Phoenix Life Assurance Ltd v The Financial Services Authority (FSA)

In these proceedings the claimant, Phoenix Life Assurance Limited (“Phoenix”), seeks a declaration about the proper interpretation of a form of single premium with-profits policy, the Pearl Freedom Bond (With Profits) policy (the “Freedom Bond”), which was sold to policyholders between 1986 and 1992. It promises three minimum benefits:

No representative policyholder is party to the proceedings because, it is said, the question needs to be resolved urgently. The Financial Services Authority (“FSA”) has therefore agreed to be the defendant and has, through Mr Tom Weitzman QC and Mr Richard Brent, sought to present all arguments that could properly have been advanced by policyholders. I am grateful to them: to some extent they took on the role of amicus curiae and it should not be assumed that all their submissions necessarily presented a position that the FSA would have adopted in other circumstances. I shall later have to consider whether I should in these circumstances make a declaration about the meaning of the Freedom Bonds, but it was agreed by Mr Weitzman and by Lord Grabiner QC, who with Ms Nehali Shah represented Phoenix, that I should decide the proper meaning of the Freedom Bond, and consider separately what, if any, relief should be ordered. One consideration then will be whether an exemplar Freedom Bond, the holder of which was referred to during the hearing as “Mr Y”, is typical in material respects of the Freedom Bonds issued by Phoenix or at least is sufficiently representative for the court properly to make a declaration, albeit possibly with some qualifications.

There is no significant factual dispute between Phoenix and the FSA. The evidence directed to the proper construction of the Freedom Bond was given in witness statements of Mr Michael Merrick, who has been the Chief Executive Officer of Phoenix since September 2009, having previously been the chief actuary to the Phoenix Group, and of Mr Mark Threipland, a solicitor and head of the relevant department of the FSA. Their evidence was not contentious (except for an unimportant complaint that Mr Threipland gave some hearsay evidence without stating its source), and they were not cross-examined.

The Freedom Bond is, as I have said, a “with profits” policy, that is to say a policy under which the policyholder is entitled to participate in profits earned by Phoenix on the relevant fund, the PLAL Pearl With-Profits Fund (the “Pearl Fund”)”. (Phoenix used to be called Pearl Assurance Plc, before it re-registered as a limited company on 25 June 2010 and changed its name on 20 September 2012. I shall refer to “Phoenix” in this judgment, sometimes anachronistically.) Over one million policyholders have an interest in the Pearl Fund, and some 22,000 have Freedom Bonds: about half the Freedom Bond holders have reached retirement age and half have not. (Originally about 27,000 Freedom Bonds were issued, but some 5,000 have either died before retirement or transferred their benefits to other insurance companies.) As is usual, Phoenix gives effect to policyholders’ entitlement to participate in the profits of the Pearl Fund by declaring bonuses, and Lord Steyn’s description of bonuses in Equitable Life v Hyman, [2002] 1 AC 408, 453C is equally apt here:

“… Two types of bonus are involved. First, there are annual or declared bonuses (sometimes called reversionary bonuses) .... They are allotted irrevocably to policies when declared, whereupon they constitute a vested legal entitlement of the policyholder. Such bonuses are reversionary in the sense that the benefit of them is enjoyed at a future date when the policy matures. Secondly, there are final bonuses. These bonuses are sometimes called terminal bonuses. They are not allotted when announced, but only vest on a policy’s maturity.”

As Lord Cooke said in that case (loc cit at p.461H), “although discretionary and uncertain, bonuses are a very significant part of the benefits which policyholders expect”.

Until it was replaced by the State Second Pension (or “S2P”), the State Earnings Related Pension Scheme (or “SERPS”) was designed to provide employees with a pension related to earnings (in addition to the basic state pension). Members of occupational pension schemes could be “contracted out” of SERPS by their employer, in which case both employer and employee paid lower National Insurance contributions, but in order to be “contracted out” the occupational pension scheme had to guarantee a minimum level of pension (a guaranteed minimum pension or “GMP”). Section 32 of the Finance Act, 1981, by amending section 20 of the Finance Act, 1970, allowed the Board of the Inland Revenue to approve a retirement benefits scheme that provided benefits “by means of an annuity contract with an insurance company of the employee’s choice” so as to afford tax reliefs under Chapter II of the 1970 Act that had previously been available only to occupational pension schemes. The Social Security Act, 1985, by amending the Social Security Pensions Act, 1975, extended this regime, and enabled all members of retirement benefit schemes who left them on or after 1 January 1986 and were entitled to preserved benefits to have their rights secured under a policy with an insurance company and require that their scheme pay a “cash equivalent” (or transfer value) to a company scheme or arrangement that met specified conditions. But where the employer’s pension scheme was contracted out of SERPS, annuities provided under the new policy had to produce at least the applicable GMP (which, of course, was determined not by the rules of the fund of the transferee insurance company but by the employer’s pension scheme). By section 52C(1) of the 1975 Act a transaction to which the section applies discharges the trustees or managers of an occupational pension scheme from their liability to provide for certain specified benefits where the specified conditions are satisfied.

