity policy and fire excess of indemnity policy, which has caused the confusion reflected in the Court's minute. The other sample policies that were omitted from the Case on Appeal followed the more common approach of including separate sums insured in the fire indemnity policy and fire excess of indemnity policy.• Sought to file the full version of an affidavit produced in the High Court annexing "as additional exhibits the remaining sample policies produced in the High Court".[36] We decline to consider this further material. It was not in the Case on Appeal and not before the Court when the appeal was argued. We are not prepared to try and navigate this material without the assistance of counsel.[37] As to the two sample split tier policies that were in the Case on Appeal, the Court does not accept it has made an incorrect assumption and is confused. Given that neither of the policies had a sum insured, we have taken the indemnity sum insured of $300 million and the excess of indemnity sum (or amount) of $400 million in the Appendix example, yielding a combined (or aggregate) suminsured of $700 million. But for s 48(7), the levies would be computed on that sum: s 48(6)(d)(i).[38] Counsel for the respondents are on firmer ground in submitting the effect of s 48(7) is to exclude, for levy computation purposes, that part of the contract of fire insurance comprising the excess of indemnity policy. That follows from the interpretation of s 48(7) we have set out in the final bullet point in [25] above.[39] We do not accept either of the two bases on which Mr Galbraith submitted s 48(7) does not apply. First, a "portion" of the contract of fire insurance is limited to an excess of the indemnity value. This is the cover in the excess of indemnity policy, which is expressly limited to a loss in excess of the (unidentified) indemnity value of the property. If a claim is made on that policy, the indemnity value of the property will need to be ascertained. But it need not be ascertained earlier, in particular for the computation of levies, because in its absence the levies are computed on the sum insured stipulated in the indemnity policy part of the contract of insurance (that is, in the Appendix example, on $300 million). We understood Mr Galbraith to argue that on the respondents' approach levies could only be computed when a claim or claims were made and the indemnity value of the property needed to be ascertained to see whether the second (excess of indemnity) tier of insurance applied. Because the levies are computed on the sum insured in the indemnity policy that argument falls away.[40] Secondly, we do not accept s 48(7) only engages when the property is insured up to its indemnity value, and applies to any insurance in excess of that indemnity value. The exemption provided by s 48(7) applies to fire insurance limited "to an excess over the indemnity value of the property". Mr Galbraith's submission requires s 48(7) to be interpreted as if it read:This section shall not apply to any contract or to any part of a contract of fire insurance that is limited to an excess over the indemnity value of the property provided the property is insured against fire for its full indemnity value.(Our emphasis.)So to interpret s 48(7) would be impermissibly to re-write it.[41] Mr Galbraith's second submission also runs into two further obstacles. The first is cl 2 of the operative clause in the excess of indemnity policy which is set out in [18](a) above. As Mr Simpson submits, this clause means the excess of indemnity policy does not "reach down" into the white (unshaded) uninsured tier in the example in the Appendix. The second obstacle is inconsistency with the Commission's acceptance that the Act permits a property owner not to insure at all, or to under insure. This is the appropriate point to cite another part of this Court's judgment in AMP:19The reasons pointing to this interpretation are further strengthened by the following considerations. If the interpretation suggested and feared by counsel for the Commission were correct, there could be cases where property is under-insured in an indemnity policy – that is to say, in his suggested terminology, the indemnity "sum" would be less than the indemnity "value" – but where separate replacement policies cover the difference between the indemnity "sum" and replacement cost. On the same interpretation the replacement policies would not be limited to an excess over the indemnity value, so they would not be excluded from the section by sec. 14(2B). It seems unlikely that Parliament would have intended this complication. It also seems unlikely that there would at all commonly be issued a separate replacement policy leaving the insured to bear the difference between the limit in his indemnity policy and the actual value of the property destroyed or damaged. The kind of contract which Parliament meant to take altogether out of the scope of the section, by sec. 14(2B), is much more likely to have been simply a contract purporting to give cover in excess of the amount of indemnity insurance.