I explain this because an instrument such as the Freedom Bond must be construed in its statutory context: “It has been said on more than one occasion that many provisions in pension schemes and insurance contracts have to be construed against their fiscal backgrounds”: National Grid Plc v Mayes, [2001] UKHL 20, para 18 per Lord Hoffmann. Sometimes courts have apparently treated this as an aspect of the more general principle of contractual interpretation, that evidence of the factual background is admissible provided that, when they entered into the contract, both or all parties knew it or are taken to have known it because information was reasonably available to them; and for this reason have not attributed to the parties detailed knowledge of the statutory background to a contract. For example, in Zoan v Rouamba, [2002] 2 All ER 620 at para 38 the Court of Appeal found it “impossible to attribute to someone in the position of [the hirer of a motor vehicle] at the time that the agreements were made the background knowledge” of consumer credit legislation. In Prenn v Simmonds, [1971] 1 WLR 1381 Lord Wilberforce (at p. 1388C) spoke of interpreting a share acquisition agreementon the basis that the parties, as businessmen, must have know about the requirements of the Companies Act, 1948 for accounts to be placed before shareholders in a general meeting “at least in general terms”. But this case does not depend upon how much detail of the relevant law is attributed to contracting parties: the minutiae of the fiscal regime behind instruments such as the Freedom Bond are not important, and Lord Grabiner relied only upon the general statutory background that I have described and in particular the requirement that policies issued in lieu of employees’ rights under contracted-out occupational pension schemes should provide the benefit of (at least) an equivalent GMP and widow’s (or, later, spouse’s) GMP. As Mr Weitzman pointed out, Phoenix provided brochures to potential policyholders that gave them much of that information.

Moreover, before issuing a Freedom Bond Phoenix required a proposal form be completed by the prospective policyholder, referred to as the “Annuitant”, and by the trustees or manager or administrator of the occupational pension scheme from which the transfer was to be made, referred to as the “Purchaser”. That form referred to the fiscal background, stating that the proposal was for “a contract approved under Section 32 of the Finance Act 1981”. As I shall explain, the proposal form was agreed to be “the basis of the contract”, and so it was the agreed basis of the contract that it was approved under section 32.

The proposal form also required that details of the proposed Freedom Bond be given, including the “Nominal Capital Sum”, the “Single Premium (Transfer Value)”, whether the bond was to be “With profits or “Without profits”, and the Pension Date and which birthday of the Annuitant that would be. In the exemplar case of Mr Y, the NCS was £34,634, the premium was £11,764, a box was ticked to indicate that the Freedom Bond was to be with profits, and the pension date was stated to be 5 December 2011, Mr Y’s sixty-fifth birthday.

Part A of the proposal form was to be completed by the Annuitant with various personal details, and having completed part A he (or she) was to sign a declaration in the following terms:

“I, the undersigned, on whose life the payment of the annuity is to depend, hereby declare that:

(1)

[Certain statements and particulars] are true and complete, and I agree that these statements and particulars shall be the basis of the proposed contract between me and the Company.

(2)

I have compared the benefits with those under the scheme from which the transfer value arises and I agree to take a Policy in the form ordinarily used by the Company.”

Part B of the form was to be completed by the Purchaser (in Mr Y’s case, the London Transport Pension Fund), who was to give details of the scheme from which a transfer was being made into the Freedom Bond. It included a statement whether the scheme was contracted out of SERPS and if so what the GMP was on the “date of cessation of contracted-out employment” and the rate at which it was to be revalued to the State Pension Age. Under part B there was a declaration to be signed by the Purchaser:

“We, the undersigned, hereby make application to [Phoenix] for a Freedom Bond in accordance with this proposal and we declare that: …. (5) the above statements and particulars are, to the best of our knowledge, true and complete, and we agree that this proposal and declaration, together with the information supplied for the quotation, shall be the basis of the Policy to be issued to the Annuitant”.

In the following operative provision, the Freedom Bond stated (inter alia) that the Proposal was the basis for the contract:

“The Purchaser and the Annuitant … have made a proposal to [Phoenix] for a Pearl Assurance Freedom Bond on the life of the Annuitant containing statements and particulars which together with any supplementary proposal or memorandum given by the Purchaser and by the Annuitant are hereby admitted to be the basis of the contract.

“This Policy which consists of the respective pages stated in the Schedule witnesses that in consideration of the payment to [Phoenix] by the Purchaser of the premium stated in the Schedule on the specified date and subject to the Conditions of the Policy the Company will on survival by the Annuitant to the Pension Date pay the annuity or annuities as described to the person or persons to whom the annuity is expressed to be payable.

“Provided Always that if the death of the Annuitant occurs before the Pension Date (and before any payment is made under this Policy) the death benefit as described in the Schedule will be paid to the Executors or Administrators of the Annuitant in full discharge and satisfaction of all claims under this Policy.”

As the operative provision indicates, the Freedom Bond also had a Schedule (which identified the pages comprising the contract) and Conditions. I shall set out most of the terms in the Schedule to Mr Y’s Freedom Bond. Each of its three pages described the “Type of Policy” as a Pearl Assurance Freedom Bond and stated that it was “issued in lieu of benefits provided by London Transport Pension Fund for the Annuitant and approved by the Board of the Inland Revenue under … the Finance Act 1970” and related to service between specified dates. Other provisions of the Schedule are these:

There were fourteen terms set out in the Conditions, all but one of which were identified for inclusion in Mr Y’s policy. Important for present purposes is condition 8, which is headed “Preservation of Guaranteed Minimum Pension and Spouse’s Guaranteed Minimum Pension”:

“At the date the annuity commences the amount of the annuity payable to the Annuitant shall be not less than the Guaranteed Minimum Pension revalued to State Pension Age stated in the Schedule and the annuity payable to the spouse of the Annuitant shall not be less than the Spouse’s Guaranteed Minimum Pension revalued to State Pension Age (if any) stated in the Schedule. Any option or election under Conditions 2, 3, 5 and 6 as is applicable to this Policy may only be exercised to the extent that there shall remain an annuity payable to the Annuitant of not less than the Guaranteed Minimum Pension revalued to State Pension Age and an annuity payable to the widow of the Annuitant of not less than the Spouse’s Guaranteed Minimum Pension revalued to State Pension Age (if any).”

As condition 8 indicates, the Conditions allow the Annuitant various options subject to the stipulated restrictions, namely:

I should also refer to conditions 10 and 11, which give the policyholder other rights. Condition 10 is headed “Open Market Option”, and it provides that, as an alternative to an annuity at the Pension Date:

“… the Policy proceeds (with the exception of any amount required to provide a Guaranteed Minimum Pension or Spouse’s Guaranteed Minimum Pension) may be used at the time the annuity commences to purchase an annuity from any reputable assurance company or society.”

Condition 11 is about transferring the benefit of the Freedom Bond to a retirement benefits scheme of a subsequent employer, and it reads as follows:

“In the event of the Annuitant being admitted to membership of a retirement benefits scheme (hereinafter referred to as “the receiving scheme”) and which [Phoenix] shall have ascertained is a scheme approved under [the relevant legislation] or any other fund scheme or arrangement approved for the purposes of this Condition by the Board of Inland Revenue, [Phoenix] may agree, on written application by the Annuitant, to pay the surrender value of this Policy (with the exception of any amount required to provide any Guaranteed Minimum Pension or Spouse’s Guaranteed Minimum Pension to be preserved in accordance with [statute]) directly to the trustees or administrator of the receiving scheme, in consideration for which the Annuitant will be granted transfer credits allowed under the rules of the receiving scheme. ...”

Finally, condition 12 provides that, “The amount of the annuity payable at the Pension Date shall not exceed the amount of the Maximum Pension stated in the Schedule”.

I have to determine the proper meaning of the NCS provision, the provision in Mr Y’s Schedule: “Nominal Capital Sum: £34634. (The Company guarantee this amount to be sufficient to cover any Guaranteed Minimum Pension for the Annuitant and his widow as described below)”. The FSA says that “this amount” refers to £34,634 and that the guarantee given by Phoenix is that the £34,634 will provide any GMPs. Mr Weitzman emphasised that this would not mean that the Freedom Bond disregards the value to Phoenix of receiving consideration some years before paying any annuity: in Mr Y’s case, the NCS of £34,634 was some three times the premium. Mr Weitzman submitted that this takes account of its time value to Phoenix.

Phoenix submits that it is wrong to focus only on the NCS provision: that the proper approach to an issue of contractual interpretation of this kind is not to examine the wording of a clause or provision in isolation (as if it were “an island”, to echo Bingham LJ in Atkins International HA v Islamic Republic of Iran Shipping Lines (The “APJ Priti”), [1987] 2 Ll L R 37, 41), but to consider “every part” of [an instrument] “in order to collect from the whole one uniform and constant sense, if it may be done” (per Lord Ellenborough CJ in Barton v Fitzgerald, (1812) 15 East 531, 541; 104 ER 944, 948): in other words, to adopt what has been called a “holistic” or sometimes an “iterative” approach, the latter expression referring to a process of “checking each of the rival meanings against the other provisions of the document and investigating its commercial consequences”: Re Sigma Finance Corp. [2008] EWCA Civ 1303 at para 98 per Lord Neuberger MR, whose dissenting judgment was upheld and approach to contractual construction endorsed in the Supreme Court, [2009] UKSC 2 at para 12. I accept this submission, but find it convenient, simply for the purposes of exposition, to start with the language of the NCS provision in Mr Y’s bond and to consider two questions:

Lord Grabiner submitted that “this amount” includes any allotted bonuses. The NCS provision, he said, is indeed about the NCS, and this observation introduces two arguments: first, that the very expression reflects that the sum is a “nominal” one, and connotes that it exists in name rather than in reality and is susceptible to change through additions or deductions. (He cited the Shorter Oxford English Dictionary, 6^th Ed, 2007, in support of this, drawing attention to one of the definitions of “nominal”: “Existing in name only, not real or actual; merely named, stated or expressed, without reference to reality or fact; minimal in relation to the true value, token; so small or insignificant as hardly to justify the name”.) Secondly, Lord Grabiner drew attention to other uses in the Freedom Bond of the term “Nominal Capital Sum”, arguing that an interpretation of the Freedom Bond that gives it a consistent use and application is to be preferred. In the Schedule, the With Profits provision states that bonuses whereby the bond participates in the profits are allocated “as additions to” the NCS. The Amount of Annuity provision uses the term twice, and on both occasions it must mean the enhanced pot, which is to be converted into an annuity by application of the MAR or some higher rate. As for condition 6 Lord Grabiner contended that the parties’ natural expectation would have been that on early retirement the policyholder could use the bonuses as well as the initial NCS for an adjusted annuity (subject to the preservation of the GMPs as required by condition 8). “Nominal Capital Sum” is also found in the proposal form, the agreed “basis of the contract”, but its use here does not, to my mind, assist to resolve this case.