[42] The relevance of this passage is twofold. First, the split tier policies in evidence here are the very type the Court considered "it unlikely would at all commonly be issued". Secondly, the "kind of contract" this Court refers to in the last sentence of the passage is exactly the type of excess of indemnity policy with which we are dealing.[43] More generally, Mr Galbraith complained Heath J's interpretation gave insured the incentive to insure only for their largest single risk, and conversely removed the incentive to insure the property for its indemnity value. That may be so, but it is not a reason for giving s 48 an interpretation it does not bear. It simply reinforces the point we made in [7] above: s 48 was drafted before the split tier policies in issue were in vogue and is not designed to deal with them. The fresh look at the legislation currently underway is the answer.19 AMP, above n 12, at 78,022.[44] Mr Galbraith readily accepted there is no anti-avoidance provision in s 48. The Commission is not suggesting the split tier policies are a sham. For that reason Heath J was right to look at the true nature of the split tier policies. As Richardson J pointed out, delivering this Court's judgment in Marac Life Assurance Ltd v Commissioner of Inland Revenue:20 at common law there is no halfway house between sham and characterisation of the transaction according to the true nature of the legal arrangements actually entered into and carried out.[45] For the reasons we have set out, which essentially endorse those of Heath J, in respect of the two sample split tier policies before us, we uphold the declarations made by the Judge. We set out those declarations, slightly re-ordered and with added introductory headings:1. Levy computed on sum insured: If a contract of fire insurance provides for the settlement of any claim for damage to or the destruction of the property upon a basis no more favourable to the insured person than its indemnity value, and specifies a sum insured for all claims during the period of the contract of fire insurance that is lower than its indemnity value, the fire service levy payable under s 48(1) of the Act is to be computed on the sum insured.2. Levy computed on indemnity value if sum insured higher: For a policy that:(a) provides cover for the indemnity value of the property; and(b) contains a capped sum insured,20 Marac Life Assurance Ltd v Commissioner of Inland Revenue [1986] 1 NZLR 694 (CA) at 706. This decision has been followed in a significant number of New Zealand authorities. See for example: Commissioner of Inland Revenue v Gulf Harbour Development Ltd [2005] 2 NZLR 162 (CA) at [19]; Tepe Holdings Ltd v Commissioner of Inland Revenue [2011] NZCA 534, (2011) 25 NZTC ¶20-091 at [18]–[31]; Commerce Commission v Fletcher Challenge Ltd [1989] 2 NZLR 554 (HC) at 626; Commissioner of Inland Revenue v New Zealand Refining Co Ltd(1997) 18 NZTC 13,187 (CA) at 13,192; and Chatham Islands Enterprise Trust v Commissioner of Inland Revenue [1999] 2 NZLR 388 (CA) at [17]. That this is the correct analytical approach (in a tax context, in the absence of an allegation of tax avoidance), was confirmed by the Supreme Court in Ben Nevis Forestry Ventures Ltd v Commissioner of Inland Revenue [2008] NZSC 115, [2009] 2 NZLR 289 at [46]–[48].the maximum levy is that computed on the sum insured. However, if the sum insured exceeds the indemnity value of the property, and the insured provides compliant declarations or valuations under s 48(6)(c), the levy is payable on the indemnity value.3. No levy on excess of indemnity value contract: If a contract or any portion of a contract of fire insurance provides for the settlement of any claim for damage to or the destruction of any item of insured property limited to that part of its value in excess of its indemnity value, then pursuant to s 48(7) of the Act, no fire service levy is payable on that contract or portion thereof.4. Application of s 48(7): To be exempt under s 48(7) it is not necessary that the insured hold a policy that insures all or any part of the indemnity value of the property. It is sufficient if the excess of indemnity policy only insures part or all of the difference between the "as new" replacement value and the indemnity value of the property.Analysis of the New Zealand Ports Collective policy[46] The key issues are whether the New Zealand Ports Collective (NZPC) policy described at [19]–[22] above is one contract of insurance or eight. And, if the latter, on what basis levies should be computed under each of the eight contracts.