I am not greatly impressed by the first of Lord Grabiner’s submissions. Of course, the policyholder’s interest in the fund or his pot is not fixed: that is the very nature of a with-profits policy, and that might be why the expression nominal capital sum is used. But this does not engage with the question whether the GMP guarantee is given by reference to the initial pot or the enhanced pot. As Mr Weitzman rightly emphasised, the GMP guarantee in the NCS provision does not refer to the NCS but to “this amount”.

The same criticism might be made of Lord Grabiner’s argument about the use of the “Nominal Capital Sum” elsewhere in the Freedom Bond, but here I consider his argument has more force. Lord Grabiner is right that elsewhere the bond provides that annuity rights are determined by reference to the enhanced pot. I agree with Mr Weitzman that the With Profits provision does not much assist Phoenix’s argument: it could mean either that allotted bonuses are added to and become part of the NCS or that they make a contribution to the pot that is additional to the NCS. But other uses of the term cannot be dismissed so readily.

The FSA accepts that in the Amount of Annuity provision “Nominal Capital Sum” refers to the enhanced pot. This must be so: otherwise, as Mr Weitzman put it, the bonuses would be left in limbo at maturity date. He submitted, however, that the “inconsistency” between the use of the expression in the NCS provision of the Schedule and the (immediately following) Amount of Annuity provision should be “resolved” by recognising that in the latter the expression is misused, and this should not colour the interpretation of other provisions. This is not the approach to contractual construction advocated by judges since Lord Ellenborough CJ (and no doubt before): to my mind this argument falls into the trap of interpreting provisions in isolation and then solving the problems so created as best one can.

Mr Weitzman made two submissions about condition 6. First, he argued that it is ambiguous as to whether the NCS refers to the initial pot or the enhanced pot, and therefore does not assist Phoenix’s contention that consistency demands that in the NCS provision it refers to the enhanced pot. He contended that by exercising the option available in condition 6 the policyholder agrees to take a reduced annuity and there is no reason that the reduced annuity should not be determined by converting the initial pot at the appropriate rate. I agree that conceptually the reduced annuity might be so determined: as a formula it could work. But the consequences would be remarkable and cannot have been intended by the parties: for example, (as Mr Weitzman acknowledged) if a policyholder were forced to retire early by incapacity (a circumstance expressly contemplated by the condition) and his circumstances required him to use the Freedom Bond to take an early annuity, he would be stripped of allotted bonuses. I see no real uncertainty about the meaning of NCS in condition 6. When interpreting contracts, the court takes account of what would make business sense not only to dispose of impossible interpretations and those which flout common sense, but as providing guidance as to the parties’ intentions more generally: see Barclays Bank Plc v HHY Luxembourg SARL, [2010] EWCA Civ 1248 at para 26, endorsed by the Supreme Court in Rainy Sky SA and ors v Kookmin Bank, [2011] UKSC 50.

Mr Weitzman also submitted that the contract should be interpreted on the basis that “in cases of conflict or inconsistency between the two, it is the Policy Schedule which prevails over the Policy wording. This reflects the fact that the Policy Schedule is specific to the Policyholder and accords with the normal principles of construction applicable to insurance contracts”. He cited the speech of Lord Bingham in the Hombuirg Houtimport BV and ors v Agrosin Private Ltd and anor (The “Starsin”), [2003] UKHL 12 at para 11. However, what Lord Bingham said does not really support Mr Weitzman’s submission but, on the contrary, draws attention to its limitation: he said, “…it is common sense that greater weight should attach to terms which the particular contracting parties have chosen to include in the contract than to pre-printed terms probably devised to cover very many situations to which the particular contracting parties have never addressed their minds” (emphasis added). Here the contract itself chose the conditions applicable to the Freedom Bond of the particular policy-holder: in the case of Mr Y thirteen were selected. Admittedly they are “pre-printed terms” and therefore excessive weight should not be given to their exact wording when interpreting its application in a particular contract. But it detracts from Mr Weitzman’s point that the operative conditions were chosen in the Schedule.

Lord Grabiner has further arguments based on the conditions: first, conditions 10, 11 and 12. The “policy proceeds” in condition 10 are the enhanced pot, not the initial pot. The “surrender value” in condition 11 also must be determined by reference to the enhanced pot. The natural connotation of both conditions, to my mind, is that the GMP, to which both conditions in parenthesis refer, is to be produced from that same enhanced pot. Neither condition directly contradicts the FSA’s argument, but in both cases the language seems to me more harmonious with Phoenix’s interpretation.

Phoenix submits that, if the “amount of annuity payable” were calculated by reference only to the initial pot (in Mr Y’s case, £34,634), it would be extremely unlikely to exceed that Maximum Pension in the Schedule and in practice condition 12 has no scope. That point does not impress me: maybe condition 12 does involve what Hoffmann J in Tea Trade Properties Ltd v CIN Properties Ltd, [1990] 1 EGLR 155 called “linguistic overkill” and sometimes contracts are construed on the basis that no provision should be discarded as surplusage, but recently, I think, the courts have more readily recognised that commercial documents sometimes do include redundant provisions: see Lewison on the Interpretation of Contracts (5^th Ed, 2011) para 7.03 and the cases cited there.

However, Phoenix has another point on condition 12. The condition is concerned with the amount of the annuity that the policyholder is to receive from his enhanced pot, not from his initial pot, and its purpose is to limit that to the Maximum Pension. If Phoenix’s interpretation of the NCS provision is adopted, this would require a parallel comparison between the annuity provided by the enhanced pot and the GMP, whereas on the FSA’s interpretation a different comparison would be required: a comparison between a part of the annuity, that from the initial pot, and the GMP. To that extent, Phoenix’s interpretation again produces the more harmonious interpretation of the bond as a whole.

Phoenix’s interpretation of the NCS provision is further corroborated by condition 8. This provides that the “amount of the annuity payable to the Annuitant”, that is to say the value of the fund converted at the MAR or some higher rate, is to be not less than the GMP, and that the amount of the annuity is determined by reference to the enhanced pot, including the initial NCS (but not confined to it, assuming that at least one bonus has been allocated to the bond). It would be strange for condition 8 and the NCS provision to provide different guarantees: as Lord Grabiner submitted, the (more modest) guarantee in condition 8 would then be redundant. It might be suggested that the more modest guarantee is given in condition 8 only in order to define the limits of the options in conditions 2, 3, 5 and 6, but that suggestion itself underlines another problem unless the NCS provision and condition 8 provide for the same guarantee: it creates uncertainly about the limits of these options, and the parties cannot have intended that. Phoenix’s interpretation means that condition 8 would be another case of “linguistic overkill” but, as I have explained, that does not seem to me important. In any case it is natural to reiterate the NCS provision in the context of the options, and the purpose is reflected in the heading to condition 8: the condition’s focus is not to provide for the GMPs, viz to introduce them into the contract, but to preserve the GMPs for the annuitant and his spouse promised in the Schedule.

Mr Weitzman disputes this reasoning: he argues that the first sentence of condition 8 is consistent with either interpretation of the NCS provision: it sets out a general promise that the annuities will be at least as much as the GMPs, but does not qualify the NCS provision which spells out how this general undertaking will be implemented. There is, he argued, no justification for “any such re-writing of the contract”. Again, to my mind this argument defies the principle of interpreting the contract as a whole. I do not say that it is impossible to reconcile the two provisions as Mr Weitzman suggests: it is possible. But each provision should inform the proper interpretation of the other, and the parties’ intention is ascertained by seeking from the start a harmonious interpretation of the contract as a whole, rather than managing to reconcile discordant interpretations of different provisions.

Returning to the NCS provision itself, for these reasons I would accept Lord Grabiner’s submission that the expression “this amount” should be interpreted to mean the enhanced pot if this were necessary in order to uphold Phoenix’s contention about what GMP guarantee they gave. But there is another route to this result which I myself find more attractive. In saying this, I acknowledge immediately that a debate of this kind about how semantically a provision comes to have a particular meaning is at best arid and might be criticised as failing to recognise that language conveys more meaning through sentences (akin to what Robert Frost called “sentence sounds”) than through isolated words or phrases. That said, to my mind “this amount” can properly be analysed as referring to the initial NCS (Mr Y’s £34,634) and the meaning of “sufficient to cover any Guaranteed Minimum Pension for the Annuitant and his widow” examined on this basis.

I do not consider that this phrase means that Phoenix guarantees that the sum of £34,634 if simply set aside until the Pension Date will prove to be enough to buy an annuity of at least the GMPs from that very sum. It does not convey that as a matter of ordinary language, especially given that the Freedom Bonds would have been issued some considerable time before the Pension Date (in Mr Y’s case, 23 years). If a benefactor is invited to endow a professorial chair, he might be told what sum will be “sufficient to cover” it: that does not mean that that cash sum will be paid out over the years and the money will run out so far into the future that it does not matter. It means that that sum will produce enough funds to defray the expenses associated with the chair, given that the endowment will be invested, and perhaps that other funds will be attracted once it is established. Of course, Mr Weitzman is right that in this case the difference between the premium and the initial NCS takes some account of the income that the premium sum is anticipated to produce, but that does not alter my understanding of the NCS provision. By it Phoenix promised that, if the initial NCS reflected the only profits from the premium and if, given that, the rate of annuity was not sufficient to produce the GMP, then it would make up the shortfall.

I observe in passing that, consistently with this understanding of the NCS provision, its wording is to be contrasted with that of conditions 10 and 11, which refer to an amount of “Policy proceeds” that is “required to provide” (not to “cover”) any GMP or spouse’s GMP. The word “provide” reflects the more immediate correspondence between the “Policy proceeds” and the amount of the annuities: that conditions 10 and 11 are about what will be provided by the enhanced pot, not the initial pot, and the guarantee is that the enhanced pot will directly provide an annuity of at least the GMP. My preference for Phoenix’s interpretation does not depend on this observation: that would place unrealistic weight on the difference in language in different parts of the Freedom Bond; but I think that it provides some support for it.