[47] In the Judge's view, analysis of the NZPC policy for the purposes of s 48 of the Act required a "pragmatic commercial approach".21 He cited the judgment of the English Court of Appeal in General Accident, Fire and Life Assurance Corporation Ltd v Midland Bank Ltd.22 That case concerned a policy of fire insurance taken out by the owner, mortgagee (a bank) and tenant (an industrial business) of a building. Delivering the main judgment, Sir Wilfred Greene MR said this of the policy:2321 High Court judgment, above n 1, at [60].22 General Accident, Fire and Life Assurance Corporation Ltd v Midland Bank Ltd [1940] 2 KB 388 (CA).23 At 405–406. There is no joint risk; there is no joint interest; the measure of loss suffered by those two parties will be different, calling for a different measure of indemnity, and, accordingly, it seems to me that there is no joint element about the thing at all.Such a policy, in my judgment, may be more accurately described as a composite policy, because it comprises, for reasons of obvious convenience, in one piece of paper the interests of a number of persons whose connection with the subject-matter of the insurance makes it natural and reasonable that the whole matter should be dealt with in one policy. The description of the insured by name, followed by the words "for their respective rights and interests," in my judgment read in its natural sense, indicates that these three persons, having interests which it is not material to investigate for the purpose of the document, are minded to combine in one policy, and each of them to obtain cover from the underwriters in respect of his right or interest, whatever it may be – and it may vary from time to time. While accepting that the port companies did not have an interest in the same property, as did the three insured in General Accident, Heath J considered they shared the commercial aims of lowering the cost, and extending the scope, of their individual insurance cover. The Judge stated:24[65] In my view, the composite policy must be interpreted as a single policy of insurance. The aggregate sum for indemnity cover against the peril of fire disclosed in the policy ($250 million) is to be regarded as the "indemnity value" for the purposes of s 48(6)(c). [48] Although, in the passage just quoted, Heath J refers uniformly to "the policy" and "a policy", he was obviously expressing the view that the NZPC composite policy was a single contract of insurance. He applied the same reasoning he had applied to the split tier policies, namely that the NZPC composite policy was "a contract by which an insurer undertakes separate and distinct obligations to the various insured".25[49] For those reasons Heath J made the declaration sought by IBANZ and Vero that:The "indemnity value" for the purposes of s 48(6)(c) was the $250 million aggregate sum for indemnity cover against fire.2624 High Court judgment, above n 1.25 At [53]. That description comes from the joint judgment of Mason CJ, Wilson, Dawson and Toohey JJ in Federation Insurance Ltd v Wasson (1987) 163 CLR 303 at 314 [Wasson]. Those Judges cited United States and Canadian authorities which had accepted the concept of a "composite" contract of insurance.26 At [23]. only one fire service levy is payable in respect of the New Zealand Ports Collective policy, which levy is to be quantified on the Sum Insured in Section 1 of the New Zealand Ports Collective Policy.[50] For four reasons, Mr Galbraith submitted Heath J had erred in categorising the NZPC policy as a single contract of insurance. The primary reason is the lack of a joint obligation to pay the premium. Mr Galbraith pointed out that an earlier draft of the policy had contained this clause:(a) the Insured first named in the Schedule will be responsible for payment, on behalf of all Insured, of any premium due or that may become due in terms of this Policy.That clause was obviously deleted because, as issued, the policy contained only the clause set out in [21](c) above.[51] In his evidence Mr D B Thomson, the Managing Director of Vero Specialists Risks Ltd, confirmed the deleted clause "was not considered necessary in this case". He added:27However, in the unlikely event that any member of the Ports Collective failed to pay its share of the premium, Vero regarded all of the members of the Ports Collective to be responsible for any shortfall.Mr Galbraith pointed also to the arrangements for payment of the premium, which we have summarised in [22] above.[52] In support of his primary reason for submitting the NZPC policy was not a single contract of insurance, Mr Galbraith relied on the (separate but concurring) judgment of Gaudron J in the High Court of Australia in Federation Insurance Ltd v Wasson.28 Her Honour started by pointing out that a transaction involving entirely separate obligations owed to and by different persons comprised several different contracts, notwithstanding that it was embodied in one document. But, where some of the promises were joint and others purely several, her Honour affirmed "there is but one contract". Turning to the policy in issue, Gaudron J said:2927 Reply affidavit of Douglas Bruce Thomson, sworn 21 November 2011, at [7].28 Wasson, above n 25, starting at 316.29 At 319. In the present case, it is clear that the obligation by the insured to pay the specified premium was a joint obligation: the premium was specified as a total sum without specification of proportions referable to the interests of the insured parties; it was payable by all or any of the parties, in the sense that payment by one constituted performance by all. The joint obligation as to payment of the premium constituted the policy a single contract, although it effected separate insurances, and in that sense constituted a "composite policy" or "composite contract" as explained in the judgment of Mason CJ, Wilson, Dawson and Toohey JJ.(Our emphasis.)[53] We agree with Gaudron J and apply her approach. The existence of a joint obligation is antithetical to the existence of separate contracts of insurance. We consider the NZPC policy contains some joint obligations, as well as some several ones. Accordingly, it is a single, composite contract of insurance and not eight separate contracts. First, as in Wasson, and contrary to Mr Galbraith's submission, the obligation in the NZPC policy to pay the premium is a joint one. The operative clause we have set out at [21](c) above contains a promise by "the insured" – that is, by the eight port companies together – to pay the premium. That is a joint obligation. There was a single premium. The apportioning of that premium between the eight insured companies was a matter of administrative convenience which did not alter the insured's joint obligation to pay the premium. The insurers gave uncontradicted evidence they were only prepared to offer the policy as a single one to the eight insured collectively. This evidence also came from Mr Thomson:30 Vero invoiced the NZPC for a single aggregate premium This reflected the fact that the insurers regarded all eight port companies as jointly responsible for the premium. Vero was only prepared to offer them a substantial discount on the basis that all eight companies purchased their material damage cover under the NZPC Policy. Marsh then invoiced each port company separately for its share of that premium. [54] Mr Simpson accepted there was nothing in the policy expressly covering the situation of one or more of the port companies not paying its portion of the premium. Given each company was financially sound and obtained significant benefits under the policy, we accept non-payment was not contemplated.[55] Secondly, the NZPC policy could only be terminated by the eight insured acting together. The cancellation clause is set out at [21](d) above. It is "all or30 Second affidavit of Douglas Bruce Thomson for the plaintiff, sworn 12 December 2011, at [13].none" in nature. Thus, cancellation was another joint obligation in the NZPC policy. Again, the position was the same in Wasson, where the High Court of Australia was unanimous in holding that one of the two co-insured was not able unilaterally to terminate the policy, and their attempt to do so ineffective.[56] By contrast, the fraud clause in the policy (it is set out in [21](e) above) imposes on each of the eight insured companies a separate – or several – obligation. This is evidenced by the fact that the fraud clause explicitly states "[f]or the purpose of this condition, each of the Insured (if more than one) will be treated as having been issued with a separate policy".[57] The explanation of a "composite contract" of insurance by Mason CJ, Wilson, Dawson and Toohey JJ referred to by Gaudron J in the passage we have set out in [52] above was this:31By a composite contract we mean a contract by which an insurer undertakes separate and distinct obligations to the various insured. This concept has been accepted and applied in the United States and Canada: Wenig v Glens Falls Indemnity Co; 32 Morgan v Greater New York Taxpayers Mutual Insurance Association;33 Rankin v North Waterloo Farmers Mutual Insurance Co.34[58] Analysing a composite policy as potentially comprising a mixture of joint and several obligations accords with the approach adopted by Eichelbaum CJ in Maulder v National Insurance Company of New Zealand Ltd. 35 Mr Maulder had deliberately set fire to the family home, incinerating himself in the process. The home was registered in the names of Mr and Mrs Maulder as joint tenants. Mr and Mrs Maulder were also the "insured" in the two fire policies. The question for Eichelbaum CJ was whether the innocent Mrs Maulder was debarred from claiming by virtue of her husband's arson. Eichelbaum CJ respectfully disagreed with the result in two earlier New Zealand High Court cases.3631 Wasson, above n 25, at 314.His Honour surveyed United States, Canadian, English and Australian authorities, including the English Court of32 Wenig v Glens Falls Indemnity Co 61 NE 2d 442 (NY 1945) at 445.33 Morgan v Greater New York Taxpayers Mutual Insurance Association 112 NE 2d 273 (NY 1953) at 275–276.34 Rankin v North Waterloo Farmers Mutual Insurance Co (1979) 25 OR (2d) 102 (ONCA).35 Maulder v National Insurance Company of New Zealand Ltd [1993] 2 NZLR 351 (HC).36 Royal Insurance Fire and General (NZ) Ltd v Renata HC Whangarei AP50/90, 12 December 1990 and McQuade v Sun Alliance Insurance Company (1992) 7 ANZ Ins Cas 77,698 (HC).Appeal's judgment in General Accident37 and the Australian High Court's decisions in Wasson38 and Advance (NSW) Insurance Agencies Pty Ltd v Matthews.39 The Chief Justice concluded:406. Following the modern approach in the North American decisions I therefore take the view that in each case the present policies should be construed as composite, severally insuring the respective interests of the insured in the insured property.7. It does not necessarily follow that the obligations (or all the obligations) cast upon the insured are several; see Advance (NSW) Insurance Agencies Pty Ltd v Matthews.41Thus, Mrs Maulder succeeded in her claim to half the value of the property destroyed by the fire.However, as in other cases of several interests insured under a composite policy, my finding relieves [Mrs Maulder] from application of the principle that an insured cannot recover in respect of loss deliberately caused by his own act.