Whichever route is taken to Phoenix’s interpretation of the NCS provision, it is also supported by the words “as described below”. Mr Weitzman submitted that this refers only to the amounts of the GMP and the widow’s GMP found later in the Schedule. But those bald statements of amounts are not, to my mind, a “description”: had they been the intended referent the NCS provision would more naturally have read “as stated below”. I find it easier to understand the words “as described below” as referring, as Lord Grabiner submitted, more generally to the following provisions of the Schedule, including the With Profits provision: I take it to be an adverbial phrase modifying how the amount will prove sufficient rather than as an adjectival phrase qualifying “any Guaranteed Minimum Pension for the Annuitant and his widow”.

I therefore uphold Phoenix’s interpretation of the Freedom Bonds simply as a matter of their natural and ordinary meaning. As Lord Grabiner observed, it avoids the somewhat surprising implication of the FSA’s interpretation that the GMP guarantee might work as a disincentive to Phoenix declaring bonuses on the Pearl Fund. More importantly, in my judgment, it also is in accordance with the statutory context of the Freedom Bonds. The reference to GMPs naturally connotes that the bonds met the requirement of ensuring that the policyholders’ rights were no less than if they received pensions through SERPS: precisely that is achieved by Phoenix’s interpretation. If the parties had intended that Phoenix should give a different and more generous guarantee (because the minimum level of pension would be met from only part of the policyholder’s entitlement), then this would, I think, have been clear from the Freedom Bonds: it was not. More specifically, the GMP guarantee would not have been given in a parenthesis in the NCS provision and a condition expressly directed, as I have said, to preserving the GMPs from being undermined by the options. Furthermore, by the time that Phoenix issued Freedom Bonds, the expression “Guaranteed Minimum Pension” had an established statutory meaning (deriving from the Social Security Pensions Act, 1975, where it was defined in section 26) that was associated with what retired people would receive from their relevant pension provision as a whole. I do not think that that same statutory expression would have been used (especially with capital letters that consistently signify in the Freedom Bond the use of a term in a technical or at least specific sense) if the GMP guarantee was to be met from only the initial pot and not from the total benefits under the bond. It would then have been more natural for the Freedom Bond to have used an expression that did not have the (misleading) associations with the statutory language (and probably to have confirmed separately that the statutory requirement was necessarily therefore met).

Before I come to other arguments that Lord Grabiner advances to bolster this conclusion, I should deal with three further points of Mr Weitzman. First, I do not overlook that the profits added to the NCS are referred to as “bonuses” (in the With Profits provision, reflecting the normal language of with profits policies). In some contexts the word “bonus” connotes a benefit, typically perhaps a monetary benefit, which is additional to what is expected. If Phoenix’s interpretation is adopted, a bonus added to the NCS might not result in any additional benefit for the policyholder: he might simply receive by way of a bonus what would in any case have come his way through the GMPs. I do not think this a weighty consideration: in the context of a with profits policy the expression “bonus” has an established meaning and, as Lord Steyn said in the Equitable Life case (loc cit at p.359E) of final bonuses, they are not paid as “bounty”, but are a recognised and significant part of the consideration for the premium(s) paid.

Secondly, Mr Weitzman submitted that part of the context in which the Freedom Bond is to be construed is that policyholders would generally not have financial advisers and would themselves “normally be financially unsophisticated”. I am prepared to suppose that many policyholders would have met this description, and since it can hardly be supposed that Freedom Bonds have a different meaning depending upon the financial sophistication (or the financial sophistication known to Phoenix) of the particular policyholder, I also accept that they are to be interpreted on the basis that neither financial advice nor financial sophistication is needed properly to understand them. But this point cannot be taken too far: I certainly distinguish between an investor being unsophisticated and him being obtuse or having no curiosity about his financial affairs. After all, the Freedom Bond is clearly described as a with-profits policy and I do not accept that it is to be interpreted on the basis that policyholders neither knew nor were bothered to enquire what such a policy is and how such policies generally operate, nor that Phoenix knew this.

Thirdly, it was rightly observed that in so far as there is scope for interpreting the Freedom Bond contra proferentem it is to be interpreted against Phoenix’s interest. Phoenix prepared the wording of the Freedom Bond as a whole, and the controversial wording in particular. But the principle of interpretation contra proferentem does not find uncertainty where none truly exists, nor is it used whenever there is some uncertainty about a contract’s meaning, but has been described as a “rule of last resort”: Mance LJ in Sinchem International Oil (London) Co Ltd v Mobil Sales & Supply Corp, [2000] 1 Lloyd’s L R 339, para 27. That is not to say that the principle is never determinative: see, for example, Billyack v Leyland Construction Co Ltd, [1968] 1 WLR 471. But I see no need to turn to it here.