[59] We summarise thus far. The inclusion in the NZPC policy of a joint obligation to pay the single premium securing the $250 million indemnity cover is fatal to the Commission's argument that the policy comprised eight separate contracts of insurance, each imposing, on the insured port company, a several obligation to pay its portion of the premium in return for $250 million indemnity cover.[60] As stated in [53] above, we align ourselves with the approach Gaudron J took in Wasson that a joint obligation constitutes a composite policy a single contract of insurance. That is the Australian approach. It is well summarised in the High Court of Australia's subsequent decision in Advance (NSW) Insurance Agencies Pty Ltd v Matthews:4237 General Accident, above n 22.38 Wasson, above n 25.39 Advance (NSW) Insurance Agencies Pty Ltd v Matthews (1989) 166 CLR 606.40 Maulder, above n 35, at 359.41 Advance (NSW) Insurance Agencies Pty Ltd, above n 39.42 Ibid at 619. The Court cites Wasson, above n 25, and also Deaves v CML Fire and General Insurance Co Ltd (1979) 143 CLR 24 at 41 and United Shoe Machinery Company of Canada v Brunet [1909] AC 330 (PC). we should point out that, even in the case of a composite contract, some obligations are joint so that it is impossible to treat the contract of insurance as involving separate contracts.[61] The distinctly different view of a composite policy of insurance taken in the English case law is well captured by Rix J in Arab Bank PLC v Zurich Insurance Co.43 Referring to the Court of Appeal's judgment in New Hampshire Insurance Co v MGN Ltd44It seems to me that the Court of Appeal were saying that in the typical case of a composite policy where there are several assureds with separate interests, the single policy is indeed a bundle of separate contracts. That is the prima facie position under a composite policy, without any need for a meticulous examination, for instance, to see whether separate premiums have been agreed for the various interests. Indeed, one can well understand that on a practical level it would be unrealistic to expect the separate interests to be divided out and severally assessed.Rix J commented:[62] Similarly, the English insurance text Good Faith in Insurance Contractsstates the essence of a composite insurance is: "each assured is insured separately under different contracts of insurance which may be embodied in a single policy". 45This text also asserts each insured has a separate contract with insurers, and that one policy (the policy being the physical document) may combine several contracts.46[63] We mentioned that Eichelbaum CJ, in Maulder, 47 surveyed overseas authorities, including those from Australia and England. This Court undertook a similar exercise in Kelly v National Insurance Company of New Zealand Limited.48[64] We note the analysis of a composite policy by the English Court of Appeal inGeneral Accident has subsequently been questioned by that Court itself, and by two commentators. In American Motorists Insurance Co v Cellstar Corporation theBut in neither case was there a need to choose between the opposing Australian and English views of a composite policy containing at least one joint obligation. And we are not aware of any New Zealand decision that has done that.43 Arab Bank PLC v Zurich Insurance Co [1998] CLC 1351 (QB).44 New Hampshire Insurance Co v MGN Ltd [1996] CLC 1728 (CA) [MGN].45 Peter MacDonald Eggers, Simon Picken and Patrick Foss Good Faith and Insurance Contracts(3rd ed, Lloyd's List, London, 2010) at [11.81].46 At [13.26] and [16.56].47 Maulder, above n 35.48 Kelly v National Insurance Co of New Zealand Limited [1995] 1 NZLR 641 (CA).issue was the correct forum for a dispute under a policy covering global risks.49Mance LJ delivered the Court's judgment. Assuming the law governing the policy adopts the same approach as English law, Mance LJ held:50 the policy is a composite policy, under which the different interests of Cellstar and its subsidiaries are insured, as and when they are at risk in respect of shipments in which they are involved: General Accident.But the Judge then went on to say:51That does not mean that the policy is severable into the series of separate policies or insurances. Cellstar, and those subsidiaries who are party to the contract, are on the face of it jointly responsible for the premium Non-disclosure in relation to any aspect of risk would (at least under English law) render the whole policy voidable. So the policy has unitary aspects, although insurance under it enures to the benefit of those interests in any particular consignments.