Although I do not think that Phoenix needs to rely on them, I should say something about two other arguments that it deployed. First, Phoenix wants to rely upon so-called “sales documents” to explain the context in which the Freedom Bond was issued and so to inform its interpretation. It is not said, as for example in R v Clowes (no 2), [1994] All ER 316, that Phoenix’s brochures were themselves contractual documents, nor that they gave rise to an estoppel by convention or otherwise restrict what interpretation of the contract could be asserted by any party. Phoenix put in evidence two brochures that they used to market Freedom Bonds, and Mr Merrick’s evidence was that they are the only relevant brochures that Phoenix has been able to find and that they are believed to be representative of its sales literature. Neither is dated but both were clearly issued after 1 January 1986 and the later version reflects changes that had been introduced with effect from 6 April 1988 whereby widowers as well as widows were given the protection of a GMP. Both versions have eight pages, including the cover page. Much of their contents is similar: they provided an explanation that those who had left an employment with a pension scheme could transfer their accrued pension benefits and could elect to transfer them to a Freedom Bond instead of to the scheme of a new employer; and they included a “glossary of terms”. They were directed to policyholders themselves rather than to brokers or financial advisers: this is clear from their general tone and from such statements as “Pearl can arrange the transfer for you. Just a few moments of your time in the comfort of your own home is all we require.”

I shall set out the parts of the documents on which Phoenix particularly relied.

The significance that Phoenix attaches to this material is not that it demonstrates that the contracting parties knew (or are to be taken to have known) facts when they entered into Freedom Bonds: to that extent it might not be covered by Lord Wilberforce’s expression “matrix of facts” in Prenn v Simmonds, [1971] 1 WLR 1381, 1384A. That does not matter: see BCCI v Ali, [2001] UKHL 8 para 37 per Lord Hoffmann. As he had said in ICS v West Bromwich Building Society, [1998] 1 WLR 896, 912H, “Subject to the requirement that it should be reasonably available to the parties” contractual interpretation is properly informed by “absolutely anything which might have affected the way in which the language of the document would have been understood by a reasonable man”, subject to the exceptions of evidence of the parties’ subjective intentions or understanding and evidence of negotiations. I accept that the material upon which Phoenix seeks to rely is covered by neither exception.

It was therefore common ground between Phoenix and the FSA that standard promotional and sales material provided by Phoenix to policyholders should be considered when interpreting the Freedom Bond. However, I am not persuaded that it properly assists in this case. First, to be admissible evidence the material must have been reasonably available to all the contracting parties. I consider that at least one view of the arrangements whereby Freedom Bonds were issued is that there was a tripartite contract between Phoenix, the policyholder (or Annuitant) and the Purchaser. The proposal form, which is the “basis of contract”, put forward a proposal by both policyholder and Purchaser and the purchaser paid the consideration, the premium. I do not say that this is the only possible analysis of the arrangements: I recognise that the proposal form refers to a contract between the “Annuitant” and Phoenix and that section 32 of the Finance Act, 1980 refers to a contract between an employee and an insurance company, but had my decision depended on this point (which arose only during oral exchanges at trial) I should have sought further assistance from counsel about it. This leads to the question whether, and if so what, material was available to the Purchasers of Freedom Bonds. The proper inference from the evidence might be that they were provided with the same brochures as policyholders: Mr Merrick exhibited to one of his witness statements a form apparently sent to Purchasers that refers to “A brochure on Pearl’s ‘Freedom bond’” being enclosed. But the point is not covered in the witness statements, and the position is not as clear as would have been desirable in a case of this kind.

There seems to me to be another difficulty in giving much weight to the sales documents. The nature of some contracts is such that the factual background known to the parties is not very significant: in Re Sigma Finance Corp, [2009] UKSC 2 the Supreme Court considered the interpretation of a security trust deed securing various creditors who held different instruments issued at different times and in different circumstances. Lord Collins said (at para 37):

“Consequently this is not the type of case where the background or matrix of fact is or ought to be relevant, except in the most generalised way. I do not consider, therefore, that there is much assistance to be derived from the principles of interpretation re-stated by Lord Hoffmann in the familiar passage in [the ICS case]. Where a security document secures a number of creditors who have advanced funds over a long period it would be quite wrong to take account of circumstances which are not known to all of them. In this type of case it is the wording of the instrument which is paramount. The instrument must be interpreted as a whole in the light of the commercial intention which may be inferred from the face of the instrument and from the nature of the debtor’s business. Detailed semantic analysis must give way to business common sense …”

The position here is in some ways analogous. Of course, all the policyholders had their own contracts with Phoenix, but (i) it will have been obvious to them that their bond had largely standard wording (subject to the individual details of the Schedule, including the selection of operative conditions), but (ii) they will not have known whether other policyholders had and were to have the same sales documents as them. They might rationally and properly have thought that they had a standard product and so what mattered is the words of the product itself, rather than the particular literature sent to them. In other words, to my mind some care is needed before interpreting what is obviously standard wording by reference to the personal circumstances of an individual consumer, such as the sales documents which, as far as any individual consumer knew, Phoenix might have provided to only some policyholders. This point is not met by evidence that all policyholders received materially similar sales documents because policyholders would not have known that.

Lord Grabiner pointed out that, had the sales documents undermined Phoenix’s interpretation of the Freedom Bond, policyholders would no doubt seek to rely upon them to assert claims. I dare say that he is right about that, but in appropriate cases they could invoke other legal principles, such as an estoppel by convention. Indeed, this is how it is put by Ivamy, General Principles of Insurance Law (6^th edition 1993) p.244:

“Where there is an ambiguity on the face of the policy, and a question, therefore, arises as to its meaning or effect, the Court may take into consideration any documents, such as the prospectus, the proposal form, a letter which the insurers have written, or even the back of the policy, if not incorporated into the policy, in which the insurers profess to set out or explain the purport and effect of their policies, and any verbal explanations given by themselves or their agents inconsistent with the contract contained in the policy. The documents and the explanations show the interpretation which the insurers themselves place on the policy, and they are, therefore, precluded by their own words from relying on any other interpretations.