[65] Later in the judgment, in relation to whether different laws can govern the contract in respect of different parties, Mance LJ said:52The parties by choice can select the law applicable to the whole or a part only of the contract. However, I have no doubt that the present composite policy would fall to be regarded as a single, probably multi-partite, contract. Neither the parties nor the Rome Convention could sensibly be taken to have intended to scissor up the policy negotiated and issued in Houston and to subject different aspects of it to different governing laws.[66] The first article critical of the English approach was written by Mr Birds in response to Arab Bank, and questioned the categorisation of a composite policy as comprising several contracts of insurance.53 Mr Birds suggested Rix J in Arab Bankregarded himself as bound by the Court of Appeal's decision in New Hampshire Insurance Co v MGN Ltd,54 which in turn relied on the House of Lords' decision in P Samuel & Co Ltd v Dumas.5549 American Motorists Insurance Co v Cellstar Corporation [2003] EWCA Civ 206, [2003] 2 CLC 599.However, Mr Birds asserts Dumas is not authority for the proposition, leaving both the MGN and Arab Bank decisions open to question. Mr Birds considered Dumas did no more than "establish, and quite unexceptionally,50 At [12].51 At [12].52 At [20].53 J R Birds "Fraud and composite insurance" (1999) Journal of Business Law 151.54 MGN, above n 44.55 P Samuel & Co Ltd v Dumas [1924] AC 431 (HL).that a fraudulent claim by one co-insured did not affect the validity of a claim by the other co-insured".56 He suggests MGN was correct to hold that the policy covered the separate interests of all the insured companies. But, in his view, it did not necessarily follow there were separate contracts of insurance. He pointed out that the Court of Appeal in MGN did not decide that, indeed declined to do so stating:57Technically one ought to enquire whether there was one contract, or as many contracts as there were companies insured. And if the former, can a contract be avoided for non-disclosure against one or some of the insured, but not against others? We feel that we are relieved from the need to answer those questions by the authority of P Samuel & Co Ltd v Dumas.[67] Mr Birds referred to the following observation in MacGillivray on Insurance Law questioning the correctness of MGN. The comment is:58In [MGN] the CA held that non-disclosure by one assured under a composite insurance did not prejudice the rights of an innocent co-assured so long as their interests were separate. This is open to question, with respect, because (1) the critical issue was whether there were separate contracts not interests, as the court itself said, (2) the authority cited did not concern itself with avoidance of the contract for non-disclosure, and (3) the insurer had been induced to accept the risk as a whole on a false basis regardless of the exact responsibility for the concealment. [68] Mr Birds concludes by questioning whether the courts are perhaps a little too willing to construe a composite policy as containing separate contracts of insurance, rather than being insurance of separate interests under one contract.[69] In a more recent article Brian Harris suggests the approach the English courts take is wrong, and that a composite contract should ordinarily be treated as contained in a single indivisible contract, rather than in multiple contracts.59[70] Relevant to the NZPC policy, Mr Harris suggests: 6056 Birds, above n 53, at 154–155.57 MGN, above n 44, at 1737.58 Nicholas Legh-Jones (ed) MacGillivray on Insurance Law (9th ed, Sweet & Maxwell, London, 1997) at [17-29], fn 79.59 Brian Harris "Insurance Policies for Multiple Insureds: the effect of a composite approach to construction?" (2011) Lloyd's Maritime and Commercial Law Quarterly 393.60 At 399. Where an indivisible premium is offered and accepted in consideration of the underwriting of collective insurable interests of various co-insured within a single policy, the resulting contractual arrangements have all the appearance of being one indivisible contract, which should not be rendered severable by neat legalistic distinctions in insurable interest. The insurance of multiple divergent insurable interests for a single premium can be usefully contrasted with the insurance of a common risk by multiple insurers for the same premium. While the former is not divisible in terms of the consideration offered by the insured parties, the latter is divisible as each co-insurer is only liable pro tanto for the proportion of the risk that it has agreed to underwrite.(Footnotes omitted.)[71] Mr Harris concludes:61It seems extraordinary that the construction of an insurance contract underwritten for multiple insured parties should be allowed to depend, not on the nature of the shared risks underwritten, or on the intention of the contracting parties, but on the extent to which the respective insurable interests of the insured parties are identical. In adopting this stance the composite approach – which regards each separate interest as being separately insured – appears to disregard the fact that in indemnity insurance the assured's insurable interest has little connection with the conclusion of the insurance contract. Furthermore, the composite approach strikes a discordant note with prevailing principles of contract law that are generally applicable to all contracts concluded under the English common law.