The onus of providing that he knew and acted on such an interpretation rests on the assured, though especially in a case where prospectuses are issued, the onus is not difficult to discharge.”

However I need not decide these questions: first I uphold Phoenix’s interpretation of the Freedom Bond without regard to the sales documents; and secondly I do not consider that the sales literature adds anything of real significance to what is in the contractual documents, interpreted by reference to indisputably legitimate aids, including the statutory background and, in broad terms, the conventional nature of with-profit policies. The FSA rightly does not suggest that the sales literature contains anything inconsistent with or contra-indicative of Phoenix’s interpretation of the Freedom Bond, and to my mind that is the highest that Phoenix can properly put its argument about sales documents.

Lord Grabiner also made submissions, based on the details of Mr Y’s bond, that the FSA’s interpretation makes no commercial sense or at least would have surprising commercial consequences. First he pointed out that if the sum of £34,634 were converted at the MAR of £76 per £1,000 it would produce an annuity of £2,632.18, a little over half of the GMP of £5,006.56. The £34,634 would have to be converted at £144.50 per £1,000 to produce an annuity to meet the GMP. According to Mr Merrick, in the case of 88% of the policies in force in mid-2010 the initial NCS, if converted to an annuity at the MAR, would produce smaller annuities than the GMPs. Lord Grabiner submitted that the MAR would have no real impact if the FSA’s interpretation of the Freedom Bond is adopted. Although of course a higher annuity rate than the MAR might have been adopted, it would on the face of it be surprising if the parties contemplated as high a rate as would have been required in Mr Y’s case: a rate of £144.5 per £1,000 would have been far higher than the prevailing rate for those retiring in 1989, which according to Mr Merrick was between £109 and £123 per £1,000.

I am not entirely convinced by this argument, and see some force in Mr Threipland’s characterisation of it as “(at least to some extent) an argument derived from hindsight”. I find it difficult to assess how optimistic actuarial assumptions were in the late 1980s. But I accept that this point lends some modest support to Phoenix’s argument.

Secondly, Lord Grabiner referred to a quotation provided to Mr Y apparently on the basis of information about his entitlement, including the transfer value of his benefits and his GMP, given to Phoenix by his former employer, London Transport, in a “Section 32 Freedom Bond Questionnaire”. The quotation was said to be consistent only with Phoenix’s contentions. The relevant part of the quotation is under the headings “Benefits” and “On survival to normal pension date”:

“The capital fund which provides the benefit is built up by means of a with profits pure endowment policy which matures at the Normal Pension Date.

The funds illustrated below includes the Nominal Capital Sum of £34634.00

The fund assumed available on the Normal Pension Date is

Lower basis of illustration £69900.00

Higher Basis of illustration £182000.00

The assumed yearly pension before tax the fund could provide assuming the pension is payable monthly in advance for life plus 50% widow’s reversionary pension to the wife at date of death is-

Lower basis of illustration £6860.00

Higher basis of illustration £20400.00

The guaranteed minimum annuity rate for each £1000 of fund is £76 p.a.”

Lord Grabiner pointed out by reference to the “lower basis” figures that, if (as the FSA submits) Mr Y’s GMP of £5,006 were met out of £34,634, the calculations suppose that the rest of the £69,900 (£35,266) would be used to buy only £1,854 (£6,860 less £5,006) of the projected annuity, which represents an annuity rate of only £53 per £1,000. Thus it is said that these figures can only have been produced on the basis that the annuity from the whole NCS including bonuses will count towards the GMP, and this is confirmed by the note on the quotation that, “The total annual pension payable from the State Pension Age is guaranteed to be not less than the Revalued Guaranteed Minimum Pension of £5006.56 p a” (emphasis added). Again, I accept that this point lends some support to Phoenix’s argument given the figures in Mr Y’s quotation, but I do not know whether his case is typical in this respect.

There was reference in Mr Merrick’s evidence to Phoenix’s internal documents, such as internal training briefings, and Phoenix’s post-contractual correspondence with Mr Y. They cannot affect the interpretation of the contract, but in any case nothing in them detracts from Phoenix’s case.

I uphold Phoenix’s interpretation of the Freedom Bond. I am only too conscious that, as Salmon LJ said in C W Dixie & Sons Ltd v Parsons, (1964) 192 EG 197, “No one is infallible [on the construction of documents], except the House of Lords [and now, presumably, the Supreme Court] and there were many points of construction upon which outstanding learned judges differed”. Nevertheless, to my mind there can be no real doubt about the proper interpretation of the Freedom Bond on the point in issue. It seems to me clear from the wording of the bond itself, given the nature of the instrument. It is reinforced by the statutory background. Phoenix does not need to rely upon the arguments about the detailed sums in Mr Y’s case and about the sales documents, but they do, I suppose, provide with some further support for this conclusion. I shall invite submissions about what (if any) relief I should order in light of this judgment.