[72] The other three reasons Mr Galbraith advanced as indicating the NZPC policy is not a single composite contract were:• no joint property interests or interests in a common property or properties;• no insurable interest in the other insured's property; and• no right to indemnity in respect of the other insured's property.[73] We accept all these are features of the NZPC policy. But we do not consider, either singly or together, they have the consequence that the NZPC policy is not a single contract of insurance.[74] We accept that most of the composite policy cases decided in New Zealand, Australia and England involve insured with a joint interest in the same property or61 At 409.properties. The MGN case62[75] Both Potter J in the Commercial Court and the Court of Appeal concluded the policy was composite in nature. Potter J said:is an exception. The companies in the Maxwell 'group' had insured the properties each owned separately under four fidelity policies. The issue was whether fraud by the late Mr Robert Maxwell and some of his associates defeated the interests of other of the companies insured under the policy.63 the words quoted [from General Accident] seem to me equally applicable in this case, where the reasons for dealing with the various parties' insurance within a single policy stemmed not so much from the subject matter of the insurance as the commercial and administrative convenience of the parties.Potter J concluded: "...there was more than one assured and each assured was insured separately".64 The Court of Appeal said the policy was not joint because "the companies that formed the Maxwell group had separate interests to insure, and not a joint interest in the same property".65 The Maxwell case is thus authority that there is no requirement for the insured to have a joint interest in the same property or properties insured under the policy. A composite policy was found to exist notwithstanding that the insurance covered a number of companies within the same "amorphous" group (Potter J had observed "the word 'group' as applied to those companies could be neither technical nor precise",66[76] For those reasons, we endorse Heath J's analysis of the NZPC policy and uphold the declaration he made in respect of it, as set out atit was not a group in any legal or technical sense), each owning separate properties.[49] above.[77] It is, however, worth considering how levies under s 48 would be computed if each port company had a separate contract of insurance. Mr Galbraith submitted the levy on each of the eight contracts would be struck on the $250 million of cover the insured company had at the commencement of the policy. That could only be correct if there existed at commencement of the eight contracts, total indemnity cover of $2 billion (8 x $250 million). Such cover did not exist. Certainly, until any claim62 MGN, above n 44.63 New Hampshire Insurance Co v MGN Ltd [1996] CLC 1692 (QB) at 1715 [MGN QB].64 At 1716.65 MGN, above n 44, at 1737.66 MGN QB, above n 63, at 1698.was accepted, each of the eight insured companies had $250 million of cover. But, as Mr Galbraith readily accepts, there was only one aggregate sum of $250 million indemnity cover at commencement. The Commission's argument that the levies would be calculated on $2 billion of indemnity cover is untenable and was rightly dismissed by Heath J.Costs[78] In his judgment Heath J said this about costs:67[67] While I did not hear specifically from counsel on questions of costs, my provisional view is that the proceeding should be treated as one in the nature of a test case and that no order as to costs should be made.[79] Counsel informed us that no party had sought costs in the High Court.[80] Despite Mr Simpson's submission to the contrary, we are firmly of the view that there should be no orders for costs in this litigation. It is a test case. Invariably, in a test case, the tester is ordered to pay the costs of the party or parties responding, irrespective of the outcome. The rationale for that is that the outcome has a value and importance for the tester, going well beyond the particular case. On the other hand, the responding party or parties have no corresponding wider interest. The question in this proceeding is: who is the tester? It was the respondents who sought declarations in the High Court, and can therefore be categorised as the "testers". Certainly, the outcome had importance and value for both respondents going beyond just this case. But we accept the Commission, at least as the appellant to this Court, but really also in the High Court, can equally be said to be a "tester". Its interest was at least as wide as that of the respondents.[81] In those circumstances, the appropriate costs result is that the parties bear their own costs. Accordingly there will be no order for the costs of this appeal, and there should be none in the High Court either.67 High Court judgment, above n 1.Result[82] The appeal is dismissed.[83] The declarations made by the High Court, as set out in [45], [49] and [76] of this judgment, are affirmed.Solicitors:DLA Phillips Fox, Wellington for AppellantBell Gully, Auckland for RespondentsAPPENDIX